1 \input texinfo @c -*-texinfo-*-
4 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
6 @c GNAT DOCUMENTATION o
10 @c Copyright (C) 1992-2014, Free Software Foundation, Inc. o
12 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
14 @setfilename gnat_ugn.info
17 Copyright @copyright{} 1995-2014 Free Software Foundation,
20 Permission is granted to copy, distribute and/or modify this document
21 under the terms of the GNU Free Documentation License, Version 1.3 or
22 any later version published by the Free Software Foundation; with no
23 Invariant Sections, with no Front-Cover Texts and with no Back-Cover
24 Texts. A copy of the license is included in the section entitled
25 ``GNU Free Documentation License''.
28 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
30 @c GNAT_UGN Style Guide
32 @c 1. Always put a @noindent on the line before the first paragraph
33 @c after any of these commands:
45 @c 2. DO NOT use @example. Use @smallexample instead.
46 @c a) DO NOT use highlighting commands (@b{}, @i{}) inside an @smallexample
47 @c context. These can interfere with the readability of the texi
48 @c source file. Instead, use one of the following annotated
49 @c @smallexample commands, and preprocess the texi file with the
50 @c ada2texi tool (which generates appropriate highlighting):
51 @c @smallexample @c ada
52 @c @smallexample @c adanocomment
53 @c @smallexample @c projectfile
54 @c b) The "@c ada" markup will result in boldface for reserved words
55 @c and italics for comments
56 @c c) The "@c adanocomment" markup will result only in boldface for
57 @c reserved words (comments are left alone)
58 @c d) The "@c projectfile" markup is like "@c ada" except that the set
59 @c of reserved words include the new reserved words for project files
61 @c 3. Each @chapter, @section, @subsection, @subsubsection, etc.
62 @c command must be preceded by two empty lines
64 @c 4. The @item command should be on a line of its own if it is in an
65 @c @itemize or @enumerate command.
67 @c 5. When talking about ALI files use "ALI" (all uppercase), not "Ali"
70 @c 6. DO NOT put trailing spaces at the end of a line. Such spaces will
71 @c cause the document build to fail.
73 @c 7. DO NOT use @cartouche for examples that are longer than around 10 lines.
74 @c This command inhibits page breaks, so long examples in a @cartouche can
75 @c lead to large, ugly patches of empty space on a page.
77 @c NOTE: This file should be submitted to xgnatugn with either the vms flag
78 @c or the unw flag set. The unw flag covers topics for both Unix and
81 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
84 @c This flag is used where the text refers to conditions that exist when the
85 @c text was entered into the document but which may change over time.
86 @c Update the setting for the flag, and (if necessary) the text surrounding,
87 @c the references to the flag, on future doc revisions:
88 @c search for @value{NOW}.
100 @set TITLESUFFIX for OpenVMS
105 @c The ARG is an optional argument. To be used for macro arguments in
106 @c their documentation (@defmac).
108 @r{[}@var{\varname\}@r{]}@c
110 @c Status as of November 2009:
111 @c Unfortunately texi2pdf and texi2html treat the trailing "@c"
112 @c differently, and faulty output is produced by one or the other
113 @c depending on whether the "@c" is present or absent.
114 @c As a result, the @ovar macro is not used, and all invocations
115 @c of the @ovar macro have been expanded inline.
118 @settitle @value{EDITION} User's Guide @value{TITLESUFFIX}
119 @dircategory GNU Ada tools
121 * @value{EDITION} User's Guide: (gnat_ugn). @value{PLATFORM}
124 @include gcc-common.texi
126 @setchapternewpage odd
131 @title @value{EDITION} User's Guide
135 @titlefont{@i{@value{PLATFORM}}}
141 @subtitle GNAT, The GNU Ada Development Environment
146 @vskip 0pt plus 1filll
153 @node Top, About This Guide, (dir), (dir)
154 @top @value{EDITION} User's Guide
157 @value{EDITION} User's Guide @value{PLATFORM}
160 GNAT, The GNU Ada Development Environment@*
161 GCC version @value{version-GCC}@*
168 * Getting Started with GNAT::
169 * The GNAT Compilation Model::
170 * Compiling with gcc::
171 * Binding with gnatbind::
172 * Linking with gnatlink::
173 * The GNAT Make Program gnatmake::
174 * Improving Performance::
175 * Renaming Files with gnatchop::
176 * Configuration Pragmas::
177 * Handling Arbitrary File Naming Conventions with gnatname::
178 * GNAT Project Manager::
179 * Tools Supporting Project Files::
180 * The Cross-Referencing Tools gnatxref and gnatfind::
182 * The GNAT Pretty-Printer gnatpp::
184 * The Ada-to-XML converter gnat2xml::
186 * The GNAT Metrics Tool gnatmetric::
188 * File Name Krunching with gnatkr::
189 * Preprocessing with gnatprep::
190 * The GNAT Library Browser gnatls::
191 * Cleaning Up with gnatclean::
193 * GNAT and Libraries::
194 * Using the GNU make Utility::
196 * Memory Management Issues::
197 * Stack Related Facilities::
199 * Verifying Properties with gnatcheck::
200 * Creating Sample Bodies with gnatstub::
201 * Creating Unit Tests with gnattest::
203 * Performing Dimensionality Analysis in GNAT::
204 * Generating Ada Bindings for C and C++ headers::
205 * Other Utility Programs::
207 * Code Coverage and Profiling::
209 * Running and Debugging Ada Programs::
211 * Compatibility with HP Ada::
213 * Platform-Specific Information for the Run-Time Libraries::
214 * Example of Binder Output File::
215 * Elaboration Order Handling in GNAT::
216 * Overflow Check Handling in GNAT::
217 * Conditional Compilation::
219 * Compatibility and Porting Guide::
220 * Microsoft Windows Topics::
222 * GNU Free Documentation License::
227 @node About This Guide
228 @unnumbered About This Guide
232 This guide describes the use of @value{EDITION},
233 a compiler and software development toolset for the full Ada
234 programming language, implemented on OpenVMS for HP's Alpha and
235 Integrity server (I64) platforms.
238 This guide describes the use of @value{EDITION},
239 a compiler and software development
240 toolset for the full Ada programming language.
242 It documents the features of the compiler and tools, and explains
243 how to use them to build Ada applications.
245 @value{EDITION} implements Ada 95, Ada 2005 and Ada 2012, and it may also be
246 invoked in Ada 83 compatibility mode.
247 By default, @value{EDITION} assumes Ada 2012, but you can override with a
248 compiler switch (@pxref{Compiling Different Versions of Ada})
249 to explicitly specify the language version.
250 Throughout this manual, references to ``Ada'' without a year suffix
251 apply to all Ada 95/2005/2012 versions of the language.
254 For ease of exposition, ``@value{EDITION}'' will be referred to simply as
255 ``GNAT'' in the remainder of this document.
260 * What This Guide Contains::
261 * What You Should Know before Reading This Guide::
262 * Related Information::
266 @node What This Guide Contains
267 @unnumberedsec What This Guide Contains
270 This guide contains the following chapters:
274 @ref{Getting Started with GNAT}, describes how to get started compiling
275 and running Ada programs with the GNAT Ada programming environment.
277 @ref{The GNAT Compilation Model}, describes the compilation model used
281 @ref{Compiling with gcc}, describes how to compile
282 Ada programs with @command{gcc}, the Ada compiler.
285 @ref{Binding with gnatbind}, describes how to
286 perform binding of Ada programs with @code{gnatbind}, the GNAT binding
290 @ref{Linking with gnatlink},
291 describes @command{gnatlink}, a
292 program that provides for linking using the GNAT run-time library to
293 construct a program. @command{gnatlink} can also incorporate foreign language
294 object units into the executable.
297 @ref{The GNAT Make Program gnatmake}, describes @command{gnatmake}, a
298 utility that automatically determines the set of sources
299 needed by an Ada compilation unit, and executes the necessary compilations
303 @ref{Improving Performance}, shows various techniques for making your
304 Ada program run faster or take less space and describes the effect of
305 the compiler's optimization switch.
308 the @command{gnatelim} tool and
310 unused subprogram/data elimination.
313 @ref{Renaming Files with gnatchop}, describes
314 @code{gnatchop}, a utility that allows you to preprocess a file that
315 contains Ada source code, and split it into one or more new files, one
316 for each compilation unit.
319 @ref{Configuration Pragmas}, describes the configuration pragmas
323 @ref{Handling Arbitrary File Naming Conventions with gnatname},
324 shows how to override the default GNAT file naming conventions,
325 either for an individual unit or globally.
328 @ref{GNAT Project Manager}, describes how to use project files
329 to organize large projects.
332 @ref{The Cross-Referencing Tools gnatxref and gnatfind}, discusses
333 @code{gnatxref} and @code{gnatfind}, two tools that provide an easy
334 way to navigate through sources.
338 @ref{The GNAT Pretty-Printer gnatpp}, shows how to produce a reformatted
339 version of an Ada source file with control over casing, indentation,
340 comment placement, and other elements of program presentation style.
346 @ref{The Ada-to-XML converter gnat2xml}, shows how to convert Ada
347 source code into XML.
353 @ref{The GNAT Metrics Tool gnatmetric}, shows how to compute various
354 metrics for an Ada source file, such as the number of types and subprograms,
355 and assorted complexity measures.
359 @ref{File Name Krunching with gnatkr}, describes the @code{gnatkr}
360 file name krunching utility, used to handle shortened
361 file names on operating systems with a limit on the length of names.
364 @ref{Preprocessing with gnatprep}, describes @code{gnatprep}, a
365 preprocessor utility that allows a single source file to be used to
366 generate multiple or parameterized source files by means of macro
370 @ref{The GNAT Library Browser gnatls}, describes @code{gnatls}, a
371 utility that displays information about compiled units, including dependences
372 on the corresponding sources files, and consistency of compilations.
375 @ref{Cleaning Up with gnatclean}, describes @code{gnatclean}, a utility
376 to delete files that are produced by the compiler, binder and linker.
380 @ref{GNAT and Libraries}, describes the process of creating and using
381 Libraries with GNAT. It also describes how to recompile the GNAT run-time
385 @ref{Using the GNU make Utility}, describes some techniques for using
386 the GNAT toolset in Makefiles.
390 @ref{Memory Management Issues}, describes some useful predefined storage pools
391 and in particular the GNAT Debug Pool facility, which helps detect incorrect
395 It also describes @command{gnatmem}, a utility that monitors dynamic
396 allocation and deallocation and helps detect ``memory leaks''.
401 @ref{Stack Related Facilities}, describes some useful tools associated with
402 stack checking and analysis.
406 @ref{Verifying Properties with gnatcheck}, discusses @code{gnatcheck},
407 a utility that checks Ada code against a set of rules.
410 @ref{Creating Sample Bodies with gnatstub}, discusses @code{gnatstub},
411 a utility that generates empty but compilable bodies for library units.
416 @ref{Creating Unit Tests with gnattest}, discusses @code{gnattest},
417 a utility that generates unit testing templates for library units.
421 @ref{Performing Dimensionality Analysis in GNAT}, describes the Ada 2012
422 facilities used in GNAT to declare dimensioned objects, and to verify that
423 uses of these objects are consistent with their given physical dimensions
424 (so that meters cannot be assigned to kilograms, and so on).
427 @ref{Generating Ada Bindings for C and C++ headers}, describes how to
428 generate automatically Ada bindings from C and C++ headers.
431 @ref{Other Utility Programs}, discusses several other GNAT utilities,
432 including @code{gnathtml}.
436 @ref{Code Coverage and Profiling}, describes how to perform a structural
437 coverage and profile the execution of Ada programs.
441 @ref{Running and Debugging Ada Programs}, describes how to run and debug
446 @ref{Compatibility with HP Ada}, details the compatibility of GNAT with
447 HP Ada 83 @footnote{``HP Ada'' refers to the legacy product originally
448 developed by Digital Equipment Corporation and currently supported by HP.}
449 for OpenVMS Alpha. This product was formerly known as DEC Ada,
452 historical compatibility reasons, the relevant libraries still use the
457 @ref{Platform-Specific Information for the Run-Time Libraries},
458 describes the various run-time
459 libraries supported by GNAT on various platforms and explains how to
460 choose a particular library.
463 @ref{Example of Binder Output File}, shows the source code for the binder
464 output file for a sample program.
467 @ref{Elaboration Order Handling in GNAT}, describes how GNAT helps
468 you deal with elaboration order issues.
471 @ref{Overflow Check Handling in GNAT}, describes how GNAT helps
472 you deal with arithmetic overflow issues.
475 @ref{Conditional Compilation}, describes how to model conditional compilation,
476 both with Ada in general and with GNAT facilities in particular.
479 @ref{Inline Assembler}, shows how to use the inline assembly facility
483 @ref{Compatibility and Porting Guide}, contains sections on compatibility
484 of GNAT with other Ada development environments (including Ada 83 systems),
485 to assist in porting code from those environments.
489 @ref{Microsoft Windows Topics}, presents information relevant to the
490 Microsoft Windows platform.
493 @ref{Mac OS Topics}, presents information relevant to Apple's OS X
498 @c *************************************************
499 @node What You Should Know before Reading This Guide
500 @c *************************************************
501 @unnumberedsec What You Should Know before Reading This Guide
503 @cindex Ada 95 Language Reference Manual
504 @cindex Ada 2005 Language Reference Manual
506 This guide assumes a basic familiarity with the Ada 95 language, as
507 described in the International Standard ANSI/ISO/IEC-8652:1995, January
509 It does not require knowledge of the new features introduced by Ada 2005,
510 (officially known as ISO/IEC 8652:1995 with Technical Corrigendum 1
512 Both reference manuals are included in the GNAT documentation
515 @node Related Information
516 @unnumberedsec Related Information
519 For further information about related tools, refer to the following
524 @xref{Top, GNAT Reference Manual, About This Guide, gnat_rm, GNAT
525 Reference Manual}, which contains all reference material for the GNAT
526 implementation of Ada.
530 @cite{Using the GNAT Programming Studio}, which describes the GPS
531 Integrated Development Environment.
534 @cite{GNAT Programming Studio Tutorial}, which introduces the
535 main GPS features through examples.
539 @cite{Ada 95 Reference Manual}, which contains reference
540 material for the Ada 95 programming language.
543 @cite{Ada 2005 Reference Manual}, which contains reference
544 material for the Ada 2005 programming language.
547 @xref{Top,, Debugging with GDB, gdb, Debugging with GDB},
549 in the GNU:[DOCS] directory,
551 for all details on the use of the GNU source-level debugger.
554 @xref{Top,, The extensible self-documenting text editor, emacs,
557 located in the GNU:[DOCS] directory if the EMACS kit is installed,
559 for full information on the extensible editor and programming
566 @unnumberedsec Conventions
568 @cindex Typographical conventions
571 Following are examples of the typographical and graphic conventions used
576 @code{Functions}, @command{utility program names}, @code{standard names},
580 @option{Option flags}
583 @file{File names}, @samp{button names}, and @samp{field names}.
586 @code{Variables}, @env{environment variables}, and @var{metasyntactic
593 @r{[}optional information or parameters@r{]}
596 Examples are described by text
598 and then shown this way.
603 Commands that are entered by the user are preceded in this manual by the
604 characters @w{``@code{$ }''} (dollar sign followed by space). If your system
605 uses this sequence as a prompt, then the commands will appear exactly as
606 you see them in the manual. If your system uses some other prompt, then
607 the command will appear with the @code{$} replaced by whatever prompt
608 character you are using.
611 Full file names are shown with the ``@code{/}'' character
612 as the directory separator; e.g., @file{parent-dir/subdir/myfile.adb}.
613 If you are using GNAT on a Windows platform, please note that
614 the ``@code{\}'' character should be used instead.
617 @c ****************************
618 @node Getting Started with GNAT
619 @chapter Getting Started with GNAT
622 This chapter describes some simple ways of using GNAT to build
623 executable Ada programs.
625 @ref{Running GNAT}, through @ref{Using the gnatmake Utility},
626 show how to use the command line environment.
627 @ref{Introduction to GPS}, provides a brief
628 introduction to the GNAT Programming Studio, a visually-oriented
629 Integrated Development Environment for GNAT.
630 GPS offers a graphical ``look and feel'', support for development in
631 other programming languages, comprehensive browsing features, and
632 many other capabilities.
633 For information on GPS please refer to
634 @cite{Using the GNAT Programming Studio}.
639 * Running a Simple Ada Program::
640 * Running a Program with Multiple Units::
641 * Using the gnatmake Utility::
643 * Editing with Emacs::
646 * Introduction to GPS::
651 @section Running GNAT
654 Three steps are needed to create an executable file from an Ada source
659 The source file(s) must be compiled.
661 The file(s) must be bound using the GNAT binder.
663 All appropriate object files must be linked to produce an executable.
667 All three steps are most commonly handled by using the @command{gnatmake}
668 utility program that, given the name of the main program, automatically
669 performs the necessary compilation, binding and linking steps.
671 @node Running a Simple Ada Program
672 @section Running a Simple Ada Program
675 Any text editor may be used to prepare an Ada program.
677 used, the optional Ada mode may be helpful in laying out the program.)
679 program text is a normal text file. We will assume in our initial
680 example that you have used your editor to prepare the following
681 standard format text file:
685 with Ada.Text_IO; use Ada.Text_IO;
688 Put_Line ("Hello WORLD!");
694 This file should be named @file{hello.adb}.
695 With the normal default file naming conventions, GNAT requires
697 contain a single compilation unit whose file name is the
699 with periods replaced by hyphens; the
700 extension is @file{ads} for a
701 spec and @file{adb} for a body.
702 You can override this default file naming convention by use of the
703 special pragma @code{Source_File_Name} (@pxref{Using Other File Names}).
704 Alternatively, if you want to rename your files according to this default
705 convention, which is probably more convenient if you will be using GNAT
706 for all your compilations, then the @code{gnatchop} utility
707 can be used to generate correctly-named source files
708 (@pxref{Renaming Files with gnatchop}).
710 You can compile the program using the following command (@code{$} is used
711 as the command prompt in the examples in this document):
718 @command{gcc} is the command used to run the compiler. This compiler is
719 capable of compiling programs in several languages, including Ada and
720 C. It assumes that you have given it an Ada program if the file extension is
721 either @file{.ads} or @file{.adb}, and it will then call
722 the GNAT compiler to compile the specified file.
725 The @option{-c} switch is required. It tells @command{gcc} to only do a
726 compilation. (For C programs, @command{gcc} can also do linking, but this
727 capability is not used directly for Ada programs, so the @option{-c}
728 switch must always be present.)
731 This compile command generates a file
732 @file{hello.o}, which is the object
733 file corresponding to your Ada program. It also generates
734 an ``Ada Library Information'' file @file{hello.ali},
735 which contains additional information used to check
736 that an Ada program is consistent.
737 To build an executable file,
738 use @code{gnatbind} to bind the program
739 and @command{gnatlink} to link it. The
740 argument to both @code{gnatbind} and @command{gnatlink} is the name of the
741 @file{ALI} file, but the default extension of @file{.ali} can
742 be omitted. This means that in the most common case, the argument
743 is simply the name of the main program:
751 A simpler method of carrying out these steps is to use
753 a master program that invokes all the required
754 compilation, binding and linking tools in the correct order. In particular,
755 @command{gnatmake} automatically recompiles any sources that have been
756 modified since they were last compiled, or sources that depend
757 on such modified sources, so that ``version skew'' is avoided.
758 @cindex Version skew (avoided by @command{gnatmake})
765 The result is an executable program called @file{hello}, which can be
773 assuming that the current directory is on the search path
774 for executable programs.
777 and, if all has gone well, you will see
784 appear in response to this command.
786 @c ****************************************
787 @node Running a Program with Multiple Units
788 @section Running a Program with Multiple Units
791 Consider a slightly more complicated example that has three files: a
792 main program, and the spec and body of a package:
802 with Ada.Text_IO; use Ada.Text_IO;
803 package body Greetings is
806 Put_Line ("Hello WORLD!");
811 Put_Line ("Goodbye WORLD!");
828 Following the one-unit-per-file rule, place this program in the
829 following three separate files:
833 spec of package @code{Greetings}
836 body of package @code{Greetings}
843 To build an executable version of
844 this program, we could use four separate steps to compile, bind, and link
845 the program, as follows:
849 $ gcc -c greetings.adb
855 Note that there is no required order of compilation when using GNAT.
856 In particular it is perfectly fine to compile the main program first.
857 Also, it is not necessary to compile package specs in the case where
858 there is an accompanying body; you only need to compile the body. If you want
859 to submit these files to the compiler for semantic checking and not code
860 generation, then use the
861 @option{-gnatc} switch:
864 $ gcc -c greetings.ads -gnatc
868 Although the compilation can be done in separate steps as in the
869 above example, in practice it is almost always more convenient
870 to use the @command{gnatmake} tool. All you need to know in this case
871 is the name of the main program's source file. The effect of the above four
872 commands can be achieved with a single one:
879 In the next section we discuss the advantages of using @command{gnatmake} in
882 @c *****************************
883 @node Using the gnatmake Utility
884 @section Using the @command{gnatmake} Utility
887 If you work on a program by compiling single components at a time using
888 @command{gcc}, you typically keep track of the units you modify. In order to
889 build a consistent system, you compile not only these units, but also any
890 units that depend on the units you have modified.
891 For example, in the preceding case,
892 if you edit @file{gmain.adb}, you only need to recompile that file. But if
893 you edit @file{greetings.ads}, you must recompile both
894 @file{greetings.adb} and @file{gmain.adb}, because both files contain
895 units that depend on @file{greetings.ads}.
897 @code{gnatbind} will warn you if you forget one of these compilation
898 steps, so that it is impossible to generate an inconsistent program as a
899 result of forgetting to do a compilation. Nevertheless it is tedious and
900 error-prone to keep track of dependencies among units.
901 One approach to handle the dependency-bookkeeping is to use a
902 makefile. However, makefiles present maintenance problems of their own:
903 if the dependencies change as you change the program, you must make
904 sure that the makefile is kept up-to-date manually, which is also an
907 The @command{gnatmake} utility takes care of these details automatically.
908 Invoke it using either one of the following forms:
912 $ gnatmake ^gmain^GMAIN^
916 The argument is the name of the file containing the main program;
917 you may omit the extension. @command{gnatmake}
918 examines the environment, automatically recompiles any files that need
919 recompiling, and binds and links the resulting set of object files,
920 generating the executable file, @file{^gmain^GMAIN.EXE^}.
921 In a large program, it
922 can be extremely helpful to use @command{gnatmake}, because working out by hand
923 what needs to be recompiled can be difficult.
925 Note that @command{gnatmake}
926 takes into account all the Ada rules that
927 establish dependencies among units. These include dependencies that result
928 from inlining subprogram bodies, and from
929 generic instantiation. Unlike some other
930 Ada make tools, @command{gnatmake} does not rely on the dependencies that were
931 found by the compiler on a previous compilation, which may possibly
932 be wrong when sources change. @command{gnatmake} determines the exact set of
933 dependencies from scratch each time it is run.
936 @node Editing with Emacs
937 @section Editing with Emacs
941 Emacs is an extensible self-documenting text editor that is available in a
942 separate VMSINSTAL kit.
944 Invoke Emacs by typing @kbd{Emacs} at the command prompt. To get started,
945 click on the Emacs Help menu and run the Emacs Tutorial.
946 In a character cell terminal, Emacs help is invoked with @kbd{Ctrl-h} (also
947 written as @kbd{C-h}), and the tutorial by @kbd{C-h t}.
949 Documentation on Emacs and other tools is available in Emacs under the
950 pull-down menu button: @code{Help - Info}. After selecting @code{Info},
951 use the middle mouse button to select a topic (e.g.@: Emacs).
953 In a character cell terminal, do @kbd{C-h i} to invoke info, and then @kbd{m}
954 (stands for menu) followed by the menu item desired, as in @kbd{m Emacs}, to
955 get to the Emacs manual.
956 Help on Emacs is also available by typing @kbd{HELP EMACS} at the DCL command
959 The tutorial is highly recommended in order to learn the intricacies of Emacs,
960 which is sufficiently extensible to provide for a complete programming
961 environment and shell for the sophisticated user.
965 @node Introduction to GPS
966 @section Introduction to GPS
967 @cindex GPS (GNAT Programming Studio)
968 @cindex GNAT Programming Studio (GPS)
970 Although the command line interface (@command{gnatmake}, etc.) alone
971 is sufficient, a graphical Interactive Development
972 Environment can make it easier for you to compose, navigate, and debug
973 programs. This section describes the main features of GPS
974 (``GNAT Programming Studio''), the GNAT graphical IDE.
975 You will see how to use GPS to build and debug an executable, and
976 you will also learn some of the basics of the GNAT ``project'' facility.
978 GPS enables you to do much more than is presented here;
979 e.g., you can produce a call graph, interface to a third-party
980 Version Control System, and inspect the generated assembly language
982 Indeed, GPS also supports languages other than Ada.
983 Such additional information, and an explanation of all of the GPS menu
984 items. may be found in the on-line help, which includes
985 a user's guide and a tutorial (these are also accessible from the GNAT
989 * Building a New Program with GPS::
990 * Simple Debugging with GPS::
993 @node Building a New Program with GPS
994 @subsection Building a New Program with GPS
996 GPS invokes the GNAT compilation tools using information
997 contained in a @emph{project} (also known as a @emph{project file}):
998 a collection of properties such
999 as source directories, identities of main subprograms, tool switches, etc.,
1000 and their associated values.
1001 See @ref{GNAT Project Manager} for details.
1002 In order to run GPS, you will need to either create a new project
1003 or else open an existing one.
1005 This section will explain how you can use GPS to create a project,
1006 to associate Ada source files with a project, and to build and run
1010 @item @emph{Creating a project}
1012 Invoke GPS, either from the command line or the platform's IDE.
1013 After it starts, GPS will display a ``Welcome'' screen with three
1018 @code{Start with default project in directory}
1021 @code{Create new project with wizard}
1024 @code{Open existing project}
1028 Select @code{Create new project with wizard} and press @code{OK}.
1029 A new window will appear. In the text box labeled with
1030 @code{Enter the name of the project to create}, type @file{sample}
1031 as the project name.
1032 In the next box, browse to choose the directory in which you
1033 would like to create the project file.
1034 After selecting an appropriate directory, press @code{Forward}.
1036 A window will appear with the title
1037 @code{Version Control System Configuration}.
1038 Simply press @code{Forward}.
1040 A window will appear with the title
1041 @code{Please select the source directories for this project}.
1042 The directory that you specified for the project file will be selected
1043 by default as the one to use for sources; simply press @code{Forward}.
1045 A window will appear with the title
1046 @code{Please select the build directory for this project}.
1047 The directory that you specified for the project file will be selected
1048 by default for object files and executables;
1049 simply press @code{Forward}.
1051 A window will appear with the title
1052 @code{Please select the main units for this project}.
1053 You will supply this information later, after creating the source file.
1054 Simply press @code{Forward} for now.
1056 A window will appear with the title
1057 @code{Please select the switches to build the project}.
1058 Press @code{Apply}. This will create a project file named
1059 @file{sample.prj} in the directory that you had specified.
1061 @item @emph{Creating and saving the source file}
1063 After you create the new project, a GPS window will appear, which is
1064 partitioned into two main sections:
1068 A @emph{Workspace area}, initially greyed out, which you will use for
1069 creating and editing source files
1072 Directly below, a @emph{Messages area}, which initially displays a
1073 ``Welcome'' message.
1074 (If the Messages area is not visible, drag its border upward to expand it.)
1078 Select @code{File} on the menu bar, and then the @code{New} command.
1079 The Workspace area will become white, and you can now
1080 enter the source program explicitly.
1081 Type the following text
1083 @smallexample @c ada
1085 with Ada.Text_IO; use Ada.Text_IO;
1088 Put_Line("Hello from GPS!");
1094 Select @code{File}, then @code{Save As}, and enter the source file name
1096 The file will be saved in the same directory you specified as the
1097 location of the default project file.
1099 @item @emph{Updating the project file}
1101 You need to add the new source file to the project.
1103 the @code{Project} menu and then @code{Edit project properties}.
1104 Click the @code{Main files} tab on the left, and then the
1106 Choose @file{hello.adb} from the list, and press @code{Open}.
1107 The project settings window will reflect this action.
1110 @item @emph{Building and running the program}
1112 In the main GPS window, now choose the @code{Build} menu, then @code{Make},
1113 and select @file{hello.adb}.
1114 The Messages window will display the resulting invocations of @command{gcc},
1115 @command{gnatbind}, and @command{gnatlink}
1116 (reflecting the default switch settings from the
1117 project file that you created) and then a ``successful compilation/build''
1120 To run the program, choose the @code{Build} menu, then @code{Run}, and
1121 select @command{hello}.
1122 An @emph{Arguments Selection} window will appear.
1123 There are no command line arguments, so just click @code{OK}.
1125 The Messages window will now display the program's output (the string
1126 @code{Hello from GPS}), and at the bottom of the GPS window a status
1127 update is displayed (@code{Run: hello}).
1128 Close the GPS window (or select @code{File}, then @code{Exit}) to
1129 terminate this GPS session.
1132 @node Simple Debugging with GPS
1133 @subsection Simple Debugging with GPS
1135 This section illustrates basic debugging techniques (setting breakpoints,
1136 examining/modifying variables, single stepping).
1139 @item @emph{Opening a project}
1141 Start GPS and select @code{Open existing project}; browse to
1142 specify the project file @file{sample.prj} that you had created in the
1145 @item @emph{Creating a source file}
1147 Select @code{File}, then @code{New}, and type in the following program:
1149 @smallexample @c ada
1151 with Ada.Text_IO; use Ada.Text_IO;
1152 procedure Example is
1153 Line : String (1..80);
1156 Put_Line("Type a line of text at each prompt; an empty line to exit");
1160 Put_Line (Line (1..N) );
1168 Select @code{File}, then @code{Save as}, and enter the file name
1171 @item @emph{Updating the project file}
1173 Add @code{Example} as a new main unit for the project:
1176 Select @code{Project}, then @code{Edit Project Properties}.
1179 Select the @code{Main files} tab, click @code{Add}, then
1180 select the file @file{example.adb} from the list, and
1182 You will see the file name appear in the list of main units
1188 @item @emph{Building/running the executable}
1190 To build the executable
1191 select @code{Build}, then @code{Make}, and then choose @file{example.adb}.
1193 Run the program to see its effect (in the Messages area).
1194 Each line that you enter is displayed; an empty line will
1195 cause the loop to exit and the program to terminate.
1197 @item @emph{Debugging the program}
1199 Note that the @option{-g} switches to @command{gcc} and @command{gnatlink},
1200 which are required for debugging, are on by default when you create
1202 Thus unless you intentionally remove these settings, you will be able
1203 to debug any program that you develop using GPS.
1206 @item @emph{Initializing}
1208 Select @code{Debug}, then @code{Initialize}, then @file{example}
1210 @item @emph{Setting a breakpoint}
1212 After performing the initialization step, you will observe a small
1213 icon to the right of each line number.
1214 This serves as a toggle for breakpoints; clicking the icon will
1215 set a breakpoint at the corresponding line (the icon will change to
1216 a red circle with an ``x''), and clicking it again
1217 will remove the breakpoint / reset the icon.
1219 For purposes of this example, set a breakpoint at line 10 (the
1220 statement @code{Put_Line@ (Line@ (1..N));}
1222 @item @emph{Starting program execution}
1224 Select @code{Debug}, then @code{Run}. When the
1225 @code{Program Arguments} window appears, click @code{OK}.
1226 A console window will appear; enter some line of text,
1227 e.g.@: @code{abcde}, at the prompt.
1228 The program will pause execution when it gets to the
1229 breakpoint, and the corresponding line is highlighted.
1231 @item @emph{Examining a variable}
1233 Move the mouse over one of the occurrences of the variable @code{N}.
1234 You will see the value (5) displayed, in ``tool tip'' fashion.
1235 Right click on @code{N}, select @code{Debug}, then select @code{Display N}.
1236 You will see information about @code{N} appear in the @code{Debugger Data}
1237 pane, showing the value as 5.
1239 @item @emph{Assigning a new value to a variable}
1241 Right click on the @code{N} in the @code{Debugger Data} pane, and
1242 select @code{Set value of N}.
1243 When the input window appears, enter the value @code{4} and click
1245 This value does not automatically appear in the @code{Debugger Data}
1246 pane; to see it, right click again on the @code{N} in the
1247 @code{Debugger Data} pane and select @code{Update value}.
1248 The new value, 4, will appear in red.
1250 @item @emph{Single stepping}
1252 Select @code{Debug}, then @code{Next}.
1253 This will cause the next statement to be executed, in this case the
1254 call of @code{Put_Line} with the string slice.
1255 Notice in the console window that the displayed string is simply
1256 @code{abcd} and not @code{abcde} which you had entered.
1257 This is because the upper bound of the slice is now 4 rather than 5.
1259 @item @emph{Removing a breakpoint}
1261 Toggle the breakpoint icon at line 10.
1263 @item @emph{Resuming execution from a breakpoint}
1265 Select @code{Debug}, then @code{Continue}.
1266 The program will reach the next iteration of the loop, and
1267 wait for input after displaying the prompt.
1268 This time, just hit the @kbd{Enter} key.
1269 The value of @code{N} will be 0, and the program will terminate.
1270 The console window will disappear.
1275 @node The GNAT Compilation Model
1276 @chapter The GNAT Compilation Model
1277 @cindex GNAT compilation model
1278 @cindex Compilation model
1281 * Source Representation::
1282 * Foreign Language Representation::
1283 * File Naming Rules::
1284 * Using Other File Names::
1285 * Alternative File Naming Schemes::
1286 * Generating Object Files::
1287 * Source Dependencies::
1288 * The Ada Library Information Files::
1289 * Binding an Ada Program::
1290 * Mixed Language Programming::
1292 * Building Mixed Ada & C++ Programs::
1293 * Comparison between GNAT and C/C++ Compilation Models::
1295 * Comparison between GNAT and Conventional Ada Library Models::
1297 * Placement of temporary files::
1302 This chapter describes the compilation model used by GNAT. Although
1303 similar to that used by other languages, such as C and C++, this model
1304 is substantially different from the traditional Ada compilation models,
1305 which are based on a library. The model is initially described without
1306 reference to the library-based model. If you have not previously used an
1307 Ada compiler, you need only read the first part of this chapter. The
1308 last section describes and discusses the differences between the GNAT
1309 model and the traditional Ada compiler models. If you have used other
1310 Ada compilers, this section will help you to understand those
1311 differences, and the advantages of the GNAT model.
1313 @node Source Representation
1314 @section Source Representation
1318 Ada source programs are represented in standard text files, using
1319 Latin-1 coding. Latin-1 is an 8-bit code that includes the familiar
1320 7-bit ASCII set, plus additional characters used for
1321 representing foreign languages (@pxref{Foreign Language Representation}
1322 for support of non-USA character sets). The format effector characters
1323 are represented using their standard ASCII encodings, as follows:
1328 Vertical tab, @code{16#0B#}
1332 Horizontal tab, @code{16#09#}
1336 Carriage return, @code{16#0D#}
1340 Line feed, @code{16#0A#}
1344 Form feed, @code{16#0C#}
1348 Source files are in standard text file format. In addition, GNAT will
1349 recognize a wide variety of stream formats, in which the end of
1350 physical lines is marked by any of the following sequences:
1351 @code{LF}, @code{CR}, @code{CR-LF}, or @code{LF-CR}. This is useful
1352 in accommodating files that are imported from other operating systems.
1354 @cindex End of source file
1355 @cindex Source file, end
1357 The end of a source file is normally represented by the physical end of
1358 file. However, the control character @code{16#1A#} (@code{SUB}) is also
1359 recognized as signalling the end of the source file. Again, this is
1360 provided for compatibility with other operating systems where this
1361 code is used to represent the end of file.
1363 Each file contains a single Ada compilation unit, including any pragmas
1364 associated with the unit. For example, this means you must place a
1365 package declaration (a package @dfn{spec}) and the corresponding body in
1366 separate files. An Ada @dfn{compilation} (which is a sequence of
1367 compilation units) is represented using a sequence of files. Similarly,
1368 you will place each subunit or child unit in a separate file.
1370 @node Foreign Language Representation
1371 @section Foreign Language Representation
1374 GNAT supports the standard character sets defined in Ada as well as
1375 several other non-standard character sets for use in localized versions
1376 of the compiler (@pxref{Character Set Control}).
1379 * Other 8-Bit Codes::
1380 * Wide Character Encodings::
1388 The basic character set is Latin-1. This character set is defined by ISO
1389 standard 8859, part 1. The lower half (character codes @code{16#00#}
1390 @dots{} @code{16#7F#)} is identical to standard ASCII coding, but the upper
1391 half is used to represent additional characters. These include extended letters
1392 used by European languages, such as French accents, the vowels with umlauts
1393 used in German, and the extra letter A-ring used in Swedish.
1395 @findex Ada.Characters.Latin_1
1396 For a complete list of Latin-1 codes and their encodings, see the source
1397 file of library unit @code{Ada.Characters.Latin_1} in file
1398 @file{a-chlat1.ads}.
1399 You may use any of these extended characters freely in character or
1400 string literals. In addition, the extended characters that represent
1401 letters can be used in identifiers.
1403 @node Other 8-Bit Codes
1404 @subsection Other 8-Bit Codes
1407 GNAT also supports several other 8-bit coding schemes:
1410 @item ISO 8859-2 (Latin-2)
1413 Latin-2 letters allowed in identifiers, with uppercase and lowercase
1416 @item ISO 8859-3 (Latin-3)
1419 Latin-3 letters allowed in identifiers, with uppercase and lowercase
1422 @item ISO 8859-4 (Latin-4)
1425 Latin-4 letters allowed in identifiers, with uppercase and lowercase
1428 @item ISO 8859-5 (Cyrillic)
1431 ISO 8859-5 letters (Cyrillic) allowed in identifiers, with uppercase and
1432 lowercase equivalence.
1434 @item ISO 8859-15 (Latin-9)
1437 ISO 8859-15 (Latin-9) letters allowed in identifiers, with uppercase and
1438 lowercase equivalence
1440 @item IBM PC (code page 437)
1441 @cindex code page 437
1442 This code page is the normal default for PCs in the U.S. It corresponds
1443 to the original IBM PC character set. This set has some, but not all, of
1444 the extended Latin-1 letters, but these letters do not have the same
1445 encoding as Latin-1. In this mode, these letters are allowed in
1446 identifiers with uppercase and lowercase equivalence.
1448 @item IBM PC (code page 850)
1449 @cindex code page 850
1450 This code page is a modification of 437 extended to include all the
1451 Latin-1 letters, but still not with the usual Latin-1 encoding. In this
1452 mode, all these letters are allowed in identifiers with uppercase and
1453 lowercase equivalence.
1455 @item Full Upper 8-bit
1456 Any character in the range 80-FF allowed in identifiers, and all are
1457 considered distinct. In other words, there are no uppercase and lowercase
1458 equivalences in this range. This is useful in conjunction with
1459 certain encoding schemes used for some foreign character sets (e.g.,
1460 the typical method of representing Chinese characters on the PC).
1463 No upper-half characters in the range 80-FF are allowed in identifiers.
1464 This gives Ada 83 compatibility for identifier names.
1468 For precise data on the encodings permitted, and the uppercase and lowercase
1469 equivalences that are recognized, see the file @file{csets.adb} in
1470 the GNAT compiler sources. You will need to obtain a full source release
1471 of GNAT to obtain this file.
1473 @node Wide Character Encodings
1474 @subsection Wide Character Encodings
1477 GNAT allows wide character codes to appear in character and string
1478 literals, and also optionally in identifiers, by means of the following
1479 possible encoding schemes:
1484 In this encoding, a wide character is represented by the following five
1492 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1493 characters (using uppercase letters) of the wide character code. For
1494 example, ESC A345 is used to represent the wide character with code
1496 This scheme is compatible with use of the full Wide_Character set.
1498 @item Upper-Half Coding
1499 @cindex Upper-Half Coding
1500 The wide character with encoding @code{16#abcd#} where the upper bit is on
1501 (in other words, ``a'' is in the range 8-F) is represented as two bytes,
1502 @code{16#ab#} and @code{16#cd#}. The second byte cannot be a format control
1503 character, but is not required to be in the upper half. This method can
1504 be also used for shift-JIS or EUC, where the internal coding matches the
1507 @item Shift JIS Coding
1508 @cindex Shift JIS Coding
1509 A wide character is represented by a two-character sequence,
1511 @code{16#cd#}, with the restrictions described for upper-half encoding as
1512 described above. The internal character code is the corresponding JIS
1513 character according to the standard algorithm for Shift-JIS
1514 conversion. Only characters defined in the JIS code set table can be
1515 used with this encoding method.
1519 A wide character is represented by a two-character sequence
1521 @code{16#cd#}, with both characters being in the upper half. The internal
1522 character code is the corresponding JIS character according to the EUC
1523 encoding algorithm. Only characters defined in the JIS code set table
1524 can be used with this encoding method.
1527 A wide character is represented using
1528 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
1529 10646-1/Am.2. Depending on the character value, the representation
1530 is a one, two, or three byte sequence:
1535 16#0000#-16#007f#: 2#0@var{xxxxxxx}#
1536 16#0080#-16#07ff#: 2#110@var{xxxxx}# 2#10@var{xxxxxx}#
1537 16#0800#-16#ffff#: 2#1110@var{xxxx}# 2#10@var{xxxxxx}# 2#10@var{xxxxxx}#
1542 where the @var{xxx} bits correspond to the left-padded bits of the
1543 16-bit character value. Note that all lower half ASCII characters
1544 are represented as ASCII bytes and all upper half characters and
1545 other wide characters are represented as sequences of upper-half
1546 (The full UTF-8 scheme allows for encoding 31-bit characters as
1547 6-byte sequences, but in this implementation, all UTF-8 sequences
1548 of four or more bytes length will be treated as illegal).
1549 @item Brackets Coding
1550 In this encoding, a wide character is represented by the following eight
1558 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1559 characters (using uppercase letters) of the wide character code. For
1560 example, [``A345''] is used to represent the wide character with code
1561 @code{16#A345#}. It is also possible (though not required) to use the
1562 Brackets coding for upper half characters. For example, the code
1563 @code{16#A3#} can be represented as @code{[``A3'']}.
1565 This scheme is compatible with use of the full Wide_Character set,
1566 and is also the method used for wide character encoding in the standard
1567 ACVC (Ada Compiler Validation Capability) test suite distributions.
1572 Note: Some of these coding schemes do not permit the full use of the
1573 Ada character set. For example, neither Shift JIS, nor EUC allow the
1574 use of the upper half of the Latin-1 set.
1576 @node File Naming Rules
1577 @section File Naming Rules
1580 The default file name is determined by the name of the unit that the
1581 file contains. The name is formed by taking the full expanded name of
1582 the unit and replacing the separating dots with hyphens and using
1583 ^lowercase^uppercase^ for all letters.
1585 An exception arises if the file name generated by the above rules starts
1586 with one of the characters
1588 @samp{A}, @samp{G}, @samp{I}, or @samp{S},
1591 @samp{a}, @samp{g}, @samp{i}, or @samp{s},
1593 and the second character is a
1594 minus. In this case, the character ^tilde^dollar sign^ is used in place
1595 of the minus. The reason for this special rule is to avoid clashes with
1596 the standard names for child units of the packages System, Ada,
1597 Interfaces, and GNAT, which use the prefixes
1599 @samp{S-}, @samp{A-}, @samp{I-}, and @samp{G-},
1602 @samp{s-}, @samp{a-}, @samp{i-}, and @samp{g-},
1606 The file extension is @file{.ads} for a spec and
1607 @file{.adb} for a body. The following list shows some
1608 examples of these rules.
1615 @item arith_functions.ads
1616 Arith_Functions (package spec)
1617 @item arith_functions.adb
1618 Arith_Functions (package body)
1620 Func.Spec (child package spec)
1622 Func.Spec (child package body)
1624 Sub (subunit of Main)
1625 @item ^a~bad.adb^A$BAD.ADB^
1626 A.Bad (child package body)
1630 Following these rules can result in excessively long
1631 file names if corresponding
1632 unit names are long (for example, if child units or subunits are
1633 heavily nested). An option is available to shorten such long file names
1634 (called file name ``krunching''). This may be particularly useful when
1635 programs being developed with GNAT are to be used on operating systems
1636 with limited file name lengths. @xref{Using gnatkr}.
1638 Of course, no file shortening algorithm can guarantee uniqueness over
1639 all possible unit names; if file name krunching is used, it is your
1640 responsibility to ensure no name clashes occur. Alternatively you
1641 can specify the exact file names that you want used, as described
1642 in the next section. Finally, if your Ada programs are migrating from a
1643 compiler with a different naming convention, you can use the gnatchop
1644 utility to produce source files that follow the GNAT naming conventions.
1645 (For details @pxref{Renaming Files with gnatchop}.)
1647 Note: in the case of @code{Windows NT/XP} or @code{OpenVMS} operating
1648 systems, case is not significant. So for example on @code{Windows XP}
1649 if the canonical name is @code{main-sub.adb}, you can use the file name
1650 @code{Main-Sub.adb} instead. However, case is significant for other
1651 operating systems, so for example, if you want to use other than
1652 canonically cased file names on a Unix system, you need to follow
1653 the procedures described in the next section.
1655 @node Using Other File Names
1656 @section Using Other File Names
1660 In the previous section, we have described the default rules used by
1661 GNAT to determine the file name in which a given unit resides. It is
1662 often convenient to follow these default rules, and if you follow them,
1663 the compiler knows without being explicitly told where to find all
1666 However, in some cases, particularly when a program is imported from
1667 another Ada compiler environment, it may be more convenient for the
1668 programmer to specify which file names contain which units. GNAT allows
1669 arbitrary file names to be used by means of the Source_File_Name pragma.
1670 The form of this pragma is as shown in the following examples:
1671 @cindex Source_File_Name pragma
1673 @smallexample @c ada
1675 pragma Source_File_Name (My_Utilities.Stacks,
1676 Spec_File_Name => "myutilst_a.ada");
1677 pragma Source_File_name (My_Utilities.Stacks,
1678 Body_File_Name => "myutilst.ada");
1683 As shown in this example, the first argument for the pragma is the unit
1684 name (in this example a child unit). The second argument has the form
1685 of a named association. The identifier
1686 indicates whether the file name is for a spec or a body;
1687 the file name itself is given by a string literal.
1689 The source file name pragma is a configuration pragma, which means that
1690 normally it will be placed in the @file{gnat.adc}
1691 file used to hold configuration
1692 pragmas that apply to a complete compilation environment.
1693 For more details on how the @file{gnat.adc} file is created and used
1694 see @ref{Handling of Configuration Pragmas}.
1695 @cindex @file{gnat.adc}
1698 GNAT allows completely arbitrary file names to be specified using the
1699 source file name pragma. However, if the file name specified has an
1700 extension other than @file{.ads} or @file{.adb} it is necessary to use
1701 a special syntax when compiling the file. The name in this case must be
1702 preceded by the special sequence @option{-x} followed by a space and the name
1703 of the language, here @code{ada}, as in:
1706 $ gcc -c -x ada peculiar_file_name.sim
1711 @command{gnatmake} handles non-standard file names in the usual manner (the
1712 non-standard file name for the main program is simply used as the
1713 argument to gnatmake). Note that if the extension is also non-standard,
1714 then it must be included in the @command{gnatmake} command, it may not
1717 @node Alternative File Naming Schemes
1718 @section Alternative File Naming Schemes
1719 @cindex File naming schemes, alternative
1722 In the previous section, we described the use of the @code{Source_File_Name}
1723 pragma to allow arbitrary names to be assigned to individual source files.
1724 However, this approach requires one pragma for each file, and especially in
1725 large systems can result in very long @file{gnat.adc} files, and also create
1726 a maintenance problem.
1728 GNAT also provides a facility for specifying systematic file naming schemes
1729 other than the standard default naming scheme previously described. An
1730 alternative scheme for naming is specified by the use of
1731 @code{Source_File_Name} pragmas having the following format:
1732 @cindex Source_File_Name pragma
1734 @smallexample @c ada
1735 pragma Source_File_Name (
1736 Spec_File_Name => FILE_NAME_PATTERN
1737 @r{[},Casing => CASING_SPEC@r{]}
1738 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
1740 pragma Source_File_Name (
1741 Body_File_Name => FILE_NAME_PATTERN
1742 @r{[},Casing => CASING_SPEC@r{]}
1743 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
1745 pragma Source_File_Name (
1746 Subunit_File_Name => FILE_NAME_PATTERN
1747 @r{[},Casing => CASING_SPEC@r{]}
1748 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
1750 FILE_NAME_PATTERN ::= STRING_LITERAL
1751 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
1755 The @code{FILE_NAME_PATTERN} string shows how the file name is constructed.
1756 It contains a single asterisk character, and the unit name is substituted
1757 systematically for this asterisk. The optional parameter
1758 @code{Casing} indicates
1759 whether the unit name is to be all upper-case letters, all lower-case letters,
1760 or mixed-case. If no
1761 @code{Casing} parameter is used, then the default is all
1762 ^lower-case^upper-case^.
1764 The optional @code{Dot_Replacement} string is used to replace any periods
1765 that occur in subunit or child unit names. If no @code{Dot_Replacement}
1766 argument is used then separating dots appear unchanged in the resulting
1768 Although the above syntax indicates that the
1769 @code{Casing} argument must appear
1770 before the @code{Dot_Replacement} argument, but it
1771 is also permissible to write these arguments in the opposite order.
1773 As indicated, it is possible to specify different naming schemes for
1774 bodies, specs, and subunits. Quite often the rule for subunits is the
1775 same as the rule for bodies, in which case, there is no need to give
1776 a separate @code{Subunit_File_Name} rule, and in this case the
1777 @code{Body_File_name} rule is used for subunits as well.
1779 The separate rule for subunits can also be used to implement the rather
1780 unusual case of a compilation environment (e.g.@: a single directory) which
1781 contains a subunit and a child unit with the same unit name. Although
1782 both units cannot appear in the same partition, the Ada Reference Manual
1783 allows (but does not require) the possibility of the two units coexisting
1784 in the same environment.
1786 The file name translation works in the following steps:
1791 If there is a specific @code{Source_File_Name} pragma for the given unit,
1792 then this is always used, and any general pattern rules are ignored.
1795 If there is a pattern type @code{Source_File_Name} pragma that applies to
1796 the unit, then the resulting file name will be used if the file exists. If
1797 more than one pattern matches, the latest one will be tried first, and the
1798 first attempt resulting in a reference to a file that exists will be used.
1801 If no pattern type @code{Source_File_Name} pragma that applies to the unit
1802 for which the corresponding file exists, then the standard GNAT default
1803 naming rules are used.
1808 As an example of the use of this mechanism, consider a commonly used scheme
1809 in which file names are all lower case, with separating periods copied
1810 unchanged to the resulting file name, and specs end with @file{.1.ada}, and
1811 bodies end with @file{.2.ada}. GNAT will follow this scheme if the following
1814 @smallexample @c ada
1815 pragma Source_File_Name
1816 (Spec_File_Name => "*.1.ada");
1817 pragma Source_File_Name
1818 (Body_File_Name => "*.2.ada");
1822 The default GNAT scheme is actually implemented by providing the following
1823 default pragmas internally:
1825 @smallexample @c ada
1826 pragma Source_File_Name
1827 (Spec_File_Name => "*.ads", Dot_Replacement => "-");
1828 pragma Source_File_Name
1829 (Body_File_Name => "*.adb", Dot_Replacement => "-");
1833 Our final example implements a scheme typically used with one of the
1834 Ada 83 compilers, where the separator character for subunits was ``__''
1835 (two underscores), specs were identified by adding @file{_.ADA}, bodies
1836 by adding @file{.ADA}, and subunits by
1837 adding @file{.SEP}. All file names were
1838 upper case. Child units were not present of course since this was an
1839 Ada 83 compiler, but it seems reasonable to extend this scheme to use
1840 the same double underscore separator for child units.
1842 @smallexample @c ada
1843 pragma Source_File_Name
1844 (Spec_File_Name => "*_.ADA",
1845 Dot_Replacement => "__",
1846 Casing = Uppercase);
1847 pragma Source_File_Name
1848 (Body_File_Name => "*.ADA",
1849 Dot_Replacement => "__",
1850 Casing = Uppercase);
1851 pragma Source_File_Name
1852 (Subunit_File_Name => "*.SEP",
1853 Dot_Replacement => "__",
1854 Casing = Uppercase);
1857 @node Generating Object Files
1858 @section Generating Object Files
1861 An Ada program consists of a set of source files, and the first step in
1862 compiling the program is to generate the corresponding object files.
1863 These are generated by compiling a subset of these source files.
1864 The files you need to compile are the following:
1868 If a package spec has no body, compile the package spec to produce the
1869 object file for the package.
1872 If a package has both a spec and a body, compile the body to produce the
1873 object file for the package. The source file for the package spec need
1874 not be compiled in this case because there is only one object file, which
1875 contains the code for both the spec and body of the package.
1878 For a subprogram, compile the subprogram body to produce the object file
1879 for the subprogram. The spec, if one is present, is as usual in a
1880 separate file, and need not be compiled.
1884 In the case of subunits, only compile the parent unit. A single object
1885 file is generated for the entire subunit tree, which includes all the
1889 Compile child units independently of their parent units
1890 (though, of course, the spec of all the ancestor unit must be present in order
1891 to compile a child unit).
1895 Compile generic units in the same manner as any other units. The object
1896 files in this case are small dummy files that contain at most the
1897 flag used for elaboration checking. This is because GNAT always handles generic
1898 instantiation by means of macro expansion. However, it is still necessary to
1899 compile generic units, for dependency checking and elaboration purposes.
1903 The preceding rules describe the set of files that must be compiled to
1904 generate the object files for a program. Each object file has the same
1905 name as the corresponding source file, except that the extension is
1908 You may wish to compile other files for the purpose of checking their
1909 syntactic and semantic correctness. For example, in the case where a
1910 package has a separate spec and body, you would not normally compile the
1911 spec. However, it is convenient in practice to compile the spec to make
1912 sure it is error-free before compiling clients of this spec, because such
1913 compilations will fail if there is an error in the spec.
1915 GNAT provides an option for compiling such files purely for the
1916 purposes of checking correctness; such compilations are not required as
1917 part of the process of building a program. To compile a file in this
1918 checking mode, use the @option{-gnatc} switch.
1920 @node Source Dependencies
1921 @section Source Dependencies
1924 A given object file clearly depends on the source file which is compiled
1925 to produce it. Here we are using @dfn{depends} in the sense of a typical
1926 @code{make} utility; in other words, an object file depends on a source
1927 file if changes to the source file require the object file to be
1929 In addition to this basic dependency, a given object may depend on
1930 additional source files as follows:
1934 If a file being compiled @code{with}'s a unit @var{X}, the object file
1935 depends on the file containing the spec of unit @var{X}. This includes
1936 files that are @code{with}'ed implicitly either because they are parents
1937 of @code{with}'ed child units or they are run-time units required by the
1938 language constructs used in a particular unit.
1941 If a file being compiled instantiates a library level generic unit, the
1942 object file depends on both the spec and body files for this generic
1946 If a file being compiled instantiates a generic unit defined within a
1947 package, the object file depends on the body file for the package as
1948 well as the spec file.
1952 @cindex @option{-gnatn} switch
1953 If a file being compiled contains a call to a subprogram for which
1954 pragma @code{Inline} applies and inlining is activated with the
1955 @option{-gnatn} switch, the object file depends on the file containing the
1956 body of this subprogram as well as on the file containing the spec. Note
1957 that for inlining to actually occur as a result of the use of this switch,
1958 it is necessary to compile in optimizing mode.
1960 @cindex @option{-gnatN} switch
1961 The use of @option{-gnatN} activates inlining optimization
1962 that is performed by the front end of the compiler. This inlining does
1963 not require that the code generation be optimized. Like @option{-gnatn},
1964 the use of this switch generates additional dependencies.
1966 When using a gcc-based back end (in practice this means using any version
1967 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
1968 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
1969 Historically front end inlining was more extensive than the gcc back end
1970 inlining, but that is no longer the case.
1973 If an object file @file{O} depends on the proper body of a subunit through
1974 inlining or instantiation, it depends on the parent unit of the subunit.
1975 This means that any modification of the parent unit or one of its subunits
1976 affects the compilation of @file{O}.
1979 The object file for a parent unit depends on all its subunit body files.
1982 The previous two rules meant that for purposes of computing dependencies and
1983 recompilation, a body and all its subunits are treated as an indivisible whole.
1986 These rules are applied transitively: if unit @code{A} @code{with}'s
1987 unit @code{B}, whose elaboration calls an inlined procedure in package
1988 @code{C}, the object file for unit @code{A} will depend on the body of
1989 @code{C}, in file @file{c.adb}.
1991 The set of dependent files described by these rules includes all the
1992 files on which the unit is semantically dependent, as dictated by the
1993 Ada language standard. However, it is a superset of what the
1994 standard describes, because it includes generic, inline, and subunit
1997 An object file must be recreated by recompiling the corresponding source
1998 file if any of the source files on which it depends are modified. For
1999 example, if the @code{make} utility is used to control compilation,
2000 the rule for an Ada object file must mention all the source files on
2001 which the object file depends, according to the above definition.
2002 The determination of the necessary
2003 recompilations is done automatically when one uses @command{gnatmake}.
2006 @node The Ada Library Information Files
2007 @section The Ada Library Information Files
2008 @cindex Ada Library Information files
2009 @cindex @file{ALI} files
2012 Each compilation actually generates two output files. The first of these
2013 is the normal object file that has a @file{.o} extension. The second is a
2014 text file containing full dependency information. It has the same
2015 name as the source file, but an @file{.ali} extension.
2016 This file is known as the Ada Library Information (@file{ALI}) file.
2017 The following information is contained in the @file{ALI} file.
2021 Version information (indicates which version of GNAT was used to compile
2022 the unit(s) in question)
2025 Main program information (including priority and time slice settings,
2026 as well as the wide character encoding used during compilation).
2029 List of arguments used in the @command{gcc} command for the compilation
2032 Attributes of the unit, including configuration pragmas used, an indication
2033 of whether the compilation was successful, exception model used etc.
2036 A list of relevant restrictions applying to the unit (used for consistency)
2040 Categorization information (e.g.@: use of pragma @code{Pure}).
2043 Information on all @code{with}'ed units, including presence of
2044 @code{Elaborate} or @code{Elaborate_All} pragmas.
2047 Information from any @code{Linker_Options} pragmas used in the unit
2050 Information on the use of @code{Body_Version} or @code{Version}
2051 attributes in the unit.
2054 Dependency information. This is a list of files, together with
2055 time stamp and checksum information. These are files on which
2056 the unit depends in the sense that recompilation is required
2057 if any of these units are modified.
2060 Cross-reference data. Contains information on all entities referenced
2061 in the unit. Used by tools like @code{gnatxref} and @code{gnatfind} to
2062 provide cross-reference information.
2067 For a full detailed description of the format of the @file{ALI} file,
2068 see the source of the body of unit @code{Lib.Writ}, contained in file
2069 @file{lib-writ.adb} in the GNAT compiler sources.
2071 @node Binding an Ada Program
2072 @section Binding an Ada Program
2075 When using languages such as C and C++, once the source files have been
2076 compiled the only remaining step in building an executable program
2077 is linking the object modules together. This means that it is possible to
2078 link an inconsistent version of a program, in which two units have
2079 included different versions of the same header.
2081 The rules of Ada do not permit such an inconsistent program to be built.
2082 For example, if two clients have different versions of the same package,
2083 it is illegal to build a program containing these two clients.
2084 These rules are enforced by the GNAT binder, which also determines an
2085 elaboration order consistent with the Ada rules.
2087 The GNAT binder is run after all the object files for a program have
2088 been created. It is given the name of the main program unit, and from
2089 this it determines the set of units required by the program, by reading the
2090 corresponding ALI files. It generates error messages if the program is
2091 inconsistent or if no valid order of elaboration exists.
2093 If no errors are detected, the binder produces a main program, in Ada by
2094 default, that contains calls to the elaboration procedures of those
2095 compilation unit that require them, followed by
2096 a call to the main program. This Ada program is compiled to generate the
2097 object file for the main program. The name of
2098 the Ada file is @file{b~@var{xxx}.adb} (with the corresponding spec
2099 @file{b~@var{xxx}.ads}) where @var{xxx} is the name of the
2102 Finally, the linker is used to build the resulting executable program,
2103 using the object from the main program from the bind step as well as the
2104 object files for the Ada units of the program.
2106 @node Mixed Language Programming
2107 @section Mixed Language Programming
2108 @cindex Mixed Language Programming
2111 This section describes how to develop a mixed-language program,
2112 specifically one that comprises units in both Ada and C.
2115 * Interfacing to C::
2116 * Calling Conventions::
2119 @node Interfacing to C
2120 @subsection Interfacing to C
2122 Interfacing Ada with a foreign language such as C involves using
2123 compiler directives to import and/or export entity definitions in each
2124 language---using @code{extern} statements in C, for instance, and the
2125 @code{Import}, @code{Export}, and @code{Convention} pragmas in Ada.
2126 A full treatment of these topics is provided in Appendix B, section 1
2127 of the Ada Reference Manual.
2129 There are two ways to build a program using GNAT that contains some Ada
2130 sources and some foreign language sources, depending on whether or not
2131 the main subprogram is written in Ada. Here is a source example with
2132 the main subprogram in Ada:
2138 void print_num (int num)
2140 printf ("num is %d.\n", num);
2146 /* num_from_Ada is declared in my_main.adb */
2147 extern int num_from_Ada;
2151 return num_from_Ada;
2155 @smallexample @c ada
2157 procedure My_Main is
2159 -- Declare then export an Integer entity called num_from_Ada
2160 My_Num : Integer := 10;
2161 pragma Export (C, My_Num, "num_from_Ada");
2163 -- Declare an Ada function spec for Get_Num, then use
2164 -- C function get_num for the implementation.
2165 function Get_Num return Integer;
2166 pragma Import (C, Get_Num, "get_num");
2168 -- Declare an Ada procedure spec for Print_Num, then use
2169 -- C function print_num for the implementation.
2170 procedure Print_Num (Num : Integer);
2171 pragma Import (C, Print_Num, "print_num");
2174 Print_Num (Get_Num);
2180 To build this example, first compile the foreign language files to
2181 generate object files:
2183 ^gcc -c file1.c^gcc -c FILE1.C^
2184 ^gcc -c file2.c^gcc -c FILE2.C^
2188 Then, compile the Ada units to produce a set of object files and ALI
2191 gnatmake ^-c^/ACTIONS=COMPILE^ my_main.adb
2195 Run the Ada binder on the Ada main program:
2197 gnatbind my_main.ali
2201 Link the Ada main program, the Ada objects and the other language
2204 gnatlink my_main.ali file1.o file2.o
2208 The last three steps can be grouped in a single command:
2210 gnatmake my_main.adb -largs file1.o file2.o
2213 @cindex Binder output file
2215 If the main program is in a language other than Ada, then you may have
2216 more than one entry point into the Ada subsystem. You must use a special
2217 binder option to generate callable routines that initialize and
2218 finalize the Ada units (@pxref{Binding with Non-Ada Main Programs}).
2219 Calls to the initialization and finalization routines must be inserted
2220 in the main program, or some other appropriate point in the code. The
2221 call to initialize the Ada units must occur before the first Ada
2222 subprogram is called, and the call to finalize the Ada units must occur
2223 after the last Ada subprogram returns. The binder will place the
2224 initialization and finalization subprograms into the
2225 @file{b~@var{xxx}.adb} file where they can be accessed by your C
2226 sources. To illustrate, we have the following example:
2230 extern void adainit (void);
2231 extern void adafinal (void);
2232 extern int add (int, int);
2233 extern int sub (int, int);
2235 int main (int argc, char *argv[])
2241 /* Should print "21 + 7 = 28" */
2242 printf ("%d + %d = %d\n", a, b, add (a, b));
2243 /* Should print "21 - 7 = 14" */
2244 printf ("%d - %d = %d\n", a, b, sub (a, b));
2250 @smallexample @c ada
2253 function Add (A, B : Integer) return Integer;
2254 pragma Export (C, Add, "add");
2258 package body Unit1 is
2259 function Add (A, B : Integer) return Integer is
2267 function Sub (A, B : Integer) return Integer;
2268 pragma Export (C, Sub, "sub");
2272 package body Unit2 is
2273 function Sub (A, B : Integer) return Integer is
2282 The build procedure for this application is similar to the last
2283 example's. First, compile the foreign language files to generate object
2286 ^gcc -c main.c^gcc -c main.c^
2290 Next, compile the Ada units to produce a set of object files and ALI
2293 gnatmake ^-c^/ACTIONS=COMPILE^ unit1.adb
2294 gnatmake ^-c^/ACTIONS=COMPILE^ unit2.adb
2298 Run the Ada binder on every generated ALI file. Make sure to use the
2299 @option{-n} option to specify a foreign main program:
2301 gnatbind ^-n^/NOMAIN^ unit1.ali unit2.ali
2305 Link the Ada main program, the Ada objects and the foreign language
2306 objects. You need only list the last ALI file here:
2308 gnatlink unit2.ali main.o -o exec_file
2311 This procedure yields a binary executable called @file{exec_file}.
2315 Depending on the circumstances (for example when your non-Ada main object
2316 does not provide symbol @code{main}), you may also need to instruct the
2317 GNAT linker not to include the standard startup objects by passing the
2318 @option{^-nostartfiles^/NOSTART_FILES^} switch to @command{gnatlink}.
2320 @node Calling Conventions
2321 @subsection Calling Conventions
2322 @cindex Foreign Languages
2323 @cindex Calling Conventions
2324 GNAT follows standard calling sequence conventions and will thus interface
2325 to any other language that also follows these conventions. The following
2326 Convention identifiers are recognized by GNAT:
2329 @cindex Interfacing to Ada
2330 @cindex Other Ada compilers
2331 @cindex Convention Ada
2333 This indicates that the standard Ada calling sequence will be
2334 used and all Ada data items may be passed without any limitations in the
2335 case where GNAT is used to generate both the caller and callee. It is also
2336 possible to mix GNAT generated code and code generated by another Ada
2337 compiler. In this case, the data types should be restricted to simple
2338 cases, including primitive types. Whether complex data types can be passed
2339 depends on the situation. Probably it is safe to pass simple arrays, such
2340 as arrays of integers or floats. Records may or may not work, depending
2341 on whether both compilers lay them out identically. Complex structures
2342 involving variant records, access parameters, tasks, or protected types,
2343 are unlikely to be able to be passed.
2345 Note that in the case of GNAT running
2346 on a platform that supports HP Ada 83, a higher degree of compatibility
2347 can be guaranteed, and in particular records are laid out in an identical
2348 manner in the two compilers. Note also that if output from two different
2349 compilers is mixed, the program is responsible for dealing with elaboration
2350 issues. Probably the safest approach is to write the main program in the
2351 version of Ada other than GNAT, so that it takes care of its own elaboration
2352 requirements, and then call the GNAT-generated adainit procedure to ensure
2353 elaboration of the GNAT components. Consult the documentation of the other
2354 Ada compiler for further details on elaboration.
2356 However, it is not possible to mix the tasking run time of GNAT and
2357 HP Ada 83, All the tasking operations must either be entirely within
2358 GNAT compiled sections of the program, or entirely within HP Ada 83
2359 compiled sections of the program.
2361 @cindex Interfacing to Assembly
2362 @cindex Convention Assembler
2364 Specifies assembler as the convention. In practice this has the
2365 same effect as convention Ada (but is not equivalent in the sense of being
2366 considered the same convention).
2368 @cindex Convention Asm
2371 Equivalent to Assembler.
2373 @cindex Interfacing to COBOL
2374 @cindex Convention COBOL
2377 Data will be passed according to the conventions described
2378 in section B.4 of the Ada Reference Manual.
2381 @cindex Interfacing to C
2382 @cindex Convention C
2384 Data will be passed according to the conventions described
2385 in section B.3 of the Ada Reference Manual.
2387 A note on interfacing to a C ``varargs'' function:
2388 @findex C varargs function
2389 @cindex Interfacing to C varargs function
2390 @cindex varargs function interfaces
2394 In C, @code{varargs} allows a function to take a variable number of
2395 arguments. There is no direct equivalent in this to Ada. One
2396 approach that can be used is to create a C wrapper for each
2397 different profile and then interface to this C wrapper. For
2398 example, to print an @code{int} value using @code{printf},
2399 create a C function @code{printfi} that takes two arguments, a
2400 pointer to a string and an int, and calls @code{printf}.
2401 Then in the Ada program, use pragma @code{Import} to
2402 interface to @code{printfi}.
2405 It may work on some platforms to directly interface to
2406 a @code{varargs} function by providing a specific Ada profile
2407 for a particular call. However, this does not work on
2408 all platforms, since there is no guarantee that the
2409 calling sequence for a two argument normal C function
2410 is the same as for calling a @code{varargs} C function with
2411 the same two arguments.
2414 @cindex Convention Default
2419 @cindex Convention External
2426 @cindex Interfacing to C++
2427 @cindex Convention C++
2428 @item C_Plus_Plus (or CPP)
2429 This stands for C++. For most purposes this is identical to C.
2430 See the separate description of the specialized GNAT pragmas relating to
2431 C++ interfacing for further details.
2435 @cindex Interfacing to Fortran
2436 @cindex Convention Fortran
2438 Data will be passed according to the conventions described
2439 in section B.5 of the Ada Reference Manual.
2442 This applies to an intrinsic operation, as defined in the Ada
2443 Reference Manual. If a pragma Import (Intrinsic) applies to a subprogram,
2444 this means that the body of the subprogram is provided by the compiler itself,
2445 usually by means of an efficient code sequence, and that the user does not
2446 supply an explicit body for it. In an application program, the pragma may
2447 be applied to the following sets of names:
2451 Rotate_Left, Rotate_Right, Shift_Left, Shift_Right,
2452 Shift_Right_Arithmetic. The corresponding subprogram declaration must have
2453 two formal parameters. The
2454 first one must be a signed integer type or a modular type with a binary
2455 modulus, and the second parameter must be of type Natural.
2456 The return type must be the same as the type of the first argument. The size
2457 of this type can only be 8, 16, 32, or 64.
2460 Binary arithmetic operators: ``+'', ``-'', ``*'', ``/''
2461 The corresponding operator declaration must have parameters and result type
2462 that have the same root numeric type (for example, all three are long_float
2463 types). This simplifies the definition of operations that use type checking
2464 to perform dimensional checks:
2466 @smallexample @c ada
2467 type Distance is new Long_Float;
2468 type Time is new Long_Float;
2469 type Velocity is new Long_Float;
2470 function "/" (D : Distance; T : Time)
2472 pragma Import (Intrinsic, "/");
2476 This common idiom is often programmed with a generic definition and an
2477 explicit body. The pragma makes it simpler to introduce such declarations.
2478 It incurs no overhead in compilation time or code size, because it is
2479 implemented as a single machine instruction.
2482 General subprogram entities, to bind an Ada subprogram declaration to
2483 a compiler builtin by name with back-ends where such interfaces are
2484 available. A typical example is the set of ``__builtin'' functions
2485 exposed by the GCC back-end, as in the following example:
2487 @smallexample @c ada
2488 function builtin_sqrt (F : Float) return Float;
2489 pragma Import (Intrinsic, builtin_sqrt, "__builtin_sqrtf");
2492 Most of the GCC builtins are accessible this way, and as for other
2493 import conventions (e.g. C), it is the user's responsibility to ensure
2494 that the Ada subprogram profile matches the underlying builtin
2502 @cindex Convention Stdcall
2504 This is relevant only to Windows XP/2000/NT implementations of GNAT,
2505 and specifies that the @code{Stdcall} calling sequence will be used,
2506 as defined by the NT API. Nevertheless, to ease building
2507 cross-platform bindings this convention will be handled as a @code{C} calling
2508 convention on non-Windows platforms.
2511 @cindex Convention DLL
2513 This is equivalent to @code{Stdcall}.
2516 @cindex Convention Win32
2518 This is equivalent to @code{Stdcall}.
2522 @cindex Convention Stubbed
2524 This is a special convention that indicates that the compiler
2525 should provide a stub body that raises @code{Program_Error}.
2529 GNAT additionally provides a useful pragma @code{Convention_Identifier}
2530 that can be used to parameterize conventions and allow additional synonyms
2531 to be specified. For example if you have legacy code in which the convention
2532 identifier Fortran77 was used for Fortran, you can use the configuration
2535 @smallexample @c ada
2536 pragma Convention_Identifier (Fortran77, Fortran);
2540 And from now on the identifier Fortran77 may be used as a convention
2541 identifier (for example in an @code{Import} pragma) with the same
2545 @node Building Mixed Ada & C++ Programs
2546 @section Building Mixed Ada and C++ Programs
2549 A programmer inexperienced with mixed-language development may find that
2550 building an application containing both Ada and C++ code can be a
2551 challenge. This section gives a few
2552 hints that should make this task easier. The first section addresses
2553 the differences between interfacing with C and interfacing with C++.
2555 looks into the delicate problem of linking the complete application from
2556 its Ada and C++ parts. The last section gives some hints on how the GNAT
2557 run-time library can be adapted in order to allow inter-language dispatching
2558 with a new C++ compiler.
2561 * Interfacing to C++::
2562 * Linking a Mixed C++ & Ada Program::
2563 * A Simple Example::
2564 * Interfacing with C++ constructors::
2565 * Interfacing with C++ at the Class Level::
2568 @node Interfacing to C++
2569 @subsection Interfacing to C++
2572 GNAT supports interfacing with the G++ compiler (or any C++ compiler
2573 generating code that is compatible with the G++ Application Binary
2574 Interface ---see http://www.codesourcery.com/archives/cxx-abi).
2577 Interfacing can be done at 3 levels: simple data, subprograms, and
2578 classes. In the first two cases, GNAT offers a specific @code{Convention
2579 C_Plus_Plus} (or @code{CPP}) that behaves exactly like @code{Convention C}.
2580 Usually, C++ mangles the names of subprograms. To generate proper mangled
2581 names automatically, see @ref{Generating Ada Bindings for C and C++ headers}).
2582 This problem can also be addressed manually in two ways:
2586 by modifying the C++ code in order to force a C convention using
2587 the @code{extern "C"} syntax.
2590 by figuring out the mangled name (using e.g. @command{nm}) and using it as the
2591 Link_Name argument of the pragma import.
2595 Interfacing at the class level can be achieved by using the GNAT specific
2596 pragmas such as @code{CPP_Constructor}. @xref{Interfacing to C++,,,
2597 gnat_rm, GNAT Reference Manual}, for additional information.
2599 @node Linking a Mixed C++ & Ada Program
2600 @subsection Linking a Mixed C++ & Ada Program
2603 Usually the linker of the C++ development system must be used to link
2604 mixed applications because most C++ systems will resolve elaboration
2605 issues (such as calling constructors on global class instances)
2606 transparently during the link phase. GNAT has been adapted to ease the
2607 use of a foreign linker for the last phase. Three cases can be
2612 Using GNAT and G++ (GNU C++ compiler) from the same GCC installation:
2613 The C++ linker can simply be called by using the C++ specific driver
2616 Note that if the C++ code uses inline functions, you will need to
2617 compile your C++ code with the @code{-fkeep-inline-functions} switch in
2618 order to provide an existing function implementation that the Ada code can
2622 $ g++ -c -fkeep-inline-functions file1.C
2623 $ g++ -c -fkeep-inline-functions file2.C
2624 $ gnatmake ada_unit -largs file1.o file2.o --LINK=g++
2628 Using GNAT and G++ from two different GCC installations: If both
2629 compilers are on the @env{PATH}, the previous method may be used. It is
2630 important to note that environment variables such as
2631 @env{C_INCLUDE_PATH}, @env{GCC_EXEC_PREFIX}, @env{BINUTILS_ROOT}, and
2632 @env{GCC_ROOT} will affect both compilers
2633 at the same time and may make one of the two compilers operate
2634 improperly if set during invocation of the wrong compiler. It is also
2635 very important that the linker uses the proper @file{libgcc.a} GCC
2636 library -- that is, the one from the C++ compiler installation. The
2637 implicit link command as suggested in the @command{gnatmake} command
2638 from the former example can be replaced by an explicit link command with
2639 the full-verbosity option in order to verify which library is used:
2642 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=c++
2644 If there is a problem due to interfering environment variables, it can
2645 be worked around by using an intermediate script. The following example
2646 shows the proper script to use when GNAT has not been installed at its
2647 default location and g++ has been installed at its default location:
2655 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=./my_script
2659 Using a non-GNU C++ compiler: The commands previously described can be
2660 used to insure that the C++ linker is used. Nonetheless, you need to add
2661 a few more parameters to the link command line, depending on the exception
2664 If the @code{setjmp/longjmp} exception mechanism is used, only the paths
2665 to the libgcc libraries are required:
2670 CC $* `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a`
2671 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
2674 Where CC is the name of the non-GNU C++ compiler.
2676 If the @code{zero cost} exception mechanism is used, and the platform
2677 supports automatic registration of exception tables (e.g.@: Solaris),
2678 paths to more objects are required:
2683 CC `gcc -print-file-name=crtbegin.o` $* \
2684 `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a` \
2685 `gcc -print-file-name=crtend.o`
2686 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
2689 If the @code{zero cost} exception mechanism is used, and the platform
2690 doesn't support automatic registration of exception tables (e.g.@: HP-UX
2691 or AIX), the simple approach described above will not work and
2692 a pre-linking phase using GNAT will be necessary.
2696 Another alternative is to use the @command{gprbuild} multi-language builder
2697 which has a large knowledge base and knows how to link Ada and C++ code
2698 together automatically in most cases.
2700 @node A Simple Example
2701 @subsection A Simple Example
2703 The following example, provided as part of the GNAT examples, shows how
2704 to achieve procedural interfacing between Ada and C++ in both
2705 directions. The C++ class A has two methods. The first method is exported
2706 to Ada by the means of an extern C wrapper function. The second method
2707 calls an Ada subprogram. On the Ada side, The C++ calls are modelled by
2708 a limited record with a layout comparable to the C++ class. The Ada
2709 subprogram, in turn, calls the C++ method. So, starting from the C++
2710 main program, the process passes back and forth between the two
2714 Here are the compilation commands:
2716 $ gnatmake -c simple_cpp_interface
2719 $ gnatbind -n simple_cpp_interface
2720 $ gnatlink simple_cpp_interface -o cpp_main --LINK=g++
2721 -lstdc++ ex7.o cpp_main.o
2725 Here are the corresponding sources:
2733 void adainit (void);
2734 void adafinal (void);
2735 void method1 (A *t);
2757 class A : public Origin @{
2759 void method1 (void);
2760 void method2 (int v);
2770 extern "C" @{ void ada_method2 (A *t, int v);@}
2772 void A::method1 (void)
2775 printf ("in A::method1, a_value = %d \n",a_value);
2779 void A::method2 (int v)
2781 ada_method2 (this, v);
2782 printf ("in A::method2, a_value = %d \n",a_value);
2789 printf ("in A::A, a_value = %d \n",a_value);
2793 @smallexample @c ada
2795 package body Simple_Cpp_Interface is
2797 procedure Ada_Method2 (This : in out A; V : Integer) is
2803 end Simple_Cpp_Interface;
2806 package Simple_Cpp_Interface is
2809 Vptr : System.Address;
2813 pragma Convention (C, A);
2815 procedure Method1 (This : in out A);
2816 pragma Import (C, Method1);
2818 procedure Ada_Method2 (This : in out A; V : Integer);
2819 pragma Export (C, Ada_Method2);
2821 end Simple_Cpp_Interface;
2824 @node Interfacing with C++ constructors
2825 @subsection Interfacing with C++ constructors
2828 In order to interface with C++ constructors GNAT provides the
2829 @code{pragma CPP_Constructor} (@xref{Interfacing to C++,,,
2830 gnat_rm, GNAT Reference Manual}, for additional information).
2831 In this section we present some common uses of C++ constructors
2832 in mixed-languages programs in GNAT.
2834 Let us assume that we need to interface with the following
2842 @b{virtual} int Get_Value ();
2843 Root(); // Default constructor
2844 Root(int v); // 1st non-default constructor
2845 Root(int v, int w); // 2nd non-default constructor
2849 For this purpose we can write the following package spec (further
2850 information on how to build this spec is available in
2851 @ref{Interfacing with C++ at the Class Level} and
2852 @ref{Generating Ada Bindings for C and C++ headers}).
2854 @smallexample @c ada
2855 with Interfaces.C; use Interfaces.C;
2857 type Root is tagged limited record
2861 pragma Import (CPP, Root);
2863 function Get_Value (Obj : Root) return int;
2864 pragma Import (CPP, Get_Value);
2866 function Constructor return Root;
2867 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ev");
2869 function Constructor (v : Integer) return Root;
2870 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ei");
2872 function Constructor (v, w : Integer) return Root;
2873 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Eii");
2877 On the Ada side the constructor is represented by a function (whose
2878 name is arbitrary) that returns the classwide type corresponding to
2879 the imported C++ class. Although the constructor is described as a
2880 function, it is typically a procedure with an extra implicit argument
2881 (the object being initialized) at the implementation level. GNAT
2882 issues the appropriate call, whatever it is, to get the object
2883 properly initialized.
2885 Constructors can only appear in the following contexts:
2889 On the right side of an initialization of an object of type @var{T}.
2891 On the right side of an initialization of a record component of type @var{T}.
2893 In an Ada 2005 limited aggregate.
2895 In an Ada 2005 nested limited aggregate.
2897 In an Ada 2005 limited aggregate that initializes an object built in
2898 place by an extended return statement.
2902 In a declaration of an object whose type is a class imported from C++,
2903 either the default C++ constructor is implicitly called by GNAT, or
2904 else the required C++ constructor must be explicitly called in the
2905 expression that initializes the object. For example:
2907 @smallexample @c ada
2909 Obj2 : Root := Constructor;
2910 Obj3 : Root := Constructor (v => 10);
2911 Obj4 : Root := Constructor (30, 40);
2914 The first two declarations are equivalent: in both cases the default C++
2915 constructor is invoked (in the former case the call to the constructor is
2916 implicit, and in the latter case the call is explicit in the object
2917 declaration). @code{Obj3} is initialized by the C++ non-default constructor
2918 that takes an integer argument, and @code{Obj4} is initialized by the
2919 non-default C++ constructor that takes two integers.
2921 Let us derive the imported C++ class in the Ada side. For example:
2923 @smallexample @c ada
2924 type DT is new Root with record
2925 C_Value : Natural := 2009;
2929 In this case the components DT inherited from the C++ side must be
2930 initialized by a C++ constructor, and the additional Ada components
2931 of type DT are initialized by GNAT. The initialization of such an
2932 object is done either by default, or by means of a function returning
2933 an aggregate of type DT, or by means of an extension aggregate.
2935 @smallexample @c ada
2937 Obj6 : DT := Function_Returning_DT (50);
2938 Obj7 : DT := (Constructor (30,40) with C_Value => 50);
2941 The declaration of @code{Obj5} invokes the default constructors: the
2942 C++ default constructor of the parent type takes care of the initialization
2943 of the components inherited from Root, and GNAT takes care of the default
2944 initialization of the additional Ada components of type DT (that is,
2945 @code{C_Value} is initialized to value 2009). The order of invocation of
2946 the constructors is consistent with the order of elaboration required by
2947 Ada and C++. That is, the constructor of the parent type is always called
2948 before the constructor of the derived type.
2950 Let us now consider a record that has components whose type is imported
2951 from C++. For example:
2953 @smallexample @c ada
2954 type Rec1 is limited record
2955 Data1 : Root := Constructor (10);
2956 Value : Natural := 1000;
2959 type Rec2 (D : Integer := 20) is limited record
2961 Data2 : Root := Constructor (D, 30);
2965 The initialization of an object of type @code{Rec2} will call the
2966 non-default C++ constructors specified for the imported components.
2969 @smallexample @c ada
2973 Using Ada 2005 we can use limited aggregates to initialize an object
2974 invoking C++ constructors that differ from those specified in the type
2975 declarations. For example:
2977 @smallexample @c ada
2978 Obj9 : Rec2 := (Rec => (Data1 => Constructor (15, 16),
2983 The above declaration uses an Ada 2005 limited aggregate to
2984 initialize @code{Obj9}, and the C++ constructor that has two integer
2985 arguments is invoked to initialize the @code{Data1} component instead
2986 of the constructor specified in the declaration of type @code{Rec1}. In
2987 Ada 2005 the box in the aggregate indicates that unspecified components
2988 are initialized using the expression (if any) available in the component
2989 declaration. That is, in this case discriminant @code{D} is initialized
2990 to value @code{20}, @code{Value} is initialized to value 1000, and the
2991 non-default C++ constructor that handles two integers takes care of
2992 initializing component @code{Data2} with values @code{20,30}.
2994 In Ada 2005 we can use the extended return statement to build the Ada
2995 equivalent to C++ non-default constructors. For example:
2997 @smallexample @c ada
2998 function Constructor (V : Integer) return Rec2 is
3000 return Obj : Rec2 := (Rec => (Data1 => Constructor (V, 20),
3003 -- Further actions required for construction of
3004 -- objects of type Rec2
3010 In this example the extended return statement construct is used to
3011 build in place the returned object whose components are initialized
3012 by means of a limited aggregate. Any further action associated with
3013 the constructor can be placed inside the construct.
3015 @node Interfacing with C++ at the Class Level
3016 @subsection Interfacing with C++ at the Class Level
3018 In this section we demonstrate the GNAT features for interfacing with
3019 C++ by means of an example making use of Ada 2005 abstract interface
3020 types. This example consists of a classification of animals; classes
3021 have been used to model our main classification of animals, and
3022 interfaces provide support for the management of secondary
3023 classifications. We first demonstrate a case in which the types and
3024 constructors are defined on the C++ side and imported from the Ada
3025 side, and latter the reverse case.
3027 The root of our derivation will be the @code{Animal} class, with a
3028 single private attribute (the @code{Age} of the animal) and two public
3029 primitives to set and get the value of this attribute.
3034 @b{virtual} void Set_Age (int New_Age);
3035 @b{virtual} int Age ();
3041 Abstract interface types are defined in C++ by means of classes with pure
3042 virtual functions and no data members. In our example we will use two
3043 interfaces that provide support for the common management of @code{Carnivore}
3044 and @code{Domestic} animals:
3047 @b{class} Carnivore @{
3049 @b{virtual} int Number_Of_Teeth () = 0;
3052 @b{class} Domestic @{
3054 @b{virtual void} Set_Owner (char* Name) = 0;
3058 Using these declarations, we can now say that a @code{Dog} is an animal that is
3059 both Carnivore and Domestic, that is:
3062 @b{class} Dog : Animal, Carnivore, Domestic @{
3064 @b{virtual} int Number_Of_Teeth ();
3065 @b{virtual} void Set_Owner (char* Name);
3067 Dog(); // Constructor
3074 In the following examples we will assume that the previous declarations are
3075 located in a file named @code{animals.h}. The following package demonstrates
3076 how to import these C++ declarations from the Ada side:
3078 @smallexample @c ada
3079 with Interfaces.C.Strings; use Interfaces.C.Strings;
3081 type Carnivore is interface;
3082 pragma Convention (C_Plus_Plus, Carnivore);
3083 function Number_Of_Teeth (X : Carnivore)
3084 return Natural is abstract;
3086 type Domestic is interface;
3087 pragma Convention (C_Plus_Plus, Set_Owner);
3089 (X : in out Domestic;
3090 Name : Chars_Ptr) is abstract;
3092 type Animal is tagged record
3095 pragma Import (C_Plus_Plus, Animal);
3097 procedure Set_Age (X : in out Animal; Age : Integer);
3098 pragma Import (C_Plus_Plus, Set_Age);
3100 function Age (X : Animal) return Integer;
3101 pragma Import (C_Plus_Plus, Age);
3103 type Dog is new Animal and Carnivore and Domestic with record
3104 Tooth_Count : Natural;
3105 Owner : String (1 .. 30);
3107 pragma Import (C_Plus_Plus, Dog);
3109 function Number_Of_Teeth (A : Dog) return Integer;
3110 pragma Import (C_Plus_Plus, Number_Of_Teeth);
3112 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3113 pragma Import (C_Plus_Plus, Set_Owner);
3115 function New_Dog return Dog;
3116 pragma CPP_Constructor (New_Dog);
3117 pragma Import (CPP, New_Dog, "_ZN3DogC2Ev");
3121 Thanks to the compatibility between GNAT run-time structures and the C++ ABI,
3122 interfacing with these C++ classes is easy. The only requirement is that all
3123 the primitives and components must be declared exactly in the same order in
3126 Regarding the abstract interfaces, we must indicate to the GNAT compiler by
3127 means of a @code{pragma Convention (C_Plus_Plus)}, the convention used to pass
3128 the arguments to the called primitives will be the same as for C++. For the
3129 imported classes we use @code{pragma Import} with convention @code{C_Plus_Plus}
3130 to indicate that they have been defined on the C++ side; this is required
3131 because the dispatch table associated with these tagged types will be built
3132 in the C++ side and therefore will not contain the predefined Ada primitives
3133 which Ada would otherwise expect.
3135 As the reader can see there is no need to indicate the C++ mangled names
3136 associated with each subprogram because it is assumed that all the calls to
3137 these primitives will be dispatching calls. The only exception is the
3138 constructor, which must be registered with the compiler by means of
3139 @code{pragma CPP_Constructor} and needs to provide its associated C++
3140 mangled name because the Ada compiler generates direct calls to it.
3142 With the above packages we can now declare objects of type Dog on the Ada side
3143 and dispatch calls to the corresponding subprograms on the C++ side. We can
3144 also extend the tagged type Dog with further fields and primitives, and
3145 override some of its C++ primitives on the Ada side. For example, here we have
3146 a type derivation defined on the Ada side that inherits all the dispatching
3147 primitives of the ancestor from the C++ side.
3150 @b{with} Animals; @b{use} Animals;
3151 @b{package} Vaccinated_Animals @b{is}
3152 @b{type} Vaccinated_Dog @b{is new} Dog @b{with null record};
3153 @b{function} Vaccination_Expired (A : Vaccinated_Dog) @b{return} Boolean;
3154 @b{end} Vaccinated_Animals;
3157 It is important to note that, because of the ABI compatibility, the programmer
3158 does not need to add any further information to indicate either the object
3159 layout or the dispatch table entry associated with each dispatching operation.
3161 Now let us define all the types and constructors on the Ada side and export
3162 them to C++, using the same hierarchy of our previous example:
3164 @smallexample @c ada
3165 with Interfaces.C.Strings;
3166 use Interfaces.C.Strings;
3168 type Carnivore is interface;
3169 pragma Convention (C_Plus_Plus, Carnivore);
3170 function Number_Of_Teeth (X : Carnivore)
3171 return Natural is abstract;
3173 type Domestic is interface;
3174 pragma Convention (C_Plus_Plus, Set_Owner);
3176 (X : in out Domestic;
3177 Name : Chars_Ptr) is abstract;
3179 type Animal is tagged record
3182 pragma Convention (C_Plus_Plus, Animal);
3184 procedure Set_Age (X : in out Animal; Age : Integer);
3185 pragma Export (C_Plus_Plus, Set_Age);
3187 function Age (X : Animal) return Integer;
3188 pragma Export (C_Plus_Plus, Age);
3190 type Dog is new Animal and Carnivore and Domestic with record
3191 Tooth_Count : Natural;
3192 Owner : String (1 .. 30);
3194 pragma Convention (C_Plus_Plus, Dog);
3196 function Number_Of_Teeth (A : Dog) return Integer;
3197 pragma Export (C_Plus_Plus, Number_Of_Teeth);
3199 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3200 pragma Export (C_Plus_Plus, Set_Owner);
3202 function New_Dog return Dog'Class;
3203 pragma Export (C_Plus_Plus, New_Dog);
3207 Compared with our previous example the only difference is the use of
3208 @code{pragma Export} to indicate to the GNAT compiler that the primitives will
3209 be available to C++. Thanks to the ABI compatibility, on the C++ side there is
3210 nothing else to be done; as explained above, the only requirement is that all
3211 the primitives and components are declared in exactly the same order.
3213 For completeness, let us see a brief C++ main program that uses the
3214 declarations available in @code{animals.h} (presented in our first example) to
3215 import and use the declarations from the Ada side, properly initializing and
3216 finalizing the Ada run-time system along the way:
3219 @b{#include} "animals.h"
3220 @b{#include} <iostream>
3221 @b{using namespace} std;
3223 void Check_Carnivore (Carnivore *obj) @{@dots{}@}
3224 void Check_Domestic (Domestic *obj) @{@dots{}@}
3225 void Check_Animal (Animal *obj) @{@dots{}@}
3226 void Check_Dog (Dog *obj) @{@dots{}@}
3229 void adainit (void);
3230 void adafinal (void);
3236 Dog *obj = new_dog(); // Ada constructor
3237 Check_Carnivore (obj); // Check secondary DT
3238 Check_Domestic (obj); // Check secondary DT
3239 Check_Animal (obj); // Check primary DT
3240 Check_Dog (obj); // Check primary DT
3245 adainit (); test(); adafinal ();
3250 @node Comparison between GNAT and C/C++ Compilation Models
3251 @section Comparison between GNAT and C/C++ Compilation Models
3254 The GNAT model of compilation is close to the C and C++ models. You can
3255 think of Ada specs as corresponding to header files in C. As in C, you
3256 don't need to compile specs; they are compiled when they are used. The
3257 Ada @code{with} is similar in effect to the @code{#include} of a C
3260 One notable difference is that, in Ada, you may compile specs separately
3261 to check them for semantic and syntactic accuracy. This is not always
3262 possible with C headers because they are fragments of programs that have
3263 less specific syntactic or semantic rules.
3265 The other major difference is the requirement for running the binder,
3266 which performs two important functions. First, it checks for
3267 consistency. In C or C++, the only defense against assembling
3268 inconsistent programs lies outside the compiler, in a makefile, for
3269 example. The binder satisfies the Ada requirement that it be impossible
3270 to construct an inconsistent program when the compiler is used in normal
3273 @cindex Elaboration order control
3274 The other important function of the binder is to deal with elaboration
3275 issues. There are also elaboration issues in C++ that are handled
3276 automatically. This automatic handling has the advantage of being
3277 simpler to use, but the C++ programmer has no control over elaboration.
3278 Where @code{gnatbind} might complain there was no valid order of
3279 elaboration, a C++ compiler would simply construct a program that
3280 malfunctioned at run time.
3283 @node Comparison between GNAT and Conventional Ada Library Models
3284 @section Comparison between GNAT and Conventional Ada Library Models
3287 This section is intended for Ada programmers who have
3288 used an Ada compiler implementing the traditional Ada library
3289 model, as described in the Ada Reference Manual.
3291 @cindex GNAT library
3292 In GNAT, there is no ``library'' in the normal sense. Instead, the set of
3293 source files themselves acts as the library. Compiling Ada programs does
3294 not generate any centralized information, but rather an object file and
3295 a ALI file, which are of interest only to the binder and linker.
3296 In a traditional system, the compiler reads information not only from
3297 the source file being compiled, but also from the centralized library.
3298 This means that the effect of a compilation depends on what has been
3299 previously compiled. In particular:
3303 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3304 to the version of the unit most recently compiled into the library.
3307 Inlining is effective only if the necessary body has already been
3308 compiled into the library.
3311 Compiling a unit may obsolete other units in the library.
3315 In GNAT, compiling one unit never affects the compilation of any other
3316 units because the compiler reads only source files. Only changes to source
3317 files can affect the results of a compilation. In particular:
3321 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3322 to the source version of the unit that is currently accessible to the
3327 Inlining requires the appropriate source files for the package or
3328 subprogram bodies to be available to the compiler. Inlining is always
3329 effective, independent of the order in which units are complied.
3332 Compiling a unit never affects any other compilations. The editing of
3333 sources may cause previous compilations to be out of date if they
3334 depended on the source file being modified.
3338 The most important result of these differences is that order of compilation
3339 is never significant in GNAT. There is no situation in which one is
3340 required to do one compilation before another. What shows up as order of
3341 compilation requirements in the traditional Ada library becomes, in
3342 GNAT, simple source dependencies; in other words, there is only a set
3343 of rules saying what source files must be present when a file is
3347 @node Placement of temporary files
3348 @section Placement of temporary files
3349 @cindex Temporary files (user control over placement)
3352 GNAT creates temporary files in the directory designated by the environment
3353 variable @env{TMPDIR}.
3354 (See the HP @emph{C RTL Reference Manual} on the function @code{getenv()}
3355 for detailed information on how environment variables are resolved.
3356 For most users the easiest way to make use of this feature is to simply
3357 define @env{TMPDIR} as a job level logical name).
3358 For example, if you wish to use a Ramdisk (assuming DECRAM is installed)
3359 for compiler temporary files, then you can include something like the
3360 following command in your @file{LOGIN.COM} file:
3363 $ define/job TMPDIR "/disk$scratchram/000000/temp/"
3367 If @env{TMPDIR} is not defined, then GNAT uses the directory designated by
3368 @env{TMP}; if @env{TMP} is not defined, then GNAT uses the directory
3369 designated by @env{TEMP}.
3370 If none of these environment variables are defined then GNAT uses the
3371 directory designated by the logical name @code{SYS$SCRATCH:}
3372 (by default the user's home directory). If all else fails
3373 GNAT uses the current directory for temporary files.
3376 @c *************************
3377 @node Compiling with gcc
3378 @chapter Compiling with @command{gcc}
3381 This chapter discusses how to compile Ada programs using the @command{gcc}
3382 command. It also describes the set of switches
3383 that can be used to control the behavior of the compiler.
3385 * Compiling Programs::
3386 * Switches for gcc::
3387 * Search Paths and the Run-Time Library (RTL)::
3388 * Order of Compilation Issues::
3392 @node Compiling Programs
3393 @section Compiling Programs
3396 The first step in creating an executable program is to compile the units
3397 of the program using the @command{gcc} command. You must compile the
3402 the body file (@file{.adb}) for a library level subprogram or generic
3406 the spec file (@file{.ads}) for a library level package or generic
3407 package that has no body
3410 the body file (@file{.adb}) for a library level package
3411 or generic package that has a body
3416 You need @emph{not} compile the following files
3421 the spec of a library unit which has a body
3428 because they are compiled as part of compiling related units. GNAT
3430 when the corresponding body is compiled, and subunits when the parent is
3433 @cindex cannot generate code
3434 If you attempt to compile any of these files, you will get one of the
3435 following error messages (where @var{fff} is the name of the file you
3439 cannot generate code for file @var{fff} (package spec)
3440 to check package spec, use -gnatc
3442 cannot generate code for file @var{fff} (missing subunits)
3443 to check parent unit, use -gnatc
3445 cannot generate code for file @var{fff} (subprogram spec)
3446 to check subprogram spec, use -gnatc
3448 cannot generate code for file @var{fff} (subunit)
3449 to check subunit, use -gnatc
3453 As indicated by the above error messages, if you want to submit
3454 one of these files to the compiler to check for correct semantics
3455 without generating code, then use the @option{-gnatc} switch.
3457 The basic command for compiling a file containing an Ada unit is
3460 @c $ gcc -c @ovar{switches} @file{file name}
3461 @c Expanding @ovar macro inline (explanation in macro def comments)
3462 $ gcc -c @r{[}@var{switches}@r{]} @file{file name}
3466 where @var{file name} is the name of the Ada file (usually
3468 @file{.ads} for a spec or @file{.adb} for a body).
3471 @option{-c} switch to tell @command{gcc} to compile, but not link, the file.
3473 The result of a successful compilation is an object file, which has the
3474 same name as the source file but an extension of @file{.o} and an Ada
3475 Library Information (ALI) file, which also has the same name as the
3476 source file, but with @file{.ali} as the extension. GNAT creates these
3477 two output files in the current directory, but you may specify a source
3478 file in any directory using an absolute or relative path specification
3479 containing the directory information.
3482 @command{gcc} is actually a driver program that looks at the extensions of
3483 the file arguments and loads the appropriate compiler. For example, the
3484 GNU C compiler is @file{cc1}, and the Ada compiler is @file{gnat1}.
3485 These programs are in directories known to the driver program (in some
3486 configurations via environment variables you set), but need not be in
3487 your path. The @command{gcc} driver also calls the assembler and any other
3488 utilities needed to complete the generation of the required object
3491 It is possible to supply several file names on the same @command{gcc}
3492 command. This causes @command{gcc} to call the appropriate compiler for
3493 each file. For example, the following command lists two separate
3494 files to be compiled:
3497 $ gcc -c x.adb y.adb
3501 calls @code{gnat1} (the Ada compiler) twice to compile @file{x.adb} and
3503 The compiler generates two object files @file{x.o} and @file{y.o}
3504 and the two ALI files @file{x.ali} and @file{y.ali}.
3505 Any switches apply to all the files ^listed,^listed.^
3507 @node Switches for gcc
3508 @section Switches for @command{gcc}
3511 The @command{gcc} command accepts switches that control the
3512 compilation process. These switches are fully described in this section.
3513 First we briefly list all the switches, in alphabetical order, then we
3514 describe the switches in more detail in functionally grouped sections.
3516 More switches exist for GCC than those documented here, especially
3517 for specific targets. However, their use is not recommended as
3518 they may change code generation in ways that are incompatible with
3519 the Ada run-time library, or can cause inconsistencies between
3523 * Output and Error Message Control::
3524 * Warning Message Control::
3525 * Debugging and Assertion Control::
3526 * Validity Checking::
3529 * Using gcc for Syntax Checking::
3530 * Using gcc for Semantic Checking::
3531 * Compiling Different Versions of Ada::
3532 * Character Set Control::
3533 * File Naming Control::
3534 * Subprogram Inlining Control::
3535 * Auxiliary Output Control::
3536 * Debugging Control::
3537 * Exception Handling Control::
3538 * Units to Sources Mapping Files::
3539 * Integrated Preprocessing::
3540 * Code Generation Control::
3549 @cindex @option{-b} (@command{gcc})
3550 @item -b @var{target}
3551 Compile your program to run on @var{target}, which is the name of a
3552 system configuration. You must have a GNAT cross-compiler built if
3553 @var{target} is not the same as your host system.
3556 @cindex @option{-B} (@command{gcc})
3557 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
3558 from @var{dir} instead of the default location. Only use this switch
3559 when multiple versions of the GNAT compiler are available.
3560 @xref{Directory Options,, Options for Directory Search, gcc, Using the
3561 GNU Compiler Collection (GCC)}, for further details. You would normally
3562 use the @option{-b} or @option{-V} switch instead.
3565 @cindex @option{-c} (@command{gcc})
3566 Compile. Always use this switch when compiling Ada programs.
3568 Note: for some other languages when using @command{gcc}, notably in
3569 the case of C and C++, it is possible to use
3570 use @command{gcc} without a @option{-c} switch to
3571 compile and link in one step. In the case of GNAT, you
3572 cannot use this approach, because the binder must be run
3573 and @command{gcc} cannot be used to run the GNAT binder.
3576 @item -fcallgraph-info@r{[}=su,da@r{]}
3577 @cindex @option{-fcallgraph-info} (@command{gcc})
3578 Makes the compiler output callgraph information for the program, on a
3579 per-file basis. The information is generated in the VCG format. It can
3580 be decorated with additional, per-node and/or per-edge information, if a
3581 list of comma-separated markers is additionally specified. When the
3582 @var{su} marker is specified, the callgraph is decorated with stack usage information; it is equivalent to @option{-fstack-usage}. When the @var{da}
3583 marker is specified, the callgraph is decorated with information about
3584 dynamically allocated objects.
3587 @cindex @option{-fdump-scos} (@command{gcc})
3588 Generates SCO (Source Coverage Obligation) information in the ALI file.
3589 This information is used by advanced coverage tools. See unit @file{SCOs}
3590 in the compiler sources for details in files @file{scos.ads} and
3594 @cindex @option{-fdump-xref} (@command{gcc})
3595 Generates cross reference information in GLI files for C and C++ sources.
3596 The GLI files have the same syntax as the ALI files for Ada, and can be used
3597 for source navigation in IDEs and on the command line using e.g. gnatxref
3598 and the @option{--ext=gli} switch.
3600 @item -flto@r{[}=n@r{]}
3601 @cindex @option{-flto} (@command{gcc})
3602 Enables Link Time Optimization. This switch must be used in conjunction
3603 with the traditional @option{-Ox} switches and instructs the compiler to
3604 defer most optimizations until the link stage. The advantage of this
3605 approach is that the compiler can do a whole-program analysis and choose
3606 the best interprocedural optimization strategy based on a complete view
3607 of the program, instead of a fragmentary view with the usual approach.
3608 This can also speed up the compilation of huge programs and reduce the
3609 size of the final executable, compared with a per-unit compilation with
3610 full inlining across modules enabled with the @option{-gnatn2} switch.
3611 The drawback of this approach is that it may require much more memory.
3612 The switch, as well as the accompanying @option{-Ox} switches, must be
3613 specified both for the compilation and the link phases.
3614 If the @var{n} parameter is specified, the optimization and final code
3615 generation at link time are executed using @var{n} parallel jobs by
3616 means of an installed @command{make} program.
3619 @cindex @option{-fno-inline} (@command{gcc})
3620 Suppresses all inlining, even if other optimization or inlining
3621 switches are set. This includes suppression of inlining that
3622 results from the use of the pragma @code{Inline_Always}.
3623 Any occurrences of pragma @code{Inline} or @code{Inline_Always}
3624 are ignored, and @option{-gnatn} and @option{-gnatN} have no
3625 effects if this switch is present. Note that inlining can also
3626 be suppressed on a finer-grained basis with pragma @code{No_Inline}.
3628 @item -fno-inline-functions
3629 @cindex @option{-fno-inline-functions} (@command{gcc})
3630 Suppresses automatic inlining of subprograms, which is enabled
3631 if @option{-O3} is used.
3633 @item -fno-inline-small-functions
3634 @cindex @option{-fno-inline-small-functions} (@command{gcc})
3635 Suppresses automatic inlining of small subprograms, which is enabled
3636 if @option{-O2} is used.
3638 @item -fno-inline-functions-called-once
3639 @cindex @option{-fno-inline-functions-called-once} (@command{gcc})
3640 Suppresses inlining of subprograms local to the unit and called once
3641 from within it, which is enabled if @option{-O1} is used.
3644 @cindex @option{-fno-ivopts} (@command{gcc})
3645 Suppresses high-level loop induction variable optimizations, which are
3646 enabled if @option{-O1} is used. These optimizations are generally
3647 profitable but, for some specific cases of loops with numerous uses
3648 of the iteration variable that follow a common pattern, they may end
3649 up destroying the regularity that could be exploited at a lower level
3650 and thus producing inferior code.
3652 @item -fno-strict-aliasing
3653 @cindex @option{-fno-strict-aliasing} (@command{gcc})
3654 Causes the compiler to avoid assumptions regarding non-aliasing
3655 of objects of different types. See
3656 @ref{Optimization and Strict Aliasing} for details.
3659 @cindex @option{-fstack-check} (@command{gcc})
3660 Activates stack checking.
3661 See @ref{Stack Overflow Checking} for details.
3664 @cindex @option{-fstack-usage} (@command{gcc})
3665 Makes the compiler output stack usage information for the program, on a
3666 per-subprogram basis. See @ref{Static Stack Usage Analysis} for details.
3669 @cindex @option{^-g^/DEBUG^} (@command{gcc})
3670 Generate debugging information. This information is stored in the object
3671 file and copied from there to the final executable file by the linker,
3672 where it can be read by the debugger. You must use the
3673 @option{^-g^/DEBUG^} switch if you plan on using the debugger.
3676 @cindex @option{-gnat83} (@command{gcc})
3677 Enforce Ada 83 restrictions.
3680 @cindex @option{-gnat95} (@command{gcc})
3681 Enforce Ada 95 restrictions.
3684 @cindex @option{-gnat05} (@command{gcc})
3685 Allow full Ada 2005 features.
3688 @cindex @option{-gnat2005} (@command{gcc})
3689 Allow full Ada 2005 features (same as @option{-gnat05})
3692 @cindex @option{-gnat12} (@command{gcc})
3695 @cindex @option{-gnat2012} (@command{gcc})
3696 Allow full Ada 2012 features (same as @option{-gnat12})
3699 @cindex @option{-gnata} (@command{gcc})
3700 Assertions enabled. @code{Pragma Assert} and @code{pragma Debug} to be
3701 activated. Note that these pragmas can also be controlled using the
3702 configuration pragmas @code{Assertion_Policy} and @code{Debug_Policy}.
3703 It also activates pragmas @code{Check}, @code{Precondition}, and
3704 @code{Postcondition}. Note that these pragmas can also be controlled
3705 using the configuration pragma @code{Check_Policy}. In Ada 2012, it
3706 also activates all assertions defined in the RM as aspects: preconditions,
3707 postconditions, type invariants and (sub)type predicates. In all Ada modes,
3708 corresponding pragmas for type invariants and (sub)type predicates are
3712 @cindex @option{-gnatA} (@command{gcc})
3713 Avoid processing @file{gnat.adc}. If a @file{gnat.adc} file is present,
3717 @cindex @option{-gnatb} (@command{gcc})
3718 Generate brief messages to @file{stderr} even if verbose mode set.
3721 @cindex @option{-gnatB} (@command{gcc})
3722 Assume no invalid (bad) values except for 'Valid attribute use
3723 (@pxref{Validity Checking}).
3726 @cindex @option{-gnatc} (@command{gcc})
3727 Check syntax and semantics only (no code generation attempted). When the
3728 compiler is invoked by @command{gnatmake}, if the switch @option{-gnatc} is
3729 only given to the compiler (after @option{-cargs} or in package Compiler of
3730 the project file, @command{gnatmake} will fail because it will not find the
3731 object file after compilation. If @command{gnatmake} is called with
3732 @option{-gnatc} as a builder switch (before @option{-cargs} or in package
3733 Builder of the project file) then @command{gnatmake} will not fail because
3734 it will not look for the object files after compilation, and it will not try
3735 to build and link. This switch may not be given if a previous @code{-gnatR}
3736 switch has been given, since @code{-gnatR} requires that the code generator
3737 be called to complete determination of representation information.
3740 @cindex @option{-gnatC} (@command{gcc})
3741 Generate CodePeer intermediate format (no code generation attempted).
3742 This switch will generate an intermediate representation suitable for
3743 use by CodePeer (@file{.scil} files). This switch is not compatible with
3744 code generation (it will, among other things, disable some switches such
3745 as -gnatn, and enable others such as -gnata).
3748 @cindex @option{-gnatd} (@command{gcc})
3749 Specify debug options for the compiler. The string of characters after
3750 the @option{-gnatd} specify the specific debug options. The possible
3751 characters are 0-9, a-z, A-Z, optionally preceded by a dot. See
3752 compiler source file @file{debug.adb} for details of the implemented
3753 debug options. Certain debug options are relevant to applications
3754 programmers, and these are documented at appropriate points in this
3759 @cindex @option{-gnatD[nn]} (@command{gcc})
3762 @item /XDEBUG /LXDEBUG=nnn
3764 Create expanded source files for source level debugging. This switch
3765 also suppress generation of cross-reference information
3766 (see @option{-gnatx}). Note that this switch is not allowed if a previous
3767 -gnatR switch has been given, since these two switches are not compatible.
3769 @item ^-gnateA^/ALIASING_CHECK^
3770 @cindex @option{-gnateA} (@command{gcc})
3771 Check that there is no aliasing between two parameters of the same subprogram.
3773 @item -gnatec=@var{path}
3774 @cindex @option{-gnatec} (@command{gcc})
3775 Specify a configuration pragma file
3777 (the equal sign is optional)
3779 (@pxref{The Configuration Pragmas Files}).
3782 @cindex @option{-gnateC} (@command{gcc})
3783 Generate CodePeer messages in a compiler-like format. This switch is only
3784 effective if @option{-gnatcC} is also specified and requires an installation
3787 @item ^-gnated^/DISABLE_ATOMIC_SYNCHRONIZATION^
3788 @cindex @option{-gnated} (@command{gcc})
3789 Disable atomic synchronization
3791 @item ^-gnateD^/DATA_PREPROCESSING=^symbol@r{[}=@var{value}@r{]}
3792 @cindex @option{-gnateD} (@command{gcc})
3793 Defines a symbol, associated with @var{value}, for preprocessing.
3794 (@pxref{Integrated Preprocessing}).
3797 @cindex @option{-gnateE} (@command{gcc})
3798 Generate extra information in exception messages. In particular, display
3799 extra column information and the value and range associated with index and
3800 range check failures, and extra column information for access checks.
3801 In cases where the compiler is able to determine at compile time that
3802 a check will fail, it gives a warning, and the extra information is not
3803 produced at run time.
3806 @cindex @option{-gnatef} (@command{gcc})
3807 Display full source path name in brief error messages.
3810 @cindex @option{-gnateF} (@command{gcc})
3811 Check for overflow on all floating-point operations, including those
3812 for unconstrained predefined types. See description of pragma
3813 @code{Check_Float_Overflow} in GNAT RM.
3816 @cindex @option{-gnateG} (@command{gcc})
3817 Save result of preprocessing in a text file.
3819 @item -gnatei@var{nnn}
3820 @cindex @option{-gnatei} (@command{gcc})
3821 Set maximum number of instantiations during compilation of a single unit to
3822 @var{nnn}. This may be useful in increasing the default maximum of 8000 for
3823 the rare case when a single unit legitimately exceeds this limit.
3825 @item -gnateI@var{nnn}
3826 @cindex @option{-gnateI} (@command{gcc})
3827 Indicates that the source is a multi-unit source and that the index of the
3828 unit to compile is @var{nnn}. @var{nnn} needs to be a positive number and need
3829 to be a valid index in the multi-unit source.
3832 @cindex @option{-gnatel} (@command{gcc})
3833 This switch can be used with the static elaboration model to issue info
3835 where implicit @code{pragma Elaborate} and @code{pragma Elaborate_All}
3836 are generated. This is useful in diagnosing elaboration circularities
3837 caused by these implicit pragmas when using the static elaboration
3838 model. See See the section in this guide on elaboration checking for
3839 further details. These messages are not generated by default, and are
3840 intended only for temporary use when debugging circularity problems.
3843 @cindex @option{-gnatel} (@command{gcc})
3844 This switch turns off the info messages about implicit elaboration pragmas.
3846 @item -gnatem=@var{path}
3847 @cindex @option{-gnatem} (@command{gcc})
3848 Specify a mapping file
3850 (the equal sign is optional)
3852 (@pxref{Units to Sources Mapping Files}).
3854 @item -gnatep=@var{file}
3855 @cindex @option{-gnatep} (@command{gcc})
3856 Specify a preprocessing data file
3858 (the equal sign is optional)
3860 (@pxref{Integrated Preprocessing}).
3863 @cindex @option{-gnateP} (@command{gcc})
3864 Turn categorization dependency errors into warnings.
3865 Ada requires that units that WITH one another have compatible categories, for
3866 example a Pure unit cannot WITH a Preelaborate unit. If this switch is used,
3867 these errors become warnings (which can be ignored, or suppressed in the usual
3868 manner). This can be useful in some specialized circumstances such as the
3869 temporary use of special test software.
3872 @cindex @option{-gnateS} (@command{gcc})
3873 Synonym of @option{-fdump-scos}, kept for backwards compatibility.
3875 @item -gnatet=@var{path}
3876 @cindex @option{-gnatet=file} (@command{gcc})
3877 Generate target dependent information. The format of the output file is
3878 described in the section about switch @option{-gnateT}.
3880 @item -gnateT=@var{path}
3881 @cindex @option{-gnateT} (@command{gcc})
3882 Read target dependent information, such as endianness or sizes and alignments
3883 of base type. If this switch is passed, the default target dependent
3884 information of the compiler is replaced by the one read from the input file.
3885 This is used by tools other than the compiler, e.g. to do
3886 semantic analysis of programs that will run on some other target than
3887 the machine on which the tool is run.
3889 The following target dependent values should be defined,
3890 where @code{Nat} denotes a natural integer value, @code{Pos} denotes a
3891 positive integer value, and fields marked with a question mark are
3892 boolean fields, where a value of 0 is False, and a value of 1 is True:
3895 Bits_BE : Nat; -- Bits stored big-endian?
3896 Bits_Per_Unit : Pos; -- Bits in a storage unit
3897 Bits_Per_Word : Pos; -- Bits in a word
3898 Bytes_BE : Nat; -- Bytes stored big-endian?
3899 Char_Size : Pos; -- Standard.Character'Size
3900 Double_Float_Alignment : Nat; -- Alignment of double float
3901 Double_Scalar_Alignment : Nat; -- Alignment of double length scalar
3902 Double_Size : Pos; -- Standard.Long_Float'Size
3903 Float_Size : Pos; -- Standard.Float'Size
3904 Float_Words_BE : Nat; -- Float words stored big-endian?
3905 Int_Size : Pos; -- Standard.Integer'Size
3906 Long_Double_Size : Pos; -- Standard.Long_Long_Float'Size
3907 Long_Long_Size : Pos; -- Standard.Long_Long_Integer'Size
3908 Long_Size : Pos; -- Standard.Long_Integer'Size
3909 Maximum_Alignment : Pos; -- Maximum permitted alignment
3910 Max_Unaligned_Field : Pos; -- Maximum size for unaligned bit field
3911 Pointer_Size : Pos; -- System.Address'Size
3912 Short_Enums : Nat; -- Short foreign convention enums?
3913 Short_Size : Pos; -- Standard.Short_Integer'Size
3914 Strict_Alignment : Nat; -- Strict alignment?
3915 System_Allocator_Alignment : Nat; -- Alignment for malloc calls
3916 Wchar_T_Size : Pos; -- Interfaces.C.wchar_t'Size
3917 Words_BE : Nat; -- Words stored big-endian?
3920 The format of the input file is as follows. First come the values of
3921 the variables defined above, with one line per value:
3927 where @code{name} is the name of the parameter, spelled out in full,
3928 and cased as in the above list, and @code{value} is an unsigned decimal
3929 integer. Two or more blanks separates the name from the value.
3931 All the variables must be present, in alphabetical order (i.e. the
3932 same order as the list above).
3934 Then there is a blank line to separate the two parts of the file. Then
3935 come the lines showing the floating-point types to be registered, with
3936 one line per registered mode:
3939 name digs float_rep size alignment
3942 where @code{name} is the string name of the type (which can have
3943 single spaces embedded in the name (e.g. long double), @code{digs} is
3944 the number of digits for the floating-point type, @code{float_rep} is
3945 the float representation (I/V/A for IEEE-754-Binary, Vax_Native,
3946 AAMP), @code{size} is the size in bits, @code{alignment} is the
3947 alignment in bits. The name is followed by at least two blanks, fields
3948 are separated by at least one blank, and a LF character immediately
3949 follows the alignment field.
3951 Here is an example of a target parameterization file:
3959 Double_Float_Alignment 0
3960 Double_Scalar_Alignment 0
3965 Long_Double_Size 128
3968 Maximum_Alignment 16
3969 Max_Unaligned_Field 64
3973 System_Allocator_Alignment 16
3979 long double 18 I 80 128
3984 @cindex @option{-gnateu} (@command{gcc})
3985 Ignore unrecognized validity, warning, and style switches that
3986 appear after this switch is given. This may be useful when
3987 compiling sources developed on a later version of the compiler
3988 with an earlier version. Of course the earlier version must
3989 support this switch.
3991 @item ^-gnateV^/PARAMETER_VALIDITY_CHECK^
3992 @cindex @option{-gnateV} (@command{gcc})
3993 Check validity of subprogram parameters.
3995 @item ^-gnateY^/IGNORE_SUPPRESS_SYLE_CHECK_PRAGMAS^
3996 @cindex @option{-gnateY} (@command{gcc})
3997 Ignore all STYLE_CHECKS pragmas. Full legality checks
3998 are still carried out, but the pragmas have no effect
3999 on what style checks are active. This allows all style
4000 checking options to be controlled from the command line.
4003 @cindex @option{-gnatE} (@command{gcc})
4004 Full dynamic elaboration checks.
4007 @cindex @option{-gnatf} (@command{gcc})
4008 Full errors. Multiple errors per line, all undefined references, do not
4009 attempt to suppress cascaded errors.
4012 @cindex @option{-gnatF} (@command{gcc})
4013 Externals names are folded to all uppercase.
4015 @item ^-gnatg^/GNAT_INTERNAL^
4016 @cindex @option{^-gnatg^/GNAT_INTERNAL^} (@command{gcc})
4017 Internal GNAT implementation mode. This should not be used for
4018 applications programs, it is intended only for use by the compiler
4019 and its run-time library. For documentation, see the GNAT sources.
4020 Note that @option{^-gnatg^/GNAT_INTERNAL^} implies
4021 @option{^-gnatw.ge^/WARNINGS=GNAT,ERRORS^} and
4022 @option{^-gnatyg^/STYLE_CHECKS=GNAT^}
4023 so that all standard warnings and all standard style options are turned on.
4024 All warnings and style messages are treated as errors.
4028 @cindex @option{-gnatG[nn]} (@command{gcc})
4031 @item /EXPAND_SOURCE, /LEXPAND_SOURCE=nnn
4033 List generated expanded code in source form.
4035 @item ^-gnath^/HELP^
4036 @cindex @option{^-gnath^/HELP^} (@command{gcc})
4037 Output usage information. The output is written to @file{stdout}.
4039 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
4040 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
4041 Identifier character set
4043 (@var{c}=1/2/3/4/8/9/p/f/n/w).
4045 For details of the possible selections for @var{c},
4046 see @ref{Character Set Control}.
4048 @item ^-gnatI^/IGNORE_REP_CLAUSES^
4049 @cindex @option{^-gnatI^IGNORE_REP_CLAUSES^} (@command{gcc})
4050 Ignore representation clauses. When this switch is used,
4051 representation clauses are treated as comments. This is useful
4052 when initially porting code where you want to ignore rep clause
4053 problems, and also for compiling foreign code (particularly
4054 for use with ASIS). The representation clauses that are ignored
4055 are: enumeration_representation_clause, record_representation_clause,
4056 and attribute_definition_clause for the following attributes:
4057 Address, Alignment, Bit_Order, Component_Size, Machine_Radix,
4058 Object_Size, Size, Small, Stream_Size, and Value_Size.
4059 Note that this option should be used only for compiling -- the
4060 code is likely to malfunction at run time.
4063 @cindex @option{-gnatjnn} (@command{gcc})
4064 Reformat error messages to fit on nn character lines
4066 @item -gnatk=@var{n}
4067 @cindex @option{-gnatk} (@command{gcc})
4068 Limit file names to @var{n} (1-999) characters ^(@code{k} = krunch)^^.
4071 @cindex @option{-gnatl} (@command{gcc})
4072 Output full source listing with embedded error messages.
4075 @cindex @option{-gnatL} (@command{gcc})
4076 Used in conjunction with -gnatG or -gnatD to intersperse original
4077 source lines (as comment lines with line numbers) in the expanded
4080 @item -gnatm=@var{n}
4081 @cindex @option{-gnatm} (@command{gcc})
4082 Limit number of detected error or warning messages to @var{n}
4083 where @var{n} is in the range 1..999999. The default setting if
4084 no switch is given is 9999. If the number of warnings reaches this
4085 limit, then a message is output and further warnings are suppressed,
4086 but the compilation is continued. If the number of error messages
4087 reaches this limit, then a message is output and the compilation
4088 is abandoned. The equal sign here is optional. A value of zero
4089 means that no limit applies.
4092 @cindex @option{-gnatn} (@command{gcc})
4093 Activate inlining for subprograms for which pragma @code{Inline} is
4094 specified. This inlining is performed by the GCC back-end. An optional
4095 digit sets the inlining level: 1 for moderate inlining across modules
4096 or 2 for full inlining across modules. If no inlining level is specified,
4097 the compiler will pick it based on the optimization level.
4100 @cindex @option{-gnatN} (@command{gcc})
4101 Activate front end inlining for subprograms for which
4102 pragma @code{Inline} is specified. This inlining is performed
4103 by the front end and will be visible in the
4104 @option{-gnatG} output.
4106 When using a gcc-based back end (in practice this means using any version
4107 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
4108 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
4109 Historically front end inlining was more extensive than the gcc back end
4110 inlining, but that is no longer the case.
4113 @cindex @option{-gnato??} (@command{gcc})
4114 Set default mode for handling generation of code to avoid intermediate
4115 arithmetic overflow. Here `@code{??}' is two digits, a
4116 single digit, or nothing. Each digit is one of the digits `@code{1}'
4121 all intermediate overflows checked against base type (@code{STRICT})
4123 minimize intermediate overflows (@code{MINIMIZED})
4125 eliminate intermediate overflows (@code{ELIMINATED})
4128 If only one digit appears then it applies to all
4129 cases; if two digits are given, then the first applies outside
4130 assertions, and the second within assertions.
4132 If no digits follow the @option{-gnato}, then it is equivalent to
4133 @option{^-gnato11^/OVERFLOW_CHECKS=11^},
4134 causing all intermediate overflows to be handled in strict mode.
4136 This switch also causes arithmetic overflow checking to be performed
4137 (as though pragma @code{Unsuppress (Overflow_Mode)} has been specified.
4139 The default if no option @option{-gnato} is given is that overflow handling
4140 is in @code{STRICT} mode (computations done using the base type), and that
4141 overflow checking is suppressed.
4143 Note that division by zero is a separate check that is not
4144 controlled by this switch (division by zero checking is on by default).
4146 See also @ref{Specifying the Desired Mode}.
4149 @cindex @option{-gnatp} (@command{gcc})
4150 Suppress all checks. See @ref{Run-Time Checks} for details. This switch
4151 has no effect if cancelled by a subsequent @option{-gnat-p} switch.
4154 @cindex @option{-gnat-p} (@command{gcc})
4155 Cancel effect of previous @option{-gnatp} switch.
4158 @cindex @option{-gnatP} (@command{gcc})
4159 Enable polling. This is required on some systems (notably Windows NT) to
4160 obtain asynchronous abort and asynchronous transfer of control capability.
4161 @xref{Pragma Polling,,, gnat_rm, GNAT Reference Manual}, for full
4165 @cindex @option{-gnatq} (@command{gcc})
4166 Don't quit. Try semantics, even if parse errors.
4169 @cindex @option{-gnatQ} (@command{gcc})
4170 Don't quit. Generate @file{ALI} and tree files even if illegalities.
4173 @cindex @option{-gnatr} (@command{gcc})
4174 Treat pragma Restrictions as Restriction_Warnings.
4176 @item ^-gnatR@r{[}0@r{/}1@r{/}2@r{/}3@r{[}s@r{]]}^/REPRESENTATION_INFO^
4177 @cindex @option{-gnatR} (@command{gcc})
4178 Output representation information for declared types and objects.
4179 Note that this switch is not allowed if a previous @code{-gnatD} switch has
4180 been given, since these two switches are not compatible. It is also not allowed
4181 if a previous @code{-gnatc} switch has been given, since we must be generating
4182 code to be able to determine representation information.
4184 @item ^-gnatRm[s]^/REPRESENTATION_INFO^
4185 Output convention and parameter passing mechanisms for all subprograms.
4186 This form is also incompatible with the use of @code{-gnatc}.
4189 @cindex @option{-gnats} (@command{gcc})
4193 @cindex @option{-gnatS} (@command{gcc})
4194 Print package Standard.
4197 @cindex @option{-gnatt} (@command{gcc})
4198 Generate tree output file.
4200 @item ^-gnatT^/TABLE_MULTIPLIER=^@var{nnn}
4201 @cindex @option{^-gnatT^/TABLE_MULTIPLIER^} (@command{gcc})
4202 All compiler tables start at @var{nnn} times usual starting size.
4205 @cindex @option{-gnatu} (@command{gcc})
4206 List units for this compilation.
4209 @cindex @option{-gnatU} (@command{gcc})
4210 Tag all error messages with the unique string ``error:''
4213 @cindex @option{-gnatv} (@command{gcc})
4214 Verbose mode. Full error output with source lines to @file{stdout}.
4217 @cindex @option{-gnatV} (@command{gcc})
4218 Control level of validity checking (@pxref{Validity Checking}).
4220 @item ^-gnatw@var{xxx}^/WARNINGS=(@var{option}@r{[},@dots{}@r{]})^
4221 @cindex @option{^-gnatw^/WARNINGS^} (@command{gcc})
4223 ^@var{xxx} is a string of option letters that^the list of options^ denotes
4224 the exact warnings that
4225 are enabled or disabled (@pxref{Warning Message Control}).
4227 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
4228 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
4229 Wide character encoding method
4231 (@var{e}=n/h/u/s/e/8).
4234 (@var{e}=@code{BRACKETS, NONE, HEX, UPPER, SHIFT_JIS, EUC, UTF8})
4238 @cindex @option{-gnatx} (@command{gcc})
4239 Suppress generation of cross-reference information.
4242 @cindex @option{-gnatX} (@command{gcc})
4243 Enable GNAT implementation extensions and latest Ada version.
4245 @item ^-gnaty^/STYLE_CHECKS=(option,option@dots{})^
4246 @cindex @option{^-gnaty^/STYLE_CHECKS^} (@command{gcc})
4247 Enable built-in style checks (@pxref{Style Checking}).
4249 @item ^-gnatz^/DISTRIBUTION_STUBS=^@var{m}
4250 @cindex @option{^-gnatz^/DISTRIBUTION_STUBS^} (@command{gcc})
4251 Distribution stub generation and compilation
4253 (@var{m}=r/c for receiver/caller stubs).
4256 (@var{m}=@code{RECEIVER} or @code{CALLER} to specify the type of stubs
4257 to be generated and compiled).
4260 @item ^-I^/SEARCH=^@var{dir}
4261 @cindex @option{^-I^/SEARCH^} (@command{gcc})
4263 Direct GNAT to search the @var{dir} directory for source files needed by
4264 the current compilation
4265 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4267 @item ^-I-^/NOCURRENT_DIRECTORY^
4268 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gcc})
4270 Except for the source file named in the command line, do not look for source
4271 files in the directory containing the source file named in the command line
4272 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4276 @cindex @option{-mbig-switch} (@command{gcc})
4277 @cindex @code{case} statement (effect of @option{-mbig-switch} option)
4278 This standard gcc switch causes the compiler to use larger offsets in its
4279 jump table representation for @code{case} statements.
4280 This may result in less efficient code, but is sometimes necessary
4281 (for example on HP-UX targets)
4282 @cindex HP-UX and @option{-mbig-switch} option
4283 in order to compile large and/or nested @code{case} statements.
4286 @cindex @option{-o} (@command{gcc})
4287 This switch is used in @command{gcc} to redirect the generated object file
4288 and its associated ALI file. Beware of this switch with GNAT, because it may
4289 cause the object file and ALI file to have different names which in turn
4290 may confuse the binder and the linker.
4294 @cindex @option{-nostdinc} (@command{gcc})
4295 Inhibit the search of the default location for the GNAT Run Time
4296 Library (RTL) source files.
4299 @cindex @option{-nostdlib} (@command{gcc})
4300 Inhibit the search of the default location for the GNAT Run Time
4301 Library (RTL) ALI files.
4305 @c Expanding @ovar macro inline (explanation in macro def comments)
4306 @item -O@r{[}@var{n}@r{]}
4307 @cindex @option{-O} (@command{gcc})
4308 @var{n} controls the optimization level.
4312 No optimization, the default setting if no @option{-O} appears
4315 Normal optimization, the default if you specify @option{-O} without
4316 an operand. A good compromise between code quality and compilation
4320 Extensive optimization, may improve execution time, possibly at the cost of
4321 substantially increased compilation time.
4324 Same as @option{-O2}, and also includes inline expansion for small subprograms
4328 Optimize space usage
4332 See also @ref{Optimization Levels}.
4337 @cindex @option{/NOOPTIMIZE} (@code{GNAT COMPILE})
4338 Equivalent to @option{/OPTIMIZE=NONE}.
4339 This is the default behavior in the absence of an @option{/OPTIMIZE}
4342 @item /OPTIMIZE@r{[}=(keyword@r{[},@dots{}@r{]})@r{]}
4343 @cindex @option{/OPTIMIZE} (@code{GNAT COMPILE})
4344 Selects the level of optimization for your program. The supported
4345 keywords are as follows:
4348 Perform most optimizations, including those that
4350 This is the default if the @option{/OPTIMIZE} qualifier is supplied
4351 without keyword options.
4354 Do not do any optimizations. Same as @code{/NOOPTIMIZE}.
4357 Perform some optimizations, but omit ones that are costly.
4360 Same as @code{SOME}.
4363 Full optimization as in @option{/OPTIMIZE=ALL}, and also attempts
4364 automatic inlining of small subprograms within a unit
4367 Try to unroll loops. This keyword may be specified together with
4368 any keyword above other than @code{NONE}. Loop unrolling
4369 usually, but not always, improves the performance of programs.
4372 Optimize space usage
4376 See also @ref{Optimization Levels}.
4380 @item -pass-exit-codes
4381 @cindex @option{-pass-exit-codes} (@command{gcc})
4382 Catch exit codes from the compiler and use the most meaningful as
4386 @item --RTS=@var{rts-path}
4387 @cindex @option{--RTS} (@command{gcc})
4388 Specifies the default location of the runtime library. Same meaning as the
4389 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
4392 @cindex @option{^-S^/ASM^} (@command{gcc})
4393 ^Used in place of @option{-c} to^Used to^
4394 cause the assembler source file to be
4395 generated, using @file{^.s^.S^} as the extension,
4396 instead of the object file.
4397 This may be useful if you need to examine the generated assembly code.
4399 @item ^-fverbose-asm^/VERBOSE_ASM^
4400 @cindex @option{^-fverbose-asm^/VERBOSE_ASM^} (@command{gcc})
4401 ^Used in conjunction with @option{-S}^Used in place of @option{/ASM}^
4402 to cause the generated assembly code file to be annotated with variable
4403 names, making it significantly easier to follow.
4406 @cindex @option{^-v^/VERBOSE^} (@command{gcc})
4407 Show commands generated by the @command{gcc} driver. Normally used only for
4408 debugging purposes or if you need to be sure what version of the
4409 compiler you are executing.
4413 @cindex @option{-V} (@command{gcc})
4414 Execute @var{ver} version of the compiler. This is the @command{gcc}
4415 version, not the GNAT version.
4418 @item ^-w^/NO_BACK_END_WARNINGS^
4419 @cindex @option{-w} (@command{gcc})
4420 Turn off warnings generated by the back end of the compiler. Use of
4421 this switch also causes the default for front end warnings to be set
4422 to suppress (as though @option{-gnatws} had appeared at the start of
4428 @c Combining qualifiers does not work on VMS
4429 You may combine a sequence of GNAT switches into a single switch. For
4430 example, the combined switch
4432 @cindex Combining GNAT switches
4438 is equivalent to specifying the following sequence of switches:
4441 -gnato -gnatf -gnati3
4446 The following restrictions apply to the combination of switches
4451 The switch @option{-gnatc} if combined with other switches must come
4452 first in the string.
4455 The switch @option{-gnats} if combined with other switches must come
4456 first in the string.
4460 ^^@option{/DISTRIBUTION_STUBS=},^
4461 @option{-gnatzc} and @option{-gnatzr} may not be combined with any other
4462 switches, and only one of them may appear in the command line.
4465 The switch @option{-gnat-p} may not be combined with any other switch.
4469 Once a ``y'' appears in the string (that is a use of the @option{-gnaty}
4470 switch), then all further characters in the switch are interpreted
4471 as style modifiers (see description of @option{-gnaty}).
4474 Once a ``d'' appears in the string (that is a use of the @option{-gnatd}
4475 switch), then all further characters in the switch are interpreted
4476 as debug flags (see description of @option{-gnatd}).
4479 Once a ``w'' appears in the string (that is a use of the @option{-gnatw}
4480 switch), then all further characters in the switch are interpreted
4481 as warning mode modifiers (see description of @option{-gnatw}).
4484 Once a ``V'' appears in the string (that is a use of the @option{-gnatV}
4485 switch), then all further characters in the switch are interpreted
4486 as validity checking options (@pxref{Validity Checking}).
4489 Option ``em'', ``ec'', ``ep'', ``l='' and ``R'' must be the last options in
4490 a combined list of options.
4494 @node Output and Error Message Control
4495 @subsection Output and Error Message Control
4499 The standard default format for error messages is called ``brief format''.
4500 Brief format messages are written to @file{stderr} (the standard error
4501 file) and have the following form:
4504 e.adb:3:04: Incorrect spelling of keyword "function"
4505 e.adb:4:20: ";" should be "is"
4509 The first integer after the file name is the line number in the file,
4510 and the second integer is the column number within the line.
4512 @code{GPS} can parse the error messages
4513 and point to the referenced character.
4515 The following switches provide control over the error message
4521 @cindex @option{-gnatv} (@command{gcc})
4524 The v stands for verbose.
4526 The effect of this setting is to write long-format error
4527 messages to @file{stdout} (the standard output file.
4528 The same program compiled with the
4529 @option{-gnatv} switch would generate:
4533 3. funcion X (Q : Integer)
4535 >>> Incorrect spelling of keyword "function"
4538 >>> ";" should be "is"
4543 The vertical bar indicates the location of the error, and the @samp{>>>}
4544 prefix can be used to search for error messages. When this switch is
4545 used the only source lines output are those with errors.
4548 @cindex @option{-gnatl} (@command{gcc})
4550 The @code{l} stands for list.
4552 This switch causes a full listing of
4553 the file to be generated. In the case where a body is
4554 compiled, the corresponding spec is also listed, along
4555 with any subunits. Typical output from compiling a package
4556 body @file{p.adb} might look like:
4558 @smallexample @c ada
4562 1. package body p is
4564 3. procedure a is separate;
4575 2. pragma Elaborate_Body
4599 When you specify the @option{-gnatv} or @option{-gnatl} switches and
4600 standard output is redirected, a brief summary is written to
4601 @file{stderr} (standard error) giving the number of error messages and
4602 warning messages generated.
4604 @item ^-gnatl^/OUTPUT_FILE^=file
4605 @cindex @option{^-gnatl^/OUTPUT_FILE^=fname} (@command{gcc})
4606 This has the same effect as @option{-gnatl} except that the output is
4607 written to a file instead of to standard output. If the given name
4608 @file{fname} does not start with a period, then it is the full name
4609 of the file to be written. If @file{fname} is an extension, it is
4610 appended to the name of the file being compiled. For example, if
4611 file @file{xyz.adb} is compiled with @option{^-gnatl^/OUTPUT_FILE^=.lst},
4612 then the output is written to file ^xyz.adb.lst^xyz.adb_lst^.
4615 @cindex @option{-gnatU} (@command{gcc})
4616 This switch forces all error messages to be preceded by the unique
4617 string ``error:''. This means that error messages take a few more
4618 characters in space, but allows easy searching for and identification
4622 @cindex @option{-gnatb} (@command{gcc})
4624 The @code{b} stands for brief.
4626 This switch causes GNAT to generate the
4627 brief format error messages to @file{stderr} (the standard error
4628 file) as well as the verbose
4629 format message or full listing (which as usual is written to
4630 @file{stdout} (the standard output file).
4632 @item -gnatm=@var{n}
4633 @cindex @option{-gnatm} (@command{gcc})
4635 The @code{m} stands for maximum.
4637 @var{n} is a decimal integer in the
4638 range of 1 to 999999 and limits the number of error or warning
4639 messages to be generated. For example, using
4640 @option{-gnatm2} might yield
4643 e.adb:3:04: Incorrect spelling of keyword "function"
4644 e.adb:5:35: missing ".."
4645 fatal error: maximum number of errors detected
4646 compilation abandoned
4650 The default setting if
4651 no switch is given is 9999. If the number of warnings reaches this
4652 limit, then a message is output and further warnings are suppressed,
4653 but the compilation is continued. If the number of error messages
4654 reaches this limit, then a message is output and the compilation
4655 is abandoned. A value of zero means that no limit applies.
4658 Note that the equal sign is optional, so the switches
4659 @option{-gnatm2} and @option{-gnatm=2} are equivalent.
4662 @cindex @option{-gnatf} (@command{gcc})
4663 @cindex Error messages, suppressing
4665 The @code{f} stands for full.
4667 Normally, the compiler suppresses error messages that are likely to be
4668 redundant. This switch causes all error
4669 messages to be generated. In particular, in the case of
4670 references to undefined variables. If a given variable is referenced
4671 several times, the normal format of messages is
4673 e.adb:7:07: "V" is undefined (more references follow)
4677 where the parenthetical comment warns that there are additional
4678 references to the variable @code{V}. Compiling the same program with the
4679 @option{-gnatf} switch yields
4682 e.adb:7:07: "V" is undefined
4683 e.adb:8:07: "V" is undefined
4684 e.adb:8:12: "V" is undefined
4685 e.adb:8:16: "V" is undefined
4686 e.adb:9:07: "V" is undefined
4687 e.adb:9:12: "V" is undefined
4691 The @option{-gnatf} switch also generates additional information for
4692 some error messages. Some examples are:
4696 Details on possibly non-portable unchecked conversion
4698 List possible interpretations for ambiguous calls
4700 Additional details on incorrect parameters
4704 @cindex @option{-gnatjnn} (@command{gcc})
4705 In normal operation mode (or if @option{-gnatj0} is used), then error messages
4706 with continuation lines are treated as though the continuation lines were
4707 separate messages (and so a warning with two continuation lines counts as
4708 three warnings, and is listed as three separate messages).
4710 If the @option{-gnatjnn} switch is used with a positive value for nn, then
4711 messages are output in a different manner. A message and all its continuation
4712 lines are treated as a unit, and count as only one warning or message in the
4713 statistics totals. Furthermore, the message is reformatted so that no line
4714 is longer than nn characters.
4717 @cindex @option{-gnatq} (@command{gcc})
4719 The @code{q} stands for quit (really ``don't quit'').
4721 In normal operation mode, the compiler first parses the program and
4722 determines if there are any syntax errors. If there are, appropriate
4723 error messages are generated and compilation is immediately terminated.
4725 GNAT to continue with semantic analysis even if syntax errors have been
4726 found. This may enable the detection of more errors in a single run. On
4727 the other hand, the semantic analyzer is more likely to encounter some
4728 internal fatal error when given a syntactically invalid tree.
4731 @cindex @option{-gnatQ} (@command{gcc})
4732 In normal operation mode, the @file{ALI} file is not generated if any
4733 illegalities are detected in the program. The use of @option{-gnatQ} forces
4734 generation of the @file{ALI} file. This file is marked as being in
4735 error, so it cannot be used for binding purposes, but it does contain
4736 reasonably complete cross-reference information, and thus may be useful
4737 for use by tools (e.g., semantic browsing tools or integrated development
4738 environments) that are driven from the @file{ALI} file. This switch
4739 implies @option{-gnatq}, since the semantic phase must be run to get a
4740 meaningful ALI file.
4742 In addition, if @option{-gnatt} is also specified, then the tree file is
4743 generated even if there are illegalities. It may be useful in this case
4744 to also specify @option{-gnatq} to ensure that full semantic processing
4745 occurs. The resulting tree file can be processed by ASIS, for the purpose
4746 of providing partial information about illegal units, but if the error
4747 causes the tree to be badly malformed, then ASIS may crash during the
4750 When @option{-gnatQ} is used and the generated @file{ALI} file is marked as
4751 being in error, @command{gnatmake} will attempt to recompile the source when it
4752 finds such an @file{ALI} file, including with switch @option{-gnatc}.
4754 Note that @option{-gnatQ} has no effect if @option{-gnats} is specified,
4755 since ALI files are never generated if @option{-gnats} is set.
4759 @node Warning Message Control
4760 @subsection Warning Message Control
4761 @cindex Warning messages
4763 In addition to error messages, which correspond to illegalities as defined
4764 in the Ada Reference Manual, the compiler detects two kinds of warning
4767 First, the compiler considers some constructs suspicious and generates a
4768 warning message to alert you to a possible error. Second, if the
4769 compiler detects a situation that is sure to raise an exception at
4770 run time, it generates a warning message. The following shows an example
4771 of warning messages:
4773 e.adb:4:24: warning: creation of object may raise Storage_Error
4774 e.adb:10:17: warning: static value out of range
4775 e.adb:10:17: warning: "Constraint_Error" will be raised at run time
4779 GNAT considers a large number of situations as appropriate
4780 for the generation of warning messages. As always, warnings are not
4781 definite indications of errors. For example, if you do an out-of-range
4782 assignment with the deliberate intention of raising a
4783 @code{Constraint_Error} exception, then the warning that may be
4784 issued does not indicate an error. Some of the situations for which GNAT
4785 issues warnings (at least some of the time) are given in the following
4786 list. This list is not complete, and new warnings are often added to
4787 subsequent versions of GNAT. The list is intended to give a general idea
4788 of the kinds of warnings that are generated.
4792 Possible infinitely recursive calls
4795 Out-of-range values being assigned
4798 Possible order of elaboration problems
4801 Size not a multiple of alignment for a record type
4804 Assertions (pragma Assert) that are sure to fail
4810 Address clauses with possibly unaligned values, or where an attempt is
4811 made to overlay a smaller variable with a larger one.
4814 Fixed-point type declarations with a null range
4817 Direct_IO or Sequential_IO instantiated with a type that has access values
4820 Variables that are never assigned a value
4823 Variables that are referenced before being initialized
4826 Task entries with no corresponding @code{accept} statement
4829 Duplicate accepts for the same task entry in a @code{select}
4832 Objects that take too much storage
4835 Unchecked conversion between types of differing sizes
4838 Missing @code{return} statement along some execution path in a function
4841 Incorrect (unrecognized) pragmas
4844 Incorrect external names
4847 Allocation from empty storage pool
4850 Potentially blocking operation in protected type
4853 Suspicious parenthesization of expressions
4856 Mismatching bounds in an aggregate
4859 Attempt to return local value by reference
4862 Premature instantiation of a generic body
4865 Attempt to pack aliased components
4868 Out of bounds array subscripts
4871 Wrong length on string assignment
4874 Violations of style rules if style checking is enabled
4877 Unused @code{with} clauses
4880 @code{Bit_Order} usage that does not have any effect
4883 @code{Standard.Duration} used to resolve universal fixed expression
4886 Dereference of possibly null value
4889 Declaration that is likely to cause storage error
4892 Internal GNAT unit @code{with}'ed by application unit
4895 Values known to be out of range at compile time
4898 Unreferenced or unmodified variables. Note that a special
4899 exemption applies to variables which contain any of the substrings
4900 @code{DISCARD, DUMMY, IGNORE, JUNK, UNUSED}, in any casing. Such variables
4901 are considered likely to be intentionally used in a situation where
4902 otherwise a warning would be given, so warnings of this kind are
4903 always suppressed for such variables.
4906 Address overlays that could clobber memory
4909 Unexpected initialization when address clause present
4912 Bad alignment for address clause
4915 Useless type conversions
4918 Redundant assignment statements and other redundant constructs
4921 Useless exception handlers
4924 Accidental hiding of name by child unit
4927 Access before elaboration detected at compile time
4930 A range in a @code{for} loop that is known to be null or might be null
4935 The following section lists compiler switches that are available
4936 to control the handling of warning messages. It is also possible
4937 to exercise much finer control over what warnings are issued and
4938 suppressed using the GNAT pragma Warnings, @xref{Pragma Warnings,,,
4939 gnat_rm, GNAT Reference manual}.
4944 @emph{Activate most optional warnings.}
4945 @cindex @option{-gnatwa} (@command{gcc})
4946 This switch activates most optional warning messages. See the remaining list
4947 in this section for details on optional warning messages that can be
4948 individually controlled. The warnings that are not turned on by this
4950 @option{-gnatwd} (implicit dereferencing),
4951 @option{-gnatwh} (hiding),
4952 @option{-gnatw.d} (tag warnings with -gnatw switch)
4953 @option{-gnatw.h} (holes (gaps) in record layouts)
4954 @option{-gnatw.i} (overlapping actuals),
4955 @option{-gnatw.k} (redefinition of names in standard),
4956 @option{-gnatwl} (elaboration warnings),
4957 @option{-gnatw.l} (inherited aspects),
4958 @option{-gnatw.o} (warn on values set by out parameters ignored),
4959 @option{-gnatwt} (tracking of deleted conditional code)
4960 and @option{-gnatw.u} (unordered enumeration),
4961 All other optional warnings are turned on.
4964 @emph{Suppress all optional errors.}
4965 @cindex @option{-gnatwA} (@command{gcc})
4966 This switch suppresses all optional warning messages, see remaining list
4967 in this section for details on optional warning messages that can be
4968 individually controlled. Note that unlike switch @option{-gnatws}, the
4969 use of switch @option{-gnatwA} does not suppress warnings that are
4970 normally given unconditionally and cannot be individually controlled
4971 (for example, the warning about a missing exit path in a function).
4972 Also, again unlike switch @option{-gnatws}, warnings suppressed by
4973 the use of switch @option{-gnatwA} can be individually turned back
4974 on. For example the use of switch @option{-gnatwA} followed by
4975 switch @option{-gnatwd} will suppress all optional warnings except
4976 the warnings for implicit dereferencing.
4979 @emph{Activate warnings on failing assertions.}
4980 @cindex @option{-gnatw.a} (@command{gcc})
4981 @cindex Assert failures
4982 This switch activates warnings for assertions where the compiler can tell at
4983 compile time that the assertion will fail. Note that this warning is given
4984 even if assertions are disabled. The default is that such warnings are
4988 @emph{Suppress warnings on failing assertions.}
4989 @cindex @option{-gnatw.A} (@command{gcc})
4990 @cindex Assert failures
4991 This switch suppresses warnings for assertions where the compiler can tell at
4992 compile time that the assertion will fail.
4995 @emph{Activate warnings on bad fixed values.}
4996 @cindex @option{-gnatwb} (@command{gcc})
4997 @cindex Bad fixed values
4998 @cindex Fixed-point Small value
5000 This switch activates warnings for static fixed-point expressions whose
5001 value is not an exact multiple of Small. Such values are implementation
5002 dependent, since an implementation is free to choose either of the multiples
5003 that surround the value. GNAT always chooses the closer one, but this is not
5004 required behavior, and it is better to specify a value that is an exact
5005 multiple, ensuring predictable execution. The default is that such warnings
5009 @emph{Suppress warnings on bad fixed values.}
5010 @cindex @option{-gnatwB} (@command{gcc})
5011 This switch suppresses warnings for static fixed-point expressions whose
5012 value is not an exact multiple of Small.
5015 @emph{Activate warnings on biased representation.}
5016 @cindex @option{-gnatw.b} (@command{gcc})
5017 @cindex Biased representation
5018 This switch activates warnings when a size clause, value size clause, component
5019 clause, or component size clause forces the use of biased representation for an
5020 integer type (e.g. representing a range of 10..11 in a single bit by using 0/1
5021 to represent 10/11). The default is that such warnings are generated.
5024 @emph{Suppress warnings on biased representation.}
5025 @cindex @option{-gnatwB} (@command{gcc})
5026 This switch suppresses warnings for representation clauses that force the use
5027 of biased representation.
5030 @emph{Activate warnings on conditionals.}
5031 @cindex @option{-gnatwc} (@command{gcc})
5032 @cindex Conditionals, constant
5033 This switch activates warnings for conditional expressions used in
5034 tests that are known to be True or False at compile time. The default
5035 is that such warnings are not generated.
5036 Note that this warning does
5037 not get issued for the use of boolean variables or constants whose
5038 values are known at compile time, since this is a standard technique
5039 for conditional compilation in Ada, and this would generate too many
5040 false positive warnings.
5042 This warning option also activates a special test for comparisons using
5043 the operators ``>='' and`` <=''.
5044 If the compiler can tell that only the equality condition is possible,
5045 then it will warn that the ``>'' or ``<'' part of the test
5046 is useless and that the operator could be replaced by ``=''.
5047 An example would be comparing a @code{Natural} variable <= 0.
5049 This warning option also generates warnings if
5050 one or both tests is optimized away in a membership test for integer
5051 values if the result can be determined at compile time. Range tests on
5052 enumeration types are not included, since it is common for such tests
5053 to include an end point.
5055 This warning can also be turned on using @option{-gnatwa}.
5058 @emph{Suppress warnings on conditionals.}
5059 @cindex @option{-gnatwC} (@command{gcc})
5060 This switch suppresses warnings for conditional expressions used in
5061 tests that are known to be True or False at compile time.
5064 @emph{Activate warnings on missing component clauses.}
5065 @cindex @option{-gnatw.c} (@command{gcc})
5066 @cindex Component clause, missing
5067 This switch activates warnings for record components where a record
5068 representation clause is present and has component clauses for the
5069 majority, but not all, of the components. A warning is given for each
5070 component for which no component clause is present.
5072 This warning can also be turned on using @option{-gnatwa}.
5075 @emph{Suppress warnings on missing component clauses.}
5076 @cindex @option{-gnatwC} (@command{gcc})
5077 This switch suppresses warnings for record components that are
5078 missing a component clause in the situation described above.
5081 @emph{Activate warnings on implicit dereferencing.}
5082 @cindex @option{-gnatwd} (@command{gcc})
5083 If this switch is set, then the use of a prefix of an access type
5084 in an indexed component, slice, or selected component without an
5085 explicit @code{.all} will generate a warning. With this warning
5086 enabled, access checks occur only at points where an explicit
5087 @code{.all} appears in the source code (assuming no warnings are
5088 generated as a result of this switch). The default is that such
5089 warnings are not generated.
5090 Note that @option{-gnatwa} does not affect the setting of
5091 this warning option.
5094 @emph{Suppress warnings on implicit dereferencing.}
5095 @cindex @option{-gnatwD} (@command{gcc})
5096 @cindex Implicit dereferencing
5097 @cindex Dereferencing, implicit
5098 This switch suppresses warnings for implicit dereferences in
5099 indexed components, slices, and selected components.
5102 @emph{Activate tagging of warning and info messages.}
5103 @cindex @option{-gnatw.d} (@command{gcc})
5104 If this switch is set, then warning messages are tagged, with one of the
5110 Used to tag warnings controlled by the switch @option{-gnatwx} where x
5114 Used to tag warnings controlled by the switch @option{-gnatw.x} where x
5118 Used to tag elaboration information (info) messages generated when the
5119 static model of elaboration is used and the @option{-gnatel} switch is set.
5121 @item [restriction warning]
5122 Used to tag warning messages for restriction violations, activated by use
5123 of the pragma @option{Restriction_Warnings}.
5125 @item [warning-as-error]
5126 Used to tag warning messages that have been converted to error messages by
5127 use of the pragma Warning_As_Error. Note that such warnings are prefixed by
5128 the string "error: " rather than "warning: ".
5130 @item [enabled by default]
5131 Used to tag all other warnings that are always given by default, unless
5132 warnings are completely suppressed using pragma @option{Warnings(Off)} or
5133 the switch @option{-gnatws}.
5138 @emph{Deactivate tagging of warning and info messages messages.}
5139 @cindex @option{-gnatw.d} (@command{gcc})
5140 If this switch is set, then warning messages return to the default
5141 mode in which warnings and info messages are not tagged as described above for
5145 @emph{Treat warnings and style checks as errors.}
5146 @cindex @option{-gnatwe} (@command{gcc})
5147 @cindex Warnings, treat as error
5148 This switch causes warning messages and style check messages to be
5150 The warning string still appears, but the warning messages are counted
5151 as errors, and prevent the generation of an object file. Note that this
5152 is the only -gnatw switch that affects the handling of style check messages.
5155 @emph{Activate every optional warning}
5156 @cindex @option{-gnatw.e} (@command{gcc})
5157 @cindex Warnings, activate every optional warning
5158 This switch activates all optional warnings, including those which
5159 are not activated by @code{-gnatwa}. The use of this switch is not
5160 recommended for normal use. If you turn this switch on, it is almost
5161 certain that you will get large numbers of useless warnings. The
5162 warnings that are excluded from @code{-gnatwa} are typically highly
5163 specialized warnings that are suitable for use only in code that has
5164 been specifically designed according to specialized coding rules.
5167 @emph{Activate warnings on unreferenced formals.}
5168 @cindex @option{-gnatwf} (@command{gcc})
5169 @cindex Formals, unreferenced
5170 This switch causes a warning to be generated if a formal parameter
5171 is not referenced in the body of the subprogram. This warning can
5172 also be turned on using @option{-gnatwa} or @option{-gnatwu}. The
5173 default is that these warnings are not generated.
5176 @emph{Suppress warnings on unreferenced formals.}
5177 @cindex @option{-gnatwF} (@command{gcc})
5178 This switch suppresses warnings for unreferenced formal
5179 parameters. Note that the
5180 combination @option{-gnatwu} followed by @option{-gnatwF} has the
5181 effect of warning on unreferenced entities other than subprogram
5185 @emph{Activate warnings on unrecognized pragmas.}
5186 @cindex @option{-gnatwg} (@command{gcc})
5187 @cindex Pragmas, unrecognized
5188 This switch causes a warning to be generated if an unrecognized
5189 pragma is encountered. Apart from issuing this warning, the
5190 pragma is ignored and has no effect. This warning can
5191 also be turned on using @option{-gnatwa}. The default
5192 is that such warnings are issued (satisfying the Ada Reference
5193 Manual requirement that such warnings appear).
5196 @emph{Suppress warnings on unrecognized pragmas.}
5197 @cindex @option{-gnatwG} (@command{gcc})
5198 This switch suppresses warnings for unrecognized pragmas.
5201 @emph{Warnings used for GNAT sources}
5202 @cindex @option{-gnatw.g} (@command{gcc})
5203 This switch sets the warning categories that are used by the standard
5204 GNAT style. Currently this is equivalent to
5205 @option{-gnatwAao.sI.C.V.X}
5206 but more warnings may be added in the future without advanced notice.
5209 @emph{Activate warnings on hiding.}
5210 @cindex @option{-gnatwh} (@command{gcc})
5211 @cindex Hiding of Declarations
5212 This switch activates warnings on hiding declarations.
5213 A declaration is considered hiding
5214 if it is for a non-overloadable entity, and it declares an entity with the
5215 same name as some other entity that is directly or use-visible. The default
5216 is that such warnings are not generated.
5217 Note that @option{-gnatwa} does not affect the setting of this warning option.
5220 @emph{Suppress warnings on hiding.}
5221 @cindex @option{-gnatwH} (@command{gcc})
5222 This switch suppresses warnings on hiding declarations.
5225 @emph{Activate warnings on holes/gaps in records.}
5226 @cindex @option{-gnatw.h} (@command{gcc})
5227 @cindex Record Representation (gaps)
5228 This switch activates warnings on component clauses in record
5229 representation clauses that leave holes (gaps) in the record layout.
5230 If this warning option is active, then record representation clauses
5231 should specify a contiguous layout, adding unused fill fields if needed.
5232 Note that @option{-gnatwa} does not affect the setting of this warning option.
5235 @emph{Suppress warnings on holes/gaps in records.}
5236 @cindex @option{-gnatw.H} (@command{gcc})
5237 This switch suppresses warnings on component clauses in record
5238 representation clauses that leave holes (haps) in the record layout.
5241 @emph{Activate warnings on implementation units.}
5242 @cindex @option{-gnatwi} (@command{gcc})
5243 This switch activates warnings for a @code{with} of an internal GNAT
5244 implementation unit, defined as any unit from the @code{Ada},
5245 @code{Interfaces}, @code{GNAT},
5246 ^^@code{DEC},^ or @code{System}
5247 hierarchies that is not
5248 documented in either the Ada Reference Manual or the GNAT
5249 Programmer's Reference Manual. Such units are intended only
5250 for internal implementation purposes and should not be @code{with}'ed
5251 by user programs. The default is that such warnings are generated
5252 This warning can also be turned on using @option{-gnatwa}.
5255 @emph{Disable warnings on implementation units.}
5256 @cindex @option{-gnatwI} (@command{gcc})
5257 This switch disables warnings for a @code{with} of an internal GNAT
5258 implementation unit.
5261 @emph{Activate warnings on overlapping actuals.}
5262 @cindex @option{-gnatw.i} (@command{gcc})
5263 This switch enables a warning on statically detectable overlapping actuals in
5264 a subprogram call, when one of the actuals is an in-out parameter, and the
5265 types of the actuals are not by-copy types. The warning is off by default,
5266 and is not included under -gnatwa.
5269 @emph{Disable warnings on overlapping actuals.}
5270 @cindex @option{-gnatw.I} (@command{gcc})
5271 This switch disables warnings on overlapping actuals in a call..
5274 @emph{Activate warnings on obsolescent features (Annex J).}
5275 @cindex @option{-gnatwj} (@command{gcc})
5276 @cindex Features, obsolescent
5277 @cindex Obsolescent features
5278 If this warning option is activated, then warnings are generated for
5279 calls to subprograms marked with @code{pragma Obsolescent} and
5280 for use of features in Annex J of the Ada Reference Manual. In the
5281 case of Annex J, not all features are flagged. In particular use
5282 of the renamed packages (like @code{Text_IO}) and use of package
5283 @code{ASCII} are not flagged, since these are very common and
5284 would generate many annoying positive warnings. The default is that
5285 such warnings are not generated. This warning is also turned on by
5286 the use of @option{-gnatwa}.
5288 In addition to the above cases, warnings are also generated for
5289 GNAT features that have been provided in past versions but which
5290 have been superseded (typically by features in the new Ada standard).
5291 For example, @code{pragma Ravenscar} will be flagged since its
5292 function is replaced by @code{pragma Profile(Ravenscar)}, and
5293 @code{pragma Interface_Name} will be flagged since its function
5294 is replaced by @code{pragma Import}.
5296 Note that this warning option functions differently from the
5297 restriction @code{No_Obsolescent_Features} in two respects.
5298 First, the restriction applies only to annex J features.
5299 Second, the restriction does flag uses of package @code{ASCII}.
5302 @emph{Suppress warnings on obsolescent features (Annex J).}
5303 @cindex @option{-gnatwJ} (@command{gcc})
5304 This switch disables warnings on use of obsolescent features.
5307 @emph{Activate warnings on variables that could be constants.}
5308 @cindex @option{-gnatwk} (@command{gcc})
5309 This switch activates warnings for variables that are initialized but
5310 never modified, and then could be declared constants. The default is that
5311 such warnings are not given.
5312 This warning can also be turned on using @option{-gnatwa}.
5315 @emph{Suppress warnings on variables that could be constants.}
5316 @cindex @option{-gnatwK} (@command{gcc})
5317 This switch disables warnings on variables that could be declared constants.
5320 @emph{Activate warnings on redefinition of names in standard.}
5321 @cindex @option{-gnatw.k} (@command{gcc})
5322 This switch activates warnings for declarations that declare a name that
5323 is defined in package Standard. Such declarations can be confusing,
5324 especially since the names in package Standard continue to be directly
5325 visible, meaning that use visibiliy on such redeclared names does not
5326 work as expected. Names of discriminants and components in records are
5327 not included in this check.
5328 This warning is not part of the warnings activated by @option{-gnatwa}.
5329 It must be explicitly activated.
5332 @emph{Suppress warnings on variables that could be constants.}
5333 @cindex @option{-gnatwK} (@command{gcc})
5334 This switch activates warnings for declarations that declare a name that
5335 is defined in package Standard.
5338 @emph{Activate warnings for elaboration pragmas.}
5339 @cindex @option{-gnatwl} (@command{gcc})
5340 @cindex Elaboration, warnings
5341 This switch activates warnings on missing
5342 for possible elaboration problems, including suspicious use
5343 of @code{Elaborate} pragmas, when using the static elaboration model, and
5344 possible situations that may raise @code{Program_Error} when using the
5345 dynamic elaboration model.
5346 See the section in this guide on elaboration checking for further details.
5347 The default is that such warnings
5349 This warning is not automatically turned on by the use of @option{-gnatwa}.
5352 @emph{Suppress warnings for elaboration pragmas.}
5353 @cindex @option{-gnatwL} (@command{gcc})
5354 This switch suppresses warnings for possible elaboration problems.
5357 @emph{List inherited aspects.}
5358 @cindex @option{-gnatw.l} (@command{gcc})
5359 This switch causes the compiler to list inherited invariants,
5360 preconditions, and postconditions from Type_Invariant'Class, Invariant'Class,
5361 Pre'Class, and Post'Class aspects. Also list inherited subtype predicates.
5362 These messages are not automatically turned on by the use of @option{-gnatwa}.
5365 @emph{Suppress listing of inherited aspects.}
5366 @cindex @option{-gnatw.L} (@command{gcc})
5367 This switch suppresses listing of inherited aspects.
5370 @emph{Activate warnings on modified but unreferenced variables.}
5371 @cindex @option{-gnatwm} (@command{gcc})
5372 This switch activates warnings for variables that are assigned (using
5373 an initialization value or with one or more assignment statements) but
5374 whose value is never read. The warning is suppressed for volatile
5375 variables and also for variables that are renamings of other variables
5376 or for which an address clause is given.
5377 This warning can also be turned on using @option{-gnatwa}.
5378 The default is that these warnings are not given.
5381 @emph{Disable warnings on modified but unreferenced variables.}
5382 @cindex @option{-gnatwM} (@command{gcc})
5383 This switch disables warnings for variables that are assigned or
5384 initialized, but never read.
5387 @emph{Activate warnings on suspicious modulus values.}
5388 @cindex @option{-gnatw.m} (@command{gcc})
5389 This switch activates warnings for modulus values that seem suspicious.
5390 The cases caught are where the size is the same as the modulus (e.g.
5391 a modulus of 7 with a size of 7 bits), and modulus values of 32 or 64
5392 with no size clause. The guess in both cases is that 2**x was intended
5393 rather than x. In addition expressions of the form 2*x for small x
5394 generate a warning (the almost certainly accurate guess being that
5395 2**x was intended). The default is that these warnings are given.
5398 @emph{Disable warnings on suspicious modulus values.}
5399 @cindex @option{-gnatw.M} (@command{gcc})
5400 This switch disables warnings for suspicious modulus values.
5403 @emph{Set normal warnings mode.}
5404 @cindex @option{-gnatwn} (@command{gcc})
5405 This switch sets normal warning mode, in which enabled warnings are
5406 issued and treated as warnings rather than errors. This is the default
5407 mode. the switch @option{-gnatwn} can be used to cancel the effect of
5408 an explicit @option{-gnatws} or
5409 @option{-gnatwe}. It also cancels the effect of the
5410 implicit @option{-gnatwe} that is activated by the
5411 use of @option{-gnatg}.
5414 @emph{Activate warnings on atomic synchronization.}
5415 @cindex @option{-gnatw.n} (@command{gcc})
5416 @cindex Atomic Synchronization, warnings
5417 This switch actives warnings when an access to an atomic variable
5418 requires the generation of atomic synchronization code. These
5419 warnings are off by default and this warning is not included
5423 @emph{Suppress warnings on atomic synchronization.}
5424 @cindex @option{-gnatw.n} (@command{gcc})
5425 @cindex Atomic Synchronization, warnings
5426 This switch suppresses warnings when an access to an atomic variable
5427 requires the generation of atomic synchronization code.
5430 @emph{Activate warnings on address clause overlays.}
5431 @cindex @option{-gnatwo} (@command{gcc})
5432 @cindex Address Clauses, warnings
5433 This switch activates warnings for possibly unintended initialization
5434 effects of defining address clauses that cause one variable to overlap
5435 another. The default is that such warnings are generated.
5436 This warning can also be turned on using @option{-gnatwa}.
5439 @emph{Suppress warnings on address clause overlays.}
5440 @cindex @option{-gnatwO} (@command{gcc})
5441 This switch suppresses warnings on possibly unintended initialization
5442 effects of defining address clauses that cause one variable to overlap
5446 @emph{Activate warnings on modified but unreferenced out parameters.}
5447 @cindex @option{-gnatw.o} (@command{gcc})
5448 This switch activates warnings for variables that are modified by using
5449 them as actuals for a call to a procedure with an out mode formal, where
5450 the resulting assigned value is never read. It is applicable in the case
5451 where there is more than one out mode formal. If there is only one out
5452 mode formal, the warning is issued by default (controlled by -gnatwu).
5453 The warning is suppressed for volatile
5454 variables and also for variables that are renamings of other variables
5455 or for which an address clause is given.
5456 The default is that these warnings are not given. Note that this warning
5457 is not included in -gnatwa, it must be activated explicitly.
5460 @emph{Disable warnings on modified but unreferenced out parameters.}
5461 @cindex @option{-gnatw.O} (@command{gcc})
5462 This switch suppresses warnings for variables that are modified by using
5463 them as actuals for a call to a procedure with an out mode formal, where
5464 the resulting assigned value is never read.
5467 @emph{Activate warnings on ineffective pragma Inlines.}
5468 @cindex @option{-gnatwp} (@command{gcc})
5469 @cindex Inlining, warnings
5470 This switch activates warnings for failure of front end inlining
5471 (activated by @option{-gnatN}) to inline a particular call. There are
5472 many reasons for not being able to inline a call, including most
5473 commonly that the call is too complex to inline. The default is
5474 that such warnings are not given.
5475 This warning can also be turned on using @option{-gnatwa}.
5476 Warnings on ineffective inlining by the gcc back-end can be activated
5477 separately, using the gcc switch -Winline.
5480 @emph{Suppress warnings on ineffective pragma Inlines.}
5481 @cindex @option{-gnatwP} (@command{gcc})
5482 This switch suppresses warnings on ineffective pragma Inlines. If the
5483 inlining mechanism cannot inline a call, it will simply ignore the
5487 @emph{Activate warnings on parameter ordering.}
5488 @cindex @option{-gnatw.p} (@command{gcc})
5489 @cindex Parameter order, warnings
5490 This switch activates warnings for cases of suspicious parameter
5491 ordering when the list of arguments are all simple identifiers that
5492 match the names of the formals, but are in a different order. The
5493 warning is suppressed if any use of named parameter notation is used,
5494 so this is the appropriate way to suppress a false positive (and
5495 serves to emphasize that the "misordering" is deliberate). The
5497 that such warnings are not given.
5498 This warning can also be turned on using @option{-gnatwa}.
5501 @emph{Suppress warnings on parameter ordering.}
5502 @cindex @option{-gnatw.P} (@command{gcc})
5503 This switch suppresses warnings on cases of suspicious parameter
5507 @emph{Activate warnings on questionable missing parentheses.}
5508 @cindex @option{-gnatwq} (@command{gcc})
5509 @cindex Parentheses, warnings
5510 This switch activates warnings for cases where parentheses are not used and
5511 the result is potential ambiguity from a readers point of view. For example
5512 (not a > b) when a and b are modular means ((not a) > b) and very likely the
5513 programmer intended (not (a > b)). Similarly (-x mod 5) means (-(x mod 5)) and
5514 quite likely ((-x) mod 5) was intended. In such situations it seems best to
5515 follow the rule of always parenthesizing to make the association clear, and
5516 this warning switch warns if such parentheses are not present. The default
5517 is that these warnings are given.
5518 This warning can also be turned on using @option{-gnatwa}.
5521 @emph{Suppress warnings on questionable missing parentheses.}
5522 @cindex @option{-gnatwQ} (@command{gcc})
5523 This switch suppresses warnings for cases where the association is not
5524 clear and the use of parentheses is preferred.
5527 @emph{Activate warnings on redundant constructs.}
5528 @cindex @option{-gnatwr} (@command{gcc})
5529 This switch activates warnings for redundant constructs. The following
5530 is the current list of constructs regarded as redundant:
5534 Assignment of an item to itself.
5536 Type conversion that converts an expression to its own type.
5538 Use of the attribute @code{Base} where @code{typ'Base} is the same
5541 Use of pragma @code{Pack} when all components are placed by a record
5542 representation clause.
5544 Exception handler containing only a reraise statement (raise with no
5545 operand) which has no effect.
5547 Use of the operator abs on an operand that is known at compile time
5550 Comparison of boolean expressions to an explicit True value.
5553 This warning can also be turned on using @option{-gnatwa}.
5554 The default is that warnings for redundant constructs are not given.
5557 @emph{Suppress warnings on redundant constructs.}
5558 @cindex @option{-gnatwR} (@command{gcc})
5559 This switch suppresses warnings for redundant constructs.
5562 @emph{Activate warnings for object renaming function.}
5563 @cindex @option{-gnatw.r} (@command{gcc})
5564 This switch activates warnings for an object renaming that renames a
5565 function call, which is equivalent to a constant declaration (as
5566 opposed to renaming the function itself). The default is that these
5567 warnings are given. This warning can also be turned on using
5571 @emph{Suppress warnings for object renaming function.}
5572 @cindex @option{-gnatwT} (@command{gcc})
5573 This switch suppresses warnings for object renaming function.
5576 @emph{Suppress all warnings.}
5577 @cindex @option{-gnatws} (@command{gcc})
5578 This switch completely suppresses the
5579 output of all warning messages from the GNAT front end, including
5580 both warnings that can be controlled by switches described in this
5581 section, and those that are normally given unconditionally. The
5582 effect of this suppress action can only be cancelled by a subsequent
5583 use of the switch @option{-gnatwn}.
5585 Note that switch @option{-gnatws} does not suppress
5586 warnings from the @command{gcc} back end.
5587 To suppress these back end warnings as well, use the switch @option{-w}
5588 in addition to @option{-gnatws}. Also this switch has no effect on the
5589 handling of style check messages.
5592 @emph{Activate warnings on overridden size clauses.}
5593 @cindex @option{-gnatw.s} (@command{gcc})
5594 @cindex Record Representation (component sizes)
5595 This switch activates warnings on component clauses in record
5596 representation clauses where the length given overrides that
5597 specified by an explicit size clause for the component type. A
5598 warning is similarly given in the array case if a specified
5599 component size overrides an explicit size clause for the array
5601 Note that @option{-gnatwa} does not affect the setting of this warning option.
5604 @emph{Suppress warnings on overridden size clauses.}
5605 @cindex @option{-gnatw.S} (@command{gcc})
5606 This switch suppresses warnings on component clauses in record
5607 representation clauses that override size clauses, and similar
5608 warnings when an array component size overrides a size clause.
5611 @emph{Activate warnings for tracking of deleted conditional code.}
5612 @cindex @option{-gnatwt} (@command{gcc})
5613 @cindex Deactivated code, warnings
5614 @cindex Deleted code, warnings
5615 This switch activates warnings for tracking of code in conditionals (IF and
5616 CASE statements) that is detected to be dead code which cannot be executed, and
5617 which is removed by the front end. This warning is off by default, and is not
5618 turned on by @option{-gnatwa}, it has to be turned on explicitly. This may be
5619 useful for detecting deactivated code in certified applications.
5622 @emph{Suppress warnings for tracking of deleted conditional code.}
5623 @cindex @option{-gnatwT} (@command{gcc})
5624 This switch suppresses warnings for tracking of deleted conditional code.
5627 @emph{Activate warnings on suspicious contracts.}
5628 @cindex @option{-gnatw.t} (@command{gcc})
5629 This switch activates warnings on suspicious postconditions (whether a
5630 pragma @code{Postcondition} or a @code{Post} aspect in Ada 2012)
5631 and suspicious contract cases (pragma @code{Contract_Cases}). A
5632 function postcondition or contract case is suspicious when no postcondition
5633 or contract case for this function mentions the result of the function.
5634 A procedure postcondition or contract case is suspicious when it only
5635 refers to the pre-state of the procedure, because in that case it should
5636 rather be expressed as a precondition. The default is that such warnings
5637 are not generated. This warning can also be turned on using @option{-gnatwa}.
5640 @emph{Suppress warnings on suspicious contracts.}
5641 @cindex @option{-gnatw.T} (@command{gcc})
5642 This switch suppresses warnings on suspicious postconditions.
5645 @emph{Activate warnings on unused entities.}
5646 @cindex @option{-gnatwu} (@command{gcc})
5647 This switch activates warnings to be generated for entities that
5648 are declared but not referenced, and for units that are @code{with}'ed
5650 referenced. In the case of packages, a warning is also generated if
5651 no entities in the package are referenced. This means that if a with'ed
5652 package is referenced but the only references are in @code{use}
5653 clauses or @code{renames}
5654 declarations, a warning is still generated. A warning is also generated
5655 for a generic package that is @code{with}'ed but never instantiated.
5656 In the case where a package or subprogram body is compiled, and there
5657 is a @code{with} on the corresponding spec
5658 that is only referenced in the body,
5659 a warning is also generated, noting that the
5660 @code{with} can be moved to the body. The default is that
5661 such warnings are not generated.
5662 This switch also activates warnings on unreferenced formals
5663 (it includes the effect of @option{-gnatwf}).
5664 This warning can also be turned on using @option{-gnatwa}.
5667 @emph{Suppress warnings on unused entities.}
5668 @cindex @option{-gnatwU} (@command{gcc})
5669 This switch suppresses warnings for unused entities and packages.
5670 It also turns off warnings on unreferenced formals (and thus includes
5671 the effect of @option{-gnatwF}).
5674 @emph{Activate warnings on unordered enumeration types.}
5675 @cindex @option{-gnatw.u} (@command{gcc})
5676 This switch causes enumeration types to be considered as conceptually
5677 unordered, unless an explicit pragma @code{Ordered} is given for the type.
5678 The effect is to generate warnings in clients that use explicit comparisons
5679 or subranges, since these constructs both treat objects of the type as
5680 ordered. (A @emph{client} is defined as a unit that is other than the unit in
5681 which the type is declared, or its body or subunits.) Please refer to
5682 the description of pragma @code{Ordered} in the
5683 @cite{@value{EDITION} Reference Manual} for further details.
5684 The default is that such warnings are not generated.
5685 This warning is not automatically turned on by the use of @option{-gnatwa}.
5688 @emph{Deactivate warnings on unordered enumeration types.}
5689 @cindex @option{-gnatw.U} (@command{gcc})
5690 This switch causes all enumeration types to be considered as ordered, so
5691 that no warnings are given for comparisons or subranges for any type.
5694 @emph{Activate warnings on unassigned variables.}
5695 @cindex @option{-gnatwv} (@command{gcc})
5696 @cindex Unassigned variable warnings
5697 This switch activates warnings for access to variables which
5698 may not be properly initialized. The default is that
5699 such warnings are generated.
5700 This warning can also be turned on using @option{-gnatwa}.
5703 @emph{Suppress warnings on unassigned variables.}
5704 @cindex @option{-gnatwV} (@command{gcc})
5705 This switch suppresses warnings for access to variables which
5706 may not be properly initialized.
5707 For variables of a composite type, the warning can also be suppressed in
5708 Ada 2005 by using a default initialization with a box. For example, if
5709 Table is an array of records whose components are only partially uninitialized,
5710 then the following code:
5712 @smallexample @c ada
5713 Tab : Table := (others => <>);
5716 will suppress warnings on subsequent statements that access components
5720 @emph{Activate info messages for non-default bit order.}
5721 @cindex @option{-gnatw.v} (@command{gcc})
5722 @cindex bit order warnings
5723 This switch activates messages (labeled "info", they are not warnings,
5724 just informational messages) about the effects of non-default bit-order
5725 on records to which a component clause is applied. The effect of specifying
5726 non-default bit ordering is a bit subtle (and changed with Ada 2005), so
5727 these messages, which are given by default, are useful in understanding the
5728 exact consequences of using this feature. These messages
5729 can also be turned on using @option{-gnatwa}
5732 @emph{Suppress info messages for non-default bit order.}
5733 @cindex @option{-gnatw.V} (@command{gcc})
5734 This switch suppresses information messages for the effects of specifying
5735 non-default bit order on record components with component clauses.
5738 @emph{Activate warnings on wrong low bound assumption.}
5739 @cindex @option{-gnatww} (@command{gcc})
5740 @cindex String indexing warnings
5741 This switch activates warnings for indexing an unconstrained string parameter
5742 with a literal or S'Length. This is a case where the code is assuming that the
5743 low bound is one, which is in general not true (for example when a slice is
5744 passed). The default is that such warnings are generated.
5745 This warning can also be turned on using @option{-gnatwa}.
5748 @emph{Suppress warnings on wrong low bound assumption.}
5749 @cindex @option{-gnatwW} (@command{gcc})
5750 This switch suppresses warnings for indexing an unconstrained string parameter
5751 with a literal or S'Length. Note that this warning can also be suppressed
5752 in a particular case by adding an
5753 assertion that the lower bound is 1,
5754 as shown in the following example.
5756 @smallexample @c ada
5757 procedure K (S : String) is
5758 pragma Assert (S'First = 1);
5763 @emph{Activate warnings on Warnings Off pragmas}
5764 @cindex @option{-gnatw.w} (@command{gcc})
5765 @cindex Warnings Off control
5766 This switch activates warnings for use of @code{pragma Warnings (Off, entity)}
5767 where either the pragma is entirely useless (because it suppresses no
5768 warnings), or it could be replaced by @code{pragma Unreferenced} or
5769 @code{pragma Unmodified}. The default is that these warnings are not given.
5770 Note that this warning is not included in -gnatwa, it must be
5771 activated explicitly. Also activates warnings for the case of
5772 Warnings (Off, String), where either there is no matching
5773 Warnings (On, String), or the Warnings (Off) did not suppress any warning.
5776 @emph{Suppress warnings on unnecessary Warnings Off pragmas}
5777 @cindex @option{-gnatw.W} (@command{gcc})
5778 This switch suppresses warnings for use of @code{pragma Warnings (Off, ...)}.
5781 @emph{Activate warnings on Export/Import pragmas.}
5782 @cindex @option{-gnatwx} (@command{gcc})
5783 @cindex Export/Import pragma warnings
5784 This switch activates warnings on Export/Import pragmas when
5785 the compiler detects a possible conflict between the Ada and
5786 foreign language calling sequences. For example, the use of
5787 default parameters in a convention C procedure is dubious
5788 because the C compiler cannot supply the proper default, so
5789 a warning is issued. The default is that such warnings are
5791 This warning can also be turned on using @option{-gnatwa}.
5794 @emph{Suppress warnings on Export/Import pragmas.}
5795 @cindex @option{-gnatwX} (@command{gcc})
5796 This switch suppresses warnings on Export/Import pragmas.
5797 The sense of this is that you are telling the compiler that
5798 you know what you are doing in writing the pragma, and it
5799 should not complain at you.
5802 @emph{Activate warnings for No_Exception_Propagation mode.}
5803 @cindex @option{-gnatwm} (@command{gcc})
5804 This switch activates warnings for exception usage when pragma Restrictions
5805 (No_Exception_Propagation) is in effect. Warnings are given for implicit or
5806 explicit exception raises which are not covered by a local handler, and for
5807 exception handlers which do not cover a local raise. The default is that these
5808 warnings are not given.
5811 @emph{Disable warnings for No_Exception_Propagation mode.}
5812 This switch disables warnings for exception usage when pragma Restrictions
5813 (No_Exception_Propagation) is in effect.
5816 @emph{Activate warnings for Ada compatibility issues.}
5817 @cindex @option{-gnatwy} (@command{gcc})
5818 @cindex Ada compatibility issues warnings
5819 For the most part, newer versions of Ada are upwards compatible
5820 with older versions. For example, Ada 2005 programs will almost
5821 always work when compiled as Ada 2012.
5822 However there are some exceptions (for example the fact that
5823 @code{some} is now a reserved word in Ada 2012). This
5824 switch activates several warnings to help in identifying
5825 and correcting such incompatibilities. The default is that
5826 these warnings are generated. Note that at one point Ada 2005
5827 was called Ada 0Y, hence the choice of character.
5828 This warning can also be turned on using @option{-gnatwa}.
5831 @emph{Disable warnings for Ada compatibility issues.}
5832 @cindex @option{-gnatwY} (@command{gcc})
5833 @cindex Ada compatibility issues warnings
5834 This switch suppresses the warnings intended to help in identifying
5835 incompatibilities between Ada language versions.
5838 @emph{Activate information messages for why package spec needs body}
5839 @cindex @option{-gnatw.y} (@command{gcc})
5840 @cindex Package spec needing body
5841 There are a number of cases in which a package spec needs a body.
5842 For example, the use of pragma Elaborate_Body, or the declaration
5843 of a procedure specification requiring a completion. This switch
5844 causes information messages to be output showing why a package
5845 specification requires a body. This can be useful in the case of
5846 a large package specification which is unexpectedly requiring a
5847 body. The default is that such information messages are not output.
5850 @emph{Disable information messages for why package spec needs body}
5851 @cindex @option{-gnatw.Y} (@command{gcc})
5852 @cindex No information messages for why package spec needs body
5853 This switch suppresses the output of information messages showing why
5854 a package specification needs a body.
5857 @emph{Activate warnings on unchecked conversions.}
5858 @cindex @option{-gnatwz} (@command{gcc})
5859 @cindex Unchecked_Conversion warnings
5860 This switch activates warnings for unchecked conversions
5861 where the types are known at compile time to have different
5863 is that such warnings are generated. Warnings are also
5864 generated for subprogram pointers with different conventions,
5865 and, on VMS only, for data pointers with different conventions.
5866 This warning can also be turned on using @option{-gnatwa}.
5869 @emph{Suppress warnings on unchecked conversions.}
5870 @cindex @option{-gnatwZ} (@command{gcc})
5871 This switch suppresses warnings for unchecked conversions
5872 where the types are known at compile time to have different
5873 sizes or conventions.
5876 @emph{Activate warnings for size not a multiple of alignment.}
5877 @cindex @option{-gnatw.z} (@command{gcc})
5878 @cindex Size/Alignment warnings
5879 This switch activates warnings for cases of record types with
5880 specified @code{Size} and @code{Alignment} attributes where the
5881 size is not a multiple of the alignment, resulting in an object
5882 size that is greater than the specified size. The default
5883 is that such warnings are generated.
5884 This warning can also be turned on using @option{-gnatwa}.
5887 @emph{Suppress warnings for size not a multiple of alignment.}
5888 @cindex @option{-gnatw.Z} (@command{gcc})
5889 @cindex Size/Alignment warnings
5890 This switch suppresses warnings for cases of record types with
5891 specified @code{Size} and @code{Alignment} attributes where the
5892 size is not a multiple of the alignment, resulting in an object
5893 size that is greater than the specified size.
5894 The warning can also be
5895 suppressed by giving an explicit @code{Object_Size} value.
5897 @item ^-Wunused^WARNINGS=UNUSED^
5898 @cindex @option{-Wunused}
5899 The warnings controlled by the @option{-gnatw} switch are generated by
5900 the front end of the compiler. The @option{GCC} back end can provide
5901 additional warnings and they are controlled by the @option{-W} switch.
5902 For example, @option{^-Wunused^WARNINGS=UNUSED^} activates back end
5903 warnings for entities that are declared but not referenced.
5905 @item ^-Wuninitialized^WARNINGS=UNINITIALIZED^
5906 @cindex @option{-Wuninitialized}
5907 Similarly, @option{^-Wuninitialized^WARNINGS=UNINITIALIZED^} activates
5908 the back end warning for uninitialized variables. This switch must be
5909 used in conjunction with an optimization level greater than zero.
5911 @item -Wstack-usage=@var{len}
5912 @cindex @option{-Wstack-usage}
5913 Warn if the stack usage of a subprogram might be larger than @var{len} bytes.
5914 See @ref{Static Stack Usage Analysis} for details.
5916 @item ^-Wall^/ALL_BACK_END_WARNINGS^
5917 @cindex @option{-Wall}
5918 This switch enables most warnings from the @option{GCC} back end.
5919 The code generator detects a number of warning situations that are missed
5920 by the @option{GNAT} front end, and this switch can be used to activate them.
5921 The use of this switch also sets the default front end warning mode to
5922 @option{-gnatwa}, that is, most front end warnings activated as well.
5924 @item ^-w^/NO_BACK_END_WARNINGS^
5926 Conversely, this switch suppresses warnings from the @option{GCC} back end.
5927 The use of this switch also sets the default front end warning mode to
5928 @option{-gnatws}, that is, front end warnings suppressed as well.
5931 @cindex @option{-Werror}
5932 This switch causes warnings from the @option{GCC} back end to be treated as
5933 errors. The warning string still appears, but the warning messages are
5934 counted as errors, and prevent the generation of an object file.
5940 A string of warning parameters can be used in the same parameter. For example:
5947 will turn on all optional warnings except for unrecognized pragma warnings,
5948 and also specify that warnings should be treated as errors.
5951 When no switch @option{^-gnatw^/WARNINGS^} is used, this is equivalent to:
5994 @node Debugging and Assertion Control
5995 @subsection Debugging and Assertion Control
5999 @cindex @option{-gnata} (@command{gcc})
6005 The pragmas @code{Assert} and @code{Debug} normally have no effect and
6006 are ignored. This switch, where @samp{a} stands for assert, causes
6007 @code{Assert} and @code{Debug} pragmas to be activated.
6009 The pragmas have the form:
6013 @b{pragma} Assert (@var{Boolean-expression} @r{[},
6014 @var{static-string-expression}@r{]})
6015 @b{pragma} Debug (@var{procedure call})
6020 The @code{Assert} pragma causes @var{Boolean-expression} to be tested.
6021 If the result is @code{True}, the pragma has no effect (other than
6022 possible side effects from evaluating the expression). If the result is
6023 @code{False}, the exception @code{Assert_Failure} declared in the package
6024 @code{System.Assertions} is
6025 raised (passing @var{static-string-expression}, if present, as the
6026 message associated with the exception). If no string expression is
6027 given the default is a string giving the file name and line number
6030 The @code{Debug} pragma causes @var{procedure} to be called. Note that
6031 @code{pragma Debug} may appear within a declaration sequence, allowing
6032 debugging procedures to be called between declarations.
6035 @item /DEBUG@r{[}=debug-level@r{]}
6037 Specifies how much debugging information is to be included in
6038 the resulting object file where 'debug-level' is one of the following:
6041 Include both debugger symbol records and traceback
6043 This is the default setting.
6045 Include both debugger symbol records and traceback in
6048 Excludes both debugger symbol records and traceback
6049 the object file. Same as /NODEBUG.
6051 Includes only debugger symbol records in the object
6052 file. Note that this doesn't include traceback information.
6057 @node Validity Checking
6058 @subsection Validity Checking
6059 @findex Validity Checking
6062 The Ada Reference Manual defines the concept of invalid values (see
6063 RM 13.9.1). The primary source of invalid values is uninitialized
6064 variables. A scalar variable that is left uninitialized may contain
6065 an invalid value; the concept of invalid does not apply to access or
6068 It is an error to read an invalid value, but the RM does not require
6069 run-time checks to detect such errors, except for some minimal
6070 checking to prevent erroneous execution (i.e. unpredictable
6071 behavior). This corresponds to the @option{-gnatVd} switch below,
6072 which is the default. For example, by default, if the expression of a
6073 case statement is invalid, it will raise Constraint_Error rather than
6074 causing a wild jump, and if an array index on the left-hand side of an
6075 assignment is invalid, it will raise Constraint_Error rather than
6076 overwriting an arbitrary memory location.
6078 The @option{-gnatVa} may be used to enable additional validity checks,
6079 which are not required by the RM. These checks are often very
6080 expensive (which is why the RM does not require them). These checks
6081 are useful in tracking down uninitialized variables, but they are
6082 not usually recommended for production builds, and in particular
6083 we do not recommend using these extra validity checking options in
6084 combination with optimization, since this can confuse the optimizer.
6085 If performance is a consideration, leading to the need to optimize,
6086 then the validity checking options should not be used.
6088 The other @option{-gnatV^@var{x}^^} switches below allow finer-grained
6089 control; you can enable whichever validity checks you desire. However,
6090 for most debugging purposes, @option{-gnatVa} is sufficient, and the
6091 default @option{-gnatVd} (i.e. standard Ada behavior) is usually
6092 sufficient for non-debugging use.
6094 The @option{-gnatB} switch tells the compiler to assume that all
6095 values are valid (that is, within their declared subtype range)
6096 except in the context of a use of the Valid attribute. This means
6097 the compiler can generate more efficient code, since the range
6098 of values is better known at compile time. However, an uninitialized
6099 variable can cause wild jumps and memory corruption in this mode.
6101 The @option{-gnatV^@var{x}^^} switch allows control over the validity
6102 checking mode as described below.
6104 The @code{x} argument is a string of letters that
6105 indicate validity checks that are performed or not performed in addition
6106 to the default checks required by Ada as described above.
6109 The options allowed for this qualifier
6110 indicate validity checks that are performed or not performed in addition
6111 to the default checks required by Ada as described above.
6117 @emph{All validity checks.}
6118 @cindex @option{-gnatVa} (@command{gcc})
6119 All validity checks are turned on.
6121 That is, @option{-gnatVa} is
6122 equivalent to @option{gnatVcdfimorst}.
6126 @emph{Validity checks for copies.}
6127 @cindex @option{-gnatVc} (@command{gcc})
6128 The right hand side of assignments, and the initializing values of
6129 object declarations are validity checked.
6132 @emph{Default (RM) validity checks.}
6133 @cindex @option{-gnatVd} (@command{gcc})
6134 Some validity checks are done by default following normal Ada semantics
6136 A check is done in case statements that the expression is within the range
6137 of the subtype. If it is not, Constraint_Error is raised.
6138 For assignments to array components, a check is done that the expression used
6139 as index is within the range. If it is not, Constraint_Error is raised.
6140 Both these validity checks may be turned off using switch @option{-gnatVD}.
6141 They are turned on by default. If @option{-gnatVD} is specified, a subsequent
6142 switch @option{-gnatVd} will leave the checks turned on.
6143 Switch @option{-gnatVD} should be used only if you are sure that all such
6144 expressions have valid values. If you use this switch and invalid values
6145 are present, then the program is erroneous, and wild jumps or memory
6146 overwriting may occur.
6149 @emph{Validity checks for elementary components.}
6150 @cindex @option{-gnatVe} (@command{gcc})
6151 In the absence of this switch, assignments to record or array components are
6152 not validity checked, even if validity checks for assignments generally
6153 (@option{-gnatVc}) are turned on. In Ada, assignment of composite values do not
6154 require valid data, but assignment of individual components does. So for
6155 example, there is a difference between copying the elements of an array with a
6156 slice assignment, compared to assigning element by element in a loop. This
6157 switch allows you to turn off validity checking for components, even when they
6158 are assigned component by component.
6161 @emph{Validity checks for floating-point values.}
6162 @cindex @option{-gnatVf} (@command{gcc})
6163 In the absence of this switch, validity checking occurs only for discrete
6164 values. If @option{-gnatVf} is specified, then validity checking also applies
6165 for floating-point values, and NaNs and infinities are considered invalid,
6166 as well as out of range values for constrained types. Note that this means
6167 that standard IEEE infinity mode is not allowed. The exact contexts
6168 in which floating-point values are checked depends on the setting of other
6169 options. For example,
6170 @option{^-gnatVif^VALIDITY_CHECKING=(IN_PARAMS,FLOATS)^} or
6171 @option{^-gnatVfi^VALIDITY_CHECKING=(FLOATS,IN_PARAMS)^}
6172 (the order does not matter) specifies that floating-point parameters of mode
6173 @code{in} should be validity checked.
6176 @emph{Validity checks for @code{in} mode parameters}
6177 @cindex @option{-gnatVi} (@command{gcc})
6178 Arguments for parameters of mode @code{in} are validity checked in function
6179 and procedure calls at the point of call.
6182 @emph{Validity checks for @code{in out} mode parameters.}
6183 @cindex @option{-gnatVm} (@command{gcc})
6184 Arguments for parameters of mode @code{in out} are validity checked in
6185 procedure calls at the point of call. The @code{'m'} here stands for
6186 modify, since this concerns parameters that can be modified by the call.
6187 Note that there is no specific option to test @code{out} parameters,
6188 but any reference within the subprogram will be tested in the usual
6189 manner, and if an invalid value is copied back, any reference to it
6190 will be subject to validity checking.
6193 @emph{No validity checks.}
6194 @cindex @option{-gnatVn} (@command{gcc})
6195 This switch turns off all validity checking, including the default checking
6196 for case statements and left hand side subscripts. Note that the use of
6197 the switch @option{-gnatp} suppresses all run-time checks, including
6198 validity checks, and thus implies @option{-gnatVn}. When this switch
6199 is used, it cancels any other @option{-gnatV} previously issued.
6202 @emph{Validity checks for operator and attribute operands.}
6203 @cindex @option{-gnatVo} (@command{gcc})
6204 Arguments for predefined operators and attributes are validity checked.
6205 This includes all operators in package @code{Standard},
6206 the shift operators defined as intrinsic in package @code{Interfaces}
6207 and operands for attributes such as @code{Pos}. Checks are also made
6208 on individual component values for composite comparisons, and on the
6209 expressions in type conversions and qualified expressions. Checks are
6210 also made on explicit ranges using @samp{..} (e.g.@: slices, loops etc).
6213 @emph{Validity checks for parameters.}
6214 @cindex @option{-gnatVp} (@command{gcc})
6215 This controls the treatment of parameters within a subprogram (as opposed
6216 to @option{-gnatVi} and @option{-gnatVm} which control validity testing
6217 of parameters on a call. If either of these call options is used, then
6218 normally an assumption is made within a subprogram that the input arguments
6219 have been validity checking at the point of call, and do not need checking
6220 again within a subprogram). If @option{-gnatVp} is set, then this assumption
6221 is not made, and parameters are not assumed to be valid, so their validity
6222 will be checked (or rechecked) within the subprogram.
6225 @emph{Validity checks for function returns.}
6226 @cindex @option{-gnatVr} (@command{gcc})
6227 The expression in @code{return} statements in functions is validity
6231 @emph{Validity checks for subscripts.}
6232 @cindex @option{-gnatVs} (@command{gcc})
6233 All subscripts expressions are checked for validity, whether they appear
6234 on the right side or left side (in default mode only left side subscripts
6235 are validity checked).
6238 @emph{Validity checks for tests.}
6239 @cindex @option{-gnatVt} (@command{gcc})
6240 Expressions used as conditions in @code{if}, @code{while} or @code{exit}
6241 statements are checked, as well as guard expressions in entry calls.
6246 The @option{-gnatV} switch may be followed by
6247 ^a string of letters^a list of options^
6248 to turn on a series of validity checking options.
6250 @option{^-gnatVcr^/VALIDITY_CHECKING=(COPIES, RETURNS)^}
6251 specifies that in addition to the default validity checking, copies and
6252 function return expressions are to be validity checked.
6253 In order to make it easier
6254 to specify the desired combination of effects,
6256 the upper case letters @code{CDFIMORST} may
6257 be used to turn off the corresponding lower case option.
6260 the prefix @code{NO} on an option turns off the corresponding validity
6263 @item @code{NOCOPIES}
6264 @item @code{NODEFAULT}
6265 @item @code{NOFLOATS}
6266 @item @code{NOIN_PARAMS}
6267 @item @code{NOMOD_PARAMS}
6268 @item @code{NOOPERANDS}
6269 @item @code{NORETURNS}
6270 @item @code{NOSUBSCRIPTS}
6271 @item @code{NOTESTS}
6275 @option{^-gnatVaM^/VALIDITY_CHECKING=(ALL, NOMOD_PARAMS)^}
6276 turns on all validity checking options except for
6277 checking of @code{@b{in out}} procedure arguments.
6279 The specification of additional validity checking generates extra code (and
6280 in the case of @option{-gnatVa} the code expansion can be substantial).
6281 However, these additional checks can be very useful in detecting
6282 uninitialized variables, incorrect use of unchecked conversion, and other
6283 errors leading to invalid values. The use of pragma @code{Initialize_Scalars}
6284 is useful in conjunction with the extra validity checking, since this
6285 ensures that wherever possible uninitialized variables have invalid values.
6287 See also the pragma @code{Validity_Checks} which allows modification of
6288 the validity checking mode at the program source level, and also allows for
6289 temporary disabling of validity checks.
6291 @node Style Checking
6292 @subsection Style Checking
6293 @findex Style checking
6296 The @option{-gnaty^x^(option,option,@dots{})^} switch
6297 @cindex @option{-gnaty} (@command{gcc})
6298 causes the compiler to
6299 enforce specified style rules. A limited set of style rules has been used
6300 in writing the GNAT sources themselves. This switch allows user programs
6301 to activate all or some of these checks. If the source program fails a
6302 specified style check, an appropriate message is given, preceded by
6303 the character sequence ``(style)''. This message does not prevent
6304 successful compilation (unless the @option{-gnatwe} switch is used).
6306 Note that this is by no means intended to be a general facility for
6307 checking arbitrary coding standards. It is simply an embedding of the
6308 style rules we have chosen for the GNAT sources. If you are starting
6309 a project which does not have established style standards, you may
6310 find it useful to adopt the entire set of GNAT coding standards, or
6311 some subset of them.
6313 If you already have an established set of coding
6314 standards, then the selected style checking options may
6315 indeed correspond to choices you have made, but for general checking
6316 of an existing set of coding rules, you should look to the gnatcheck
6317 tool, which is designed for that purpose.
6321 @code{(option,option,@dots{})} is a sequence of keywords
6324 The string @var{x} is a sequence of letters or digits
6326 indicating the particular style
6327 checks to be performed. The following checks are defined:
6332 @emph{Specify indentation level.}
6333 If a digit from 1-9 appears
6334 ^in the string after @option{-gnaty}^as an option for /STYLE_CHECKS^
6335 then proper indentation is checked, with the digit indicating the
6336 indentation level required. A value of zero turns off this style check.
6337 The general style of required indentation is as specified by
6338 the examples in the Ada Reference Manual. Full line comments must be
6339 aligned with the @code{--} starting on a column that is a multiple of
6340 the alignment level, or they may be aligned the same way as the following
6341 non-blank line (this is useful when full line comments appear in the middle
6342 of a statement, or they may be aligned with the source line on the previous
6346 @emph{Check attribute casing.}
6347 Attribute names, including the case of keywords such as @code{digits}
6348 used as attributes names, must be written in mixed case, that is, the
6349 initial letter and any letter following an underscore must be uppercase.
6350 All other letters must be lowercase.
6352 @item ^A^ARRAY_INDEXES^
6353 @emph{Use of array index numbers in array attributes.}
6354 When using the array attributes First, Last, Range,
6355 or Length, the index number must be omitted for one-dimensional arrays
6356 and is required for multi-dimensional arrays.
6359 @emph{Blanks not allowed at statement end.}
6360 Trailing blanks are not allowed at the end of statements. The purpose of this
6361 rule, together with h (no horizontal tabs), is to enforce a canonical format
6362 for the use of blanks to separate source tokens.
6364 @item ^B^BOOLEAN_OPERATORS^
6365 @emph{Check Boolean operators.}
6366 The use of AND/OR operators is not permitted except in the cases of modular
6367 operands, array operands, and simple stand-alone boolean variables or
6368 boolean constants. In all other cases @code{and then}/@code{or else} are
6372 @emph{Check comments, double space.}
6373 Comments must meet the following set of rules:
6378 The ``@code{--}'' that starts the column must either start in column one,
6379 or else at least one blank must precede this sequence.
6382 Comments that follow other tokens on a line must have at least one blank
6383 following the ``@code{--}'' at the start of the comment.
6386 Full line comments must have at least two blanks following the
6387 ``@code{--}'' that starts the comment, with the following exceptions.
6390 A line consisting only of the ``@code{--}'' characters, possibly preceded
6391 by blanks is permitted.
6394 A comment starting with ``@code{--x}'' where @code{x} is a special character
6396 This allows proper processing of the output generated by specialized tools
6397 including @command{gnatprep} (where ``@code{--!}'' is used) and the SPARK
6399 language (where ``@code{--#}'' is used). For the purposes of this rule, a
6400 special character is defined as being in one of the ASCII ranges
6401 @code{16#21#@dots{}16#2F#} or @code{16#3A#@dots{}16#3F#}.
6402 Note that this usage is not permitted
6403 in GNAT implementation units (i.e., when @option{-gnatg} is used).
6406 A line consisting entirely of minus signs, possibly preceded by blanks, is
6407 permitted. This allows the construction of box comments where lines of minus
6408 signs are used to form the top and bottom of the box.
6411 A comment that starts and ends with ``@code{--}'' is permitted as long as at
6412 least one blank follows the initial ``@code{--}''. Together with the preceding
6413 rule, this allows the construction of box comments, as shown in the following
6416 ---------------------------
6417 -- This is a box comment --
6418 -- with two text lines. --
6419 ---------------------------
6424 @emph{Check comments, single space.}
6425 This is identical to @code{^c^COMMENTS^} except that only one space
6426 is required following the @code{--} of a comment instead of two.
6428 @item ^d^DOS_LINE_ENDINGS^
6429 @emph{Check no DOS line terminators present.}
6430 All lines must be terminated by a single ASCII.LF
6431 character (in particular the DOS line terminator sequence CR/LF is not
6435 @emph{Check end/exit labels.}
6436 Optional labels on @code{end} statements ending subprograms and on
6437 @code{exit} statements exiting named loops, are required to be present.
6440 @emph{No form feeds or vertical tabs.}
6441 Neither form feeds nor vertical tab characters are permitted
6445 @emph{GNAT style mode.}
6446 The set of style check switches is set to match that used by the GNAT sources.
6447 This may be useful when developing code that is eventually intended to be
6448 incorporated into GNAT. Currently this is equivalent to @option{-gnatwydISux})
6449 but additional style switches may be added to this set in the future without
6453 @emph{No horizontal tabs.}
6454 Horizontal tab characters are not permitted in the source text.
6455 Together with the b (no blanks at end of line) check, this
6456 enforces a canonical form for the use of blanks to separate
6460 @emph{Check if-then layout.}
6461 The keyword @code{then} must appear either on the same
6462 line as corresponding @code{if}, or on a line on its own, lined
6463 up under the @code{if}.
6466 @emph{check mode IN keywords.}
6467 Mode @code{in} (the default mode) is not
6468 allowed to be given explicitly. @code{in out} is fine,
6469 but not @code{in} on its own.
6472 @emph{Check keyword casing.}
6473 All keywords must be in lower case (with the exception of keywords
6474 such as @code{digits} used as attribute names to which this check
6478 @emph{Check layout.}
6479 Layout of statement and declaration constructs must follow the
6480 recommendations in the Ada Reference Manual, as indicated by the
6481 form of the syntax rules. For example an @code{else} keyword must
6482 be lined up with the corresponding @code{if} keyword.
6484 There are two respects in which the style rule enforced by this check
6485 option are more liberal than those in the Ada Reference Manual. First
6486 in the case of record declarations, it is permissible to put the
6487 @code{record} keyword on the same line as the @code{type} keyword, and
6488 then the @code{end} in @code{end record} must line up under @code{type}.
6489 This is also permitted when the type declaration is split on two lines.
6490 For example, any of the following three layouts is acceptable:
6492 @smallexample @c ada
6515 Second, in the case of a block statement, a permitted alternative
6516 is to put the block label on the same line as the @code{declare} or
6517 @code{begin} keyword, and then line the @code{end} keyword up under
6518 the block label. For example both the following are permitted:
6520 @smallexample @c ada
6538 The same alternative format is allowed for loops. For example, both of
6539 the following are permitted:
6541 @smallexample @c ada
6543 Clear : while J < 10 loop
6554 @item ^Lnnn^MAX_NESTING=nnn^
6555 @emph{Set maximum nesting level.}
6556 The maximum level of nesting of constructs (including subprograms, loops,
6557 blocks, packages, and conditionals) may not exceed the given value
6558 @option{nnn}. A value of zero disconnects this style check.
6560 @item ^m^LINE_LENGTH^
6561 @emph{Check maximum line length.}
6562 The length of source lines must not exceed 79 characters, including
6563 any trailing blanks. The value of 79 allows convenient display on an
6564 80 character wide device or window, allowing for possible special
6565 treatment of 80 character lines. Note that this count is of
6566 characters in the source text. This means that a tab character counts
6567 as one character in this count and a wide character sequence counts as
6568 a single character (however many bytes are needed in the encoding).
6570 @item ^Mnnn^MAX_LENGTH=nnn^
6571 @emph{Set maximum line length.}
6572 The length of lines must not exceed the
6573 given value @option{nnn}. The maximum value that can be specified is 32767.
6574 If neither style option for setting the line length is used, then the
6575 default is 255. This also controls the maximum length of lexical elements,
6576 where the only restriction is that they must fit on a single line.
6578 @item ^n^STANDARD_CASING^
6579 @emph{Check casing of entities in Standard.}
6580 Any identifier from Standard must be cased
6581 to match the presentation in the Ada Reference Manual (for example,
6582 @code{Integer} and @code{ASCII.NUL}).
6585 @emph{Turn off all style checks.}
6586 All style check options are turned off.
6588 @item ^o^ORDERED_SUBPROGRAMS^
6589 @emph{Check order of subprogram bodies.}
6590 All subprogram bodies in a given scope
6591 (e.g.@: a package body) must be in alphabetical order. The ordering
6592 rule uses normal Ada rules for comparing strings, ignoring casing
6593 of letters, except that if there is a trailing numeric suffix, then
6594 the value of this suffix is used in the ordering (e.g.@: Junk2 comes
6597 @item ^O^OVERRIDING_INDICATORS^
6598 @emph{Check that overriding subprograms are explicitly marked as such.}
6599 The declaration of a primitive operation of a type extension that overrides
6600 an inherited operation must carry an overriding indicator.
6603 @emph{Check pragma casing.}
6604 Pragma names must be written in mixed case, that is, the
6605 initial letter and any letter following an underscore must be uppercase.
6606 All other letters must be lowercase. An exception is that SPARK_Mode is
6607 allowed as an alternative for Spark_Mode.
6609 @item ^r^REFERENCES^
6610 @emph{Check references.}
6611 All identifier references must be cased in the same way as the
6612 corresponding declaration. No specific casing style is imposed on
6613 identifiers. The only requirement is for consistency of references
6617 @emph{Check separate specs.}
6618 Separate declarations (``specs'') are required for subprograms (a
6619 body is not allowed to serve as its own declaration). The only
6620 exception is that parameterless library level procedures are
6621 not required to have a separate declaration. This exception covers
6622 the most frequent form of main program procedures.
6624 @item ^S^STATEMENTS_AFTER_THEN_ELSE^
6625 @emph{Check no statements after @code{then}/@code{else}.}
6626 No statements are allowed
6627 on the same line as a @code{then} or @code{else} keyword following the
6628 keyword in an @code{if} statement. @code{or else} and @code{and then} are not
6629 affected, and a special exception allows a pragma to appear after @code{else}.
6632 @emph{Check token spacing.}
6633 The following token spacing rules are enforced:
6638 The keywords @code{abs} and @code{not} must be followed by a space.
6641 The token @code{=>} must be surrounded by spaces.
6644 The token @code{<>} must be preceded by a space or a left parenthesis.
6647 Binary operators other than @code{**} must be surrounded by spaces.
6648 There is no restriction on the layout of the @code{**} binary operator.
6651 Colon must be surrounded by spaces.
6654 Colon-equal (assignment, initialization) must be surrounded by spaces.
6657 Comma must be the first non-blank character on the line, or be
6658 immediately preceded by a non-blank character, and must be followed
6662 If the token preceding a left parenthesis ends with a letter or digit, then
6663 a space must separate the two tokens.
6666 if the token following a right parenthesis starts with a letter or digit, then
6667 a space must separate the two tokens.
6670 A right parenthesis must either be the first non-blank character on
6671 a line, or it must be preceded by a non-blank character.
6674 A semicolon must not be preceded by a space, and must not be followed by
6675 a non-blank character.
6678 A unary plus or minus may not be followed by a space.
6681 A vertical bar must be surrounded by spaces.
6685 Exactly one blank (and no other white space) must appear between
6686 a @code{not} token and a following @code{in} token.
6688 @item ^u^UNNECESSARY_BLANK_LINES^
6689 @emph{Check unnecessary blank lines.}
6690 Unnecessary blank lines are not allowed. A blank line is considered
6691 unnecessary if it appears at the end of the file, or if more than
6692 one blank line occurs in sequence.
6694 @item ^x^XTRA_PARENS^
6695 @emph{Check extra parentheses.}
6696 Unnecessary extra level of parentheses (C-style) are not allowed
6697 around conditions in @code{if} statements, @code{while} statements and
6698 @code{exit} statements.
6700 @item ^y^ALL_BUILTIN^
6701 @emph{Set all standard style check options}
6702 This is equivalent to @code{gnaty3aAbcefhiklmnprst}, that is all checking
6703 options enabled with the exception of @option{-gnatyB}, @option{-gnatyd},
6704 @option{-gnatyI}, @option{-gnatyLnnn}, @option{-gnatyo}, @option{-gnatyO},
6705 @option{-gnatyS}, @option{-gnatyu}, and @option{-gnatyx}.
6709 @emph{Remove style check options}
6710 This causes any subsequent options in the string to act as canceling the
6711 corresponding style check option. To cancel maximum nesting level control,
6712 use @option{L} parameter witout any integer value after that, because any
6713 digit following @option{-} in the parameter string of the @option{-gnaty}
6714 option will be threated as canceling indentation check. The same is true
6715 for @option{M} parameter. @option{y} and @option{N} parameters are not
6716 allowed after @option{-}.
6719 This causes any subsequent options in the string to enable the corresponding
6720 style check option. That is, it cancels the effect of a previous ^-^REMOVE^,
6726 @emph{Removing style check options}
6727 If the name of a style check is preceded by @option{NO} then the corresponding
6728 style check is turned off. For example @option{NOCOMMENTS} turns off style
6729 checking for comments.
6734 In the above rules, appearing in column one is always permitted, that is,
6735 counts as meeting either a requirement for a required preceding space,
6736 or as meeting a requirement for no preceding space.
6738 Appearing at the end of a line is also always permitted, that is, counts
6739 as meeting either a requirement for a following space, or as meeting
6740 a requirement for no following space.
6743 If any of these style rules is violated, a message is generated giving
6744 details on the violation. The initial characters of such messages are
6745 always ``@code{(style)}''. Note that these messages are treated as warning
6746 messages, so they normally do not prevent the generation of an object
6747 file. The @option{-gnatwe} switch can be used to treat warning messages,
6748 including style messages, as fatal errors.
6752 @option{-gnaty} on its own (that is not
6753 followed by any letters or digits) is equivalent
6754 to the use of @option{-gnatyy} as described above, that is all
6755 built-in standard style check options are enabled.
6759 /STYLE_CHECKS=ALL_BUILTIN enables all checking options with
6760 the exception of ORDERED_SUBPROGRAMS, UNNECESSARY_BLANK_LINES,
6761 XTRA_PARENS, and DOS_LINE_ENDINGS. In addition
6771 clears any previously set style checks.
6773 @node Run-Time Checks
6774 @subsection Run-Time Checks
6775 @cindex Division by zero
6776 @cindex Access before elaboration
6777 @cindex Checks, division by zero
6778 @cindex Checks, access before elaboration
6779 @cindex Checks, stack overflow checking
6782 By default, the following checks are suppressed: integer overflow
6783 checks, stack overflow checks, and checks for access before
6784 elaboration on subprogram calls. All other checks, including range
6785 checks and array bounds checks, are turned on by default. The
6786 following @command{gcc} switches refine this default behavior.
6791 @cindex @option{-gnatp} (@command{gcc})
6792 @cindex Suppressing checks
6793 @cindex Checks, suppressing
6795 This switch causes the unit to be compiled
6796 as though @code{pragma Suppress (All_checks)}
6797 had been present in the source. Validity checks are also eliminated (in
6798 other words @option{-gnatp} also implies @option{-gnatVn}.
6799 Use this switch to improve the performance
6800 of the code at the expense of safety in the presence of invalid data or
6803 Note that when checks are suppressed, the compiler is allowed, but not
6804 required, to omit the checking code. If the run-time cost of the
6805 checking code is zero or near-zero, the compiler will generate it even
6806 if checks are suppressed. In particular, if the compiler can prove
6807 that a certain check will necessarily fail, it will generate code to
6808 do an unconditional ``raise'', even if checks are suppressed. The
6809 compiler warns in this case. Another case in which checks may not be
6810 eliminated is when they are embedded in certain run time routines such
6811 as math library routines.
6813 Of course, run-time checks are omitted whenever the compiler can prove
6814 that they will not fail, whether or not checks are suppressed.
6816 Note that if you suppress a check that would have failed, program
6817 execution is erroneous, which means the behavior is totally
6818 unpredictable. The program might crash, or print wrong answers, or
6819 do anything else. It might even do exactly what you wanted it to do
6820 (and then it might start failing mysteriously next week or next
6821 year). The compiler will generate code based on the assumption that
6822 the condition being checked is true, which can result in erroneous
6823 execution if that assumption is wrong.
6825 The checks subject to suppression include all the checks defined by
6826 the Ada standard, the additional implementation defined checks
6827 @code{Alignment_Check},
6828 @code{Duplicated_Tag_Check}, @code{Predicate_Check}, and
6829 @code{Validity_Check}, as well as any checks introduced using
6830 @code{pragma Check_Name}. Note that @code{Atomic_Synchronization}
6831 is not automatically suppressed by use of this option.
6833 If the code depends on certain checks being active, you can use
6834 pragma @code{Unsuppress} either as a configuration pragma or as
6835 a local pragma to make sure that a specified check is performed
6836 even if @option{gnatp} is specified.
6838 The @option{-gnatp} switch has no effect if a subsequent
6839 @option{-gnat-p} switch appears.
6842 @cindex @option{-gnat-p} (@command{gcc})
6843 @cindex Suppressing checks
6844 @cindex Checks, suppressing
6846 This switch cancels the effect of a previous @option{gnatp} switch.
6849 @cindex @option{-gnato??} (@command{gcc})
6850 @cindex Overflow checks
6851 @cindex Overflow mode
6852 @cindex Check, overflow
6853 This switch controls the mode used for computing intermediate
6854 arithmetic integer operations, and also enables overflow checking.
6855 For a full description of overflow mode and checking control, see
6856 the ``Overflow Check Handling in GNAT'' appendix in this
6859 Overflow checks are always enabled by this switch. The argument
6860 controls the mode, using the codes
6864 In STRICT mode, intermediate operations are always done using the
6865 base type, and overflow checking ensures that the result is within
6866 the base type range.
6869 In MINIMIZED mode, overflows in intermediate operations are avoided
6870 where possible by using a larger integer type for the computation
6871 (typically @code{Long_Long_Integer}). Overflow checking ensures that
6872 the result fits in this larger integer type.
6874 @item 3 = ELIMINATED
6875 In ELIMINATED mode, overflows in intermediate operations are avoided
6876 by using multi-precision arithmetic. In this case, overflow checking
6877 has no effect on intermediate operations (since overflow is impossible).
6880 If two digits are present after @option{-gnato} then the first digit
6881 sets the mode for expressions outside assertions, and the second digit
6882 sets the mode for expressions within assertions. Here assertions is used
6883 in the technical sense (which includes for example precondition and
6884 postcondition expressions).
6886 If one digit is present, the corresponding mode is applicable to both
6887 expressions within and outside assertion expressions.
6889 If no digits are present, the default is to enable overflow checks
6890 and set STRICT mode for both kinds of expressions. This is compatible
6891 with the use of @option{-gnato} in previous versions of GNAT.
6893 @findex Machine_Overflows
6894 Note that the @option{-gnato??} switch does not affect the code generated
6895 for any floating-point operations; it applies only to integer semantics.
6896 For floating-point, @value{EDITION} has the @code{Machine_Overflows}
6897 attribute set to @code{False} and the normal mode of operation is to
6898 generate IEEE NaN and infinite values on overflow or invalid operations
6899 (such as dividing 0.0 by 0.0).
6901 The reason that we distinguish overflow checking from other kinds of
6902 range constraint checking is that a failure of an overflow check, unlike
6903 for example the failure of a range check, can result in an incorrect
6904 value, but cannot cause random memory destruction (like an out of range
6905 subscript), or a wild jump (from an out of range case value). Overflow
6906 checking is also quite expensive in time and space, since in general it
6907 requires the use of double length arithmetic.
6909 Note again that the default is @option{^-gnato00^/OVERFLOW_CHECKS=00^},
6910 so overflow checking is not performed in default mode. This means that out of
6911 the box, with the default settings, @value{EDITION} does not do all the checks
6912 expected from the language description in the Ada Reference Manual.
6913 If you want all constraint checks to be performed, as described in this Manual,
6914 then you must explicitly use the @option{-gnato??}
6915 switch either on the @command{gnatmake} or @command{gcc} command.
6918 @cindex @option{-gnatE} (@command{gcc})
6919 @cindex Elaboration checks
6920 @cindex Check, elaboration
6921 Enables dynamic checks for access-before-elaboration
6922 on subprogram calls and generic instantiations.
6923 Note that @option{-gnatE} is not necessary for safety, because in the
6924 default mode, GNAT ensures statically that the checks would not fail.
6925 For full details of the effect and use of this switch,
6926 @xref{Compiling with gcc}.
6929 @cindex @option{-fstack-check} (@command{gcc})
6930 @cindex Stack Overflow Checking
6931 @cindex Checks, stack overflow checking
6932 Activates stack overflow checking. For full details of the effect and use of
6933 this switch see @ref{Stack Overflow Checking}.
6938 The setting of these switches only controls the default setting of the
6939 checks. You may modify them using either @code{Suppress} (to remove
6940 checks) or @code{Unsuppress} (to add back suppressed checks) pragmas in
6943 @node Using gcc for Syntax Checking
6944 @subsection Using @command{gcc} for Syntax Checking
6947 @cindex @option{-gnats} (@command{gcc})
6951 The @code{s} stands for ``syntax''.
6954 Run GNAT in syntax checking only mode. For
6955 example, the command
6958 $ gcc -c -gnats x.adb
6962 compiles file @file{x.adb} in syntax-check-only mode. You can check a
6963 series of files in a single command
6965 , and can use wild cards to specify such a group of files.
6966 Note that you must specify the @option{-c} (compile
6967 only) flag in addition to the @option{-gnats} flag.
6970 You may use other switches in conjunction with @option{-gnats}. In
6971 particular, @option{-gnatl} and @option{-gnatv} are useful to control the
6972 format of any generated error messages.
6974 When the source file is empty or contains only empty lines and/or comments,
6975 the output is a warning:
6978 $ gcc -c -gnats -x ada toto.txt
6979 toto.txt:1:01: warning: empty file, contains no compilation units
6983 Otherwise, the output is simply the error messages, if any. No object file or
6984 ALI file is generated by a syntax-only compilation. Also, no units other
6985 than the one specified are accessed. For example, if a unit @code{X}
6986 @code{with}'s a unit @code{Y}, compiling unit @code{X} in syntax
6987 check only mode does not access the source file containing unit
6990 @cindex Multiple units, syntax checking
6991 Normally, GNAT allows only a single unit in a source file. However, this
6992 restriction does not apply in syntax-check-only mode, and it is possible
6993 to check a file containing multiple compilation units concatenated
6994 together. This is primarily used by the @code{gnatchop} utility
6995 (@pxref{Renaming Files with gnatchop}).
6998 @node Using gcc for Semantic Checking
6999 @subsection Using @command{gcc} for Semantic Checking
7002 @cindex @option{-gnatc} (@command{gcc})
7006 The @code{c} stands for ``check''.
7008 Causes the compiler to operate in semantic check mode,
7009 with full checking for all illegalities specified in the
7010 Ada Reference Manual, but without generation of any object code
7011 (no object file is generated).
7013 Because dependent files must be accessed, you must follow the GNAT
7014 semantic restrictions on file structuring to operate in this mode:
7018 The needed source files must be accessible
7019 (@pxref{Search Paths and the Run-Time Library (RTL)}).
7022 Each file must contain only one compilation unit.
7025 The file name and unit name must match (@pxref{File Naming Rules}).
7028 The output consists of error messages as appropriate. No object file is
7029 generated. An @file{ALI} file is generated for use in the context of
7030 cross-reference tools, but this file is marked as not being suitable
7031 for binding (since no object file is generated).
7032 The checking corresponds exactly to the notion of
7033 legality in the Ada Reference Manual.
7035 Any unit can be compiled in semantics-checking-only mode, including
7036 units that would not normally be compiled (subunits,
7037 and specifications where a separate body is present).
7040 @node Compiling Different Versions of Ada
7041 @subsection Compiling Different Versions of Ada
7044 The switches described in this section allow you to explicitly specify
7045 the version of the Ada language that your programs are written in.
7046 The default mode is Ada 2012,
7047 but you can also specify Ada 95, Ada 2005 mode, or
7048 indicate Ada 83 compatibility mode.
7051 @cindex Compatibility with Ada 83
7053 @item -gnat83 (Ada 83 Compatibility Mode)
7054 @cindex @option{-gnat83} (@command{gcc})
7055 @cindex ACVC, Ada 83 tests
7059 Although GNAT is primarily an Ada 95 / Ada 2005 compiler, this switch
7060 specifies that the program is to be compiled in Ada 83 mode. With
7061 @option{-gnat83}, GNAT rejects most post-Ada 83 extensions and applies Ada 83
7062 semantics where this can be done easily.
7063 It is not possible to guarantee this switch does a perfect
7064 job; some subtle tests, such as are
7065 found in earlier ACVC tests (and that have been removed from the ACATS suite
7066 for Ada 95), might not compile correctly.
7067 Nevertheless, this switch may be useful in some circumstances, for example
7068 where, due to contractual reasons, existing code needs to be maintained
7069 using only Ada 83 features.
7071 With few exceptions (most notably the need to use @code{<>} on
7072 @cindex Generic formal parameters
7073 unconstrained generic formal parameters, the use of the new Ada 95 / Ada 2005
7074 reserved words, and the use of packages
7075 with optional bodies), it is not necessary to specify the
7076 @option{-gnat83} switch when compiling Ada 83 programs, because, with rare
7077 exceptions, Ada 95 and Ada 2005 are upwardly compatible with Ada 83. Thus
7078 a correct Ada 83 program is usually also a correct program
7079 in these later versions of the language standard.
7080 For further information, please refer to @ref{Compatibility and Porting Guide}.
7082 @item -gnat95 (Ada 95 mode)
7083 @cindex @option{-gnat95} (@command{gcc})
7087 This switch directs the compiler to implement the Ada 95 version of the
7089 Since Ada 95 is almost completely upwards
7090 compatible with Ada 83, Ada 83 programs may generally be compiled using
7091 this switch (see the description of the @option{-gnat83} switch for further
7092 information about Ada 83 mode).
7093 If an Ada 2005 program is compiled in Ada 95 mode,
7094 uses of the new Ada 2005 features will cause error
7095 messages or warnings.
7097 This switch also can be used to cancel the effect of a previous
7098 @option{-gnat83}, @option{-gnat05/2005}, or @option{-gnat12/2012}
7099 switch earlier in the command line.
7101 @item -gnat05 or -gnat2005 (Ada 2005 mode)
7102 @cindex @option{-gnat05} (@command{gcc})
7103 @cindex @option{-gnat2005} (@command{gcc})
7104 @cindex Ada 2005 mode
7107 This switch directs the compiler to implement the Ada 2005 version of the
7108 language, as documented in the official Ada standards document.
7109 Since Ada 2005 is almost completely upwards
7110 compatible with Ada 95 (and thus also with Ada 83), Ada 83 and Ada 95 programs
7111 may generally be compiled using this switch (see the description of the
7112 @option{-gnat83} and @option{-gnat95} switches for further
7115 @item -gnat12 or -gnat2012 (Ada 2012 mode)
7116 @cindex @option{-gnat12} (@command{gcc})
7117 @cindex @option{-gnat2012} (@command{gcc})
7118 @cindex Ada 2012 mode
7121 This switch directs the compiler to implement the Ada 2012 version of the
7122 language (also the default).
7123 Since Ada 2012 is almost completely upwards
7124 compatible with Ada 2005 (and thus also with Ada 83, and Ada 95),
7125 Ada 83 and Ada 95 programs
7126 may generally be compiled using this switch (see the description of the
7127 @option{-gnat83}, @option{-gnat95}, and @option{-gnat05/2005} switches
7128 for further information).
7130 @item -gnatX (Enable GNAT Extensions)
7131 @cindex @option{-gnatX} (@command{gcc})
7132 @cindex Ada language extensions
7133 @cindex GNAT extensions
7136 This switch directs the compiler to implement the latest version of the
7137 language (currently Ada 2012) and also to enable certain GNAT implementation
7138 extensions that are not part of any Ada standard. For a full list of these
7139 extensions, see the GNAT reference manual.
7143 @node Character Set Control
7144 @subsection Character Set Control
7146 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
7147 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
7150 Normally GNAT recognizes the Latin-1 character set in source program
7151 identifiers, as described in the Ada Reference Manual.
7153 GNAT to recognize alternate character sets in identifiers. @var{c} is a
7154 single character ^^or word^ indicating the character set, as follows:
7158 ISO 8859-1 (Latin-1) identifiers
7161 ISO 8859-2 (Latin-2) letters allowed in identifiers
7164 ISO 8859-3 (Latin-3) letters allowed in identifiers
7167 ISO 8859-4 (Latin-4) letters allowed in identifiers
7170 ISO 8859-5 (Cyrillic) letters allowed in identifiers
7173 ISO 8859-15 (Latin-9) letters allowed in identifiers
7176 IBM PC letters (code page 437) allowed in identifiers
7179 IBM PC letters (code page 850) allowed in identifiers
7181 @item ^f^FULL_UPPER^
7182 Full upper-half codes allowed in identifiers
7185 No upper-half codes allowed in identifiers
7188 Wide-character codes (that is, codes greater than 255)
7189 allowed in identifiers
7192 @xref{Foreign Language Representation}, for full details on the
7193 implementation of these character sets.
7195 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
7196 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
7197 Specify the method of encoding for wide characters.
7198 @var{e} is one of the following:
7203 Hex encoding (brackets coding also recognized)
7206 Upper half encoding (brackets encoding also recognized)
7209 Shift/JIS encoding (brackets encoding also recognized)
7212 EUC encoding (brackets encoding also recognized)
7215 UTF-8 encoding (brackets encoding also recognized)
7218 Brackets encoding only (default value)
7220 For full details on these encoding
7221 methods see @ref{Wide Character Encodings}.
7222 Note that brackets coding is always accepted, even if one of the other
7223 options is specified, so for example @option{-gnatW8} specifies that both
7224 brackets and UTF-8 encodings will be recognized. The units that are
7225 with'ed directly or indirectly will be scanned using the specified
7226 representation scheme, and so if one of the non-brackets scheme is
7227 used, it must be used consistently throughout the program. However,
7228 since brackets encoding is always recognized, it may be conveniently
7229 used in standard libraries, allowing these libraries to be used with
7230 any of the available coding schemes.
7232 Note that brackets encoding only applies to program text. Within comments,
7233 brackets are considered to be normal graphic characters, and bracket sequences
7234 are never recognized as wide characters.
7236 If no @option{-gnatW?} parameter is present, then the default
7237 representation is normally Brackets encoding only. However, if the
7238 first three characters of the file are 16#EF# 16#BB# 16#BF# (the standard
7239 byte order mark or BOM for UTF-8), then these three characters are
7240 skipped and the default representation for the file is set to UTF-8.
7242 Note that the wide character representation that is specified (explicitly
7243 or by default) for the main program also acts as the default encoding used
7244 for Wide_Text_IO files if not specifically overridden by a WCEM form
7249 When no @option{-gnatW?} is specified, then characters (other than wide
7250 characters represented using brackets notation) are treated as 8-bit
7251 Latin-1 codes. The codes recognized are the Latin-1 graphic characters,
7252 and ASCII format effectors (CR, LF, HT, VT). Other lower half control
7253 characters in the range 16#00#..16#1F# are not accepted in program text
7254 or in comments. Upper half control characters (16#80#..16#9F#) are rejected
7255 in program text, but allowed and ignored in comments. Note in particular
7256 that the Next Line (NEL) character whose encoding is 16#85# is not recognized
7257 as an end of line in this default mode. If your source program contains
7258 instances of the NEL character used as a line terminator,
7259 you must use UTF-8 encoding for the whole
7260 source program. In default mode, all lines must be ended by a standard
7261 end of line sequence (CR, CR/LF, or LF).
7263 Note that the convention of simply accepting all upper half characters in
7264 comments means that programs that use standard ASCII for program text, but
7265 UTF-8 encoding for comments are accepted in default mode, providing that the
7266 comments are ended by an appropriate (CR, or CR/LF, or LF) line terminator.
7267 This is a common mode for many programs with foreign language comments.
7269 @node File Naming Control
7270 @subsection File Naming Control
7273 @item ^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{n}
7274 @cindex @option{-gnatk} (@command{gcc})
7275 Activates file name ``krunching''. @var{n}, a decimal integer in the range
7276 1-999, indicates the maximum allowable length of a file name (not
7277 including the @file{.ads} or @file{.adb} extension). The default is not
7278 to enable file name krunching.
7280 For the source file naming rules, @xref{File Naming Rules}.
7283 @node Subprogram Inlining Control
7284 @subsection Subprogram Inlining Control
7289 @cindex @option{-gnatn} (@command{gcc})
7291 The @code{n} here is intended to suggest the first syllable of the
7294 GNAT recognizes and processes @code{Inline} pragmas. However, for the
7295 inlining to actually occur, optimization must be enabled and, in order
7296 to enable inlining of subprograms specified by pragma @code{Inline},
7297 you must also specify this switch.
7298 In the absence of this switch, GNAT does not attempt
7299 inlining and does not need to access the bodies of
7300 subprograms for which @code{pragma Inline} is specified if they are not
7301 in the current unit.
7303 You can optionally specify the inlining level: 1 for moderate inlining across
7304 modules, which is a good compromise between compilation times and performances
7305 at run time, or 2 for full inlining across modules, which may bring about
7306 longer compilation times. If no inlining level is specified, the compiler will
7307 pick it based on the optimization level: 1 for @option{-O1}, @option{-O2} or
7308 @option{-Os} and 2 for @option{-O3}.
7310 If you specify this switch the compiler will access these bodies,
7311 creating an extra source dependency for the resulting object file, and
7312 where possible, the call will be inlined.
7313 For further details on when inlining is possible
7314 see @ref{Inlining of Subprograms}.
7317 @cindex @option{-gnatN} (@command{gcc})
7318 This switch activates front-end inlining which also
7319 generates additional dependencies.
7321 When using a gcc-based back end (in practice this means using any version
7322 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
7323 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
7324 Historically front end inlining was more extensive than the gcc back end
7325 inlining, but that is no longer the case.
7328 @node Auxiliary Output Control
7329 @subsection Auxiliary Output Control
7333 @cindex @option{-gnatt} (@command{gcc})
7334 @cindex Writing internal trees
7335 @cindex Internal trees, writing to file
7336 Causes GNAT to write the internal tree for a unit to a file (with the
7337 extension @file{.adt}.
7338 This not normally required, but is used by separate analysis tools.
7340 these tools do the necessary compilations automatically, so you should
7341 not have to specify this switch in normal operation.
7342 Note that the combination of switches @option{-gnatct}
7343 generates a tree in the form required by ASIS applications.
7346 @cindex @option{-gnatu} (@command{gcc})
7347 Print a list of units required by this compilation on @file{stdout}.
7348 The listing includes all units on which the unit being compiled depends
7349 either directly or indirectly.
7352 @item -pass-exit-codes
7353 @cindex @option{-pass-exit-codes} (@command{gcc})
7354 If this switch is not used, the exit code returned by @command{gcc} when
7355 compiling multiple files indicates whether all source files have
7356 been successfully used to generate object files or not.
7358 When @option{-pass-exit-codes} is used, @command{gcc} exits with an extended
7359 exit status and allows an integrated development environment to better
7360 react to a compilation failure. Those exit status are:
7364 There was an error in at least one source file.
7366 At least one source file did not generate an object file.
7368 The compiler died unexpectedly (internal error for example).
7370 An object file has been generated for every source file.
7375 @node Debugging Control
7376 @subsection Debugging Control
7380 @cindex Debugging options
7383 @cindex @option{-gnatd} (@command{gcc})
7384 Activate internal debugging switches. @var{x} is a letter or digit, or
7385 string of letters or digits, which specifies the type of debugging
7386 outputs desired. Normally these are used only for internal development
7387 or system debugging purposes. You can find full documentation for these
7388 switches in the body of the @code{Debug} unit in the compiler source
7389 file @file{debug.adb}.
7393 @cindex @option{-gnatG} (@command{gcc})
7394 This switch causes the compiler to generate auxiliary output containing
7395 a pseudo-source listing of the generated expanded code. Like most Ada
7396 compilers, GNAT works by first transforming the high level Ada code into
7397 lower level constructs. For example, tasking operations are transformed
7398 into calls to the tasking run-time routines. A unique capability of GNAT
7399 is to list this expanded code in a form very close to normal Ada source.
7400 This is very useful in understanding the implications of various Ada
7401 usage on the efficiency of the generated code. There are many cases in
7402 Ada (e.g.@: the use of controlled types), where simple Ada statements can
7403 generate a lot of run-time code. By using @option{-gnatG} you can identify
7404 these cases, and consider whether it may be desirable to modify the coding
7405 approach to improve efficiency.
7407 The optional parameter @code{nn} if present after -gnatG specifies an
7408 alternative maximum line length that overrides the normal default of 72.
7409 This value is in the range 40-999999, values less than 40 being silently
7410 reset to 40. The equal sign is optional.
7412 The format of the output is very similar to standard Ada source, and is
7413 easily understood by an Ada programmer. The following special syntactic
7414 additions correspond to low level features used in the generated code that
7415 do not have any exact analogies in pure Ada source form. The following
7416 is a partial list of these special constructions. See the spec
7417 of package @code{Sprint} in file @file{sprint.ads} for a full list.
7419 If the switch @option{-gnatL} is used in conjunction with
7420 @cindex @option{-gnatL} (@command{gcc})
7421 @option{-gnatG}, then the original source lines are interspersed
7422 in the expanded source (as comment lines with the original line number).
7425 @item new @var{xxx} @r{[}storage_pool = @var{yyy}@r{]}
7426 Shows the storage pool being used for an allocator.
7428 @item at end @var{procedure-name};
7429 Shows the finalization (cleanup) procedure for a scope.
7431 @item (if @var{expr} then @var{expr} else @var{expr})
7432 Conditional expression equivalent to the @code{x?y:z} construction in C.
7434 @item @var{target}^^^(@var{source})
7435 A conversion with floating-point truncation instead of rounding.
7437 @item @var{target}?(@var{source})
7438 A conversion that bypasses normal Ada semantic checking. In particular
7439 enumeration types and fixed-point types are treated simply as integers.
7441 @item @var{target}?^^^(@var{source})
7442 Combines the above two cases.
7444 @item @var{x} #/ @var{y}
7445 @itemx @var{x} #mod @var{y}
7446 @itemx @var{x} #* @var{y}
7447 @itemx @var{x} #rem @var{y}
7448 A division or multiplication of fixed-point values which are treated as
7449 integers without any kind of scaling.
7451 @item free @var{expr} @r{[}storage_pool = @var{xxx}@r{]}
7452 Shows the storage pool associated with a @code{free} statement.
7454 @item [subtype or type declaration]
7455 Used to list an equivalent declaration for an internally generated
7456 type that is referenced elsewhere in the listing.
7458 @c @item freeze @var{type-name} @ovar{actions}
7459 @c Expanding @ovar macro inline (explanation in macro def comments)
7460 @item freeze @var{type-name} @r{[}@var{actions}@r{]}
7461 Shows the point at which @var{type-name} is frozen, with possible
7462 associated actions to be performed at the freeze point.
7464 @item reference @var{itype}
7465 Reference (and hence definition) to internal type @var{itype}.
7467 @item @var{function-name}! (@var{arg}, @var{arg}, @var{arg})
7468 Intrinsic function call.
7470 @item @var{label-name} : label
7471 Declaration of label @var{labelname}.
7473 @item #$ @var{subprogram-name}
7474 An implicit call to a run-time support routine
7475 (to meet the requirement of H.3.1(9) in a
7478 @item @var{expr} && @var{expr} && @var{expr} @dots{} && @var{expr}
7479 A multiple concatenation (same effect as @var{expr} & @var{expr} &
7480 @var{expr}, but handled more efficiently).
7482 @item [constraint_error]
7483 Raise the @code{Constraint_Error} exception.
7485 @item @var{expression}'reference
7486 A pointer to the result of evaluating @var{expression}.
7488 @item @var{target-type}!(@var{source-expression})
7489 An unchecked conversion of @var{source-expression} to @var{target-type}.
7491 @item [@var{numerator}/@var{denominator}]
7492 Used to represent internal real literals (that) have no exact
7493 representation in base 2-16 (for example, the result of compile time
7494 evaluation of the expression 1.0/27.0).
7498 @cindex @option{-gnatD} (@command{gcc})
7499 When used in conjunction with @option{-gnatG}, this switch causes
7500 the expanded source, as described above for
7501 @option{-gnatG} to be written to files with names
7502 @file{^xxx.dg^XXX_DG^}, where @file{xxx} is the normal file name,
7503 instead of to the standard output file. For
7504 example, if the source file name is @file{hello.adb}, then a file
7505 @file{^hello.adb.dg^HELLO.ADB_DG^} will be written. The debugging
7506 information generated by the @command{gcc} @option{^-g^/DEBUG^} switch
7507 will refer to the generated @file{^xxx.dg^XXX_DG^} file. This allows
7508 you to do source level debugging using the generated code which is
7509 sometimes useful for complex code, for example to find out exactly
7510 which part of a complex construction raised an exception. This switch
7511 also suppress generation of cross-reference information (see
7512 @option{-gnatx}) since otherwise the cross-reference information
7513 would refer to the @file{^.dg^.DG^} file, which would cause
7514 confusion since this is not the original source file.
7516 Note that @option{-gnatD} actually implies @option{-gnatG}
7517 automatically, so it is not necessary to give both options.
7518 In other words @option{-gnatD} is equivalent to @option{-gnatDG}).
7520 If the switch @option{-gnatL} is used in conjunction with
7521 @cindex @option{-gnatL} (@command{gcc})
7522 @option{-gnatDG}, then the original source lines are interspersed
7523 in the expanded source (as comment lines with the original line number).
7525 The optional parameter @code{nn} if present after -gnatD specifies an
7526 alternative maximum line length that overrides the normal default of 72.
7527 This value is in the range 40-999999, values less than 40 being silently
7528 reset to 40. The equal sign is optional.
7531 @cindex @option{-gnatr} (@command{gcc})
7532 @cindex pragma Restrictions
7533 This switch causes pragma Restrictions to be treated as Restriction_Warnings
7534 so that violation of restrictions causes warnings rather than illegalities.
7535 This is useful during the development process when new restrictions are added
7536 or investigated. The switch also causes pragma Profile to be treated as
7537 Profile_Warnings, and pragma Restricted_Run_Time and pragma Ravenscar set
7538 restriction warnings rather than restrictions.
7541 @item -gnatR@r{[}0@r{|}1@r{|}2@r{|}3@r{[}s@r{]]}
7542 @cindex @option{-gnatR} (@command{gcc})
7543 This switch controls output from the compiler of a listing showing
7544 representation information for declared types and objects. For
7545 @option{-gnatR0}, no information is output (equivalent to omitting
7546 the @option{-gnatR} switch). For @option{-gnatR1} (which is the default,
7547 so @option{-gnatR} with no parameter has the same effect), size and alignment
7548 information is listed for declared array and record types. For
7549 @option{-gnatR2}, size and alignment information is listed for all
7550 declared types and objects. The @code{Linker_Section} is also listed for any
7551 entity for which the @code{Linker_Section} is set explicitly or implicitly (the
7552 latter case occurs for objects of a type for which a @code{Linker_Section}
7555 Finally @option{-gnatR3} includes symbolic
7556 expressions for values that are computed at run time for
7557 variant records. These symbolic expressions have a mostly obvious
7558 format with #n being used to represent the value of the n'th
7559 discriminant. See source files @file{repinfo.ads/adb} in the
7560 @code{GNAT} sources for full details on the format of @option{-gnatR3}
7561 output. If the switch is followed by an s (e.g.@: @option{-gnatR2s}), then
7562 the output is to a file with the name @file{^file.rep^file_REP^} where
7563 file is the name of the corresponding source file.
7566 This form of the switch controls output of subprogram conventions
7567 and parameter passing mechanisms for all subprograms. A following
7568 @code{s} means output to a file as described above.
7571 @item /REPRESENTATION_INFO
7572 @cindex @option{/REPRESENTATION_INFO} (@command{gcc})
7573 This qualifier controls output from the compiler of a listing showing
7574 representation information for declared types and objects. For
7575 @option{/REPRESENTATION_INFO=NONE}, no information is output
7576 (equivalent to omitting the @option{/REPRESENTATION_INFO} qualifier).
7577 @option{/REPRESENTATION_INFO} without option is equivalent to
7578 @option{/REPRESENTATION_INFO=ARRAYS}.
7579 For @option{/REPRESENTATION_INFO=ARRAYS}, size and alignment
7580 information is listed for declared array and record types. For
7581 @option{/REPRESENTATION_INFO=OBJECTS}, size and alignment information
7582 is listed for all expression information for values that are computed
7583 at run time for variant records. These symbolic expressions have a mostly
7584 obvious format with #n being used to represent the value of the n'th
7585 discriminant. See source files @file{REPINFO.ADS/ADB} in the
7586 @code{GNAT} sources for full details on the format of
7587 @option{/REPRESENTATION_INFO=SYMBOLIC} output.
7588 If _FILE is added at the end of an option
7589 (e.g.@: @option{/REPRESENTATION_INFO=ARRAYS_FILE}),
7590 then the output is to a file with the name @file{file_REP} where
7591 file is the name of the corresponding source file.
7593 @item /REPRESENTATION_INFO=MECHANISMS
7594 This qualifier form controls output of subprogram conventions
7595 and parameter passing mechanisms for all subprograms. It is
7596 possible to append _FILE as described above to cause information
7597 to be written to a file.
7600 Note that it is possible for record components to have zero size. In
7601 this case, the component clause uses an obvious extension of permitted
7602 Ada syntax, for example @code{at 0 range 0 .. -1}.
7604 Representation information requires that code be generated (since it is the
7605 code generator that lays out complex data structures). If an attempt is made
7606 to output representation information when no code is generated, for example
7607 when a subunit is compiled on its own, then no information can be generated
7608 and the compiler outputs a message to this effect.
7611 @cindex @option{-gnatS} (@command{gcc})
7612 The use of the switch @option{-gnatS} for an
7613 Ada compilation will cause the compiler to output a
7614 representation of package Standard in a form very
7615 close to standard Ada. It is not quite possible to
7616 do this entirely in standard Ada (since new
7617 numeric base types cannot be created in standard
7618 Ada), but the output is easily
7619 readable to any Ada programmer, and is useful to
7620 determine the characteristics of target dependent
7621 types in package Standard.
7624 @cindex @option{-gnatx} (@command{gcc})
7625 Normally the compiler generates full cross-referencing information in
7626 the @file{ALI} file. This information is used by a number of tools,
7627 including @code{gnatfind} and @code{gnatxref}. The @option{-gnatx} switch
7628 suppresses this information. This saves some space and may slightly
7629 speed up compilation, but means that these tools cannot be used.
7632 @node Exception Handling Control
7633 @subsection Exception Handling Control
7636 GNAT uses two methods for handling exceptions at run-time. The
7637 @code{setjmp/longjmp} method saves the context when entering
7638 a frame with an exception handler. Then when an exception is
7639 raised, the context can be restored immediately, without the
7640 need for tracing stack frames. This method provides very fast
7641 exception propagation, but introduces significant overhead for
7642 the use of exception handlers, even if no exception is raised.
7644 The other approach is called ``zero cost'' exception handling.
7645 With this method, the compiler builds static tables to describe
7646 the exception ranges. No dynamic code is required when entering
7647 a frame containing an exception handler. When an exception is
7648 raised, the tables are used to control a back trace of the
7649 subprogram invocation stack to locate the required exception
7650 handler. This method has considerably poorer performance for
7651 the propagation of exceptions, but there is no overhead for
7652 exception handlers if no exception is raised. Note that in this
7653 mode and in the context of mixed Ada and C/C++ programming,
7654 to propagate an exception through a C/C++ code, the C/C++ code
7655 must be compiled with the @option{-funwind-tables} GCC's
7658 The following switches may be used to control which of the
7659 two exception handling methods is used.
7665 @cindex @option{--RTS=sjlj} (@command{gnatmake})
7666 This switch causes the setjmp/longjmp run-time (when available) to be used
7667 for exception handling. If the default
7668 mechanism for the target is zero cost exceptions, then
7669 this switch can be used to modify this default, and must be
7670 used for all units in the partition.
7671 This option is rarely used. One case in which it may be
7672 advantageous is if you have an application where exception
7673 raising is common and the overall performance of the
7674 application is improved by favoring exception propagation.
7677 @cindex @option{--RTS=zcx} (@command{gnatmake})
7678 @cindex Zero Cost Exceptions
7679 This switch causes the zero cost approach to be used
7680 for exception handling. If this is the default mechanism for the
7681 target (see below), then this switch is unneeded. If the default
7682 mechanism for the target is setjmp/longjmp exceptions, then
7683 this switch can be used to modify this default, and must be
7684 used for all units in the partition.
7685 This option can only be used if the zero cost approach
7686 is available for the target in use, otherwise it will generate an error.
7690 The same option @option{--RTS} must be used both for @command{gcc}
7691 and @command{gnatbind}. Passing this option to @command{gnatmake}
7692 (@pxref{Switches for gnatmake}) will ensure the required consistency
7693 through the compilation and binding steps.
7695 @node Units to Sources Mapping Files
7696 @subsection Units to Sources Mapping Files
7700 @item -gnatem=@var{path}
7701 @cindex @option{-gnatem} (@command{gcc})
7702 A mapping file is a way to communicate to the compiler two mappings:
7703 from unit names to file names (without any directory information) and from
7704 file names to path names (with full directory information). These mappings
7705 are used by the compiler to short-circuit the path search.
7707 The use of mapping files is not required for correct operation of the
7708 compiler, but mapping files can improve efficiency, particularly when
7709 sources are read over a slow network connection. In normal operation,
7710 you need not be concerned with the format or use of mapping files,
7711 and the @option{-gnatem} switch is not a switch that you would use
7712 explicitly. It is intended primarily for use by automatic tools such as
7713 @command{gnatmake} running under the project file facility. The
7714 description here of the format of mapping files is provided
7715 for completeness and for possible use by other tools.
7717 A mapping file is a sequence of sets of three lines. In each set, the
7718 first line is the unit name, in lower case, with @code{%s} appended
7719 for specs and @code{%b} appended for bodies; the second line is the
7720 file name; and the third line is the path name.
7726 /gnat/project1/sources/main.2.ada
7729 When the switch @option{-gnatem} is specified, the compiler will
7730 create in memory the two mappings from the specified file. If there is
7731 any problem (nonexistent file, truncated file or duplicate entries),
7732 no mapping will be created.
7734 Several @option{-gnatem} switches may be specified; however, only the
7735 last one on the command line will be taken into account.
7737 When using a project file, @command{gnatmake} creates a temporary
7738 mapping file and communicates it to the compiler using this switch.
7742 @node Integrated Preprocessing
7743 @subsection Integrated Preprocessing
7746 GNAT sources may be preprocessed immediately before compilation.
7747 In this case, the actual
7748 text of the source is not the text of the source file, but is derived from it
7749 through a process called preprocessing. Integrated preprocessing is specified
7750 through switches @option{-gnatep} and/or @option{-gnateD}. @option{-gnatep}
7751 indicates, through a text file, the preprocessing data to be used.
7752 @option{-gnateD} specifies or modifies the values of preprocessing symbol.
7753 Note that integrated preprocessing applies only to Ada source files, it is
7754 not available for configuration pragma files.
7757 Note that when integrated preprocessing is used, the output from the
7758 preprocessor is not written to any external file. Instead it is passed
7759 internally to the compiler. If you need to preserve the result of
7760 preprocessing in a file, then you should use @command{gnatprep}
7761 to perform the desired preprocessing in stand-alone mode.
7764 It is recommended that @command{gnatmake} switch ^-s^/SWITCH_CHECK^ should be
7765 used when Integrated Preprocessing is used. The reason is that preprocessing
7766 with another Preprocessing Data file without changing the sources will
7767 not trigger recompilation without this switch.
7770 Note that @command{gnatmake} switch ^-m^/MINIMAL_RECOMPILATION^ will almost
7771 always trigger recompilation for sources that are preprocessed,
7772 because @command{gnatmake} cannot compute the checksum of the source after
7776 The actual preprocessing function is described in details in section
7777 @ref{Preprocessing with gnatprep}. This section only describes how integrated
7778 preprocessing is triggered and parameterized.
7782 @item -gnatep=@var{file}
7783 @cindex @option{-gnatep} (@command{gcc})
7784 This switch indicates to the compiler the file name (without directory
7785 information) of the preprocessor data file to use. The preprocessor data file
7786 should be found in the source directories. Note that when the compiler is
7787 called by a builder such as (@command{gnatmake} with a project
7788 file, if the object directory is not also a source directory, the builder needs
7789 to be called with @option{-x}.
7792 A preprocessing data file is a text file with significant lines indicating
7793 how should be preprocessed either a specific source or all sources not
7794 mentioned in other lines. A significant line is a nonempty, non-comment line.
7795 Comments are similar to Ada comments.
7798 Each significant line starts with either a literal string or the character '*'.
7799 A literal string is the file name (without directory information) of the source
7800 to preprocess. A character '*' indicates the preprocessing for all the sources
7801 that are not specified explicitly on other lines (order of the lines is not
7802 significant). It is an error to have two lines with the same file name or two
7803 lines starting with the character '*'.
7806 After the file name or the character '*', another optional literal string
7807 indicating the file name of the definition file to be used for preprocessing
7808 (@pxref{Form of Definitions File}). The definition files are found by the
7809 compiler in one of the source directories. In some cases, when compiling
7810 a source in a directory other than the current directory, if the definition
7811 file is in the current directory, it may be necessary to add the current
7812 directory as a source directory through switch ^-I.^/SEARCH=[]^, otherwise
7813 the compiler would not find the definition file.
7816 Then, optionally, ^switches^switches^ similar to those of @code{gnatprep} may
7817 be found. Those ^switches^switches^ are:
7822 Causes both preprocessor lines and the lines deleted by
7823 preprocessing to be replaced by blank lines, preserving the line number.
7824 This ^switch^switch^ is always implied; however, if specified after @option{-c}
7825 it cancels the effect of @option{-c}.
7828 Causes both preprocessor lines and the lines deleted
7829 by preprocessing to be retained as comments marked
7830 with the special string ``@code{--! }''.
7832 @item -Dsymbol=value
7833 Define or redefine a symbol, associated with value. A symbol is an Ada
7834 identifier, or an Ada reserved word, with the exception of @code{if},
7835 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
7836 @code{value} is either a literal string, an Ada identifier or any Ada reserved
7837 word. A symbol declared with this ^switch^switch^ replaces a symbol with the
7838 same name defined in a definition file.
7841 Causes a sorted list of symbol names and values to be
7842 listed on the standard output file.
7845 Causes undefined symbols to be treated as having the value @code{FALSE}
7847 of a preprocessor test. In the absence of this option, an undefined symbol in
7848 a @code{#if} or @code{#elsif} test will be treated as an error.
7853 Examples of valid lines in a preprocessor data file:
7856 "toto.adb" "prep.def" -u
7857 -- preprocess "toto.adb", using definition file "prep.def",
7858 -- undefined symbol are False.
7861 -- preprocess all other sources without a definition file;
7862 -- suppressed lined are commented; symbol VERSION has the value V101.
7864 "titi.adb" "prep2.def" -s
7865 -- preprocess "titi.adb", using definition file "prep2.def";
7866 -- list all symbols with their values.
7869 @item ^-gnateD^/DATA_PREPROCESSING=^symbol@r{[}=value@r{]}
7870 @cindex @option{-gnateD} (@command{gcc})
7871 Define or redefine a preprocessing symbol, associated with value. If no value
7872 is given on the command line, then the value of the symbol is @code{True}.
7873 A symbol is an identifier, following normal Ada (case-insensitive)
7874 rules for its syntax, and value is either an arbitrary string between double
7875 quotes or any sequence (including an empty sequence) of characters from the
7876 set (letters, digits, period, underline).
7877 Ada reserved words may be used as symbols, with the exceptions of @code{if},
7878 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
7887 -gnateDFoo=\"Foo-Bar\"
7892 A symbol declared with this ^switch^switch^ on the command line replaces a
7893 symbol with the same name either in a definition file or specified with a
7894 ^switch^switch^ -D in the preprocessor data file.
7897 This switch is similar to switch @option{^-D^/ASSOCIATE^} of @code{gnatprep}.
7900 When integrated preprocessing is performed and the preprocessor modifies
7901 the source text, write the result of this preprocessing into a file
7902 <source>^.prep^_prep^.
7906 @node Code Generation Control
7907 @subsection Code Generation Control
7911 The GCC technology provides a wide range of target dependent
7912 @option{-m} switches for controlling
7913 details of code generation with respect to different versions of
7914 architectures. This includes variations in instruction sets (e.g.@:
7915 different members of the power pc family), and different requirements
7916 for optimal arrangement of instructions (e.g.@: different members of
7917 the x86 family). The list of available @option{-m} switches may be
7918 found in the GCC documentation.
7920 Use of these @option{-m} switches may in some cases result in improved
7923 The @value{EDITION} technology is tested and qualified without any
7924 @option{-m} switches,
7925 so generally the most reliable approach is to avoid the use of these
7926 switches. However, we generally expect most of these switches to work
7927 successfully with @value{EDITION}, and many customers have reported successful
7928 use of these options.
7930 Our general advice is to avoid the use of @option{-m} switches unless
7931 special needs lead to requirements in this area. In particular,
7932 there is no point in using @option{-m} switches to improve performance
7933 unless you actually see a performance improvement.
7937 @subsection Return Codes
7938 @cindex Return Codes
7939 @cindex @option{/RETURN_CODES=VMS}
7942 On VMS, GNAT compiled programs return POSIX-style codes by default,
7943 e.g.@: @option{/RETURN_CODES=POSIX}.
7945 To enable VMS style return codes, use GNAT BIND and LINK with the option
7946 @option{/RETURN_CODES=VMS}. For example:
7949 GNAT BIND MYMAIN.ALI /RETURN_CODES=VMS
7950 GNAT LINK MYMAIN.ALI /RETURN_CODES=VMS
7954 Programs built with /RETURN_CODES=VMS are suitable to be called in
7955 VMS DCL scripts. Programs compiled with the default /RETURN_CODES=POSIX
7956 are suitable for spawning with appropriate GNAT RTL routines.
7960 @node Search Paths and the Run-Time Library (RTL)
7961 @section Search Paths and the Run-Time Library (RTL)
7964 With the GNAT source-based library system, the compiler must be able to
7965 find source files for units that are needed by the unit being compiled.
7966 Search paths are used to guide this process.
7968 The compiler compiles one source file whose name must be given
7969 explicitly on the command line. In other words, no searching is done
7970 for this file. To find all other source files that are needed (the most
7971 common being the specs of units), the compiler examines the following
7972 directories, in the following order:
7976 The directory containing the source file of the main unit being compiled
7977 (the file name on the command line).
7980 Each directory named by an @option{^-I^/SOURCE_SEARCH^} switch given on the
7981 @command{gcc} command line, in the order given.
7984 @findex ADA_PRJ_INCLUDE_FILE
7985 Each of the directories listed in the text file whose name is given
7986 by the @env{ADA_PRJ_INCLUDE_FILE} ^environment variable^logical name^.
7989 @env{ADA_PRJ_INCLUDE_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
7990 driver when project files are used. It should not normally be set
7994 @findex ADA_INCLUDE_PATH
7995 Each of the directories listed in the value of the
7996 @env{ADA_INCLUDE_PATH} ^environment variable^logical name^.
7998 Construct this value
7999 exactly as the @env{PATH} environment variable: a list of directory
8000 names separated by colons (semicolons when working with the NT version).
8003 Normally, define this value as a logical name containing a comma separated
8004 list of directory names.
8006 This variable can also be defined by means of an environment string
8007 (an argument to the HP C exec* set of functions).
8011 DEFINE ANOTHER_PATH FOO:[BAG]
8012 DEFINE ADA_INCLUDE_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
8015 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
8016 first, followed by the standard Ada
8017 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADAINCLUDE].
8018 If this is not redefined, the user will obtain the HP Ada 83 IO packages
8019 (Text_IO, Sequential_IO, etc)
8020 instead of the standard Ada packages. Thus, in order to get the standard Ada
8021 packages by default, ADA_INCLUDE_PATH must be redefined.
8025 The content of the @file{ada_source_path} file which is part of the GNAT
8026 installation tree and is used to store standard libraries such as the
8027 GNAT Run Time Library (RTL) source files.
8029 @ref{Installing a library}
8034 Specifying the switch @option{^-I-^/NOCURRENT_DIRECTORY^}
8035 inhibits the use of the directory
8036 containing the source file named in the command line. You can still
8037 have this directory on your search path, but in this case it must be
8038 explicitly requested with a @option{^-I^/SOURCE_SEARCH^} switch.
8040 Specifying the switch @option{-nostdinc}
8041 inhibits the search of the default location for the GNAT Run Time
8042 Library (RTL) source files.
8044 The compiler outputs its object files and ALI files in the current
8047 Caution: The object file can be redirected with the @option{-o} switch;
8048 however, @command{gcc} and @code{gnat1} have not been coordinated on this
8049 so the @file{ALI} file will not go to the right place. Therefore, you should
8050 avoid using the @option{-o} switch.
8054 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
8055 children make up the GNAT RTL, together with the simple @code{System.IO}
8056 package used in the @code{"Hello World"} example. The sources for these units
8057 are needed by the compiler and are kept together in one directory. Not
8058 all of the bodies are needed, but all of the sources are kept together
8059 anyway. In a normal installation, you need not specify these directory
8060 names when compiling or binding. Either the environment variables or
8061 the built-in defaults cause these files to be found.
8063 In addition to the language-defined hierarchies (@code{System}, @code{Ada} and
8064 @code{Interfaces}), the GNAT distribution provides a fourth hierarchy,
8065 consisting of child units of @code{GNAT}. This is a collection of generally
8066 useful types, subprograms, etc. @xref{Top, GNAT Reference Manual, About
8067 This Guid, gnat_rm, GNAT Reference Manual}, for further details.
8069 Besides simplifying access to the RTL, a major use of search paths is
8070 in compiling sources from multiple directories. This can make
8071 development environments much more flexible.
8073 @node Order of Compilation Issues
8074 @section Order of Compilation Issues
8077 If, in our earlier example, there was a spec for the @code{hello}
8078 procedure, it would be contained in the file @file{hello.ads}; yet this
8079 file would not have to be explicitly compiled. This is the result of the
8080 model we chose to implement library management. Some of the consequences
8081 of this model are as follows:
8085 There is no point in compiling specs (except for package
8086 specs with no bodies) because these are compiled as needed by clients. If
8087 you attempt a useless compilation, you will receive an error message.
8088 It is also useless to compile subunits because they are compiled as needed
8092 There are no order of compilation requirements: performing a
8093 compilation never obsoletes anything. The only way you can obsolete
8094 something and require recompilations is to modify one of the
8095 source files on which it depends.
8098 There is no library as such, apart from the ALI files
8099 (@pxref{The Ada Library Information Files}, for information on the format
8100 of these files). For now we find it convenient to create separate ALI files,
8101 but eventually the information therein may be incorporated into the object
8105 When you compile a unit, the source files for the specs of all units
8106 that it @code{with}'s, all its subunits, and the bodies of any generics it
8107 instantiates must be available (reachable by the search-paths mechanism
8108 described above), or you will receive a fatal error message.
8115 The following are some typical Ada compilation command line examples:
8118 @item $ gcc -c xyz.adb
8119 Compile body in file @file{xyz.adb} with all default options.
8122 @item $ gcc -c -O2 -gnata xyz-def.adb
8125 @item $ GNAT COMPILE /OPTIMIZE=ALL -gnata xyz-def.adb
8128 Compile the child unit package in file @file{xyz-def.adb} with extensive
8129 optimizations, and pragma @code{Assert}/@code{Debug} statements
8132 @item $ gcc -c -gnatc abc-def.adb
8133 Compile the subunit in file @file{abc-def.adb} in semantic-checking-only
8137 @node Binding with gnatbind
8138 @chapter Binding with @code{gnatbind}
8142 * Running gnatbind::
8143 * Switches for gnatbind::
8144 * Command-Line Access::
8145 * Search Paths for gnatbind::
8146 * Examples of gnatbind Usage::
8150 This chapter describes the GNAT binder, @code{gnatbind}, which is used
8151 to bind compiled GNAT objects.
8153 Note: to invoke @code{gnatbind} with a project file, use the @code{gnat}
8154 driver (see @ref{The GNAT Driver and Project Files}).
8156 The @code{gnatbind} program performs four separate functions:
8160 Checks that a program is consistent, in accordance with the rules in
8161 Chapter 10 of the Ada Reference Manual. In particular, error
8162 messages are generated if a program uses inconsistent versions of a
8166 Checks that an acceptable order of elaboration exists for the program
8167 and issues an error message if it cannot find an order of elaboration
8168 that satisfies the rules in Chapter 10 of the Ada Language Manual.
8171 Generates a main program incorporating the given elaboration order.
8172 This program is a small Ada package (body and spec) that
8173 must be subsequently compiled
8174 using the GNAT compiler. The necessary compilation step is usually
8175 performed automatically by @command{gnatlink}. The two most important
8176 functions of this program
8177 are to call the elaboration routines of units in an appropriate order
8178 and to call the main program.
8181 Determines the set of object files required by the given main program.
8182 This information is output in the forms of comments in the generated program,
8183 to be read by the @command{gnatlink} utility used to link the Ada application.
8186 @node Running gnatbind
8187 @section Running @code{gnatbind}
8190 The form of the @code{gnatbind} command is
8193 @c $ gnatbind @ovar{switches} @var{mainprog}@r{[}.ali@r{]} @ovar{switches}
8194 @c Expanding @ovar macro inline (explanation in macro def comments)
8195 $ gnatbind @r{[}@var{switches}@r{]} @var{mainprog}@r{[}.ali@r{]} @r{[}@var{switches}@r{]}
8199 where @file{@var{mainprog}.adb} is the Ada file containing the main program
8200 unit body. @code{gnatbind} constructs an Ada
8201 package in two files whose names are
8202 @file{b~@var{mainprog}.ads}, and @file{b~@var{mainprog}.adb}.
8203 For example, if given the
8204 parameter @file{hello.ali}, for a main program contained in file
8205 @file{hello.adb}, the binder output files would be @file{b~hello.ads}
8206 and @file{b~hello.adb}.
8208 When doing consistency checking, the binder takes into consideration
8209 any source files it can locate. For example, if the binder determines
8210 that the given main program requires the package @code{Pack}, whose
8212 file is @file{pack.ali} and whose corresponding source spec file is
8213 @file{pack.ads}, it attempts to locate the source file @file{pack.ads}
8214 (using the same search path conventions as previously described for the
8215 @command{gcc} command). If it can locate this source file, it checks that
8217 or source checksums of the source and its references to in @file{ALI} files
8218 match. In other words, any @file{ALI} files that mentions this spec must have
8219 resulted from compiling this version of the source file (or in the case
8220 where the source checksums match, a version close enough that the
8221 difference does not matter).
8223 @cindex Source files, use by binder
8224 The effect of this consistency checking, which includes source files, is
8225 that the binder ensures that the program is consistent with the latest
8226 version of the source files that can be located at bind time. Editing a
8227 source file without compiling files that depend on the source file cause
8228 error messages to be generated by the binder.
8230 For example, suppose you have a main program @file{hello.adb} and a
8231 package @code{P}, from file @file{p.ads} and you perform the following
8236 Enter @code{gcc -c hello.adb} to compile the main program.
8239 Enter @code{gcc -c p.ads} to compile package @code{P}.
8242 Edit file @file{p.ads}.
8245 Enter @code{gnatbind hello}.
8249 At this point, the file @file{p.ali} contains an out-of-date time stamp
8250 because the file @file{p.ads} has been edited. The attempt at binding
8251 fails, and the binder generates the following error messages:
8254 error: "hello.adb" must be recompiled ("p.ads" has been modified)
8255 error: "p.ads" has been modified and must be recompiled
8259 Now both files must be recompiled as indicated, and then the bind can
8260 succeed, generating a main program. You need not normally be concerned
8261 with the contents of this file, but for reference purposes a sample
8262 binder output file is given in @ref{Example of Binder Output File}.
8264 In most normal usage, the default mode of @command{gnatbind} which is to
8265 generate the main package in Ada, as described in the previous section.
8266 In particular, this means that any Ada programmer can read and understand
8267 the generated main program. It can also be debugged just like any other
8268 Ada code provided the @option{^-g^/DEBUG^} switch is used for
8269 @command{gnatbind} and @command{gnatlink}.
8271 @node Switches for gnatbind
8272 @section Switches for @command{gnatbind}
8275 The following switches are available with @code{gnatbind}; details will
8276 be presented in subsequent sections.
8279 * Consistency-Checking Modes::
8280 * Binder Error Message Control::
8281 * Elaboration Control::
8283 * Dynamic Allocation Control::
8284 * Binding with Non-Ada Main Programs::
8285 * Binding Programs with No Main Subprogram::
8292 @cindex @option{--version} @command{gnatbind}
8293 Display Copyright and version, then exit disregarding all other options.
8296 @cindex @option{--help} @command{gnatbind}
8297 If @option{--version} was not used, display usage, then exit disregarding
8301 @cindex @option{-a} @command{gnatbind}
8302 Indicates that, if supported by the platform, the adainit procedure should
8303 be treated as an initialisation routine by the linker (a constructor). This
8304 is intended to be used by the Project Manager to automatically initialize
8305 shared Stand-Alone Libraries.
8307 @item ^-aO^/OBJECT_SEARCH^
8308 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatbind})
8309 Specify directory to be searched for ALI files.
8311 @item ^-aI^/SOURCE_SEARCH^
8312 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
8313 Specify directory to be searched for source file.
8315 @item ^-A^/ALI_LIST^@r{[=}@var{filename}@r{]}
8316 @cindex @option{^-A^/ALI_LIST^} (@command{gnatbind})
8317 Output ALI list (to standard output or to the named file).
8319 @item ^-b^/REPORT_ERRORS=BRIEF^
8320 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@command{gnatbind})
8321 Generate brief messages to @file{stderr} even if verbose mode set.
8323 @item ^-c^/NOOUTPUT^
8324 @cindex @option{^-c^/NOOUTPUT^} (@command{gnatbind})
8325 Check only, no generation of binder output file.
8327 @item ^-d^/DEFAULT_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}
8328 @cindex @option{^-d^/DEFAULT_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}} (@command{gnatbind})
8329 This switch can be used to change the default task stack size value
8330 to a specified size @var{nn}, which is expressed in bytes by default, or
8331 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
8333 In the absence of a @samp{@r{[}k@r{|}m@r{]}} suffix, this switch is equivalent,
8334 in effect, to completing all task specs with
8335 @smallexample @c ada
8336 pragma Storage_Size (nn);
8338 When they do not already have such a pragma.
8340 @item ^-D^/DEFAULT_SECONDARY_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}
8341 @cindex @option{^-D^/DEFAULT_SECONDARY_STACK_SIZE=nnnnn^} (@command{gnatbind})
8342 This switch can be used to change the default secondary stack size value
8343 to a specified size @var{nn}, which is expressed in bytes by default, or
8344 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
8347 The secondary stack is used to deal with functions that return a variable
8348 sized result, for example a function returning an unconstrained
8349 String. There are two ways in which this secondary stack is allocated.
8351 For most targets, the secondary stack is growing on demand and is allocated
8352 as a chain of blocks in the heap. The -D option is not very
8353 relevant. It only give some control over the size of the allocated
8354 blocks (whose size is the minimum of the default secondary stack size value,
8355 and the actual size needed for the current allocation request).
8357 For certain targets, notably VxWorks 653,
8358 the secondary stack is allocated by carving off a fixed ratio chunk of the
8359 primary task stack. The -D option is used to define the
8360 size of the environment task's secondary stack.
8362 @item ^-e^/ELABORATION_DEPENDENCIES^
8363 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@command{gnatbind})
8364 Output complete list of elaboration-order dependencies.
8366 @item ^-E^/STORE_TRACEBACKS^
8367 @cindex @option{^-E^/STORE_TRACEBACKS^} (@command{gnatbind})
8368 Store tracebacks in exception occurrences when the target supports it.
8370 @c The following may get moved to an appendix
8371 This option is currently supported on the following targets:
8372 all x86 ports, Solaris, Windows, HP-UX, AIX, PowerPC VxWorks and Alpha VxWorks.
8374 See also the packages @code{GNAT.Traceback} and
8375 @code{GNAT.Traceback.Symbolic} for more information.
8377 Note that on x86 ports, you must not use @option{-fomit-frame-pointer}
8378 @command{gcc} option.
8381 @item ^-F^/FORCE_ELABS_FLAGS^
8382 @cindex @option{^-F^/FORCE_ELABS_FLAGS^} (@command{gnatbind})
8383 Force the checks of elaboration flags. @command{gnatbind} does not normally
8384 generate checks of elaboration flags for the main executable, except when
8385 a Stand-Alone Library is used. However, there are cases when this cannot be
8386 detected by gnatbind. An example is importing an interface of a Stand-Alone
8387 Library through a pragma Import and only specifying through a linker switch
8388 this Stand-Alone Library. This switch is used to guarantee that elaboration
8389 flag checks are generated.
8392 @cindex @option{^-h^/HELP^} (@command{gnatbind})
8393 Output usage (help) information
8395 @item ^-H32^/32_MALLOC^
8396 @cindex @option{^-H32^/32_MALLOC^} (@command{gnatbind})
8397 Use 32-bit allocations for @code{__gnat_malloc} (and thus for access types).
8398 For further details see @ref{Dynamic Allocation Control}.
8400 @item ^-H64^/64_MALLOC^
8401 @cindex @option{^-H64^/64_MALLOC^} (@command{gnatbind})
8402 Use 64-bit allocations for @code{__gnat_malloc} (and thus for access types).
8403 @cindex @code{__gnat_malloc}
8404 For further details see @ref{Dynamic Allocation Control}.
8407 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
8408 Specify directory to be searched for source and ALI files.
8410 @item ^-I-^/NOCURRENT_DIRECTORY^
8411 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatbind})
8412 Do not look for sources in the current directory where @code{gnatbind} was
8413 invoked, and do not look for ALI files in the directory containing the
8414 ALI file named in the @code{gnatbind} command line.
8416 @item ^-l^/ORDER_OF_ELABORATION^
8417 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@command{gnatbind})
8418 Output chosen elaboration order.
8420 @item ^-L@var{xxx}^/BUILD_LIBRARY=@var{xxx}^
8421 @cindex @option{^-L^/BUILD_LIBRARY^} (@command{gnatbind})
8422 Bind the units for library building. In this case the adainit and
8423 adafinal procedures (@pxref{Binding with Non-Ada Main Programs})
8424 are renamed to ^@var{xxx}init^@var{XXX}INIT^ and
8425 ^@var{xxx}final^@var{XXX}FINAL^.
8426 Implies ^-n^/NOCOMPILE^.
8428 (@xref{GNAT and Libraries}, for more details.)
8431 On OpenVMS, these init and final procedures are exported in uppercase
8432 letters. For example if /BUILD_LIBRARY=toto is used, the exported name of
8433 the init procedure will be "TOTOINIT" and the exported name of the final
8434 procedure will be "TOTOFINAL".
8437 @item ^-Mxyz^/RENAME_MAIN=xyz^
8438 @cindex @option{^-M^/RENAME_MAIN^} (@command{gnatbind})
8439 Rename generated main program from main to xyz. This option is
8440 supported on cross environments only.
8442 @item ^-m^/ERROR_LIMIT=^@var{n}
8443 @cindex @option{^-m^/ERROR_LIMIT^} (@command{gnatbind})
8444 Limit number of detected errors or warnings to @var{n}, where @var{n} is
8445 in the range 1..999999. The default value if no switch is
8446 given is 9999. If the number of warnings reaches this limit, then a
8447 message is output and further warnings are suppressed, the bind
8448 continues in this case. If the number of errors reaches this
8449 limit, then a message is output and the bind is abandoned.
8450 A value of zero means that no limit is enforced. The equal
8454 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
8458 @cindex @option{-nostdinc} (@command{gnatbind})
8459 Do not look for sources in the system default directory.
8462 @cindex @option{-nostdlib} (@command{gnatbind})
8463 Do not look for library files in the system default directory.
8465 @item --RTS=@var{rts-path}
8466 @cindex @option{--RTS} (@code{gnatbind})
8467 Specifies the default location of the runtime library. Same meaning as the
8468 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
8470 @item ^-o ^/OUTPUT=^@var{file}
8471 @cindex @option{^-o ^/OUTPUT^} (@command{gnatbind})
8472 Name the output file @var{file} (default is @file{b~@var{xxx}.adb}).
8473 Note that if this option is used, then linking must be done manually,
8474 gnatlink cannot be used.
8476 @item ^-O^/OBJECT_LIST^@r{[=}@var{filename}@r{]}
8477 @cindex @option{^-O^/OBJECT_LIST^} (@command{gnatbind})
8478 Output object list (to standard output or to the named file).
8480 @item ^-p^/PESSIMISTIC_ELABORATION^
8481 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@command{gnatbind})
8482 Pessimistic (worst-case) elaboration order
8485 @cindex @option{^-P^/CODEPEER^} (@command{gnatbind})
8486 Generate binder file suitable for CodePeer.
8489 @cindex @option{^-R^-R^} (@command{gnatbind})
8490 Output closure source list, which includes all non-time-units that are
8491 included in the bind.
8494 @cindex @option{^-Ra^-Ra^} (@command{gnatbind})
8495 Like @option{-R} but the list includes run-time units.
8497 @item ^-s^/READ_SOURCES=ALL^
8498 @cindex @option{^-s^/READ_SOURCES=ALL^} (@command{gnatbind})
8499 Require all source files to be present.
8501 @item ^-S@var{xxx}^/INITIALIZE_SCALARS=@var{xxx}^
8502 @cindex @option{^-S^/INITIALIZE_SCALARS^} (@command{gnatbind})
8503 Specifies the value to be used when detecting uninitialized scalar
8504 objects with pragma Initialize_Scalars.
8505 The @var{xxx} ^string specified with the switch^option^ may be either
8507 @item ``@option{^in^INVALID^}'' requesting an invalid value where possible
8508 @item ``@option{^lo^LOW^}'' for the lowest possible value
8509 @item ``@option{^hi^HIGH^}'' for the highest possible value
8510 @item ``@option{@var{xx}}'' for a value consisting of repeated bytes with the
8511 value @code{16#@var{xx}#} (i.e., @var{xx} is a string of two hexadecimal digits).
8514 In addition, you can specify @option{-Sev} to indicate that the value is
8515 to be set at run time. In this case, the program will look for an environment
8516 @cindex GNAT_INIT_SCALARS
8517 variable of the form @env{GNAT_INIT_SCALARS=@var{xx}}, where @var{xx} is one
8518 of @option{in/lo/hi/@var{xx}} with the same meanings as above.
8519 If no environment variable is found, or if it does not have a valid value,
8520 then the default is @option{in} (invalid values).
8524 @cindex @option{-static} (@code{gnatbind})
8525 Link against a static GNAT run time.
8528 @cindex @option{-shared} (@code{gnatbind})
8529 Link against a shared GNAT run time when available.
8532 @item ^-t^/NOTIME_STAMP_CHECK^
8533 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
8534 Tolerate time stamp and other consistency errors
8536 @item ^-T@var{n}^/TIME_SLICE=@var{n}^
8537 @cindex @option{^-T^/TIME_SLICE^} (@code{gnatbind})
8538 Set the time slice value to @var{n} milliseconds. If the system supports
8539 the specification of a specific time slice value, then the indicated value
8540 is used. If the system does not support specific time slice values, but
8541 does support some general notion of round-robin scheduling, then any
8542 nonzero value will activate round-robin scheduling.
8544 A value of zero is treated specially. It turns off time
8545 slicing, and in addition, indicates to the tasking run time that the
8546 semantics should match as closely as possible the Annex D
8547 requirements of the Ada RM, and in particular sets the default
8548 scheduling policy to @code{FIFO_Within_Priorities}.
8550 @item ^-u@var{n}^/DYNAMIC_STACK_USAGE=@var{n}^
8551 @cindex @option{^-u^/DYNAMIC_STACK_USAGE^} (@code{gnatbind})
8552 Enable dynamic stack usage, with @var{n} results stored and displayed
8553 at program termination. A result is generated when a task
8554 terminates. Results that can't be stored are displayed on the fly, at
8555 task termination. This option is currently not supported on Itanium
8556 platforms. (See @ref{Dynamic Stack Usage Analysis} for details.)
8558 @item ^-v^/REPORT_ERRORS=VERBOSE^
8559 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
8560 Verbose mode. Write error messages, header, summary output to
8565 @cindex @option{-w} (@code{gnatbind})
8566 Warning mode (@var{x}=s/e for suppress/treat as error)
8570 @item /WARNINGS=NORMAL
8571 @cindex @option{/WARNINGS} (@code{gnatbind})
8572 Normal warnings mode. Warnings are issued but ignored
8574 @item /WARNINGS=SUPPRESS
8575 @cindex @option{/WARNINGS} (@code{gnatbind})
8576 All warning messages are suppressed
8578 @item /WARNINGS=ERROR
8579 @cindex @option{/WARNINGS} (@code{gnatbind})
8580 Warning messages are treated as fatal errors
8583 @item ^-Wx^/WIDE_CHARACTER_ENCODING=^@var{e}
8584 @cindex @option{^-Wx^/WIDE_CHARACTER_ENCODING^} (@code{gnatbind})
8585 Override default wide character encoding for standard Text_IO files.
8587 @item ^-x^/READ_SOURCES=NONE^
8588 @cindex @option{^-x^/READ_SOURCES^} (@code{gnatbind})
8589 Exclude source files (check object consistency only).
8592 @item /READ_SOURCES=AVAILABLE
8593 @cindex @option{/READ_SOURCES} (@code{gnatbind})
8594 Default mode, in which sources are checked for consistency only if
8598 @item ^-X@var{nnn}^/RETURN_CODES=POSIX^
8599 @cindex @option{^-X@var{nnn}^/RETURN_CODES=POSIX^} (@code{gnatbind})
8600 Set default exit status value, normally 0 for POSIX compliance.
8603 @item /RETURN_CODES=VMS
8604 @cindex @option{/RETURN_CODES=VMS} (@code{gnatbind})
8605 VMS default normal successful return value is 1.
8608 @item ^-y^/ENABLE_LEAP_SECONDS^
8609 @cindex @option{^-y^/ENABLE_LEAP_SECONDS^} (@code{gnatbind})
8610 Enable leap seconds support in @code{Ada.Calendar} and its children.
8612 @item ^-z^/ZERO_MAIN^
8613 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
8619 You may obtain this listing of switches by running @code{gnatbind} with
8623 @node Consistency-Checking Modes
8624 @subsection Consistency-Checking Modes
8627 As described earlier, by default @code{gnatbind} checks
8628 that object files are consistent with one another and are consistent
8629 with any source files it can locate. The following switches control binder
8634 @item ^-s^/READ_SOURCES=ALL^
8635 @cindex @option{^-s^/READ_SOURCES=ALL^} (@code{gnatbind})
8636 Require source files to be present. In this mode, the binder must be
8637 able to locate all source files that are referenced, in order to check
8638 their consistency. In normal mode, if a source file cannot be located it
8639 is simply ignored. If you specify this switch, a missing source
8642 @item ^-Wx^/WIDE_CHARACTER_ENCODING=^@var{e}
8643 @cindex @option{^-Wx^/WIDE_CHARACTER_ENCODING^} (@code{gnatbind})
8644 Override default wide character encoding for standard Text_IO files.
8645 Normally the default wide character encoding method used for standard
8646 [Wide_[Wide_]]Text_IO files is taken from the encoding specified for
8647 the main source input (see description of switch
8648 @option{^-gnatWx^/WIDE_CHARACTER_ENCODING^} for the compiler). The
8649 use of this switch for the binder (which has the same set of
8650 possible arguments) overrides this default as specified.
8652 @item ^-x^/READ_SOURCES=NONE^
8653 @cindex @option{^-x^/READ_SOURCES=NONE^} (@code{gnatbind})
8654 Exclude source files. In this mode, the binder only checks that ALI
8655 files are consistent with one another. Source files are not accessed.
8656 The binder runs faster in this mode, and there is still a guarantee that
8657 the resulting program is self-consistent.
8658 If a source file has been edited since it was last compiled, and you
8659 specify this switch, the binder will not detect that the object
8660 file is out of date with respect to the source file. Note that this is the
8661 mode that is automatically used by @command{gnatmake} because in this
8662 case the checking against sources has already been performed by
8663 @command{gnatmake} in the course of compilation (i.e.@: before binding).
8666 @item /READ_SOURCES=AVAILABLE
8667 @cindex @code{/READ_SOURCES=AVAILABLE} (@code{gnatbind})
8668 This is the default mode in which source files are checked if they are
8669 available, and ignored if they are not available.
8673 @node Binder Error Message Control
8674 @subsection Binder Error Message Control
8677 The following switches provide control over the generation of error
8678 messages from the binder:
8682 @item ^-v^/REPORT_ERRORS=VERBOSE^
8683 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
8684 Verbose mode. In the normal mode, brief error messages are generated to
8685 @file{stderr}. If this switch is present, a header is written
8686 to @file{stdout} and any error messages are directed to @file{stdout}.
8687 All that is written to @file{stderr} is a brief summary message.
8689 @item ^-b^/REPORT_ERRORS=BRIEF^
8690 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@code{gnatbind})
8691 Generate brief error messages to @file{stderr} even if verbose mode is
8692 specified. This is relevant only when used with the
8693 @option{^-v^/REPORT_ERRORS=VERBOSE^} switch.
8697 @cindex @option{-m} (@code{gnatbind})
8698 Limits the number of error messages to @var{n}, a decimal integer in the
8699 range 1-999. The binder terminates immediately if this limit is reached.
8702 @cindex @option{-M} (@code{gnatbind})
8703 Renames the generated main program from @code{main} to @code{xxx}.
8704 This is useful in the case of some cross-building environments, where
8705 the actual main program is separate from the one generated
8709 @item ^-ws^/WARNINGS=SUPPRESS^
8710 @cindex @option{^-ws^/WARNINGS=SUPPRESS^} (@code{gnatbind})
8712 Suppress all warning messages.
8714 @item ^-we^/WARNINGS=ERROR^
8715 @cindex @option{^-we^/WARNINGS=ERROR^} (@code{gnatbind})
8716 Treat any warning messages as fatal errors.
8719 @item /WARNINGS=NORMAL
8720 Standard mode with warnings generated, but warnings do not get treated
8724 @item ^-t^/NOTIME_STAMP_CHECK^
8725 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
8726 @cindex Time stamp checks, in binder
8727 @cindex Binder consistency checks
8728 @cindex Consistency checks, in binder
8729 The binder performs a number of consistency checks including:
8733 Check that time stamps of a given source unit are consistent
8735 Check that checksums of a given source unit are consistent
8737 Check that consistent versions of @code{GNAT} were used for compilation
8739 Check consistency of configuration pragmas as required
8743 Normally failure of such checks, in accordance with the consistency
8744 requirements of the Ada Reference Manual, causes error messages to be
8745 generated which abort the binder and prevent the output of a binder
8746 file and subsequent link to obtain an executable.
8748 The @option{^-t^/NOTIME_STAMP_CHECK^} switch converts these error messages
8749 into warnings, so that
8750 binding and linking can continue to completion even in the presence of such
8751 errors. The result may be a failed link (due to missing symbols), or a
8752 non-functional executable which has undefined semantics.
8753 @emph{This means that
8754 @option{^-t^/NOTIME_STAMP_CHECK^} should be used only in unusual situations,
8758 @node Elaboration Control
8759 @subsection Elaboration Control
8762 The following switches provide additional control over the elaboration
8763 order. For full details see @ref{Elaboration Order Handling in GNAT}.
8766 @item ^-p^/PESSIMISTIC_ELABORATION^
8767 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@code{gnatbind})
8768 Normally the binder attempts to choose an elaboration order that is
8769 likely to minimize the likelihood of an elaboration order error resulting
8770 in raising a @code{Program_Error} exception. This switch reverses the
8771 action of the binder, and requests that it deliberately choose an order
8772 that is likely to maximize the likelihood of an elaboration error.
8773 This is useful in ensuring portability and avoiding dependence on
8774 accidental fortuitous elaboration ordering.
8776 Normally it only makes sense to use the @option{^-p^/PESSIMISTIC_ELABORATION^}
8778 elaboration checking is used (@option{-gnatE} switch used for compilation).
8779 This is because in the default static elaboration mode, all necessary
8780 @code{Elaborate} and @code{Elaborate_All} pragmas are implicitly inserted.
8781 These implicit pragmas are still respected by the binder in
8782 @option{^-p^/PESSIMISTIC_ELABORATION^} mode, so a
8783 safe elaboration order is assured.
8785 Note that @option{^-p^/PESSIMISTIC_ELABORATION^} is not intended for
8786 production use; it is more for debugging/experimental use.
8789 @node Output Control
8790 @subsection Output Control
8793 The following switches allow additional control over the output
8794 generated by the binder.
8799 @item ^-c^/NOOUTPUT^
8800 @cindex @option{^-c^/NOOUTPUT^} (@code{gnatbind})
8801 Check only. Do not generate the binder output file. In this mode the
8802 binder performs all error checks but does not generate an output file.
8804 @item ^-e^/ELABORATION_DEPENDENCIES^
8805 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@code{gnatbind})
8806 Output complete list of elaboration-order dependencies, showing the
8807 reason for each dependency. This output can be rather extensive but may
8808 be useful in diagnosing problems with elaboration order. The output is
8809 written to @file{stdout}.
8812 @cindex @option{^-h^/HELP^} (@code{gnatbind})
8813 Output usage information. The output is written to @file{stdout}.
8815 @item ^-K^/LINKER_OPTION_LIST^
8816 @cindex @option{^-K^/LINKER_OPTION_LIST^} (@code{gnatbind})
8817 Output linker options to @file{stdout}. Includes library search paths,
8818 contents of pragmas Ident and Linker_Options, and libraries added
8821 @item ^-l^/ORDER_OF_ELABORATION^
8822 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@code{gnatbind})
8823 Output chosen elaboration order. The output is written to @file{stdout}.
8825 @item ^-O^/OBJECT_LIST^
8826 @cindex @option{^-O^/OBJECT_LIST^} (@code{gnatbind})
8827 Output full names of all the object files that must be linked to provide
8828 the Ada component of the program. The output is written to @file{stdout}.
8829 This list includes the files explicitly supplied and referenced by the user
8830 as well as implicitly referenced run-time unit files. The latter are
8831 omitted if the corresponding units reside in shared libraries. The
8832 directory names for the run-time units depend on the system configuration.
8834 @item ^-o ^/OUTPUT=^@var{file}
8835 @cindex @option{^-o^/OUTPUT^} (@code{gnatbind})
8836 Set name of output file to @var{file} instead of the normal
8837 @file{b~@var{mainprog}.adb} default. Note that @var{file} denote the Ada
8838 binder generated body filename.
8839 Note that if this option is used, then linking must be done manually.
8840 It is not possible to use gnatlink in this case, since it cannot locate
8843 @item ^-r^/RESTRICTION_LIST^
8844 @cindex @option{^-r^/RESTRICTION_LIST^} (@code{gnatbind})
8845 Generate list of @code{pragma Restrictions} that could be applied to
8846 the current unit. This is useful for code audit purposes, and also may
8847 be used to improve code generation in some cases.
8851 @node Dynamic Allocation Control
8852 @subsection Dynamic Allocation Control
8855 The heap control switches -- @option{-H32} and @option{-H64} --
8856 determine whether dynamic allocation uses 32-bit or 64-bit memory.
8857 They only affect compiler-generated allocations via @code{__gnat_malloc};
8858 explicit calls to @code{malloc} and related functions from the C
8859 run-time library are unaffected.
8863 Allocate memory on 32-bit heap
8866 Allocate memory on 64-bit heap. This is the default
8867 unless explicitly overridden by a @code{'Size} clause on the access type.
8872 See also @ref{Access types and 32/64-bit allocation}.
8876 These switches are only effective on VMS platforms.
8880 @node Binding with Non-Ada Main Programs
8881 @subsection Binding with Non-Ada Main Programs
8884 In our description so far we have assumed that the main
8885 program is in Ada, and that the task of the binder is to generate a
8886 corresponding function @code{main} that invokes this Ada main
8887 program. GNAT also supports the building of executable programs where
8888 the main program is not in Ada, but some of the called routines are
8889 written in Ada and compiled using GNAT (@pxref{Mixed Language Programming}).
8890 The following switch is used in this situation:
8894 @cindex @option{^-n^/NOMAIN^} (@code{gnatbind})
8895 No main program. The main program is not in Ada.
8899 In this case, most of the functions of the binder are still required,
8900 but instead of generating a main program, the binder generates a file
8901 containing the following callable routines:
8906 You must call this routine to initialize the Ada part of the program by
8907 calling the necessary elaboration routines. A call to @code{adainit} is
8908 required before the first call to an Ada subprogram.
8910 Note that it is assumed that the basic execution environment must be setup
8911 to be appropriate for Ada execution at the point where the first Ada
8912 subprogram is called. In particular, if the Ada code will do any
8913 floating-point operations, then the FPU must be setup in an appropriate
8914 manner. For the case of the x86, for example, full precision mode is
8915 required. The procedure GNAT.Float_Control.Reset may be used to ensure
8916 that the FPU is in the right state.
8920 You must call this routine to perform any library-level finalization
8921 required by the Ada subprograms. A call to @code{adafinal} is required
8922 after the last call to an Ada subprogram, and before the program
8927 If the @option{^-n^/NOMAIN^} switch
8928 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
8929 @cindex Binder, multiple input files
8930 is given, more than one ALI file may appear on
8931 the command line for @code{gnatbind}. The normal @dfn{closure}
8932 calculation is performed for each of the specified units. Calculating
8933 the closure means finding out the set of units involved by tracing
8934 @code{with} references. The reason it is necessary to be able to
8935 specify more than one ALI file is that a given program may invoke two or
8936 more quite separate groups of Ada units.
8938 The binder takes the name of its output file from the last specified ALI
8939 file, unless overridden by the use of the @option{^-o file^/OUTPUT=file^}.
8940 @cindex @option{^-o^/OUTPUT^} (@command{gnatbind})
8941 The output is an Ada unit in source form that can be compiled with GNAT.
8942 This compilation occurs automatically as part of the @command{gnatlink}
8945 Currently the GNAT run time requires a FPU using 80 bits mode
8946 precision. Under targets where this is not the default it is required to
8947 call GNAT.Float_Control.Reset before using floating point numbers (this
8948 include float computation, float input and output) in the Ada code. A
8949 side effect is that this could be the wrong mode for the foreign code
8950 where floating point computation could be broken after this call.
8952 @node Binding Programs with No Main Subprogram
8953 @subsection Binding Programs with No Main Subprogram
8956 It is possible to have an Ada program which does not have a main
8957 subprogram. This program will call the elaboration routines of all the
8958 packages, then the finalization routines.
8960 The following switch is used to bind programs organized in this manner:
8963 @item ^-z^/ZERO_MAIN^
8964 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
8965 Normally the binder checks that the unit name given on the command line
8966 corresponds to a suitable main subprogram. When this switch is used,
8967 a list of ALI files can be given, and the execution of the program
8968 consists of elaboration of these units in an appropriate order. Note
8969 that the default wide character encoding method for standard Text_IO
8970 files is always set to Brackets if this switch is set (you can use
8972 @option{^-Wx^WIDE_CHARACTER_ENCODING^} to override this default).
8975 @node Command-Line Access
8976 @section Command-Line Access
8979 The package @code{Ada.Command_Line} provides access to the command-line
8980 arguments and program name. In order for this interface to operate
8981 correctly, the two variables
8993 are declared in one of the GNAT library routines. These variables must
8994 be set from the actual @code{argc} and @code{argv} values passed to the
8995 main program. With no @option{^n^/NOMAIN^} present, @code{gnatbind}
8996 generates the C main program to automatically set these variables.
8997 If the @option{^n^/NOMAIN^} switch is used, there is no automatic way to
8998 set these variables. If they are not set, the procedures in
8999 @code{Ada.Command_Line} will not be available, and any attempt to use
9000 them will raise @code{Constraint_Error}. If command line access is
9001 required, your main program must set @code{gnat_argc} and
9002 @code{gnat_argv} from the @code{argc} and @code{argv} values passed to
9005 @node Search Paths for gnatbind
9006 @section Search Paths for @code{gnatbind}
9009 The binder takes the name of an ALI file as its argument and needs to
9010 locate source files as well as other ALI files to verify object consistency.
9012 For source files, it follows exactly the same search rules as @command{gcc}
9013 (@pxref{Search Paths and the Run-Time Library (RTL)}). For ALI files the
9014 directories searched are:
9018 The directory containing the ALI file named in the command line, unless
9019 the switch @option{^-I-^/NOCURRENT_DIRECTORY^} is specified.
9022 All directories specified by @option{^-I^/SEARCH^}
9023 switches on the @code{gnatbind}
9024 command line, in the order given.
9027 @findex ADA_PRJ_OBJECTS_FILE
9028 Each of the directories listed in the text file whose name is given
9029 by the @env{ADA_PRJ_OBJECTS_FILE} ^environment variable^logical name^.
9032 @env{ADA_PRJ_OBJECTS_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
9033 driver when project files are used. It should not normally be set
9037 @findex ADA_OBJECTS_PATH
9038 Each of the directories listed in the value of the
9039 @env{ADA_OBJECTS_PATH} ^environment variable^logical name^.
9041 Construct this value
9042 exactly as the @env{PATH} environment variable: a list of directory
9043 names separated by colons (semicolons when working with the NT version
9047 Normally, define this value as a logical name containing a comma separated
9048 list of directory names.
9050 This variable can also be defined by means of an environment string
9051 (an argument to the HP C exec* set of functions).
9055 DEFINE ANOTHER_PATH FOO:[BAG]
9056 DEFINE ADA_OBJECTS_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
9059 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
9060 first, followed by the standard Ada
9061 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADALIB].
9062 If this is not redefined, the user will obtain the HP Ada 83 IO packages
9063 (Text_IO, Sequential_IO, etc)
9064 instead of the standard Ada packages. Thus, in order to get the standard Ada
9065 packages by default, ADA_OBJECTS_PATH must be redefined.
9069 The content of the @file{ada_object_path} file which is part of the GNAT
9070 installation tree and is used to store standard libraries such as the
9071 GNAT Run Time Library (RTL) unless the switch @option{-nostdlib} is
9074 @ref{Installing a library}
9079 In the binder the switch @option{^-I^/SEARCH^}
9080 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
9081 is used to specify both source and
9082 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
9083 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
9084 instead if you want to specify
9085 source paths only, and @option{^-aO^/LIBRARY_SEARCH^}
9086 @cindex @option{^-aO^/LIBRARY_SEARCH^} (@command{gnatbind})
9087 if you want to specify library paths
9088 only. This means that for the binder
9089 @option{^-I^/SEARCH=^}@var{dir} is equivalent to
9090 @option{^-aI^/SOURCE_SEARCH=^}@var{dir}
9091 @option{^-aO^/OBJECT_SEARCH=^}@var{dir}.
9092 The binder generates the bind file (a C language source file) in the
9093 current working directory.
9099 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
9100 children make up the GNAT Run-Time Library, together with the package
9101 GNAT and its children, which contain a set of useful additional
9102 library functions provided by GNAT. The sources for these units are
9103 needed by the compiler and are kept together in one directory. The ALI
9104 files and object files generated by compiling the RTL are needed by the
9105 binder and the linker and are kept together in one directory, typically
9106 different from the directory containing the sources. In a normal
9107 installation, you need not specify these directory names when compiling
9108 or binding. Either the environment variables or the built-in defaults
9109 cause these files to be found.
9111 Besides simplifying access to the RTL, a major use of search paths is
9112 in compiling sources from multiple directories. This can make
9113 development environments much more flexible.
9115 @node Examples of gnatbind Usage
9116 @section Examples of @code{gnatbind} Usage
9119 This section contains a number of examples of using the GNAT binding
9120 utility @code{gnatbind}.
9123 @item gnatbind hello
9124 The main program @code{Hello} (source program in @file{hello.adb}) is
9125 bound using the standard switch settings. The generated main program is
9126 @file{b~hello.adb}. This is the normal, default use of the binder.
9129 @item gnatbind hello -o mainprog.adb
9132 @item gnatbind HELLO.ALI /OUTPUT=Mainprog.ADB
9134 The main program @code{Hello} (source program in @file{hello.adb}) is
9135 bound using the standard switch settings. The generated main program is
9136 @file{mainprog.adb} with the associated spec in
9137 @file{mainprog.ads}. Note that you must specify the body here not the
9138 spec. Note that if this option is used, then linking must be done manually,
9139 since gnatlink will not be able to find the generated file.
9142 @c ------------------------------------
9143 @node Linking with gnatlink
9144 @chapter Linking with @command{gnatlink}
9145 @c ------------------------------------
9149 This chapter discusses @command{gnatlink}, a tool that links
9150 an Ada program and builds an executable file. This utility
9151 invokes the system linker ^(via the @command{gcc} command)^^
9152 with a correct list of object files and library references.
9153 @command{gnatlink} automatically determines the list of files and
9154 references for the Ada part of a program. It uses the binder file
9155 generated by the @command{gnatbind} to determine this list.
9157 Note: to invoke @code{gnatlink} with a project file, use the @code{gnat}
9158 driver (see @ref{The GNAT Driver and Project Files}).
9161 * Running gnatlink::
9162 * Switches for gnatlink::
9165 @node Running gnatlink
9166 @section Running @command{gnatlink}
9169 The form of the @command{gnatlink} command is
9172 @c $ gnatlink @ovar{switches} @var{mainprog}@r{[}.ali@r{]}
9173 @c @ovar{non-Ada objects} @ovar{linker options}
9174 @c Expanding @ovar macro inline (explanation in macro def comments)
9175 $ gnatlink @r{[}@var{switches}@r{]} @var{mainprog}@r{[}.ali@r{]}
9176 @r{[}@var{non-Ada objects}@r{]} @r{[}@var{linker options}@r{]}
9181 The arguments of @command{gnatlink} (switches, main @file{ALI} file,
9183 or linker options) may be in any order, provided that no non-Ada object may
9184 be mistaken for a main @file{ALI} file.
9185 Any file name @file{F} without the @file{.ali}
9186 extension will be taken as the main @file{ALI} file if a file exists
9187 whose name is the concatenation of @file{F} and @file{.ali}.
9190 @file{@var{mainprog}.ali} references the ALI file of the main program.
9191 The @file{.ali} extension of this file can be omitted. From this
9192 reference, @command{gnatlink} locates the corresponding binder file
9193 @file{b~@var{mainprog}.adb} and, using the information in this file along
9194 with the list of non-Ada objects and linker options, constructs a
9195 linker command file to create the executable.
9197 The arguments other than the @command{gnatlink} switches and the main
9198 @file{ALI} file are passed to the linker uninterpreted.
9199 They typically include the names of
9200 object files for units written in other languages than Ada and any library
9201 references required to resolve references in any of these foreign language
9202 units, or in @code{Import} pragmas in any Ada units.
9204 @var{linker options} is an optional list of linker specific
9206 The default linker called by gnatlink is @command{gcc} which in
9207 turn calls the appropriate system linker.
9209 One useful option for the linker is @option{-s}: it reduces the size of the
9210 executable by removing all symbol table and relocation information from the
9213 Standard options for the linker such as @option{-lmy_lib} or
9214 @option{-Ldir} can be added as is.
9215 For options that are not recognized by
9216 @command{gcc} as linker options, use the @command{gcc} switches
9217 @option{-Xlinker} or @option{-Wl,}.
9219 Refer to the GCC documentation for
9222 Here is an example showing how to generate a linker map:
9225 $ ^gnatlink my_prog -Wl,-Map,MAPFILE^GNAT LINK my_prog.ali /MAP^
9228 Using @var{linker options} it is possible to set the program stack and
9231 See @ref{Setting Stack Size from gnatlink} and
9232 @ref{Setting Heap Size from gnatlink}.
9235 @command{gnatlink} determines the list of objects required by the Ada
9236 program and prepends them to the list of objects passed to the linker.
9237 @command{gnatlink} also gathers any arguments set by the use of
9238 @code{pragma Linker_Options} and adds them to the list of arguments
9239 presented to the linker.
9242 @command{gnatlink} accepts the following types of extra files on the command
9243 line: objects (@file{.OBJ}), libraries (@file{.OLB}), sharable images
9244 (@file{.EXE}), and options files (@file{.OPT}). These are recognized and
9245 handled according to their extension.
9248 @node Switches for gnatlink
9249 @section Switches for @command{gnatlink}
9252 The following switches are available with the @command{gnatlink} utility:
9258 @cindex @option{--version} @command{gnatlink}
9259 Display Copyright and version, then exit disregarding all other options.
9262 @cindex @option{--help} @command{gnatlink}
9263 If @option{--version} was not used, display usage, then exit disregarding
9266 @item ^-f^/FORCE_OBJECT_FILE_LIST^
9267 @cindex Command line length
9268 @cindex @option{^-f^/FORCE_OBJECT_FILE_LIST^} (@command{gnatlink})
9269 On some targets, the command line length is limited, and @command{gnatlink}
9270 will generate a separate file for the linker if the list of object files
9272 The @option{^-f^/FORCE_OBJECT_FILE_LIST^} switch forces this file
9273 to be generated even if
9274 the limit is not exceeded. This is useful in some cases to deal with
9275 special situations where the command line length is exceeded.
9278 @cindex Debugging information, including
9279 @cindex @option{^-g^/DEBUG^} (@command{gnatlink})
9280 The option to include debugging information causes the Ada bind file (in
9281 other words, @file{b~@var{mainprog}.adb}) to be compiled with
9282 @option{^-g^/DEBUG^}.
9283 In addition, the binder does not delete the @file{b~@var{mainprog}.adb},
9284 @file{b~@var{mainprog}.o} and @file{b~@var{mainprog}.ali} files.
9285 Without @option{^-g^/DEBUG^}, the binder removes these files by
9286 default. The same procedure apply if a C bind file was generated using
9287 @option{^-C^/BIND_FILE=C^} @code{gnatbind} option, in this case the filenames
9288 are @file{b_@var{mainprog}.c} and @file{b_@var{mainprog}.o}.
9290 @item ^-n^/NOCOMPILE^
9291 @cindex @option{^-n^/NOCOMPILE^} (@command{gnatlink})
9292 Do not compile the file generated by the binder. This may be used when
9293 a link is rerun with different options, but there is no need to recompile
9297 @cindex @option{^-v^/VERBOSE^} (@command{gnatlink})
9298 Causes additional information to be output, including a full list of the
9299 included object files. This switch option is most useful when you want
9300 to see what set of object files are being used in the link step.
9302 @item ^-v -v^/VERBOSE/VERBOSE^
9303 @cindex @option{^-v -v^/VERBOSE/VERBOSE^} (@command{gnatlink})
9304 Very verbose mode. Requests that the compiler operate in verbose mode when
9305 it compiles the binder file, and that the system linker run in verbose mode.
9307 @item ^-o ^/EXECUTABLE=^@var{exec-name}
9308 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatlink})
9309 @var{exec-name} specifies an alternate name for the generated
9310 executable program. If this switch is omitted, the executable has the same
9311 name as the main unit. For example, @code{gnatlink try.ali} creates
9312 an executable called @file{^try^TRY.EXE^}.
9315 @item -b @var{target}
9316 @cindex @option{-b} (@command{gnatlink})
9317 Compile your program to run on @var{target}, which is the name of a
9318 system configuration. You must have a GNAT cross-compiler built if
9319 @var{target} is not the same as your host system.
9322 @cindex @option{-B} (@command{gnatlink})
9323 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
9324 from @var{dir} instead of the default location. Only use this switch
9325 when multiple versions of the GNAT compiler are available.
9326 @xref{Directory Options,,, gcc, The GNU Compiler Collection},
9327 for further details. You would normally use the @option{-b} or
9328 @option{-V} switch instead.
9331 When linking an executable, create a map file. The name of the map file
9332 has the same name as the executable with extension ".map".
9335 When linking an executable, create a map file. The name of the map file is
9338 @item --GCC=@var{compiler_name}
9339 @cindex @option{--GCC=compiler_name} (@command{gnatlink})
9340 Program used for compiling the binder file. The default is
9341 @command{gcc}. You need to use quotes around @var{compiler_name} if
9342 @code{compiler_name} contains spaces or other separator characters.
9343 As an example @option{--GCC="foo -x -y"} will instruct @command{gnatlink} to
9344 use @code{foo -x -y} as your compiler. Note that switch @option{-c} is always
9345 inserted after your command name. Thus in the above example the compiler
9346 command that will be used by @command{gnatlink} will be @code{foo -c -x -y}.
9347 A limitation of this syntax is that the name and path name of the executable
9348 itself must not include any embedded spaces. If the compiler executable is
9349 different from the default one (gcc or <prefix>-gcc), then the back-end
9350 switches in the ALI file are not used to compile the binder generated source.
9351 For example, this is the case with @option{--GCC="foo -x -y"}. But the back end
9352 switches will be used for @option{--GCC="gcc -gnatv"}. If several
9353 @option{--GCC=compiler_name} are used, only the last @var{compiler_name}
9354 is taken into account. However, all the additional switches are also taken
9356 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
9357 @option{--GCC="bar -x -y -z -t"}.
9359 @item --LINK=@var{name}
9360 @cindex @option{--LINK=} (@command{gnatlink})
9361 @var{name} is the name of the linker to be invoked. This is especially
9362 useful in mixed language programs since languages such as C++ require
9363 their own linker to be used. When this switch is omitted, the default
9364 name for the linker is @command{gcc}. When this switch is used, the
9365 specified linker is called instead of @command{gcc} with exactly the same
9366 parameters that would have been passed to @command{gcc} so if the desired
9367 linker requires different parameters it is necessary to use a wrapper
9368 script that massages the parameters before invoking the real linker. It
9369 may be useful to control the exact invocation by using the verbose
9375 @item /DEBUG=TRACEBACK
9376 @cindex @code{/DEBUG=TRACEBACK} (@command{gnatlink})
9377 This qualifier causes sufficient information to be included in the
9378 executable file to allow a traceback, but does not include the full
9379 symbol information needed by the debugger.
9381 @item /IDENTIFICATION="<string>"
9382 @code{"<string>"} specifies the string to be stored in the image file
9383 identification field in the image header.
9384 It overrides any pragma @code{Ident} specified string.
9386 @item /NOINHIBIT-EXEC
9387 Generate the executable file even if there are linker warnings.
9389 @item /NOSTART_FILES
9390 Don't link in the object file containing the ``main'' transfer address.
9391 Used when linking with a foreign language main program compiled with an
9395 Prefer linking with object libraries over sharable images, even without
9401 @node The GNAT Make Program gnatmake
9402 @chapter The GNAT Make Program @command{gnatmake}
9406 * Running gnatmake::
9407 * Switches for gnatmake::
9408 * Mode Switches for gnatmake::
9409 * Notes on the Command Line::
9410 * How gnatmake Works::
9411 * Examples of gnatmake Usage::
9414 A typical development cycle when working on an Ada program consists of
9415 the following steps:
9419 Edit some sources to fix bugs.
9425 Compile all sources affected.
9435 The third step can be tricky, because not only do the modified files
9436 @cindex Dependency rules
9437 have to be compiled, but any files depending on these files must also be
9438 recompiled. The dependency rules in Ada can be quite complex, especially
9439 in the presence of overloading, @code{use} clauses, generics and inlined
9442 @command{gnatmake} automatically takes care of the third and fourth steps
9443 of this process. It determines which sources need to be compiled,
9444 compiles them, and binds and links the resulting object files.
9446 Unlike some other Ada make programs, the dependencies are always
9447 accurately recomputed from the new sources. The source based approach of
9448 the GNAT compilation model makes this possible. This means that if
9449 changes to the source program cause corresponding changes in
9450 dependencies, they will always be tracked exactly correctly by
9453 @node Running gnatmake
9454 @section Running @command{gnatmake}
9457 The usual form of the @command{gnatmake} command is
9460 @c $ gnatmake @ovar{switches} @var{file_name}
9461 @c @ovar{file_names} @ovar{mode_switches}
9462 @c Expanding @ovar macro inline (explanation in macro def comments)
9463 $ gnatmake @r{[}@var{switches}@r{]} @var{file_name}
9464 @r{[}@var{file_names}@r{]} @r{[}@var{mode_switches}@r{]}
9468 The only required argument is one @var{file_name}, which specifies
9469 a compilation unit that is a main program. Several @var{file_names} can be
9470 specified: this will result in several executables being built.
9471 If @code{switches} are present, they can be placed before the first
9472 @var{file_name}, between @var{file_names} or after the last @var{file_name}.
9473 If @var{mode_switches} are present, they must always be placed after
9474 the last @var{file_name} and all @code{switches}.
9476 If you are using standard file extensions (@file{.adb} and @file{.ads}), then the
9477 extension may be omitted from the @var{file_name} arguments. However, if
9478 you are using non-standard extensions, then it is required that the
9479 extension be given. A relative or absolute directory path can be
9480 specified in a @var{file_name}, in which case, the input source file will
9481 be searched for in the specified directory only. Otherwise, the input
9482 source file will first be searched in the directory where
9483 @command{gnatmake} was invoked and if it is not found, it will be search on
9484 the source path of the compiler as described in
9485 @ref{Search Paths and the Run-Time Library (RTL)}.
9487 All @command{gnatmake} output (except when you specify
9488 @option{^-M^/DEPENDENCIES_LIST^}) is to
9489 @file{stderr}. The output produced by the
9490 @option{^-M^/DEPENDENCIES_LIST^} switch is send to
9493 @node Switches for gnatmake
9494 @section Switches for @command{gnatmake}
9497 You may specify any of the following switches to @command{gnatmake}:
9503 @cindex @option{--version} @command{gnatmake}
9504 Display Copyright and version, then exit disregarding all other options.
9507 @cindex @option{--help} @command{gnatmake}
9508 If @option{--version} was not used, display usage, then exit disregarding
9512 @item --GCC=@var{compiler_name}
9513 @cindex @option{--GCC=compiler_name} (@command{gnatmake})
9514 Program used for compiling. The default is `@command{gcc}'. You need to use
9515 quotes around @var{compiler_name} if @code{compiler_name} contains
9516 spaces or other separator characters. As an example @option{--GCC="foo -x
9517 -y"} will instruct @command{gnatmake} to use @code{foo -x -y} as your
9518 compiler. A limitation of this syntax is that the name and path name of
9519 the executable itself must not include any embedded spaces. Note that
9520 switch @option{-c} is always inserted after your command name. Thus in the
9521 above example the compiler command that will be used by @command{gnatmake}
9522 will be @code{foo -c -x -y}. If several @option{--GCC=compiler_name} are
9523 used, only the last @var{compiler_name} is taken into account. However,
9524 all the additional switches are also taken into account. Thus,
9525 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
9526 @option{--GCC="bar -x -y -z -t"}.
9528 @item --GNATBIND=@var{binder_name}
9529 @cindex @option{--GNATBIND=binder_name} (@command{gnatmake})
9530 Program used for binding. The default is `@code{gnatbind}'. You need to
9531 use quotes around @var{binder_name} if @var{binder_name} contains spaces
9532 or other separator characters. As an example @option{--GNATBIND="bar -x
9533 -y"} will instruct @command{gnatmake} to use @code{bar -x -y} as your
9534 binder. Binder switches that are normally appended by @command{gnatmake}
9535 to `@code{gnatbind}' are now appended to the end of @code{bar -x -y}.
9536 A limitation of this syntax is that the name and path name of the executable
9537 itself must not include any embedded spaces.
9539 @item --GNATLINK=@var{linker_name}
9540 @cindex @option{--GNATLINK=linker_name} (@command{gnatmake})
9541 Program used for linking. The default is `@command{gnatlink}'. You need to
9542 use quotes around @var{linker_name} if @var{linker_name} contains spaces
9543 or other separator characters. As an example @option{--GNATLINK="lan -x
9544 -y"} will instruct @command{gnatmake} to use @code{lan -x -y} as your
9545 linker. Linker switches that are normally appended by @command{gnatmake} to
9546 `@command{gnatlink}' are now appended to the end of @code{lan -x -y}.
9547 A limitation of this syntax is that the name and path name of the executable
9548 itself must not include any embedded spaces.
9552 @item ^--subdirs^/SUBDIRS^=subdir
9553 Actual object directory of each project file is the subdirectory subdir of the
9554 object directory specified or defaulted in the project file.
9556 @item ^--single-compile-per-obj-dir^/SINGLE_COMPILE_PER_OBJ_DIR^
9557 Disallow simultaneous compilations in the same object directory when
9558 project files are used.
9560 @item ^--unchecked-shared-lib-imports^/UNCHECKED_SHARED_LIB_IMPORTS^
9561 By default, shared library projects are not allowed to import static library
9562 projects. When this switch is used on the command line, this restriction is
9565 @item ^--source-info=<source info file>^/SRC_INFO=source-info-file^
9566 Specify a source info file. This switch is active only when project files
9567 are used. If the source info file is specified as a relative path, then it is
9568 relative to the object directory of the main project. If the source info file
9569 does not exist, then after the Project Manager has successfully parsed and
9570 processed the project files and found the sources, it creates the source info
9571 file. If the source info file already exists and can be read successfully,
9572 then the Project Manager will get all the needed information about the sources
9573 from the source info file and will not look for them. This reduces the time
9574 to process the project files, especially when looking for sources that take a
9575 long time. If the source info file exists but cannot be parsed successfully,
9576 the Project Manager will attempt to recreate it. If the Project Manager fails
9577 to create the source info file, a message is issued, but gnatmake does not
9578 fail. @command{gnatmake} "trusts" the source info file. This means that
9579 if the source files have changed (addition, deletion, moving to a different
9580 source directory), then the source info file need to be deleted and recreated.
9583 @item --create-map-file
9584 When linking an executable, create a map file. The name of the map file
9585 has the same name as the executable with extension ".map".
9587 @item --create-map-file=mapfile
9588 When linking an executable, create a map file. The name of the map file is
9593 @item ^-a^/ALL_FILES^
9594 @cindex @option{^-a^/ALL_FILES^} (@command{gnatmake})
9595 Consider all files in the make process, even the GNAT internal system
9596 files (for example, the predefined Ada library files), as well as any
9597 locked files. Locked files are files whose ALI file is write-protected.
9599 @command{gnatmake} does not check these files,
9600 because the assumption is that the GNAT internal files are properly up
9601 to date, and also that any write protected ALI files have been properly
9602 installed. Note that if there is an installation problem, such that one
9603 of these files is not up to date, it will be properly caught by the
9605 You may have to specify this switch if you are working on GNAT
9606 itself. The switch @option{^-a^/ALL_FILES^} is also useful
9607 in conjunction with @option{^-f^/FORCE_COMPILE^}
9608 if you need to recompile an entire application,
9609 including run-time files, using special configuration pragmas,
9610 such as a @code{Normalize_Scalars} pragma.
9613 @code{gnatmake ^-a^/ALL_FILES^} compiles all GNAT
9616 @code{gcc -c -gnatpg} rather than @code{gcc -c}.
9619 the @code{/CHECKS=SUPPRESS_ALL /STYLE_CHECKS=GNAT} switch.
9622 @item ^-b^/ACTIONS=BIND^
9623 @cindex @option{^-b^/ACTIONS=BIND^} (@command{gnatmake})
9624 Bind only. Can be combined with @option{^-c^/ACTIONS=COMPILE^} to do
9625 compilation and binding, but no link.
9626 Can be combined with @option{^-l^/ACTIONS=LINK^}
9627 to do binding and linking. When not combined with
9628 @option{^-c^/ACTIONS=COMPILE^}
9629 all the units in the closure of the main program must have been previously
9630 compiled and must be up to date. The root unit specified by @var{file_name}
9631 may be given without extension, with the source extension or, if no GNAT
9632 Project File is specified, with the ALI file extension.
9634 @item ^-c^/ACTIONS=COMPILE^
9635 @cindex @option{^-c^/ACTIONS=COMPILE^} (@command{gnatmake})
9636 Compile only. Do not perform binding, except when @option{^-b^/ACTIONS=BIND^}
9637 is also specified. Do not perform linking, except if both
9638 @option{^-b^/ACTIONS=BIND^} and
9639 @option{^-l^/ACTIONS=LINK^} are also specified.
9640 If the root unit specified by @var{file_name} is not a main unit, this is the
9641 default. Otherwise @command{gnatmake} will attempt binding and linking
9642 unless all objects are up to date and the executable is more recent than
9646 @cindex @option{^-C^/MAPPING^} (@command{gnatmake})
9647 Use a temporary mapping file. A mapping file is a way to communicate
9648 to the compiler two mappings: from unit names to file names (without
9649 any directory information) and from file names to path names (with
9650 full directory information). A mapping file can make the compiler's
9651 file searches faster, especially if there are many source directories,
9652 or the sources are read over a slow network connection. If
9653 @option{^-P^/PROJECT_FILE^} is used, a mapping file is always used, so
9654 @option{^-C^/MAPPING^} is unnecessary; in this case the mapping file
9655 is initially populated based on the project file. If
9656 @option{^-C^/MAPPING^} is used without
9657 @option{^-P^/PROJECT_FILE^},
9658 the mapping file is initially empty. Each invocation of the compiler
9659 will add any newly accessed sources to the mapping file.
9661 @item ^-C=^/USE_MAPPING_FILE=^@var{file}
9662 @cindex @option{^-C=^/USE_MAPPING^} (@command{gnatmake})
9663 Use a specific mapping file. The file, specified as a path name (absolute or
9664 relative) by this switch, should already exist, otherwise the switch is
9665 ineffective. The specified mapping file will be communicated to the compiler.
9666 This switch is not compatible with a project file
9667 (^-P^/PROJECT_FILE=^@var{file}) or with multiple compiling processes
9668 (^-j^/PROCESSES=^nnn, when nnn is greater than 1).
9670 @item ^-d^/DISPLAY_PROGRESS^
9671 @cindex @option{^-d^/DISPLAY_PROGRESS^} (@command{gnatmake})
9672 Display progress for each source, up to date or not, as a single line
9675 completed x out of y (zz%)
9678 If the file needs to be compiled this is displayed after the invocation of
9679 the compiler. These lines are displayed even in quiet output mode.
9681 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
9682 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@command{gnatmake})
9683 Put all object files and ALI file in directory @var{dir}.
9684 If the @option{^-D^/DIRECTORY_OBJECTS^} switch is not used, all object files
9685 and ALI files go in the current working directory.
9687 This switch cannot be used when using a project file.
9690 @cindex @option{-eI} (@command{gnatmake})
9691 Indicates that the main source is a multi-unit source and the rank of the unit
9692 in the source file is nnn. nnn needs to be a positive number and a valid
9693 index in the source. This switch cannot be used when @command{gnatmake} is
9694 invoked for several mains.
9698 @cindex @option{-eL} (@command{gnatmake})
9699 @cindex symbolic links
9700 Follow all symbolic links when processing project files.
9701 This should be used if your project uses symbolic links for files or
9702 directories, but is not needed in other cases.
9704 @cindex naming scheme
9705 This also assumes that no directory matches the naming scheme for files (for
9706 instance that you do not have a directory called "sources.ads" when using the
9707 default GNAT naming scheme).
9709 When you do not have to use this switch (i.e.@: by default), gnatmake is able to
9710 save a lot of system calls (several per source file and object file), which
9711 can result in a significant speed up to load and manipulate a project file,
9712 especially when using source files from a remote system.
9716 @item ^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^
9717 @cindex @option{^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^} (@command{gnatmake})
9718 Output the commands for the compiler, the binder and the linker
9719 on ^standard output^SYS$OUTPUT^,
9720 instead of ^standard error^SYS$ERROR^.
9722 @item ^-f^/FORCE_COMPILE^
9723 @cindex @option{^-f^/FORCE_COMPILE^} (@command{gnatmake})
9724 Force recompilations. Recompile all sources, even though some object
9725 files may be up to date, but don't recompile predefined or GNAT internal
9726 files or locked files (files with a write-protected ALI file),
9727 unless the @option{^-a^/ALL_FILES^} switch is also specified.
9729 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
9730 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@command{gnatmake})
9731 When using project files, if some errors or warnings are detected during
9732 parsing and verbose mode is not in effect (no use of switch
9733 ^-v^/VERBOSE^), then error lines start with the full path name of the project
9734 file, rather than its simple file name.
9737 @cindex @option{^-g^/DEBUG^} (@command{gnatmake})
9738 Enable debugging. This switch is simply passed to the compiler and to the
9741 @item ^-i^/IN_PLACE^
9742 @cindex @option{^-i^/IN_PLACE^} (@command{gnatmake})
9743 In normal mode, @command{gnatmake} compiles all object files and ALI files
9744 into the current directory. If the @option{^-i^/IN_PLACE^} switch is used,
9745 then instead object files and ALI files that already exist are overwritten
9746 in place. This means that once a large project is organized into separate
9747 directories in the desired manner, then @command{gnatmake} will automatically
9748 maintain and update this organization. If no ALI files are found on the
9749 Ada object path (@ref{Search Paths and the Run-Time Library (RTL)}),
9750 the new object and ALI files are created in the
9751 directory containing the source being compiled. If another organization
9752 is desired, where objects and sources are kept in different directories,
9753 a useful technique is to create dummy ALI files in the desired directories.
9754 When detecting such a dummy file, @command{gnatmake} will be forced to
9755 recompile the corresponding source file, and it will be put the resulting
9756 object and ALI files in the directory where it found the dummy file.
9758 @item ^-j^/PROCESSES=^@var{n}
9759 @cindex @option{^-j^/PROCESSES^} (@command{gnatmake})
9760 @cindex Parallel make
9761 Use @var{n} processes to carry out the (re)compilations. On a multiprocessor
9762 machine compilations will occur in parallel. If @var{n} is 0, then the
9763 maximum number of parallel compilations is the number of core processors
9764 on the platform. In the event of compilation errors, messages from various
9765 compilations might get interspersed (but @command{gnatmake} will give you the
9766 full ordered list of failing compiles at the end). If this is problematic,
9767 rerun the make process with n set to 1 to get a clean list of messages.
9769 @item ^-k^/CONTINUE_ON_ERROR^
9770 @cindex @option{^-k^/CONTINUE_ON_ERROR^} (@command{gnatmake})
9771 Keep going. Continue as much as possible after a compilation error. To
9772 ease the programmer's task in case of compilation errors, the list of
9773 sources for which the compile fails is given when @command{gnatmake}
9776 If @command{gnatmake} is invoked with several @file{file_names} and with this
9777 switch, if there are compilation errors when building an executable,
9778 @command{gnatmake} will not attempt to build the following executables.
9780 @item ^-l^/ACTIONS=LINK^
9781 @cindex @option{^-l^/ACTIONS=LINK^} (@command{gnatmake})
9782 Link only. Can be combined with @option{^-b^/ACTIONS=BIND^} to binding
9783 and linking. Linking will not be performed if combined with
9784 @option{^-c^/ACTIONS=COMPILE^}
9785 but not with @option{^-b^/ACTIONS=BIND^}.
9786 When not combined with @option{^-b^/ACTIONS=BIND^}
9787 all the units in the closure of the main program must have been previously
9788 compiled and must be up to date, and the main program needs to have been bound.
9789 The root unit specified by @var{file_name}
9790 may be given without extension, with the source extension or, if no GNAT
9791 Project File is specified, with the ALI file extension.
9793 @item ^-m^/MINIMAL_RECOMPILATION^
9794 @cindex @option{^-m^/MINIMAL_RECOMPILATION^} (@command{gnatmake})
9795 Specify that the minimum necessary amount of recompilations
9796 be performed. In this mode @command{gnatmake} ignores time
9797 stamp differences when the only
9798 modifications to a source file consist in adding/removing comments,
9799 empty lines, spaces or tabs. This means that if you have changed the
9800 comments in a source file or have simply reformatted it, using this
9801 switch will tell @command{gnatmake} not to recompile files that depend on it
9802 (provided other sources on which these files depend have undergone no
9803 semantic modifications). Note that the debugging information may be
9804 out of date with respect to the sources if the @option{-m} switch causes
9805 a compilation to be switched, so the use of this switch represents a
9806 trade-off between compilation time and accurate debugging information.
9808 @item ^-M^/DEPENDENCIES_LIST^
9809 @cindex Dependencies, producing list
9810 @cindex @option{^-M^/DEPENDENCIES_LIST^} (@command{gnatmake})
9811 Check if all objects are up to date. If they are, output the object
9812 dependences to @file{stdout} in a form that can be directly exploited in
9813 a @file{Makefile}. By default, each source file is prefixed with its
9814 (relative or absolute) directory name. This name is whatever you
9815 specified in the various @option{^-aI^/SOURCE_SEARCH^}
9816 and @option{^-I^/SEARCH^} switches. If you use
9817 @code{gnatmake ^-M^/DEPENDENCIES_LIST^}
9818 @option{^-q^/QUIET^}
9819 (see below), only the source file names,
9820 without relative paths, are output. If you just specify the
9821 @option{^-M^/DEPENDENCIES_LIST^}
9822 switch, dependencies of the GNAT internal system files are omitted. This
9823 is typically what you want. If you also specify
9824 the @option{^-a^/ALL_FILES^} switch,
9825 dependencies of the GNAT internal files are also listed. Note that
9826 dependencies of the objects in external Ada libraries (see switch
9827 @option{^-aL^/SKIP_MISSING=^}@var{dir} in the following list)
9830 @item ^-n^/DO_OBJECT_CHECK^
9831 @cindex @option{^-n^/DO_OBJECT_CHECK^} (@command{gnatmake})
9832 Don't compile, bind, or link. Checks if all objects are up to date.
9833 If they are not, the full name of the first file that needs to be
9834 recompiled is printed.
9835 Repeated use of this option, followed by compiling the indicated source
9836 file, will eventually result in recompiling all required units.
9838 @item ^-o ^/EXECUTABLE=^@var{exec_name}
9839 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatmake})
9840 Output executable name. The name of the final executable program will be
9841 @var{exec_name}. If the @option{^-o^/EXECUTABLE^} switch is omitted the default
9842 name for the executable will be the name of the input file in appropriate form
9843 for an executable file on the host system.
9845 This switch cannot be used when invoking @command{gnatmake} with several
9848 @item ^-p or --create-missing-dirs^/CREATE_MISSING_DIRS^
9849 @cindex @option{^-p^/CREATE_MISSING_DIRS^} (@command{gnatmake})
9850 When using project files (^-P^/PROJECT_FILE=^@var{project}), create
9851 automatically missing object directories, library directories and exec
9854 @item ^-P^/PROJECT_FILE=^@var{project}
9855 @cindex @option{^-P^/PROJECT_FILE^} (@command{gnatmake})
9856 Use project file @var{project}. Only one such switch can be used.
9857 @xref{gnatmake and Project Files}.
9860 @cindex @option{^-q^/QUIET^} (@command{gnatmake})
9861 Quiet. When this flag is not set, the commands carried out by
9862 @command{gnatmake} are displayed.
9864 @item ^-s^/SWITCH_CHECK/^
9865 @cindex @option{^-s^/SWITCH_CHECK^} (@command{gnatmake})
9866 Recompile if compiler switches have changed since last compilation.
9867 All compiler switches but -I and -o are taken into account in the
9869 orders between different ``first letter'' switches are ignored, but
9870 orders between same switches are taken into account. For example,
9871 @option{-O -O2} is different than @option{-O2 -O}, but @option{-g -O}
9872 is equivalent to @option{-O -g}.
9874 This switch is recommended when Integrated Preprocessing is used.
9877 @cindex @option{^-u^/UNIQUE^} (@command{gnatmake})
9878 Unique. Recompile at most the main files. It implies -c. Combined with
9879 -f, it is equivalent to calling the compiler directly. Note that using
9880 ^-u^/UNIQUE^ with a project file and no main has a special meaning
9881 (@pxref{Project Files and Main Subprograms}).
9883 @item ^-U^/ALL_PROJECTS^
9884 @cindex @option{^-U^/ALL_PROJECTS^} (@command{gnatmake})
9885 When used without a project file or with one or several mains on the command
9886 line, is equivalent to ^-u^/UNIQUE^. When used with a project file and no main
9887 on the command line, all sources of all project files are checked and compiled
9888 if not up to date, and libraries are rebuilt, if necessary.
9891 @cindex @option{^-v^/REASONS^} (@command{gnatmake})
9892 Verbose. Display the reason for all recompilations @command{gnatmake}
9893 decides are necessary, with the highest verbosity level.
9895 @item ^-vl^/LOW_VERBOSITY^
9896 @cindex @option{^-vl^/LOW_VERBOSITY^} (@command{gnatmake})
9897 Verbosity level Low. Display fewer lines than in verbosity Medium.
9899 @item ^-vm^/MEDIUM_VERBOSITY^
9900 @cindex @option{^-vm^/MEDIUM_VERBOSITY^} (@command{gnatmake})
9901 Verbosity level Medium. Potentially display fewer lines than in verbosity High.
9903 @item ^-vh^/HIGH_VERBOSITY^
9904 @cindex @option{^-vm^/HIGH_VERBOSITY^} (@command{gnatmake})
9905 Verbosity level High. Equivalent to ^-v^/REASONS^.
9907 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
9908 Indicate the verbosity of the parsing of GNAT project files.
9909 @xref{Switches Related to Project Files}.
9911 @item ^-x^/NON_PROJECT_UNIT_COMPILATION^
9912 @cindex @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} (@command{gnatmake})
9913 Indicate that sources that are not part of any Project File may be compiled.
9914 Normally, when using Project Files, only sources that are part of a Project
9915 File may be compile. When this switch is used, a source outside of all Project
9916 Files may be compiled. The ALI file and the object file will be put in the
9917 object directory of the main Project. The compilation switches used will only
9918 be those specified on the command line. Even when
9919 @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} is used, mains specified on the
9920 command line need to be sources of a project file.
9922 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
9923 Indicate that external variable @var{name} has the value @var{value}.
9924 The Project Manager will use this value for occurrences of
9925 @code{external(name)} when parsing the project file.
9926 @xref{Switches Related to Project Files}.
9929 @cindex @option{^-z^/NOMAIN^} (@command{gnatmake})
9930 No main subprogram. Bind and link the program even if the unit name
9931 given on the command line is a package name. The resulting executable
9932 will execute the elaboration routines of the package and its closure,
9933 then the finalization routines.
9938 @item @command{gcc} @asis{switches}
9940 Any uppercase or multi-character switch that is not a @command{gnatmake} switch
9941 is passed to @command{gcc} (e.g.@: @option{-O}, @option{-gnato,} etc.)
9944 Any qualifier that cannot be recognized as a qualifier for @code{GNAT MAKE}
9945 but is recognizable as a valid qualifier for @code{GNAT COMPILE} is
9946 automatically treated as a compiler switch, and passed on to all
9947 compilations that are carried out.
9952 Source and library search path switches:
9956 @item ^-aI^/SOURCE_SEARCH=^@var{dir}
9957 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatmake})
9958 When looking for source files also look in directory @var{dir}.
9959 The order in which source files search is undertaken is
9960 described in @ref{Search Paths and the Run-Time Library (RTL)}.
9962 @item ^-aL^/SKIP_MISSING=^@var{dir}
9963 @cindex @option{^-aL^/SKIP_MISSING^} (@command{gnatmake})
9964 Consider @var{dir} as being an externally provided Ada library.
9965 Instructs @command{gnatmake} to skip compilation units whose @file{.ALI}
9966 files have been located in directory @var{dir}. This allows you to have
9967 missing bodies for the units in @var{dir} and to ignore out of date bodies
9968 for the same units. You still need to specify
9969 the location of the specs for these units by using the switches
9970 @option{^-aI^/SOURCE_SEARCH=^@var{dir}}
9971 or @option{^-I^/SEARCH=^@var{dir}}.
9972 Note: this switch is provided for compatibility with previous versions
9973 of @command{gnatmake}. The easier method of causing standard libraries
9974 to be excluded from consideration is to write-protect the corresponding
9977 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
9978 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatmake})
9979 When searching for library and object files, look in directory
9980 @var{dir}. The order in which library files are searched is described in
9981 @ref{Search Paths for gnatbind}.
9983 @item ^-A^/CONDITIONAL_SOURCE_SEARCH=^@var{dir}
9984 @cindex Search paths, for @command{gnatmake}
9985 @cindex @option{^-A^/CONDITIONAL_SOURCE_SEARCH^} (@command{gnatmake})
9986 Equivalent to @option{^-aL^/SKIP_MISSING=^@var{dir}
9987 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
9989 @item ^-I^/SEARCH=^@var{dir}
9990 @cindex @option{^-I^/SEARCH^} (@command{gnatmake})
9991 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}
9992 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
9994 @item ^-I-^/NOCURRENT_DIRECTORY^
9995 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatmake})
9996 @cindex Source files, suppressing search
9997 Do not look for source files in the directory containing the source
9998 file named in the command line.
9999 Do not look for ALI or object files in the directory
10000 where @command{gnatmake} was invoked.
10002 @item ^-L^/LIBRARY_SEARCH=^@var{dir}
10003 @cindex @option{^-L^/LIBRARY_SEARCH^} (@command{gnatmake})
10004 @cindex Linker libraries
10005 Add directory @var{dir} to the list of directories in which the linker
10006 will search for libraries. This is equivalent to
10007 @option{-largs ^-L^/LIBRARY_SEARCH=^}@var{dir}.
10009 Furthermore, under Windows, the sources pointed to by the libraries path
10010 set in the registry are not searched for.
10014 @cindex @option{-nostdinc} (@command{gnatmake})
10015 Do not look for source files in the system default directory.
10018 @cindex @option{-nostdlib} (@command{gnatmake})
10019 Do not look for library files in the system default directory.
10021 @item --RTS=@var{rts-path}
10022 @cindex @option{--RTS} (@command{gnatmake})
10023 Specifies the default location of the runtime library. GNAT looks for the
10025 in the following directories, and stops as soon as a valid runtime is found
10026 (@file{adainclude} or @file{ada_source_path}, and @file{adalib} or
10027 @file{ada_object_path} present):
10030 @item <current directory>/$rts_path
10032 @item <default-search-dir>/$rts_path
10034 @item <default-search-dir>/rts-$rts_path
10038 The selected path is handled like a normal RTS path.
10042 @node Mode Switches for gnatmake
10043 @section Mode Switches for @command{gnatmake}
10046 The mode switches (referred to as @code{mode_switches}) allow the
10047 inclusion of switches that are to be passed to the compiler itself, the
10048 binder or the linker. The effect of a mode switch is to cause all
10049 subsequent switches up to the end of the switch list, or up to the next
10050 mode switch, to be interpreted as switches to be passed on to the
10051 designated component of GNAT.
10055 @item -cargs @var{switches}
10056 @cindex @option{-cargs} (@command{gnatmake})
10057 Compiler switches. Here @var{switches} is a list of switches
10058 that are valid switches for @command{gcc}. They will be passed on to
10059 all compile steps performed by @command{gnatmake}.
10061 @item -bargs @var{switches}
10062 @cindex @option{-bargs} (@command{gnatmake})
10063 Binder switches. Here @var{switches} is a list of switches
10064 that are valid switches for @code{gnatbind}. They will be passed on to
10065 all bind steps performed by @command{gnatmake}.
10067 @item -largs @var{switches}
10068 @cindex @option{-largs} (@command{gnatmake})
10069 Linker switches. Here @var{switches} is a list of switches
10070 that are valid switches for @command{gnatlink}. They will be passed on to
10071 all link steps performed by @command{gnatmake}.
10073 @item -margs @var{switches}
10074 @cindex @option{-margs} (@command{gnatmake})
10075 Make switches. The switches are directly interpreted by @command{gnatmake},
10076 regardless of any previous occurrence of @option{-cargs}, @option{-bargs}
10077 or @option{-largs}.
10080 @node Notes on the Command Line
10081 @section Notes on the Command Line
10084 This section contains some additional useful notes on the operation
10085 of the @command{gnatmake} command.
10089 @cindex Recompilation, by @command{gnatmake}
10090 If @command{gnatmake} finds no ALI files, it recompiles the main program
10091 and all other units required by the main program.
10092 This means that @command{gnatmake}
10093 can be used for the initial compile, as well as during subsequent steps of
10094 the development cycle.
10097 If you enter @code{gnatmake @var{file}.adb}, where @file{@var{file}.adb}
10098 is a subunit or body of a generic unit, @command{gnatmake} recompiles
10099 @file{@var{file}.adb} (because it finds no ALI) and stops, issuing a
10103 In @command{gnatmake} the switch @option{^-I^/SEARCH^}
10104 is used to specify both source and
10105 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
10106 instead if you just want to specify
10107 source paths only and @option{^-aO^/OBJECT_SEARCH^}
10108 if you want to specify library paths
10112 @command{gnatmake} will ignore any files whose ALI file is write-protected.
10113 This may conveniently be used to exclude standard libraries from
10114 consideration and in particular it means that the use of the
10115 @option{^-f^/FORCE_COMPILE^} switch will not recompile these files
10116 unless @option{^-a^/ALL_FILES^} is also specified.
10119 @command{gnatmake} has been designed to make the use of Ada libraries
10120 particularly convenient. Assume you have an Ada library organized
10121 as follows: @i{^obj-dir^[OBJ_DIR]^} contains the objects and ALI files for
10122 of your Ada compilation units,
10123 whereas @i{^include-dir^[INCLUDE_DIR]^} contains the
10124 specs of these units, but no bodies. Then to compile a unit
10125 stored in @code{main.adb}, which uses this Ada library you would just type
10129 $ gnatmake -aI@var{include-dir} -aL@var{obj-dir} main
10132 $ gnatmake /SOURCE_SEARCH=@i{[INCLUDE_DIR]}
10133 /SKIP_MISSING=@i{[OBJ_DIR]} main
10138 Using @command{gnatmake} along with the
10139 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}
10140 switch provides a mechanism for avoiding unnecessary recompilations. Using
10142 you can update the comments/format of your
10143 source files without having to recompile everything. Note, however, that
10144 adding or deleting lines in a source files may render its debugging
10145 info obsolete. If the file in question is a spec, the impact is rather
10146 limited, as that debugging info will only be useful during the
10147 elaboration phase of your program. For bodies the impact can be more
10148 significant. In all events, your debugger will warn you if a source file
10149 is more recent than the corresponding object, and alert you to the fact
10150 that the debugging information may be out of date.
10153 @node How gnatmake Works
10154 @section How @command{gnatmake} Works
10157 Generally @command{gnatmake} automatically performs all necessary
10158 recompilations and you don't need to worry about how it works. However,
10159 it may be useful to have some basic understanding of the @command{gnatmake}
10160 approach and in particular to understand how it uses the results of
10161 previous compilations without incorrectly depending on them.
10163 First a definition: an object file is considered @dfn{up to date} if the
10164 corresponding ALI file exists and if all the source files listed in the
10165 dependency section of this ALI file have time stamps matching those in
10166 the ALI file. This means that neither the source file itself nor any
10167 files that it depends on have been modified, and hence there is no need
10168 to recompile this file.
10170 @command{gnatmake} works by first checking if the specified main unit is up
10171 to date. If so, no compilations are required for the main unit. If not,
10172 @command{gnatmake} compiles the main program to build a new ALI file that
10173 reflects the latest sources. Then the ALI file of the main unit is
10174 examined to find all the source files on which the main program depends,
10175 and @command{gnatmake} recursively applies the above procedure on all these
10178 This process ensures that @command{gnatmake} only trusts the dependencies
10179 in an existing ALI file if they are known to be correct. Otherwise it
10180 always recompiles to determine a new, guaranteed accurate set of
10181 dependencies. As a result the program is compiled ``upside down'' from what may
10182 be more familiar as the required order of compilation in some other Ada
10183 systems. In particular, clients are compiled before the units on which
10184 they depend. The ability of GNAT to compile in any order is critical in
10185 allowing an order of compilation to be chosen that guarantees that
10186 @command{gnatmake} will recompute a correct set of new dependencies if
10189 When invoking @command{gnatmake} with several @var{file_names}, if a unit is
10190 imported by several of the executables, it will be recompiled at most once.
10192 Note: when using non-standard naming conventions
10193 (@pxref{Using Other File Names}), changing through a configuration pragmas
10194 file the version of a source and invoking @command{gnatmake} to recompile may
10195 have no effect, if the previous version of the source is still accessible
10196 by @command{gnatmake}. It may be necessary to use the switch
10197 ^-f^/FORCE_COMPILE^.
10199 @node Examples of gnatmake Usage
10200 @section Examples of @command{gnatmake} Usage
10203 @item gnatmake hello.adb
10204 Compile all files necessary to bind and link the main program
10205 @file{hello.adb} (containing unit @code{Hello}) and bind and link the
10206 resulting object files to generate an executable file @file{^hello^HELLO.EXE^}.
10208 @item gnatmake main1 main2 main3
10209 Compile all files necessary to bind and link the main programs
10210 @file{main1.adb} (containing unit @code{Main1}), @file{main2.adb}
10211 (containing unit @code{Main2}) and @file{main3.adb}
10212 (containing unit @code{Main3}) and bind and link the resulting object files
10213 to generate three executable files @file{^main1^MAIN1.EXE^},
10214 @file{^main2^MAIN2.EXE^}
10215 and @file{^main3^MAIN3.EXE^}.
10218 @item gnatmake -q Main_Unit -cargs -O2 -bargs -l
10222 @item gnatmake Main_Unit /QUIET
10223 /COMPILER_QUALIFIERS /OPTIMIZE=ALL
10224 /BINDER_QUALIFIERS /ORDER_OF_ELABORATION
10226 Compile all files necessary to bind and link the main program unit
10227 @code{Main_Unit} (from file @file{main_unit.adb}). All compilations will
10228 be done with optimization level 2 and the order of elaboration will be
10229 listed by the binder. @command{gnatmake} will operate in quiet mode, not
10230 displaying commands it is executing.
10233 @c *************************
10234 @node Improving Performance
10235 @chapter Improving Performance
10236 @cindex Improving performance
10239 This chapter presents several topics related to program performance.
10240 It first describes some of the tradeoffs that need to be considered
10241 and some of the techniques for making your program run faster.
10243 @ifclear FSFEDITION
10244 the @command{gnatelim} tool and
10246 unused subprogram/data
10247 elimination feature, which can reduce the size of program executables.
10251 * Performance Considerations::
10252 * Text_IO Suggestions::
10253 @ifclear FSFEDITION
10254 * Reducing Size of Ada Executables with gnatelim::
10256 * Reducing Size of Executables with unused subprogram/data elimination::
10260 @c *****************************
10261 @node Performance Considerations
10262 @section Performance Considerations
10265 The GNAT system provides a number of options that allow a trade-off
10270 performance of the generated code
10273 speed of compilation
10276 minimization of dependences and recompilation
10279 the degree of run-time checking.
10283 The defaults (if no options are selected) aim at improving the speed
10284 of compilation and minimizing dependences, at the expense of performance
10285 of the generated code:
10292 no inlining of subprogram calls
10295 all run-time checks enabled except overflow and elaboration checks
10299 These options are suitable for most program development purposes. This
10300 chapter describes how you can modify these choices, and also provides
10301 some guidelines on debugging optimized code.
10304 * Controlling Run-Time Checks::
10305 * Use of Restrictions::
10306 * Optimization Levels::
10307 * Debugging Optimized Code::
10308 * Inlining of Subprograms::
10309 * Vectorization of loops::
10310 * Other Optimization Switches::
10311 * Optimization and Strict Aliasing::
10312 * Aliased Variables and Optimization::
10313 * Atomic Variables and Optimization::
10314 * Passive Task Optimization::
10317 * Coverage Analysis::
10321 @node Controlling Run-Time Checks
10322 @subsection Controlling Run-Time Checks
10325 By default, GNAT generates all run-time checks, except integer overflow
10326 checks, stack overflow checks, and checks for access before elaboration on
10327 subprogram calls. The latter are not required in default mode, because all
10328 necessary checking is done at compile time.
10329 @cindex @option{-gnatp} (@command{gcc})
10330 @cindex @option{-gnato} (@command{gcc})
10331 Two gnat switches, @option{-gnatp} and @option{-gnato} allow this default to
10332 be modified. @xref{Run-Time Checks}.
10334 Our experience is that the default is suitable for most development
10337 We treat integer overflow specially because these
10338 are quite expensive and in our experience are not as important as other
10339 run-time checks in the development process. Note that division by zero
10340 is not considered an overflow check, and divide by zero checks are
10341 generated where required by default.
10343 Elaboration checks are off by default, and also not needed by default, since
10344 GNAT uses a static elaboration analysis approach that avoids the need for
10345 run-time checking. This manual contains a full chapter discussing the issue
10346 of elaboration checks, and if the default is not satisfactory for your use,
10347 you should read this chapter.
10349 For validity checks, the minimal checks required by the Ada Reference
10350 Manual (for case statements and assignments to array elements) are on
10351 by default. These can be suppressed by use of the @option{-gnatVn} switch.
10352 Note that in Ada 83, there were no validity checks, so if the Ada 83 mode
10353 is acceptable (or when comparing GNAT performance with an Ada 83 compiler),
10354 it may be reasonable to routinely use @option{-gnatVn}. Validity checks
10355 are also suppressed entirely if @option{-gnatp} is used.
10357 @cindex Overflow checks
10358 @cindex Checks, overflow
10361 @cindex pragma Suppress
10362 @cindex pragma Unsuppress
10363 Note that the setting of the switches controls the default setting of
10364 the checks. They may be modified using either @code{pragma Suppress} (to
10365 remove checks) or @code{pragma Unsuppress} (to add back suppressed
10366 checks) in the program source.
10368 @node Use of Restrictions
10369 @subsection Use of Restrictions
10372 The use of pragma Restrictions allows you to control which features are
10373 permitted in your program. Apart from the obvious point that if you avoid
10374 relatively expensive features like finalization (enforceable by the use
10375 of pragma Restrictions (No_Finalization), the use of this pragma does not
10376 affect the generated code in most cases.
10378 One notable exception to this rule is that the possibility of task abort
10379 results in some distributed overhead, particularly if finalization or
10380 exception handlers are used. The reason is that certain sections of code
10381 have to be marked as non-abortable.
10383 If you use neither the @code{abort} statement, nor asynchronous transfer
10384 of control (@code{select @dots{} then abort}), then this distributed overhead
10385 is removed, which may have a general positive effect in improving
10386 overall performance. Especially code involving frequent use of tasking
10387 constructs and controlled types will show much improved performance.
10388 The relevant restrictions pragmas are
10390 @smallexample @c ada
10391 pragma Restrictions (No_Abort_Statements);
10392 pragma Restrictions (Max_Asynchronous_Select_Nesting => 0);
10396 It is recommended that these restriction pragmas be used if possible. Note
10397 that this also means that you can write code without worrying about the
10398 possibility of an immediate abort at any point.
10400 @node Optimization Levels
10401 @subsection Optimization Levels
10402 @cindex @option{^-O^/OPTIMIZE^} (@command{gcc})
10405 Without any optimization ^option,^qualifier,^
10406 the compiler's goal is to reduce the cost of
10407 compilation and to make debugging produce the expected results.
10408 Statements are independent: if you stop the program with a breakpoint between
10409 statements, you can then assign a new value to any variable or change
10410 the program counter to any other statement in the subprogram and get exactly
10411 the results you would expect from the source code.
10413 Turning on optimization makes the compiler attempt to improve the
10414 performance and/or code size at the expense of compilation time and
10415 possibly the ability to debug the program.
10417 If you use multiple
10418 ^-O options, with or without level numbers,^/OPTIMIZE qualifiers,^
10419 the last such option is the one that is effective.
10422 The default is optimization off. This results in the fastest compile
10423 times, but GNAT makes absolutely no attempt to optimize, and the
10424 generated programs are considerably larger and slower than when
10425 optimization is enabled. You can use the
10427 @option{-O} switch (the permitted forms are @option{-O0}, @option{-O1}
10428 @option{-O2}, @option{-O3}, and @option{-Os})
10431 @code{OPTIMIZE} qualifier
10433 to @command{gcc} to control the optimization level:
10436 @item ^-O0^/OPTIMIZE=NONE^
10437 No optimization (the default);
10438 generates unoptimized code but has
10439 the fastest compilation time.
10441 Note that many other compilers do fairly extensive optimization
10442 even if ``no optimization'' is specified. With gcc, it is
10443 very unusual to use ^-O0^/OPTIMIZE=NONE^ for production if
10444 execution time is of any concern, since ^-O0^/OPTIMIZE=NONE^
10445 really does mean no optimization at all. This difference between
10446 gcc and other compilers should be kept in mind when doing
10447 performance comparisons.
10449 @item ^-O1^/OPTIMIZE=SOME^
10450 Moderate optimization;
10451 optimizes reasonably well but does not
10452 degrade compilation time significantly.
10454 @item ^-O2^/OPTIMIZE=ALL^
10456 @itemx /OPTIMIZE=DEVELOPMENT
10459 generates highly optimized code and has
10460 the slowest compilation time.
10462 @item ^-O3^/OPTIMIZE=INLINING^
10463 Full optimization as in @option{-O2};
10464 also uses more aggressive automatic inlining of subprograms within a unit
10465 (@pxref{Inlining of Subprograms}) and attempts to vectorize loops.
10467 @item ^-Os^/OPTIMIZE=SPACE^
10468 Optimize space usage (code and data) of resulting program.
10472 Higher optimization levels perform more global transformations on the
10473 program and apply more expensive analysis algorithms in order to generate
10474 faster and more compact code. The price in compilation time, and the
10475 resulting improvement in execution time,
10476 both depend on the particular application and the hardware environment.
10477 You should experiment to find the best level for your application.
10479 Since the precise set of optimizations done at each level will vary from
10480 release to release (and sometime from target to target), it is best to think
10481 of the optimization settings in general terms.
10482 @xref{Optimize Options,, Options That Control Optimization, gcc, Using
10483 the GNU Compiler Collection (GCC)}, for details about
10484 ^the @option{-O} settings and a number of @option{-f} options that^how to^
10485 individually enable or disable specific optimizations.
10487 Unlike some other compilation systems, ^@command{gcc}^GNAT^ has
10488 been tested extensively at all optimization levels. There are some bugs
10489 which appear only with optimization turned on, but there have also been
10490 bugs which show up only in @emph{unoptimized} code. Selecting a lower
10491 level of optimization does not improve the reliability of the code
10492 generator, which in practice is highly reliable at all optimization
10495 Note regarding the use of @option{-O3}: The use of this optimization level
10496 is generally discouraged with GNAT, since it often results in larger
10497 executables which may run more slowly. See further discussion of this point
10498 in @ref{Inlining of Subprograms}.
10500 @node Debugging Optimized Code
10501 @subsection Debugging Optimized Code
10502 @cindex Debugging optimized code
10503 @cindex Optimization and debugging
10506 Although it is possible to do a reasonable amount of debugging at
10508 nonzero optimization levels,
10509 the higher the level the more likely that
10512 @option{/OPTIMIZE} settings other than @code{NONE},
10513 such settings will make it more likely that
10515 source-level constructs will have been eliminated by optimization.
10516 For example, if a loop is strength-reduced, the loop
10517 control variable may be completely eliminated and thus cannot be
10518 displayed in the debugger.
10519 This can only happen at @option{-O2} or @option{-O3}.
10520 Explicit temporary variables that you code might be eliminated at
10521 ^level^setting^ @option{-O1} or higher.
10523 The use of the @option{^-g^/DEBUG^} switch,
10524 @cindex @option{^-g^/DEBUG^} (@command{gcc})
10525 which is needed for source-level debugging,
10526 affects the size of the program executable on disk,
10527 and indeed the debugging information can be quite large.
10528 However, it has no effect on the generated code (and thus does not
10529 degrade performance)
10531 Since the compiler generates debugging tables for a compilation unit before
10532 it performs optimizations, the optimizing transformations may invalidate some
10533 of the debugging data. You therefore need to anticipate certain
10534 anomalous situations that may arise while debugging optimized code.
10535 These are the most common cases:
10539 @i{The ``hopping Program Counter'':} Repeated @code{step} or @code{next}
10541 the PC bouncing back and forth in the code. This may result from any of
10542 the following optimizations:
10546 @i{Common subexpression elimination:} using a single instance of code for a
10547 quantity that the source computes several times. As a result you
10548 may not be able to stop on what looks like a statement.
10551 @i{Invariant code motion:} moving an expression that does not change within a
10552 loop, to the beginning of the loop.
10555 @i{Instruction scheduling:} moving instructions so as to
10556 overlap loads and stores (typically) with other code, or in
10557 general to move computations of values closer to their uses. Often
10558 this causes you to pass an assignment statement without the assignment
10559 happening and then later bounce back to the statement when the
10560 value is actually needed. Placing a breakpoint on a line of code
10561 and then stepping over it may, therefore, not always cause all the
10562 expected side-effects.
10566 @i{The ``big leap'':} More commonly known as @emph{cross-jumping}, in which
10567 two identical pieces of code are merged and the program counter suddenly
10568 jumps to a statement that is not supposed to be executed, simply because
10569 it (and the code following) translates to the same thing as the code
10570 that @emph{was} supposed to be executed. This effect is typically seen in
10571 sequences that end in a jump, such as a @code{goto}, a @code{return}, or
10572 a @code{break} in a C @code{^switch^switch^} statement.
10575 @i{The ``roving variable'':} The symptom is an unexpected value in a variable.
10576 There are various reasons for this effect:
10580 In a subprogram prologue, a parameter may not yet have been moved to its
10584 A variable may be dead, and its register re-used. This is
10585 probably the most common cause.
10588 As mentioned above, the assignment of a value to a variable may
10592 A variable may be eliminated entirely by value propagation or
10593 other means. In this case, GCC may incorrectly generate debugging
10594 information for the variable
10598 In general, when an unexpected value appears for a local variable or parameter
10599 you should first ascertain if that value was actually computed by
10600 your program, as opposed to being incorrectly reported by the debugger.
10602 array elements in an object designated by an access value
10603 are generally less of a problem, once you have ascertained that the access
10605 Typically, this means checking variables in the preceding code and in the
10606 calling subprogram to verify that the value observed is explainable from other
10607 values (one must apply the procedure recursively to those
10608 other values); or re-running the code and stopping a little earlier
10609 (perhaps before the call) and stepping to better see how the variable obtained
10610 the value in question; or continuing to step @emph{from} the point of the
10611 strange value to see if code motion had simply moved the variable's
10616 In light of such anomalies, a recommended technique is to use @option{-O0}
10617 early in the software development cycle, when extensive debugging capabilities
10618 are most needed, and then move to @option{-O1} and later @option{-O2} as
10619 the debugger becomes less critical.
10620 Whether to use the @option{^-g^/DEBUG^} switch in the release version is
10621 a release management issue.
10623 Note that if you use @option{-g} you can then use the @command{strip} program
10624 on the resulting executable,
10625 which removes both debugging information and global symbols.
10628 @node Inlining of Subprograms
10629 @subsection Inlining of Subprograms
10632 A call to a subprogram in the current unit is inlined if all the
10633 following conditions are met:
10637 The optimization level is at least @option{-O1}.
10640 The called subprogram is suitable for inlining: It must be small enough
10641 and not contain something that @command{gcc} cannot support in inlined
10645 @cindex pragma Inline
10647 Any one of the following applies: @code{pragma Inline} is applied to the
10648 subprogram and the @option{^-gnatn^/INLINE^} switch is specified; the
10649 subprogram is local to the unit and called once from within it; the
10650 subprogram is small and optimization level @option{-O2} is specified;
10651 optimization level @option{-O3} is specified.
10655 Calls to subprograms in @code{with}'ed units are normally not inlined.
10656 To achieve actual inlining (that is, replacement of the call by the code
10657 in the body of the subprogram), the following conditions must all be true:
10661 The optimization level is at least @option{-O1}.
10664 The called subprogram is suitable for inlining: It must be small enough
10665 and not contain something that @command{gcc} cannot support in inlined
10669 The call appears in a body (not in a package spec).
10672 There is a @code{pragma Inline} for the subprogram.
10675 The @option{^-gnatn^/INLINE^} switch is used on the command line.
10678 Even if all these conditions are met, it may not be possible for
10679 the compiler to inline the call, due to the length of the body,
10680 or features in the body that make it impossible for the compiler
10681 to do the inlining.
10683 Note that specifying the @option{-gnatn} switch causes additional
10684 compilation dependencies. Consider the following:
10686 @smallexample @c ada
10706 With the default behavior (no @option{-gnatn} switch specified), the
10707 compilation of the @code{Main} procedure depends only on its own source,
10708 @file{main.adb}, and the spec of the package in file @file{r.ads}. This
10709 means that editing the body of @code{R} does not require recompiling
10712 On the other hand, the call @code{R.Q} is not inlined under these
10713 circumstances. If the @option{-gnatn} switch is present when @code{Main}
10714 is compiled, the call will be inlined if the body of @code{Q} is small
10715 enough, but now @code{Main} depends on the body of @code{R} in
10716 @file{r.adb} as well as on the spec. This means that if this body is edited,
10717 the main program must be recompiled. Note that this extra dependency
10718 occurs whether or not the call is in fact inlined by @command{gcc}.
10720 The use of front end inlining with @option{-gnatN} generates similar
10721 additional dependencies.
10723 @cindex @option{^-fno-inline^/INLINE=SUPPRESS^} (@command{gcc})
10724 Note: The @option{^-fno-inline^/INLINE=SUPPRESS^} switch
10725 can be used to prevent
10726 all inlining. This switch overrides all other conditions and ensures
10727 that no inlining occurs. The extra dependences resulting from
10728 @option{-gnatn} will still be active, even if
10729 this switch is used to suppress the resulting inlining actions.
10731 @cindex @option{-fno-inline-functions} (@command{gcc})
10732 Note: The @option{-fno-inline-functions} switch can be used to prevent
10733 automatic inlining of subprograms if @option{-O3} is used.
10735 @cindex @option{-fno-inline-small-functions} (@command{gcc})
10736 Note: The @option{-fno-inline-small-functions} switch can be used to prevent
10737 automatic inlining of small subprograms if @option{-O2} is used.
10739 @cindex @option{-fno-inline-functions-called-once} (@command{gcc})
10740 Note: The @option{-fno-inline-functions-called-once} switch
10741 can be used to prevent inlining of subprograms local to the unit
10742 and called once from within it if @option{-O1} is used.
10744 Note regarding the use of @option{-O3}: @option{-gnatn} is made up of two
10745 sub-switches @option{-gnatn1} and @option{-gnatn2} that can be directly
10746 specified in lieu of it, @option{-gnatn} being translated into one of them
10747 based on the optimization level. With @option{-O2} or below, @option{-gnatn}
10748 is equivalent to @option{-gnatn1} which activates pragma @code{Inline} with
10749 moderate inlining across modules. With @option{-O3}, @option{-gnatn} is
10750 equivalent to @option{-gnatn2} which activates pragma @code{Inline} with
10751 full inlining across modules. If you have used pragma @code{Inline} in appropriate cases, then it is usually much better to use @option{-O2} and @option{-gnatn} and avoid the use of @option{-O3} which has the additional
10752 effect of inlining subprograms you did not think should be inlined. We have
10753 found that the use of @option{-O3} may slow down the compilation and increase
10754 the code size by performing excessive inlining, leading to increased
10755 instruction cache pressure from the increased code size and thus minor
10756 performance improvements. So the bottom line here is that you should not
10757 automatically assume that @option{-O3} is better than @option{-O2}, and
10758 indeed you should use @option{-O3} only if tests show that it actually
10759 improves performance for your program.
10761 @node Vectorization of loops
10762 @subsection Vectorization of loops
10763 @cindex Optimization Switches
10765 You can take advantage of the auto-vectorizer present in the @command{gcc}
10766 back end to vectorize loops with GNAT. The corresponding command line switch
10767 is @option{-ftree-vectorize} but, as it is enabled by default at @option{-O3}
10768 and other aggressive optimizations helpful for vectorization also are enabled
10769 by default at this level, using @option{-O3} directly is recommended.
10771 You also need to make sure that the target architecture features a supported
10772 SIMD instruction set. For example, for the x86 architecture, you should at
10773 least specify @option{-msse2} to get significant vectorization (but you don't
10774 need to specify it for x86-64 as it is part of the base 64-bit architecture).
10775 Similarly, for the PowerPC architecture, you should specify @option{-maltivec}.
10777 The preferred loop form for vectorization is the @code{for} iteration scheme.
10778 Loops with a @code{while} iteration scheme can also be vectorized if they are
10779 very simple, but the vectorizer will quickly give up otherwise. With either
10780 iteration scheme, the flow of control must be straight, in particular no
10781 @code{exit} statement may appear in the loop body. The loop may however
10782 contain a single nested loop, if it can be vectorized when considered alone:
10784 @smallexample @c ada
10786 A : array (1..4, 1..4) of Long_Float;
10787 S : array (1..4) of Long_Float;
10791 for I in A'Range(1) loop
10792 for J in A'Range(2) loop
10793 S (I) := S (I) + A (I, J);
10800 The vectorizable operations depend on the targeted SIMD instruction set, but
10801 the adding and some of the multiplying operators are generally supported, as
10802 well as the logical operators for modular types. Note that, in the former
10803 case, enabling overflow checks, for example with @option{-gnato}, totally
10804 disables vectorization. The other checks are not supposed to have the same
10805 definitive effect, although compiling with @option{-gnatp} might well reveal
10806 cases where some checks do thwart vectorization.
10808 Type conversions may also prevent vectorization if they involve semantics that
10809 are not directly supported by the code generator or the SIMD instruction set.
10810 A typical example is direct conversion from floating-point to integer types.
10811 The solution in this case is to use the following idiom:
10813 @smallexample @c ada
10814 Integer (S'Truncation (F))
10818 if @code{S} is the subtype of floating-point object @code{F}.
10820 In most cases, the vectorizable loops are loops that iterate over arrays.
10821 All kinds of array types are supported, i.e. constrained array types with
10824 @smallexample @c ada
10825 type Array_Type is array (1 .. 4) of Long_Float;
10829 constrained array types with dynamic bounds:
10831 @smallexample @c ada
10832 type Array_Type is array (1 .. Q.N) of Long_Float;
10834 type Array_Type is array (Q.K .. 4) of Long_Float;
10836 type Array_Type is array (Q.K .. Q.N) of Long_Float;
10840 or unconstrained array types:
10842 @smallexample @c ada
10843 type Array_Type is array (Positive range <>) of Long_Float;
10847 The quality of the generated code decreases when the dynamic aspect of the
10848 array type increases, the worst code being generated for unconstrained array
10849 types. This is so because, the less information the compiler has about the
10850 bounds of the array, the more fallback code it needs to generate in order to
10851 fix things up at run time.
10853 It is possible to specify that a given loop should be subject to vectorization
10854 preferably to other optimizations by means of pragma @code{Loop_Optimize}:
10856 @smallexample @c ada
10857 pragma Loop_Optimize (Vector);
10861 placed immediately within the loop will convey the appropriate hint to the
10862 compiler for this loop.
10864 It is also possible to help the compiler generate better vectorized code
10865 for a given loop by asserting that there are no loop-carried dependencies
10866 in the loop. Consider for example the procedure:
10868 @smallexample @c ada
10869 type Arr is array (1 .. 4) of Long_Float;
10871 procedure Add (X, Y : not null access Arr; R : not null access Arr) is
10873 for I in Arr'Range loop
10874 R(I) := X(I) + Y(I);
10880 By default, the compiler cannot unconditionally vectorize the loop because
10881 assigning to a component of the array designated by R in one iteration could
10882 change the value read from the components of the arrays designated by X or Y
10883 in a later iteration. As a result, the compiler will generate two versions
10884 of the loop in the object code, one vectorized and the other not vectorized,
10885 as well as a test to select the appropriate version at run time. This can
10886 be overcome by another hint:
10888 @smallexample @c ada
10889 pragma Loop_Optimize (Ivdep);
10893 placed immediately within the loop will tell the compiler that it can safely
10894 omit the non-vectorized version of the loop as well as the run-time test.
10896 @node Other Optimization Switches
10897 @subsection Other Optimization Switches
10898 @cindex Optimization Switches
10900 Since @code{GNAT} uses the @command{gcc} back end, all the specialized
10901 @command{gcc} optimization switches are potentially usable. These switches
10902 have not been extensively tested with GNAT but can generally be expected
10903 to work. Examples of switches in this category are @option{-funroll-loops}
10904 and the various target-specific @option{-m} options (in particular, it has
10905 been observed that @option{-march=xxx} can significantly improve performance
10906 on appropriate machines). For full details of these switches, see
10907 @ref{Submodel Options,, Hardware Models and Configurations, gcc, Using
10908 the GNU Compiler Collection (GCC)}.
10910 @node Optimization and Strict Aliasing
10911 @subsection Optimization and Strict Aliasing
10913 @cindex Strict Aliasing
10914 @cindex No_Strict_Aliasing
10917 The strong typing capabilities of Ada allow an optimizer to generate
10918 efficient code in situations where other languages would be forced to
10919 make worst case assumptions preventing such optimizations. Consider
10920 the following example:
10922 @smallexample @c ada
10925 type Int1 is new Integer;
10926 type Int2 is new Integer;
10927 type Int1A is access Int1;
10928 type Int2A is access Int2;
10935 for J in Data'Range loop
10936 if Data (J) = Int1V.all then
10937 Int2V.all := Int2V.all + 1;
10946 In this example, since the variable @code{Int1V} can only access objects
10947 of type @code{Int1}, and @code{Int2V} can only access objects of type
10948 @code{Int2}, there is no possibility that the assignment to
10949 @code{Int2V.all} affects the value of @code{Int1V.all}. This means that
10950 the compiler optimizer can "know" that the value @code{Int1V.all} is constant
10951 for all iterations of the loop and avoid the extra memory reference
10952 required to dereference it each time through the loop.
10954 This kind of optimization, called strict aliasing analysis, is
10955 triggered by specifying an optimization level of @option{-O2} or
10956 higher or @option{-Os} and allows @code{GNAT} to generate more efficient code
10957 when access values are involved.
10959 However, although this optimization is always correct in terms of
10960 the formal semantics of the Ada Reference Manual, difficulties can
10961 arise if features like @code{Unchecked_Conversion} are used to break
10962 the typing system. Consider the following complete program example:
10964 @smallexample @c ada
10967 type int1 is new integer;
10968 type int2 is new integer;
10969 type a1 is access int1;
10970 type a2 is access int2;
10975 function to_a2 (Input : a1) return a2;
10978 with Unchecked_Conversion;
10980 function to_a2 (Input : a1) return a2 is
10982 new Unchecked_Conversion (a1, a2);
10984 return to_a2u (Input);
10990 with Text_IO; use Text_IO;
10992 v1 : a1 := new int1;
10993 v2 : a2 := to_a2 (v1);
10997 put_line (int1'image (v1.all));
11003 This program prints out 0 in @option{-O0} or @option{-O1}
11004 mode, but it prints out 1 in @option{-O2} mode. That's
11005 because in strict aliasing mode, the compiler can and
11006 does assume that the assignment to @code{v2.all} could not
11007 affect the value of @code{v1.all}, since different types
11010 This behavior is not a case of non-conformance with the standard, since
11011 the Ada RM specifies that an unchecked conversion where the resulting
11012 bit pattern is not a correct value of the target type can result in an
11013 abnormal value and attempting to reference an abnormal value makes the
11014 execution of a program erroneous. That's the case here since the result
11015 does not point to an object of type @code{int2}. This means that the
11016 effect is entirely unpredictable.
11018 However, although that explanation may satisfy a language
11019 lawyer, in practice an applications programmer expects an
11020 unchecked conversion involving pointers to create true
11021 aliases and the behavior of printing 1 seems plain wrong.
11022 In this case, the strict aliasing optimization is unwelcome.
11024 Indeed the compiler recognizes this possibility, and the
11025 unchecked conversion generates a warning:
11028 p2.adb:5:07: warning: possible aliasing problem with type "a2"
11029 p2.adb:5:07: warning: use -fno-strict-aliasing switch for references
11030 p2.adb:5:07: warning: or use "pragma No_Strict_Aliasing (a2);"
11034 Unfortunately the problem is recognized when compiling the body of
11035 package @code{p2}, but the actual "bad" code is generated while
11036 compiling the body of @code{m} and this latter compilation does not see
11037 the suspicious @code{Unchecked_Conversion}.
11039 As implied by the warning message, there are approaches you can use to
11040 avoid the unwanted strict aliasing optimization in a case like this.
11042 One possibility is to simply avoid the use of @option{-O2}, but
11043 that is a bit drastic, since it throws away a number of useful
11044 optimizations that do not involve strict aliasing assumptions.
11046 A less drastic approach is to compile the program using the
11047 option @option{-fno-strict-aliasing}. Actually it is only the
11048 unit containing the dereferencing of the suspicious pointer
11049 that needs to be compiled. So in this case, if we compile
11050 unit @code{m} with this switch, then we get the expected
11051 value of zero printed. Analyzing which units might need
11052 the switch can be painful, so a more reasonable approach
11053 is to compile the entire program with options @option{-O2}
11054 and @option{-fno-strict-aliasing}. If the performance is
11055 satisfactory with this combination of options, then the
11056 advantage is that the entire issue of possible "wrong"
11057 optimization due to strict aliasing is avoided.
11059 To avoid the use of compiler switches, the configuration
11060 pragma @code{No_Strict_Aliasing} with no parameters may be
11061 used to specify that for all access types, the strict
11062 aliasing optimization should be suppressed.
11064 However, these approaches are still overkill, in that they causes
11065 all manipulations of all access values to be deoptimized. A more
11066 refined approach is to concentrate attention on the specific
11067 access type identified as problematic.
11069 First, if a careful analysis of uses of the pointer shows
11070 that there are no possible problematic references, then
11071 the warning can be suppressed by bracketing the
11072 instantiation of @code{Unchecked_Conversion} to turn
11075 @smallexample @c ada
11076 pragma Warnings (Off);
11078 new Unchecked_Conversion (a1, a2);
11079 pragma Warnings (On);
11083 Of course that approach is not appropriate for this particular
11084 example, since indeed there is a problematic reference. In this
11085 case we can take one of two other approaches.
11087 The first possibility is to move the instantiation of unchecked
11088 conversion to the unit in which the type is declared. In
11089 this example, we would move the instantiation of
11090 @code{Unchecked_Conversion} from the body of package
11091 @code{p2} to the spec of package @code{p1}. Now the
11092 warning disappears. That's because any use of the
11093 access type knows there is a suspicious unchecked
11094 conversion, and the strict aliasing optimization
11095 is automatically suppressed for the type.
11097 If it is not practical to move the unchecked conversion to the same unit
11098 in which the destination access type is declared (perhaps because the
11099 source type is not visible in that unit), you may use pragma
11100 @code{No_Strict_Aliasing} for the type. This pragma must occur in the
11101 same declarative sequence as the declaration of the access type:
11103 @smallexample @c ada
11104 type a2 is access int2;
11105 pragma No_Strict_Aliasing (a2);
11109 Here again, the compiler now knows that the strict aliasing optimization
11110 should be suppressed for any reference to type @code{a2} and the
11111 expected behavior is obtained.
11113 Finally, note that although the compiler can generate warnings for
11114 simple cases of unchecked conversions, there are tricker and more
11115 indirect ways of creating type incorrect aliases which the compiler
11116 cannot detect. Examples are the use of address overlays and unchecked
11117 conversions involving composite types containing access types as
11118 components. In such cases, no warnings are generated, but there can
11119 still be aliasing problems. One safe coding practice is to forbid the
11120 use of address clauses for type overlaying, and to allow unchecked
11121 conversion only for primitive types. This is not really a significant
11122 restriction since any possible desired effect can be achieved by
11123 unchecked conversion of access values.
11125 The aliasing analysis done in strict aliasing mode can certainly
11126 have significant benefits. We have seen cases of large scale
11127 application code where the time is increased by up to 5% by turning
11128 this optimization off. If you have code that includes significant
11129 usage of unchecked conversion, you might want to just stick with
11130 @option{-O1} and avoid the entire issue. If you get adequate
11131 performance at this level of optimization level, that's probably
11132 the safest approach. If tests show that you really need higher
11133 levels of optimization, then you can experiment with @option{-O2}
11134 and @option{-O2 -fno-strict-aliasing} to see how much effect this
11135 has on size and speed of the code. If you really need to use
11136 @option{-O2} with strict aliasing in effect, then you should
11137 review any uses of unchecked conversion of access types,
11138 particularly if you are getting the warnings described above.
11140 @node Aliased Variables and Optimization
11141 @subsection Aliased Variables and Optimization
11143 There are scenarios in which programs may
11144 use low level techniques to modify variables
11145 that otherwise might be considered to be unassigned. For example,
11146 a variable can be passed to a procedure by reference, which takes
11147 the address of the parameter and uses the address to modify the
11148 variable's value, even though it is passed as an IN parameter.
11149 Consider the following example:
11151 @smallexample @c ada
11153 Max_Length : constant Natural := 16;
11154 type Char_Ptr is access all Character;
11156 procedure Get_String(Buffer: Char_Ptr; Size : Integer);
11157 pragma Import (C, Get_String, "get_string");
11159 Name : aliased String (1 .. Max_Length) := (others => ' ');
11162 function Addr (S : String) return Char_Ptr is
11163 function To_Char_Ptr is
11164 new Ada.Unchecked_Conversion (System.Address, Char_Ptr);
11166 return To_Char_Ptr (S (S'First)'Address);
11170 Temp := Addr (Name);
11171 Get_String (Temp, Max_Length);
11176 where Get_String is a C function that uses the address in Temp to
11177 modify the variable @code{Name}. This code is dubious, and arguably
11178 erroneous, and the compiler would be entitled to assume that
11179 @code{Name} is never modified, and generate code accordingly.
11181 However, in practice, this would cause some existing code that
11182 seems to work with no optimization to start failing at high
11183 levels of optimzization.
11185 What the compiler does for such cases is to assume that marking
11186 a variable as aliased indicates that some "funny business" may
11187 be going on. The optimizer recognizes the aliased keyword and
11188 inhibits optimizations that assume the value cannot be assigned.
11189 This means that the above example will in fact "work" reliably,
11190 that is, it will produce the expected results.
11192 @node Atomic Variables and Optimization
11193 @subsection Atomic Variables and Optimization
11195 There are two considerations with regard to performance when
11196 atomic variables are used.
11198 First, the RM only guarantees that access to atomic variables
11199 be atomic, it has nothing to say about how this is achieved,
11200 though there is a strong implication that this should not be
11201 achieved by explicit locking code. Indeed GNAT will never
11202 generate any locking code for atomic variable access (it will
11203 simply reject any attempt to make a variable or type atomic
11204 if the atomic access cannot be achieved without such locking code).
11206 That being said, it is important to understand that you cannot
11207 assume that the entire variable will always be accessed. Consider
11210 @smallexample @c ada
11212 A,B,C,D : Character;
11215 for R'Alignment use 4;
11218 pragma Atomic (RV);
11225 You cannot assume that the reference to @code{RV.B}
11226 will read the entire 32-bit
11227 variable with a single load instruction. It is perfectly legitimate if
11228 the hardware allows it to do a byte read of just the B field. This read
11229 is still atomic, which is all the RM requires. GNAT can and does take
11230 advantage of this, depending on the architecture and optimization level.
11231 Any assumption to the contrary is non-portable and risky. Even if you
11232 examine the assembly language and see a full 32-bit load, this might
11233 change in a future version of the compiler.
11235 If your application requires that all accesses to @code{RV} in this
11236 example be full 32-bit loads, you need to make a copy for the access
11239 @smallexample @c ada
11241 RV_Copy : constant R := RV;
11249 Now the reference to RV must read the whole variable.
11250 Actually one can imagine some compiler which figures
11251 out that the whole copy is not required (because only
11252 the B field is actually accessed), but GNAT
11253 certainly won't do that, and we don't know of any
11254 compiler that would not handle this right, and the
11255 above code will in practice work portably across
11256 all architectures (that permit the Atomic declaration).
11258 The second issue with atomic variables has to do with
11259 the possible requirement of generating synchronization
11260 code. For more details on this, consult the sections on
11261 the pragmas Enable/Disable_Atomic_Synchronization in the
11262 GNAT Reference Manual. If performance is critical, and
11263 such synchronization code is not required, it may be
11264 useful to disable it.
11266 @node Passive Task Optimization
11267 @subsection Passive Task Optimization
11268 @cindex Passive Task
11270 A passive task is one which is sufficiently simple that
11271 in theory a compiler could recognize it an implement it
11272 efficiently without creating a new thread. The original design
11273 of Ada 83 had in mind this kind of passive task optimization, but
11274 only a few Ada 83 compilers attempted it. The problem was that
11275 it was difficult to determine the exact conditions under which
11276 the optimization was possible. The result is a very fragile
11277 optimization where a very minor change in the program can
11278 suddenly silently make a task non-optimizable.
11280 With the revisiting of this issue in Ada 95, there was general
11281 agreement that this approach was fundamentally flawed, and the
11282 notion of protected types was introduced. When using protected
11283 types, the restrictions are well defined, and you KNOW that the
11284 operations will be optimized, and furthermore this optimized
11285 performance is fully portable.
11287 Although it would theoretically be possible for GNAT to attempt to
11288 do this optimization, but it really doesn't make sense in the
11289 context of Ada 95, and none of the Ada 95 compilers implement
11290 this optimization as far as we know. In particular GNAT never
11291 attempts to perform this optimization.
11293 In any new Ada 95 code that is written, you should always
11294 use protected types in place of tasks that might be able to
11295 be optimized in this manner.
11296 Of course this does not help if you have legacy Ada 83 code
11297 that depends on this optimization, but it is unusual to encounter
11298 a case where the performance gains from this optimization
11301 Your program should work correctly without this optimization. If
11302 you have performance problems, then the most practical
11303 approach is to figure out exactly where these performance problems
11304 arise, and update those particular tasks to be protected types. Note
11305 that typically clients of the tasks who call entries, will not have
11306 to be modified, only the task definition itself.
11309 @node Coverage Analysis
11310 @subsection Coverage Analysis
11313 GNAT supports the HP Performance Coverage Analyzer (PCA), which allows
11314 the user to determine the distribution of execution time across a program,
11315 @pxref{Profiling} for details of usage.
11319 @node Text_IO Suggestions
11320 @section @code{Text_IO} Suggestions
11321 @cindex @code{Text_IO} and performance
11324 The @code{Ada.Text_IO} package has fairly high overheads due in part to
11325 the requirement of maintaining page and line counts. If performance
11326 is critical, a recommendation is to use @code{Stream_IO} instead of
11327 @code{Text_IO} for volume output, since this package has less overhead.
11329 If @code{Text_IO} must be used, note that by default output to the standard
11330 output and standard error files is unbuffered (this provides better
11331 behavior when output statements are used for debugging, or if the
11332 progress of a program is observed by tracking the output, e.g. by
11333 using the Unix @command{tail -f} command to watch redirected output.
11335 If you are generating large volumes of output with @code{Text_IO} and
11336 performance is an important factor, use a designated file instead
11337 of the standard output file, or change the standard output file to
11338 be buffered using @code{Interfaces.C_Streams.setvbuf}.
11341 @ifclear FSFEDITION
11342 @node Reducing Size of Ada Executables with gnatelim
11343 @section Reducing Size of Ada Executables with @code{gnatelim}
11347 This section describes @command{gnatelim}, a tool which detects unused
11348 subprograms and helps the compiler to create a smaller executable for your
11353 * Running gnatelim::
11354 * Processing Precompiled Libraries::
11355 * Correcting the List of Eliminate Pragmas::
11356 * Making Your Executables Smaller::
11357 * Summary of the gnatelim Usage Cycle::
11360 @node About gnatelim
11361 @subsection About @code{gnatelim}
11364 When a program shares a set of Ada
11365 packages with other programs, it may happen that this program uses
11366 only a fraction of the subprograms defined in these packages. The code
11367 created for these unused subprograms increases the size of the executable.
11369 @code{gnatelim} tracks unused subprograms in an Ada program and
11370 outputs a list of GNAT-specific pragmas @code{Eliminate} marking all the
11371 subprograms that are declared but never called. By placing the list of
11372 @code{Eliminate} pragmas in the GNAT configuration file @file{gnat.adc} and
11373 recompiling your program, you may decrease the size of its executable,
11374 because the compiler will not generate the code for 'eliminated' subprograms.
11375 @xref{Pragma Eliminate,,, gnat_rm, GNAT Reference Manual}, for more
11376 information about this pragma.
11378 @code{gnatelim} needs as its input data the name of the main subprogram.
11380 If a set of source files is specified as @code{gnatelim} arguments, it
11381 treats these files as a complete set of sources making up a program to
11382 analyse, and analyses only these sources.
11384 After a full successful build of the main subprogram @code{gnatelim} can be
11385 called without specifying sources to analyse, in this case it computes
11386 the source closure of the main unit from the @file{ALI} files.
11388 The following command will create the set of @file{ALI} files needed for
11392 $ gnatmake ^-c Main_Prog^/ACTIONS=COMPILE MAIN_PROG^
11395 Note that @code{gnatelim} does not need object files.
11397 @node Running gnatelim
11398 @subsection Running @code{gnatelim}
11401 @code{gnatelim} has the following command-line interface:
11404 $ gnatelim [@var{switches}] ^-main^?MAIN^=@var{main_unit_name} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
11408 @var{main_unit_name} should be a name of a source file that contains the main
11409 subprogram of a program (partition).
11411 Each @var{filename} is the name (including the extension) of a source
11412 file to process. ``Wildcards'' are allowed, and
11413 the file name may contain path information.
11415 @samp{@var{gcc_switches}} is a list of switches for
11416 @command{gcc}. They will be passed on to all compiler invocations made by
11417 @command{gnatelim} to generate the ASIS trees. Here you can provide
11418 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
11419 use the @option{-gnatec} switch to set the configuration file,
11420 use the @option{-gnat05} switch if sources should be compiled in
11423 @code{gnatelim} has the following switches:
11428 @cindex @option{--version} @command{gnatelim}
11429 Display Copyright and version, then exit disregarding all other options.
11432 @cindex @option{--help} @command{gnatelim}
11433 Display usage, then exit disregarding all other options.
11435 @item -P @var{file}
11436 @cindex @option{-P} @command{gnatelim}
11437 Indicates the name of the project file that describes the set of sources
11440 @item -X@var{name}=@var{value}
11441 @cindex @option{-X} @command{gnatelim}
11442 Indicates that external variable @var{name} in the argument project
11443 has the value @var{value}. Has no effect if no project is specified as
11446 @item ^-files^/FILES^=@var{filename}
11447 @cindex @option{^-files^/FILES^} (@code{gnatelim})
11448 Take the argument source files from the specified file. This file should be an
11449 ordinary text file containing file names separated by spaces or
11450 line breaks. You can use this switch more than once in the same call to
11451 @command{gnatelim}. You also can combine this switch with
11452 an explicit list of files.
11455 @cindex @option{^-log^/LOG^} (@command{gnatelim})
11456 Duplicate all the output sent to @file{stderr} into a log file. The log file
11457 is named @file{gnatelim.log} and is located in the current directory.
11460 @item ^-log^/LOGFILE^=@var{filename}
11461 @cindex @option{^-log^/LOGFILE^} (@command{gnatelim})
11462 Duplicate all the output sent to @file{stderr} into a specified log file.
11465 @cindex @option{^--no-elim-dispatch^/NO_DISPATCH^} (@command{gnatelim})
11466 @item ^--no-elim-dispatch^/NO_DISPATCH^
11467 Do not generate pragmas for dispatching operations.
11469 @item ^--ignore^/IGNORE^=@var{filename}
11470 @cindex @option{^--ignore^/IGNORE^} (@command{gnatelim})
11471 Do not generate pragmas for subprograms declared in the sources
11472 listed in a specified file
11474 @cindex @option{^-o^/OUTPUT^} (@command{gnatelim})
11475 @item ^-o^/OUTPUT^=@var{report_file}
11476 Put @command{gnatelim} output into a specified file. If this file already exists,
11477 it is overridden. If this switch is not used, @command{gnatelim} outputs its results
11480 @item ^-j^/PROCESSES=^@var{n}
11481 @cindex @option{^-j^/PROCESSES^} (@command{gnatelim})
11482 Use @var{n} processes to carry out the tree creations (internal representations
11483 of the argument sources). On a multiprocessor machine this speeds up processing
11484 of big sets of argument sources. If @var{n} is 0, then the maximum number of
11485 parallel tree creations is the number of core processors on the platform.
11488 @cindex @option{^-q^/QUIET^} (@command{gnatelim})
11489 Quiet mode: by default @code{gnatelim} outputs to the standard error
11490 stream the number of program units left to be processed. This option turns
11493 @cindex @option{^-t^/TIME^} (@command{gnatelim})
11495 Print out execution time.
11497 @item ^-v^/VERBOSE^
11498 @cindex @option{^-v^/VERBOSE^} (@command{gnatelim})
11499 Verbose mode: @code{gnatelim} version information is printed as Ada
11500 comments to the standard output stream. Also, in addition to the number of
11501 program units left @code{gnatelim} will output the name of the current unit
11504 @item ^-wq^/WARNINGS=QUIET^
11505 @cindex @option{^-wq^/WARNINGS=QUIET^} (@command{gnatelim})
11506 Quiet warning mode - some warnings are suppressed. In particular warnings that
11507 indicate that the analysed set of sources is incomplete to make up a
11508 partition and that some subprogram bodies are missing are not generated.
11512 Note: to invoke @command{gnatelim} with a project file, use the @code{gnat}
11513 driver (see @ref{The GNAT Driver and Project Files}).
11515 @node Processing Precompiled Libraries
11516 @subsection Processing Precompiled Libraries
11519 If some program uses a precompiled Ada library, it can be processed by
11520 @code{gnatelim} in a usual way. @code{gnatelim} will newer generate an
11521 Eliminate pragma for a subprogram if the body of this subprogram has not
11522 been analysed, this is a typical case for subprograms from precompiled
11523 libraries. Switch @option{^-wq^/WARNINGS=QUIET^} may be used to suppress
11524 warnings about missing source files and non-analyzed subprogram bodies
11525 that can be generated when processing precompiled Ada libraries.
11527 @node Correcting the List of Eliminate Pragmas
11528 @subsection Correcting the List of Eliminate Pragmas
11531 In some rare cases @code{gnatelim} may try to eliminate
11532 subprograms that are actually called in the program. In this case, the
11533 compiler will generate an error message of the form:
11536 main.adb:4:08: cannot reference subprogram "P" eliminated at elim.out:5
11540 You will need to manually remove the wrong @code{Eliminate} pragmas from
11541 the configuration file indicated in the error message. You should recompile
11542 your program from scratch after that, because you need a consistent
11543 configuration file(s) during the entire compilation.
11545 @node Making Your Executables Smaller
11546 @subsection Making Your Executables Smaller
11549 In order to get a smaller executable for your program you now have to
11550 recompile the program completely with the configuration file containing
11551 pragmas Eliminate generated by gnatelim. If these pragmas are placed in
11552 @file{gnat.adc} file located in your current directory, just do:
11555 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
11559 (Use the @option{^-f^/FORCE_COMPILE^} option for @command{gnatmake} to
11560 recompile everything
11561 with the set of pragmas @code{Eliminate} that you have obtained with
11562 @command{gnatelim}).
11564 Be aware that the set of @code{Eliminate} pragmas is specific to each
11565 program. It is not recommended to merge sets of @code{Eliminate}
11566 pragmas created for different programs in one configuration file.
11568 @node Summary of the gnatelim Usage Cycle
11569 @subsection Summary of the @code{gnatelim} Usage Cycle
11572 Here is a quick summary of the steps to be taken in order to reduce
11573 the size of your executables with @code{gnatelim}. You may use
11574 other GNAT options to control the optimization level,
11575 to produce the debugging information, to set search path, etc.
11579 Create a complete set of @file{ALI} files (if the program has not been
11583 $ gnatmake ^-c main_prog^/ACTIONS=COMPILE MAIN_PROG^
11587 Generate a list of @code{Eliminate} pragmas in default configuration file
11588 @file{gnat.adc} in the current directory
11591 $ PIPE GNAT ELIM MAIN_PROG > GNAT.ADC
11594 $ gnatelim main_prog >@r{[}>@r{]} gnat.adc
11599 Recompile the application
11602 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
11608 @node Reducing Size of Executables with unused subprogram/data elimination
11609 @section Reducing Size of Executables with Unused Subprogram/Data Elimination
11610 @findex unused subprogram/data elimination
11613 This section describes how you can eliminate unused subprograms and data from
11614 your executable just by setting options at compilation time.
11617 * About unused subprogram/data elimination::
11618 * Compilation options::
11619 * Example of unused subprogram/data elimination::
11622 @node About unused subprogram/data elimination
11623 @subsection About unused subprogram/data elimination
11626 By default, an executable contains all code and data of its composing objects
11627 (directly linked or coming from statically linked libraries), even data or code
11628 never used by this executable.
11630 This feature will allow you to eliminate such unused code from your
11631 executable, making it smaller (in disk and in memory).
11633 This functionality is available on all Linux platforms except for the IA-64
11634 architecture and on all cross platforms using the ELF binary file format.
11635 In both cases GNU binutils version 2.16 or later are required to enable it.
11637 @node Compilation options
11638 @subsection Compilation options
11641 The operation of eliminating the unused code and data from the final executable
11642 is directly performed by the linker.
11644 In order to do this, it has to work with objects compiled with the
11646 @option{-ffunction-sections} @option{-fdata-sections}.
11647 @cindex @option{-ffunction-sections} (@command{gcc})
11648 @cindex @option{-fdata-sections} (@command{gcc})
11649 These options are usable with C and Ada files.
11650 They will place respectively each
11651 function or data in a separate section in the resulting object file.
11653 Once the objects and static libraries are created with these options, the
11654 linker can perform the dead code elimination. You can do this by setting
11655 the @option{-Wl,--gc-sections} option to gcc command or in the
11656 @option{-largs} section of @command{gnatmake}. This will perform a
11657 garbage collection of code and data never referenced.
11659 If the linker performs a partial link (@option{-r} linker option), then you
11660 will need to provide the entry point using the @option{-e} / @option{--entry}
11663 Note that objects compiled without the @option{-ffunction-sections} and
11664 @option{-fdata-sections} options can still be linked with the executable.
11665 However, no dead code elimination will be performed on those objects (they will
11668 The GNAT static library is now compiled with -ffunction-sections and
11669 -fdata-sections on some platforms. This allows you to eliminate the unused code
11670 and data of the GNAT library from your executable.
11672 @node Example of unused subprogram/data elimination
11673 @subsection Example of unused subprogram/data elimination
11676 Here is a simple example:
11678 @smallexample @c ada
11687 Used_Data : Integer;
11688 Unused_Data : Integer;
11690 procedure Used (Data : Integer);
11691 procedure Unused (Data : Integer);
11694 package body Aux is
11695 procedure Used (Data : Integer) is
11700 procedure Unused (Data : Integer) is
11702 Unused_Data := Data;
11708 @code{Unused} and @code{Unused_Data} are never referenced in this code
11709 excerpt, and hence they may be safely removed from the final executable.
11714 $ nm test | grep used
11715 020015f0 T aux__unused
11716 02005d88 B aux__unused_data
11717 020015cc T aux__used
11718 02005d84 B aux__used_data
11720 $ gnatmake test -cargs -fdata-sections -ffunction-sections \
11721 -largs -Wl,--gc-sections
11723 $ nm test | grep used
11724 02005350 T aux__used
11725 0201ffe0 B aux__used_data
11729 It can be observed that the procedure @code{Unused} and the object
11730 @code{Unused_Data} are removed by the linker when using the
11731 appropriate options.
11733 @c ********************************
11734 @node Renaming Files with gnatchop
11735 @chapter Renaming Files with @code{gnatchop}
11739 This chapter discusses how to handle files with multiple units by using
11740 the @code{gnatchop} utility. This utility is also useful in renaming
11741 files to meet the standard GNAT default file naming conventions.
11744 * Handling Files with Multiple Units::
11745 * Operating gnatchop in Compilation Mode::
11746 * Command Line for gnatchop::
11747 * Switches for gnatchop::
11748 * Examples of gnatchop Usage::
11751 @node Handling Files with Multiple Units
11752 @section Handling Files with Multiple Units
11755 The basic compilation model of GNAT requires that a file submitted to the
11756 compiler have only one unit and there be a strict correspondence
11757 between the file name and the unit name.
11759 The @code{gnatchop} utility allows both of these rules to be relaxed,
11760 allowing GNAT to process files which contain multiple compilation units
11761 and files with arbitrary file names. @code{gnatchop}
11762 reads the specified file and generates one or more output files,
11763 containing one unit per file. The unit and the file name correspond,
11764 as required by GNAT.
11766 If you want to permanently restructure a set of ``foreign'' files so that
11767 they match the GNAT rules, and do the remaining development using the
11768 GNAT structure, you can simply use @command{gnatchop} once, generate the
11769 new set of files and work with them from that point on.
11771 Alternatively, if you want to keep your files in the ``foreign'' format,
11772 perhaps to maintain compatibility with some other Ada compilation
11773 system, you can set up a procedure where you use @command{gnatchop} each
11774 time you compile, regarding the source files that it writes as temporary
11775 files that you throw away.
11777 Note that if your file containing multiple units starts with a byte order
11778 mark (BOM) specifying UTF-8 encoding, then the files generated by gnatchop
11779 will each start with a copy of this BOM, meaning that they can be compiled
11780 automatically in UTF-8 mode without needing to specify an explicit encoding.
11782 @node Operating gnatchop in Compilation Mode
11783 @section Operating gnatchop in Compilation Mode
11786 The basic function of @code{gnatchop} is to take a file with multiple units
11787 and split it into separate files. The boundary between files is reasonably
11788 clear, except for the issue of comments and pragmas. In default mode, the
11789 rule is that any pragmas between units belong to the previous unit, except
11790 that configuration pragmas always belong to the following unit. Any comments
11791 belong to the following unit. These rules
11792 almost always result in the right choice of
11793 the split point without needing to mark it explicitly and most users will
11794 find this default to be what they want. In this default mode it is incorrect to
11795 submit a file containing only configuration pragmas, or one that ends in
11796 configuration pragmas, to @code{gnatchop}.
11798 However, using a special option to activate ``compilation mode'',
11800 can perform another function, which is to provide exactly the semantics
11801 required by the RM for handling of configuration pragmas in a compilation.
11802 In the absence of configuration pragmas (at the main file level), this
11803 option has no effect, but it causes such configuration pragmas to be handled
11804 in a quite different manner.
11806 First, in compilation mode, if @code{gnatchop} is given a file that consists of
11807 only configuration pragmas, then this file is appended to the
11808 @file{gnat.adc} file in the current directory. This behavior provides
11809 the required behavior described in the RM for the actions to be taken
11810 on submitting such a file to the compiler, namely that these pragmas
11811 should apply to all subsequent compilations in the same compilation
11812 environment. Using GNAT, the current directory, possibly containing a
11813 @file{gnat.adc} file is the representation
11814 of a compilation environment. For more information on the
11815 @file{gnat.adc} file, see @ref{Handling of Configuration Pragmas}.
11817 Second, in compilation mode, if @code{gnatchop}
11818 is given a file that starts with
11819 configuration pragmas, and contains one or more units, then these
11820 configuration pragmas are prepended to each of the chopped files. This
11821 behavior provides the required behavior described in the RM for the
11822 actions to be taken on compiling such a file, namely that the pragmas
11823 apply to all units in the compilation, but not to subsequently compiled
11826 Finally, if configuration pragmas appear between units, they are appended
11827 to the previous unit. This results in the previous unit being illegal,
11828 since the compiler does not accept configuration pragmas that follow
11829 a unit. This provides the required RM behavior that forbids configuration
11830 pragmas other than those preceding the first compilation unit of a
11833 For most purposes, @code{gnatchop} will be used in default mode. The
11834 compilation mode described above is used only if you need exactly
11835 accurate behavior with respect to compilations, and you have files
11836 that contain multiple units and configuration pragmas. In this
11837 circumstance the use of @code{gnatchop} with the compilation mode
11838 switch provides the required behavior, and is for example the mode
11839 in which GNAT processes the ACVC tests.
11841 @node Command Line for gnatchop
11842 @section Command Line for @code{gnatchop}
11845 The @code{gnatchop} command has the form:
11848 @c $ gnatchop switches @var{file name} @r{[}@var{file name} @dots{}@r{]}
11849 @c @ovar{directory}
11850 @c Expanding @ovar macro inline (explanation in macro def comments)
11851 $ gnatchop switches @var{file name} @r{[}@var{file name} @dots{}@r{]}
11852 @r{[}@var{directory}@r{]}
11856 The only required argument is the file name of the file to be chopped.
11857 There are no restrictions on the form of this file name. The file itself
11858 contains one or more Ada units, in normal GNAT format, concatenated
11859 together. As shown, more than one file may be presented to be chopped.
11861 When run in default mode, @code{gnatchop} generates one output file in
11862 the current directory for each unit in each of the files.
11864 @var{directory}, if specified, gives the name of the directory to which
11865 the output files will be written. If it is not specified, all files are
11866 written to the current directory.
11868 For example, given a
11869 file called @file{hellofiles} containing
11871 @smallexample @c ada
11876 with Text_IO; use Text_IO;
11879 Put_Line ("Hello");
11889 $ gnatchop ^hellofiles^HELLOFILES.^
11893 generates two files in the current directory, one called
11894 @file{hello.ads} containing the single line that is the procedure spec,
11895 and the other called @file{hello.adb} containing the remaining text. The
11896 original file is not affected. The generated files can be compiled in
11900 When gnatchop is invoked on a file that is empty or that contains only empty
11901 lines and/or comments, gnatchop will not fail, but will not produce any
11904 For example, given a
11905 file called @file{toto.txt} containing
11907 @smallexample @c ada
11919 $ gnatchop ^toto.txt^TOT.TXT^
11923 will not produce any new file and will result in the following warnings:
11926 toto.txt:1:01: warning: empty file, contains no compilation units
11927 no compilation units found
11928 no source files written
11931 @node Switches for gnatchop
11932 @section Switches for @code{gnatchop}
11935 @command{gnatchop} recognizes the following switches:
11941 @cindex @option{--version} @command{gnatchop}
11942 Display Copyright and version, then exit disregarding all other options.
11945 @cindex @option{--help} @command{gnatchop}
11946 If @option{--version} was not used, display usage, then exit disregarding
11949 @item ^-c^/COMPILATION^
11950 @cindex @option{^-c^/COMPILATION^} (@code{gnatchop})
11951 Causes @code{gnatchop} to operate in compilation mode, in which
11952 configuration pragmas are handled according to strict RM rules. See
11953 previous section for a full description of this mode.
11956 @item -gnat@var{xxx}
11957 This passes the given @option{-gnat@var{xxx}} switch to @code{gnat} which is
11958 used to parse the given file. Not all @var{xxx} options make sense,
11959 but for example, the use of @option{-gnati2} allows @code{gnatchop} to
11960 process a source file that uses Latin-2 coding for identifiers.
11964 Causes @code{gnatchop} to generate a brief help summary to the standard
11965 output file showing usage information.
11967 @item ^-k@var{mm}^/FILE_NAME_MAX_LENGTH=@var{mm}^
11968 @cindex @option{^-k^/FILE_NAME_MAX_LENGTH^} (@code{gnatchop})
11969 Limit generated file names to the specified number @code{mm}
11971 This is useful if the
11972 resulting set of files is required to be interoperable with systems
11973 which limit the length of file names.
11975 If no value is given, or
11976 if no @code{/FILE_NAME_MAX_LENGTH} qualifier is given,
11977 a default of 39, suitable for OpenVMS Alpha
11978 Systems, is assumed
11981 No space is allowed between the @option{-k} and the numeric value. The numeric
11982 value may be omitted in which case a default of @option{-k8},
11984 with DOS-like file systems, is used. If no @option{-k} switch
11986 there is no limit on the length of file names.
11989 @item ^-p^/PRESERVE^
11990 @cindex @option{^-p^/PRESERVE^} (@code{gnatchop})
11991 Causes the file ^modification^creation^ time stamp of the input file to be
11992 preserved and used for the time stamp of the output file(s). This may be
11993 useful for preserving coherency of time stamps in an environment where
11994 @code{gnatchop} is used as part of a standard build process.
11997 @cindex @option{^-q^/QUIET^} (@code{gnatchop})
11998 Causes output of informational messages indicating the set of generated
11999 files to be suppressed. Warnings and error messages are unaffected.
12001 @item ^-r^/REFERENCE^
12002 @cindex @option{^-r^/REFERENCE^} (@code{gnatchop})
12003 @findex Source_Reference
12004 Generate @code{Source_Reference} pragmas. Use this switch if the output
12005 files are regarded as temporary and development is to be done in terms
12006 of the original unchopped file. This switch causes
12007 @code{Source_Reference} pragmas to be inserted into each of the
12008 generated files to refers back to the original file name and line number.
12009 The result is that all error messages refer back to the original
12011 In addition, the debugging information placed into the object file (when
12012 the @option{^-g^/DEBUG^} switch of @command{gcc} or @command{gnatmake} is
12014 also refers back to this original file so that tools like profilers and
12015 debuggers will give information in terms of the original unchopped file.
12017 If the original file to be chopped itself contains
12018 a @code{Source_Reference}
12019 pragma referencing a third file, then gnatchop respects
12020 this pragma, and the generated @code{Source_Reference} pragmas
12021 in the chopped file refer to the original file, with appropriate
12022 line numbers. This is particularly useful when @code{gnatchop}
12023 is used in conjunction with @code{gnatprep} to compile files that
12024 contain preprocessing statements and multiple units.
12026 @item ^-v^/VERBOSE^
12027 @cindex @option{^-v^/VERBOSE^} (@code{gnatchop})
12028 Causes @code{gnatchop} to operate in verbose mode. The version
12029 number and copyright notice are output, as well as exact copies of
12030 the gnat1 commands spawned to obtain the chop control information.
12032 @item ^-w^/OVERWRITE^
12033 @cindex @option{^-w^/OVERWRITE^} (@code{gnatchop})
12034 Overwrite existing file names. Normally @code{gnatchop} regards it as a
12035 fatal error if there is already a file with the same name as a
12036 file it would otherwise output, in other words if the files to be
12037 chopped contain duplicated units. This switch bypasses this
12038 check, and causes all but the last instance of such duplicated
12039 units to be skipped.
12042 @item --GCC=@var{xxxx}
12043 @cindex @option{--GCC=} (@code{gnatchop})
12044 Specify the path of the GNAT parser to be used. When this switch is used,
12045 no attempt is made to add the prefix to the GNAT parser executable.
12049 @node Examples of gnatchop Usage
12050 @section Examples of @code{gnatchop} Usage
12054 @item gnatchop /OVERWRITE HELLO_S.ADA [PRERELEASE.FILES]
12057 @item gnatchop -w hello_s.ada prerelease/files
12060 Chops the source file @file{hello_s.ada}. The output files will be
12061 placed in the directory @file{^prerelease/files^[PRERELEASE.FILES]^},
12063 files with matching names in that directory (no files in the current
12064 directory are modified).
12066 @item gnatchop ^archive^ARCHIVE.^
12067 Chops the source file @file{^archive^ARCHIVE.^}
12068 into the current directory. One
12069 useful application of @code{gnatchop} is in sending sets of sources
12070 around, for example in email messages. The required sources are simply
12071 concatenated (for example, using a ^Unix @code{cat}^VMS @code{APPEND/NEW}^
12073 @command{gnatchop} is used at the other end to reconstitute the original
12076 @item gnatchop file1 file2 file3 direc
12077 Chops all units in files @file{file1}, @file{file2}, @file{file3}, placing
12078 the resulting files in the directory @file{direc}. Note that if any units
12079 occur more than once anywhere within this set of files, an error message
12080 is generated, and no files are written. To override this check, use the
12081 @option{^-w^/OVERWRITE^} switch,
12082 in which case the last occurrence in the last file will
12083 be the one that is output, and earlier duplicate occurrences for a given
12084 unit will be skipped.
12087 @node Configuration Pragmas
12088 @chapter Configuration Pragmas
12089 @cindex Configuration pragmas
12090 @cindex Pragmas, configuration
12093 * Handling of Configuration Pragmas::
12094 * The Configuration Pragmas Files::
12098 Configuration pragmas include those pragmas described as
12099 such in the Ada Reference Manual, as well as
12100 implementation-dependent pragmas that are configuration pragmas.
12101 @xref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
12102 for details on these additional GNAT-specific configuration pragmas.
12103 Most notably, the pragma @code{Source_File_Name}, which allows
12104 specifying non-default names for source files, is a configuration
12105 pragma. The following is a complete list of configuration pragmas
12106 recognized by GNAT:
12115 Allow_Integer_Address
12118 Assume_No_Invalid_Values
12123 Compile_Time_Warning
12125 Component_Alignment
12126 Convention_Identifier
12129 Default_Storage_Pool
12135 External_Name_Casing
12138 Float_Representation
12151 Priority_Specific_Dispatching
12154 Propagate_Exceptions
12157 Restricted_Run_Time
12159 Restrictions_Warnings
12161 Short_Circuit_And_Or
12163 Source_File_Name_Project
12167 Suppress_Exception_Locations
12168 Task_Dispatching_Policy
12174 Wide_Character_Encoding
12177 @node Handling of Configuration Pragmas
12178 @section Handling of Configuration Pragmas
12180 Configuration pragmas may either appear at the start of a compilation
12181 unit, or they can appear in a configuration pragma file to apply to
12182 all compilations performed in a given compilation environment.
12184 GNAT also provides the @code{gnatchop} utility to provide an automatic
12185 way to handle configuration pragmas following the semantics for
12186 compilations (that is, files with multiple units), described in the RM.
12187 See @ref{Operating gnatchop in Compilation Mode} for details.
12188 However, for most purposes, it will be more convenient to edit the
12189 @file{gnat.adc} file that contains configuration pragmas directly,
12190 as described in the following section.
12192 In the case of @code{Restrictions} pragmas appearing as configuration
12193 pragmas in individual compilation units, the exact handling depends on
12194 the type of restriction.
12196 Restrictions that require partition-wide consistency (like
12197 @code{No_Tasking}) are
12198 recognized wherever they appear
12199 and can be freely inherited, e.g. from a with'ed unit to the with'ing
12200 unit. This makes sense since the binder will in any case insist on seeing
12201 consistent use, so any unit not conforming to any restrictions that are
12202 anywhere in the partition will be rejected, and you might as well find
12203 that out at compile time rather than at bind time.
12205 For restrictions that do not require partition-wide consistency, e.g.
12206 SPARK or No_Implementation_Attributes, in general the restriction applies
12207 only to the unit in which the pragma appears, and not to any other units.
12209 The exception is No_Elaboration_Code which always applies to the entire
12210 object file from a compilation, i.e. to the body, spec, and all subunits.
12211 This restriction can be specified in a configuration pragma file, or it
12212 can be on the body and/or the spec (in eithe case it applies to all the
12213 relevant units). It can appear on a subunit only if it has previously
12214 appeared in the body of spec.
12216 @node The Configuration Pragmas Files
12217 @section The Configuration Pragmas Files
12218 @cindex @file{gnat.adc}
12221 In GNAT a compilation environment is defined by the current
12222 directory at the time that a compile command is given. This current
12223 directory is searched for a file whose name is @file{gnat.adc}. If
12224 this file is present, it is expected to contain one or more
12225 configuration pragmas that will be applied to the current compilation.
12226 However, if the switch @option{-gnatA} is used, @file{gnat.adc} is not
12229 Configuration pragmas may be entered into the @file{gnat.adc} file
12230 either by running @code{gnatchop} on a source file that consists only of
12231 configuration pragmas, or more conveniently by
12232 direct editing of the @file{gnat.adc} file, which is a standard format
12235 In addition to @file{gnat.adc}, additional files containing configuration
12236 pragmas may be applied to the current compilation using the switch
12237 @option{-gnatec}@var{path}. @var{path} must designate an existing file that
12238 contains only configuration pragmas. These configuration pragmas are
12239 in addition to those found in @file{gnat.adc} (provided @file{gnat.adc}
12240 is present and switch @option{-gnatA} is not used).
12242 It is allowed to specify several switches @option{-gnatec}, all of which
12243 will be taken into account.
12245 If you are using project file, a separate mechanism is provided using
12246 project attributes, see @ref{Specifying Configuration Pragmas} for more
12250 Of special interest to GNAT OpenVMS Alpha is the following
12251 configuration pragma:
12253 @smallexample @c ada
12255 pragma Extend_System (Aux_DEC);
12260 In the presence of this pragma, GNAT adds to the definition of the
12261 predefined package SYSTEM all the additional types and subprograms that are
12262 defined in HP Ada. See @ref{Compatibility with HP Ada} for details.
12265 @node Handling Arbitrary File Naming Conventions with gnatname
12266 @chapter Handling Arbitrary File Naming Conventions with @code{gnatname}
12267 @cindex Arbitrary File Naming Conventions
12270 * Arbitrary File Naming Conventions::
12271 * Running gnatname::
12272 * Switches for gnatname::
12273 * Examples of gnatname Usage::
12276 @node Arbitrary File Naming Conventions
12277 @section Arbitrary File Naming Conventions
12280 The GNAT compiler must be able to know the source file name of a compilation
12281 unit. When using the standard GNAT default file naming conventions
12282 (@code{.ads} for specs, @code{.adb} for bodies), the GNAT compiler
12283 does not need additional information.
12286 When the source file names do not follow the standard GNAT default file naming
12287 conventions, the GNAT compiler must be given additional information through
12288 a configuration pragmas file (@pxref{Configuration Pragmas})
12290 When the non-standard file naming conventions are well-defined,
12291 a small number of pragmas @code{Source_File_Name} specifying a naming pattern
12292 (@pxref{Alternative File Naming Schemes}) may be sufficient. However,
12293 if the file naming conventions are irregular or arbitrary, a number
12294 of pragma @code{Source_File_Name} for individual compilation units
12296 To help maintain the correspondence between compilation unit names and
12297 source file names within the compiler,
12298 GNAT provides a tool @code{gnatname} to generate the required pragmas for a
12301 @node Running gnatname
12302 @section Running @code{gnatname}
12305 The usual form of the @code{gnatname} command is
12308 @c $ gnatname @ovar{switches} @var{naming_pattern} @ovar{naming_patterns}
12309 @c @r{[}--and @ovar{switches} @var{naming_pattern} @ovar{naming_patterns}@r{]}
12310 @c Expanding @ovar macro inline (explanation in macro def comments)
12311 $ gnatname @r{[}@var{switches}@r{]} @var{naming_pattern} @r{[}@var{naming_patterns}@r{]}
12312 @r{[}--and @r{[}@var{switches}@r{]} @var{naming_pattern} @r{[}@var{naming_patterns}@r{]}@r{]}
12316 All of the arguments are optional. If invoked without any argument,
12317 @code{gnatname} will display its usage.
12320 When used with at least one naming pattern, @code{gnatname} will attempt to
12321 find all the compilation units in files that follow at least one of the
12322 naming patterns. To find these compilation units,
12323 @code{gnatname} will use the GNAT compiler in syntax-check-only mode on all
12327 One or several Naming Patterns may be given as arguments to @code{gnatname}.
12328 Each Naming Pattern is enclosed between double quotes (or single
12329 quotes on Windows).
12330 A Naming Pattern is a regular expression similar to the wildcard patterns
12331 used in file names by the Unix shells or the DOS prompt.
12334 @code{gnatname} may be called with several sections of directories/patterns.
12335 Sections are separated by switch @code{--and}. In each section, there must be
12336 at least one pattern. If no directory is specified in a section, the current
12337 directory (or the project directory is @code{-P} is used) is implied.
12338 The options other that the directory switches and the patterns apply globally
12339 even if they are in different sections.
12342 Examples of Naming Patterns are
12351 For a more complete description of the syntax of Naming Patterns,
12352 see the second kind of regular expressions described in @file{g-regexp.ads}
12353 (the ``Glob'' regular expressions).
12356 When invoked with no switch @code{-P}, @code{gnatname} will create a
12357 configuration pragmas file @file{gnat.adc} in the current working directory,
12358 with pragmas @code{Source_File_Name} for each file that contains a valid Ada
12361 @node Switches for gnatname
12362 @section Switches for @code{gnatname}
12365 Switches for @code{gnatname} must precede any specified Naming Pattern.
12368 You may specify any of the following switches to @code{gnatname}:
12374 @cindex @option{--version} @command{gnatname}
12375 Display Copyright and version, then exit disregarding all other options.
12378 @cindex @option{--help} @command{gnatname}
12379 If @option{--version} was not used, display usage, then exit disregarding
12382 @item --subdirs=<dir>
12383 Real object, library or exec directories are subdirectories <dir> of the
12387 Do not create a backup copy of an existing project file.
12390 Start another section of directories/patterns.
12392 @item ^-c^/CONFIG_FILE=^@file{file}
12393 @cindex @option{^-c^/CONFIG_FILE^} (@code{gnatname})
12394 Create a configuration pragmas file @file{file} (instead of the default
12397 There may be zero, one or more space between @option{-c} and
12400 @file{file} may include directory information. @file{file} must be
12401 writable. There may be only one switch @option{^-c^/CONFIG_FILE^}.
12402 When a switch @option{^-c^/CONFIG_FILE^} is
12403 specified, no switch @option{^-P^/PROJECT_FILE^} may be specified (see below).
12405 @item ^-d^/SOURCE_DIRS=^@file{dir}
12406 @cindex @option{^-d^/SOURCE_DIRS^} (@code{gnatname})
12407 Look for source files in directory @file{dir}. There may be zero, one or more
12408 spaces between @option{^-d^/SOURCE_DIRS=^} and @file{dir}.
12409 @file{dir} may end with @code{/**}, that is it may be of the form
12410 @code{root_dir/**}. In this case, the directory @code{root_dir} and all of its
12411 subdirectories, recursively, have to be searched for sources.
12412 When a switch @option{^-d^/SOURCE_DIRS^}
12413 is specified, the current working directory will not be searched for source
12414 files, unless it is explicitly specified with a @option{^-d^/SOURCE_DIRS^}
12415 or @option{^-D^/DIR_FILES^} switch.
12416 Several switches @option{^-d^/SOURCE_DIRS^} may be specified.
12417 If @file{dir} is a relative path, it is relative to the directory of
12418 the configuration pragmas file specified with switch
12419 @option{^-c^/CONFIG_FILE^},
12420 or to the directory of the project file specified with switch
12421 @option{^-P^/PROJECT_FILE^} or,
12422 if neither switch @option{^-c^/CONFIG_FILE^}
12423 nor switch @option{^-P^/PROJECT_FILE^} are specified, it is relative to the
12424 current working directory. The directory
12425 specified with switch @option{^-d^/SOURCE_DIRS^} must exist and be readable.
12427 @item ^-D^/DIRS_FILE=^@file{file}
12428 @cindex @option{^-D^/DIRS_FILE^} (@code{gnatname})
12429 Look for source files in all directories listed in text file @file{file}.
12430 There may be zero, one or more spaces between @option{^-D^/DIRS_FILE=^}
12432 @file{file} must be an existing, readable text file.
12433 Each nonempty line in @file{file} must be a directory.
12434 Specifying switch @option{^-D^/DIRS_FILE^} is equivalent to specifying as many
12435 switches @option{^-d^/SOURCE_DIRS^} as there are nonempty lines in
12439 Follow symbolic links when processing project files.
12441 @item ^-f^/FOREIGN_PATTERN=^@file{pattern}
12442 @cindex @option{^-f^/FOREIGN_PATTERN^} (@code{gnatname})
12443 Foreign patterns. Using this switch, it is possible to add sources of languages
12444 other than Ada to the list of sources of a project file.
12445 It is only useful if a ^-P^/PROJECT_FILE^ switch is used.
12448 gnatname ^-Pprj -f"*.c"^/PROJECT_FILE=PRJ /FOREIGN_PATTERN=*.C^ "*.ada"
12451 will look for Ada units in all files with the @file{.ada} extension,
12452 and will add to the list of file for project @file{prj.gpr} the C files
12453 with extension @file{.^c^C^}.
12456 @cindex @option{^-h^/HELP^} (@code{gnatname})
12457 Output usage (help) information. The output is written to @file{stdout}.
12459 @item ^-P^/PROJECT_FILE=^@file{proj}
12460 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatname})
12461 Create or update project file @file{proj}. There may be zero, one or more space
12462 between @option{-P} and @file{proj}. @file{proj} may include directory
12463 information. @file{proj} must be writable.
12464 There may be only one switch @option{^-P^/PROJECT_FILE^}.
12465 When a switch @option{^-P^/PROJECT_FILE^} is specified,
12466 no switch @option{^-c^/CONFIG_FILE^} may be specified.
12467 On all platforms, except on VMS, when @code{gnatname} is invoked for an
12468 existing project file <proj>.gpr, a backup copy of the project file is created
12469 in the project directory with file name <proj>.gpr.saved_x. 'x' is the first
12470 non negative number that makes this backup copy a new file.
12472 @item ^-v^/VERBOSE^
12473 @cindex @option{^-v^/VERBOSE^} (@code{gnatname})
12474 Verbose mode. Output detailed explanation of behavior to @file{stdout}.
12475 This includes name of the file written, the name of the directories to search
12476 and, for each file in those directories whose name matches at least one of
12477 the Naming Patterns, an indication of whether the file contains a unit,
12478 and if so the name of the unit.
12480 @item ^-v -v^/VERBOSE /VERBOSE^
12481 @cindex @option{^-v -v^/VERBOSE /VERBOSE^} (@code{gnatname})
12482 Very Verbose mode. In addition to the output produced in verbose mode,
12483 for each file in the searched directories whose name matches none of
12484 the Naming Patterns, an indication is given that there is no match.
12486 @item ^-x^/EXCLUDED_PATTERN=^@file{pattern}
12487 @cindex @option{^-x^/EXCLUDED_PATTERN^} (@code{gnatname})
12488 Excluded patterns. Using this switch, it is possible to exclude some files
12489 that would match the name patterns. For example,
12491 gnatname ^-x "*_nt.ada"^/EXCLUDED_PATTERN=*_nt.ada^ "*.ada"
12494 will look for Ada units in all files with the @file{.ada} extension,
12495 except those whose names end with @file{_nt.ada}.
12499 @node Examples of gnatname Usage
12500 @section Examples of @code{gnatname} Usage
12504 $ gnatname /CONFIG_FILE=[HOME.ME]NAMES.ADC /SOURCE_DIRS=SOURCES "[a-z]*.ada*"
12510 $ gnatname -c /home/me/names.adc -d sources "[a-z]*.ada*"
12515 In this example, the directory @file{^/home/me^[HOME.ME]^} must already exist
12516 and be writable. In addition, the directory
12517 @file{^/home/me/sources^[HOME.ME.SOURCES]^} (specified by
12518 @option{^-d sources^/SOURCE_DIRS=SOURCES^}) must exist and be readable.
12521 Note the optional spaces after @option{-c} and @option{-d}.
12526 $ gnatname -P/home/me/proj -x "*_nt_body.ada"
12527 -dsources -dsources/plus -Dcommon_dirs.txt "body_*" "spec_*"
12530 $ gnatname /PROJECT_FILE=[HOME.ME]PROJ
12531 /EXCLUDED_PATTERN=*_nt_body.ada
12532 /SOURCE_DIRS=(SOURCES,[SOURCES.PLUS])
12533 /DIRS_FILE=COMMON_DIRS.TXT "body_*" "spec_*"
12537 Note that several switches @option{^-d^/SOURCE_DIRS^} may be used,
12538 even in conjunction with one or several switches
12539 @option{^-D^/DIRS_FILE^}. Several Naming Patterns and one excluded pattern
12540 are used in this example.
12542 @c *****************************************
12543 @c * G N A T P r o j e c t M a n a g e r *
12544 @c *****************************************
12546 @c ------ macros for projects.texi
12547 @c These macros are needed when building the gprbuild documentation, but
12548 @c should have no effect in the gnat user's guide
12550 @macro CODESAMPLE{TXT}
12558 @macro PROJECTFILE{TXT}
12562 @c simulates a newline when in a @CODESAMPLE
12573 @macro TIPHTML{TXT}
12577 @macro IMPORTANT{TXT}
12592 @include projects.texi
12594 @c ---------------------------------------------
12595 @c Tools Supporting Project Files
12596 @c ---------------------------------------------
12598 @node Tools Supporting Project Files
12599 @chapter Tools Supporting Project Files
12604 * gnatmake and Project Files::
12605 * The GNAT Driver and Project Files::
12608 @c ---------------------------------------------
12609 @node gnatmake and Project Files
12610 @section gnatmake and Project Files
12611 @c ---------------------------------------------
12614 This section covers several topics related to @command{gnatmake} and
12615 project files: defining ^switches^switches^ for @command{gnatmake}
12616 and for the tools that it invokes; specifying configuration pragmas;
12617 the use of the @code{Main} attribute; building and rebuilding library project
12621 * Switches Related to Project Files::
12622 * Switches and Project Files::
12623 * Specifying Configuration Pragmas::
12624 * Project Files and Main Subprograms::
12625 * Library Project Files::
12628 @c ---------------------------------------------
12629 @node Switches Related to Project Files
12630 @subsection Switches Related to Project Files
12631 @c ---------------------------------------------
12634 The following switches are used by GNAT tools that support project files:
12638 @item ^-P^/PROJECT_FILE=^@var{project}
12639 @cindex @option{^-P^/PROJECT_FILE^} (any project-aware tool)
12640 Indicates the name of a project file. This project file will be parsed with
12641 the verbosity indicated by @option{^-vP^MESSAGE_PROJECT_FILES=^@emph{x}},
12642 if any, and using the external references indicated
12643 by @option{^-X^/EXTERNAL_REFERENCE^} switches, if any.
12645 There may zero, one or more spaces between @option{-P} and @var{project}.
12648 There must be only one @option{^-P^/PROJECT_FILE^} switch on the command line.
12650 Since the Project Manager parses the project file only after all the switches
12651 on the command line are checked, the order of the switches
12652 @option{^-P^/PROJECT_FILE^},
12653 @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}}
12654 or @option{^-X^/EXTERNAL_REFERENCE^} is not significant.
12656 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
12657 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (any project-aware tool)
12658 Indicates that external variable @var{name} has the value @var{value}.
12659 The Project Manager will use this value for occurrences of
12660 @code{external(name)} when parsing the project file.
12663 If @var{name} or @var{value} includes a space, then @var{name=value} should be
12664 put between quotes.
12671 Several @option{^-X^/EXTERNAL_REFERENCE^} switches can be used simultaneously.
12672 If several @option{^-X^/EXTERNAL_REFERENCE^} switches specify the same
12673 @var{name}, only the last one is used.
12675 An external variable specified with a @option{^-X^/EXTERNAL_REFERENCE^} switch
12676 takes precedence over the value of the same name in the environment.
12678 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
12679 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (any project-aware tool)
12680 Indicates the verbosity of the parsing of GNAT project files.
12683 @option{-vP0} means Default;
12684 @option{-vP1} means Medium;
12685 @option{-vP2} means High.
12689 There are three possible options for this qualifier: DEFAULT, MEDIUM and
12693 The default is ^Default^DEFAULT^: no output for syntactically correct
12695 If several @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}} switches are present,
12696 only the last one is used.
12698 @item ^-aP^/ADD_PROJECT_SEARCH_DIR=^<dir>
12699 @cindex @option{^-aP^/ADD_PROJECT_SEARCH_DIR=^} (any project-aware tool)
12700 Add directory <dir> at the beginning of the project search path, in order,
12701 after the current working directory.
12705 @cindex @option{-eL} (any project-aware tool)
12706 Follow all symbolic links when processing project files.
12709 @item ^--subdirs^/SUBDIRS^=<subdir>
12710 @cindex @option{^--subdirs^/SUBDIRS^=} (gnatmake and gnatclean)
12711 This switch is recognized by @command{gnatmake} and @command{gnatclean}. It
12712 indicate that the real directories (except the source directories) are the
12713 subdirectories <subdir> of the directories specified in the project files.
12714 This applies in particular to object directories, library directories and
12715 exec directories. If the subdirectories do not exist, they are created
12720 @c ---------------------------------------------
12721 @node Switches and Project Files
12722 @subsection Switches and Project Files
12723 @c ---------------------------------------------
12727 It is not currently possible to specify VMS style qualifiers in the project
12728 files; only Unix style ^switches^switches^ may be specified.
12731 For each of the packages @code{Builder}, @code{Compiler}, @code{Binder}, and
12732 @code{Linker}, you can specify a @code{^Default_Switches^Default_Switches^}
12733 attribute, a @code{Switches} attribute, or both;
12734 as their names imply, these ^switch^switch^-related
12735 attributes affect the ^switches^switches^ that are used for each of these GNAT
12737 @command{gnatmake} is invoked. As will be explained below, these
12738 component-specific ^switches^switches^ precede
12739 the ^switches^switches^ provided on the @command{gnatmake} command line.
12741 The @code{^Default_Switches^Default_Switches^} attribute is an attribute
12742 indexed by language name (case insensitive) whose value is a string list.
12745 @smallexample @c projectfile
12747 package Compiler is
12748 for ^Default_Switches^Default_Switches^ ("Ada")
12749 use ("^-gnaty^-gnaty^",
12756 The @code{Switches} attribute is indexed on a file name (which may or may
12757 not be case sensitive, depending
12758 on the operating system) whose value is a string list. For example:
12760 @smallexample @c projectfile
12763 for Switches ("main1.adb")
12765 for Switches ("main2.adb")
12772 For the @code{Builder} package, the file names must designate source files
12773 for main subprograms. For the @code{Binder} and @code{Linker} packages, the
12774 file names must designate @file{ALI} or source files for main subprograms.
12775 In each case just the file name without an explicit extension is acceptable.
12777 For each tool used in a program build (@command{gnatmake}, the compiler, the
12778 binder, and the linker), the corresponding package @dfn{contributes} a set of
12779 ^switches^switches^ for each file on which the tool is invoked, based on the
12780 ^switch^switch^-related attributes defined in the package.
12781 In particular, the ^switches^switches^
12782 that each of these packages contributes for a given file @var{f} comprise:
12785 @item the value of attribute @code{Switches (@var{f})},
12786 if it is specified in the package for the given file,
12787 @item otherwise, the value of @code{^Default_Switches^Default_Switches^ ("Ada")},
12788 if it is specified in the package.
12793 If neither of these attributes is defined in the package, then the package does
12794 not contribute any ^switches^switches^ for the given file.
12796 When @command{gnatmake} is invoked on a file, the ^switches^switches^ comprise
12797 two sets, in the following order: those contributed for the file
12798 by the @code{Builder} package;
12799 and the switches passed on the command line.
12801 When @command{gnatmake} invokes a tool (compiler, binder, linker) on a file,
12802 the ^switches^switches^ passed to the tool comprise three sets,
12803 in the following order:
12807 the applicable ^switches^switches^ contributed for the file
12808 by the @code{Builder} package in the project file supplied on the command line;
12811 those contributed for the file by the package (in the relevant project file --
12812 see below) corresponding to the tool; and
12815 the applicable switches passed on the command line.
12818 The term @emph{applicable ^switches^switches^} reflects the fact that
12819 @command{gnatmake} ^switches^switches^ may or may not be passed to individual
12820 tools, depending on the individual ^switch^switch^.
12822 @command{gnatmake} may invoke the compiler on source files from different
12823 projects. The Project Manager will use the appropriate project file to
12824 determine the @code{Compiler} package for each source file being compiled.
12825 Likewise for the @code{Binder} and @code{Linker} packages.
12827 As an example, consider the following package in a project file:
12829 @smallexample @c projectfile
12832 package Compiler is
12833 for ^Default_Switches^Default_Switches^ ("Ada")
12835 for Switches ("a.adb")
12837 for Switches ("b.adb")
12839 "^-gnaty^-gnaty^");
12846 If @command{gnatmake} is invoked with this project file, and it needs to
12847 compile, say, the files @file{a.adb}, @file{b.adb}, and @file{c.adb}, then
12848 @file{a.adb} will be compiled with the ^switch^switch^
12849 @option{^-O1^-O1^},
12850 @file{b.adb} with ^switches^switches^
12852 and @option{^-gnaty^-gnaty^},
12853 and @file{c.adb} with @option{^-g^-g^}.
12855 The following example illustrates the ordering of the ^switches^switches^
12856 contributed by different packages:
12858 @smallexample @c projectfile
12862 for Switches ("main.adb")
12870 package Compiler is
12871 for Switches ("main.adb")
12879 If you issue the command:
12882 gnatmake ^-Pproj2^/PROJECT_FILE=PROJ2^ -O0 main
12886 then the compiler will be invoked on @file{main.adb} with the following
12887 sequence of ^switches^switches^
12890 ^-g -O1 -O2 -O0^-g -O1 -O2 -O0^
12894 with the last @option{^-O^-O^}
12895 ^switch^switch^ having precedence over the earlier ones;
12896 several other ^switches^switches^
12897 (such as @option{^-c^-c^}) are added implicitly.
12899 The ^switches^switches^
12901 and @option{^-O1^-O1^} are contributed by package
12902 @code{Builder}, @option{^-O2^-O2^} is contributed
12903 by the package @code{Compiler}
12904 and @option{^-O0^-O0^} comes from the command line.
12906 The @option{^-g^-g^}
12907 ^switch^switch^ will also be passed in the invocation of
12908 @command{Gnatlink.}
12910 A final example illustrates switch contributions from packages in different
12913 @smallexample @c projectfile
12916 for Source_Files use ("pack.ads", "pack.adb");
12917 package Compiler is
12918 for ^Default_Switches^Default_Switches^ ("Ada")
12919 use ("^-gnata^-gnata^");
12927 for Source_Files use ("foo_main.adb", "bar_main.adb");
12929 for Switches ("foo_main.adb")
12937 -- Ada source file:
12939 procedure Foo_Main is
12948 gnatmake ^-PProj4^/PROJECT_FILE=PROJ4^ foo_main.adb -cargs -gnato
12952 then the ^switches^switches^ passed to the compiler for @file{foo_main.adb} are
12953 @option{^-g^-g^} (contributed by the package @code{Proj4.Builder}) and
12954 @option{^-gnato^-gnato^} (passed on the command line).
12955 When the imported package @code{Pack} is compiled, the ^switches^switches^ used
12956 are @option{^-g^-g^} from @code{Proj4.Builder},
12957 @option{^-gnata^-gnata^} (contributed from package @code{Proj3.Compiler},
12958 and @option{^-gnato^-gnato^} from the command line.
12960 When using @command{gnatmake} with project files, some ^switches^switches^ or
12961 arguments may be expressed as relative paths. As the working directory where
12962 compilation occurs may change, these relative paths are converted to absolute
12963 paths. For the ^switches^switches^ found in a project file, the relative paths
12964 are relative to the project file directory, for the switches on the command
12965 line, they are relative to the directory where @command{gnatmake} is invoked.
12966 The ^switches^switches^ for which this occurs are:
12972 ^-aI^-aI^, as well as all arguments that are not switches (arguments to
12974 ^-o^-o^, object files specified in package @code{Linker} or after
12975 -largs on the command line). The exception to this rule is the ^switch^switch^
12976 ^--RTS=^--RTS=^ for which a relative path argument is never converted.
12978 @c ---------------------------------------------
12979 @node Specifying Configuration Pragmas
12980 @subsection Specifying Configuration Pragmas
12981 @c ---------------------------------------------
12984 When using @command{gnatmake} with project files, if there exists a file
12985 @file{gnat.adc} that contains configuration pragmas, this file will be
12988 Configuration pragmas can be defined by means of the following attributes in
12989 project files: @code{Global_Configuration_Pragmas} in package @code{Builder}
12990 and @code{Local_Configuration_Pragmas} in package @code{Compiler}.
12992 Both these attributes are single string attributes. Their values is the path
12993 name of a file containing configuration pragmas. If a path name is relative,
12994 then it is relative to the project directory of the project file where the
12995 attribute is defined.
12997 When compiling a source, the configuration pragmas used are, in order,
12998 those listed in the file designated by attribute
12999 @code{Global_Configuration_Pragmas} in package @code{Builder} of the main
13000 project file, if it is specified, and those listed in the file designated by
13001 attribute @code{Local_Configuration_Pragmas} in package @code{Compiler} of
13002 the project file of the source, if it exists.
13004 @c ---------------------------------------------
13005 @node Project Files and Main Subprograms
13006 @subsection Project Files and Main Subprograms
13007 @c ---------------------------------------------
13010 When using a project file, you can invoke @command{gnatmake}
13011 with one or several main subprograms, by specifying their source files on the
13015 gnatmake ^-P^/PROJECT_FILE=^prj main1.adb main2.adb main3.adb
13019 Each of these needs to be a source file of the same project, except
13020 when the switch ^-u^/UNIQUE^ is used.
13022 When ^-u^/UNIQUE^ is not used, all the mains need to be sources of the
13023 same project, one of the project in the tree rooted at the project specified
13024 on the command line. The package @code{Builder} of this common project, the
13025 "main project" is the one that is considered by @command{gnatmake}.
13027 When ^-u^/UNIQUE^ is used, the specified source files may be in projects
13028 imported directly or indirectly by the project specified on the command line.
13029 Note that if such a source file is not part of the project specified on the
13030 command line, the ^switches^switches^ found in package @code{Builder} of the
13031 project specified on the command line, if any, that are transmitted
13032 to the compiler will still be used, not those found in the project file of
13035 When using a project file, you can also invoke @command{gnatmake} without
13036 explicitly specifying any main, and the effect depends on whether you have
13037 defined the @code{Main} attribute. This attribute has a string list value,
13038 where each element in the list is the name of a source file (the file
13039 extension is optional) that contains a unit that can be a main subprogram.
13041 If the @code{Main} attribute is defined in a project file as a non-empty
13042 string list and the switch @option{^-u^/UNIQUE^} is not used on the command
13043 line, then invoking @command{gnatmake} with this project file but without any
13044 main on the command line is equivalent to invoking @command{gnatmake} with all
13045 the file names in the @code{Main} attribute on the command line.
13048 @smallexample @c projectfile
13051 for Main use ("main1.adb", "main2.adb", "main3.adb");
13057 With this project file, @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^"}
13059 @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^ main1.adb main2.adb main3.adb"}.
13061 When the project attribute @code{Main} is not specified, or is specified
13062 as an empty string list, or when the switch @option{-u} is used on the command
13063 line, then invoking @command{gnatmake} with no main on the command line will
13064 result in all immediate sources of the project file being checked, and
13065 potentially recompiled. Depending on the presence of the switch @option{-u},
13066 sources from other project files on which the immediate sources of the main
13067 project file depend are also checked and potentially recompiled. In other
13068 words, the @option{-u} switch is applied to all of the immediate sources of the
13071 When no main is specified on the command line and attribute @code{Main} exists
13072 and includes several mains, or when several mains are specified on the
13073 command line, the default ^switches^switches^ in package @code{Builder} will
13074 be used for all mains, even if there are specific ^switches^switches^
13075 specified for one or several mains.
13077 But the ^switches^switches^ from package @code{Binder} or @code{Linker} will be
13078 the specific ^switches^switches^ for each main, if they are specified.
13080 @c ---------------------------------------------
13081 @node Library Project Files
13082 @subsection Library Project Files
13083 @c ---------------------------------------------
13086 When @command{gnatmake} is invoked with a main project file that is a library
13087 project file, it is not allowed to specify one or more mains on the command
13090 When a library project file is specified, switches ^-b^/ACTION=BIND^ and
13091 ^-l^/ACTION=LINK^ have special meanings.
13094 @item ^-b^/ACTION=BIND^ is only allowed for stand-alone libraries. It indicates
13095 to @command{gnatmake} that @command{gnatbind} should be invoked for the
13098 @item ^-l^/ACTION=LINK^ may be used for all library projects. It indicates
13099 to @command{gnatmake} that the binder generated file should be compiled
13100 (in the case of a stand-alone library) and that the library should be built.
13103 @c ---------------------------------------------
13104 @node The GNAT Driver and Project Files
13105 @section The GNAT Driver and Project Files
13106 @c ---------------------------------------------
13109 A number of GNAT tools, other than @command{^gnatmake^gnatmake^}
13110 can benefit from project files:
13111 (@command{^gnatbind^gnatbind^},
13112 @ifclear FSFEDITION
13113 @command{^gnatcheck^gnatcheck^},
13115 @command{^gnatclean^gnatclean^},
13116 @ifclear FSFEDITION
13117 @command{^gnatelim^gnatelim^},
13119 @command{^gnatfind^gnatfind^},
13120 @command{^gnatlink^gnatlink^},
13121 @command{^gnatls^gnatls^},
13122 @ifclear FSFEDITION
13123 @command{^gnatmetric^gnatmetric^},
13124 @command{^gnatpp^gnatpp^},
13125 @command{^gnatstub^gnatstub^},
13127 and @command{^gnatxref^gnatxref^}). However, none of these tools can be invoked
13128 directly with a project file switch (@option{^-P^/PROJECT_FILE=^}).
13129 They must be invoked through the @command{gnat} driver.
13131 The @command{gnat} driver is a wrapper that accepts a number of commands and
13132 calls the corresponding tool. It was designed initially for VMS platforms (to
13133 convert VMS qualifiers to Unix-style switches), but it is now available on all
13136 On non-VMS platforms, the @command{gnat} driver accepts the following commands
13137 (case insensitive):
13140 @item BIND to invoke @command{^gnatbind^gnatbind^}
13141 @item CHOP to invoke @command{^gnatchop^gnatchop^}
13142 @item CLEAN to invoke @command{^gnatclean^gnatclean^}
13143 @item COMP or COMPILE to invoke the compiler
13144 @ifclear FSFEDITION
13145 @item ELIM to invoke @command{^gnatelim^gnatelim^}
13147 @item FIND to invoke @command{^gnatfind^gnatfind^}
13148 @item KR or KRUNCH to invoke @command{^gnatkr^gnatkr^}
13149 @item LINK to invoke @command{^gnatlink^gnatlink^}
13150 @item LS or LIST to invoke @command{^gnatls^gnatls^}
13151 @item MAKE to invoke @command{^gnatmake^gnatmake^}
13152 @item NAME to invoke @command{^gnatname^gnatname^}
13153 @item PREP or PREPROCESS to invoke @command{^gnatprep^gnatprep^}
13154 @ifclear FSFEDITION
13155 @item PP or PRETTY to invoke @command{^gnatpp^gnatpp^}
13156 @item METRIC to invoke @command{^gnatmetric^gnatmetric^}
13157 @item STUB to invoke @command{^gnatstub^gnatstub^}
13159 @item XREF to invoke @command{^gnatxref^gnatxref^}
13164 (note that the compiler is invoked using the command
13165 @command{^gnatmake -f -u -c^gnatmake -f -u -c^}).
13167 On non-VMS platforms, between @command{gnat} and the command, two
13168 special switches may be used:
13171 @item @command{-v} to display the invocation of the tool.
13172 @item @command{-dn} to prevent the @command{gnat} driver from removing
13173 the temporary files it has created. These temporary files are
13174 configuration files and temporary file list files.
13179 The command may be followed by switches and arguments for the invoked
13183 gnat bind -C main.ali
13189 Switches may also be put in text files, one switch per line, and the text
13190 files may be specified with their path name preceded by '@@'.
13193 gnat bind @@args.txt main.ali
13197 In addition, for commands BIND, COMP or COMPILE, FIND,
13198 @ifclear FSFEDITION
13202 @ifclear FSFEDITION
13207 and XREF, the project file related switches
13208 (@option{^-P^/PROJECT_FILE^},
13209 @option{^-X^/EXTERNAL_REFERENCE^} and
13210 @option{^-vP^/MESSAGES_PROJECT_FILE=^x}) may be used in addition to
13211 the switches of the invoking tool.
13213 @ifclear FSFEDITION
13214 When GNAT PP or GNAT PRETTY is used with a project file, but with no source
13215 specified on the command line, it invokes @command{^gnatpp^gnatpp^} with all
13216 the immediate sources of the specified project file.
13219 @ifclear FSFEDITION
13220 When GNAT METRIC is used with a project file, but with no source
13221 specified on the command line, it invokes @command{^gnatmetric^gnatmetric^}
13222 with all the immediate sources of the specified project file and with
13223 @option{^-d^/DIRECTORY^} with the parameter pointing to the object directory
13227 @ifclear FSFEDITION
13228 In addition, when GNAT PP, GNAT PRETTY or GNAT METRIC is used with
13229 a project file, no source is specified on the command line and
13230 switch ^-U^/ALL_PROJECTS^ is specified on the command line, then
13231 the underlying tool (^gnatpp^gnatpp^ or
13232 ^gnatmetric^gnatmetric^) is invoked for all sources of all projects,
13233 not only for the immediate sources of the main project.
13235 (-U stands for Universal or Union of the project files of the project tree)
13239 For each of the following commands, there is optionally a corresponding
13240 package in the main project.
13243 @item package @code{Binder} for command BIND (invoking @code{^gnatbind^gnatbind^})
13245 @ifclear FSFEDITION
13246 @item package @code{Check} for command CHECK (invoking
13247 @code{^gnatcheck^gnatcheck^})
13250 @item package @code{Compiler} for command COMP or COMPILE (invoking the compiler)
13252 @item package @code{Cross_Reference} for command XREF (invoking
13253 @code{^gnatxref^gnatxref^})
13255 @ifclear FSFEDITION
13256 @item package @code{Eliminate} for command ELIM (invoking
13257 @code{^gnatelim^gnatelim^})
13260 @item package @code{Finder} for command FIND (invoking @code{^gnatfind^gnatfind^})
13262 @item package @code{Gnatls} for command LS or LIST (invoking @code{^gnatls^gnatls^})
13264 @ifclear FSFEDITION
13265 @item package @code{Gnatstub} for command STUB
13266 (invoking @code{^gnatstub^gnatstub^})
13269 @item package @code{Linker} for command LINK (invoking @code{^gnatlink^gnatlink^})
13271 @ifclear FSFEDITION
13272 @item package @code{Check} for command CHECK
13273 (invoking @code{^gnatcheck^gnatcheck^})
13276 @ifclear FSFEDITION
13277 @item package @code{Metrics} for command METRIC
13278 (invoking @code{^gnatmetric^gnatmetric^})
13281 @ifclear FSFEDITION
13282 @item package @code{Pretty_Printer} for command PP or PRETTY
13283 (invoking @code{^gnatpp^gnatpp^})
13289 Package @code{Gnatls} has a unique attribute @code{Switches},
13290 a simple variable with a string list value. It contains ^switches^switches^
13291 for the invocation of @code{^gnatls^gnatls^}.
13293 @smallexample @c projectfile
13306 All other packages have two attribute @code{Switches} and
13307 @code{^Default_Switches^Default_Switches^}.
13309 @code{Switches} is an indexed attribute, indexed by the
13310 source file name, that has a string list value: the ^switches^switches^ to be
13311 used when the tool corresponding to the package is invoked for the specific
13314 @code{^Default_Switches^Default_Switches^} is an attribute,
13315 indexed by the programming language that has a string list value.
13316 @code{^Default_Switches^Default_Switches^ ("Ada")} contains the
13317 ^switches^switches^ for the invocation of the tool corresponding
13318 to the package, except if a specific @code{Switches} attribute
13319 is specified for the source file.
13321 @smallexample @c projectfile
13325 for Source_Dirs use ("**");
13335 package Compiler is
13336 for ^Default_Switches^Default_Switches^ ("Ada")
13337 use ("^-gnatv^-gnatv^",
13338 "^-gnatwa^-gnatwa^");
13344 for ^Default_Switches^Default_Switches^ ("Ada")
13352 for ^Default_Switches^Default_Switches^ ("Ada")
13354 for Switches ("main.adb")
13363 for ^Default_Switches^Default_Switches^ ("Ada")
13370 package Cross_Reference is
13371 for ^Default_Switches^Default_Switches^ ("Ada")
13376 end Cross_Reference;
13382 With the above project file, commands such as
13385 ^gnat comp -Pproj main^GNAT COMP /PROJECT_FILE=PROJ MAIN^
13386 ^gnat ls -Pproj main^GNAT LIST /PROJECT_FILE=PROJ MAIN^
13387 ^gnat xref -Pproj main^GNAT XREF /PROJECT_FILE=PROJ MAIN^
13388 ^gnat bind -Pproj main.ali^GNAT BIND /PROJECT_FILE=PROJ MAIN.ALI^
13389 ^gnat link -Pproj main.ali^GNAT LINK /PROJECT_FILE=PROJ MAIN.ALI^
13393 will set up the environment properly and invoke the tool with the switches
13394 found in the package corresponding to the tool:
13395 @code{^Default_Switches^Default_Switches^ ("Ada")} for all tools,
13396 except @code{Switches ("main.adb")}
13397 for @code{^gnatlink^gnatlink^}.
13398 @ifclear FSFEDITION
13399 It is also possible to invoke some of the tools,
13400 (@code{^gnatcheck^gnatcheck^},
13401 @code{^gnatmetric^gnatmetric^},
13402 and @code{^gnatpp^gnatpp^})
13403 on a set of project units thanks to the combination of the switches
13404 @option{-P}, @option{-U} and possibly the main unit when one is interested
13405 in its closure. For instance,
13411 will compute the metrics for all the immediate units of project
13414 gnat metric -Pproj -U
13418 will compute the metrics for all the units of the closure of projects
13419 rooted at @code{proj}.
13421 gnat metric -Pproj -U main_unit
13425 will compute the metrics for the closure of units rooted at
13426 @code{main_unit}. This last possibility relies implicitly
13427 on @command{gnatbind}'s option @option{-R}. But if the argument files for the
13428 tool invoked by the @command{gnat} driver are explicitly specified
13429 either directly or through the tool @option{-files} option, then the tool
13430 is called only for these explicitly specified files.
13433 @c *****************************************
13434 @c * Cross-referencing tools
13435 @c *****************************************
13437 @node The Cross-Referencing Tools gnatxref and gnatfind
13438 @chapter The Cross-Referencing Tools @code{gnatxref} and @code{gnatfind}
13443 The compiler generates cross-referencing information (unless
13444 you set the @samp{-gnatx} switch), which are saved in the @file{.ali} files.
13445 This information indicates where in the source each entity is declared and
13446 referenced. Note that entities in package Standard are not included, but
13447 entities in all other predefined units are included in the output.
13449 Before using any of these two tools, you need to compile successfully your
13450 application, so that GNAT gets a chance to generate the cross-referencing
13453 The two tools @code{gnatxref} and @code{gnatfind} take advantage of this
13454 information to provide the user with the capability to easily locate the
13455 declaration and references to an entity. These tools are quite similar,
13456 the difference being that @code{gnatfind} is intended for locating
13457 definitions and/or references to a specified entity or entities, whereas
13458 @code{gnatxref} is oriented to generating a full report of all
13461 To use these tools, you must not compile your application using the
13462 @option{-gnatx} switch on the @command{gnatmake} command line
13463 (@pxref{The GNAT Make Program gnatmake}). Otherwise, cross-referencing
13464 information will not be generated.
13466 Note: to invoke @code{gnatxref} or @code{gnatfind} with a project file,
13467 use the @code{gnat} driver (see @ref{The GNAT Driver and Project Files}).
13470 * Switches for gnatxref::
13471 * Switches for gnatfind::
13472 * Project Files for gnatxref and gnatfind::
13473 * Regular Expressions in gnatfind and gnatxref::
13474 * Examples of gnatxref Usage::
13475 * Examples of gnatfind Usage::
13478 @node Switches for gnatxref
13479 @section @code{gnatxref} Switches
13482 The command invocation for @code{gnatxref} is:
13484 @c $ gnatxref @ovar{switches} @var{sourcefile1} @r{[}@var{sourcefile2} @dots{}@r{]}
13485 @c Expanding @ovar macro inline (explanation in macro def comments)
13486 $ gnatxref @r{[}@var{switches}@r{]} @var{sourcefile1} @r{[}@var{sourcefile2} @dots{}@r{]}
13495 identifies the source files for which a report is to be generated. The
13496 ``with''ed units will be processed too. You must provide at least one file.
13498 These file names are considered to be regular expressions, so for instance
13499 specifying @file{source*.adb} is the same as giving every file in the current
13500 directory whose name starts with @file{source} and whose extension is
13503 You shouldn't specify any directory name, just base names. @command{gnatxref}
13504 and @command{gnatfind} will be able to locate these files by themselves using
13505 the source path. If you specify directories, no result is produced.
13510 The switches can be:
13514 @cindex @option{--version} @command{gnatxref}
13515 Display Copyright and version, then exit disregarding all other options.
13518 @cindex @option{--help} @command{gnatxref}
13519 If @option{--version} was not used, display usage, then exit disregarding
13522 @item ^-a^/ALL_FILES^
13523 @cindex @option{^-a^/ALL_FILES^} (@command{gnatxref})
13524 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
13525 the read-only files found in the library search path. Otherwise, these files
13526 will be ignored. This option can be used to protect Gnat sources or your own
13527 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
13528 much faster, and their output much smaller. Read-only here refers to access
13529 or permissions status in the file system for the current user.
13532 @cindex @option{-aIDIR} (@command{gnatxref})
13533 When looking for source files also look in directory DIR. The order in which
13534 source file search is undertaken is the same as for @command{gnatmake}.
13537 @cindex @option{-aODIR} (@command{gnatxref})
13538 When searching for library and object files, look in directory
13539 DIR. The order in which library files are searched is the same as for
13540 @command{gnatmake}.
13543 @cindex @option{-nostdinc} (@command{gnatxref})
13544 Do not look for sources in the system default directory.
13547 @cindex @option{-nostdlib} (@command{gnatxref})
13548 Do not look for library files in the system default directory.
13550 @item --ext=@var{extension}
13551 @cindex @option{--ext} (@command{gnatxref})
13552 Specify an alternate ali file extension. The default is @code{ali} and other
13553 extensions (e.g. @code{gli} for C/C++ sources when using @option{-fdump-xref})
13554 may be specified via this switch. Note that if this switch overrides the
13555 default, which means that only the new extension will be considered.
13557 @item --RTS=@var{rts-path}
13558 @cindex @option{--RTS} (@command{gnatxref})
13559 Specifies the default location of the runtime library. Same meaning as the
13560 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
13562 @item ^-d^/DERIVED_TYPES^
13563 @cindex @option{^-d^/DERIVED_TYPES^} (@command{gnatxref})
13564 If this switch is set @code{gnatxref} will output the parent type
13565 reference for each matching derived types.
13567 @item ^-f^/FULL_PATHNAME^
13568 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatxref})
13569 If this switch is set, the output file names will be preceded by their
13570 directory (if the file was found in the search path). If this switch is
13571 not set, the directory will not be printed.
13573 @item ^-g^/IGNORE_LOCALS^
13574 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatxref})
13575 If this switch is set, information is output only for library-level
13576 entities, ignoring local entities. The use of this switch may accelerate
13577 @code{gnatfind} and @code{gnatxref}.
13580 @cindex @option{-IDIR} (@command{gnatxref})
13581 Equivalent to @samp{-aODIR -aIDIR}.
13584 @cindex @option{-pFILE} (@command{gnatxref})
13585 Specify a project file to use @xref{GNAT Project Manager}.
13586 If you need to use the @file{.gpr}
13587 project files, you should use gnatxref through the GNAT driver
13588 (@command{gnat xref -Pproject}).
13590 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
13591 project file in the current directory.
13593 If a project file is either specified or found by the tools, then the content
13594 of the source directory and object directory lines are added as if they
13595 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^}
13596 and @samp{^-aO^OBJECT_SEARCH^}.
13598 Output only unused symbols. This may be really useful if you give your
13599 main compilation unit on the command line, as @code{gnatxref} will then
13600 display every unused entity and 'with'ed package.
13604 Instead of producing the default output, @code{gnatxref} will generate a
13605 @file{tags} file that can be used by vi. For examples how to use this
13606 feature, see @ref{Examples of gnatxref Usage}. The tags file is output
13607 to the standard output, thus you will have to redirect it to a file.
13613 All these switches may be in any order on the command line, and may even
13614 appear after the file names. They need not be separated by spaces, thus
13615 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
13616 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
13618 @node Switches for gnatfind
13619 @section @code{gnatfind} Switches
13622 The command line for @code{gnatfind} is:
13625 @c $ gnatfind @ovar{switches} @var{pattern}@r{[}:@var{sourcefile}@r{[}:@var{line}@r{[}:@var{column}@r{]]]}
13626 @c @r{[}@var{file1} @var{file2} @dots{}]
13627 @c Expanding @ovar macro inline (explanation in macro def comments)
13628 $ gnatfind @r{[}@var{switches}@r{]} @var{pattern}@r{[}:@var{sourcefile}@r{[}:@var{line}@r{[}:@var{column}@r{]]]}
13629 @r{[}@var{file1} @var{file2} @dots{}@r{]}
13637 An entity will be output only if it matches the regular expression found
13638 in @var{pattern}, see @ref{Regular Expressions in gnatfind and gnatxref}.
13640 Omitting the pattern is equivalent to specifying @samp{*}, which
13641 will match any entity. Note that if you do not provide a pattern, you
13642 have to provide both a sourcefile and a line.
13644 Entity names are given in Latin-1, with uppercase/lowercase equivalence
13645 for matching purposes. At the current time there is no support for
13646 8-bit codes other than Latin-1, or for wide characters in identifiers.
13649 @code{gnatfind} will look for references, bodies or declarations
13650 of symbols referenced in @file{@var{sourcefile}}, at line @var{line}
13651 and column @var{column}. See @ref{Examples of gnatfind Usage}
13652 for syntax examples.
13655 is a decimal integer identifying the line number containing
13656 the reference to the entity (or entities) to be located.
13659 is a decimal integer identifying the exact location on the
13660 line of the first character of the identifier for the
13661 entity reference. Columns are numbered from 1.
13663 @item file1 file2 @dots{}
13664 The search will be restricted to these source files. If none are given, then
13665 the search will be done for every library file in the search path.
13666 These file must appear only after the pattern or sourcefile.
13668 These file names are considered to be regular expressions, so for instance
13669 specifying @file{source*.adb} is the same as giving every file in the current
13670 directory whose name starts with @file{source} and whose extension is
13673 The location of the spec of the entity will always be displayed, even if it
13674 isn't in one of @file{@var{file1}}, @file{@var{file2}},@enddots{} The
13675 occurrences of the entity in the separate units of the ones given on the
13676 command line will also be displayed.
13678 Note that if you specify at least one file in this part, @code{gnatfind} may
13679 sometimes not be able to find the body of the subprograms.
13684 At least one of 'sourcefile' or 'pattern' has to be present on
13687 The following switches are available:
13691 @cindex @option{--version} @command{gnatfind}
13692 Display Copyright and version, then exit disregarding all other options.
13695 @cindex @option{--help} @command{gnatfind}
13696 If @option{--version} was not used, display usage, then exit disregarding
13699 @item ^-a^/ALL_FILES^
13700 @cindex @option{^-a^/ALL_FILES^} (@command{gnatfind})
13701 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
13702 the read-only files found in the library search path. Otherwise, these files
13703 will be ignored. This option can be used to protect Gnat sources or your own
13704 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
13705 much faster, and their output much smaller. Read-only here refers to access
13706 or permission status in the file system for the current user.
13709 @cindex @option{-aIDIR} (@command{gnatfind})
13710 When looking for source files also look in directory DIR. The order in which
13711 source file search is undertaken is the same as for @command{gnatmake}.
13714 @cindex @option{-aODIR} (@command{gnatfind})
13715 When searching for library and object files, look in directory
13716 DIR. The order in which library files are searched is the same as for
13717 @command{gnatmake}.
13720 @cindex @option{-nostdinc} (@command{gnatfind})
13721 Do not look for sources in the system default directory.
13724 @cindex @option{-nostdlib} (@command{gnatfind})
13725 Do not look for library files in the system default directory.
13727 @item --ext=@var{extension}
13728 @cindex @option{--ext} (@command{gnatfind})
13729 Specify an alternate ali file extension. The default is @code{ali} and other
13730 extensions (e.g. @code{gli} for C/C++ sources when using @option{-fdump-xref})
13731 may be specified via this switch. Note that if this switch overrides the
13732 default, which means that only the new extension will be considered.
13734 @item --RTS=@var{rts-path}
13735 @cindex @option{--RTS} (@command{gnatfind})
13736 Specifies the default location of the runtime library. Same meaning as the
13737 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
13739 @item ^-d^/DERIVED_TYPE_INFORMATION^
13740 @cindex @option{^-d^/DERIVED_TYPE_INFORMATION^} (@code{gnatfind})
13741 If this switch is set, then @code{gnatfind} will output the parent type
13742 reference for each matching derived types.
13744 @item ^-e^/EXPRESSIONS^
13745 @cindex @option{^-e^/EXPRESSIONS^} (@command{gnatfind})
13746 By default, @code{gnatfind} accept the simple regular expression set for
13747 @samp{pattern}. If this switch is set, then the pattern will be
13748 considered as full Unix-style regular expression.
13750 @item ^-f^/FULL_PATHNAME^
13751 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatfind})
13752 If this switch is set, the output file names will be preceded by their
13753 directory (if the file was found in the search path). If this switch is
13754 not set, the directory will not be printed.
13756 @item ^-g^/IGNORE_LOCALS^
13757 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatfind})
13758 If this switch is set, information is output only for library-level
13759 entities, ignoring local entities. The use of this switch may accelerate
13760 @code{gnatfind} and @code{gnatxref}.
13763 @cindex @option{-IDIR} (@command{gnatfind})
13764 Equivalent to @samp{-aODIR -aIDIR}.
13767 @cindex @option{-pFILE} (@command{gnatfind})
13768 Specify a project file (@pxref{GNAT Project Manager}) to use.
13769 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
13770 project file in the current directory.
13772 If a project file is either specified or found by the tools, then the content
13773 of the source directory and object directory lines are added as if they
13774 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^} and
13775 @samp{^-aO^/OBJECT_SEARCH^}.
13777 @item ^-r^/REFERENCES^
13778 @cindex @option{^-r^/REFERENCES^} (@command{gnatfind})
13779 By default, @code{gnatfind} will output only the information about the
13780 declaration, body or type completion of the entities. If this switch is
13781 set, the @code{gnatfind} will locate every reference to the entities in
13782 the files specified on the command line (or in every file in the search
13783 path if no file is given on the command line).
13785 @item ^-s^/PRINT_LINES^
13786 @cindex @option{^-s^/PRINT_LINES^} (@command{gnatfind})
13787 If this switch is set, then @code{gnatfind} will output the content
13788 of the Ada source file lines were the entity was found.
13790 @item ^-t^/TYPE_HIERARCHY^
13791 @cindex @option{^-t^/TYPE_HIERARCHY^} (@command{gnatfind})
13792 If this switch is set, then @code{gnatfind} will output the type hierarchy for
13793 the specified type. It act like -d option but recursively from parent
13794 type to parent type. When this switch is set it is not possible to
13795 specify more than one file.
13800 All these switches may be in any order on the command line, and may even
13801 appear after the file names. They need not be separated by spaces, thus
13802 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
13803 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
13805 As stated previously, gnatfind will search in every directory in the
13806 search path. You can force it to look only in the current directory if
13807 you specify @code{*} at the end of the command line.
13809 @node Project Files for gnatxref and gnatfind
13810 @section Project Files for @command{gnatxref} and @command{gnatfind}
13813 Project files allow a programmer to specify how to compile its
13814 application, where to find sources, etc. These files are used
13816 primarily by GPS, but they can also be used
13819 @code{gnatxref} and @code{gnatfind}.
13821 A project file name must end with @file{.gpr}. If a single one is
13822 present in the current directory, then @code{gnatxref} and @code{gnatfind} will
13823 extract the information from it. If multiple project files are found, none of
13824 them is read, and you have to use the @samp{-p} switch to specify the one
13827 The following lines can be included, even though most of them have default
13828 values which can be used in most cases.
13829 The lines can be entered in any order in the file.
13830 Except for @file{src_dir} and @file{obj_dir}, you can only have one instance of
13831 each line. If you have multiple instances, only the last one is taken into
13836 [default: @code{"^./^[]^"}]
13837 specifies a directory where to look for source files. Multiple @code{src_dir}
13838 lines can be specified and they will be searched in the order they
13842 [default: @code{"^./^[]^"}]
13843 specifies a directory where to look for object and library files. Multiple
13844 @code{obj_dir} lines can be specified, and they will be searched in the order
13847 @item comp_opt=SWITCHES
13848 [default: @code{""}]
13849 creates a variable which can be referred to subsequently by using
13850 the @code{$@{comp_opt@}} notation. This is intended to store the default
13851 switches given to @command{gnatmake} and @command{gcc}.
13853 @item bind_opt=SWITCHES
13854 [default: @code{""}]
13855 creates a variable which can be referred to subsequently by using
13856 the @samp{$@{bind_opt@}} notation. This is intended to store the default
13857 switches given to @command{gnatbind}.
13859 @item link_opt=SWITCHES
13860 [default: @code{""}]
13861 creates a variable which can be referred to subsequently by using
13862 the @samp{$@{link_opt@}} notation. This is intended to store the default
13863 switches given to @command{gnatlink}.
13865 @item main=EXECUTABLE
13866 [default: @code{""}]
13867 specifies the name of the executable for the application. This variable can
13868 be referred to in the following lines by using the @samp{$@{main@}} notation.
13871 @item comp_cmd=COMMAND
13872 [default: @code{"GNAT COMPILE /SEARCH=$@{src_dir@} /DEBUG /TRY_SEMANTICS"}]
13875 @item comp_cmd=COMMAND
13876 [default: @code{"gcc -c -I$@{src_dir@} -g -gnatq"}]
13878 specifies the command used to compile a single file in the application.
13881 @item make_cmd=COMMAND
13882 [default: @code{"GNAT MAKE $@{main@}
13883 /SOURCE_SEARCH=$@{src_dir@} /OBJECT_SEARCH=$@{obj_dir@}
13884 /DEBUG /TRY_SEMANTICS /COMPILER_QUALIFIERS $@{comp_opt@}
13885 /BINDER_QUALIFIERS $@{bind_opt@} /LINKER_QUALIFIERS $@{link_opt@}"}]
13888 @item make_cmd=COMMAND
13889 [default: @code{"gnatmake $@{main@} -aI$@{src_dir@}
13890 -aO$@{obj_dir@} -g -gnatq -cargs $@{comp_opt@}
13891 -bargs $@{bind_opt@} -largs $@{link_opt@}"}]
13893 specifies the command used to recompile the whole application.
13895 @item run_cmd=COMMAND
13896 [default: @code{"$@{main@}"}]
13897 specifies the command used to run the application.
13899 @item debug_cmd=COMMAND
13900 [default: @code{"gdb $@{main@}"}]
13901 specifies the command used to debug the application
13906 @command{gnatxref} and @command{gnatfind} only take into account the
13907 @code{src_dir} and @code{obj_dir} lines, and ignore the others.
13909 @node Regular Expressions in gnatfind and gnatxref
13910 @section Regular Expressions in @code{gnatfind} and @code{gnatxref}
13913 As specified in the section about @command{gnatfind}, the pattern can be a
13914 regular expression. Actually, there are to set of regular expressions
13915 which are recognized by the program:
13918 @item globbing patterns
13919 These are the most usual regular expression. They are the same that you
13920 generally used in a Unix shell command line, or in a DOS session.
13922 Here is a more formal grammar:
13929 term ::= elmt -- matches elmt
13930 term ::= elmt elmt -- concatenation (elmt then elmt)
13931 term ::= * -- any string of 0 or more characters
13932 term ::= ? -- matches any character
13933 term ::= [char @{char@}] -- matches any character listed
13934 term ::= [char - char] -- matches any character in range
13938 @item full regular expression
13939 The second set of regular expressions is much more powerful. This is the
13940 type of regular expressions recognized by utilities such a @file{grep}.
13942 The following is the form of a regular expression, expressed in Ada
13943 reference manual style BNF is as follows
13950 regexp ::= term @{| term@} -- alternation (term or term @dots{})
13952 term ::= item @{item@} -- concatenation (item then item)
13954 item ::= elmt -- match elmt
13955 item ::= elmt * -- zero or more elmt's
13956 item ::= elmt + -- one or more elmt's
13957 item ::= elmt ? -- matches elmt or nothing
13960 elmt ::= nschar -- matches given character
13961 elmt ::= [nschar @{nschar@}] -- matches any character listed
13962 elmt ::= [^^^ nschar @{nschar@}] -- matches any character not listed
13963 elmt ::= [char - char] -- matches chars in given range
13964 elmt ::= \ char -- matches given character
13965 elmt ::= . -- matches any single character
13966 elmt ::= ( regexp ) -- parens used for grouping
13968 char ::= any character, including special characters
13969 nschar ::= any character except ()[].*+?^^^
13973 Following are a few examples:
13977 will match any of the two strings @samp{abcde} and @samp{fghi},
13980 will match any string like @samp{abd}, @samp{abcd}, @samp{abccd},
13981 @samp{abcccd}, and so on,
13984 will match any string which has only lowercase characters in it (and at
13985 least one character.
13990 @node Examples of gnatxref Usage
13991 @section Examples of @code{gnatxref} Usage
13993 @subsection General Usage
13996 For the following examples, we will consider the following units:
13998 @smallexample @c ada
14004 3: procedure Foo (B : in Integer);
14011 1: package body Main is
14012 2: procedure Foo (B : in Integer) is
14023 2: procedure Print (B : Integer);
14032 The first thing to do is to recompile your application (for instance, in
14033 that case just by doing a @samp{gnatmake main}, so that GNAT generates
14034 the cross-referencing information.
14035 You can then issue any of the following commands:
14037 @item gnatxref main.adb
14038 @code{gnatxref} generates cross-reference information for main.adb
14039 and every unit 'with'ed by main.adb.
14041 The output would be:
14049 Decl: main.ads 3:20
14050 Body: main.adb 2:20
14051 Ref: main.adb 4:13 5:13 6:19
14054 Ref: main.adb 6:8 7:8
14064 Decl: main.ads 3:15
14065 Body: main.adb 2:15
14068 Body: main.adb 1:14
14071 Ref: main.adb 6:12 7:12
14075 that is the entity @code{Main} is declared in main.ads, line 2, column 9,
14076 its body is in main.adb, line 1, column 14 and is not referenced any where.
14078 The entity @code{Print} is declared in bar.ads, line 2, column 15 and it
14079 is referenced in main.adb, line 6 column 12 and line 7 column 12.
14081 @item gnatxref package1.adb package2.ads
14082 @code{gnatxref} will generates cross-reference information for
14083 package1.adb, package2.ads and any other package 'with'ed by any
14089 @subsection Using gnatxref with vi
14091 @code{gnatxref} can generate a tags file output, which can be used
14092 directly from @command{vi}. Note that the standard version of @command{vi}
14093 will not work properly with overloaded symbols. Consider using another
14094 free implementation of @command{vi}, such as @command{vim}.
14097 $ gnatxref -v gnatfind.adb > tags
14101 will generate the tags file for @code{gnatfind} itself (if the sources
14102 are in the search path!).
14104 From @command{vi}, you can then use the command @samp{:tag @var{entity}}
14105 (replacing @var{entity} by whatever you are looking for), and vi will
14106 display a new file with the corresponding declaration of entity.
14109 @node Examples of gnatfind Usage
14110 @section Examples of @code{gnatfind} Usage
14114 @item gnatfind ^-f^/FULL_PATHNAME^ xyz:main.adb
14115 Find declarations for all entities xyz referenced at least once in
14116 main.adb. The references are search in every library file in the search
14119 The directories will be printed as well (as the @samp{^-f^/FULL_PATHNAME^}
14122 The output will look like:
14124 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
14125 ^directory/^[directory]^main.adb:24:10: xyz <= body
14126 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
14130 that is to say, one of the entities xyz found in main.adb is declared at
14131 line 12 of main.ads (and its body is in main.adb), and another one is
14132 declared at line 45 of foo.ads
14134 @item gnatfind ^-fs^/FULL_PATHNAME/SOURCE_LINE^ xyz:main.adb
14135 This is the same command as the previous one, instead @code{gnatfind} will
14136 display the content of the Ada source file lines.
14138 The output will look like:
14141 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
14143 ^directory/^[directory]^main.adb:24:10: xyz <= body
14145 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
14150 This can make it easier to find exactly the location your are looking
14153 @item gnatfind ^-r^/REFERENCES^ "*x*":main.ads:123 foo.adb
14154 Find references to all entities containing an x that are
14155 referenced on line 123 of main.ads.
14156 The references will be searched only in main.ads and foo.adb.
14158 @item gnatfind main.ads:123
14159 Find declarations and bodies for all entities that are referenced on
14160 line 123 of main.ads.
14162 This is the same as @code{gnatfind "*":main.adb:123}.
14164 @item gnatfind ^mydir/^[mydir]^main.adb:123:45
14165 Find the declaration for the entity referenced at column 45 in
14166 line 123 of file main.adb in directory mydir. Note that it
14167 is usual to omit the identifier name when the column is given,
14168 since the column position identifies a unique reference.
14170 The column has to be the beginning of the identifier, and should not
14171 point to any character in the middle of the identifier.
14175 @ifclear FSFEDITION
14176 @c *********************************
14177 @node The GNAT Pretty-Printer gnatpp
14178 @chapter The GNAT Pretty-Printer @command{gnatpp}
14180 @cindex Pretty-Printer
14183 * Switches for gnatpp::
14184 * Formatting Rules::
14188 ^The @command{gnatpp} tool^GNAT PRETTY^ is an ASIS-based utility
14189 for source reformatting / pretty-printing.
14190 It takes an Ada source file as input and generates a reformatted
14192 You can specify various style directives via switches; e.g.,
14193 identifier case conventions, rules of indentation, and comment layout.
14195 Note: A newly-redesigned set of formatting algorithms used by gnatpp
14197 To invoke the old formatting algorithms, use the @option{--pp-old} switch.
14198 Support for @option{--pp-old} will be removed in some future version.
14200 To produce a reformatted file, @command{gnatpp} invokes the Ada
14201 compiler and generates and uses the ASIS tree for the input source;
14202 thus the input must be legal Ada code, and the tool should have all the
14203 information needed to compile the input source. To provide this information,
14204 you may specify as a tool parameter the project file the input source belongs to
14205 (or you may call @command{gnatpp}
14206 through the @command{gnat} driver (see @ref{The GNAT Driver and
14207 Project Files}). Another possibility is to specify the source search
14208 path and needed configuration files in @option{-cargs} section of @command{gnatpp}
14209 call, see the description of the @command{gnatpp} switches below.
14211 @command{gnatpp} cannot process sources that contain
14212 preprocessing directives.
14214 The @command{gnatpp} command has the form
14217 @c $ gnatpp @ovar{switches} @var{filename}
14218 @c Expanding @ovar macro inline (explanation in macro def comments)
14219 $ gnatpp @r{[}@var{switches}@r{]} @var{filename} @r{[}-cargs @var{gcc_switches}@r{]}
14226 @var{switches} is an optional sequence of switches defining such properties as
14227 the formatting rules, the source search path, and the destination for the
14231 @var{filename} is the name (including the extension) of the source file to
14232 reformat; wildcards or several file names on the same gnatpp command are
14233 allowed. The file name may contain path information; it does not have to
14234 follow the GNAT file naming rules
14237 @samp{@var{gcc_switches}} is a list of switches for
14238 @command{gcc}. They will be passed on to all compiler invocations made by
14239 @command{gnatpp} to generate the ASIS trees. Here you can provide
14240 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
14241 use the @option{-gnatec} switch to set the configuration file, etc.
14244 @node Switches for gnatpp
14245 @section Switches for @command{gnatpp}
14248 The following subsections describe the various switches accepted by
14249 @command{gnatpp}, organized by category.
14252 You specify a switch by supplying a name and generally also a value.
14253 In many cases the values for a switch with a given name are incompatible with
14255 (for example the switch that controls the casing of a reserved word may have
14256 exactly one value: upper case, lower case, or
14257 mixed case) and thus exactly one such switch can be in effect for an
14258 invocation of @command{gnatpp}.
14259 If more than one is supplied, the last one is used.
14260 However, some values for the same switch are mutually compatible.
14261 You may supply several such switches to @command{gnatpp}, but then
14262 each must be specified in full, with both the name and the value.
14263 Abbreviated forms (the name appearing once, followed by each value) are
14268 In many cases the set of options for a given qualifier are incompatible with
14269 each other (for example the qualifier that controls the casing of a reserved
14270 word may have exactly one option, which specifies either upper case, lower
14271 case, or mixed case), and thus exactly one such option can be in effect for
14272 an invocation of @command{gnatpp}.
14273 If more than one is supplied, the last one is used.
14277 * Alignment Control::
14279 * General Text Layout Control::
14280 * Other Formatting Options::
14281 * Setting the Source Search Path::
14282 * Output File Control::
14283 * Other gnatpp Switches::
14286 @node Alignment Control
14287 @subsection Alignment Control
14288 @cindex Alignment control in @command{gnatpp}
14291 Programs can be easier to read if certain constructs are vertically aligned.
14292 By default alignment of the following constructs is set ON:
14293 @code{:} in declarations, @code{:=} in initializations in declarations
14294 @code{:=} in assignment statements, @code{=>} in associations, and
14295 @code{at} keywords in the component clauses in record
14296 representation clauses.
14299 @cindex @option{^-A@var{n}^/ALIGN^} (@command{gnatpp})
14301 @item ^-A0^/ALIGN=OFF^
14302 Set alignment to OFF
14304 @item ^-A1^/ALIGN=ON^
14305 Set alignment to ON
14308 @node Casing Control
14309 @subsection Casing Control
14310 @cindex Casing control in @command{gnatpp}
14313 @command{gnatpp} allows you to specify the casing for reserved words,
14314 pragma names, attribute designators and identifiers.
14315 For identifiers you may define a
14316 general rule for name casing but also override this rule
14317 via a set of dictionary files.
14319 Three types of casing are supported: lower case, upper case, and mixed case.
14320 ``Mixed case'' means that the first letter, and also each letter immediately
14321 following an underscore, are converted to their uppercase forms;
14322 all the other letters are converted to their lowercase forms.
14325 @cindex @option{^-a@var{x}^/ATTRIBUTE^} (@command{gnatpp})
14326 @item ^-aL^/ATTRIBUTE_CASING=LOWER_CASE^
14327 Attribute designators are lower case
14329 @item ^-aU^/ATTRIBUTE_CASING=UPPER_CASE^
14330 Attribute designators are upper case
14332 @item ^-aM^/ATTRIBUTE_CASING=MIXED_CASE^
14333 Attribute designators are mixed case (this is the default)
14335 @cindex @option{^-k@var{x}^/KEYWORD_CASING^} (@command{gnatpp})
14336 @item ^-kL^/KEYWORD_CASING=LOWER_CASE^
14337 Keywords (technically, these are known in Ada as @emph{reserved words}) are
14338 lower case (this is the default)
14340 @item ^-kU^/KEYWORD_CASING=UPPER_CASE^
14341 Keywords are upper case
14343 @cindex @option{^-n@var{x}^/NAME_CASING^} (@command{gnatpp})
14344 @item ^-nD^/NAME_CASING=AS_DECLARED^
14345 Name casing for defining occurrences are as they appear in the source file
14346 (this is the default)
14348 @item ^-nU^/NAME_CASING=UPPER_CASE^
14349 Names are in upper case
14351 @item ^-nL^/NAME_CASING=LOWER_CASE^
14352 Names are in lower case
14354 @item ^-nM^/NAME_CASING=MIXED_CASE^
14355 Names are in mixed case
14357 @cindex @option{^-ne@var{x}^/ENUM_CASING^} (@command{gnatpp})
14358 @item ^-neD^/ENUM_CASING=AS_DECLARED^
14359 Enumeration literal casing for defining occurrences are as they appear in the
14360 source file. Overrides ^-n^/NAME_CASING^ casing setting.
14362 @item ^-neU^/ENUM_CASING=UPPER_CASE^
14363 Enumeration literals are in upper case. Overrides ^-n^/NAME_CASING^ casing
14366 @item ^-neL^/ENUM_CASING=LOWER_CASE^
14367 Enumeration literals are in lower case. Overrides ^-n^/NAME_CASING^ casing
14370 @item ^-neM^/ENUM_CASING=MIXED_CASE^
14371 Enumeration literals are in mixed case. Overrides ^-n^/NAME_CASING^ casing
14374 @cindex @option{^-nt@var{x}^/TYPE_CASING^} (@command{gnatpp})
14375 @item ^-neD^/TYPE_CASING=AS_DECLARED^
14376 Names introduced by type and subtype declarations are always
14377 cased as they appear in the declaration in the source file.
14378 Overrides ^-n^/NAME_CASING^ casing setting.
14380 @item ^-ntU^/TYPE_CASING=UPPER_CASE^
14381 Names introduced by type and subtype declarations are always in
14382 upper case. Overrides ^-n^/NAME_CASING^ casing setting.
14384 @item ^-ntL^/TYPE_CASING=LOWER_CASE^
14385 Names introduced by type and subtype declarations are always in
14386 lower case. Overrides ^-n^/NAME_CASING^ casing setting.
14388 @item ^-ntM^/TYPE_CASING=MIXED_CASE^
14389 Names introduced by type and subtype declarations are always in
14390 mixed case. Overrides ^-n^/NAME_CASING^ casing setting.
14392 @item ^-nnU^/NUMBER_CASING=UPPER_CASE^
14393 Names introduced by number declarations are always in
14394 upper case. Overrides ^-n^/NAME_CASING^ casing setting.
14396 @item ^-nnL^/NUMBER_CASING=LOWER_CASE^
14397 Names introduced by number declarations are always in
14398 lower case. Overrides ^-n^/NAME_CASING^ casing setting.
14400 @item ^-nnM^/NUMBER_CASING=MIXED_CASE^
14401 Names introduced by number declarations are always in
14402 mixed case. Overrides ^-n^/NAME_CASING^ casing setting.
14404 @cindex @option{^-p@var{x}^/PRAGMA_CASING^} (@command{gnatpp})
14405 @item ^-pL^/PRAGMA_CASING=LOWER_CASE^
14406 Pragma names are lower case
14408 @item ^-pU^/PRAGMA_CASING=UPPER_CASE^
14409 Pragma names are upper case
14411 @item ^-pM^/PRAGMA_CASING=MIXED_CASE^
14412 Pragma names are mixed case (this is the default)
14414 @item ^-D@var{file}^/DICTIONARY=@var{file}^
14415 @cindex @option{^-D^/DICTIONARY^} (@command{gnatpp})
14416 Use @var{file} as a @emph{dictionary file} that defines
14417 the casing for a set of specified names,
14418 thereby overriding the effect on these names by
14419 any explicit or implicit
14420 ^-n^/NAME_CASING^ switch.
14421 To supply more than one dictionary file,
14422 use ^several @option{-D} switches^a list of files as options^.
14425 @option{gnatpp} implicitly uses a @emph{default dictionary file}
14426 to define the casing for the Ada predefined names and
14427 the names declared in the GNAT libraries.
14429 @item ^-D-^/SPECIFIC_CASING^
14430 @cindex @option{^-D-^/SPECIFIC_CASING^} (@command{gnatpp})
14431 Do not use the default dictionary file;
14432 instead, use the casing
14433 defined by a @option{^-n^/NAME_CASING^} switch and any explicit
14438 The structure of a dictionary file, and details on the conventions
14439 used in the default dictionary file, are defined in @ref{Name Casing}.
14441 The @option{^-D-^/SPECIFIC_CASING^} and
14442 @option{^-D@var{file}^/DICTIONARY=@var{file}^} switches are mutually
14446 This group of @command{gnatpp} switches controls the layout of comments and
14447 complex syntactic constructs. See @ref{Formatting Comments} for details
14451 @cindex @option{^-c@var{n}^/COMMENTS_LAYOUT^} (@command{gnatpp})
14452 @item ^-c0^/COMMENTS_LAYOUT=UNTOUCHED^
14453 All comments remain unchanged.
14455 @item ^-c1^/COMMENTS_LAYOUT=DEFAULT^
14456 GNAT-style comment line indentation.
14457 This is the default.
14459 @item ^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^
14460 GNAT-style comment beginning.
14462 @item ^-c4^/COMMENTS_LAYOUT=REFORMAT^
14463 Fill comment blocks.
14465 @item ^-c5^/COMMENTS_LAYOUT=KEEP_SPECIAL^
14466 Keep unchanged special form comments.
14467 This is the default.
14469 @item --comments-only
14470 @cindex @option{--comments-only} @command{gnatpp}
14471 Format just the comments.
14473 @cindex @option{^--no-separate-is^/NO_SEPARATE_IS^} (@command{gnatpp})
14474 @item ^--no-separate-is^/NO_SEPARATE_IS^
14475 Do not place the keyword @code{is} on a separate line in a subprogram body in
14476 case if the spec occupies more than one line.
14478 @cindex @option{^--separate-loop-then^/SEPARATE_LOOP_THEN^} (@command{gnatpp})
14479 @item ^--separate-loop-then^/SEPARATE_LOOP_THEN^
14480 Place the keyword @code{loop} in FOR and WHILE loop statements and the
14481 keyword @code{then} in IF statements on a separate line.
14483 @cindex @option{^--no-separate-loop-then^/NO_SEPARATE_LOOP_THEN^} (@command{gnatpp})
14484 @item ^--no-separate-loop-then^/NO_SEPARATE_LOOP_THEN^
14485 Do not place the keyword @code{loop} in FOR and WHILE loop statements and the
14486 keyword @code{then} in IF statements on a separate line. This option is
14487 incompatible with @option{^--separate-loop-then^/SEPARATE_LOOP_THEN^} option.
14489 @cindex @option{^--use-on-new-line^/USE_ON_NEW_LINE^} (@command{gnatpp})
14490 @item ^--use-on-new-line^/USE_ON_NEW_LINE^
14491 Start each USE clause in a context clause from a separate line.
14493 @cindex @option{^--insert-blank-lines^/INSERT_BLANK_LINES^} (@command{gnatpp})
14494 @item ^--insert-blank-lines^/INSERT_BLANK_LINES^
14495 Insert blank lines where appropriate (between bodies and other large
14498 @cindex @option{^--preserve-blank-lines^/PRESERVE_BLANK_LINES^} (@command{gnatpp})
14499 @item ^--preserve-blank-lines^/PRESERVE_BLANK_LINES^
14500 Preserve blank lines in the input. By default, gnatpp will squeeze
14501 multiple blank lines down to one.
14507 The @option{-c} switches are compatible with one another, except that
14508 the @option{-c0} switch disables all other comment formatting
14514 For the @option{/COMMENTS_LAYOUT} qualifier,
14515 The @option{GNAT_BEGINNING}, @option{REFORMAT}, and @option{DEFAULT}
14516 options are compatible with one another.
14519 @node General Text Layout Control
14520 @subsection General Text Layout Control
14523 These switches allow control over line length and indentation.
14526 @item ^-M@var{nnn}^/LINE_LENGTH_MAX=@var{nnn}^
14527 @cindex @option{^-M^/LINE_LENGTH^} (@command{gnatpp})
14528 Maximum line length, @var{nnn} from 32@dots{}256, the default value is 79
14530 @item ^-i@var{nnn}^/INDENTATION_LEVEL=@var{nnn}^
14531 @cindex @option{^-i^/INDENTATION_LEVEL^} (@command{gnatpp})
14532 Indentation level, @var{nnn} from 1@dots{}9, the default value is 3
14534 @item ^-cl@var{nnn}^/CONTINUATION_INDENT=@var{nnn}^
14535 @cindex @option{^-cl^/CONTINUATION_INDENT^} (@command{gnatpp})
14536 Indentation level for continuation lines (relative to the line being
14537 continued), @var{nnn} from 1@dots{}9.
14539 value is one less than the (normal) indentation level, unless the
14540 indentation is set to 1 (in which case the default value for continuation
14541 line indentation is also 1)
14544 @node Other Formatting Options
14545 @subsection Other Formatting Options
14548 These switches control other formatting not listed above.
14551 @item --decimal-grouping=@var{n}
14552 @cindex @option{--decimal-grouping} @command{gnatpp}
14553 Put underscores in decimal literals (numeric literals without a base)
14554 every @var{n} characters. If a literal already has one or more
14555 underscores, it is not modified. For example, with
14556 @code{--decimal-grouping=3}, @code{1000000} will be changed to
14559 @item --based-grouping=@var{n}
14560 @cindex @option{--based-grouping} @command{gnatpp}
14561 Same as @code{--decimal-grouping}, but for based literals. For
14562 example, with @code{--based-grouping=4}, @code{16#0001FFFE#} will be
14563 changed to @code{16#0001_FFFE#}.
14565 @item ^--RM-style-spacing^/RM_STYLE_SPACING^
14566 @cindex @option{^--RM-style-spacing^/RM_STYLE_SPACING^} (@command{gnatpp})
14567 Do not insert an extra blank before various occurrences of
14568 `(' and `:'. This also turns off alignment.
14570 @item ^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^
14571 @cindex @option{^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^} (@command{gnatpp})
14572 Insert a Form Feed character after a pragma Page.
14574 @item ^--call_threshold=@var{nnn}^/MAX_ACT=@var{nnn}^
14575 @cindex @option{^--call_threshold^/MAX_ACT^} (@command{gnatpp})
14576 If the number of parameter associations is greater than @var{nnn} and if at
14577 least one association uses named notation, start each association from
14578 a new line. If @var{nnn} is 0, no check for the number of associations
14579 is made; this is the default.
14581 @item ^--par_threshold=@var{nnn}^/MAX_PAR=@var{nnn}^
14582 @cindex @option{^--par_threshold^/MAX_PAR^} (@command{gnatpp})
14583 If the number of parameter specifications is greater than @var{nnn}
14584 (or equal to @var{nnn} in case of a function), start each specification from
14585 a new line. This feature is disabled by default.
14588 @node Setting the Source Search Path
14589 @subsection Setting the Source Search Path
14592 To define the search path for the input source file, @command{gnatpp}
14593 uses the same switches as the GNAT compiler, with the same effects:
14596 @item ^-I^/SEARCH=^@var{dir}
14597 @cindex @option{^-I^/SEARCH^} (@command{gnatpp})
14599 @item ^-I-^/NOCURRENT_DIRECTORY^
14600 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatpp})
14602 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE^=@var{path}
14603 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@command{gnatpp})
14605 @item ^--RTS^/RUNTIME_SYSTEM^=@var{path}
14606 @cindex @option{^--RTS^/RUNTIME_SYSTEM^} (@command{gnatpp})
14610 @node Output File Control
14611 @subsection Output File Control
14614 By default the output is sent to a file whose name is obtained by appending
14615 the ^@file{.pp}^@file{$PP}^ suffix to the name of the input file.
14616 If the file with this name already exists, it is overwritten.
14617 Thus if the input file is @file{^my_ada_proc.adb^MY_ADA_PROC.ADB^} then
14618 @command{gnatpp} will produce @file{^my_ada_proc.adb.pp^MY_ADA_PROC.ADB$PP^}
14620 The output may be redirected by the following switches:
14623 @item ^--output-dir=@var{dir}^/OUTPUT_DIR=@var{dir}^
14624 @cindex @option{^--output-dir^/OUTPUT_DIR^} (@command{gnatpp})
14625 Generate output file in directory @file{dir} with the same name as the input
14626 file. If @file{dir} is the same as the directory containing the input file,
14627 the input file is not processed; use @option{^-rnb^/REPLACE_NO_BACKUP^}
14628 if you want to update the input file in place.
14630 @item ^-pipe^/STANDARD_OUTPUT^
14631 @cindex @option{^-pipe^/STANDARD_OUTPUT^} (@command{gnatpp})
14632 Send the output to @code{Standard_Output}
14634 @item ^-o @var{output_file}^/OUTPUT=@var{output_file}^
14635 @cindex @option{^-o^/OUTPUT^} (@code{gnatpp})
14636 Write the output into @var{output_file}.
14637 If @var{output_file} already exists, @command{gnatpp} terminates without
14638 reading or processing the input file.
14640 @item ^-of ^/FORCED_OUTPUT=^@var{output_file}
14641 @cindex @option{^-of^/FORCED_OUTPUT^} (@command{gnatpp})
14642 Write the output into @var{output_file}, overwriting the existing file
14643 (if one is present).
14645 @item ^-r^/REPLACE^
14646 @cindex @option{^-r^/REPLACE^} (@command{gnatpp})
14647 Replace the input source file with the reformatted output, and copy the
14648 original input source into the file whose name is obtained by appending the
14649 ^@file{.npp}^@file{$NPP}^ suffix to the name of the input file.
14650 If a file with this name already exists, @command{gnatpp} terminates without
14651 reading or processing the input file.
14653 @item ^-rf^/OVERRIDING_REPLACE^
14654 @cindex @option{^-rf^/OVERRIDING_REPLACE^} (@code{gnatpp})
14655 Like @option{^-r^/REPLACE^} except that if the file with the specified name
14656 already exists, it is overwritten.
14658 @item ^-rnb^/REPLACE_NO_BACKUP^
14659 @cindex @option{^-rnb^/REPLACE_NO_BACKUP^} (@command{gnatpp})
14660 Replace the input source file with the reformatted output without
14661 creating any backup copy of the input source.
14663 @item ^--eol=@var{xxx}^/END_OF_LINE=@var{xxx}^
14664 @cindex @option{^--eol^/END_OF_LINE^} (@code{gnatpp})
14665 Specifies the line-ending style of the reformatted output file. The @var{xxx}
14666 ^string specified with the switch^option^ may be:
14668 @item ``@option{^dos^DOS^}'' MS DOS style, lines end with CR LF characters
14669 @item ``@option{^crlf^CRLF^}''
14670 the same as @option{^dos^DOS^}
14671 @item ``@option{^unix^UNIX^}'' UNIX style, lines end with LF character
14672 @item ``@option{^lf^LF^}''
14673 the same as @option{^unix^UNIX^}
14676 @item ^-W^/RESULT_ENCODING=^@var{e}
14677 @cindex @option{^-W^/RESULT_ENCODING=^} (@command{gnatpp})
14678 Specify the wide character encoding method for the input and output files.
14679 @var{e} is one of the following:
14687 Upper half encoding
14689 @item ^s^SHIFT_JIS^
14699 Brackets encoding (default value)
14705 Options @option{^-o^/OUTPUT^} and
14706 @option{^-of^/FORCED_OUTPUT^} are allowed only if the call to gnatpp
14707 contains only one file to reformat.
14709 @option{^--eol^/END_OF_LINE^}
14711 @option{^-W^/RESULT_ENCODING^}
14712 cannot be used together
14713 with @option{^-pipe^/STANDARD_OUTPUT^} option.
14715 @node Other gnatpp Switches
14716 @subsection Other @code{gnatpp} Switches
14719 The additional @command{gnatpp} switches are defined in this subsection.
14723 @cindex @option{--version} @command{gnatpp}
14724 Display copyright and version, then exit disregarding all other options.
14727 @cindex @option{--help} @command{gnatpp}
14728 Display usage, then exit disregarding all other options.
14730 @item -P @var{file}
14731 @cindex @option{-P} @command{gnatpp}
14732 Indicates the name of the project file that describes the set of sources
14733 to be processed. The exact set of argument sources depends on other options
14734 specified; see below.
14737 @cindex @option{-U} @command{gnatpp}
14738 If a project file is specified and no argument source is explicitly
14739 specified (either directly or by means of @option{-files} option), process
14740 all the units of the closure of the argument project. Otherwise this option
14743 @item -U @var{main_unit}
14744 If a project file is specified and no argument source is explicitly
14745 specified (either directly or by means of @option{-files} option), process
14746 the closure of units rooted at @var{main_unit}. Otherwise this option
14749 @item -X@var{name}=@var{value}
14750 @cindex @option{-X} @command{gnatpp}
14751 Indicates that external variable @var{name} in the argument project
14752 has the value @var{value}. Has no effect if no project is specified as
14755 @item --incremental
14756 @cindex @option{--incremental} @command{gnatpp}
14757 Incremental processing on a per-file basis. Source files are only
14758 processed if they have been modified, or if files they depend on have
14759 been modified. This is similar to the way gnatmake/gprbuild only
14760 compiles files that need to be recompiled.
14762 @item --pp-off=@var{xxx}
14763 @cindex @option{--pp-off} @command{gnatpp}
14764 Use @code{--xxx} as the command to turn off pretty printing, instead
14765 of the default @code{--!pp off}.
14767 @item --pp-on=@var{xxx}
14768 @cindex @option{--pp-on} @command{gnatpp}
14769 Use @code{--xxx} as the command to turn pretty printing back on, instead
14770 of the default @code{--!pp on}.
14773 @cindex @option{--pp-old} @command{gnatpp}
14774 Use the old formatting algorithms.
14776 @item ^-files @var{filename}^/FILES=@var{filename}^
14777 @cindex @option{^-files^/FILES^} (@code{gnatpp})
14778 Take the argument source files from the specified file. This file should be an
14779 ordinary text file containing file names separated by spaces or
14780 line breaks. You can use this switch more than once in the same call to
14781 @command{gnatpp}. You also can combine this switch with an explicit list of
14784 @item ^-j^/PROCESSES=^@var{n}
14785 @cindex @option{^-j^/PROCESSES^} (@command{gnatpp})
14786 Without @option{--incremental}, use @var{n} processes to carry out the
14787 tree creations (internal representations of the argument sources). On
14788 a multiprocessor machine this speeds up processing of big sets of
14789 argument sources. If @var{n} is 0, then the maximum number of parallel
14790 tree creations is the number of core processors on the platform. This
14791 option cannot be used together with @option{^-r^/REPLACE^},
14792 @option{^-rf^/OVERRIDING_REPLACE^} or
14793 @option{^-rnb^/REPLACE_NO_BACKUP^} option.
14795 With @option{--incremental}, use @var{n} @command{gnatpp} processes to
14796 perform pretty-printing in parallel. @var{n} = 0 means the same as
14797 above. In this case, @option{^-r^/REPLACE^},
14798 @option{^-rf^/OVERRIDING_REPLACE^} or
14799 @option{^-rnb^/REPLACE_NO_BACKUP^} options are allowed.
14801 @cindex @option{^-t^/TIME^} (@command{gnatpp})
14803 Print out execution time.
14805 @item ^-v^/VERBOSE^
14806 @cindex @option{^-v^/VERBOSE^} (@command{gnatpp})
14810 @cindex @option{^-q^/QUIET^} (@command{gnatpp})
14815 If a project file is specified and no argument source is explicitly
14816 specified (either directly or by means of @option{-files} option), and no
14817 @option{-U} is specified, then the set of processed sources is
14818 all the immediate units of the argument project.
14821 @node Formatting Rules
14822 @section Formatting Rules
14825 The following subsections show how @command{gnatpp} treats white space,
14826 comments, program layout, and name casing.
14827 They provide detailed descriptions of the switches shown above.
14830 * Disabling Pretty Printing::
14831 * White Space and Empty Lines::
14832 * Formatting Comments::
14836 @node Disabling Pretty Printing
14837 @subsection Disabling Pretty Printing
14840 Pretty printing is highly heuristic in nature, and sometimes doesn't
14841 do exactly what you want. If you wish to format a certain region of
14842 code by hand, you can turn off pretty printing in that region by
14843 surrounding it with special comments that start with @code{--!pp off}
14844 and @code{--!pp on}. The text in that region will then be reproduced
14845 verbatim in the output with no formatting.
14847 To disable pretty printing for the whole file, put @code{--!pp off} at
14848 the top, with no following @code{--!pp on}.
14850 The comments must appear on a line by themselves, with nothing
14851 preceding except spaces. The initial text of the comment must be
14852 exactly @code{--!pp off} or @code{--!pp on} (case sensitive), but may
14853 be followed by arbitrary additional text. For example:
14855 @smallexample @c ada
14857 package Interrupts is
14858 --!pp off -- turn off pretty printing so "Interrupt_Kind" lines up
14859 type Interrupt_Kind is
14860 (Asynchronous_Interrupt_Kind,
14861 Synchronous_Interrupt_Kind,
14862 Green_Interrupt_Kind);
14863 --!pp on -- reenable pretty printing
14869 You can specify different comment strings using the @code{--pp-off}
14870 and @code{--pp-on} switches. For example, if you say @code{gnatpp
14871 --pp-off=' pp-' *.ad?} then gnatpp will recognize comments of the form
14872 @code{-- pp-} instead of @code{--!pp off} for disabling pretty
14873 printing. Note that the leading @code{--} of the comment is not
14874 included in the argument to these switches.
14876 @node White Space and Empty Lines
14877 @subsection White Space and Empty Lines
14880 @command{gnatpp} does not have an option to control space characters.
14881 It will add or remove spaces according to the style illustrated by the
14882 examples in the @cite{Ada Reference Manual}.
14883 The output file will contain no lines with trailing white space.
14885 By default, a sequence of one or more blank lines in the input is
14886 converted to a single blank line in the output; multiple blank lines
14887 are squeezed down to one.
14888 The @option{^--preserve-blank-lines^/PRESERVE_BLANK_LINES^} option
14889 turns off the squeezing; each blank line in the input is copied
14891 The @option{^--insert-blank-lines^/INSERT_BLANK_LINES^} option
14892 causes additional blank lines to be inserted if not already
14893 present in the input (e.g. between bodies).
14895 @node Formatting Comments
14896 @subsection Formatting Comments
14899 Comments in Ada code are of two kinds:
14902 a @emph{whole-line comment}, which appears by itself (possibly preceded by
14903 white space) on a line
14906 an @emph{end-of-line comment}, which follows some other Ada code on
14911 A whole-line comment is indented according to the surrounding code,
14912 with some exceptions.
14913 Comments that start in column 1 are kept there.
14914 If possible, comments are not moved so far to the right that the maximum
14915 line length is exceeded.
14916 The @option{^-c0^/COMMENTS_LAYOUT=UNTOUCHED^} option
14917 turns off comment formatting.
14918 Special-form comments such as SPARK-style @code{--#...} are left alone.
14920 For an end-of-line comment, @command{gnatpp} tries to leave the same
14921 number of spaces between the end of the preceding Ada code and the
14922 beginning of the comment as appear in the original source.
14925 The @option{^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^} switch
14926 (GNAT style comment beginning) has the following
14931 For each whole-line comment that does not end with two hyphens,
14932 @command{gnatpp} inserts spaces if necessary after the starting two hyphens
14933 to ensure that there are at least two spaces between these hyphens and the
14934 first non-blank character of the comment.
14938 The @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} switch specifies that
14939 whole-line comments that form a paragraph will be filled in typical
14940 word processor style (that is, moving words between lines to make the
14941 lines other than the last similar in length ).
14944 The @option{--comments-only} switch specifies that only the comments
14945 are formatted; the rest of the program text is left alone. The
14946 comments are formatted according to the -c3 and -c4 switches; other
14947 formatting switches are ignored. For example, @option{--comments-only
14948 -c4} means to fill comment paragraphs, and do nothing else. Likewise,
14949 @option{--comments-only -c3} ensures comments start with at least two
14950 spaces after @code{--}, and @option{--comments-only -c3 -c4} does
14951 both. If @option{--comments-only} is given without @option{-c3} or
14952 @option{-c4}, then gnatpp doesn't format anything.
14955 @subsection Name Casing
14958 @command{gnatpp} always converts the usage occurrence of a (simple) name to
14959 the same casing as the corresponding defining identifier.
14961 You control the casing for defining occurrences via the
14962 @option{^-n^/NAME_CASING^} switch.
14964 With @option{-nD} (``as declared'', which is the default),
14967 With @option{/NAME_CASING=AS_DECLARED}, which is the default,
14969 defining occurrences appear exactly as in the source file
14970 where they are declared.
14971 The other ^values for this switch^options for this qualifier^ ---
14972 @option{^-nU^UPPER_CASE^},
14973 @option{^-nL^LOWER_CASE^},
14974 @option{^-nM^MIXED_CASE^} ---
14976 ^upper, lower, or mixed case, respectively^the corresponding casing^.
14977 If @command{gnatpp} changes the casing of a defining
14978 occurrence, it analogously changes the casing of all the
14979 usage occurrences of this name.
14981 If the defining occurrence of a name is not in the source compilation unit
14982 currently being processed by @command{gnatpp}, the casing of each reference to
14983 this name is changed according to the value of the @option{^-n^/NAME_CASING^}
14984 switch (subject to the dictionary file mechanism described below).
14985 Thus @command{gnatpp} acts as though the @option{^-n^/NAME_CASING^} switch
14987 casing for the defining occurrence of the name.
14990 @option{^-a@var{x}^/ATTRIBUTE^},
14991 @option{^-k@var{x}^/KEYWORD_CASING^},
14992 @option{^-ne@var{x}^/ENUM_CASING^},
14993 @option{^-nt@var{x}^/TYPE_CASING^},
14994 @option{^-nn@var{x}^/NUMBER_CASING^}, and
14995 @option{^-p@var{x}^/PRAGMA_CASING^}
14996 allow finer-grained control over casing for
14997 attributes, keywords, enumeration literals,
14998 types, named numbers and pragmas, respectively.
14999 @option{^-nt@var{x}^/TYPE_CASING^} covers subtypes and
15000 task and protected bodies as well.
15002 Some names may need to be spelled with casing conventions that are not
15003 covered by the upper-, lower-, and mixed-case transformations.
15004 You can arrange correct casing by placing such names in a
15005 @emph{dictionary file},
15006 and then supplying a @option{^-D^/DICTIONARY^} switch.
15007 The casing of names from dictionary files overrides
15008 any @option{^-n^/NAME_CASING^} switch.
15010 To handle the casing of Ada predefined names and the names from GNAT libraries,
15011 @command{gnatpp} assumes a default dictionary file.
15012 The name of each predefined entity is spelled with the same casing as is used
15013 for the entity in the @cite{Ada Reference Manual} (usually mixed case).
15014 The name of each entity in the GNAT libraries is spelled with the same casing
15015 as is used in the declaration of that entity.
15017 The @w{@option{^-D-^/SPECIFIC_CASING^}} switch suppresses the use of
15018 the default dictionary file. Instead, the casing for predefined and
15019 GNAT-defined names will be established by the
15020 @option{^-n^/NAME_CASING^} switch or explicit dictionary files. For
15021 example, by default the names @code{Ada.Text_IO} and
15022 @code{GNAT.OS_Lib} will appear as just shown, even in the presence of
15023 a @option{^-nU^/NAME_CASING=UPPER_CASE^} switch. To ensure that even
15024 such names are rendered in uppercase, additionally supply the
15025 @w{@option{^-D-^/SPECIFIC_CASING^}} switch (or else place these names
15026 in upper case in a dictionary file).
15028 A dictionary file is a plain text file; each line in this file can be
15029 either a blank line (containing only space characters), an Ada comment
15030 line, or the specification of exactly one @emph{casing schema}.
15032 A casing schema is a string that has the following syntax:
15036 @var{casing_schema} ::= @var{identifier} | *@var{simple_identifier}*
15038 @var{simple_identifier} ::= @var{letter}@{@var{letter_or_digit}@}
15043 (See @cite{Ada Reference Manual}, Section 2.3) for the definition of the
15044 @var{identifier} lexical element and the @var{letter_or_digit} category.)
15046 The casing schema string can be followed by white space and/or an Ada-style
15047 comment; any amount of white space is allowed before the string.
15049 If a dictionary file is passed as
15051 the value of a @option{-D@var{file}} switch
15054 an option to the @option{/DICTIONARY} qualifier
15057 simple name and every identifier, @command{gnatpp} checks if the dictionary
15058 defines the casing for the name or for some of its parts (the term ``subword''
15059 is used below to denote the part of a name which is delimited by ``_'' or by
15060 the beginning or end of the word and which does not contain any ``_'' inside):
15064 if the whole name is in the dictionary, @command{gnatpp} uses for this name
15065 the casing defined by the dictionary; no subwords are checked for this word
15068 for every subword @command{gnatpp} checks if the dictionary contains the
15069 corresponding string of the form @code{*@var{simple_identifier}*},
15070 and if it does, the casing of this @var{simple_identifier} is used
15074 if the whole name does not contain any ``_'' inside, and if for this name
15075 the dictionary contains two entries - one of the form @var{identifier},
15076 and another - of the form *@var{simple_identifier}*, then the first one
15077 is applied to define the casing of this name
15080 if more than one dictionary file is passed as @command{gnatpp} switches, each
15081 dictionary adds new casing exceptions and overrides all the existing casing
15082 exceptions set by the previous dictionaries
15085 when @command{gnatpp} checks if the word or subword is in the dictionary,
15086 this check is not case sensitive
15090 For example, suppose we have the following source to reformat:
15092 @smallexample @c ada
15095 name1 : integer := 1;
15096 name4_name3_name2 : integer := 2;
15097 name2_name3_name4 : Boolean;
15100 name2_name3_name4 := name4_name3_name2 > name1;
15106 And suppose we have two dictionaries:
15123 If @command{gnatpp} is called with the following switches:
15127 @command{gnatpp -nM -D dict1 -D dict2 test.adb}
15130 @command{gnatpp test.adb /NAME_CASING=MIXED_CASE /DICTIONARY=(dict1, dict2)}
15135 then we will get the following name casing in the @command{gnatpp} output:
15137 @smallexample @c ada
15140 NAME1 : Integer := 1;
15141 Name4_NAME3_Name2 : Integer := 2;
15142 Name2_NAME3_Name4 : Boolean;
15145 Name2_NAME3_Name4 := Name4_NAME3_Name2 > NAME1;
15151 @ifclear FSFEDITION
15153 @c *********************************
15154 @node The Ada-to-XML converter gnat2xml
15155 @chapter The Ada-to-XML converter @command{gnat2xml}
15157 @cindex XML generation
15160 The @command{gnat2xml} tool is an ASIS-based utility that converts
15161 Ada source code into XML.
15164 * Switches for gnat2xml::
15166 * Structure of the XML::
15169 @node Switches for gnat2xml
15170 @section Switches for @command{gnat2xml}
15173 @command{gnat2xml} takes Ada source code as input, and produces XML
15174 that conforms to the schema.
15179 gnat2xml [options] filenames [-files filename] [-cargs gcc_switches]
15186 --help -- generate usage information and quit, ignoring all other options
15187 --version -- print version and quit, ignoring all other options
15189 -P @file{file} -- indicates the name of the project file that describes
15190 the set of sources to be processed. The exact set of argument
15191 sources depends on other options specified, see below.
15193 -U -- if a project file is specified and no argument source is explicitly
15194 specified, process all the units of the closure of the argument project.
15195 Otherwise this option has no effect.
15197 -U @var{main_unit} -- if a project file is specified and no argument source
15198 is explicitly specified (either directly or by means of @option{-files}
15199 option), process the closure of units rooted at @var{main_unit}.
15200 Otherwise this option has no effect.
15202 -X@var{name}=@var{value} -- indicates that external variable @var{name} in
15203 the argument project has the value @var{value}. Has no effect if no
15204 project is specified as tool argument.
15206 --incremental -- incremental processing on a per-file basis. Source files are
15207 only processed if they have been modified, or if files they depend
15208 on have been modified. This is similar to the way gnatmake/gprbuild
15209 only compiles files that need to be recompiled.
15211 --output-dir=@var{dir} -- generate one .xml file for each Ada source file, in
15212 directory @file{dir}. (Default is to generate the XML to standard
15216 directories to search for dependencies
15217 You can also set the ADA_INCLUDE_PATH environment variable for this.
15219 --compact -- debugging version, with interspersed source, and a more
15220 compact representation of "sloc". This version does not conform
15223 -files=filename - the name of a text file containing a list
15224 of Ada source files to process
15229 -cargs ... -- options to pass to gcc
15233 If a project file is specified and no argument source is explicitly
15234 specified, and no @option{-U} is specified, then the set of processed
15235 sources is all the immediate units of the argument project.
15240 gnat2xml -v -output-dir=xml-files *.ad[sb]
15244 The above will create *.xml files in the @file{xml-files} subdirectory.
15245 For example, if there is an Ada package Mumble.Dumble, whose spec and
15246 body source code lives in mumble-dumble.ads and mumble-dumble.adb,
15247 the above will produce xml-files/mumble-dumble.ads.xml and
15248 xml-files/mumble-dumble.adb.xml.
15250 @node Other Programs
15251 @section Other Programs
15254 The distribution includes two other programs that are related to
15255 @command{gnat2xml}:
15257 @command{gnat2xsd} is the schema generator, which generates the schema
15258 to standard output, based on the structure of Ada as encoded by
15259 ASIS. You don't need to run @command{gnat2xsd} in order to use
15260 @command{gnat2xml}. To generate the schema, type:
15263 gnat2xsd > ada-schema.xsd
15267 @command{gnat2xml} generates XML files that will validate against
15268 @file{ada-schema.xsd}.
15270 @command{xml2gnat} is a back-translator that translates the XML back
15271 into Ada source code. The Ada generated by @command{xml2gnat} has
15272 identical semantics to the original Ada code passed to
15273 @command{gnat2xml}. It is not textually identical, however --- for
15274 example, no attempt is made to preserve the original indentation.
15276 @node Structure of the XML
15277 @section Structure of the XML
15280 The primary documentation for the structure of the XML generated by
15281 @command{gnat2xml} is the schema (see @command{gnat2xsd} above). The
15282 following documentation gives additional details needed to understand
15283 the schema and therefore the XML.
15285 The elements listed under Defining Occurrences, Usage Occurrences, and
15286 Other Elements represent the syntactic structure of the Ada program.
15287 Element names are given in lower case, with the corresponding element
15288 type Capitalized_Like_This. The element and element type names are
15289 derived directly from the ASIS enumeration type Flat_Element_Kinds,
15290 declared in Asis.Extensions.Flat_Kinds, with the leading ``An_'' or ``A_''
15291 removed. For example, the ASIS enumeration literal
15292 An_Assignment_Statement corresponds to the XML element
15293 assignment_statement of XML type Assignment_Statement.
15295 To understand the details of the schema and the corresponding XML, it is
15296 necessary to understand the ASIS standard, as well as the GNAT-specific
15299 A defining occurrence is an identifier (or character literal or operator
15300 symbol) declared by a declaration. A usage occurrence is an identifier
15301 (or ...) that references such a declared entity. For example, in:
15304 type T is range 1..10;
15305 X, Y : constant T := 1;
15309 The first ``T'' is the defining occurrence of a type. The ``X'' is the
15310 defining occurrence of a constant, as is the ``Y'', and the second ``T'' is
15311 a usage occurrence referring to the defining occurrence of T.
15313 Each element has a 'sloc' (source location), and subelements for each
15314 syntactic subtree, reflecting the Ada grammar as implemented by ASIS.
15315 The types of subelements are as defined in the ASIS standard. For
15316 example, for the right-hand side of an assignment_statement we have
15317 the following comment in asis-statements.ads:
15320 ------------------------------------------------------------------------------
15321 -- 18.3 function Assignment_Expression
15322 ------------------------------------------------------------------------------
15324 function Assignment_Expression
15325 (Statement : Asis.Statement)
15326 return Asis.Expression;
15328 ------------------------------------------------------------------------------
15330 -- Returns the expression from the right hand side of the assignment.
15332 -- Returns Element_Kinds:
15337 The corresponding sub-element of type Assignment_Statement is:
15340 <xsd:element name="assignment_expression_q" type="Expression_Class"/>
15344 where Expression_Class is defined by an xsd:choice of all the
15345 various kinds of expression.
15347 The 'sloc' of each element indicates the starting and ending line and
15348 column numbers. Column numbers are character counts; that is, a tab
15349 counts as 1, not as however many spaces it might expand to.
15351 Subelements of type Element have names ending in ``_q'' (for ASIS
15352 ``Query''), and those of type Element_List end in ``_ql'' (``Query returning
15355 Some subelements are ``Boolean''. For example, Private_Type_Definition
15356 has has_abstract_q and has_limited_q, to indicate whether those
15357 keywords are present, as in @code{type T is abstract limited
15358 private;}. False is represented by a Nil_Element. True is represented
15359 by an element type specific to that query (for example, Abstract and
15362 The root of the tree is a Compilation_Unit, with attributes:
15366 unit_kind, unit_class, and unit_origin. These are strings that match the
15367 enumeration literals of types Unit_Kinds, Unit_Classes, and Unit_Origins
15371 unit_full_name is the full expanded name of the unit, starting from a
15372 root library unit. So for @code{package P.Q.R is ...},
15373 @code{unit_full_name="P.Q.R"}. Same for @code{separate (P.Q) package R is ...}.
15376 def_name is the same as unit_full_name for library units; for subunits,
15377 it is just the simple name.
15380 source_file is the name of the Ada source file. For example, for
15381 the spec of @code{P.Q.R}, @code{source_file="p-q-r.ads"}. This allows one to
15382 interpret the source locations --- the ``sloc'' of all elements
15383 within this Compilation_Unit refers to line and column numbers
15384 within the named file.
15388 Defining occurrences have these attributes:
15392 def_name is the simple name of the declared entity, as written in the Ada
15396 def is a unique URI of the form:
15398 ada://kind/fully/qualified/name
15402 kind indicates the kind of Ada entity being declared (see below), and
15404 fully/qualified/name, is the fully qualified name of the Ada
15405 entity, with each of ``fully'', ``qualified'', and ``name'' being
15406 mangled for uniqueness. We do not document the mangling
15407 algorithm, which is subject to change; we just guarantee that the
15408 names are unique in the face of overloading.
15411 type is the type of the declared object, or @code{null} for
15412 declarations of things other than objects.
15416 Usage occurrences have these attributes:
15420 ref_name is the same as the def_name of the corresponding defining
15421 occurrence. This attribute is not of much use, because of
15422 overloading; use ref for lookups, instead.
15425 ref is the same as the def of the corresponding defining
15430 In summary, @code{def_name} and @code{ref_name} are as in the source
15431 code of the declaration, possibly overloaded, whereas @code{def} and
15432 @code{ref} are unique-ified.
15434 Literal elements have this attribute:
15438 lit_val is the value of the literal as written in the source text,
15439 appropriately escaped (e.g. @code{"} ---> @code{"}). This applies
15440 only to numeric and string literals. Enumeration literals in Ada are
15441 not really "literals" in the usual sense; they are usage occurrences,
15442 and have ref_name and ref as described above. Note also that string
15443 literals used as operator symbols are treated as defining or usage
15444 occurrences, not as literals.
15448 Elements that can syntactically represent names and expressions (which
15449 includes usage occurrences, plus function calls and so forth) have this
15454 type. If the element represents an expression or the name of an object,
15455 'type' is the 'def' for the defining occurrence of the type of that
15456 expression or name. Names of other kinds of entities, such as package
15457 names and type names, do not have a type in Ada; these have type="null"
15462 Pragma elements have this attribute:
15466 pragma_name is the name of the pragma. For language-defined pragmas, the
15467 pragma name is redundant with the element kind (for example, an
15468 assert_pragma element necessarily has pragma_name="Assert"). However, all
15469 implementation-defined pragmas are lumped together in ASIS as a single
15470 element kind (for example, the GNAT-specific pragma Unreferenced is
15471 represented by an implementation_defined_pragma element with
15472 pragma_name="Unreferenced").
15476 Defining occurrences of formal parameters and generic formal objects have this
15481 mode indicates that the parameter is of mode 'in', 'in out', or 'out'.
15485 All elements other than Not_An_Element have this attribute:
15489 checks is a comma-separated list of run-time checks that are needed
15490 for that element. The possible checks are: do_accessibility_check,
15491 do_discriminant_check,do_division_check,do_length_check,
15492 do_overflow_check,do_range_check,do_storage_check,do_tag_check.
15496 The "kind" part of the "def" and "ref" attributes is taken from the ASIS
15497 enumeration type Flat_Declaration_Kinds, declared in
15498 Asis.Extensions.Flat_Kinds, with the leading "An_" or "A_" removed, and
15499 any trailing "_Declaration" or "_Specification" removed. Thus, the
15500 possible kinds are as follows:
15507 tagged_incomplete_type
15518 enumeration_literal
15522 generalized_iterator
15532 expression_function
15540 generic_package_renaming
15541 generic_procedure_renaming
15542 generic_function_renaming
15548 procedure_body_stub
15552 protected_body_stub
15558 package_instantiation
15559 procedure_instantiation
15560 function_instantiation
15563 formal_incomplete_type
15567 formal_package_declaration_with_box
15573 @ifclear FSFEDITION
15574 @c *********************************
15575 @node The GNAT Metrics Tool gnatmetric
15576 @chapter The GNAT Metrics Tool @command{gnatmetric}
15578 @cindex Metric tool
15581 ^The @command{gnatmetric} tool^@command{GNAT METRIC}^ is an ASIS-based utility
15582 for computing various program metrics.
15583 It takes an Ada source file as input and generates a file containing the
15584 metrics data as output. Various switches control which
15585 metrics are computed and output.
15588 * Switches for gnatmetric::
15591 To compute program metrics, @command{gnatmetric} invokes the Ada
15592 compiler and generates and uses the ASIS tree for the input source;
15593 thus the input must be legal Ada code, and the tool should have all the
15594 information needed to compile the input source. To provide this information,
15595 you may specify as a tool parameter the project file the input source belongs to
15596 (or you may call @command{gnatmetric}
15597 through the @command{gnat} driver (see @ref{The GNAT Driver and
15598 Project Files}). Another possibility is to specify the source search
15599 path and needed configuration files in @option{-cargs} section of @command{gnatmetric}
15600 call, see the description of the @command{gnatmetric} switches below.
15602 The @command{gnatmetric} command has the form
15605 @c $ gnatmetric @ovar{switches} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
15606 @c Expanding @ovar macro inline (explanation in macro def comments)
15607 $ gnatmetric @r{[}@var{switches}@r{]} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
15614 @var{switches} specify the metrics to compute and define the destination for
15618 Each @var{filename} is the name (including the extension) of a source
15619 file to process. ``Wildcards'' are allowed, and
15620 the file name may contain path information.
15621 If no @var{filename} is supplied, then the @var{switches} list must contain
15623 @option{-files} switch (@pxref{Other gnatmetric Switches}).
15624 Including both a @option{-files} switch and one or more
15625 @var{filename} arguments is permitted.
15628 @samp{@var{gcc_switches}} is a list of switches for
15629 @command{gcc}. They will be passed on to all compiler invocations made by
15630 @command{gnatmetric} to generate the ASIS trees. Here you can provide
15631 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
15632 and use the @option{-gnatec} switch to set the configuration file,
15633 use the @option{-gnat05} switch if sources should be compiled in
15637 @node Switches for gnatmetric
15638 @section Switches for @command{gnatmetric}
15641 The following subsections describe the various switches accepted by
15642 @command{gnatmetric}, organized by category.
15645 * Output Files Control::
15646 * Disable Metrics For Local Units::
15647 * Specifying a set of metrics to compute::
15648 * Other gnatmetric Switches::
15650 * Generate project-wide metrics::
15654 @node Output Files Control
15655 @subsection Output File Control
15656 @cindex Output file control in @command{gnatmetric}
15659 @command{gnatmetric} has two output formats. It can generate a
15660 textual (human-readable) form, and also XML. By default only textual
15661 output is generated.
15663 When generating the output in textual form, @command{gnatmetric} creates
15664 for each Ada source file a corresponding text file
15665 containing the computed metrics, except for the case when the set of metrics
15666 specified by gnatmetric parameters consists only of metrics that are computed
15667 for the whole set of analyzed sources, but not for each Ada source.
15668 By default, the name of the file containing metric information for a source
15669 is obtained by appending the ^@file{.metrix}^@file{$METRIX}^ suffix to the
15670 name of the input source file. If not otherwise specified and no project file
15671 is specified as @command{gnatmetric} option this file is placed in the same
15672 directory as where the source file is located. If @command{gnatmetric} has a
15673 project file as its parameter, it places all the generated files in the
15674 object directory of the project (or in the project source directory if the
15675 project does not define an objects directory), if @option{--subdirs} option
15676 is specified, the files are placed in the subrirectory of this directory
15677 specified by this option.
15679 All the output information generated in XML format is placed in a single
15680 file. By default the name of this file is ^@file{metrix.xml}^@file{METRIX$XML}^.
15681 If not otherwise specified and if no project file is specified
15682 as @command{gnatmetric} option this file is placed in the
15685 Some of the computed metrics are summed over the units passed to
15686 @command{gnatmetric}; for example, the total number of lines of code.
15687 By default this information is sent to @file{stdout}, but a file
15688 can be specified with the @option{-og} switch.
15690 The following switches control the @command{gnatmetric} output:
15693 @cindex @option{^-x^/XML^} (@command{gnatmetric})
15695 Generate the XML output
15697 @cindex @option{^-xs^/XSD^} (@command{gnatmetric})
15699 Generate the XML output and the XML schema file that describes the structure
15700 of the XML metric report, this schema is assigned to the XML file. The schema
15701 file has the same name as the XML output file with @file{.xml} suffix replaced
15704 @cindex @option{^-nt^/NO_TEXT^} (@command{gnatmetric})
15705 @item ^-nt^/NO_TEXT^
15706 Do not generate the output in text form (implies @option{^-x^/XML^})
15708 @cindex @option{^-d^/DIRECTORY^} (@command{gnatmetric})
15709 @item ^-d @var{output_dir}^/DIRECTORY=@var{output_dir}^
15710 Put text files with detailed metrics into @var{output_dir}
15712 @cindex @option{^-o^/SUFFIX_DETAILS^} (@command{gnatmetric})
15713 @item ^-o @var{file_suffix}^/SUFFIX_DETAILS=@var{file_suffix}^
15714 Use @var{file_suffix}, instead of ^@file{.metrix}^@file{$METRIX}^
15715 in the name of the output file.
15717 @cindex @option{^-og^/GLOBAL_OUTPUT^} (@command{gnatmetric})
15718 @item ^-og @var{file_name}^/GLOBAL_OUTPUT=@var{file_name}^
15719 Put global metrics into @var{file_name}
15721 @cindex @option{^-ox^/XML_OUTPUT^} (@command{gnatmetric})
15722 @item ^-ox @var{file_name}^/XML_OUTPUT=@var{file_name}^
15723 Put the XML output into @var{file_name} (also implies @option{^-x^/XML^})
15725 @cindex @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} (@command{gnatmetric})
15726 @item ^-sfn^/SHORT_SOURCE_FILE_NAME^
15727 Use ``short'' source file names in the output. (The @command{gnatmetric}
15728 output includes the name(s) of the Ada source file(s) from which the metrics
15729 are computed. By default each name includes the absolute path. The
15730 @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} switch causes @command{gnatmetric}
15731 to exclude all directory information from the file names that are output.)
15735 @node Disable Metrics For Local Units
15736 @subsection Disable Metrics For Local Units
15737 @cindex Disable Metrics For Local Units in @command{gnatmetric}
15740 @command{gnatmetric} relies on the GNAT compilation model @minus{}
15742 unit per one source file. It computes line metrics for the whole source
15743 file, and it also computes syntax
15744 and complexity metrics for the file's outermost unit.
15746 By default, @command{gnatmetric} will also compute all metrics for certain
15747 kinds of locally declared program units:
15751 subprogram (and generic subprogram) bodies;
15754 package (and generic package) specs and bodies;
15757 task object and type specifications and bodies;
15760 protected object and type specifications and bodies.
15764 These kinds of entities will be referred to as
15765 @emph{eligible local program units}, or simply @emph{eligible local units},
15766 @cindex Eligible local unit (for @command{gnatmetric})
15767 in the discussion below.
15769 Note that a subprogram declaration, generic instantiation,
15770 or renaming declaration only receives metrics
15771 computation when it appear as the outermost entity
15774 Suppression of metrics computation for eligible local units can be
15775 obtained via the following switch:
15778 @cindex @option{^-nolocal^/SUPPRESS^} (@command{gnatmetric})
15779 @item ^-nolocal^/SUPPRESS=LOCAL_DETAILS^
15780 Do not compute detailed metrics for eligible local program units
15784 @node Specifying a set of metrics to compute
15785 @subsection Specifying a set of metrics to compute
15788 By default all the metrics are computed and reported. The switches
15789 described in this subsection allow you to control, on an individual
15790 basis, whether metrics are computed and
15791 reported. If at least one positive metric
15792 switch is specified (that is, a switch that defines that a given
15793 metric or set of metrics is to be computed), then only
15794 explicitly specified metrics are reported.
15797 * Line Metrics Control::
15798 * Syntax Metrics Control::
15799 * Complexity Metrics Control::
15800 * Coupling Metrics Control::
15803 @node Line Metrics Control
15804 @subsubsection Line Metrics Control
15805 @cindex Line metrics control in @command{gnatmetric}
15808 For any (legal) source file, and for each of its
15809 eligible local program units, @command{gnatmetric} computes the following
15814 the total number of lines;
15817 the total number of code lines (i.e., non-blank lines that are not comments)
15820 the number of comment lines
15823 the number of code lines containing end-of-line comments;
15826 the comment percentage: the ratio between the number of lines that contain
15827 comments and the number of all non-blank lines, expressed as a percentage;
15830 the number of empty lines and lines containing only space characters and/or
15831 format effectors (blank lines)
15834 the average number of code lines in subprogram bodies, task bodies, entry
15835 bodies and statement sequences in package bodies (this metric is only computed
15836 across the whole set of the analyzed units)
15841 @command{gnatmetric} sums the values of the line metrics for all the
15842 files being processed and then generates the cumulative results. The tool
15843 also computes for all the files being processed the average number of code
15846 You can use the following switches to select the specific line metrics
15847 to be computed and reported.
15850 @cindex @option{^--lines@var{x}^/LINE_COUNT_METRICS^} (@command{gnatmetric})
15853 @cindex @option{--no-lines@var{x}}
15856 @item ^--lines-all^/LINE_COUNT_METRICS=ALL^
15857 Report all the line metrics
15859 @item ^--no-lines-all^/LINE_COUNT_METRICS=NONE^
15860 Do not report any of line metrics
15862 @item ^--lines^/LINE_COUNT_METRICS=ALL_LINES^
15863 Report the number of all lines
15865 @item ^--no-lines^/LINE_COUNT_METRICS=NOALL_LINES^
15866 Do not report the number of all lines
15868 @item ^--lines-code^/LINE_COUNT_METRICS=CODE_LINES^
15869 Report the number of code lines
15871 @item ^--no-lines-code^/LINE_COUNT_METRICS=NOCODE_LINES^
15872 Do not report the number of code lines
15874 @item ^--lines-comment^/LINE_COUNT_METRICS=COMMENT_LINES^
15875 Report the number of comment lines
15877 @item ^--no-lines-comment^/LINE_COUNT_METRICS=NOCOMMENT_LINES^
15878 Do not report the number of comment lines
15880 @item ^--lines-eol-comment^/LINE_COUNT_METRICS=CODE_COMMENT_LINES^
15881 Report the number of code lines containing
15882 end-of-line comments
15884 @item ^--no-lines-eol-comment^/LINE_COUNT_METRICS=NOCODE_COMMENT_LINES^
15885 Do not report the number of code lines containing
15886 end-of-line comments
15888 @item ^--lines-ratio^/LINE_COUNT_METRICS=COMMENT_PERCENTAGE^
15889 Report the comment percentage in the program text
15891 @item ^--no-lines-ratio^/LINE_COUNT_METRICS=NOCOMMENT_PERCENTAGE^
15892 Do not report the comment percentage in the program text
15894 @item ^--lines-blank^/LINE_COUNT_METRICS=BLANK_LINES^
15895 Report the number of blank lines
15897 @item ^--no-lines-blank^/LINE_COUNT_METRICS=NOBLANK_LINES^
15898 Do not report the number of blank lines
15900 @item ^--lines-average^/LINE_COUNT_METRICS=AVERAGE_BODY_LINES^
15901 Report the average number of code lines in subprogram bodies, task bodies,
15902 entry bodies and statement sequences in package bodies. The metric is computed
15903 and reported for the whole set of processed Ada sources only.
15905 @item ^--no-lines-average^/LINE_COUNT_METRICS=NOAVERAGE_BODY_LINES^
15906 Do not report the average number of code lines in subprogram bodies,
15907 task bodies, entry bodies and statement sequences in package bodies.
15911 @node Syntax Metrics Control
15912 @subsubsection Syntax Metrics Control
15913 @cindex Syntax metrics control in @command{gnatmetric}
15916 @command{gnatmetric} computes various syntactic metrics for the
15917 outermost unit and for each eligible local unit:
15920 @item LSLOC (``Logical Source Lines Of Code'')
15921 The total number of declarations and the total number of statements. Note
15922 that the definition of declarations is the one given in the reference
15926 ``Each of the following is defined to be a declaration: any basic_declaration;
15927 an enumeration_literal_specification; a discriminant_specification;
15928 a component_declaration; a loop_parameter_specification; a
15929 parameter_specification; a subprogram_body; an entry_declaration;
15930 an entry_index_specification; a choice_parameter_specification;
15931 a generic_formal_parameter_declaration.''
15933 This means for example that each enumeration literal adds one to the count,
15934 as well as each subprogram parameter.
15936 Thus the results from this metric will be significantly greater than might
15937 be expected from a naive view of counting semicolons.
15939 @item Maximal static nesting level of inner program units
15941 @cite{Ada Reference Manual}, 10.1(1), ``A program unit is either a
15942 package, a task unit, a protected unit, a
15943 protected entry, a generic unit, or an explicitly declared subprogram other
15944 than an enumeration literal.''
15946 @item Maximal nesting level of composite syntactic constructs
15947 This corresponds to the notion of the
15948 maximum nesting level in the GNAT built-in style checks
15949 (@pxref{Style Checking})
15953 For the outermost unit in the file, @command{gnatmetric} additionally computes
15954 the following metrics:
15957 @item Public subprograms
15958 This metric is computed for package specs. It is the
15959 number of subprograms and generic subprograms declared in the visible
15960 part (including the visible part of nested packages, protected objects, and
15963 @item All subprograms
15964 This metric is computed for bodies and subunits. The
15965 metric is equal to a total number of subprogram bodies in the compilation
15967 Neither generic instantiations nor renamings-as-a-body nor body stubs
15968 are counted. Any subprogram body is counted, independently of its nesting
15969 level and enclosing constructs. Generic bodies and bodies of protected
15970 subprograms are counted in the same way as ``usual'' subprogram bodies.
15973 This metric is computed for package specs and
15974 generic package declarations. It is the total number of types
15975 that can be referenced from outside this compilation unit, plus the
15976 number of types from all the visible parts of all the visible generic
15977 packages. Generic formal types are not counted. Only types, not subtypes,
15981 Along with the total number of public types, the following
15982 types are counted and reported separately:
15989 Root tagged types (abstract, non-abstract, private, non-private). Type
15990 extensions are @emph{not} counted
15993 Private types (including private extensions)
16004 This metric is computed for any compilation unit. It is equal to the total
16005 number of the declarations of different types given in the compilation unit.
16006 The private and the corresponding full type declaration are counted as one
16007 type declaration. Incomplete type declarations and generic formal types
16009 No distinction is made among different kinds of types (abstract,
16010 private etc.); the total number of types is computed and reported.
16015 By default, all the syntax metrics are computed and reported. You can use the
16016 following switches to select specific syntax metrics.
16020 @cindex @option{^--syntax@var{x}^/SYNTAX_METRICS^} (@command{gnatmetric})
16023 @cindex @option{--no-syntax@var{x}} (@command{gnatmetric})
16026 @item ^--syntax-all^/SYNTAX_METRICS=ALL^
16027 Report all the syntax metrics
16029 @item ^--no-syntax-all^/SYNTAX_METRICS=NONE^
16030 Do not report any of syntax metrics
16032 @item ^--declarations^/SYNTAX_METRICS=DECLARATIONS^
16033 Report the total number of declarations
16035 @item ^--no-declarations^/SYNTAX_METRICS=NODECLARATIONS^
16036 Do not report the total number of declarations
16038 @item ^--statements^/SYNTAX_METRICS=STATEMENTS^
16039 Report the total number of statements
16041 @item ^--no-statements^/SYNTAX_METRICS=NOSTATEMENTS^
16042 Do not report the total number of statements
16044 @item ^--public-subprograms^/SYNTAX_METRICS=PUBLIC_SUBPROGRAMS^
16045 Report the number of public subprograms in a compilation unit
16047 @item ^--no-public-subprograms^/SYNTAX_METRICS=NOPUBLIC_SUBPROGRAMS^
16048 Do not report the number of public subprograms in a compilation unit
16050 @item ^--all-subprograms^/SYNTAX_METRICS=ALL_SUBPROGRAMS^
16051 Report the number of all the subprograms in a compilation unit
16053 @item ^--no-all-subprograms^/SYNTAX_METRICS=NOALL_SUBPROGRAMS^
16054 Do not report the number of all the subprograms in a compilation unit
16056 @item ^--public-types^/SYNTAX_METRICS=PUBLIC_TYPES^
16057 Report the number of public types in a compilation unit
16059 @item ^--no-public-types^/SYNTAX_METRICS=NOPUBLIC_TYPES^
16060 Do not report the number of public types in a compilation unit
16062 @item ^--all-types^/SYNTAX_METRICS=ALL_TYPES^
16063 Report the number of all the types in a compilation unit
16065 @item ^--no-all-types^/SYNTAX_METRICS=NOALL_TYPES^
16066 Do not report the number of all the types in a compilation unit
16068 @item ^--unit-nesting^/SYNTAX_METRICS=UNIT_NESTING^
16069 Report the maximal program unit nesting level
16071 @item ^--no-unit-nesting^/SYNTAX_METRICS=UNIT_NESTING_OFF^
16072 Do not report the maximal program unit nesting level
16074 @item ^--construct-nesting^/SYNTAX_METRICS=CONSTRUCT_NESTING^
16075 Report the maximal construct nesting level
16077 @item ^--no-construct-nesting^/SYNTAX_METRICS=NOCONSTRUCT_NESTING^
16078 Do not report the maximal construct nesting level
16082 @node Complexity Metrics Control
16083 @subsubsection Complexity Metrics Control
16084 @cindex Complexity metrics control in @command{gnatmetric}
16087 For a program unit that is an executable body (a subprogram body (including
16088 generic bodies), task body, entry body or a package body containing
16089 its own statement sequence) @command{gnatmetric} computes the following
16090 complexity metrics:
16094 McCabe cyclomatic complexity;
16097 McCabe essential complexity;
16100 maximal loop nesting level;
16103 extra exit points (for subprograms);
16107 The McCabe cyclomatic complexity metric is defined
16108 in @url{http://www.mccabe.com/pdf/mccabe-nist235r.pdf}
16110 According to McCabe, both control statements and short-circuit control forms
16111 should be taken into account when computing cyclomatic complexity.
16112 For Ada 2012 we have also take into account conditional expressions
16113 and quantified expressions. For each body, we compute three metric values:
16117 the complexity introduced by control
16118 statements only, without taking into account short-circuit forms,
16121 the complexity introduced by short-circuit control forms only, and
16125 cyclomatic complexity, which is the sum of these two values.
16130 The cyclomatic complexity is also computed for Ada 2012 expression functions.
16131 An expression function cannot have statements as its components, so only one
16132 metric value is computed as a cyclomatic complexity of an expression function.
16134 The origin of cyclomatic complexity metric is the need to estimate the number
16135 of independent paths in the control flow graph that in turn gives the number
16136 of tests needed to satisfy paths coverage testing completeness criterion.
16137 Considered from the testing point of view, a static Ada @code{loop} (that is,
16138 the @code{loop} statement having static subtype in loop parameter
16139 specification) does not add to cyclomatic complexity. By providing
16140 @option{^--no-static-loop^NO_STATIC_LOOP^} option a user
16141 may specify that such loops should not be counted when computing the
16142 cyclomatic complexity metric
16144 The Ada essential complexity metric is a McCabe cyclomatic complexity metric
16145 counted for the code that is reduced by excluding all the pure structural Ada
16146 control statements. An compound statement is considered as a non-structural
16147 if it contains a @code{raise} or @code{return} statement as it subcomponent,
16148 or if it contains a @code{goto} statement that transfers the control outside
16149 the operator. A selective accept statement with @code{terminate} alternative
16150 is considered as non-structural statement. When computing this metric,
16151 @code{exit} statements are treated in the same way as @code{goto}
16152 statements unless @option{^-ne^NO_EXITS_AS_GOTOS^} option is specified.
16154 The Ada essential complexity metric defined here is intended to quantify
16155 the extent to which the software is unstructured. It is adapted from
16156 the McCabe essential complexity metric defined in
16157 @url{http://www.mccabe.com/pdf/mccabe-nist235r.pdf} but is modified to be more
16158 suitable for typical Ada usage. For example, short circuit forms
16159 are not penalized as unstructured in the Ada essential complexity metric.
16161 When computing cyclomatic and essential complexity, @command{gnatmetric} skips
16162 the code in the exception handlers and in all the nested program units. The
16163 code of assertions and predicates (that is, subprogram preconditions and
16164 postconditions, subtype predicates and type invariants) is also skipped.
16166 By default, all the complexity metrics are computed and reported.
16167 For more fine-grained control you can use
16168 the following switches:
16171 @cindex @option{^-complexity@var{x}^/COMPLEXITY_METRICS^} (@command{gnatmetric})
16174 @cindex @option{--no-complexity@var{x}}
16177 @item ^--complexity-all^/COMPLEXITY_METRICS=ALL^
16178 Report all the complexity metrics
16180 @item ^--no-complexity-all^/COMPLEXITY_METRICS=NONE^
16181 Do not report any of complexity metrics
16183 @item ^--complexity-cyclomatic^/COMPLEXITY_METRICS=CYCLOMATIC^
16184 Report the McCabe Cyclomatic Complexity
16186 @item ^--no-complexity-cyclomatic^/COMPLEXITY_METRICS=NOCYCLOMATIC^
16187 Do not report the McCabe Cyclomatic Complexity
16189 @item ^--complexity-essential^/COMPLEXITY_METRICS=ESSENTIAL^
16190 Report the Essential Complexity
16192 @item ^--no-complexity-essential^/COMPLEXITY_METRICS=NOESSENTIAL^
16193 Do not report the Essential Complexity
16195 @item ^--loop-nesting^/COMPLEXITY_METRICS=LOOP_NESTING_ON^
16196 Report maximal loop nesting level
16198 @item ^--no-loop-nesting^/COMPLEXITY_METRICS=NOLOOP_NESTING^
16199 Do not report maximal loop nesting level
16201 @item ^--complexity-average^/COMPLEXITY_METRICS=AVERAGE_COMPLEXITY^
16202 Report the average McCabe Cyclomatic Complexity for all the subprogram bodies,
16203 task bodies, entry bodies and statement sequences in package bodies.
16204 The metric is computed and reported for whole set of processed Ada sources
16207 @item ^--no-complexity-average^/COMPLEXITY_METRICS=NOAVERAGE_COMPLEXITY^
16208 Do not report the average McCabe Cyclomatic Complexity for all the subprogram
16209 bodies, task bodies, entry bodies and statement sequences in package bodies
16211 @cindex @option{^-ne^/NO_EXITS_AS_GOTOS^} (@command{gnatmetric})
16212 @item ^-ne^/NO_EXITS_AS_GOTOS^
16213 Do not consider @code{exit} statements as @code{goto}s when
16214 computing Essential Complexity
16216 @cindex @option{^--no-static-loop^/NO_STATIC_LOOP^} (@command{gnatmetric})
16217 @item ^--no-static-loop^/NO_STATIC_LOOP^
16218 Do not consider static loops when computing cyclomatic complexity
16220 @item ^--extra-exit-points^/EXTRA_EXIT_POINTS^
16221 Report the extra exit points for subprogram bodies. As an exit point, this
16222 metric counts @code{return} statements and raise statements in case when the
16223 raised exception is not handled in the same body. In case of a function this
16224 metric subtracts 1 from the number of exit points, because a function body
16225 must contain at least one @code{return} statement.
16227 @item ^--no-extra-exit-points^/NOEXTRA_EXIT_POINTS^
16228 Do not report the extra exit points for subprogram bodies
16232 @node Coupling Metrics Control
16233 @subsubsection Coupling Metrics Control
16234 @cindex Coupling metrics control in @command{gnatmetric}
16237 @cindex Coupling metrics (in @command{gnatmetric})
16238 Coupling metrics measure the dependencies between a given entity and other
16239 entities in the program. This information is useful since high coupling
16240 may signal potential issues with maintainability as the program evolves.
16242 @command{gnatmetric} computes the following coupling metrics:
16247 @emph{object-oriented coupling}, for classes in traditional object-oriented
16251 @emph{unit coupling}, for all the program units making up a program;
16254 @emph{control coupling}, reflecting dependencies between a unit and
16255 other units that contain subprograms.
16259 Two kinds of coupling metrics are computed:
16262 @item fan-out coupling (``efferent coupling''):
16263 @cindex fan-out coupling
16264 @cindex efferent coupling
16265 the number of entities the given entity depends upon. This metric
16266 reflects how the given entity depends on the changes in the
16267 ``external world''.
16269 @item fan-in coupling (``afferent'' coupling):
16270 @cindex fan-in coupling
16271 @cindex afferent coupling
16272 the number of entities that depend on a given entity.
16273 This metric reflects how the ``external world'' depends on the changes in a
16278 Object-oriented coupling metrics measure the dependencies
16279 between a given class (or a group of classes) and the other classes in the
16280 program. In this subsection the term ``class'' is used in its traditional
16281 object-oriented programming sense (an instantiable module that contains data
16282 and/or method members). A @emph{category} (of classes) is a group of closely
16283 related classes that are reused and/or modified together.
16285 A class @code{K}'s fan-out coupling is the number of classes
16286 that @code{K} depends upon.
16287 A category's fan-out coupling is the number of classes outside the
16288 category that the classes inside the category depend upon.
16290 A class @code{K}'s fan-in coupling is the number of classes
16291 that depend upon @code{K}.
16292 A category's fan-in coupling is the number of classes outside the
16293 category that depend on classes belonging to the category.
16295 Ada's object-oriented paradigm separates the instantiable entity
16296 (type) from the module (package), so the definition of the coupling
16297 metrics for Ada maps the class and class category notions
16298 onto Ada constructs.
16300 For the coupling metrics, several kinds of modules that define a tagged type
16301 or an interface type -- library packages, library generic packages, and
16302 library generic package instantiations -- are considered to be classes.
16303 A category consists of a library package (or
16304 a library generic package) that defines a tagged or an interface type,
16305 together with all its descendant (generic) packages that define tagged
16306 or interface types. Thus a
16307 category is an Ada hierarchy of library-level program units. Class
16308 coupling in Ada is referred to as ``tagged coupling'', and category coupling
16309 is referred to as ``hierarchy coupling''.
16311 For any package serving as a class, its body and subunits (if any) are
16312 considered together with its spec when computing dependencies, and coupling
16313 metrics are reported for spec units only. Dependencies between classes
16314 mean Ada semantic dependencies. For object-oriented coupling
16315 metrics, only dependencies on units treated as classes are
16318 Similarly, for unit and control coupling an entity is considered to be the
16319 conceptual construct consisting of the entity's specification, body, and
16320 any subunits (transitively).
16321 @command{gnatmetric} computes
16322 the dependencies of all these units as a whole, but
16323 metrics are only reported for spec
16324 units (or for a subprogram body unit in case if there is no
16325 separate spec for the given subprogram).
16327 For unit coupling, dependencies are computed between all kinds of program
16328 units. For control coupling, the dependencies of a given unit are limited to
16329 those units that define subprograms. Thus control fan-out coupling is reported
16330 for all units, but control fan-in coupling is only reported for units
16331 that define subprograms.
16333 The following simple example illustrates the difference between unit coupling
16334 and control coupling metrics:
16336 @smallexample @c ada
16339 function F_1 (I : Integer) return Integer;
16345 type T_2 is new Integer;
16350 package body Lib_1 is
16351 function F_1 (I : Integer) return Integer is
16359 with Lib_2; use Lib_2;
16362 function Fun (I : Integer) return Integer;
16367 with Lib_1; use Lib_1;
16368 package body Pack is
16369 function Fun (I : Integer) return Integer is
16378 If we apply @command{gnatmetric} with the @option{--coupling-all} option to
16379 these units, the result will be:
16385 Unit Lib_1 (C:\customers\662\L406-007\lib_1.ads)
16386 control fan-out coupling : 0
16387 control fan-in coupling : 1
16388 unit fan-out coupling : 0
16389 unit fan-in coupling : 1
16393 Unit Pack (C:\customers\662\L406-007\pack.ads)
16394 control fan-out coupling : 1
16395 control fan-in coupling : 0
16396 unit fan-out coupling : 2
16397 unit fan-in coupling : 0
16401 Unit Lib_2 (C:\customers\662\L406-007\lib_2.ads)
16402 control fan-out coupling : 0
16403 unit fan-out coupling : 0
16404 unit fan-in coupling : 1
16409 The result does not contain values for object-oriented
16410 coupling because none of the argument units contains a tagged type and
16411 therefore none of these units can be treated as a class.
16413 The @code{Pack} package (spec and body) depends on two
16414 units -- @code{Lib_1} @code{and Lib_2} -- and so its unit fan-out coupling
16415 is 2. Since nothing depends on it, its unit fan-in coupling is 0, as
16416 is its control fan-in coupling. Only one of the units @code{Pack} depends
16417 upon defines a subprogram, so its control fan-out coupling is 1.
16419 @code{Lib_2} depends on nothing, so its fan-out metrics are 0. It does
16420 not define any subprograms, so it has no control fan-in metric.
16421 One unit (@code{Pack}) depends on it , so its unit fan-in coupling is 1.
16423 @code{Lib_1} is similar to @code{Lib_2}, but it does define a subprogram.
16424 Its control fan-in coupling is 1 (because there is one unit
16427 When computing coupling metrics, @command{gnatmetric} counts only
16428 dependencies between units that are arguments of the @command{gnatmetric}
16429 invocation. Coupling metrics are program-wide (or project-wide) metrics, so
16430 you should invoke @command{gnatmetric} for
16431 the complete set of sources comprising your program. This can be done
16432 by invoking @command{gnatmetric} with the corresponding project file
16433 and with the @option{-U} option.
16435 By default, all the coupling metrics are disabled. You can use the following
16436 switches to specify the coupling metrics to be computed and reported:
16441 @cindex @option{--tagged-coupling@var{x}} (@command{gnatmetric})
16442 @cindex @option{--hierarchy-coupling@var{x}} (@command{gnatmetric})
16443 @cindex @option{--unit-coupling@var{x}} (@command{gnatmetric})
16444 @cindex @option{--control-coupling@var{x}} (@command{gnatmetric})
16448 @cindex @option{/COUPLING_METRICS} (@command{gnatmetric})
16451 @item ^--coupling-all^/COUPLING_METRICS=ALL^
16452 Report all the coupling metrics
16454 @item ^--tagged-coupling-out^/COUPLING_METRICS=TAGGED_OUT^
16455 Report tagged (class) fan-out coupling
16457 @item ^--tagged-coupling-in^/COUPLING_METRICS=TAGGED_IN^
16458 Report tagged (class) fan-in coupling
16460 @item ^--hierarchy-coupling-out^/COUPLING_METRICS=HIERARCHY_OUT^
16461 Report hierarchy (category) fan-out coupling
16463 @item ^--hierarchy-coupling-in^/COUPLING_METRICS=HIERARCHY_IN^
16464 Report hierarchy (category) fan-in coupling
16466 @item ^--unit-coupling-out^/COUPLING_METRICS=UNIT_OUT^
16467 Report unit fan-out coupling
16469 @item ^--unit-coupling-in^/COUPLING_METRICS=UNIT_IN^
16470 Report unit fan-in coupling
16472 @item ^--control-coupling-out^/COUPLING_METRICS=CONTROL_OUT^
16473 Report control fan-out coupling
16475 @item ^--control-coupling-in^/COUPLING_METRICS=CONTROL_IN^
16476 Report control fan-in coupling
16479 @node Other gnatmetric Switches
16480 @subsection Other @code{gnatmetric} Switches
16483 Additional @command{gnatmetric} switches are as follows:
16487 @cindex @option{--version} @command{gnatmetric}
16488 Display Copyright and version, then exit disregarding all other options.
16491 @cindex @option{--help} @command{gnatmetric}
16492 Display usage, then exit disregarding all other options.
16494 @item -P @var{file}
16495 @cindex @option{-P} @command{gnatmetric}
16496 Indicates the name of the project file that describes the set of sources
16497 to be processed. The exact set of argument sources depends on other options
16498 specified, see below.
16501 @cindex @option{-U} @command{gnatmetric}
16502 If a project file is specified and no argument source is explicitly
16503 specified (either directly or by means of @option{-files} option), process
16504 all the units of the closure of the argument project. Otherwise this option
16507 @item -U @var{main_unit}
16508 If a project file is specified and no argument source is explicitly
16509 specified (either directly or by means of @option{-files} option), process
16510 the closure of units rooted at @var{main_unit}. Otherwise this option
16513 @item -X@var{name}=@var{value}
16514 @cindex @option{-X} @command{gnatmetric}
16515 Indicates that external variable @var{name} in the argument project
16516 has the value @var{value}. Has no effect if no project is specified as
16519 @item --subdirs=@var{dir}
16520 @cindex @option{--subdirs=@var{dir}} @command{gnatmetric}
16521 Use the specified subdirectory of the project objects file (or of the
16522 project file directory if the project does not specify an object directory)
16523 for tool output files. Has no effect if no project is specified as
16524 tool argument r if @option{--no_objects_dir} is specified.
16526 @item --no_objects_dir
16527 @cindex @option{--no_objects_dir} @command{gnatmetric}
16528 Place all the result files into the current directory instead of
16529 project objects directory. This corresponds to the @command{gnatcheck}
16530 behavior when it is called with the project file from the
16531 GNAT driver. Has no effect if no project is specified.
16533 @item ^-files @var{filename}^/FILES=@var{filename}^
16534 @cindex @option{^-files^/FILES^} (@code{gnatmetric})
16535 Take the argument source files from the specified file. This file should be an
16536 ordinary text file containing file names separated by spaces or
16537 line breaks. You can use this switch more than once in the same call to
16538 @command{gnatmetric}. You also can combine this switch with
16539 an explicit list of files.
16541 @item ^-j^/PROCESSES=^@var{n}
16542 @cindex @option{^-j^/PROCESSES^} (@command{gnatmetric})
16543 Use @var{n} processes to carry out the tree creations (internal representations
16544 of the argument sources). On a multiprocessor machine this speeds up processing
16545 of big sets of argument sources. If @var{n} is 0, then the maximum number of
16546 parallel tree creations is the number of core processors on the platform.
16548 @cindex @option{^-t^/TIME^} (@command{gnatmetric})
16550 Print out execution time.
16552 @item ^-v^/VERBOSE^
16553 @cindex @option{^-v^/VERBOSE^} (@command{gnatmetric})
16555 @command{gnatmetric} generates version information and then
16556 a trace of sources being processed.
16559 @cindex @option{^-q^/QUIET^} (@command{gnatmetric})
16564 If a project file is specified and no argument source is explicitly
16565 specified (either directly or by means of @option{-files} option), and no
16566 @option{-U} is specified, then the set of processed sources is
16567 all the immediate units of the argument project.
16571 @node Generate project-wide metrics
16572 @subsection Generate project-wide metrics
16574 In order to compute metrics on all units of a given project, you can use
16575 the @command{gnat} driver along with the @option{-P} option:
16581 If the project @code{proj} depends upon other projects, you can compute
16582 the metrics on the project closure using the @option{-U} option:
16584 gnat metric -Pproj -U
16588 Finally, if not all the units are relevant to a particular main
16589 program in the project closure, you can generate metrics for the set
16590 of units needed to create a given main program (unit closure) using
16591 the @option{-U} option followed by the name of the main unit:
16593 gnat metric -Pproj -U main
16599 @c ***********************************
16600 @node File Name Krunching with gnatkr
16601 @chapter File Name Krunching with @code{gnatkr}
16605 This chapter discusses the method used by the compiler to shorten
16606 the default file names chosen for Ada units so that they do not
16607 exceed the maximum length permitted. It also describes the
16608 @code{gnatkr} utility that can be used to determine the result of
16609 applying this shortening.
16613 * Krunching Method::
16614 * Examples of gnatkr Usage::
16618 @section About @code{gnatkr}
16621 The default file naming rule in GNAT
16622 is that the file name must be derived from
16623 the unit name. The exact default rule is as follows:
16626 Take the unit name and replace all dots by hyphens.
16628 If such a replacement occurs in the
16629 second character position of a name, and the first character is
16630 ^@samp{a}, @samp{g}, @samp{s}, or @samp{i}, ^@samp{A}, @samp{G}, @samp{S}, or @samp{I},^
16631 then replace the dot by the character
16632 ^@samp{~} (tilde)^@samp{$} (dollar sign)^
16633 instead of a minus.
16635 The reason for this exception is to avoid clashes
16636 with the standard names for children of System, Ada, Interfaces,
16637 and GNAT, which use the prefixes
16638 ^@samp{s-}, @samp{a-}, @samp{i-}, and @samp{g-},^@samp{S-}, @samp{A-}, @samp{I-}, and @samp{G-},^
16641 The @option{^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{nn}}
16642 switch of the compiler activates a ``krunching''
16643 circuit that limits file names to nn characters (where nn is a decimal
16644 integer). For example, using OpenVMS,
16645 where the maximum file name length is
16646 39, the value of nn is usually set to 39, but if you want to generate
16647 a set of files that would be usable if ported to a system with some
16648 different maximum file length, then a different value can be specified.
16649 The default value of 39 for OpenVMS need not be specified.
16651 The @code{gnatkr} utility can be used to determine the krunched name for
16652 a given file, when krunched to a specified maximum length.
16655 @section Using @code{gnatkr}
16658 The @code{gnatkr} command has the form
16662 @c $ gnatkr @var{name} @ovar{length}
16663 @c Expanding @ovar macro inline (explanation in macro def comments)
16664 $ gnatkr @var{name} @r{[}@var{length}@r{]}
16670 $ gnatkr @var{name} /COUNT=nn
16675 @var{name} is the uncrunched file name, derived from the name of the unit
16676 in the standard manner described in the previous section (i.e., in particular
16677 all dots are replaced by hyphens). The file name may or may not have an
16678 extension (defined as a suffix of the form period followed by arbitrary
16679 characters other than period). If an extension is present then it will
16680 be preserved in the output. For example, when krunching @file{hellofile.ads}
16681 to eight characters, the result will be hellofil.ads.
16683 Note: for compatibility with previous versions of @code{gnatkr} dots may
16684 appear in the name instead of hyphens, but the last dot will always be
16685 taken as the start of an extension. So if @code{gnatkr} is given an argument
16686 such as @file{Hello.World.adb} it will be treated exactly as if the first
16687 period had been a hyphen, and for example krunching to eight characters
16688 gives the result @file{hellworl.adb}.
16690 Note that the result is always all lower case (except on OpenVMS where it is
16691 all upper case). Characters of the other case are folded as required.
16693 @var{length} represents the length of the krunched name. The default
16694 when no argument is given is ^8^39^ characters. A length of zero stands for
16695 unlimited, in other words do not chop except for system files where the
16696 implied crunching length is always eight characters.
16699 The output is the krunched name. The output has an extension only if the
16700 original argument was a file name with an extension.
16702 @node Krunching Method
16703 @section Krunching Method
16706 The initial file name is determined by the name of the unit that the file
16707 contains. The name is formed by taking the full expanded name of the
16708 unit and replacing the separating dots with hyphens and
16709 using ^lowercase^uppercase^
16710 for all letters, except that a hyphen in the second character position is
16711 replaced by a ^tilde^dollar sign^ if the first character is
16712 ^@samp{a}, @samp{i}, @samp{g}, or @samp{s}^@samp{A}, @samp{I}, @samp{G}, or @samp{S}^.
16713 The extension is @code{.ads} for a
16714 spec and @code{.adb} for a body.
16715 Krunching does not affect the extension, but the file name is shortened to
16716 the specified length by following these rules:
16720 The name is divided into segments separated by hyphens, tildes or
16721 underscores and all hyphens, tildes, and underscores are
16722 eliminated. If this leaves the name short enough, we are done.
16725 If the name is too long, the longest segment is located (left-most
16726 if there are two of equal length), and shortened by dropping
16727 its last character. This is repeated until the name is short enough.
16729 As an example, consider the krunching of @*@file{our-strings-wide_fixed.adb}
16730 to fit the name into 8 characters as required by some operating systems.
16733 our-strings-wide_fixed 22
16734 our strings wide fixed 19
16735 our string wide fixed 18
16736 our strin wide fixed 17
16737 our stri wide fixed 16
16738 our stri wide fixe 15
16739 our str wide fixe 14
16740 our str wid fixe 13
16746 Final file name: oustwifi.adb
16750 The file names for all predefined units are always krunched to eight
16751 characters. The krunching of these predefined units uses the following
16752 special prefix replacements:
16756 replaced by @file{^a^A^-}
16759 replaced by @file{^g^G^-}
16762 replaced by @file{^i^I^-}
16765 replaced by @file{^s^S^-}
16768 These system files have a hyphen in the second character position. That
16769 is why normal user files replace such a character with a
16770 ^tilde^dollar sign^, to
16771 avoid confusion with system file names.
16773 As an example of this special rule, consider
16774 @*@file{ada-strings-wide_fixed.adb}, which gets krunched as follows:
16777 ada-strings-wide_fixed 22
16778 a- strings wide fixed 18
16779 a- string wide fixed 17
16780 a- strin wide fixed 16
16781 a- stri wide fixed 15
16782 a- stri wide fixe 14
16783 a- str wide fixe 13
16789 Final file name: a-stwifi.adb
16793 Of course no file shortening algorithm can guarantee uniqueness over all
16794 possible unit names, and if file name krunching is used then it is your
16795 responsibility to ensure that no name clashes occur. The utility
16796 program @code{gnatkr} is supplied for conveniently determining the
16797 krunched name of a file.
16799 @node Examples of gnatkr Usage
16800 @section Examples of @code{gnatkr} Usage
16807 $ gnatkr very_long_unit_name.ads --> velounna.ads
16808 $ gnatkr grandparent-parent-child.ads --> grparchi.ads
16809 $ gnatkr Grandparent.Parent.Child.ads --> grparchi.ads
16810 $ gnatkr grandparent-parent-child --> grparchi
16812 $ gnatkr very_long_unit_name.ads/count=6 --> vlunna.ads
16813 $ gnatkr very_long_unit_name.ads/count=0 --> very_long_unit_name.ads
16816 @node Preprocessing with gnatprep
16817 @chapter Preprocessing with @code{gnatprep}
16821 This chapter discusses how to use GNAT's @code{gnatprep} utility for simple
16823 Although designed for use with GNAT, @code{gnatprep} does not depend on any
16824 special GNAT features.
16825 For further discussion of conditional compilation in general, see
16826 @ref{Conditional Compilation}.
16829 * Preprocessing Symbols::
16831 * Switches for gnatprep::
16832 * Form of Definitions File::
16833 * Form of Input Text for gnatprep::
16836 @node Preprocessing Symbols
16837 @section Preprocessing Symbols
16840 Preprocessing symbols are defined in definition files and referred to in
16841 sources to be preprocessed. A Preprocessing symbol is an identifier, following
16842 normal Ada (case-insensitive) rules for its syntax, with the restriction that
16843 all characters need to be in the ASCII set (no accented letters).
16845 @node Using gnatprep
16846 @section Using @code{gnatprep}
16849 To call @code{gnatprep} use
16852 @c $ gnatprep @ovar{switches} @var{infile} @var{outfile} @ovar{deffile}
16853 @c Expanding @ovar macro inline (explanation in macro def comments)
16854 $ gnatprep @r{[}@var{switches}@r{]} @var{infile} @var{outfile} @r{[}@var{deffile}@r{]}
16861 is an optional sequence of switches as described in the next section.
16864 is the full name of the input file, which is an Ada source
16865 file containing preprocessor directives.
16868 is the full name of the output file, which is an Ada source
16869 in standard Ada form. When used with GNAT, this file name will
16870 normally have an ads or adb suffix.
16873 is the full name of a text file containing definitions of
16874 preprocessing symbols to be referenced by the preprocessor. This argument is
16875 optional, and can be replaced by the use of the @option{-D} switch.
16879 @node Switches for gnatprep
16880 @section Switches for @code{gnatprep}
16885 @item ^-b^/BLANK_LINES^
16886 @cindex @option{^-b^/BLANK_LINES^} (@command{gnatprep})
16887 Causes both preprocessor lines and the lines deleted by
16888 preprocessing to be replaced by blank lines in the output source file,
16889 preserving line numbers in the output file.
16891 @item ^-c^/COMMENTS^
16892 @cindex @option{^-c^/COMMENTS^} (@command{gnatprep})
16893 Causes both preprocessor lines and the lines deleted
16894 by preprocessing to be retained in the output source as comments marked
16895 with the special string @code{"--! "}. This option will result in line numbers
16896 being preserved in the output file.
16898 @item ^-C^/REPLACE_IN_COMMENTS^
16899 @cindex @option{^-C^/REPLACE_IN_COMMENTS^} (@command{gnatprep})
16900 Causes comments to be scanned. Normally comments are ignored by gnatprep.
16901 If this option is specified, then comments are scanned and any $symbol
16902 substitutions performed as in program text. This is particularly useful
16903 when structured comments are used (e.g., when writing programs in the
16904 SPARK dialect of Ada). Note that this switch is not available when
16905 doing integrated preprocessing (it would be useless in this context
16906 since comments are ignored by the compiler in any case).
16908 @item ^-Dsymbol=value^/ASSOCIATE="symbol=value"^
16909 @cindex @option{^-D^/ASSOCIATE^} (@command{gnatprep})
16910 Defines a new preprocessing symbol, associated with value. If no value is given
16911 on the command line, then symbol is considered to be @code{True}. This switch
16912 can be used in place of a definition file.
16916 @cindex @option{/REMOVE} (@command{gnatprep})
16917 This is the default setting which causes lines deleted by preprocessing
16918 to be entirely removed from the output file.
16921 @item ^-r^/REFERENCE^
16922 @cindex @option{^-r^/REFERENCE^} (@command{gnatprep})
16923 Causes a @code{Source_Reference} pragma to be generated that
16924 references the original input file, so that error messages will use
16925 the file name of this original file. The use of this switch implies
16926 that preprocessor lines are not to be removed from the file, so its
16927 use will force @option{^-b^/BLANK_LINES^} mode if
16928 @option{^-c^/COMMENTS^}
16929 has not been specified explicitly.
16931 Note that if the file to be preprocessed contains multiple units, then
16932 it will be necessary to @code{gnatchop} the output file from
16933 @code{gnatprep}. If a @code{Source_Reference} pragma is present
16934 in the preprocessed file, it will be respected by
16935 @code{gnatchop ^-r^/REFERENCE^}
16936 so that the final chopped files will correctly refer to the original
16937 input source file for @code{gnatprep}.
16939 @item ^-s^/SYMBOLS^
16940 @cindex @option{^-s^/SYMBOLS^} (@command{gnatprep})
16941 Causes a sorted list of symbol names and values to be
16942 listed on the standard output file.
16944 @item ^-u^/UNDEFINED^
16945 @cindex @option{^-u^/UNDEFINED^} (@command{gnatprep})
16946 Causes undefined symbols to be treated as having the value FALSE in the context
16947 of a preprocessor test. In the absence of this option, an undefined symbol in
16948 a @code{#if} or @code{#elsif} test will be treated as an error.
16954 Note: if neither @option{-b} nor @option{-c} is present,
16955 then preprocessor lines and
16956 deleted lines are completely removed from the output, unless -r is
16957 specified, in which case -b is assumed.
16960 @node Form of Definitions File
16961 @section Form of Definitions File
16964 The definitions file contains lines of the form
16971 where symbol is a preprocessing symbol, and value is one of the following:
16975 Empty, corresponding to a null substitution
16977 A string literal using normal Ada syntax
16979 Any sequence of characters from the set
16980 (letters, digits, period, underline).
16984 Comment lines may also appear in the definitions file, starting with
16985 the usual @code{--},
16986 and comments may be added to the definitions lines.
16988 @node Form of Input Text for gnatprep
16989 @section Form of Input Text for @code{gnatprep}
16992 The input text may contain preprocessor conditional inclusion lines,
16993 as well as general symbol substitution sequences.
16995 The preprocessor conditional inclusion commands have the form
17000 #if @i{expression} @r{[}then@r{]}
17002 #elsif @i{expression} @r{[}then@r{]}
17004 #elsif @i{expression} @r{[}then@r{]}
17015 In this example, @i{expression} is defined by the following grammar:
17017 @i{expression} ::= <symbol>
17018 @i{expression} ::= <symbol> = "<value>"
17019 @i{expression} ::= <symbol> = <symbol>
17020 @i{expression} ::= <symbol> = <integer>
17021 @i{expression} ::= <symbol> > <integer>
17022 @i{expression} ::= <symbol> >= <integer>
17023 @i{expression} ::= <symbol> < <integer>
17024 @i{expression} ::= <symbol> <= <integer>
17025 @i{expression} ::= <symbol> 'Defined
17026 @i{expression} ::= not @i{expression}
17027 @i{expression} ::= @i{expression} and @i{expression}
17028 @i{expression} ::= @i{expression} or @i{expression}
17029 @i{expression} ::= @i{expression} and then @i{expression}
17030 @i{expression} ::= @i{expression} or else @i{expression}
17031 @i{expression} ::= ( @i{expression} )
17034 The following restriction exists: it is not allowed to have "and" or "or"
17035 following "not" in the same expression without parentheses. For example, this
17042 This should be one of the following:
17050 For the first test (@i{expression} ::= <symbol>) the symbol must have
17051 either the value true or false, that is to say the right-hand of the
17052 symbol definition must be one of the (case-insensitive) literals
17053 @code{True} or @code{False}. If the value is true, then the
17054 corresponding lines are included, and if the value is false, they are
17057 When comparing a symbol to an integer, the integer is any non negative
17058 literal integer as defined in the Ada Reference Manual, such as 3, 16#FF# or
17059 2#11#. The symbol value must also be a non negative integer. Integer values
17060 in the range 0 .. 2**31-1 are supported.
17062 The test (@i{expression} ::= <symbol> @code{'Defined}) is true only if
17063 the symbol has been defined in the definition file or by a @option{-D}
17064 switch on the command line. Otherwise, the test is false.
17066 The equality tests are case insensitive, as are all the preprocessor lines.
17068 If the symbol referenced is not defined in the symbol definitions file,
17069 then the effect depends on whether or not switch @option{-u}
17070 is specified. If so, then the symbol is treated as if it had the value
17071 false and the test fails. If this switch is not specified, then
17072 it is an error to reference an undefined symbol. It is also an error to
17073 reference a symbol that is defined with a value other than @code{True}
17076 The use of the @code{not} operator inverts the sense of this logical test.
17077 The @code{not} operator cannot be combined with the @code{or} or @code{and}
17078 operators, without parentheses. For example, "if not X or Y then" is not
17079 allowed, but "if (not X) or Y then" and "if not (X or Y) then" are.
17081 The @code{then} keyword is optional as shown
17083 The @code{#} must be the first non-blank character on a line, but
17084 otherwise the format is free form. Spaces or tabs may appear between
17085 the @code{#} and the keyword. The keywords and the symbols are case
17086 insensitive as in normal Ada code. Comments may be used on a
17087 preprocessor line, but other than that, no other tokens may appear on a
17088 preprocessor line. Any number of @code{elsif} clauses can be present,
17089 including none at all. The @code{else} is optional, as in Ada.
17091 The @code{#} marking the start of a preprocessor line must be the first
17092 non-blank character on the line, i.e., it must be preceded only by
17093 spaces or horizontal tabs.
17095 Symbol substitution outside of preprocessor lines is obtained by using
17103 anywhere within a source line, except in a comment or within a
17104 string literal. The identifier
17105 following the @code{$} must match one of the symbols defined in the symbol
17106 definition file, and the result is to substitute the value of the
17107 symbol in place of @code{$symbol} in the output file.
17109 Note that although the substitution of strings within a string literal
17110 is not possible, it is possible to have a symbol whose defined value is
17111 a string literal. So instead of setting XYZ to @code{hello} and writing:
17114 Header : String := "$XYZ";
17118 you should set XYZ to @code{"hello"} and write:
17121 Header : String := $XYZ;
17125 and then the substitution will occur as desired.
17127 @node The GNAT Library Browser gnatls
17128 @chapter The GNAT Library Browser @code{gnatls}
17130 @cindex Library browser
17133 @code{gnatls} is a tool that outputs information about compiled
17134 units. It gives the relationship between objects, unit names and source
17135 files. It can also be used to check the source dependencies of a unit
17136 as well as various characteristics.
17138 Note: to invoke @code{gnatls} with a project file, use the @code{gnat}
17139 driver (see @ref{The GNAT Driver and Project Files}).
17143 * Switches for gnatls::
17144 * Examples of gnatls Usage::
17147 @node Running gnatls
17148 @section Running @code{gnatls}
17151 The @code{gnatls} command has the form
17154 $ gnatls switches @var{object_or_ali_file}
17158 The main argument is the list of object or @file{ali} files
17159 (@pxref{The Ada Library Information Files})
17160 for which information is requested.
17162 In normal mode, without additional option, @code{gnatls} produces a
17163 four-column listing. Each line represents information for a specific
17164 object. The first column gives the full path of the object, the second
17165 column gives the name of the principal unit in this object, the third
17166 column gives the status of the source and the fourth column gives the
17167 full path of the source representing this unit.
17168 Here is a simple example of use:
17172 ^./^[]^demo1.o demo1 DIF demo1.adb
17173 ^./^[]^demo2.o demo2 OK demo2.adb
17174 ^./^[]^hello.o h1 OK hello.adb
17175 ^./^[]^instr-child.o instr.child MOK instr-child.adb
17176 ^./^[]^instr.o instr OK instr.adb
17177 ^./^[]^tef.o tef DIF tef.adb
17178 ^./^[]^text_io_example.o text_io_example OK text_io_example.adb
17179 ^./^[]^tgef.o tgef DIF tgef.adb
17183 The first line can be interpreted as follows: the main unit which is
17185 object file @file{demo1.o} is demo1, whose main source is in
17186 @file{demo1.adb}. Furthermore, the version of the source used for the
17187 compilation of demo1 has been modified (DIF). Each source file has a status
17188 qualifier which can be:
17191 @item OK (unchanged)
17192 The version of the source file used for the compilation of the
17193 specified unit corresponds exactly to the actual source file.
17195 @item MOK (slightly modified)
17196 The version of the source file used for the compilation of the
17197 specified unit differs from the actual source file but not enough to
17198 require recompilation. If you use gnatmake with the qualifier
17199 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}, a file marked
17200 MOK will not be recompiled.
17202 @item DIF (modified)
17203 No version of the source found on the path corresponds to the source
17204 used to build this object.
17206 @item ??? (file not found)
17207 No source file was found for this unit.
17209 @item HID (hidden, unchanged version not first on PATH)
17210 The version of the source that corresponds exactly to the source used
17211 for compilation has been found on the path but it is hidden by another
17212 version of the same source that has been modified.
17216 @node Switches for gnatls
17217 @section Switches for @code{gnatls}
17220 @code{gnatls} recognizes the following switches:
17224 @cindex @option{--version} @command{gnatls}
17225 Display Copyright and version, then exit disregarding all other options.
17228 @cindex @option{--help} @command{gnatls}
17229 If @option{--version} was not used, display usage, then exit disregarding
17232 @item ^-a^/ALL_UNITS^
17233 @cindex @option{^-a^/ALL_UNITS^} (@code{gnatls})
17234 Consider all units, including those of the predefined Ada library.
17235 Especially useful with @option{^-d^/DEPENDENCIES^}.
17237 @item ^-d^/DEPENDENCIES^
17238 @cindex @option{^-d^/DEPENDENCIES^} (@code{gnatls})
17239 List sources from which specified units depend on.
17241 @item ^-h^/OUTPUT=OPTIONS^
17242 @cindex @option{^-h^/OUTPUT=OPTIONS^} (@code{gnatls})
17243 Output the list of options.
17245 @item ^-o^/OUTPUT=OBJECTS^
17246 @cindex @option{^-o^/OUTPUT=OBJECTS^} (@code{gnatls})
17247 Only output information about object files.
17249 @item ^-s^/OUTPUT=SOURCES^
17250 @cindex @option{^-s^/OUTPUT=SOURCES^} (@code{gnatls})
17251 Only output information about source files.
17253 @item ^-u^/OUTPUT=UNITS^
17254 @cindex @option{^-u^/OUTPUT=UNITS^} (@code{gnatls})
17255 Only output information about compilation units.
17257 @item ^-files^/FILES^=@var{file}
17258 @cindex @option{^-files^/FILES^} (@code{gnatls})
17259 Take as arguments the files listed in text file @var{file}.
17260 Text file @var{file} may contain empty lines that are ignored.
17261 Each nonempty line should contain the name of an existing file.
17262 Several such switches may be specified simultaneously.
17264 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
17265 @itemx ^-aI^/SOURCE_SEARCH=^@var{dir}
17266 @itemx ^-I^/SEARCH=^@var{dir}
17267 @itemx ^-I-^/NOCURRENT_DIRECTORY^
17269 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatls})
17270 @cindex @option{^-aI^/SOURCE_SEARCH^} (@code{gnatls})
17271 @cindex @option{^-I^/SEARCH^} (@code{gnatls})
17272 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatls})
17273 Source path manipulation. Same meaning as the equivalent @command{gnatmake}
17274 flags (@pxref{Switches for gnatmake}).
17276 @item ^-aP^/ADD_PROJECT_SEARCH_DIR=^@var{dir}
17277 @cindex @option{^-aP^/ADD_PROJECT_SEARCH_DIR=^} (@code{gnatls})
17278 Add @var{dir} at the beginning of the project search dir.
17280 @item --RTS=@var{rts-path}
17281 @cindex @option{--RTS} (@code{gnatls})
17282 Specifies the default location of the runtime library. Same meaning as the
17283 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
17285 @item ^-v^/OUTPUT=VERBOSE^
17286 @cindex @option{^-v^/OUTPUT=VERBOSE^} (@code{gnatls})
17287 Verbose mode. Output the complete source, object and project paths. Do not use
17288 the default column layout but instead use long format giving as much as
17289 information possible on each requested units, including special
17290 characteristics such as:
17293 @item Preelaborable
17294 The unit is preelaborable in the Ada sense.
17297 No elaboration code has been produced by the compiler for this unit.
17300 The unit is pure in the Ada sense.
17302 @item Elaborate_Body
17303 The unit contains a pragma Elaborate_Body.
17306 The unit contains a pragma Remote_Types.
17308 @item Shared_Passive
17309 The unit contains a pragma Shared_Passive.
17312 This unit is part of the predefined environment and cannot be modified
17315 @item Remote_Call_Interface
17316 The unit contains a pragma Remote_Call_Interface.
17322 @node Examples of gnatls Usage
17323 @section Example of @code{gnatls} Usage
17327 Example of using the verbose switch. Note how the source and
17328 object paths are affected by the -I switch.
17331 $ gnatls -v -I.. demo1.o
17333 GNATLS 5.03w (20041123-34)
17334 Copyright 1997-2004 Free Software Foundation, Inc.
17336 Source Search Path:
17337 <Current_Directory>
17339 /home/comar/local/adainclude/
17341 Object Search Path:
17342 <Current_Directory>
17344 /home/comar/local/lib/gcc-lib/x86-linux/3.4.3/adalib/
17346 Project Search Path:
17347 <Current_Directory>
17348 /home/comar/local/lib/gnat/
17353 Kind => subprogram body
17354 Flags => No_Elab_Code
17355 Source => demo1.adb modified
17359 The following is an example of use of the dependency list.
17360 Note the use of the -s switch
17361 which gives a straight list of source files. This can be useful for
17362 building specialized scripts.
17365 $ gnatls -d demo2.o
17366 ./demo2.o demo2 OK demo2.adb
17372 $ gnatls -d -s -a demo1.o
17374 /home/comar/local/adainclude/ada.ads
17375 /home/comar/local/adainclude/a-finali.ads
17376 /home/comar/local/adainclude/a-filico.ads
17377 /home/comar/local/adainclude/a-stream.ads
17378 /home/comar/local/adainclude/a-tags.ads
17381 /home/comar/local/adainclude/gnat.ads
17382 /home/comar/local/adainclude/g-io.ads
17384 /home/comar/local/adainclude/system.ads
17385 /home/comar/local/adainclude/s-exctab.ads
17386 /home/comar/local/adainclude/s-finimp.ads
17387 /home/comar/local/adainclude/s-finroo.ads
17388 /home/comar/local/adainclude/s-secsta.ads
17389 /home/comar/local/adainclude/s-stalib.ads
17390 /home/comar/local/adainclude/s-stoele.ads
17391 /home/comar/local/adainclude/s-stratt.ads
17392 /home/comar/local/adainclude/s-tasoli.ads
17393 /home/comar/local/adainclude/s-unstyp.ads
17394 /home/comar/local/adainclude/unchconv.ads
17400 GNAT LIST /DEPENDENCIES /OUTPUT=SOURCES /ALL_UNITS DEMO1.ADB
17402 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]ada.ads
17403 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-finali.ads
17404 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-filico.ads
17405 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-stream.ads
17406 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-tags.ads
17410 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]gnat.ads
17411 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]g-io.ads
17413 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]system.ads
17414 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-exctab.ads
17415 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finimp.ads
17416 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finroo.ads
17417 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-secsta.ads
17418 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stalib.ads
17419 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stoele.ads
17420 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stratt.ads
17421 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-tasoli.ads
17422 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-unstyp.ads
17423 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]unchconv.ads
17427 @node Cleaning Up with gnatclean
17428 @chapter Cleaning Up with @code{gnatclean}
17430 @cindex Cleaning tool
17433 @code{gnatclean} is a tool that allows the deletion of files produced by the
17434 compiler, binder and linker, including ALI files, object files, tree files,
17435 expanded source files, library files, interface copy source files, binder
17436 generated files and executable files.
17439 * Running gnatclean::
17440 * Switches for gnatclean::
17441 @c * Examples of gnatclean Usage::
17444 @node Running gnatclean
17445 @section Running @code{gnatclean}
17448 The @code{gnatclean} command has the form:
17451 $ gnatclean switches @var{names}
17455 @var{names} is a list of source file names. Suffixes @code{.^ads^ADS^} and
17456 @code{^adb^ADB^} may be omitted. If a project file is specified using switch
17457 @code{^-P^/PROJECT_FILE=^}, then @var{names} may be completely omitted.
17460 In normal mode, @code{gnatclean} delete the files produced by the compiler and,
17461 if switch @code{^-c^/COMPILER_FILES_ONLY^} is not specified, by the binder and
17462 the linker. In informative-only mode, specified by switch
17463 @code{^-n^/NODELETE^}, the list of files that would have been deleted in
17464 normal mode is listed, but no file is actually deleted.
17466 @node Switches for gnatclean
17467 @section Switches for @code{gnatclean}
17470 @code{gnatclean} recognizes the following switches:
17474 @cindex @option{--version} @command{gnatclean}
17475 Display Copyright and version, then exit disregarding all other options.
17478 @cindex @option{--help} @command{gnatclean}
17479 If @option{--version} was not used, display usage, then exit disregarding
17482 @item ^--subdirs^/SUBDIRS^=subdir
17483 Actual object directory of each project file is the subdirectory subdir of the
17484 object directory specified or defaulted in the project file.
17486 @item ^--unchecked-shared-lib-imports^/UNCHECKED_SHARED_LIB_IMPORTS^
17487 By default, shared library projects are not allowed to import static library
17488 projects. When this switch is used on the command line, this restriction is
17491 @item ^-c^/COMPILER_FILES_ONLY^
17492 @cindex @option{^-c^/COMPILER_FILES_ONLY^} (@code{gnatclean})
17493 Only attempt to delete the files produced by the compiler, not those produced
17494 by the binder or the linker. The files that are not to be deleted are library
17495 files, interface copy files, binder generated files and executable files.
17497 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
17498 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@code{gnatclean})
17499 Indicate that ALI and object files should normally be found in directory
17502 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
17503 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@code{gnatclean})
17504 When using project files, if some errors or warnings are detected during
17505 parsing and verbose mode is not in effect (no use of switch
17506 ^-v^/VERBOSE^), then error lines start with the full path name of the project
17507 file, rather than its simple file name.
17510 @cindex @option{^-h^/HELP^} (@code{gnatclean})
17511 Output a message explaining the usage of @code{^gnatclean^gnatclean^}.
17513 @item ^-n^/NODELETE^
17514 @cindex @option{^-n^/NODELETE^} (@code{gnatclean})
17515 Informative-only mode. Do not delete any files. Output the list of the files
17516 that would have been deleted if this switch was not specified.
17518 @item ^-P^/PROJECT_FILE=^@var{project}
17519 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatclean})
17520 Use project file @var{project}. Only one such switch can be used.
17521 When cleaning a project file, the files produced by the compilation of the
17522 immediate sources or inherited sources of the project files are to be
17523 deleted. This is not depending on the presence or not of executable names
17524 on the command line.
17527 @cindex @option{^-q^/QUIET^} (@code{gnatclean})
17528 Quiet output. If there are no errors, do not output anything, except in
17529 verbose mode (switch ^-v^/VERBOSE^) or in informative-only mode
17530 (switch ^-n^/NODELETE^).
17532 @item ^-r^/RECURSIVE^
17533 @cindex @option{^-r^/RECURSIVE^} (@code{gnatclean})
17534 When a project file is specified (using switch ^-P^/PROJECT_FILE=^),
17535 clean all imported and extended project files, recursively. If this switch
17536 is not specified, only the files related to the main project file are to be
17537 deleted. This switch has no effect if no project file is specified.
17539 @item ^-v^/VERBOSE^
17540 @cindex @option{^-v^/VERBOSE^} (@code{gnatclean})
17543 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
17544 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (@code{gnatclean})
17545 Indicates the verbosity of the parsing of GNAT project files.
17546 @xref{Switches Related to Project Files}.
17548 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
17549 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (@code{gnatclean})
17550 Indicates that external variable @var{name} has the value @var{value}.
17551 The Project Manager will use this value for occurrences of
17552 @code{external(name)} when parsing the project file.
17553 @xref{Switches Related to Project Files}.
17555 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
17556 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatclean})
17557 When searching for ALI and object files, look in directory
17560 @item ^-I^/SEARCH=^@var{dir}
17561 @cindex @option{^-I^/SEARCH^} (@code{gnatclean})
17562 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}}.
17564 @item ^-I-^/NOCURRENT_DIRECTORY^
17565 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatclean})
17566 @cindex Source files, suppressing search
17567 Do not look for ALI or object files in the directory
17568 where @code{gnatclean} was invoked.
17572 @c @node Examples of gnatclean Usage
17573 @c @section Examples of @code{gnatclean} Usage
17576 @node GNAT and Libraries
17577 @chapter GNAT and Libraries
17578 @cindex Library, building, installing, using
17581 This chapter describes how to build and use libraries with GNAT, and also shows
17582 how to recompile the GNAT run-time library. You should be familiar with the
17583 Project Manager facility (@pxref{GNAT Project Manager}) before reading this
17587 * Introduction to Libraries in GNAT::
17588 * General Ada Libraries::
17589 * Stand-alone Ada Libraries::
17590 * Rebuilding the GNAT Run-Time Library::
17593 @node Introduction to Libraries in GNAT
17594 @section Introduction to Libraries in GNAT
17597 A library is, conceptually, a collection of objects which does not have its
17598 own main thread of execution, but rather provides certain services to the
17599 applications that use it. A library can be either statically linked with the
17600 application, in which case its code is directly included in the application,
17601 or, on platforms that support it, be dynamically linked, in which case
17602 its code is shared by all applications making use of this library.
17604 GNAT supports both types of libraries.
17605 In the static case, the compiled code can be provided in different ways. The
17606 simplest approach is to provide directly the set of objects resulting from
17607 compilation of the library source files. Alternatively, you can group the
17608 objects into an archive using whatever commands are provided by the operating
17609 system. For the latter case, the objects are grouped into a shared library.
17611 In the GNAT environment, a library has three types of components:
17617 @xref{The Ada Library Information Files}.
17619 Object files, an archive or a shared library.
17623 A GNAT library may expose all its source files, which is useful for
17624 documentation purposes. Alternatively, it may expose only the units needed by
17625 an external user to make use of the library. That is to say, the specs
17626 reflecting the library services along with all the units needed to compile
17627 those specs, which can include generic bodies or any body implementing an
17628 inlined routine. In the case of @emph{stand-alone libraries} those exposed
17629 units are called @emph{interface units} (@pxref{Stand-alone Ada Libraries}).
17631 All compilation units comprising an application, including those in a library,
17632 need to be elaborated in an order partially defined by Ada's semantics. GNAT
17633 computes the elaboration order from the @file{ALI} files and this is why they
17634 constitute a mandatory part of GNAT libraries.
17635 @emph{Stand-alone libraries} are the exception to this rule because a specific
17636 library elaboration routine is produced independently of the application(s)
17639 @node General Ada Libraries
17640 @section General Ada Libraries
17643 * Building a library::
17644 * Installing a library::
17645 * Using a library::
17648 @node Building a library
17649 @subsection Building a library
17652 The easiest way to build a library is to use the Project Manager,
17653 which supports a special type of project called a @emph{Library Project}
17654 (@pxref{Library Projects}).
17656 A project is considered a library project, when two project-level attributes
17657 are defined in it: @code{Library_Name} and @code{Library_Dir}. In order to
17658 control different aspects of library configuration, additional optional
17659 project-level attributes can be specified:
17662 This attribute controls whether the library is to be static or dynamic
17664 @item Library_Version
17665 This attribute specifies the library version; this value is used
17666 during dynamic linking of shared libraries to determine if the currently
17667 installed versions of the binaries are compatible.
17669 @item Library_Options
17671 These attributes specify additional low-level options to be used during
17672 library generation, and redefine the actual application used to generate
17677 The GNAT Project Manager takes full care of the library maintenance task,
17678 including recompilation of the source files for which objects do not exist
17679 or are not up to date, assembly of the library archive, and installation of
17680 the library (i.e., copying associated source, object and @file{ALI} files
17681 to the specified location).
17683 Here is a simple library project file:
17684 @smallexample @c ada
17686 for Source_Dirs use ("src1", "src2");
17687 for Object_Dir use "obj";
17688 for Library_Name use "mylib";
17689 for Library_Dir use "lib";
17690 for Library_Kind use "dynamic";
17695 and the compilation command to build and install the library:
17697 @smallexample @c ada
17698 $ gnatmake -Pmy_lib
17702 It is not entirely trivial to perform manually all the steps required to
17703 produce a library. We recommend that you use the GNAT Project Manager
17704 for this task. In special cases where this is not desired, the necessary
17705 steps are discussed below.
17707 There are various possibilities for compiling the units that make up the
17708 library: for example with a Makefile (@pxref{Using the GNU make Utility}) or
17709 with a conventional script. For simple libraries, it is also possible to create
17710 a dummy main program which depends upon all the packages that comprise the
17711 interface of the library. This dummy main program can then be given to
17712 @command{gnatmake}, which will ensure that all necessary objects are built.
17714 After this task is accomplished, you should follow the standard procedure
17715 of the underlying operating system to produce the static or shared library.
17717 Here is an example of such a dummy program:
17718 @smallexample @c ada
17720 with My_Lib.Service1;
17721 with My_Lib.Service2;
17722 with My_Lib.Service3;
17723 procedure My_Lib_Dummy is
17731 Here are the generic commands that will build an archive or a shared library.
17734 # compiling the library
17735 $ gnatmake -c my_lib_dummy.adb
17737 # we don't need the dummy object itself
17738 $ rm my_lib_dummy.o my_lib_dummy.ali
17740 # create an archive with the remaining objects
17741 $ ar rc libmy_lib.a *.o
17742 # some systems may require "ranlib" to be run as well
17744 # or create a shared library
17745 $ gcc -shared -o libmy_lib.so *.o
17746 # some systems may require the code to have been compiled with -fPIC
17748 # remove the object files that are now in the library
17751 # Make the ALI files read-only so that gnatmake will not try to
17752 # regenerate the objects that are in the library
17757 Please note that the library must have a name of the form @file{lib@var{xxx}.a}
17758 or @file{lib@var{xxx}.so} (or @file{lib@var{xxx}.dll} on Windows) in order to
17759 be accessed by the directive @option{-l@var{xxx}} at link time.
17761 @node Installing a library
17762 @subsection Installing a library
17763 @cindex @code{ADA_PROJECT_PATH}
17764 @cindex @code{GPR_PROJECT_PATH}
17767 If you use project files, library installation is part of the library build
17768 process (@pxref{Installing a library with project files}).
17770 When project files are not an option, it is also possible, but not recommended,
17771 to install the library so that the sources needed to use the library are on the
17772 Ada source path and the ALI files & libraries be on the Ada Object path (see
17773 @ref{Search Paths and the Run-Time Library (RTL)}. Alternatively, the system
17774 administrator can place general-purpose libraries in the default compiler
17775 paths, by specifying the libraries' location in the configuration files
17776 @file{ada_source_path} and @file{ada_object_path}. These configuration files
17777 must be located in the GNAT installation tree at the same place as the gcc spec
17778 file. The location of the gcc spec file can be determined as follows:
17784 The configuration files mentioned above have a simple format: each line
17785 must contain one unique directory name.
17786 Those names are added to the corresponding path
17787 in their order of appearance in the file. The names can be either absolute
17788 or relative; in the latter case, they are relative to where theses files
17791 The files @file{ada_source_path} and @file{ada_object_path} might not be
17793 GNAT installation, in which case, GNAT will look for its run-time library in
17794 the directories @file{adainclude} (for the sources) and @file{adalib} (for the
17795 objects and @file{ALI} files). When the files exist, the compiler does not
17796 look in @file{adainclude} and @file{adalib}, and thus the
17797 @file{ada_source_path} file
17798 must contain the location for the GNAT run-time sources (which can simply
17799 be @file{adainclude}). In the same way, the @file{ada_object_path} file must
17800 contain the location for the GNAT run-time objects (which can simply
17803 You can also specify a new default path to the run-time library at compilation
17804 time with the switch @option{--RTS=rts-path}. You can thus choose / change
17805 the run-time library you want your program to be compiled with. This switch is
17806 recognized by @command{gcc}, @command{gnatmake}, @command{gnatbind},
17807 @command{gnatls}, @command{gnatfind} and @command{gnatxref}.
17809 It is possible to install a library before or after the standard GNAT
17810 library, by reordering the lines in the configuration files. In general, a
17811 library must be installed before the GNAT library if it redefines
17814 @node Using a library
17815 @subsection Using a library
17817 @noindent Once again, the project facility greatly simplifies the use of
17818 libraries. In this context, using a library is just a matter of adding a
17819 @code{with} clause in the user project. For instance, to make use of the
17820 library @code{My_Lib} shown in examples in earlier sections, you can
17823 @smallexample @c projectfile
17830 Even if you have a third-party, non-Ada library, you can still use GNAT's
17831 Project Manager facility to provide a wrapper for it. For example, the
17832 following project, when @code{with}ed by your main project, will link with the
17833 third-party library @file{liba.a}:
17835 @smallexample @c projectfile
17838 for Externally_Built use "true";
17839 for Source_Files use ();
17840 for Library_Dir use "lib";
17841 for Library_Name use "a";
17842 for Library_Kind use "static";
17846 This is an alternative to the use of @code{pragma Linker_Options}. It is
17847 especially interesting in the context of systems with several interdependent
17848 static libraries where finding a proper linker order is not easy and best be
17849 left to the tools having visibility over project dependence information.
17852 In order to use an Ada library manually, you need to make sure that this
17853 library is on both your source and object path
17854 (see @ref{Search Paths and the Run-Time Library (RTL)}
17855 and @ref{Search Paths for gnatbind}). Furthermore, when the objects are grouped
17856 in an archive or a shared library, you need to specify the desired
17857 library at link time.
17859 For example, you can use the library @file{mylib} installed in
17860 @file{/dir/my_lib_src} and @file{/dir/my_lib_obj} with the following commands:
17863 $ gnatmake -aI/dir/my_lib_src -aO/dir/my_lib_obj my_appl \
17868 This can be expressed more simply:
17873 when the following conditions are met:
17876 @file{/dir/my_lib_src} has been added by the user to the environment
17877 variable @env{ADA_INCLUDE_PATH}, or by the administrator to the file
17878 @file{ada_source_path}
17880 @file{/dir/my_lib_obj} has been added by the user to the environment
17881 variable @env{ADA_OBJECTS_PATH}, or by the administrator to the file
17882 @file{ada_object_path}
17884 a pragma @code{Linker_Options} has been added to one of the sources.
17887 @smallexample @c ada
17888 pragma Linker_Options ("-lmy_lib");
17892 @node Stand-alone Ada Libraries
17893 @section Stand-alone Ada Libraries
17894 @cindex Stand-alone library, building, using
17897 * Introduction to Stand-alone Libraries::
17898 * Building a Stand-alone Library::
17899 * Creating a Stand-alone Library to be used in a non-Ada context::
17900 * Restrictions in Stand-alone Libraries::
17903 @node Introduction to Stand-alone Libraries
17904 @subsection Introduction to Stand-alone Libraries
17907 A Stand-alone Library (abbreviated ``SAL'') is a library that contains the
17909 elaborate the Ada units that are included in the library. In contrast with
17910 an ordinary library, which consists of all sources, objects and @file{ALI}
17912 library, a SAL may specify a restricted subset of compilation units
17913 to serve as a library interface. In this case, the fully
17914 self-sufficient set of files will normally consist of an objects
17915 archive, the sources of interface units' specs, and the @file{ALI}
17916 files of interface units.
17917 If an interface spec contains a generic unit or an inlined subprogram,
17919 source must also be provided; if the units that must be provided in the source
17920 form depend on other units, the source and @file{ALI} files of those must
17923 The main purpose of a SAL is to minimize the recompilation overhead of client
17924 applications when a new version of the library is installed. Specifically,
17925 if the interface sources have not changed, client applications do not need to
17926 be recompiled. If, furthermore, a SAL is provided in the shared form and its
17927 version, controlled by @code{Library_Version} attribute, is not changed,
17928 then the clients do not need to be relinked.
17930 SALs also allow the library providers to minimize the amount of library source
17931 text exposed to the clients. Such ``information hiding'' might be useful or
17932 necessary for various reasons.
17934 Stand-alone libraries are also well suited to be used in an executable whose
17935 main routine is not written in Ada.
17937 @node Building a Stand-alone Library
17938 @subsection Building a Stand-alone Library
17941 GNAT's Project facility provides a simple way of building and installing
17942 stand-alone libraries; see @ref{Stand-alone Library Projects}.
17943 To be a Stand-alone Library Project, in addition to the two attributes
17944 that make a project a Library Project (@code{Library_Name} and
17945 @code{Library_Dir}; see @ref{Library Projects}), the attribute
17946 @code{Library_Interface} must be defined. For example:
17948 @smallexample @c projectfile
17950 for Library_Dir use "lib_dir";
17951 for Library_Name use "dummy";
17952 for Library_Interface use ("int1", "int1.child");
17957 Attribute @code{Library_Interface} has a non-empty string list value,
17958 each string in the list designating a unit contained in an immediate source
17959 of the project file.
17961 When a Stand-alone Library is built, first the binder is invoked to build
17962 a package whose name depends on the library name
17963 (@file{^b~dummy.ads/b^B$DUMMY.ADS/B^} in the example above).
17964 This binder-generated package includes initialization and
17965 finalization procedures whose
17966 names depend on the library name (@code{dummyinit} and @code{dummyfinal}
17968 above). The object corresponding to this package is included in the library.
17970 You must ensure timely (e.g., prior to any use of interfaces in the SAL)
17971 calling of these procedures if a static SAL is built, or if a shared SAL
17973 with the project-level attribute @code{Library_Auto_Init} set to
17976 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
17977 (those that are listed in attribute @code{Library_Interface}) are copied to
17978 the Library Directory. As a consequence, only the Interface Units may be
17979 imported from Ada units outside of the library. If other units are imported,
17980 the binding phase will fail.
17983 It is also possible to build an encapsulated library where not only
17984 the code to elaborate and finalize the library is embedded but also
17985 ensuring that the library is linked only against static
17986 libraries. So an encapsulated library only depends on system
17987 libraries, all other code, including the GNAT runtime, is embedded. To
17988 build an encapsulated library the attribute
17989 @code{Library_Standalone} must be set to @code{encapsulated}:
17991 @smallexample @c projectfile
17993 for Library_Dir use "lib_dir";
17994 for Library_Name use "dummy";
17995 for Library_Kind use "dynamic";
17996 for Library_Interface use ("int1", "int1.child");
17997 for Library_Standalone use "encapsulated";
18002 The default value for this attribute is @code{standard} in which case
18003 a stand-alone library is built.
18005 The attribute @code{Library_Src_Dir} may be specified for a
18006 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
18007 single string value. Its value must be the path (absolute or relative to the
18008 project directory) of an existing directory. This directory cannot be the
18009 object directory or one of the source directories, but it can be the same as
18010 the library directory. The sources of the Interface
18011 Units of the library that are needed by an Ada client of the library will be
18012 copied to the designated directory, called the Interface Copy directory.
18013 These sources include the specs of the Interface Units, but they may also
18014 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
18015 are used, or when there is a generic unit in the spec. Before the sources
18016 are copied to the Interface Copy directory, an attempt is made to delete all
18017 files in the Interface Copy directory.
18019 Building stand-alone libraries by hand is somewhat tedious, but for those
18020 occasions when it is necessary here are the steps that you need to perform:
18023 Compile all library sources.
18026 Invoke the binder with the switch @option{-n} (No Ada main program),
18027 with all the @file{ALI} files of the interfaces, and
18028 with the switch @option{-L} to give specific names to the @code{init}
18029 and @code{final} procedures. For example:
18031 gnatbind -n int1.ali int2.ali -Lsal1
18035 Compile the binder generated file:
18041 Link the dynamic library with all the necessary object files,
18042 indicating to the linker the names of the @code{init} (and possibly
18043 @code{final}) procedures for automatic initialization (and finalization).
18044 The built library should be placed in a directory different from
18045 the object directory.
18048 Copy the @code{ALI} files of the interface to the library directory,
18049 add in this copy an indication that it is an interface to a SAL
18050 (i.e., add a word @option{SL} on the line in the @file{ALI} file that starts
18051 with letter ``P'') and make the modified copy of the @file{ALI} file
18056 Using SALs is not different from using other libraries
18057 (see @ref{Using a library}).
18059 @node Creating a Stand-alone Library to be used in a non-Ada context
18060 @subsection Creating a Stand-alone Library to be used in a non-Ada context
18063 It is easy to adapt the SAL build procedure discussed above for use of a SAL in
18066 The only extra step required is to ensure that library interface subprograms
18067 are compatible with the main program, by means of @code{pragma Export}
18068 or @code{pragma Convention}.
18070 Here is an example of simple library interface for use with C main program:
18072 @smallexample @c ada
18073 package My_Package is
18075 procedure Do_Something;
18076 pragma Export (C, Do_Something, "do_something");
18078 procedure Do_Something_Else;
18079 pragma Export (C, Do_Something_Else, "do_something_else");
18085 On the foreign language side, you must provide a ``foreign'' view of the
18086 library interface; remember that it should contain elaboration routines in
18087 addition to interface subprograms.
18089 The example below shows the content of @code{mylib_interface.h} (note
18090 that there is no rule for the naming of this file, any name can be used)
18092 /* the library elaboration procedure */
18093 extern void mylibinit (void);
18095 /* the library finalization procedure */
18096 extern void mylibfinal (void);
18098 /* the interface exported by the library */
18099 extern void do_something (void);
18100 extern void do_something_else (void);
18104 Libraries built as explained above can be used from any program, provided
18105 that the elaboration procedures (named @code{mylibinit} in the previous
18106 example) are called before the library services are used. Any number of
18107 libraries can be used simultaneously, as long as the elaboration
18108 procedure of each library is called.
18110 Below is an example of a C program that uses the @code{mylib} library.
18113 #include "mylib_interface.h"
18118 /* First, elaborate the library before using it */
18121 /* Main program, using the library exported entities */
18123 do_something_else ();
18125 /* Library finalization at the end of the program */
18132 Note that invoking any library finalization procedure generated by
18133 @code{gnatbind} shuts down the Ada run-time environment.
18135 finalization of all Ada libraries must be performed at the end of the program.
18136 No call to these libraries or to the Ada run-time library should be made
18137 after the finalization phase.
18139 @node Restrictions in Stand-alone Libraries
18140 @subsection Restrictions in Stand-alone Libraries
18143 The pragmas listed below should be used with caution inside libraries,
18144 as they can create incompatibilities with other Ada libraries:
18146 @item pragma @code{Locking_Policy}
18147 @item pragma @code{Partition_Elaboration_Policy}
18148 @item pragma @code{Queuing_Policy}
18149 @item pragma @code{Task_Dispatching_Policy}
18150 @item pragma @code{Unreserve_All_Interrupts}
18154 When using a library that contains such pragmas, the user must make sure
18155 that all libraries use the same pragmas with the same values. Otherwise,
18156 @code{Program_Error} will
18157 be raised during the elaboration of the conflicting
18158 libraries. The usage of these pragmas and its consequences for the user
18159 should therefore be well documented.
18161 Similarly, the traceback in the exception occurrence mechanism should be
18162 enabled or disabled in a consistent manner across all libraries.
18163 Otherwise, Program_Error will be raised during the elaboration of the
18164 conflicting libraries.
18166 If the @code{Version} or @code{Body_Version}
18167 attributes are used inside a library, then you need to
18168 perform a @code{gnatbind} step that specifies all @file{ALI} files in all
18169 libraries, so that version identifiers can be properly computed.
18170 In practice these attributes are rarely used, so this is unlikely
18171 to be a consideration.
18173 @node Rebuilding the GNAT Run-Time Library
18174 @section Rebuilding the GNAT Run-Time Library
18175 @cindex GNAT Run-Time Library, rebuilding
18176 @cindex Building the GNAT Run-Time Library
18177 @cindex Rebuilding the GNAT Run-Time Library
18178 @cindex Run-Time Library, rebuilding
18181 It may be useful to recompile the GNAT library in various contexts, the
18182 most important one being the use of partition-wide configuration pragmas
18183 such as @code{Normalize_Scalars}. A special Makefile called
18184 @code{Makefile.adalib} is provided to that effect and can be found in
18185 the directory containing the GNAT library. The location of this
18186 directory depends on the way the GNAT environment has been installed and can
18187 be determined by means of the command:
18194 The last entry in the object search path usually contains the
18195 gnat library. This Makefile contains its own documentation and in
18196 particular the set of instructions needed to rebuild a new library and
18199 @node Using the GNU make Utility
18200 @chapter Using the GNU @code{make} Utility
18204 This chapter offers some examples of makefiles that solve specific
18205 problems. It does not explain how to write a makefile (@pxref{Top,, GNU
18206 make, make, GNU @code{make}}), nor does it try to replace the
18207 @command{gnatmake} utility (@pxref{The GNAT Make Program gnatmake}).
18209 All the examples in this section are specific to the GNU version of
18210 make. Although @command{make} is a standard utility, and the basic language
18211 is the same, these examples use some advanced features found only in
18215 * Using gnatmake in a Makefile::
18216 * Automatically Creating a List of Directories::
18217 * Generating the Command Line Switches::
18218 * Overcoming Command Line Length Limits::
18221 @node Using gnatmake in a Makefile
18222 @section Using gnatmake in a Makefile
18227 Complex project organizations can be handled in a very powerful way by
18228 using GNU make combined with gnatmake. For instance, here is a Makefile
18229 which allows you to build each subsystem of a big project into a separate
18230 shared library. Such a makefile allows you to significantly reduce the link
18231 time of very big applications while maintaining full coherence at
18232 each step of the build process.
18234 The list of dependencies are handled automatically by
18235 @command{gnatmake}. The Makefile is simply used to call gnatmake in each of
18236 the appropriate directories.
18238 Note that you should also read the example on how to automatically
18239 create the list of directories
18240 (@pxref{Automatically Creating a List of Directories})
18241 which might help you in case your project has a lot of subdirectories.
18246 @font@heightrm=cmr8
18249 ## This Makefile is intended to be used with the following directory
18251 ## - The sources are split into a series of csc (computer software components)
18252 ## Each of these csc is put in its own directory.
18253 ## Their name are referenced by the directory names.
18254 ## They will be compiled into shared library (although this would also work
18255 ## with static libraries
18256 ## - The main program (and possibly other packages that do not belong to any
18257 ## csc is put in the top level directory (where the Makefile is).
18258 ## toplevel_dir __ first_csc (sources) __ lib (will contain the library)
18259 ## \_ second_csc (sources) __ lib (will contain the library)
18261 ## Although this Makefile is build for shared library, it is easy to modify
18262 ## to build partial link objects instead (modify the lines with -shared and
18265 ## With this makefile, you can change any file in the system or add any new
18266 ## file, and everything will be recompiled correctly (only the relevant shared
18267 ## objects will be recompiled, and the main program will be re-linked).
18269 # The list of computer software component for your project. This might be
18270 # generated automatically.
18273 # Name of the main program (no extension)
18276 # If we need to build objects with -fPIC, uncomment the following line
18279 # The following variable should give the directory containing libgnat.so
18280 # You can get this directory through 'gnatls -v'. This is usually the last
18281 # directory in the Object_Path.
18284 # The directories for the libraries
18285 # (This macro expands the list of CSC to the list of shared libraries, you
18286 # could simply use the expanded form:
18287 # LIB_DIR=aa/lib/libaa.so bb/lib/libbb.so cc/lib/libcc.so
18288 LIB_DIR=$@{foreach dir,$@{CSC_LIST@},$@{dir@}/lib/lib$@{dir@}.so@}
18290 $@{MAIN@}: objects $@{LIB_DIR@}
18291 gnatbind $@{MAIN@} $@{CSC_LIST:%=-aO%/lib@} -shared
18292 gnatlink $@{MAIN@} $@{CSC_LIST:%=-l%@}
18295 # recompile the sources
18296 gnatmake -c -i $@{MAIN@}.adb $@{NEED_FPIC@} $@{CSC_LIST:%=-I%@}
18298 # Note: In a future version of GNAT, the following commands will be simplified
18299 # by a new tool, gnatmlib
18301 mkdir -p $@{dir $@@ @}
18302 cd $@{dir $@@ @} && gcc -shared -o $@{notdir $@@ @} ../*.o -L$@{GLIB@} -lgnat
18303 cd $@{dir $@@ @} && cp -f ../*.ali .
18305 # The dependencies for the modules
18306 # Note that we have to force the expansion of *.o, since in some cases
18307 # make won't be able to do it itself.
18308 aa/lib/libaa.so: $@{wildcard aa/*.o@}
18309 bb/lib/libbb.so: $@{wildcard bb/*.o@}
18310 cc/lib/libcc.so: $@{wildcard cc/*.o@}
18312 # Make sure all of the shared libraries are in the path before starting the
18315 LD_LIBRARY_PATH=`pwd`/aa/lib:`pwd`/bb/lib:`pwd`/cc/lib ./$@{MAIN@}
18318 $@{RM@} -rf $@{CSC_LIST:%=%/lib@}
18319 $@{RM@} $@{CSC_LIST:%=%/*.ali@}
18320 $@{RM@} $@{CSC_LIST:%=%/*.o@}
18321 $@{RM@} *.o *.ali $@{MAIN@}
18324 @node Automatically Creating a List of Directories
18325 @section Automatically Creating a List of Directories
18328 In most makefiles, you will have to specify a list of directories, and
18329 store it in a variable. For small projects, it is often easier to
18330 specify each of them by hand, since you then have full control over what
18331 is the proper order for these directories, which ones should be
18334 However, in larger projects, which might involve hundreds of
18335 subdirectories, it might be more convenient to generate this list
18338 The example below presents two methods. The first one, although less
18339 general, gives you more control over the list. It involves wildcard
18340 characters, that are automatically expanded by @command{make}. Its
18341 shortcoming is that you need to explicitly specify some of the
18342 organization of your project, such as for instance the directory tree
18343 depth, whether some directories are found in a separate tree, @enddots{}
18345 The second method is the most general one. It requires an external
18346 program, called @command{find}, which is standard on all Unix systems. All
18347 the directories found under a given root directory will be added to the
18353 @font@heightrm=cmr8
18356 # The examples below are based on the following directory hierarchy:
18357 # All the directories can contain any number of files
18358 # ROOT_DIRECTORY -> a -> aa -> aaa
18361 # -> b -> ba -> baa
18364 # This Makefile creates a variable called DIRS, that can be reused any time
18365 # you need this list (see the other examples in this section)
18367 # The root of your project's directory hierarchy
18371 # First method: specify explicitly the list of directories
18372 # This allows you to specify any subset of all the directories you need.
18375 DIRS := a/aa/ a/ab/ b/ba/
18378 # Second method: use wildcards
18379 # Note that the argument(s) to wildcard below should end with a '/'.
18380 # Since wildcards also return file names, we have to filter them out
18381 # to avoid duplicate directory names.
18382 # We thus use make's @code{dir} and @code{sort} functions.
18383 # It sets DIRs to the following value (note that the directories aaa and baa
18384 # are not given, unless you change the arguments to wildcard).
18385 # DIRS= ./a/a/ ./b/ ./a/aa/ ./a/ab/ ./a/ac/ ./b/ba/ ./b/bb/ ./b/bc/
18388 DIRS := $@{sort $@{dir $@{wildcard $@{ROOT_DIRECTORY@}/*/
18389 $@{ROOT_DIRECTORY@}/*/*/@}@}@}
18392 # Third method: use an external program
18393 # This command is much faster if run on local disks, avoiding NFS slowdowns.
18394 # This is the most complete command: it sets DIRs to the following value:
18395 # DIRS= ./a ./a/aa ./a/aa/aaa ./a/ab ./a/ac ./b ./b/ba ./b/ba/baa ./b/bb ./b/bc
18398 DIRS := $@{shell find $@{ROOT_DIRECTORY@} -type d -print@}
18402 @node Generating the Command Line Switches
18403 @section Generating the Command Line Switches
18406 Once you have created the list of directories as explained in the
18407 previous section (@pxref{Automatically Creating a List of Directories}),
18408 you can easily generate the command line arguments to pass to gnatmake.
18410 For the sake of completeness, this example assumes that the source path
18411 is not the same as the object path, and that you have two separate lists
18415 # see "Automatically creating a list of directories" to create
18420 GNATMAKE_SWITCHES := $@{patsubst %,-aI%,$@{SOURCE_DIRS@}@}
18421 GNATMAKE_SWITCHES += $@{patsubst %,-aO%,$@{OBJECT_DIRS@}@}
18424 gnatmake $@{GNATMAKE_SWITCHES@} main_unit
18427 @node Overcoming Command Line Length Limits
18428 @section Overcoming Command Line Length Limits
18431 One problem that might be encountered on big projects is that many
18432 operating systems limit the length of the command line. It is thus hard to give
18433 gnatmake the list of source and object directories.
18435 This example shows how you can set up environment variables, which will
18436 make @command{gnatmake} behave exactly as if the directories had been
18437 specified on the command line, but have a much higher length limit (or
18438 even none on most systems).
18440 It assumes that you have created a list of directories in your Makefile,
18441 using one of the methods presented in
18442 @ref{Automatically Creating a List of Directories}.
18443 For the sake of completeness, we assume that the object
18444 path (where the ALI files are found) is different from the sources patch.
18446 Note a small trick in the Makefile below: for efficiency reasons, we
18447 create two temporary variables (SOURCE_LIST and OBJECT_LIST), that are
18448 expanded immediately by @code{make}. This way we overcome the standard
18449 make behavior which is to expand the variables only when they are
18452 On Windows, if you are using the standard Windows command shell, you must
18453 replace colons with semicolons in the assignments to these variables.
18458 @font@heightrm=cmr8
18461 # In this example, we create both ADA_INCLUDE_PATH and ADA_OBJECTS_PATH.
18462 # This is the same thing as putting the -I arguments on the command line.
18463 # (the equivalent of using -aI on the command line would be to define
18464 # only ADA_INCLUDE_PATH, the equivalent of -aO is ADA_OBJECTS_PATH).
18465 # You can of course have different values for these variables.
18467 # Note also that we need to keep the previous values of these variables, since
18468 # they might have been set before running 'make' to specify where the GNAT
18469 # library is installed.
18471 # see "Automatically creating a list of directories" to create these
18477 space:=$@{empty@} $@{empty@}
18478 SOURCE_LIST := $@{subst $@{space@},:,$@{SOURCE_DIRS@}@}
18479 OBJECT_LIST := $@{subst $@{space@},:,$@{OBJECT_DIRS@}@}
18480 ADA_INCLUDE_PATH += $@{SOURCE_LIST@}
18481 ADA_OBJECTS_PATH += $@{OBJECT_LIST@}
18482 export ADA_INCLUDE_PATH
18483 export ADA_OBJECTS_PATH
18490 @node Memory Management Issues
18491 @chapter Memory Management Issues
18494 This chapter describes some useful memory pools provided in the GNAT library
18495 and in particular the GNAT Debug Pool facility, which can be used to detect
18496 incorrect uses of access values (including ``dangling references'').
18498 @ifclear FSFEDITION
18499 It also describes the @command{gnatmem} tool, which can be used to track down
18505 * Some Useful Memory Pools::
18506 * The GNAT Debug Pool Facility::
18508 @ifclear FSFEDITION
18509 * The gnatmem Tool::
18514 @node Some Useful Memory Pools
18515 @section Some Useful Memory Pools
18516 @findex Memory Pool
18517 @cindex storage, pool
18520 The @code{System.Pool_Global} package offers the Unbounded_No_Reclaim_Pool
18521 storage pool. Allocations use the standard system call @code{malloc} while
18522 deallocations use the standard system call @code{free}. No reclamation is
18523 performed when the pool goes out of scope. For performance reasons, the
18524 standard default Ada allocators/deallocators do not use any explicit storage
18525 pools but if they did, they could use this storage pool without any change in
18526 behavior. That is why this storage pool is used when the user
18527 manages to make the default implicit allocator explicit as in this example:
18528 @smallexample @c ada
18529 type T1 is access Something;
18530 -- no Storage pool is defined for T2
18531 type T2 is access Something_Else;
18532 for T2'Storage_Pool use T1'Storage_Pool;
18533 -- the above is equivalent to
18534 for T2'Storage_Pool use System.Pool_Global.Global_Pool_Object;
18538 The @code{System.Pool_Local} package offers the Unbounded_Reclaim_Pool storage
18539 pool. The allocation strategy is similar to @code{Pool_Local}'s
18540 except that the all
18541 storage allocated with this pool is reclaimed when the pool object goes out of
18542 scope. This pool provides a explicit mechanism similar to the implicit one
18543 provided by several Ada 83 compilers for allocations performed through a local
18544 access type and whose purpose was to reclaim memory when exiting the
18545 scope of a given local access. As an example, the following program does not
18546 leak memory even though it does not perform explicit deallocation:
18548 @smallexample @c ada
18549 with System.Pool_Local;
18550 procedure Pooloc1 is
18551 procedure Internal is
18552 type A is access Integer;
18553 X : System.Pool_Local.Unbounded_Reclaim_Pool;
18554 for A'Storage_Pool use X;
18557 for I in 1 .. 50 loop
18562 for I in 1 .. 100 loop
18569 The @code{System.Pool_Size} package implements the Stack_Bounded_Pool used when
18570 @code{Storage_Size} is specified for an access type.
18571 The whole storage for the pool is
18572 allocated at once, usually on the stack at the point where the access type is
18573 elaborated. It is automatically reclaimed when exiting the scope where the
18574 access type is defined. This package is not intended to be used directly by the
18575 user and it is implicitly used for each such declaration:
18577 @smallexample @c ada
18578 type T1 is access Something;
18579 for T1'Storage_Size use 10_000;
18582 @node The GNAT Debug Pool Facility
18583 @section The GNAT Debug Pool Facility
18585 @cindex storage, pool, memory corruption
18588 The use of unchecked deallocation and unchecked conversion can easily
18589 lead to incorrect memory references. The problems generated by such
18590 references are usually difficult to tackle because the symptoms can be
18591 very remote from the origin of the problem. In such cases, it is
18592 very helpful to detect the problem as early as possible. This is the
18593 purpose of the Storage Pool provided by @code{GNAT.Debug_Pools}.
18595 In order to use the GNAT specific debugging pool, the user must
18596 associate a debug pool object with each of the access types that may be
18597 related to suspected memory problems. See Ada Reference Manual 13.11.
18598 @smallexample @c ada
18599 type Ptr is access Some_Type;
18600 Pool : GNAT.Debug_Pools.Debug_Pool;
18601 for Ptr'Storage_Pool use Pool;
18605 @code{GNAT.Debug_Pools} is derived from a GNAT-specific kind of
18606 pool: the @code{Checked_Pool}. Such pools, like standard Ada storage pools,
18607 allow the user to redefine allocation and deallocation strategies. They
18608 also provide a checkpoint for each dereference, through the use of
18609 the primitive operation @code{Dereference} which is implicitly called at
18610 each dereference of an access value.
18612 Once an access type has been associated with a debug pool, operations on
18613 values of the type may raise four distinct exceptions,
18614 which correspond to four potential kinds of memory corruption:
18617 @code{GNAT.Debug_Pools.Accessing_Not_Allocated_Storage}
18619 @code{GNAT.Debug_Pools.Accessing_Deallocated_Storage}
18621 @code{GNAT.Debug_Pools.Freeing_Not_Allocated_Storage}
18623 @code{GNAT.Debug_Pools.Freeing_Deallocated_Storage }
18627 For types associated with a Debug_Pool, dynamic allocation is performed using
18628 the standard GNAT allocation routine. References to all allocated chunks of
18629 memory are kept in an internal dictionary. Several deallocation strategies are
18630 provided, whereupon the user can choose to release the memory to the system,
18631 keep it allocated for further invalid access checks, or fill it with an easily
18632 recognizable pattern for debug sessions. The memory pattern is the old IBM
18633 hexadecimal convention: @code{16#DEADBEEF#}.
18635 See the documentation in the file g-debpoo.ads for more information on the
18636 various strategies.
18638 Upon each dereference, a check is made that the access value denotes a
18639 properly allocated memory location. Here is a complete example of use of
18640 @code{Debug_Pools}, that includes typical instances of memory corruption:
18641 @smallexample @c ada
18645 with Gnat.Io; use Gnat.Io;
18646 with Unchecked_Deallocation;
18647 with Unchecked_Conversion;
18648 with GNAT.Debug_Pools;
18649 with System.Storage_Elements;
18650 with Ada.Exceptions; use Ada.Exceptions;
18651 procedure Debug_Pool_Test is
18653 type T is access Integer;
18654 type U is access all T;
18656 P : GNAT.Debug_Pools.Debug_Pool;
18657 for T'Storage_Pool use P;
18659 procedure Free is new Unchecked_Deallocation (Integer, T);
18660 function UC is new Unchecked_Conversion (U, T);
18663 procedure Info is new GNAT.Debug_Pools.Print_Info(Put_Line);
18673 Put_Line (Integer'Image(B.all));
18675 when E : others => Put_Line ("raised: " & Exception_Name (E));
18680 when E : others => Put_Line ("raised: " & Exception_Name (E));
18684 Put_Line (Integer'Image(B.all));
18686 when E : others => Put_Line ("raised: " & Exception_Name (E));
18691 when E : others => Put_Line ("raised: " & Exception_Name (E));
18694 end Debug_Pool_Test;
18698 The debug pool mechanism provides the following precise diagnostics on the
18699 execution of this erroneous program:
18702 Total allocated bytes : 0
18703 Total deallocated bytes : 0
18704 Current Water Mark: 0
18708 Total allocated bytes : 8
18709 Total deallocated bytes : 0
18710 Current Water Mark: 8
18713 raised: GNAT.DEBUG_POOLS.ACCESSING_DEALLOCATED_STORAGE
18714 raised: GNAT.DEBUG_POOLS.FREEING_DEALLOCATED_STORAGE
18715 raised: GNAT.DEBUG_POOLS.ACCESSING_NOT_ALLOCATED_STORAGE
18716 raised: GNAT.DEBUG_POOLS.FREEING_NOT_ALLOCATED_STORAGE
18718 Total allocated bytes : 8
18719 Total deallocated bytes : 4
18720 Current Water Mark: 4
18725 @ifclear FSFEDITION
18726 @node The gnatmem Tool
18727 @section The @command{gnatmem} Tool
18731 The @code{gnatmem} utility monitors dynamic allocation and
18732 deallocation activity in a program, and displays information about
18733 incorrect deallocations and possible sources of memory leaks.
18734 It is designed to work in association with a static runtime library
18735 only and in this context provides three types of information:
18738 General information concerning memory management, such as the total
18739 number of allocations and deallocations, the amount of allocated
18740 memory and the high water mark, i.e.@: the largest amount of allocated
18741 memory in the course of program execution.
18744 Backtraces for all incorrect deallocations, that is to say deallocations
18745 which do not correspond to a valid allocation.
18748 Information on each allocation that is potentially the origin of a memory
18753 * Running gnatmem::
18754 * Switches for gnatmem::
18755 * Example of gnatmem Usage::
18758 @node Running gnatmem
18759 @subsection Running @code{gnatmem}
18762 @code{gnatmem} makes use of the output created by the special version of
18763 allocation and deallocation routines that record call information. This
18764 allows to obtain accurate dynamic memory usage history at a minimal cost to
18765 the execution speed. Note however, that @code{gnatmem} is not supported on
18766 all platforms (currently, it is supported on AIX, HP-UX, GNU/Linux,
18767 Solaris and Windows NT/2000/XP (x86).
18770 The @code{gnatmem} command has the form
18773 @c $ gnatmem @ovar{switches} user_program
18774 @c Expanding @ovar macro inline (explanation in macro def comments)
18775 $ gnatmem @r{[}@var{switches}@r{]} @var{user_program}
18779 The program must have been linked with the instrumented version of the
18780 allocation and deallocation routines. This is done by linking with the
18781 @file{libgmem.a} library. For correct symbolic backtrace information,
18782 the user program should be compiled with debugging options
18783 (see @ref{Switches for gcc}). For example to build @file{my_program}:
18786 $ gnatmake -g my_program -largs -lgmem
18790 As library @file{libgmem.a} contains an alternate body for package
18791 @code{System.Memory}, @file{s-memory.adb} should not be compiled and linked
18792 when an executable is linked with library @file{libgmem.a}. It is then not
18793 recommended to use @command{gnatmake} with switch @option{^-a^/ALL_FILES^}.
18796 When @file{my_program} is executed, the file @file{gmem.out} is produced.
18797 This file contains information about all allocations and deallocations
18798 performed by the program. It is produced by the instrumented allocations and
18799 deallocations routines and will be used by @code{gnatmem}.
18801 In order to produce symbolic backtrace information for allocations and
18802 deallocations performed by the GNAT run-time library, you need to use a
18803 version of that library that has been compiled with the @option{-g} switch
18804 (see @ref{Rebuilding the GNAT Run-Time Library}).
18806 Gnatmem must be supplied with the @file{gmem.out} file and the executable to
18807 examine. If the location of @file{gmem.out} file was not explicitly supplied by
18808 @option{-i} switch, gnatmem will assume that this file can be found in the
18809 current directory. For example, after you have executed @file{my_program},
18810 @file{gmem.out} can be analyzed by @code{gnatmem} using the command:
18813 $ gnatmem my_program
18817 This will produce the output with the following format:
18819 *************** debut cc
18821 $ gnatmem my_program
18825 Total number of allocations : 45
18826 Total number of deallocations : 6
18827 Final Water Mark (non freed mem) : 11.29 Kilobytes
18828 High Water Mark : 11.40 Kilobytes
18833 Allocation Root # 2
18834 -------------------
18835 Number of non freed allocations : 11
18836 Final Water Mark (non freed mem) : 1.16 Kilobytes
18837 High Water Mark : 1.27 Kilobytes
18839 my_program.adb:23 my_program.alloc
18845 The first block of output gives general information. In this case, the
18846 Ada construct ``@code{@b{new}}'' was executed 45 times, and only 6 calls to an
18847 Unchecked_Deallocation routine occurred.
18850 Subsequent paragraphs display information on all allocation roots.
18851 An allocation root is a specific point in the execution of the program
18852 that generates some dynamic allocation, such as a ``@code{@b{new}}''
18853 construct. This root is represented by an execution backtrace (or subprogram
18854 call stack). By default the backtrace depth for allocations roots is 1, so
18855 that a root corresponds exactly to a source location. The backtrace can
18856 be made deeper, to make the root more specific.
18858 @node Switches for gnatmem
18859 @subsection Switches for @code{gnatmem}
18862 @code{gnatmem} recognizes the following switches:
18867 @cindex @option{-q} (@code{gnatmem})
18868 Quiet. Gives the minimum output needed to identify the origin of the
18869 memory leaks. Omits statistical information.
18872 @cindex @var{N} (@code{gnatmem})
18873 N is an integer literal (usually between 1 and 10) which controls the
18874 depth of the backtraces defining allocation root. The default value for
18875 N is 1. The deeper the backtrace, the more precise the localization of
18876 the root. Note that the total number of roots can depend on this
18877 parameter. This parameter must be specified @emph{before} the name of the
18878 executable to be analyzed, to avoid ambiguity.
18881 @cindex @option{-b} (@code{gnatmem})
18882 This switch has the same effect as just depth parameter.
18884 @item -i @var{file}
18885 @cindex @option{-i} (@code{gnatmem})
18886 Do the @code{gnatmem} processing starting from @file{file}, rather than
18887 @file{gmem.out} in the current directory.
18890 @cindex @option{-m} (@code{gnatmem})
18891 This switch causes @code{gnatmem} to mask the allocation roots that have less
18892 than n leaks. The default value is 1. Specifying the value of 0 will allow to
18893 examine even the roots that didn't result in leaks.
18896 @cindex @option{-s} (@code{gnatmem})
18897 This switch causes @code{gnatmem} to sort the allocation roots according to the
18898 specified order of sort criteria, each identified by a single letter. The
18899 currently supported criteria are @code{n, h, w} standing respectively for
18900 number of unfreed allocations, high watermark, and final watermark
18901 corresponding to a specific root. The default order is @code{nwh}.
18905 @node Example of gnatmem Usage
18906 @subsection Example of @code{gnatmem} Usage
18909 The following example shows the use of @code{gnatmem}
18910 on a simple memory-leaking program.
18911 Suppose that we have the following Ada program:
18913 @smallexample @c ada
18916 with Unchecked_Deallocation;
18917 procedure Test_Gm is
18919 type T is array (1..1000) of Integer;
18920 type Ptr is access T;
18921 procedure Free is new Unchecked_Deallocation (T, Ptr);
18924 procedure My_Alloc is
18929 procedure My_DeAlloc is
18937 for I in 1 .. 5 loop
18938 for J in I .. 5 loop
18949 The program needs to be compiled with debugging option and linked with
18950 @code{gmem} library:
18953 $ gnatmake -g test_gm -largs -lgmem
18957 Then we execute the program as usual:
18964 Then @code{gnatmem} is invoked simply with
18970 which produces the following output (result may vary on different platforms):
18975 Total number of allocations : 18
18976 Total number of deallocations : 5
18977 Final Water Mark (non freed mem) : 53.00 Kilobytes
18978 High Water Mark : 56.90 Kilobytes
18980 Allocation Root # 1
18981 -------------------
18982 Number of non freed allocations : 11
18983 Final Water Mark (non freed mem) : 42.97 Kilobytes
18984 High Water Mark : 46.88 Kilobytes
18986 test_gm.adb:11 test_gm.my_alloc
18988 Allocation Root # 2
18989 -------------------
18990 Number of non freed allocations : 1
18991 Final Water Mark (non freed mem) : 10.02 Kilobytes
18992 High Water Mark : 10.02 Kilobytes
18994 s-secsta.adb:81 system.secondary_stack.ss_init
18996 Allocation Root # 3
18997 -------------------
18998 Number of non freed allocations : 1
18999 Final Water Mark (non freed mem) : 12 Bytes
19000 High Water Mark : 12 Bytes
19002 s-secsta.adb:181 system.secondary_stack.ss_init
19006 Note that the GNAT run time contains itself a certain number of
19007 allocations that have no corresponding deallocation,
19008 as shown here for root #2 and root
19009 #3. This is a normal behavior when the number of non-freed allocations
19010 is one, it allocates dynamic data structures that the run time needs for
19011 the complete lifetime of the program. Note also that there is only one
19012 allocation root in the user program with a single line back trace:
19013 test_gm.adb:11 test_gm.my_alloc, whereas a careful analysis of the
19014 program shows that 'My_Alloc' is called at 2 different points in the
19015 source (line 21 and line 24). If those two allocation roots need to be
19016 distinguished, the backtrace depth parameter can be used:
19019 $ gnatmem 3 test_gm
19023 which will give the following output:
19028 Total number of allocations : 18
19029 Total number of deallocations : 5
19030 Final Water Mark (non freed mem) : 53.00 Kilobytes
19031 High Water Mark : 56.90 Kilobytes
19033 Allocation Root # 1
19034 -------------------
19035 Number of non freed allocations : 10
19036 Final Water Mark (non freed mem) : 39.06 Kilobytes
19037 High Water Mark : 42.97 Kilobytes
19039 test_gm.adb:11 test_gm.my_alloc
19040 test_gm.adb:24 test_gm
19041 b_test_gm.c:52 main
19043 Allocation Root # 2
19044 -------------------
19045 Number of non freed allocations : 1
19046 Final Water Mark (non freed mem) : 10.02 Kilobytes
19047 High Water Mark : 10.02 Kilobytes
19049 s-secsta.adb:81 system.secondary_stack.ss_init
19050 s-secsta.adb:283 <system__secondary_stack___elabb>
19051 b_test_gm.c:33 adainit
19053 Allocation Root # 3
19054 -------------------
19055 Number of non freed allocations : 1
19056 Final Water Mark (non freed mem) : 3.91 Kilobytes
19057 High Water Mark : 3.91 Kilobytes
19059 test_gm.adb:11 test_gm.my_alloc
19060 test_gm.adb:21 test_gm
19061 b_test_gm.c:52 main
19063 Allocation Root # 4
19064 -------------------
19065 Number of non freed allocations : 1
19066 Final Water Mark (non freed mem) : 12 Bytes
19067 High Water Mark : 12 Bytes
19069 s-secsta.adb:181 system.secondary_stack.ss_init
19070 s-secsta.adb:283 <system__secondary_stack___elabb>
19071 b_test_gm.c:33 adainit
19075 The allocation root #1 of the first example has been split in 2 roots #1
19076 and #3 thanks to the more precise associated backtrace.
19080 @node Stack Related Facilities
19081 @chapter Stack Related Facilities
19084 This chapter describes some useful tools associated with stack
19085 checking and analysis. In
19086 particular, it deals with dynamic and static stack usage measurements.
19089 * Stack Overflow Checking::
19090 * Static Stack Usage Analysis::
19091 * Dynamic Stack Usage Analysis::
19094 @node Stack Overflow Checking
19095 @section Stack Overflow Checking
19096 @cindex Stack Overflow Checking
19097 @cindex -fstack-check
19100 For most operating systems, @command{gcc} does not perform stack overflow
19101 checking by default. This means that if the main environment task or
19102 some other task exceeds the available stack space, then unpredictable
19103 behavior will occur. Most native systems offer some level of protection by
19104 adding a guard page at the end of each task stack. This mechanism is usually
19105 not enough for dealing properly with stack overflow situations because
19106 a large local variable could ``jump'' above the guard page.
19107 Furthermore, when the
19108 guard page is hit, there may not be any space left on the stack for executing
19109 the exception propagation code. Enabling stack checking avoids
19112 To activate stack checking, compile all units with the gcc option
19113 @option{-fstack-check}. For example:
19116 gcc -c -fstack-check package1.adb
19120 Units compiled with this option will generate extra instructions to check
19121 that any use of the stack (for procedure calls or for declaring local
19122 variables in declare blocks) does not exceed the available stack space.
19123 If the space is exceeded, then a @code{Storage_Error} exception is raised.
19125 For declared tasks, the stack size is controlled by the size
19126 given in an applicable @code{Storage_Size} pragma or by the value specified
19127 at bind time with @option{-d} (@pxref{Switches for gnatbind}) or is set to
19128 the default size as defined in the GNAT runtime otherwise.
19130 For the environment task, the stack size depends on
19131 system defaults and is unknown to the compiler. Stack checking
19132 may still work correctly if a fixed
19133 size stack is allocated, but this cannot be guaranteed.
19135 To ensure that a clean exception is signalled for stack
19136 overflow, set the environment variable
19137 @env{GNAT_STACK_LIMIT} to indicate the maximum
19138 stack area that can be used, as in:
19139 @cindex GNAT_STACK_LIMIT
19142 SET GNAT_STACK_LIMIT 1600
19146 The limit is given in kilobytes, so the above declaration would
19147 set the stack limit of the environment task to 1.6 megabytes.
19148 Note that the only purpose of this usage is to limit the amount
19149 of stack used by the environment task. If it is necessary to
19150 increase the amount of stack for the environment task, then this
19151 is an operating systems issue, and must be addressed with the
19152 appropriate operating systems commands.
19155 To have a fixed size stack in the environment task, the stack must be put
19156 in the P0 address space and its size specified. Use these switches to
19160 gnatmake my_progs -largs "-Wl,--opt=STACK=4000,/p0image"
19164 The quotes are required to keep case. The number after @samp{STACK=} is the
19165 size of the environmental task stack in pagelets (512 bytes). In this example
19166 the stack size is about 2 megabytes.
19169 A consequence of the @option{/p0image} qualifier is also to makes RMS buffers
19170 be placed in P0 space. Refer to @cite{HP OpenVMS Linker Utility Manual} for
19171 more details about the @option{/p0image} qualifier and the @option{stack}
19175 On Itanium platforms, you can instead assign the @samp{GNAT_STACK_SIZE} and
19176 @samp{GNAT_RBS_SIZE} logicals to the size of the primary and register
19177 stack in kilobytes. For example:
19180 $ define GNAT_RBS_SIZE 1024 ! Limit the RBS size to 1MB.
19184 @node Static Stack Usage Analysis
19185 @section Static Stack Usage Analysis
19186 @cindex Static Stack Usage Analysis
19187 @cindex -fstack-usage
19190 A unit compiled with @option{-fstack-usage} will generate an extra file
19192 the maximum amount of stack used, on a per-function basis.
19193 The file has the same
19194 basename as the target object file with a @file{.su} extension.
19195 Each line of this file is made up of three fields:
19199 The name of the function.
19203 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
19206 The second field corresponds to the size of the known part of the function
19209 The qualifier @code{static} means that the function frame size
19211 It usually means that all local variables have a static size.
19212 In this case, the second field is a reliable measure of the function stack
19215 The qualifier @code{dynamic} means that the function frame size is not static.
19216 It happens mainly when some local variables have a dynamic size. When this
19217 qualifier appears alone, the second field is not a reliable measure
19218 of the function stack analysis. When it is qualified with @code{bounded}, it
19219 means that the second field is a reliable maximum of the function stack
19222 A unit compiled with @option{-Wstack-usage} will issue a warning for each
19223 subprogram whose stack usage might be larger than the specified amount of
19224 bytes. The wording is in keeping with the qualifier documented above.
19226 @node Dynamic Stack Usage Analysis
19227 @section Dynamic Stack Usage Analysis
19230 It is possible to measure the maximum amount of stack used by a task, by
19231 adding a switch to @command{gnatbind}, as:
19234 $ gnatbind -u0 file
19238 With this option, at each task termination, its stack usage is output on
19240 It is not always convenient to output the stack usage when the program
19241 is still running. Hence, it is possible to delay this output until program
19242 termination. for a given number of tasks specified as the argument of the
19243 @option{-u} option. For instance:
19246 $ gnatbind -u100 file
19250 will buffer the stack usage information of the first 100 tasks to terminate and
19251 output this info at program termination. Results are displayed in four
19255 Index | Task Name | Stack Size | Stack Usage
19262 is a number associated with each task.
19265 is the name of the task analyzed.
19268 is the maximum size for the stack.
19271 is the measure done by the stack analyzer. In order to prevent overflow, the stack
19272 is not entirely analyzed, and it's not possible to know exactly how
19273 much has actually been used.
19278 The environment task stack, e.g., the stack that contains the main unit, is
19279 only processed when the environment variable GNAT_STACK_LIMIT is set.
19282 The package @code{GNAT.Task_Stack_Usage} provides facilities to get
19283 stack usage reports at run-time. See its body for the details.
19285 @ifclear FSFEDITION
19286 @c *********************************
19288 @c *********************************
19289 @node Verifying Properties with gnatcheck
19290 @chapter Verifying Properties with @command{gnatcheck}
19292 @cindex @command{gnatcheck}
19295 The @command{gnatcheck} tool is an ASIS-based utility that checks properties
19296 of Ada source files according to a given set of semantic rules.
19299 In order to check compliance with a given rule, @command{gnatcheck} has to
19300 semantically analyze the Ada sources.
19301 Therefore, checks can only be performed on
19302 legal Ada units. Moreover, when a unit depends semantically upon units located
19303 outside the current directory, the source search path has to be provided when
19304 calling @command{gnatcheck}, either through a specified project file or
19305 through @command{gnatcheck} switches.
19307 For full details, refer to @cite{GNATcheck Reference Manual} document.
19310 @ifclear FSFEDITION
19311 @c *********************************
19312 @node Creating Sample Bodies with gnatstub
19313 @chapter Creating Sample Bodies with @command{gnatstub}
19317 @command{gnatstub} creates body stubs, that is, empty but compilable bodies
19318 for library unit declarations.
19320 To create a body stub, @command{gnatstub} invokes the Ada
19321 compiler and generates and uses the ASIS tree for the input source;
19322 thus the input must be legal Ada code, and the tool should have all the
19323 information needed to compile the input source. To provide this information,
19324 you may specify as a tool parameter the project file the input source belongs to
19325 (or you may call @command{gnatstub}
19326 through the @command{gnat} driver (see @ref{The GNAT Driver and
19327 Project Files}). Another possibility is to specify the source search
19328 path and needed configuration files in @option{-cargs} section of @command{gnatstub}
19329 call, see the description of the @command{gnatstub} switches below.
19331 By default, all the program unit body stubs generated by @code{gnatstub}
19332 raise the predefined @code{Program_Error} exception, which will catch
19333 accidental calls of generated stubs. This behavior can be changed with
19334 option @option{^--no-exception^/NO_EXCEPTION^} (see below).
19337 * Running gnatstub::
19338 * Switches for gnatstub::
19341 @node Running gnatstub
19342 @section Running @command{gnatstub}
19345 @command{gnatstub} has a command-line interface of the form:
19348 @c $ gnatstub @ovar{switches} @var{filename} @ovar{directory}
19349 @c Expanding @ovar macro inline (explanation in macro def comments)
19350 $ gnatstub @r{[}@var{switches}@r{]} @var{filename} @r{[}@var{directory}@r{]} @r{[}-cargs @var{gcc_switches}@r{]}
19357 is the name of the source file that contains a library unit declaration
19358 for which a body must be created. The file name may contain the path
19360 The file name does not have to follow the GNAT file name conventions. If the
19362 does not follow GNAT file naming conventions, the name of the body file must
19364 explicitly as the value of the @option{^-o^/BODY=^@var{body-name}} option.
19365 If the file name follows the GNAT file naming
19366 conventions and the name of the body file is not provided,
19369 of the body file from the argument file name by replacing the @file{.ads}
19371 with the @file{.adb} suffix.
19374 indicates the directory in which the body stub is to be placed (the default
19378 @item @samp{@var{gcc_switches}} is a list of switches for
19379 @command{gcc}. They will be passed on to all compiler invocations made by
19380 @command{gnatstub} to generate the ASIS trees. Here you can provide
19381 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
19382 use the @option{-gnatec} switch to set the configuration file,
19383 use the @option{-gnat05} switch if sources should be compiled in
19387 is an optional sequence of switches as described in the next section
19390 @node Switches for gnatstub
19391 @section Switches for @command{gnatstub}
19397 @cindex @option{--version} @command{gnatstub}
19398 Display Copyright and version, then exit disregarding all other options.
19401 @cindex @option{--help} @command{gnatstub}
19402 Display usage, then exit disregarding all other options.
19404 @item -P @var{file}
19405 @cindex @option{-P} @command{gnatstub}
19406 Indicates the name of the project file that describes the set of sources
19409 @item -X@var{name}=@var{value}
19410 @cindex @option{-X} @command{gnatstub}
19411 Indicates that external variable @var{name} in the argument project
19412 has the value @var{value}. Has no effect if no project is specified as
19416 @cindex @option{^-f^/FULL^} (@command{gnatstub})
19417 If the destination directory already contains a file with the name of the
19419 for the argument spec file, replace it with the generated body stub.
19421 @item ^-hs^/HEADER=SPEC^
19422 @cindex @option{^-hs^/HEADER=SPEC^} (@command{gnatstub})
19423 Put the comment header (i.e., all the comments preceding the
19424 compilation unit) from the source of the library unit declaration
19425 into the body stub.
19427 @item ^-hg^/HEADER=GENERAL^
19428 @cindex @option{^-hg^/HEADER=GENERAL^} (@command{gnatstub})
19429 Put a sample comment header into the body stub.
19431 @item ^--header-file=@var{filename}^/FROM_HEADER_FILE=@var{filename}^
19432 @cindex @option{^--header-file^/FROM_HEADER_FILE=^} (@command{gnatstub})
19433 Use the content of the file as the comment header for a generated body stub.
19437 @cindex @option{-IDIR} (@command{gnatstub})
19439 @cindex @option{-I-} (@command{gnatstub})
19442 @item /NOCURRENT_DIRECTORY
19443 @cindex @option{/NOCURRENT_DIRECTORY} (@command{gnatstub})
19445 ^These switches have ^This switch has^ the same meaning as in calls to
19447 ^They define ^It defines ^ the source search path in the call to
19448 @command{gcc} issued
19449 by @command{gnatstub} to compile an argument source file.
19451 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE=^@var{PATH}
19452 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@command{gnatstub})
19453 This switch has the same meaning as in calls to @command{gcc}.
19454 It defines the additional configuration file to be passed to the call to
19455 @command{gcc} issued
19456 by @command{gnatstub} to compile an argument source file.
19458 @item ^-gnatyM^/MAX_LINE_LENGTH=^@var{n}
19459 @cindex @option{^-gnatyM^/MAX_LINE_LENGTH^} (@command{gnatstub})
19460 (@var{n} is a non-negative integer). Set the maximum line length that is
19461 allowed in a source file. The default is 79. The maximum value that can be
19462 specified is 32767. Note that in the special case of configuration
19463 pragma files, the maximum is always 32767 regardless of whether or
19464 not this switch appears.
19466 @item ^-gnaty^/STYLE_CHECKS=^@var{n}
19467 @cindex @option{^-gnaty^/STYLE_CHECKS=^} (@command{gnatstub})
19468 (@var{n} is a non-negative integer from 1 to 9). Set the indentation level in
19469 the generated body sample to @var{n}.
19470 The default indentation is 3.
19472 @item ^-gnatyo^/ORDERED_SUBPROGRAMS^
19473 @cindex @option{^-gnatyo^/ORDERED_SUBPROGRAMS^} (@command{gnatstub})
19474 Order local bodies alphabetically. (By default local bodies are ordered
19475 in the same way as the corresponding local specs in the argument spec file.)
19477 @item ^-i^/INDENTATION=^@var{n}
19478 @cindex @option{^-i^/INDENTATION^} (@command{gnatstub})
19479 Same as @option{^-gnaty^/STYLE_CHECKS=^@var{n}}
19481 @item ^-k^/TREE_FILE=SAVE^
19482 @cindex @option{^-k^/TREE_FILE=SAVE^} (@command{gnatstub})
19483 Do not remove the tree file (i.e., the snapshot of the compiler internal
19484 structures used by @command{gnatstub}) after creating the body stub.
19486 @item ^-l^/LINE_LENGTH=^@var{n}
19487 @cindex @option{^-l^/LINE_LENGTH^} (@command{gnatstub})
19488 Same as @option{^-gnatyM^/MAX_LINE_LENGTH=^@var{n}}
19490 @item ^--no-exception^/NO_EXCEPTION^
19491 @cindex @option{^--no-exception^/NO_EXCEPTION^} (@command{gnatstub})
19492 Avoid raising PROGRAM_ERROR in the generated bodies of program unit stubs.
19493 This is not always possible for function stubs.
19495 @item ^--no-local-header^/NO_LOCAL_HEADER^
19496 @cindex @option{^--no-local-header^/NO_LOCAL_HEADER^} (@command{gnatstub})
19497 Do not place local comment header with unit name before body stub for a
19500 @item ^-o ^/BODY=^@var{body-name}
19501 @cindex @option{^-o^/BODY^} (@command{gnatstub})
19502 Body file name. This should be set if the argument file name does not
19504 the GNAT file naming
19505 conventions. If this switch is omitted the default name for the body will be
19507 from the argument file name according to the GNAT file naming conventions.
19509 @item ^-W^/RESULT_ENCODING=^@var{e}
19510 @cindex @option{^-W^/RESULT_ENCODING=^} (@command{gnatstub})
19511 Specify the wide character encoding method for the output body file.
19512 @var{e} is one of the following:
19520 Upper half encoding
19522 @item ^s^SHIFT_JIS^
19532 Brackets encoding (default value)
19536 @cindex @option{^-q^/QUIET^} (@command{gnatstub})
19537 Quiet mode: do not generate a confirmation when a body is
19538 successfully created, and do not generate a message when a body is not
19542 @item ^-r^/TREE_FILE=REUSE^
19543 @cindex @option{^-r^/TREE_FILE=REUSE^} (@command{gnatstub})
19544 Reuse the tree file (if it exists) instead of creating it. Instead of
19545 creating the tree file for the library unit declaration, @command{gnatstub}
19546 tries to find it in the current directory and use it for creating
19547 a body. If the tree file is not found, no body is created. This option
19548 also implies @option{^-k^/SAVE^}, whether or not
19549 the latter is set explicitly.
19551 @item ^-t^/TREE_FILE=OVERWRITE^
19552 @cindex @option{^-t^/TREE_FILE=OVERWRITE^} (@command{gnatstub})
19553 Overwrite the existing tree file. If the current directory already
19554 contains the file which, according to the GNAT file naming rules should
19555 be considered as a tree file for the argument source file,
19557 will refuse to create the tree file needed to create a sample body
19558 unless this option is set.
19560 @item ^-v^/VERBOSE^
19561 @cindex @option{^-v^/VERBOSE^} (@command{gnatstub})
19562 Verbose mode: generate version information.
19567 @ifclear FSFEDITION
19568 @c *********************************
19569 @node Creating Unit Tests with gnattest
19570 @chapter Creating Unit Tests with @command{gnattest}
19574 @command{gnattest} is an ASIS-based utility that creates unit-test skeletons
19575 as well as a test driver infrastructure (harness). @command{gnattest} creates
19576 a skeleton for each visible subprogram in the packages under consideration when
19577 they do not exist already.
19579 In order to process source files from a project, @command{gnattest} has to
19580 semantically analyze the sources. Therefore, test skeletons can only be
19581 generated for legal Ada units. If a unit is dependent on other units,
19582 those units should be among the source files of the project or of other projects
19583 imported by this one.
19585 Generated skeletons and harnesses are based on the AUnit testing framework.
19586 AUnit is an Ada adaptation of the xxxUnit testing frameworks, similar to JUnit
19587 for Java or CppUnit for C++. While it is advised that gnattest users read
19588 the AUnit manual, deep knowledge of AUnit is not necessary for using gnattest.
19589 For correct operation of @command{gnattest}, AUnit should be installed and
19590 aunit.gpr must be on the project path. This happens automatically when Aunit
19591 is installed at its default location.
19593 * Running gnattest::
19594 * Switches for gnattest::
19595 * Project Attributes for gnattest::
19597 * Setting Up and Tearing Down the Testing Environment::
19598 * Regenerating Tests::
19599 * Default Test Behavior::
19600 * Testing Primitive Operations of Tagged Types::
19601 * Testing Inheritance::
19602 * Tagged Types Substitutability Testing::
19603 * Testing with Contracts::
19604 * Additional Tests::
19606 * Support for other platforms/run-times::
19608 * Current Limitations::
19611 @node Running gnattest
19612 @section Running @command{gnattest}
19615 @command{gnattest} has a command-line interface of the form
19618 @c $ gnattest @var{-Pprojname} @ovar{switches} @ovar{filename} @ovar{directory}
19619 @c Expanding @ovar macro inline (explanation in macro def comments)
19620 $ gnattest @var{-Pprojname} @r{[}@var{--harness-dir=dirname}@r{]} @r{[}@var{switches}@r{]} @r{[}@var{filename}@r{]} @r{[}-cargs @var{gcc_switches}@r{]}
19628 specifies the project defining the location of source files. When no
19629 file names are provided on the command line, all sources in the project
19630 are used as input. This switch is required.
19633 is the name of the source file containing the library unit package declaration
19634 for which a test package will be created. The file name may be given with a
19637 @item @samp{@var{gcc_switches}}
19638 is a list of switches for
19639 @command{gcc}. These switches will be passed on to all compiler invocations
19640 made by @command{gnattest} to generate a set of ASIS trees. Here you can provide
19641 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
19642 use the @option{-gnatec} switch to set the configuration file,
19643 use the @option{-gnat05} switch if sources should be compiled in
19644 Ada 2005 mode, etc.
19647 is an optional sequence of switches as described in the next section.
19651 @command{gnattest} results can be found in two different places.
19654 @item automatic harness:
19655 the harness code, which is located by default in "gnattest/harness" directory
19656 that is created in the object directory of corresponding project file. All of
19657 this code is generated completely automatically and can be destroyed and
19658 regenerated at will. It is not recommended to modify this code manually, since
19659 it could easily be overridden by mistake. The entry point in the harness code is
19660 the project file named @command{test_driver.gpr}. Tests can be compiled and run
19661 using a command such as:
19664 gnatmake -P<harness-dir>/test_driver
19668 Note that you might need to specify the necessary values of scenario variables
19669 when you are not using the AUnit defaults.
19671 @item actual unit test skeletons:
19672 a test skeleton for each visible subprogram is created in a separate file, if it
19673 doesn't exist already. By default, those separate test files are located in a
19674 "gnattest/tests" directory that is created in the object directory of
19675 corresponding project file. For example, if a source file my_unit.ads in
19676 directory src contains a visible subprogram Proc, then the corresponding unit
19677 test will be found in file src/tests/my_unit-test_data-tests-proc_<code>.adb.
19678 <code> is a signature encoding used to differentiate test names in case of
19681 Note that if the project already has both my_unit.ads and my_unit-test_data.ads,
19682 this will cause a name conflict with the generated test package.
19685 @node Switches for gnattest
19686 @section Switches for @command{gnattest}
19691 @item --harness-only
19692 @cindex @option{--harness-only} (@command{gnattest})
19693 When this option is given, @command{gnattest} creates a harness for all
19694 sources, treating them as test packages.
19696 @item --additional-tests=@var{projname}
19697 @cindex @option{--additional-tests} (@command{gnattest})
19698 Sources described in @var{projname} are considered potential additional
19699 manual tests to be added to the test suite.
19702 @cindex @option{-r} (@command{gnattest})
19703 Recursively consider all sources from all projects.
19705 @item -X@var{name=value}
19706 @cindex @option{-X} (@command{gnattest})
19707 Indicate that external variable @var{name} has the value @var{value}.
19710 @cindex @option{-q} (@command{gnattest})
19711 Suppresses noncritical output messages.
19714 @cindex @option{-v} (@command{gnattest})
19715 Verbose mode: generates version information.
19717 @item --validate-type-extensions
19718 @cindex @option{--validate-type-extensions} (@command{gnattest})
19719 Enables substitution check: run all tests from all parents in order
19720 to check substitutability.
19722 @item --skeleton-default=@var{val}
19723 @cindex @option{--skeleton-default} (@command{gnattest})
19724 Specifies the default behavior of generated skeletons. @var{val} can be either
19725 "fail" or "pass", "fail" being the default.
19727 @item --passed-tests=@var{val}
19728 @cindex @option{--skeleton-default} (@command{gnattest})
19729 Specifies whether or not passed tests should be shown. @var{val} can be either
19730 "show" or "hide", "show" being the default.
19733 @item --tests-root=@var{dirname}
19734 @cindex @option{--tests-root} (@command{gnattest})
19735 The directory hierarchy of tested sources is recreated in the @var{dirname}
19736 directory, and test packages are placed in corresponding directories.
19737 If the @var{dirname} is a relative path, it is considered relative to the object
19738 directory of the project file. When all sources from all projects are taken
19739 recursively from all projects, directory hierarchies of tested sources are
19740 recreated for each project in their object directories and test packages are
19741 placed accordingly.
19743 @item --subdir=@var{dirname}
19744 @cindex @option{--subdir} (@command{gnattest})
19745 Test packages are placed in subdirectories.
19747 @item --tests-dir=@var{dirname}
19748 @cindex @option{--tests-dir} (@command{gnattest})
19749 All test packages are placed in the @var{dirname} directory.
19750 If the @var{dirname} is a relative path, it is considered relative to the object
19751 directory of the project file. When all sources from all projects are taken
19752 recursively from all projects, @var{dirname} directories are created for each
19753 project in their object directories and test packages are placed accordingly.
19755 @item --harness-dir=@var{dirname}
19756 @cindex @option{--harness-dir} (@command{gnattest})
19757 specifies the directory that will hold the harness packages and project file
19758 for the test driver. If the @var{dirname} is a relative path, it is considered
19759 relative to the object directory of the project file.
19762 @cindex @option{--separates} (@command{gnattest})
19763 Bodies of all test routines are generated as separates. Note that this mode is
19764 kept for compatibility reasons only and it is not advised to use it due to
19765 possible problems with hash in names of test skeletons when using an
19766 inconsistent casing. Separate test skeletons can be incorporated to monolith
19767 test package with improved hash being used by using @option{--transition}
19772 @cindex @option{--transition} (@command{gnattest})
19773 This allows transition from separate test routines to monolith test packages.
19774 All matching test routines are overwritten with contents of corresponding
19775 separates. Note that if separate test routines had any manually added with
19776 clauses they will be moved to the test package body as is and have to be moved
19781 @option{--tests_root}, @option{--subdir} and @option{--tests-dir} switches are
19782 mutually exclusive.
19784 @node Project Attributes for gnattest
19785 @section Project Attributes for @command{gnattest}
19789 Most of the command-line options can also be passed to the tool by adding
19790 special attributes to the project file. Those attributes should be put in
19791 package gnattest. Here is the list of attributes:
19796 is used to select the same output mode as with the --tests-root option.
19797 This attribute cannot be used together with Subdir or Tests_Dir.
19800 is used to select the same output mode as with the --subdir option.
19801 This attribute cannot be used together with Tests_Root or Tests_Dir.
19804 is used to select the same output mode as with the --tests-dir option.
19805 This attribute cannot be used together with Subdir or Tests_Root.
19808 is used to specify the directory in which to place harness packages and project
19809 file for the test driver, otherwise specified by --harness-dir.
19811 @item Additional_Tests
19812 is used to specify the project file, otherwise given by
19813 --additional-tests switch.
19815 @item Skeletons_Default
19816 is used to specify the default behaviour of test skeletons, otherwise
19817 specified by --skeleton-default option. The value of this attribute
19818 should be either "pass" or "fail".
19822 Each of those attributes can be overridden from the command line if needed.
19823 Other @command{gnattest} switches can also be passed via the project
19824 file as an attribute list called GNATtest_Switches.
19826 @node Simple Example
19827 @section Simple Example
19831 Let's take a very simple example using the first @command{gnattest} example
19835 <install_prefix>/share/examples/gnattest/simple
19838 This project contains a simple package containing one subprogram. By running gnattest:
19841 $ gnattest --harness-dir=driver -Psimple.gpr
19844 a test driver is created in directory "driver". It can be compiled and run:
19848 $ gnatmake -Ptest_driver
19852 One failed test with diagnosis "test not implemented" is reported.
19853 Since no special output option was specified, the test package Simple.Tests
19857 <install_prefix>/share/examples/gnattest/simple/obj/gnattest/tests
19860 For each package containing visible subprograms, a child test package is
19861 generated. It contains one test routine per tested subprogram. Each
19862 declaration of a test subprogram has a comment specifying which tested
19863 subprogram it corresponds to. Bodies of test routines are placed in test package
19864 bodies and are surrounded by special comment sections. Those comment sections
19865 should not be removed or modified in order for gnattest to be able to regenerate
19866 test packages and keep already written tests in place.
19867 The test routine Test_Inc_5eaee3 located at simple-test_data-tests.adb contains
19868 a single statement: a call to procedure Assert. It has two arguments:
19869 the Boolean expression we want to check and the diagnosis message to display if
19870 the condition is false.
19872 That is where actual testing code should be written after a proper setup.
19873 An actual check can be performed by replacing the Assert call with:
19875 @smallexample @c ada
19876 Assert (Inc (1) = 2, "wrong incrementation");
19879 After recompiling and running the test driver, one successfully passed test
19882 @node Setting Up and Tearing Down the Testing Environment
19883 @section Setting Up and Tearing Down the Testing Environment
19887 Besides test routines themselves, each test package has a parent package
19888 Test_Data that has two procedures: Set_Up and Tear_Down. This package is never
19889 overwritten by the tool. Set_Up is called before each test routine of the
19890 package and Tear_Down is called after each test routine. Those two procedures
19891 can be used to perform necessary initialization and finalization,
19892 memory allocation, etc. Test type declared in Test_Data package is parent type
19893 for the test type of test package and can have user-defined components whose
19894 values can be set by Set_Up routine and used in test routines afterwards.
19896 @node Regenerating Tests
19897 @section Regenerating Tests
19901 Bodies of test routines and test_data packages are never overridden after they
19902 have been created once. As long as the name of the subprogram, full expanded Ada
19903 names, and the order of its parameters is the same, and comment sections are
19904 intact the old test routine will fit in its place and no test skeleton will be
19905 generated for the subprogram.
19907 This can be demonstrated with the previous example. By uncommenting declaration
19908 and body of function Dec in simple.ads and simple.adb, running
19909 @command{gnattest} on the project, and then running the test driver:
19912 gnattest --harness-dir=driver -Psimple.gpr
19914 gnatmake -Ptest_driver
19918 the old test is not replaced with a stub, nor is it lost, but a new test
19919 skeleton is created for function Dec.
19921 The only way of regenerating tests skeletons is to remove the previously created
19922 tests together with corresponding comment sections.
19924 @node Default Test Behavior
19925 @section Default Test Behavior
19929 The generated test driver can treat unimplemented tests in two ways:
19930 either count them all as failed (this is useful to see which tests are still
19931 left to implement) or as passed (to sort out unimplemented ones from those
19934 The test driver accepts a switch to specify this behavior:
19935 --skeleton-default=val, where val is either "pass" or "fail" (exactly as for
19936 @command{gnattest}).
19938 The default behavior of the test driver is set with the same switch
19939 as passed to gnattest when generating the test driver.
19941 Passing it to the driver generated on the first example:
19944 test_runner --skeleton-default=pass
19947 makes both tests pass, even the unimplemented one.
19949 @node Testing Primitive Operations of Tagged Types
19950 @section Testing Primitive Operations of Tagged Types
19954 Creation of test skeletons for primitive operations of tagged types entails
19955 a number of features. Test routines for all primitives of a given tagged type
19956 are placed in a separate child package named according to the tagged type. For
19957 example, if you have tagged type T in package P, all tests for primitives
19958 of T will be in P.T_Test_Data.T_Tests.
19960 Consider running gnattest on the second example (note: actual tests for this
19961 example already exist, so there's no need to worry if the tool reports that
19962 no new stubs were generated):
19965 cd <install_prefix>/share/examples/gnattest/tagged_rec
19966 gnattest --harness-dir=driver -Ptagged_rec.gpr
19969 Taking a closer look at the test type declared in the test package
19970 Speed1.Controller_Test_Data is necessary. It is declared in:
19973 <install_prefix>/share/examples/gnattest/tagged_rec/obj/gnattest/tests
19976 Test types are direct or indirect descendants of
19977 AUnit.Test_Fixtures.Test_Fixture type. In the case of nonprimitive tested
19978 subprograms, the user doesn't need to be concerned with them. However,
19979 when generating test packages for primitive operations, there are some things
19980 the user needs to know.
19982 Type Test_Controller has components that allow assignment of various
19983 derivations of type Controller. And if you look at the specification of
19984 package Speed2.Auto_Controller, you will see that Test_Auto_Controller
19985 actually derives from Test_Controller rather than AUnit type Test_Fixture.
19986 Thus, test types mirror the hierarchy of tested types.
19988 The Set_Up procedure of Test_Data package corresponding to a test package
19989 of primitive operations of type T assigns to Fixture a reference to an
19990 object of that exact type T. Notice, however, that if the tagged type has
19991 discriminants, the Set_Up only has a commented template for setting
19992 up the fixture, since filling the discriminant with actual value is up
19995 The knowledge of the structure of test types allows additional testing
19996 without additional effort. Those possibilities are described below.
19998 @node Testing Inheritance
19999 @section Testing Inheritance
20003 Since the test type hierarchy mimics the hierarchy of tested types, the
20004 inheritance of tests takes place. An example of such inheritance can be
20005 seen by running the test driver generated for the second example. As previously
20006 mentioned, actual tests are already written for this example.
20010 gnatmake -Ptest_driver
20014 There are 6 passed tests while there are only 5 testable subprograms. The test
20015 routine for function Speed has been inherited and run against objects of the
20018 @node Tagged Types Substitutability Testing
20019 @section Tagged Types Substitutability Testing
20023 Tagged Types Substitutability Testing is a way of verifying the global type
20024 consistency by testing. Global type consistency is a principle stating that if
20025 S is a subtype of T (in Ada, S is a derived type of tagged type T),
20026 then objects of type T may be replaced with objects of type S (that is,
20027 objects of type S may be substituted for objects of type T), without
20028 altering any of the desirable properties of the program. When the properties
20029 of the program are expressed in the form of subprogram preconditions and
20030 postconditions (let's call them pre and post), the principle is formulated as
20031 relations between the pre and post of primitive operations and the pre and post
20032 of their derived operations. The pre of a derived operation should not be
20033 stronger than the original pre, and the post of the derived operation should
20034 not be weaker than the original post. Those relations ensure that verifying if
20035 a dispatching call is safe can be done just by using the pre and post of the
20038 Verifying global type consistency by testing consists of running all the unit
20039 tests associated with the primitives of a given tagged type with objects of its
20042 In the example used in the previous section, there was clearly a violation of
20043 type consistency. The overriding primitive Adjust_Speed in package Speed2
20044 removes the functionality of the overridden primitive and thus doesn't respect
20045 the consistency principle.
20046 Gnattest has a special option to run overridden parent tests against objects
20047 of the type which have overriding primitives:
20050 gnattest --harness-dir=driver --validate-type-extensions -Ptagged_rec.gpr
20052 gnatmake -Ptest_driver
20056 While all the tests pass by themselves, the parent test for Adjust_Speed fails
20057 against objects of the derived type.
20059 Non-overridden tests are already inherited for derived test types, so the
20060 --validate-type-extensions enables the application of overriden tests to objects
20063 @node Testing with Contracts
20064 @section Testing with Contracts
20068 @command{gnattest} supports pragmas Precondition, Postcondition, and Test_Case,
20069 as well as corresponding aspects.
20070 Test routines are generated, one per each Test_Case associated with a tested
20071 subprogram. Those test routines have special wrappers for tested functions
20072 that have composition of pre- and postcondition of the subprogram with
20073 "requires" and "ensures" of the Test_Case (depending on the mode, pre and post
20074 either count for Nominal mode or do not count for Robustness mode).
20076 The third example demonstrates how this works:
20079 cd <install_prefix>/share/examples/gnattest/contracts
20080 gnattest --harness-dir=driver -Pcontracts.gpr
20083 Putting actual checks within the range of the contract does not cause any
20084 error reports. For example, for the test routine which corresponds to
20087 @smallexample @c ada
20088 Assert (Sqrt (9.0) = 3.0, "wrong sqrt");
20091 and for the test routine corresponding to test case 2:
20093 @smallexample @c ada
20094 Assert (Sqrt (-5.0) = -1.0, "wrong error indication");
20101 gnatmake -Ptest_driver
20105 However, by changing 9.0 to 25.0 and 3.0 to 5.0, for example, you can get
20106 a precondition violation for test case one. Also, by using any otherwise
20107 correct but positive pair of numbers in the second test routine, you can also
20108 get a precondition violation. Postconditions are checked and reported
20111 @node Additional Tests
20112 @section Additional Tests
20115 @command{gnattest} can add user-written tests to the main suite of the test
20116 driver. @command{gnattest} traverses the given packages and searches for test
20117 routines. All procedures with a single in out parameter of a type which is
20118 derived from AUnit.Test_Fixtures.Test_Fixture and that are declared in package
20119 specifications are added to the suites and are then executed by the test driver.
20120 (Set_Up and Tear_Down are filtered out.)
20122 An example illustrates two ways of creating test harnesses for user-written
20123 tests. Directory additional_tests contains an AUnit-based test driver written
20127 <install_prefix>/share/examples/gnattest/additional_tests/
20130 To create a test driver for already-written tests, use the --harness-only
20134 gnattest -Padditional/harness/harness.gpr --harness-dir=harness_only \
20136 gnatmake -Pharness_only/test_driver.gpr
20137 harness_only/test_runner
20140 Additional tests can also be executed together with generated tests:
20143 gnattest -Psimple.gpr --additional-tests=additional/harness/harness.gpr \
20144 --harness-dir=mixing
20145 gnatmake -Pmixing/test_driver.gpr
20150 @node Support for other platforms/run-times
20151 @section Support for other platforms/run-times
20154 @command{gnattest} can be used to generate the test harness for platforms
20155 and run-time libraries others than the default native target with the
20156 default full run-time. For example, when using a limited run-time library
20157 such as Zero FootPrint (ZFP), a simplified harness is generated.
20159 Two variables are used to tell the underlying AUnit framework how to generate
20160 the test harness: @code{PLATFORM}, which identifies the target, and
20161 @code{RUNTIME}, used to determine the run-time library for which the harness
20162 is generated. Corresponding prefix should also be used when calling
20163 @command{gnattest} for non-native targets. For example, the following options
20164 are used to generate the AUnit test harness for a PowerPC ELF target using
20165 the ZFP run-time library:
20168 powerpc-elf-gnattest -Psimple.gpr -XPLATFORM=powerpc-elf -XRUNTIME=zfp
20172 @node Current Limitations
20173 @section Current Limitations
20177 The tool currently does not support following features:
20180 @item generic tests for generic packages and package instantiations
20181 @item tests for protected subprograms and entries
20187 @c *********************************
20188 @node Performing Dimensionality Analysis in GNAT
20189 @chapter Performing Dimensionality Analysis in GNAT
20190 @cindex Dimensionality analysis
20193 The GNAT compiler now supports dimensionality checking. The user can
20194 specify physical units for objects, and the compiler will verify that uses
20195 of these objects are compatible with their dimensions, in a fashion that is
20196 familiar to engineering practice. The dimensions of algebraic expressions
20197 (including powers with static exponents) are computed from their constituents.
20199 This feature depends on Ada 2012 aspect specifications, and is available from
20200 version 7.0.1 of GNAT onwards.
20201 The GNAT-specific aspect @code{Dimension_System}
20202 @cindex @code{Dimension_System} aspect
20203 allows you to define a system of units; the aspect @code{Dimension}
20204 @cindex @code{Dimension} aspect
20205 then allows the user to declare dimensioned quantities within a given system.
20206 (These aspects are described in the @i{Implementation Defined Aspects}
20207 chapter of the @i{GNAT Reference Manual}).
20209 The major advantage of this model is that it does not require the declaration of
20210 multiple operators for all possible combinations of types: it is only necessary
20211 to use the proper subtypes in object declarations.
20213 The simplest way to impose dimensionality checking on a computation is to make
20214 use of the package @code{System.Dim.Mks},
20215 @cindex @code{System.Dim.Mks} package (GNAT library)
20216 which is part of the GNAT library. This
20217 package defines a floating-point type @code{MKS_Type},
20218 @cindex @code{MKS_Type} type
20219 for which a sequence of
20220 dimension names are specified, together with their conventional abbreviations.
20221 The following should be read together with the full specification of the
20222 package, in file @file{s-dimmks.ads}.
20223 @cindex @file{s-dimmks.ads} file
20225 @smallexample @c ada
20227 type Mks_Type is new Long_Long_Float
20229 Dimension_System => (
20230 (Unit_Name => Meter, Unit_Symbol => 'm', Dim_Symbol => 'L'),
20231 (Unit_Name => Kilogram, Unit_Symbol => "kg", Dim_Symbol => 'M'),
20232 (Unit_Name => Second, Unit_Symbol => 's', Dim_Symbol => 'T'),
20233 (Unit_Name => Ampere, Unit_Symbol => 'A', Dim_Symbol => 'I'),
20234 (Unit_Name => Kelvin, Unit_Symbol => 'K', Dim_Symbol => "Theta"),
20235 (Unit_Name => Mole, Unit_Symbol => "mol", Dim_Symbol => 'N'),
20236 (Unit_Name => Candela, Unit_Symbol => "cd", Dim_Symbol => 'J'));
20241 The package then defines a series of subtypes that correspond to these
20242 conventional units. For example:
20244 @smallexample @c ada
20246 subtype Length is Mks_Type
20248 Dimension => (Symbol => 'm', Meter => 1, others => 0);
20253 and similarly for @code{Mass}, @code{Time}, @code{Electric_Current},
20254 @code{Thermodynamic_Temperature}, @code{Amount_Of_Substance}, and
20255 @code{Luminous_Intensity} (the standard set of units of the SI system).
20257 The package also defines conventional names for values of each unit, for
20260 @smallexample @c ada
20262 m : constant Length := 1.0;
20263 kg : constant Mass := 1.0;
20264 s : constant Time := 1.0;
20265 A : constant Electric_Current := 1.0;
20270 as well as useful multiples of these units:
20272 @smallexample @c ada
20274 cm : constant Length := 1.0E-02;
20275 g : constant Mass := 1.0E-03;
20276 min : constant Time := 60.0;
20277 day : constant Time := 60.0 * 24.0 * min;
20283 Using this package, you can then define a derived unit by
20284 providing the aspect that
20285 specifies its dimensions within the MKS system, as well as the string to
20286 be used for output of a value of that unit:
20288 @smallexample @c ada
20290 subtype Acceleration is Mks_Type
20291 with Dimension => ("m/sec^^^2",
20299 Here is a complete example of use:
20301 @smallexample @c ada
20303 with System.Dim.MKS; use System.Dim.Mks;
20304 with System.Dim.Mks_IO; use System.Dim.Mks_IO;
20305 with Text_IO; use Text_IO;
20306 procedure Free_Fall is
20307 subtype Acceleration is Mks_Type
20308 with Dimension => ("m/sec^^^2", 1, 0, -2, others => 0);
20309 G : constant acceleration := 9.81 * m / (s ** 2);
20310 T : Time := 10.0*s;
20315 Put ("Gravitational constant: ");
20316 Put (G, Aft => 2, Exp => 0); Put_Line ("");
20317 Distance := 0.5 * G * T ** 2;
20318 Put ("distance travelled in 10 seconds of free fall ");
20319 Put (Distance, Aft => 2, Exp => 0);
20326 Execution of this program yields:
20329 Gravitational constant: 9.81 m/sec^^^2
20330 distance travelled in 10 seconds of free fall 490.50 m
20335 However, incorrect assignments such as:
20337 @smallexample @c ada
20340 Distance := 5.0 * kg:
20345 are rejected with the following diagnoses:
20350 >>> dimensions mismatch in assignment
20351 >>> left-hand side has dimension [L]
20352 >>> right-hand side is dimensionless
20356 Distance := 5.0 * kg:
20357 >>> dimensions mismatch in assignment
20358 >>> left-hand side has dimension [L]
20359 >>> right-hand side has dimension [M]
20364 The dimensions of an expression are properly displayed, even if there is
20365 no explicit subtype for it. If we add to the program:
20367 @smallexample @c ada
20369 Put ("Final velocity: ");
20370 Put (G * T, Aft =>2, Exp =>0);
20376 then the output includes:
20378 Final velocity: 98.10 m.s**(-1)
20382 @c *********************************
20383 @node Generating Ada Bindings for C and C++ headers
20384 @chapter Generating Ada Bindings for C and C++ headers
20388 GNAT now comes with a binding generator for C and C++ headers which is
20389 intended to do 95% of the tedious work of generating Ada specs from C
20390 or C++ header files.
20392 Note that this capability is not intended to generate 100% correct Ada specs,
20393 and will is some cases require manual adjustments, although it can often
20394 be used out of the box in practice.
20396 Some of the known limitations include:
20399 @item only very simple character constant macros are translated into Ada
20400 constants. Function macros (macros with arguments) are partially translated
20401 as comments, to be completed manually if needed.
20402 @item some extensions (e.g. vector types) are not supported
20403 @item pointers to pointers or complex structures are mapped to System.Address
20404 @item identifiers with identical name (except casing) will generate compilation
20405 errors (e.g. @code{shm_get} vs @code{SHM_GET}).
20408 The code generated is using the Ada 2005 syntax, which makes it
20409 easier to interface with other languages than previous versions of Ada.
20412 * Running the binding generator::
20413 * Generating bindings for C++ headers::
20417 @node Running the binding generator
20418 @section Running the binding generator
20421 The binding generator is part of the @command{gcc} compiler and can be
20422 invoked via the @option{-fdump-ada-spec} switch, which will generate Ada
20423 spec files for the header files specified on the command line, and all
20424 header files needed by these files transitively. For example:
20427 $ g++ -c -fdump-ada-spec -C /usr/include/time.h
20428 $ gcc -c -gnat05 *.ads
20431 will generate, under GNU/Linux, the following files: @file{time_h.ads},
20432 @file{bits_time_h.ads}, @file{stddef_h.ads}, @file{bits_types_h.ads} which
20433 correspond to the files @file{/usr/include/time.h},
20434 @file{/usr/include/bits/time.h}, etc@dots{}, and will then compile in Ada 2005
20435 mode these Ada specs.
20437 The @code{-C} switch tells @command{gcc} to extract comments from headers,
20438 and will attempt to generate corresponding Ada comments.
20440 If you want to generate a single Ada file and not the transitive closure, you
20441 can use instead the @option{-fdump-ada-spec-slim} switch.
20443 You can optionally specify a parent unit, of which all generated units will
20444 be children, using @code{-fada-spec-parent=}@var{unit}.
20446 Note that we recommend when possible to use the @command{g++} driver to
20447 generate bindings, even for most C headers, since this will in general
20448 generate better Ada specs. For generating bindings for C++ headers, it is
20449 mandatory to use the @command{g++} command, or @command{gcc -x c++} which
20450 is equivalent in this case. If @command{g++} cannot work on your C headers
20451 because of incompatibilities between C and C++, then you can fallback to
20452 @command{gcc} instead.
20454 For an example of better bindings generated from the C++ front-end,
20455 the name of the parameters (when available) are actually ignored by the C
20456 front-end. Consider the following C header:
20459 extern void foo (int variable);
20462 with the C front-end, @code{variable} is ignored, and the above is handled as:
20465 extern void foo (int);
20468 generating a generic:
20471 procedure foo (param1 : int);
20474 with the C++ front-end, the name is available, and we generate:
20477 procedure foo (variable : int);
20480 In some cases, the generated bindings will be more complete or more meaningful
20481 when defining some macros, which you can do via the @option{-D} switch. This
20482 is for example the case with @file{Xlib.h} under GNU/Linux:
20485 g++ -c -fdump-ada-spec -DXLIB_ILLEGAL_ACCESS -C /usr/include/X11/Xlib.h
20488 The above will generate more complete bindings than a straight call without
20489 the @option{-DXLIB_ILLEGAL_ACCESS} switch.
20491 In other cases, it is not possible to parse a header file in a stand-alone
20492 manner, because other include files need to be included first. In this
20493 case, the solution is to create a small header file including the needed
20494 @code{#include} and possible @code{#define} directives. For example, to
20495 generate Ada bindings for @file{readline/readline.h}, you need to first
20496 include @file{stdio.h}, so you can create a file with the following two
20497 lines in e.g. @file{readline1.h}:
20501 #include <readline/readline.h>
20504 and then generate Ada bindings from this file:
20507 $ g++ -c -fdump-ada-spec readline1.h
20510 @node Generating bindings for C++ headers
20511 @section Generating bindings for C++ headers
20514 Generating bindings for C++ headers is done using the same options, always
20515 with the @command{g++} compiler.
20517 In this mode, C++ classes will be mapped to Ada tagged types, constructors
20518 will be mapped using the @code{CPP_Constructor} pragma, and when possible,
20519 multiple inheritance of abstract classes will be mapped to Ada interfaces
20520 (@xref{Interfacing to C++,,,gnat_rm, GNAT Reference Manual}, for additional
20521 information on interfacing to C++).
20523 For example, given the following C++ header file:
20530 virtual int Number_Of_Teeth () = 0;
20535 virtual void Set_Owner (char* Name) = 0;
20541 virtual void Set_Age (int New_Age);
20544 class Dog : Animal, Carnivore, Domestic @{
20549 virtual int Number_Of_Teeth ();
20550 virtual void Set_Owner (char* Name);
20558 The corresponding Ada code is generated:
20560 @smallexample @c ada
20563 package Class_Carnivore is
20564 type Carnivore is limited interface;
20565 pragma Import (CPP, Carnivore);
20567 function Number_Of_Teeth (this : access Carnivore) return int is abstract;
20569 use Class_Carnivore;
20571 package Class_Domestic is
20572 type Domestic is limited interface;
20573 pragma Import (CPP, Domestic);
20575 procedure Set_Owner
20576 (this : access Domestic;
20577 Name : Interfaces.C.Strings.chars_ptr) is abstract;
20579 use Class_Domestic;
20581 package Class_Animal is
20582 type Animal is tagged limited record
20583 Age_Count : aliased int;
20585 pragma Import (CPP, Animal);
20587 procedure Set_Age (this : access Animal; New_Age : int);
20588 pragma Import (CPP, Set_Age, "_ZN6Animal7Set_AgeEi");
20592 package Class_Dog is
20593 type Dog is new Animal and Carnivore and Domestic with record
20594 Tooth_Count : aliased int;
20595 Owner : Interfaces.C.Strings.chars_ptr;
20597 pragma Import (CPP, Dog);
20599 function Number_Of_Teeth (this : access Dog) return int;
20600 pragma Import (CPP, Number_Of_Teeth, "_ZN3Dog15Number_Of_TeethEv");
20602 procedure Set_Owner
20603 (this : access Dog; Name : Interfaces.C.Strings.chars_ptr);
20604 pragma Import (CPP, Set_Owner, "_ZN3Dog9Set_OwnerEPc");
20606 function New_Dog return Dog;
20607 pragma CPP_Constructor (New_Dog);
20608 pragma Import (CPP, New_Dog, "_ZN3DogC1Ev");
20619 @item -fdump-ada-spec
20620 @cindex @option{-fdump-ada-spec} (@command{gcc})
20621 Generate Ada spec files for the given header files transitively (including
20622 all header files that these headers depend upon).
20624 @item -fdump-ada-spec-slim
20625 @cindex @option{-fdump-ada-spec-slim} (@command{gcc})
20626 Generate Ada spec files for the header files specified on the command line
20629 @item -fada-spec-parent=@var{unit}
20630 @cindex -fada-spec-parent (@command{gcc})
20631 Specifies that all files generated by @option{-fdump-ada-spec*} are
20632 to be child units of the specified parent unit.
20635 @cindex @option{-C} (@command{gcc})
20636 Extract comments from headers and generate Ada comments in the Ada spec files.
20639 @node Other Utility Programs
20640 @chapter Other Utility Programs
20643 This chapter discusses some other utility programs available in the Ada
20647 * Using Other Utility Programs with GNAT::
20648 * The External Symbol Naming Scheme of GNAT::
20649 * Converting Ada Files to html with gnathtml::
20650 * Installing gnathtml::
20657 @node Using Other Utility Programs with GNAT
20658 @section Using Other Utility Programs with GNAT
20661 The object files generated by GNAT are in standard system format and in
20662 particular the debugging information uses this format. This means
20663 programs generated by GNAT can be used with existing utilities that
20664 depend on these formats.
20667 In general, any utility program that works with C will also often work with
20668 Ada programs generated by GNAT. This includes software utilities such as
20669 gprof (a profiling program), @code{gdb} (the FSF debugger), and utilities such
20673 @node The External Symbol Naming Scheme of GNAT
20674 @section The External Symbol Naming Scheme of GNAT
20677 In order to interpret the output from GNAT, when using tools that are
20678 originally intended for use with other languages, it is useful to
20679 understand the conventions used to generate link names from the Ada
20682 All link names are in all lowercase letters. With the exception of library
20683 procedure names, the mechanism used is simply to use the full expanded
20684 Ada name with dots replaced by double underscores. For example, suppose
20685 we have the following package spec:
20687 @smallexample @c ada
20698 The variable @code{MN} has a full expanded Ada name of @code{QRS.MN}, so
20699 the corresponding link name is @code{qrs__mn}.
20701 Of course if a @code{pragma Export} is used this may be overridden:
20703 @smallexample @c ada
20708 pragma Export (Var1, C, External_Name => "var1_name");
20710 pragma Export (Var2, C, Link_Name => "var2_link_name");
20717 In this case, the link name for @var{Var1} is whatever link name the
20718 C compiler would assign for the C function @var{var1_name}. This typically
20719 would be either @var{var1_name} or @var{_var1_name}, depending on operating
20720 system conventions, but other possibilities exist. The link name for
20721 @var{Var2} is @var{var2_link_name}, and this is not operating system
20725 One exception occurs for library level procedures. A potential ambiguity
20726 arises between the required name @code{_main} for the C main program,
20727 and the name we would otherwise assign to an Ada library level procedure
20728 called @code{Main} (which might well not be the main program).
20730 To avoid this ambiguity, we attach the prefix @code{_ada_} to such
20731 names. So if we have a library level procedure such as
20733 @smallexample @c ada
20736 procedure Hello (S : String);
20742 the external name of this procedure will be @var{_ada_hello}.
20745 @node Converting Ada Files to html with gnathtml
20746 @section Converting Ada Files to HTML with @code{gnathtml}
20749 This @code{Perl} script allows Ada source files to be browsed using
20750 standard Web browsers. For installation procedure, see the section
20751 @xref{Installing gnathtml}.
20753 Ada reserved keywords are highlighted in a bold font and Ada comments in
20754 a blue font. Unless your program was compiled with the gcc @option{-gnatx}
20755 switch to suppress the generation of cross-referencing information, user
20756 defined variables and types will appear in a different color; you will
20757 be able to click on any identifier and go to its declaration.
20759 The command line is as follow:
20761 @c $ perl gnathtml.pl @ovar{^switches^options^} @var{ada-files}
20762 @c Expanding @ovar macro inline (explanation in macro def comments)
20763 $ perl gnathtml.pl @r{[}@var{^switches^options^}@r{]} @var{ada-files}
20767 You can pass it as many Ada files as you want. @code{gnathtml} will generate
20768 an html file for every ada file, and a global file called @file{index.htm}.
20769 This file is an index of every identifier defined in the files.
20771 The available ^switches^options^ are the following ones:
20775 @cindex @option{-83} (@code{gnathtml})
20776 Only the Ada 83 subset of keywords will be highlighted.
20778 @item -cc @var{color}
20779 @cindex @option{-cc} (@code{gnathtml})
20780 This option allows you to change the color used for comments. The default
20781 value is green. The color argument can be any name accepted by html.
20784 @cindex @option{-d} (@code{gnathtml})
20785 If the Ada files depend on some other files (for instance through
20786 @code{with} clauses, the latter files will also be converted to html.
20787 Only the files in the user project will be converted to html, not the files
20788 in the run-time library itself.
20791 @cindex @option{-D} (@code{gnathtml})
20792 This command is the same as @option{-d} above, but @command{gnathtml} will
20793 also look for files in the run-time library, and generate html files for them.
20795 @item -ext @var{extension}
20796 @cindex @option{-ext} (@code{gnathtml})
20797 This option allows you to change the extension of the generated HTML files.
20798 If you do not specify an extension, it will default to @file{htm}.
20801 @cindex @option{-f} (@code{gnathtml})
20802 By default, gnathtml will generate html links only for global entities
20803 ('with'ed units, global variables and types,@dots{}). If you specify
20804 @option{-f} on the command line, then links will be generated for local
20807 @item -l @var{number}
20808 @cindex @option{-l} (@code{gnathtml})
20809 If this ^switch^option^ is provided and @var{number} is not 0, then
20810 @code{gnathtml} will number the html files every @var{number} line.
20813 @cindex @option{-I} (@code{gnathtml})
20814 Specify a directory to search for library files (@file{.ALI} files) and
20815 source files. You can provide several -I switches on the command line,
20816 and the directories will be parsed in the order of the command line.
20819 @cindex @option{-o} (@code{gnathtml})
20820 Specify the output directory for html files. By default, gnathtml will
20821 saved the generated html files in a subdirectory named @file{html/}.
20823 @item -p @var{file}
20824 @cindex @option{-p} (@code{gnathtml})
20825 If you are using Emacs and the most recent Emacs Ada mode, which provides
20826 a full Integrated Development Environment for compiling, checking,
20827 running and debugging applications, you may use @file{.gpr} files
20828 to give the directories where Emacs can find sources and object files.
20830 Using this ^switch^option^, you can tell gnathtml to use these files.
20831 This allows you to get an html version of your application, even if it
20832 is spread over multiple directories.
20834 @item -sc @var{color}
20835 @cindex @option{-sc} (@code{gnathtml})
20836 This ^switch^option^ allows you to change the color used for symbol
20838 The default value is red. The color argument can be any name accepted by html.
20840 @item -t @var{file}
20841 @cindex @option{-t} (@code{gnathtml})
20842 This ^switch^option^ provides the name of a file. This file contains a list of
20843 file names to be converted, and the effect is exactly as though they had
20844 appeared explicitly on the command line. This
20845 is the recommended way to work around the command line length limit on some
20850 @node Installing gnathtml
20851 @section Installing @code{gnathtml}
20854 @code{Perl} needs to be installed on your machine to run this script.
20855 @code{Perl} is freely available for almost every architecture and
20856 Operating System via the Internet.
20858 On Unix systems, you may want to modify the first line of the script
20859 @code{gnathtml}, to explicitly tell the Operating system where Perl
20860 is. The syntax of this line is:
20862 #!full_path_name_to_perl
20866 Alternatively, you may run the script using the following command line:
20869 @c $ perl gnathtml.pl @ovar{switches} @var{files}
20870 @c Expanding @ovar macro inline (explanation in macro def comments)
20871 $ perl gnathtml.pl @r{[}@var{switches}@r{]} @var{files}
20880 The GNAT distribution provides an Ada 95 template for the HP Language
20881 Sensitive Editor (LSE), a component of DECset. In order to
20882 access it, invoke LSE with the qualifier /ENVIRONMENT=GNU:[LIB]ADA95.ENV.
20889 GNAT supports The HP Performance Coverage Analyzer (PCA), a component
20890 of DECset. To use it proceed as outlined under ``HELP PCA'', except for running
20891 the collection phase with the /DEBUG qualifier.
20894 $ GNAT MAKE /DEBUG <PROGRAM_NAME>
20895 $ DEFINE LIB$DEBUG PCA$COLLECTOR
20896 $ RUN/DEBUG <PROGRAM_NAME>
20902 @c ******************************
20903 @node Code Coverage and Profiling
20904 @chapter Code Coverage and Profiling
20905 @cindex Code Coverage
20909 This chapter describes how to use @code{gcov} - coverage testing tool - and
20910 @code{gprof} - profiler tool - on your Ada programs.
20913 * Code Coverage of Ada Programs with gcov::
20914 * Profiling an Ada Program with gprof::
20917 @node Code Coverage of Ada Programs with gcov
20918 @section Code Coverage of Ada Programs with gcov
20920 @cindex -fprofile-arcs
20921 @cindex -ftest-coverage
20923 @cindex Code Coverage
20926 @code{gcov} is a test coverage program: it analyzes the execution of a given
20927 program on selected tests, to help you determine the portions of the program
20928 that are still untested.
20930 @code{gcov} is part of the GCC suite, and is described in detail in the GCC
20931 User's Guide. You can refer to this documentation for a more complete
20934 This chapter provides a quick startup guide, and
20935 details some Gnat-specific features.
20938 * Quick startup guide::
20942 @node Quick startup guide
20943 @subsection Quick startup guide
20945 In order to perform coverage analysis of a program using @code{gcov}, 3
20950 Code instrumentation during the compilation process
20952 Execution of the instrumented program
20954 Execution of the @code{gcov} tool to generate the result.
20957 The code instrumentation needed by gcov is created at the object level:
20958 The source code is not modified in any way, because the instrumentation code is
20959 inserted by gcc during the compilation process. To compile your code with code
20960 coverage activated, you need to recompile your whole project using the
20962 @code{-fprofile-arcs} and @code{-ftest-coverage}, and link it using
20963 @code{-fprofile-arcs}.
20966 $ gnatmake -P my_project.gpr -f -cargs -fprofile-arcs -ftest-coverage \
20967 -largs -fprofile-arcs
20970 This compilation process will create @file{.gcno} files together with
20971 the usual object files.
20973 Once the program is compiled with coverage instrumentation, you can
20974 run it as many times as needed - on portions of a test suite for
20975 example. The first execution will produce @file{.gcda} files at the
20976 same location as the @file{.gcno} files. The following executions
20977 will update those files, so that a cumulative result of the covered
20978 portions of the program is generated.
20980 Finally, you need to call the @code{gcov} tool. The different options of
20981 @code{gcov} are available in the GCC User's Guide, section 'Invoking gcov'.
20983 This will create annotated source files with a @file{.gcov} extension:
20984 @file{my_main.adb} file will be analysed in @file{my_main.adb.gcov}.
20986 @node Gnat specifics
20987 @subsection Gnat specifics
20989 Because Ada semantics, portions of the source code may be shared among
20990 several object files. This is the case for example when generics are
20991 involved, when inlining is active or when declarations generate initialisation
20992 calls. In order to take
20993 into account this shared code, you need to call @code{gcov} on all
20994 source files of the tested program at once.
20996 The list of source files might exceed the system's maximum command line
20997 length. In order to bypass this limitation, a new mechanism has been
20998 implemented in @code{gcov}: you can now list all your project's files into a
20999 text file, and provide this file to gcov as a parameter, preceded by a @@
21000 (e.g. @samp{gcov @@mysrclist.txt}).
21002 Note that on AIX compiling a static library with @code{-fprofile-arcs} is
21003 not supported as there can be unresolved symbols during the final link.
21005 @node Profiling an Ada Program with gprof
21006 @section Profiling an Ada Program with gprof
21012 This section is not meant to be an exhaustive documentation of @code{gprof}.
21013 Full documentation for it can be found in the GNU Profiler User's Guide
21014 documentation that is part of this GNAT distribution.
21016 Profiling a program helps determine the parts of a program that are executed
21017 most often, and are therefore the most time-consuming.
21019 @code{gprof} is the standard GNU profiling tool; it has been enhanced to
21020 better handle Ada programs and multitasking.
21021 It is currently supported on the following platforms
21026 solaris sparc/sparc64/x86
21032 In order to profile a program using @code{gprof}, 3 steps are needed:
21036 Code instrumentation, requiring a full recompilation of the project with the
21039 Execution of the program under the analysis conditions, i.e. with the desired
21042 Analysis of the results using the @code{gprof} tool.
21046 The following sections detail the different steps, and indicate how
21047 to interpret the results:
21049 * Compilation for profiling::
21050 * Program execution::
21052 * Interpretation of profiling results::
21055 @node Compilation for profiling
21056 @subsection Compilation for profiling
21060 In order to profile a program the first step is to tell the compiler
21061 to generate the necessary profiling information. The compiler switch to be used
21062 is @code{-pg}, which must be added to other compilation switches. This
21063 switch needs to be specified both during compilation and link stages, and can
21064 be specified once when using gnatmake:
21067 gnatmake -f -pg -P my_project
21071 Note that only the objects that were compiled with the @samp{-pg} switch will
21072 be profiled; if you need to profile your whole project, use the @samp{-f}
21073 gnatmake switch to force full recompilation.
21075 @node Program execution
21076 @subsection Program execution
21079 Once the program has been compiled for profiling, you can run it as usual.
21081 The only constraint imposed by profiling is that the program must terminate
21082 normally. An interrupted program (via a Ctrl-C, kill, etc.) will not be
21085 Once the program completes execution, a data file called @file{gmon.out} is
21086 generated in the directory where the program was launched from. If this file
21087 already exists, it will be overwritten.
21089 @node Running gprof
21090 @subsection Running gprof
21093 The @code{gprof} tool is called as follow:
21096 gprof my_prog gmon.out
21107 The complete form of the gprof command line is the following:
21110 gprof [^switches^options^] [executable [data-file]]
21114 @code{gprof} supports numerous ^switch^options^. The order of these
21115 ^switch^options^ does not matter. The full list of options can be found in
21116 the GNU Profiler User's Guide documentation that comes with this documentation.
21118 The following is the subset of those switches that is most relevant:
21122 @item --demangle[=@var{style}]
21123 @itemx --no-demangle
21124 @cindex @option{--demangle} (@code{gprof})
21125 These options control whether symbol names should be demangled when
21126 printing output. The default is to demangle C++ symbols. The
21127 @code{--no-demangle} option may be used to turn off demangling. Different
21128 compilers have different mangling styles. The optional demangling style
21129 argument can be used to choose an appropriate demangling style for your
21130 compiler, in particular Ada symbols generated by GNAT can be demangled using
21131 @code{--demangle=gnat}.
21133 @item -e @var{function_name}
21134 @cindex @option{-e} (@code{gprof})
21135 The @samp{-e @var{function}} option tells @code{gprof} not to print
21136 information about the function @var{function_name} (and its
21137 children@dots{}) in the call graph. The function will still be listed
21138 as a child of any functions that call it, but its index number will be
21139 shown as @samp{[not printed]}. More than one @samp{-e} option may be
21140 given; only one @var{function_name} may be indicated with each @samp{-e}
21143 @item -E @var{function_name}
21144 @cindex @option{-E} (@code{gprof})
21145 The @code{-E @var{function}} option works like the @code{-e} option, but
21146 execution time spent in the function (and children who were not called from
21147 anywhere else), will not be used to compute the percentages-of-time for
21148 the call graph. More than one @samp{-E} option may be given; only one
21149 @var{function_name} may be indicated with each @samp{-E} option.
21151 @item -f @var{function_name}
21152 @cindex @option{-f} (@code{gprof})
21153 The @samp{-f @var{function}} option causes @code{gprof} to limit the
21154 call graph to the function @var{function_name} and its children (and
21155 their children@dots{}). More than one @samp{-f} option may be given;
21156 only one @var{function_name} may be indicated with each @samp{-f}
21159 @item -F @var{function_name}
21160 @cindex @option{-F} (@code{gprof})
21161 The @samp{-F @var{function}} option works like the @code{-f} option, but
21162 only time spent in the function and its children (and their
21163 children@dots{}) will be used to determine total-time and
21164 percentages-of-time for the call graph. More than one @samp{-F} option
21165 may be given; only one @var{function_name} may be indicated with each
21166 @samp{-F} option. The @samp{-F} option overrides the @samp{-E} option.
21170 @node Interpretation of profiling results
21171 @subsection Interpretation of profiling results
21175 The results of the profiling analysis are represented by two arrays: the
21176 'flat profile' and the 'call graph'. Full documentation of those outputs
21177 can be found in the GNU Profiler User's Guide.
21179 The flat profile shows the time spent in each function of the program, and how
21180 many time it has been called. This allows you to locate easily the most
21181 time-consuming functions.
21183 The call graph shows, for each subprogram, the subprograms that call it,
21184 and the subprograms that it calls. It also provides an estimate of the time
21185 spent in each of those callers/called subprograms.
21188 @c ******************************
21189 @node Running and Debugging Ada Programs
21190 @chapter Running and Debugging Ada Programs
21194 This chapter discusses how to debug Ada programs.
21196 It applies to GNAT on the Alpha OpenVMS platform;
21197 for I64 OpenVMS please refer to the @cite{OpenVMS Debugger Manual},
21198 since HP has implemented Ada support in the OpenVMS debugger on I64.
21201 An incorrect Ada program may be handled in three ways by the GNAT compiler:
21205 The illegality may be a violation of the static semantics of Ada. In
21206 that case GNAT diagnoses the constructs in the program that are illegal.
21207 It is then a straightforward matter for the user to modify those parts of
21211 The illegality may be a violation of the dynamic semantics of Ada. In
21212 that case the program compiles and executes, but may generate incorrect
21213 results, or may terminate abnormally with some exception.
21216 When presented with a program that contains convoluted errors, GNAT
21217 itself may terminate abnormally without providing full diagnostics on
21218 the incorrect user program.
21222 * The GNAT Debugger GDB::
21224 * Introduction to GDB Commands::
21225 * Using Ada Expressions::
21226 * Calling User-Defined Subprograms::
21227 * Using the Next Command in a Function::
21230 * Debugging Generic Units::
21231 * Remote Debugging with gdbserver::
21232 * GNAT Abnormal Termination or Failure to Terminate::
21233 * Naming Conventions for GNAT Source Files::
21234 * Getting Internal Debugging Information::
21235 * Stack Traceback::
21241 @node The GNAT Debugger GDB
21242 @section The GNAT Debugger GDB
21245 @code{GDB} is a general purpose, platform-independent debugger that
21246 can be used to debug mixed-language programs compiled with @command{gcc},
21247 and in particular is capable of debugging Ada programs compiled with
21248 GNAT. The latest versions of @code{GDB} are Ada-aware and can handle
21249 complex Ada data structures.
21251 @xref{Top,, Debugging with GDB, gdb, Debugging with GDB},
21253 located in the GNU:[DOCS] directory,
21255 for full details on the usage of @code{GDB}, including a section on
21256 its usage on programs. This manual should be consulted for full
21257 details. The section that follows is a brief introduction to the
21258 philosophy and use of @code{GDB}.
21260 When GNAT programs are compiled, the compiler optionally writes debugging
21261 information into the generated object file, including information on
21262 line numbers, and on declared types and variables. This information is
21263 separate from the generated code. It makes the object files considerably
21264 larger, but it does not add to the size of the actual executable that
21265 will be loaded into memory, and has no impact on run-time performance. The
21266 generation of debug information is triggered by the use of the
21267 ^-g^/DEBUG^ switch in the @command{gcc} or @command{gnatmake} command
21268 used to carry out the compilations. It is important to emphasize that
21269 the use of these options does not change the generated code.
21271 The debugging information is written in standard system formats that
21272 are used by many tools, including debuggers and profilers. The format
21273 of the information is typically designed to describe C types and
21274 semantics, but GNAT implements a translation scheme which allows full
21275 details about Ada types and variables to be encoded into these
21276 standard C formats. Details of this encoding scheme may be found in
21277 the file exp_dbug.ads in the GNAT source distribution. However, the
21278 details of this encoding are, in general, of no interest to a user,
21279 since @code{GDB} automatically performs the necessary decoding.
21281 When a program is bound and linked, the debugging information is
21282 collected from the object files, and stored in the executable image of
21283 the program. Again, this process significantly increases the size of
21284 the generated executable file, but it does not increase the size of
21285 the executable program itself. Furthermore, if this program is run in
21286 the normal manner, it runs exactly as if the debug information were
21287 not present, and takes no more actual memory.
21289 However, if the program is run under control of @code{GDB}, the
21290 debugger is activated. The image of the program is loaded, at which
21291 point it is ready to run. If a run command is given, then the program
21292 will run exactly as it would have if @code{GDB} were not present. This
21293 is a crucial part of the @code{GDB} design philosophy. @code{GDB} is
21294 entirely non-intrusive until a breakpoint is encountered. If no
21295 breakpoint is ever hit, the program will run exactly as it would if no
21296 debugger were present. When a breakpoint is hit, @code{GDB} accesses
21297 the debugging information and can respond to user commands to inspect
21298 variables, and more generally to report on the state of execution.
21302 @section Running GDB
21305 This section describes how to initiate the debugger.
21306 @c The above sentence is really just filler, but it was otherwise
21307 @c clumsy to get the first paragraph nonindented given the conditional
21308 @c nature of the description
21311 The debugger can be launched from a @code{GPS} menu or
21312 directly from the command line. The description below covers the latter use.
21313 All the commands shown can be used in the @code{GPS} debug console window,
21314 but there are usually more GUI-based ways to achieve the same effect.
21317 The command to run @code{GDB} is
21320 $ ^gdb program^GDB PROGRAM^
21324 where @code{^program^PROGRAM^} is the name of the executable file. This
21325 activates the debugger and results in a prompt for debugger commands.
21326 The simplest command is simply @code{run}, which causes the program to run
21327 exactly as if the debugger were not present. The following section
21328 describes some of the additional commands that can be given to @code{GDB}.
21330 @c *******************************
21331 @node Introduction to GDB Commands
21332 @section Introduction to GDB Commands
21335 @code{GDB} contains a large repertoire of commands. @xref{Top,,
21336 Debugging with GDB, gdb, Debugging with GDB},
21338 located in the GNU:[DOCS] directory,
21340 for extensive documentation on the use
21341 of these commands, together with examples of their use. Furthermore,
21342 the command @command{help} invoked from within GDB activates a simple help
21343 facility which summarizes the available commands and their options.
21344 In this section we summarize a few of the most commonly
21345 used commands to give an idea of what @code{GDB} is about. You should create
21346 a simple program with debugging information and experiment with the use of
21347 these @code{GDB} commands on the program as you read through the
21351 @item set args @var{arguments}
21352 The @var{arguments} list above is a list of arguments to be passed to
21353 the program on a subsequent run command, just as though the arguments
21354 had been entered on a normal invocation of the program. The @code{set args}
21355 command is not needed if the program does not require arguments.
21358 The @code{run} command causes execution of the program to start from
21359 the beginning. If the program is already running, that is to say if
21360 you are currently positioned at a breakpoint, then a prompt will ask
21361 for confirmation that you want to abandon the current execution and
21364 @item breakpoint @var{location}
21365 The breakpoint command sets a breakpoint, that is to say a point at which
21366 execution will halt and @code{GDB} will await further
21367 commands. @var{location} is
21368 either a line number within a file, given in the format @code{file:linenumber},
21369 or it is the name of a subprogram. If you request that a breakpoint be set on
21370 a subprogram that is overloaded, a prompt will ask you to specify on which of
21371 those subprograms you want to breakpoint. You can also
21372 specify that all of them should be breakpointed. If the program is run
21373 and execution encounters the breakpoint, then the program
21374 stops and @code{GDB} signals that the breakpoint was encountered by
21375 printing the line of code before which the program is halted.
21377 @item catch exception @var{name}
21378 This command causes the program execution to stop whenever exception
21379 @var{name} is raised. If @var{name} is omitted, then the execution is
21380 suspended when any exception is raised.
21382 @item print @var{expression}
21383 This will print the value of the given expression. Most simple
21384 Ada expression formats are properly handled by @code{GDB}, so the expression
21385 can contain function calls, variables, operators, and attribute references.
21388 Continues execution following a breakpoint, until the next breakpoint or the
21389 termination of the program.
21392 Executes a single line after a breakpoint. If the next statement
21393 is a subprogram call, execution continues into (the first statement of)
21394 the called subprogram.
21397 Executes a single line. If this line is a subprogram call, executes and
21398 returns from the call.
21401 Lists a few lines around the current source location. In practice, it
21402 is usually more convenient to have a separate edit window open with the
21403 relevant source file displayed. Successive applications of this command
21404 print subsequent lines. The command can be given an argument which is a
21405 line number, in which case it displays a few lines around the specified one.
21408 Displays a backtrace of the call chain. This command is typically
21409 used after a breakpoint has occurred, to examine the sequence of calls that
21410 leads to the current breakpoint. The display includes one line for each
21411 activation record (frame) corresponding to an active subprogram.
21414 At a breakpoint, @code{GDB} can display the values of variables local
21415 to the current frame. The command @code{up} can be used to
21416 examine the contents of other active frames, by moving the focus up
21417 the stack, that is to say from callee to caller, one frame at a time.
21420 Moves the focus of @code{GDB} down from the frame currently being
21421 examined to the frame of its callee (the reverse of the previous command),
21423 @item frame @var{n}
21424 Inspect the frame with the given number. The value 0 denotes the frame
21425 of the current breakpoint, that is to say the top of the call stack.
21428 Kills the child process in which the program is running under GDB.
21429 This may be useful for several purposes:
21432 It allows you to recompile and relink your program, since on many systems
21433 you cannot regenerate an executable file while it is running in a process.
21435 You can run your program outside the debugger, on systems that do not
21436 permit executing a program outside GDB while breakpoints are set
21439 It allows you to debug a core dump rather than a running process.
21444 The above list is a very short introduction to the commands that
21445 @code{GDB} provides. Important additional capabilities, including conditional
21446 breakpoints, the ability to execute command sequences on a breakpoint,
21447 the ability to debug at the machine instruction level and many other
21448 features are described in detail in @ref{Top,, Debugging with GDB, gdb,
21449 Debugging with GDB}. Note that most commands can be abbreviated
21450 (for example, c for continue, bt for backtrace).
21452 @node Using Ada Expressions
21453 @section Using Ada Expressions
21454 @cindex Ada expressions
21457 @code{GDB} supports a fairly large subset of Ada expression syntax, with some
21458 extensions. The philosophy behind the design of this subset is
21462 That @code{GDB} should provide basic literals and access to operations for
21463 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
21464 leaving more sophisticated computations to subprograms written into the
21465 program (which therefore may be called from @code{GDB}).
21468 That type safety and strict adherence to Ada language restrictions
21469 are not particularly important to the @code{GDB} user.
21472 That brevity is important to the @code{GDB} user.
21476 Thus, for brevity, the debugger acts as if there were
21477 implicit @code{with} and @code{use} clauses in effect for all user-written
21478 packages, thus making it unnecessary to fully qualify most names with
21479 their packages, regardless of context. Where this causes ambiguity,
21480 @code{GDB} asks the user's intent.
21482 For details on the supported Ada syntax, see @ref{Top,, Debugging with
21483 GDB, gdb, Debugging with GDB}.
21485 @node Calling User-Defined Subprograms
21486 @section Calling User-Defined Subprograms
21489 An important capability of @code{GDB} is the ability to call user-defined
21490 subprograms while debugging. This is achieved simply by entering
21491 a subprogram call statement in the form:
21494 call subprogram-name (parameters)
21498 The keyword @code{call} can be omitted in the normal case where the
21499 @code{subprogram-name} does not coincide with any of the predefined
21500 @code{GDB} commands.
21502 The effect is to invoke the given subprogram, passing it the
21503 list of parameters that is supplied. The parameters can be expressions and
21504 can include variables from the program being debugged. The
21505 subprogram must be defined
21506 at the library level within your program, and @code{GDB} will call the
21507 subprogram within the environment of your program execution (which
21508 means that the subprogram is free to access or even modify variables
21509 within your program).
21511 The most important use of this facility is in allowing the inclusion of
21512 debugging routines that are tailored to particular data structures
21513 in your program. Such debugging routines can be written to provide a suitably
21514 high-level description of an abstract type, rather than a low-level dump
21515 of its physical layout. After all, the standard
21516 @code{GDB print} command only knows the physical layout of your
21517 types, not their abstract meaning. Debugging routines can provide information
21518 at the desired semantic level and are thus enormously useful.
21520 For example, when debugging GNAT itself, it is crucial to have access to
21521 the contents of the tree nodes used to represent the program internally.
21522 But tree nodes are represented simply by an integer value (which in turn
21523 is an index into a table of nodes).
21524 Using the @code{print} command on a tree node would simply print this integer
21525 value, which is not very useful. But the PN routine (defined in file
21526 treepr.adb in the GNAT sources) takes a tree node as input, and displays
21527 a useful high level representation of the tree node, which includes the
21528 syntactic category of the node, its position in the source, the integers
21529 that denote descendant nodes and parent node, as well as varied
21530 semantic information. To study this example in more detail, you might want to
21531 look at the body of the PN procedure in the stated file.
21533 @node Using the Next Command in a Function
21534 @section Using the Next Command in a Function
21537 When you use the @code{next} command in a function, the current source
21538 location will advance to the next statement as usual. A special case
21539 arises in the case of a @code{return} statement.
21541 Part of the code for a return statement is the ``epilog'' of the function.
21542 This is the code that returns to the caller. There is only one copy of
21543 this epilog code, and it is typically associated with the last return
21544 statement in the function if there is more than one return. In some
21545 implementations, this epilog is associated with the first statement
21548 The result is that if you use the @code{next} command from a return
21549 statement that is not the last return statement of the function you
21550 may see a strange apparent jump to the last return statement or to
21551 the start of the function. You should simply ignore this odd jump.
21552 The value returned is always that from the first return statement
21553 that was stepped through.
21555 @node Ada Exceptions
21556 @section Stopping when Ada Exceptions are Raised
21560 You can set catchpoints that stop the program execution when your program
21561 raises selected exceptions.
21564 @item catch exception
21565 Set a catchpoint that stops execution whenever (any task in the) program
21566 raises any exception.
21568 @item catch exception @var{name}
21569 Set a catchpoint that stops execution whenever (any task in the) program
21570 raises the exception @var{name}.
21572 @item catch exception unhandled
21573 Set a catchpoint that stops executing whenever (any task in the) program
21574 raises an exception for which there is no handler.
21576 @item info exceptions
21577 @itemx info exceptions @var{regexp}
21578 The @code{info exceptions} command permits the user to examine all defined
21579 exceptions within Ada programs. With a regular expression, @var{regexp}, as
21580 argument, prints out only those exceptions whose name matches @var{regexp}.
21588 @code{GDB} allows the following task-related commands:
21592 This command shows a list of current Ada tasks, as in the following example:
21599 ID TID P-ID Thread Pri State Name
21600 1 8088000 0 807e000 15 Child Activation Wait main_task
21601 2 80a4000 1 80ae000 15 Accept/Select Wait b
21602 3 809a800 1 80a4800 15 Child Activation Wait a
21603 * 4 80ae800 3 80b8000 15 Running c
21607 In this listing, the asterisk before the first task indicates it to be the
21608 currently running task. The first column lists the task ID that is used
21609 to refer to tasks in the following commands.
21611 @item break @var{linespec} task @var{taskid}
21612 @itemx break @var{linespec} task @var{taskid} if @dots{}
21613 @cindex Breakpoints and tasks
21614 These commands are like the @code{break @dots{} thread @dots{}}.
21615 @var{linespec} specifies source lines.
21617 Use the qualifier @samp{task @var{taskid}} with a breakpoint command
21618 to specify that you only want @code{GDB} to stop the program when a
21619 particular Ada task reaches this breakpoint. @var{taskid} is one of the
21620 numeric task identifiers assigned by @code{GDB}, shown in the first
21621 column of the @samp{info tasks} display.
21623 If you do not specify @samp{task @var{taskid}} when you set a
21624 breakpoint, the breakpoint applies to @emph{all} tasks of your
21627 You can use the @code{task} qualifier on conditional breakpoints as
21628 well; in this case, place @samp{task @var{taskid}} before the
21629 breakpoint condition (before the @code{if}).
21631 @item task @var{taskno}
21632 @cindex Task switching
21634 This command allows to switch to the task referred by @var{taskno}. In
21635 particular, This allows to browse the backtrace of the specified
21636 task. It is advised to switch back to the original task before
21637 continuing execution otherwise the scheduling of the program may be
21642 For more detailed information on the tasking support,
21643 see @ref{Top,, Debugging with GDB, gdb, Debugging with GDB}.
21645 @node Debugging Generic Units
21646 @section Debugging Generic Units
21647 @cindex Debugging Generic Units
21651 GNAT always uses code expansion for generic instantiation. This means that
21652 each time an instantiation occurs, a complete copy of the original code is
21653 made, with appropriate substitutions of formals by actuals.
21655 It is not possible to refer to the original generic entities in
21656 @code{GDB}, but it is always possible to debug a particular instance of
21657 a generic, by using the appropriate expanded names. For example, if we have
21659 @smallexample @c ada
21664 generic package k is
21665 procedure kp (v1 : in out integer);
21669 procedure kp (v1 : in out integer) is
21675 package k1 is new k;
21676 package k2 is new k;
21678 var : integer := 1;
21691 Then to break on a call to procedure kp in the k2 instance, simply
21695 (gdb) break g.k2.kp
21699 When the breakpoint occurs, you can step through the code of the
21700 instance in the normal manner and examine the values of local variables, as for
21703 @node Remote Debugging with gdbserver
21704 @section Remote Debugging with gdbserver
21705 @cindex Remote Debugging with gdbserver
21708 On platforms where gdbserver is supported, it is possible to use this tool
21709 to debug your application remotely. This can be useful in situations
21710 where the program needs to be run on a target host that is different
21711 from the host used for development, particularly when the target has
21712 a limited amount of resources (either CPU and/or memory).
21714 To do so, start your program using gdbserver on the target machine.
21715 gdbserver then automatically suspends the execution of your program
21716 at its entry point, waiting for a debugger to connect to it. The
21717 following commands starts an application and tells gdbserver to
21718 wait for a connection with the debugger on localhost port 4444.
21721 $ gdbserver localhost:4444 program
21722 Process program created; pid = 5685
21723 Listening on port 4444
21726 Once gdbserver has started listening, we can tell the debugger to establish
21727 a connection with this gdbserver, and then start the same debugging session
21728 as if the program was being debugged on the same host, directly under
21729 the control of GDB.
21733 (gdb) target remote targethost:4444
21734 Remote debugging using targethost:4444
21735 0x00007f29936d0af0 in ?? () from /lib64/ld-linux-x86-64.so.
21737 Breakpoint 1 at 0x401f0c: file foo.adb, line 3.
21741 Breakpoint 1, foo () at foo.adb:4
21745 It is also possible to use gdbserver to attach to an already running
21746 program, in which case the execution of that program is simply suspended
21747 until the connection between the debugger and gdbserver is established.
21749 For more information on how to use gdbserver, @ref{Top, Server, Using
21750 the gdbserver Program, gdb, Debugging with GDB}. @value{EDITION} provides support
21751 for gdbserver on x86-linux, x86-windows and x86_64-linux.
21753 @node GNAT Abnormal Termination or Failure to Terminate
21754 @section GNAT Abnormal Termination or Failure to Terminate
21755 @cindex GNAT Abnormal Termination or Failure to Terminate
21758 When presented with programs that contain serious errors in syntax
21760 GNAT may on rare occasions experience problems in operation, such
21762 segmentation fault or illegal memory access, raising an internal
21763 exception, terminating abnormally, or failing to terminate at all.
21764 In such cases, you can activate
21765 various features of GNAT that can help you pinpoint the construct in your
21766 program that is the likely source of the problem.
21768 The following strategies are presented in increasing order of
21769 difficulty, corresponding to your experience in using GNAT and your
21770 familiarity with compiler internals.
21774 Run @command{gcc} with the @option{-gnatf}. This first
21775 switch causes all errors on a given line to be reported. In its absence,
21776 only the first error on a line is displayed.
21778 The @option{-gnatdO} switch causes errors to be displayed as soon as they
21779 are encountered, rather than after compilation is terminated. If GNAT
21780 terminates prematurely or goes into an infinite loop, the last error
21781 message displayed may help to pinpoint the culprit.
21784 Run @command{gcc} with the @option{^-v (verbose)^/VERBOSE^} switch. In this
21785 mode, @command{gcc} produces ongoing information about the progress of the
21786 compilation and provides the name of each procedure as code is
21787 generated. This switch allows you to find which Ada procedure was being
21788 compiled when it encountered a code generation problem.
21791 @cindex @option{-gnatdc} switch
21792 Run @command{gcc} with the @option{-gnatdc} switch. This is a GNAT specific
21793 switch that does for the front-end what @option{^-v^VERBOSE^} does
21794 for the back end. The system prints the name of each unit,
21795 either a compilation unit or nested unit, as it is being analyzed.
21797 Finally, you can start
21798 @code{gdb} directly on the @code{gnat1} executable. @code{gnat1} is the
21799 front-end of GNAT, and can be run independently (normally it is just
21800 called from @command{gcc}). You can use @code{gdb} on @code{gnat1} as you
21801 would on a C program (but @pxref{The GNAT Debugger GDB} for caveats). The
21802 @code{where} command is the first line of attack; the variable
21803 @code{lineno} (seen by @code{print lineno}), used by the second phase of
21804 @code{gnat1} and by the @command{gcc} backend, indicates the source line at
21805 which the execution stopped, and @code{input_file name} indicates the name of
21809 @node Naming Conventions for GNAT Source Files
21810 @section Naming Conventions for GNAT Source Files
21813 In order to examine the workings of the GNAT system, the following
21814 brief description of its organization may be helpful:
21818 Files with prefix @file{^sc^SC^} contain the lexical scanner.
21821 All files prefixed with @file{^par^PAR^} are components of the parser. The
21822 numbers correspond to chapters of the Ada Reference Manual. For example,
21823 parsing of select statements can be found in @file{par-ch9.adb}.
21826 All files prefixed with @file{^sem^SEM^} perform semantic analysis. The
21827 numbers correspond to chapters of the Ada standard. For example, all
21828 issues involving context clauses can be found in @file{sem_ch10.adb}. In
21829 addition, some features of the language require sufficient special processing
21830 to justify their own semantic files: sem_aggr for aggregates, sem_disp for
21831 dynamic dispatching, etc.
21834 All files prefixed with @file{^exp^EXP^} perform normalization and
21835 expansion of the intermediate representation (abstract syntax tree, or AST).
21836 these files use the same numbering scheme as the parser and semantics files.
21837 For example, the construction of record initialization procedures is done in
21838 @file{exp_ch3.adb}.
21841 The files prefixed with @file{^bind^BIND^} implement the binder, which
21842 verifies the consistency of the compilation, determines an order of
21843 elaboration, and generates the bind file.
21846 The files @file{atree.ads} and @file{atree.adb} detail the low-level
21847 data structures used by the front-end.
21850 The files @file{sinfo.ads} and @file{sinfo.adb} detail the structure of
21851 the abstract syntax tree as produced by the parser.
21854 The files @file{einfo.ads} and @file{einfo.adb} detail the attributes of
21855 all entities, computed during semantic analysis.
21858 Library management issues are dealt with in files with prefix
21864 Ada files with the prefix @file{^a-^A-^} are children of @code{Ada}, as
21865 defined in Annex A.
21870 Files with prefix @file{^i-^I-^} are children of @code{Interfaces}, as
21871 defined in Annex B.
21875 Files with prefix @file{^s-^S-^} are children of @code{System}. This includes
21876 both language-defined children and GNAT run-time routines.
21880 Files with prefix @file{^g-^G-^} are children of @code{GNAT}. These are useful
21881 general-purpose packages, fully documented in their specs. All
21882 the other @file{.c} files are modifications of common @command{gcc} files.
21885 @node Getting Internal Debugging Information
21886 @section Getting Internal Debugging Information
21889 Most compilers have internal debugging switches and modes. GNAT
21890 does also, except GNAT internal debugging switches and modes are not
21891 secret. A summary and full description of all the compiler and binder
21892 debug flags are in the file @file{debug.adb}. You must obtain the
21893 sources of the compiler to see the full detailed effects of these flags.
21895 The switches that print the source of the program (reconstructed from
21896 the internal tree) are of general interest for user programs, as are the
21898 the full internal tree, and the entity table (the symbol table
21899 information). The reconstructed source provides a readable version of the
21900 program after the front-end has completed analysis and expansion,
21901 and is useful when studying the performance of specific constructs.
21902 For example, constraint checks are indicated, complex aggregates
21903 are replaced with loops and assignments, and tasking primitives
21904 are replaced with run-time calls.
21906 @node Stack Traceback
21907 @section Stack Traceback
21909 @cindex stack traceback
21910 @cindex stack unwinding
21913 Traceback is a mechanism to display the sequence of subprogram calls that
21914 leads to a specified execution point in a program. Often (but not always)
21915 the execution point is an instruction at which an exception has been raised.
21916 This mechanism is also known as @i{stack unwinding} because it obtains
21917 its information by scanning the run-time stack and recovering the activation
21918 records of all active subprograms. Stack unwinding is one of the most
21919 important tools for program debugging.
21921 The first entry stored in traceback corresponds to the deepest calling level,
21922 that is to say the subprogram currently executing the instruction
21923 from which we want to obtain the traceback.
21925 Note that there is no runtime performance penalty when stack traceback
21926 is enabled, and no exception is raised during program execution.
21929 * Non-Symbolic Traceback::
21930 * Symbolic Traceback::
21933 @node Non-Symbolic Traceback
21934 @subsection Non-Symbolic Traceback
21935 @cindex traceback, non-symbolic
21938 Note: this feature is not supported on all platforms. See
21939 @file{GNAT.Traceback spec in g-traceb.ads} for a complete list of supported
21943 * Tracebacks From an Unhandled Exception::
21944 * Tracebacks From Exception Occurrences (non-symbolic)::
21945 * Tracebacks From Anywhere in a Program (non-symbolic)::
21948 @node Tracebacks From an Unhandled Exception
21949 @subsubsection Tracebacks From an Unhandled Exception
21952 A runtime non-symbolic traceback is a list of addresses of call instructions.
21953 To enable this feature you must use the @option{-E}
21954 @code{gnatbind}'s option. With this option a stack traceback is stored as part
21955 of exception information. You can retrieve this information using the
21956 @code{addr2line} tool.
21958 Here is a simple example:
21960 @smallexample @c ada
21966 raise Constraint_Error;
21981 $ gnatmake stb -bargs -E
21984 Execution terminated by unhandled exception
21985 Exception name: CONSTRAINT_ERROR
21987 Call stack traceback locations:
21988 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
21992 As we see the traceback lists a sequence of addresses for the unhandled
21993 exception @code{CONSTRAINT_ERROR} raised in procedure P1. It is easy to
21994 guess that this exception come from procedure P1. To translate these
21995 addresses into the source lines where the calls appear, the
21996 @code{addr2line} tool, described below, is invaluable. The use of this tool
21997 requires the program to be compiled with debug information.
22000 $ gnatmake -g stb -bargs -E
22003 Execution terminated by unhandled exception
22004 Exception name: CONSTRAINT_ERROR
22006 Call stack traceback locations:
22007 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
22009 $ addr2line --exe=stb 0x401373 0x40138b 0x40139c 0x401335 0x4011c4
22010 0x4011f1 0x77e892a4
22012 00401373 at d:/stb/stb.adb:5
22013 0040138B at d:/stb/stb.adb:10
22014 0040139C at d:/stb/stb.adb:14
22015 00401335 at d:/stb/b~stb.adb:104
22016 004011C4 at /build/@dots{}/crt1.c:200
22017 004011F1 at /build/@dots{}/crt1.c:222
22018 77E892A4 in ?? at ??:0
22022 The @code{addr2line} tool has several other useful options:
22026 to get the function name corresponding to any location
22028 @item --demangle=gnat
22029 to use the gnat decoding mode for the function names. Note that
22030 for binutils version 2.9.x the option is simply @option{--demangle}.
22034 $ addr2line --exe=stb --functions --demangle=gnat 0x401373 0x40138b
22035 0x40139c 0x401335 0x4011c4 0x4011f1
22037 00401373 in stb.p1 at d:/stb/stb.adb:5
22038 0040138B in stb.p2 at d:/stb/stb.adb:10
22039 0040139C in stb at d:/stb/stb.adb:14
22040 00401335 in main at d:/stb/b~stb.adb:104
22041 004011C4 in <__mingw_CRTStartup> at /build/@dots{}/crt1.c:200
22042 004011F1 in <mainCRTStartup> at /build/@dots{}/crt1.c:222
22046 From this traceback we can see that the exception was raised in
22047 @file{stb.adb} at line 5, which was reached from a procedure call in
22048 @file{stb.adb} at line 10, and so on. The @file{b~std.adb} is the binder file,
22049 which contains the call to the main program.
22050 @xref{Running gnatbind}. The remaining entries are assorted runtime routines,
22051 and the output will vary from platform to platform.
22053 It is also possible to use @code{GDB} with these traceback addresses to debug
22054 the program. For example, we can break at a given code location, as reported
22055 in the stack traceback:
22061 Furthermore, this feature is not implemented inside Windows DLL. Only
22062 the non-symbolic traceback is reported in this case.
22065 (gdb) break *0x401373
22066 Breakpoint 1 at 0x401373: file stb.adb, line 5.
22070 It is important to note that the stack traceback addresses
22071 do not change when debug information is included. This is particularly useful
22072 because it makes it possible to release software without debug information (to
22073 minimize object size), get a field report that includes a stack traceback
22074 whenever an internal bug occurs, and then be able to retrieve the sequence
22075 of calls with the same program compiled with debug information.
22077 @node Tracebacks From Exception Occurrences (non-symbolic)
22078 @subsubsection Tracebacks From Exception Occurrences
22081 Non-symbolic tracebacks are obtained by using the @option{-E} binder argument.
22082 The stack traceback is attached to the exception information string, and can
22083 be retrieved in an exception handler within the Ada program, by means of the
22084 Ada facilities defined in @code{Ada.Exceptions}. Here is a simple example:
22086 @smallexample @c ada
22088 with Ada.Exceptions;
22093 use Ada.Exceptions;
22101 Text_IO.Put_Line (Exception_Information (E));
22115 This program will output:
22120 Exception name: CONSTRAINT_ERROR
22121 Message: stb.adb:12
22122 Call stack traceback locations:
22123 0x4015e4 0x401633 0x401644 0x401461 0x4011c4 0x4011f1 0x77e892a4
22126 @node Tracebacks From Anywhere in a Program (non-symbolic)
22127 @subsubsection Tracebacks From Anywhere in a Program
22130 It is also possible to retrieve a stack traceback from anywhere in a
22131 program. For this you need to
22132 use the @code{GNAT.Traceback} API. This package includes a procedure called
22133 @code{Call_Chain} that computes a complete stack traceback, as well as useful
22134 display procedures described below. It is not necessary to use the
22135 @option{-E gnatbind} option in this case, because the stack traceback mechanism
22136 is invoked explicitly.
22139 In the following example we compute a traceback at a specific location in
22140 the program, and we display it using @code{GNAT.Debug_Utilities.Image} to
22141 convert addresses to strings:
22143 @smallexample @c ada
22145 with GNAT.Traceback;
22146 with GNAT.Debug_Utilities;
22152 use GNAT.Traceback;
22155 TB : Tracebacks_Array (1 .. 10);
22156 -- We are asking for a maximum of 10 stack frames.
22158 -- Len will receive the actual number of stack frames returned.
22160 Call_Chain (TB, Len);
22162 Text_IO.Put ("In STB.P1 : ");
22164 for K in 1 .. Len loop
22165 Text_IO.Put (Debug_Utilities.Image (TB (K)));
22186 In STB.P1 : 16#0040_F1E4# 16#0040_14F2# 16#0040_170B# 16#0040_171C#
22187 16#0040_1461# 16#0040_11C4# 16#0040_11F1# 16#77E8_92A4#
22191 You can then get further information by invoking the @code{addr2line}
22192 tool as described earlier (note that the hexadecimal addresses
22193 need to be specified in C format, with a leading ``0x'').
22195 @node Symbolic Traceback
22196 @subsection Symbolic Traceback
22197 @cindex traceback, symbolic
22200 A symbolic traceback is a stack traceback in which procedure names are
22201 associated with each code location.
22204 Note that this feature is not supported on all platforms. See
22205 @file{GNAT.Traceback.Symbolic spec in g-trasym.ads} for a complete
22206 list of currently supported platforms.
22209 Note that the symbolic traceback requires that the program be compiled
22210 with debug information. If it is not compiled with debug information
22211 only the non-symbolic information will be valid.
22214 * Tracebacks From Exception Occurrences (symbolic)::
22215 * Tracebacks From Anywhere in a Program (symbolic)::
22218 @node Tracebacks From Exception Occurrences (symbolic)
22219 @subsubsection Tracebacks From Exception Occurrences
22221 @smallexample @c ada
22223 with GNAT.Traceback.Symbolic;
22229 raise Constraint_Error;
22246 Ada.Text_IO.Put_Line (GNAT.Traceback.Symbolic.Symbolic_Traceback (E));
22251 $ gnatmake -g .\stb -bargs -E
22254 0040149F in stb.p1 at stb.adb:8
22255 004014B7 in stb.p2 at stb.adb:13
22256 004014CF in stb.p3 at stb.adb:18
22257 004015DD in ada.stb at stb.adb:22
22258 00401461 in main at b~stb.adb:168
22259 004011C4 in __mingw_CRTStartup at crt1.c:200
22260 004011F1 in mainCRTStartup at crt1.c:222
22261 77E892A4 in ?? at ??:0
22265 In the above example the ``.\'' syntax in the @command{gnatmake} command
22266 is currently required by @command{addr2line} for files that are in
22267 the current working directory.
22268 Moreover, the exact sequence of linker options may vary from platform
22270 The above @option{-largs} section is for Windows platforms. By contrast,
22271 under Unix there is no need for the @option{-largs} section.
22272 Differences across platforms are due to details of linker implementation.
22274 @node Tracebacks From Anywhere in a Program (symbolic)
22275 @subsubsection Tracebacks From Anywhere in a Program
22278 It is possible to get a symbolic stack traceback
22279 from anywhere in a program, just as for non-symbolic tracebacks.
22280 The first step is to obtain a non-symbolic
22281 traceback, and then call @code{Symbolic_Traceback} to compute the symbolic
22282 information. Here is an example:
22284 @smallexample @c ada
22286 with GNAT.Traceback;
22287 with GNAT.Traceback.Symbolic;
22292 use GNAT.Traceback;
22293 use GNAT.Traceback.Symbolic;
22296 TB : Tracebacks_Array (1 .. 10);
22297 -- We are asking for a maximum of 10 stack frames.
22299 -- Len will receive the actual number of stack frames returned.
22301 Call_Chain (TB, Len);
22302 Text_IO.Put_Line (Symbolic_Traceback (TB (1 .. Len)));
22315 @c ******************************
22317 @node Compatibility with HP Ada
22318 @chapter Compatibility with HP Ada
22319 @cindex Compatibility
22324 @cindex Compatibility between GNAT and HP Ada
22325 This chapter compares HP Ada (formerly known as ``DEC Ada'')
22326 for OpenVMS Alpha and GNAT for OpenVMS for Alpha and for I64.
22327 GNAT is highly compatible
22328 with HP Ada, and it should generally be straightforward to port code
22329 from the HP Ada environment to GNAT. However, there are a few language
22330 and implementation differences of which the user must be aware. These
22331 differences are discussed in this chapter. In
22332 addition, the operating environment and command structure for the
22333 compiler are different, and these differences are also discussed.
22335 For further details on these and other compatibility issues,
22336 see Appendix E of the HP publication
22337 @cite{HP Ada, Technical Overview and Comparison on HP Platforms}.
22339 Except where otherwise indicated, the description of GNAT for OpenVMS
22340 applies to both the Alpha and I64 platforms.
22342 For information on porting Ada code from GNAT on Alpha OpenVMS to GNAT on
22343 I64 OpenVMS, see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
22345 The discussion in this chapter addresses specifically the implementation
22346 of Ada 83 for HP OpenVMS Alpha Systems. In cases where the implementation
22347 of HP Ada differs between OpenVMS Alpha Systems and OpenVMS VAX Systems,
22348 GNAT always follows the Alpha implementation.
22350 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
22351 attributes are recognized, although only a subset of them can sensibly
22352 be implemented. The description of pragmas in
22353 @xref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
22354 indicates whether or not they are applicable to non-VMS systems.
22357 * Ada Language Compatibility::
22358 * Differences in the Definition of Package System::
22359 * Language-Related Features::
22360 * The Package STANDARD::
22361 * The Package SYSTEM::
22362 * Tasking and Task-Related Features::
22363 * Pragmas and Pragma-Related Features::
22364 * Library of Predefined Units::
22366 * Main Program Definition::
22367 * Implementation-Defined Attributes::
22368 * Compiler and Run-Time Interfacing::
22369 * Program Compilation and Library Management::
22371 * Implementation Limits::
22372 * Tools and Utilities::
22375 @node Ada Language Compatibility
22376 @section Ada Language Compatibility
22379 GNAT handles Ada 95 and Ada 2005 as well as Ada 83, whereas HP Ada is only
22380 for Ada 83. Ada 95 and Ada 2005 are almost completely upwards compatible
22381 with Ada 83, and therefore Ada 83 programs will compile
22382 and run under GNAT with
22383 no changes or only minor changes. The @cite{Annotated Ada Reference Manual}
22384 provides details on specific incompatibilities.
22386 GNAT provides the switch @option{/83} on the @command{GNAT COMPILE} command,
22387 as well as the pragma @code{ADA_83}, to force the compiler to
22388 operate in Ada 83 mode. This mode does not guarantee complete
22389 conformance to Ada 83, but in practice is sufficient to
22390 eliminate most sources of incompatibilities.
22391 In particular, it eliminates the recognition of the
22392 additional Ada 95 and Ada 2005 keywords, so that their use as identifiers
22393 in Ada 83 programs is legal, and handles the cases of packages
22394 with optional bodies, and generics that instantiate unconstrained
22395 types without the use of @code{(<>)}.
22397 @node Differences in the Definition of Package System
22398 @section Differences in the Definition of Package @code{System}
22401 An Ada compiler is allowed to add
22402 implementation-dependent declarations to package @code{System}.
22404 GNAT does not take advantage of this permission, and the version of
22405 @code{System} provided by GNAT exactly matches that defined in the Ada
22408 However, HP Ada adds an extensive set of declarations to package
22410 as fully documented in the HP Ada manuals. To minimize changes required
22411 for programs that make use of these extensions, GNAT provides the pragma
22412 @code{Extend_System} for extending the definition of package System. By using:
22413 @cindex pragma @code{Extend_System}
22414 @cindex @code{Extend_System} pragma
22416 @smallexample @c ada
22419 pragma Extend_System (Aux_DEC);
22425 the set of definitions in @code{System} is extended to include those in
22426 package @code{System.Aux_DEC}.
22427 @cindex @code{System.Aux_DEC} package
22428 @cindex @code{Aux_DEC} package (child of @code{System})
22429 These definitions are incorporated directly into package @code{System},
22430 as though they had been declared there. For a
22431 list of the declarations added, see the spec of this package,
22432 which can be found in the file @file{s-auxdec.ads} in the GNAT library.
22433 @cindex @file{s-auxdec.ads} file
22434 The pragma @code{Extend_System} is a configuration pragma, which means that
22435 it can be placed in the file @file{gnat.adc}, so that it will automatically
22436 apply to all subsequent compilations. See @ref{Configuration Pragmas},
22437 for further details.
22439 An alternative approach that avoids the use of the non-standard
22440 @code{Extend_System} pragma is to add a context clause to the unit that
22441 references these facilities:
22443 @smallexample @c ada
22445 with System.Aux_DEC;
22446 use System.Aux_DEC;
22451 The effect is not quite semantically identical to incorporating
22452 the declarations directly into package @code{System},
22453 but most programs will not notice a difference
22454 unless they use prefix notation (e.g.@: @code{System.Integer_8})
22455 to reference the entities directly in package @code{System}.
22456 For units containing such references,
22457 the prefixes must either be removed, or the pragma @code{Extend_System}
22460 @node Language-Related Features
22461 @section Language-Related Features
22464 The following sections highlight differences in types,
22465 representations of types, operations, alignment, and
22469 * Integer Types and Representations::
22470 * Floating-Point Types and Representations::
22471 * Pragmas Float_Representation and Long_Float::
22472 * Fixed-Point Types and Representations::
22473 * Record and Array Component Alignment::
22474 * Address Clauses::
22475 * Other Representation Clauses::
22478 @node Integer Types and Representations
22479 @subsection Integer Types and Representations
22482 The set of predefined integer types is identical in HP Ada and GNAT.
22483 Furthermore the representation of these integer types is also identical,
22484 including the capability of size clauses forcing biased representation.
22487 HP Ada for OpenVMS Alpha systems has defined the
22488 following additional integer types in package @code{System}:
22505 @code{LARGEST_INTEGER}
22509 In GNAT, the first four of these types may be obtained from the
22510 standard Ada package @code{Interfaces}.
22511 Alternatively, by use of the pragma @code{Extend_System}, identical
22512 declarations can be referenced directly in package @code{System}.
22513 On both GNAT and HP Ada, the maximum integer size is 64 bits.
22515 @node Floating-Point Types and Representations
22516 @subsection Floating-Point Types and Representations
22517 @cindex Floating-Point types
22520 The set of predefined floating-point types is identical in HP Ada and GNAT.
22521 Furthermore the representation of these floating-point
22522 types is also identical. One important difference is that the default
22523 representation for HP Ada is @code{VAX_Float}, but the default representation
22526 Specific types may be declared to be @code{VAX_Float} or IEEE, using the
22527 pragma @code{Float_Representation} as described in the HP Ada
22529 For example, the declarations:
22531 @smallexample @c ada
22533 type F_Float is digits 6;
22534 pragma Float_Representation (VAX_Float, F_Float);
22539 declares a type @code{F_Float} that will be represented in @code{VAX_Float}
22541 This set of declarations actually appears in @code{System.Aux_DEC},
22543 the full set of additional floating-point declarations provided in
22544 the HP Ada version of package @code{System}.
22545 This and similar declarations may be accessed in a user program
22546 by using pragma @code{Extend_System}. The use of this
22547 pragma, and the related pragma @code{Long_Float} is described in further
22548 detail in the following section.
22550 @node Pragmas Float_Representation and Long_Float
22551 @subsection Pragmas @code{Float_Representation} and @code{Long_Float}
22554 HP Ada provides the pragma @code{Float_Representation}, which
22555 acts as a program library switch to allow control over
22556 the internal representation chosen for the predefined
22557 floating-point types declared in the package @code{Standard}.
22558 The format of this pragma is as follows:
22560 @smallexample @c ada
22562 pragma Float_Representation(VAX_Float | IEEE_Float);
22567 This pragma controls the representation of floating-point
22572 @code{VAX_Float} specifies that floating-point
22573 types are represented by default with the VAX system hardware types
22574 @code{F-floating}, @code{D-floating}, @code{G-floating}.
22575 Note that the @code{H-floating}
22576 type was available only on VAX systems, and is not available
22577 in either HP Ada or GNAT.
22580 @code{IEEE_Float} specifies that floating-point
22581 types are represented by default with the IEEE single and
22582 double floating-point types.
22586 GNAT provides an identical implementation of the pragma
22587 @code{Float_Representation}, except that it functions as a
22588 configuration pragma. Note that the
22589 notion of configuration pragma corresponds closely to the
22590 HP Ada notion of a program library switch.
22592 When no pragma is used in GNAT, the default is @code{IEEE_Float},
22594 from HP Ada 83, where the default is @code{VAX_Float}. In addition, the
22595 predefined libraries in GNAT are built using @code{IEEE_Float}, so it is not
22596 advisable to change the format of numbers passed to standard library
22597 routines, and if necessary explicit type conversions may be needed.
22599 The use of @code{IEEE_Float} is recommended in GNAT since it is more
22600 efficient, and (given that it conforms to an international standard)
22601 potentially more portable.
22602 The situation in which @code{VAX_Float} may be useful is in interfacing
22603 to existing code and data that expect the use of @code{VAX_Float}.
22604 In such a situation use the predefined @code{VAX_Float}
22605 types in package @code{System}, as extended by
22606 @code{Extend_System}. For example, use @code{System.F_Float}
22607 to specify the 32-bit @code{F-Float} format.
22610 On OpenVMS systems, HP Ada provides the pragma @code{Long_Float}
22611 to allow control over the internal representation chosen
22612 for the predefined type @code{Long_Float} and for floating-point
22613 type declarations with digits specified in the range 7 .. 15.
22614 The format of this pragma is as follows:
22616 @smallexample @c ada
22618 pragma Long_Float (D_FLOAT | G_FLOAT);
22622 @node Fixed-Point Types and Representations
22623 @subsection Fixed-Point Types and Representations
22626 On HP Ada for OpenVMS Alpha systems, rounding is
22627 away from zero for both positive and negative numbers.
22628 Therefore, @code{+0.5} rounds to @code{1},
22629 and @code{-0.5} rounds to @code{-1}.
22631 On GNAT the results of operations
22632 on fixed-point types are in accordance with the Ada
22633 rules. In particular, results of operations on decimal
22634 fixed-point types are truncated.
22636 @node Record and Array Component Alignment
22637 @subsection Record and Array Component Alignment
22640 On HP Ada for OpenVMS Alpha, all non-composite components
22641 are aligned on natural boundaries. For example, 1-byte
22642 components are aligned on byte boundaries, 2-byte
22643 components on 2-byte boundaries, 4-byte components on 4-byte
22644 byte boundaries, and so on. The OpenVMS Alpha hardware
22645 runs more efficiently with naturally aligned data.
22647 On GNAT, alignment rules are compatible
22648 with HP Ada for OpenVMS Alpha.
22650 @node Address Clauses
22651 @subsection Address Clauses
22654 In HP Ada and GNAT, address clauses are supported for
22655 objects and imported subprograms.
22656 The predefined type @code{System.Address} is a private type
22657 in both compilers on Alpha OpenVMS, with the same representation
22658 (it is simply a machine pointer). Addition, subtraction, and comparison
22659 operations are available in the standard Ada package
22660 @code{System.Storage_Elements}, or in package @code{System}
22661 if it is extended to include @code{System.Aux_DEC} using a
22662 pragma @code{Extend_System} as previously described.
22664 Note that code that @code{with}'s both this extended package @code{System}
22665 and the package @code{System.Storage_Elements} should not @code{use}
22666 both packages, or ambiguities will result. In general it is better
22667 not to mix these two sets of facilities. The Ada package was
22668 designed specifically to provide the kind of features that HP Ada
22669 adds directly to package @code{System}.
22671 The type @code{System.Address} is a 64-bit integer type in GNAT for
22672 I64 OpenVMS. For more information,
22673 see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
22675 GNAT is compatible with HP Ada in its handling of address
22676 clauses, except for some limitations in
22677 the form of address clauses for composite objects with
22678 initialization. Such address clauses are easily replaced
22679 by the use of an explicitly-defined constant as described
22680 in the Ada Reference Manual (13.1(22)). For example, the sequence
22683 @smallexample @c ada
22685 X, Y : Integer := Init_Func;
22686 Q : String (X .. Y) := "abc";
22688 for Q'Address use Compute_Address;
22693 will be rejected by GNAT, since the address cannot be computed at the time
22694 that @code{Q} is declared. To achieve the intended effect, write instead:
22696 @smallexample @c ada
22699 X, Y : Integer := Init_Func;
22700 Q_Address : constant Address := Compute_Address;
22701 Q : String (X .. Y) := "abc";
22703 for Q'Address use Q_Address;
22709 which will be accepted by GNAT (and other Ada compilers), and is also
22710 compatible with Ada 83. A fuller description of the restrictions
22711 on address specifications is found in @ref{Top, GNAT Reference Manual,
22712 About This Guide, gnat_rm, GNAT Reference Manual}.
22714 @node Other Representation Clauses
22715 @subsection Other Representation Clauses
22718 GNAT implements in a compatible manner all the representation
22719 clauses supported by HP Ada. In addition, GNAT
22720 implements the representation clause forms that were introduced in Ada 95,
22721 including @code{COMPONENT_SIZE} and @code{SIZE} clauses for objects.
22723 @node The Package STANDARD
22724 @section The Package @code{STANDARD}
22727 The package @code{STANDARD}, as implemented by HP Ada, is fully
22728 described in the @cite{Ada Reference Manual} and in the
22729 @cite{HP Ada Language Reference Manual}. As implemented by GNAT, the
22730 package @code{STANDARD} is described in the @cite{Ada Reference Manual}.
22732 In addition, HP Ada supports the Latin-1 character set in
22733 the type @code{CHARACTER}. GNAT supports the Latin-1 character set
22734 in the type @code{CHARACTER} and also Unicode (ISO 10646 BMP) in
22735 the type @code{WIDE_CHARACTER}.
22737 The floating-point types supported by GNAT are those
22738 supported by HP Ada, but the defaults are different, and are controlled by
22739 pragmas. See @ref{Floating-Point Types and Representations}, for details.
22741 @node The Package SYSTEM
22742 @section The Package @code{SYSTEM}
22745 HP Ada provides a specific version of the package
22746 @code{SYSTEM} for each platform on which the language is implemented.
22747 For the complete spec of the package @code{SYSTEM}, see
22748 Appendix F of the @cite{HP Ada Language Reference Manual}.
22750 On HP Ada, the package @code{SYSTEM} includes the following conversion
22753 @item @code{TO_ADDRESS(INTEGER)}
22755 @item @code{TO_ADDRESS(UNSIGNED_LONGWORD)}
22757 @item @code{TO_ADDRESS(}@i{universal_integer}@code{)}
22759 @item @code{TO_INTEGER(ADDRESS)}
22761 @item @code{TO_UNSIGNED_LONGWORD(ADDRESS)}
22763 @item Function @code{IMPORT_VALUE return UNSIGNED_LONGWORD} and the
22764 functions @code{IMPORT_ADDRESS} and @code{IMPORT_LARGEST_VALUE}
22768 By default, GNAT supplies a version of @code{SYSTEM} that matches
22769 the definition given in the @cite{Ada Reference Manual}.
22771 is a subset of the HP system definitions, which is as
22772 close as possible to the original definitions. The only difference
22773 is that the definition of @code{SYSTEM_NAME} is different:
22775 @smallexample @c ada
22777 type Name is (SYSTEM_NAME_GNAT);
22778 System_Name : constant Name := SYSTEM_NAME_GNAT;
22783 Also, GNAT adds the Ada declarations for
22784 @code{BIT_ORDER} and @code{DEFAULT_BIT_ORDER}.
22786 However, the use of the following pragma causes GNAT
22787 to extend the definition of package @code{SYSTEM} so that it
22788 encompasses the full set of HP-specific extensions,
22789 including the functions listed above:
22791 @smallexample @c ada
22793 pragma Extend_System (Aux_DEC);
22798 The pragma @code{Extend_System} is a configuration pragma that
22799 is most conveniently placed in the @file{gnat.adc} file. @xref{Pragma
22800 Extend_System,,, gnat_rm, GNAT Reference Manual}, for further details.
22802 HP Ada does not allow the recompilation of the package
22803 @code{SYSTEM}. Instead HP Ada provides several pragmas
22804 (@code{SYSTEM_NAME}, @code{STORAGE_UNIT}, and @code{MEMORY_SIZE})
22805 to modify values in the package @code{SYSTEM}.
22806 On OpenVMS Alpha systems, the pragma
22807 @code{SYSTEM_NAME} takes the enumeration literal @code{OPENVMS_AXP} as
22808 its single argument.
22810 GNAT does permit the recompilation of package @code{SYSTEM} using
22811 the special switch @option{-gnatg}, and this switch can be used if
22812 it is necessary to modify the definitions in @code{SYSTEM}. GNAT does
22813 not permit the specification of @code{SYSTEM_NAME}, @code{STORAGE_UNIT}
22814 or @code{MEMORY_SIZE} by any other means.
22816 On GNAT systems, the pragma @code{SYSTEM_NAME} takes the
22817 enumeration literal @code{SYSTEM_NAME_GNAT}.
22819 The definitions provided by the use of
22821 @smallexample @c ada
22822 pragma Extend_System (AUX_Dec);
22826 are virtually identical to those provided by the HP Ada 83 package
22827 @code{SYSTEM}. One important difference is that the name of the
22829 function for type @code{UNSIGNED_LONGWORD} is changed to
22830 @code{TO_ADDRESS_LONG}.
22831 @xref{Address Clauses,,, gnat_rm, GNAT Reference Manual}, for a
22832 discussion of why this change was necessary.
22835 The version of @code{TO_ADDRESS} taking a @i{universal_integer} argument
22837 an extension to Ada 83 not strictly compatible with the reference manual.
22838 GNAT, in order to be exactly compatible with the standard,
22839 does not provide this capability. In HP Ada 83, the
22840 point of this definition is to deal with a call like:
22842 @smallexample @c ada
22843 TO_ADDRESS (16#12777#);
22847 Normally, according to Ada 83 semantics, one would expect this to be
22848 ambiguous, since it matches both the @code{INTEGER} and
22849 @code{UNSIGNED_LONGWORD} forms of @code{TO_ADDRESS}.
22850 However, in HP Ada 83, there is no ambiguity, since the
22851 definition using @i{universal_integer} takes precedence.
22853 In GNAT, since the version with @i{universal_integer} cannot be supplied,
22855 not possible to be 100% compatible. Since there are many programs using
22856 numeric constants for the argument to @code{TO_ADDRESS}, the decision in
22858 to change the name of the function in the @code{UNSIGNED_LONGWORD} case,
22859 so the declarations provided in the GNAT version of @code{AUX_Dec} are:
22861 @smallexample @c ada
22862 function To_Address (X : Integer) return Address;
22863 pragma Pure_Function (To_Address);
22865 function To_Address_Long (X : Unsigned_Longword) return Address;
22866 pragma Pure_Function (To_Address_Long);
22870 This means that programs using @code{TO_ADDRESS} for
22871 @code{UNSIGNED_LONGWORD} must change the name to @code{TO_ADDRESS_LONG}.
22873 @node Tasking and Task-Related Features
22874 @section Tasking and Task-Related Features
22877 This section compares the treatment of tasking in GNAT
22878 and in HP Ada for OpenVMS Alpha.
22879 The GNAT description applies to both Alpha and I64 OpenVMS.
22880 For detailed information on tasking in
22881 HP Ada, see the @cite{HP Ada Language Reference Manual} and the
22882 relevant run-time reference manual.
22885 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
22886 * Assigning Task IDs::
22887 * Task IDs and Delays::
22888 * Task-Related Pragmas::
22889 * Scheduling and Task Priority::
22891 * External Interrupts::
22894 @node Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
22895 @subsection Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
22898 On OpenVMS Alpha systems, each Ada task (except a passive
22899 task) is implemented as a single stream of execution
22900 that is created and managed by the kernel. On these
22901 systems, HP Ada tasking support is based on DECthreads,
22902 an implementation of the POSIX standard for threads.
22904 Also, on OpenVMS Alpha systems, HP Ada tasks and foreign
22905 code that calls DECthreads routines can be used together.
22906 The interaction between Ada tasks and DECthreads routines
22907 can have some benefits. For example when on OpenVMS Alpha,
22908 HP Ada can call C code that is already threaded.
22910 GNAT uses the facilities of DECthreads,
22911 and Ada tasks are mapped to threads.
22913 @node Assigning Task IDs
22914 @subsection Assigning Task IDs
22917 The HP Ada Run-Time Library always assigns @code{%TASK 1} to
22918 the environment task that executes the main program. On
22919 OpenVMS Alpha systems, @code{%TASK 0} is often used for tasks
22920 that have been created but are not yet activated.
22922 On OpenVMS Alpha systems, task IDs are assigned at
22923 activation. On GNAT systems, task IDs are also assigned at
22924 task creation but do not have the same form or values as
22925 task ID values in HP Ada. There is no null task, and the
22926 environment task does not have a specific task ID value.
22928 @node Task IDs and Delays
22929 @subsection Task IDs and Delays
22932 On OpenVMS Alpha systems, tasking delays are implemented
22933 using Timer System Services. The Task ID is used for the
22934 identification of the timer request (the @code{REQIDT} parameter).
22935 If Timers are used in the application take care not to use
22936 @code{0} for the identification, because cancelling such a timer
22937 will cancel all timers and may lead to unpredictable results.
22939 @node Task-Related Pragmas
22940 @subsection Task-Related Pragmas
22943 Ada supplies the pragma @code{TASK_STORAGE}, which allows
22944 specification of the size of the guard area for a task
22945 stack. (The guard area forms an area of memory that has no
22946 read or write access and thus helps in the detection of
22947 stack overflow.) On OpenVMS Alpha systems, if the pragma
22948 @code{TASK_STORAGE} specifies a value of zero, a minimal guard
22949 area is created. In the absence of a pragma @code{TASK_STORAGE},
22950 a default guard area is created.
22952 GNAT supplies the following task-related pragma:
22955 @item @code{TASK_STORAGE}
22957 GNAT implements pragma @code{TASK_STORAGE} in the same way as HP Ada.
22958 Both HP Ada and GNAT supply the pragmas @code{PASSIVE},
22959 @code{SUPPRESS}, and @code{VOLATILE}.
22962 @node Scheduling and Task Priority
22963 @subsection Scheduling and Task Priority
22966 HP Ada implements the Ada language requirement that
22967 when two tasks are eligible for execution and they have
22968 different priorities, the lower priority task does not
22969 execute while the higher priority task is waiting. The HP
22970 Ada Run-Time Library keeps a task running until either the
22971 task is suspended or a higher priority task becomes ready.
22973 On OpenVMS Alpha systems, the default strategy is round-
22974 robin with preemption. Tasks of equal priority take turns
22975 at the processor. A task is run for a certain period of
22976 time and then placed at the tail of the ready queue for
22977 its priority level.
22979 HP Ada provides the implementation-defined pragma @code{TIME_SLICE},
22980 which can be used to enable or disable round-robin
22981 scheduling of tasks with the same priority.
22982 See the relevant HP Ada run-time reference manual for
22983 information on using the pragmas to control HP Ada task
22986 GNAT follows the scheduling rules of Annex D (Real-Time
22987 Annex) of the @cite{Ada Reference Manual}. In general, this
22988 scheduling strategy is fully compatible with HP Ada
22989 although it provides some additional constraints (as
22990 fully documented in Annex D).
22991 GNAT implements time slicing control in a manner compatible with
22992 HP Ada 83, by means of the pragma @code{Time_Slice}, whose semantics
22993 are identical to the HP Ada 83 pragma of the same name.
22994 Note that it is not possible to mix GNAT tasking and
22995 HP Ada 83 tasking in the same program, since the two run-time
22996 libraries are not compatible.
22998 @node The Task Stack
22999 @subsection The Task Stack
23002 In HP Ada, a task stack is allocated each time a
23003 non-passive task is activated. As soon as the task is
23004 terminated, the storage for the task stack is deallocated.
23005 If you specify a size of zero (bytes) with @code{T'STORAGE_SIZE},
23006 a default stack size is used. Also, regardless of the size
23007 specified, some additional space is allocated for task
23008 management purposes. On OpenVMS Alpha systems, at least
23009 one page is allocated.
23011 GNAT handles task stacks in a similar manner. In accordance with
23012 the Ada rules, it provides the pragma @code{STORAGE_SIZE} as
23013 an alternative method for controlling the task stack size.
23014 The specification of the attribute @code{T'STORAGE_SIZE} is also
23015 supported in a manner compatible with HP Ada.
23017 @node External Interrupts
23018 @subsection External Interrupts
23021 On HP Ada, external interrupts can be associated with task entries.
23022 GNAT is compatible with HP Ada in its handling of external interrupts.
23024 @node Pragmas and Pragma-Related Features
23025 @section Pragmas and Pragma-Related Features
23028 Both HP Ada and GNAT supply all language-defined pragmas
23029 as specified by the Ada 83 standard. GNAT also supplies all
23030 language-defined pragmas introduced by Ada 95 and Ada 2005.
23031 In addition, GNAT implements the implementation-defined pragmas
23035 @item @code{AST_ENTRY}
23037 @item @code{COMMON_OBJECT}
23039 @item @code{COMPONENT_ALIGNMENT}
23041 @item @code{EXPORT_EXCEPTION}
23043 @item @code{EXPORT_FUNCTION}
23045 @item @code{EXPORT_OBJECT}
23047 @item @code{EXPORT_PROCEDURE}
23049 @item @code{EXPORT_VALUED_PROCEDURE}
23051 @item @code{FLOAT_REPRESENTATION}
23055 @item @code{IMPORT_EXCEPTION}
23057 @item @code{IMPORT_FUNCTION}
23059 @item @code{IMPORT_OBJECT}
23061 @item @code{IMPORT_PROCEDURE}
23063 @item @code{IMPORT_VALUED_PROCEDURE}
23065 @item @code{INLINE_GENERIC}
23067 @item @code{INTERFACE_NAME}
23069 @item @code{LONG_FLOAT}
23071 @item @code{MAIN_STORAGE}
23073 @item @code{PASSIVE}
23075 @item @code{PSECT_OBJECT}
23077 @item @code{SHARE_GENERIC}
23079 @item @code{SUPPRESS_ALL}
23081 @item @code{TASK_STORAGE}
23083 @item @code{TIME_SLICE}
23089 These pragmas are all fully implemented, with the exception of @code{TITLE},
23090 @code{PASSIVE}, and @code{SHARE_GENERIC}, which are
23091 recognized, but which have no
23092 effect in GNAT. The effect of @code{PASSIVE} may be obtained by the
23093 use of Ada protected objects. In GNAT, all generics are inlined.
23095 Unlike HP Ada, the GNAT ``@code{EXPORT_}@i{subprogram}'' pragmas require
23096 a separate subprogram specification which must appear before the
23099 GNAT also supplies a number of implementation-defined pragmas including the
23103 @item @code{ABORT_DEFER}
23105 @item @code{ADA_83}
23107 @item @code{ADA_95}
23109 @item @code{ADA_05}
23111 @item @code{Ada_2005}
23113 @item @code{Ada_12}
23115 @item @code{Ada_2012}
23117 @item @code{ALLOW_INTEGER_ADDRESS}
23119 @item @code{ANNOTATE}
23121 @item @code{ASSERT}
23123 @item @code{C_PASS_BY_COPY}
23125 @item @code{CPP_CLASS}
23127 @item @code{CPP_CONSTRUCTOR}
23129 @item @code{CPP_DESTRUCTOR}
23133 @item @code{EXTEND_SYSTEM}
23135 @item @code{LINKER_ALIAS}
23137 @item @code{LINKER_SECTION}
23139 @item @code{MACHINE_ATTRIBUTE}
23141 @item @code{NO_RETURN}
23143 @item @code{PURE_FUNCTION}
23145 @item @code{SOURCE_FILE_NAME}
23147 @item @code{SOURCE_REFERENCE}
23149 @item @code{UNCHECKED_UNION}
23151 @item @code{UNIMPLEMENTED_UNIT}
23153 @item @code{UNIVERSAL_DATA}
23155 @item @code{UNSUPPRESS}
23157 @item @code{WARNINGS}
23159 @item @code{WEAK_EXTERNAL}
23163 For full details on these and other GNAT implementation-defined pragmas,
23164 see @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference
23168 * Restrictions on the Pragma INLINE::
23169 * Restrictions on the Pragma INTERFACE::
23170 * Restrictions on the Pragma SYSTEM_NAME::
23173 @node Restrictions on the Pragma INLINE
23174 @subsection Restrictions on Pragma @code{INLINE}
23177 HP Ada enforces the following restrictions on the pragma @code{INLINE}:
23179 @item Parameters cannot have a task type.
23181 @item Function results cannot be task types, unconstrained
23182 array types, or unconstrained types with discriminants.
23184 @item Bodies cannot declare the following:
23186 @item Subprogram body or stub (imported subprogram is allowed)
23190 @item Generic declarations
23192 @item Instantiations
23196 @item Access types (types derived from access types allowed)
23198 @item Array or record types
23200 @item Dependent tasks
23202 @item Direct recursive calls of subprogram or containing
23203 subprogram, directly or via a renaming
23209 In GNAT, the only restriction on pragma @code{INLINE} is that the
23210 body must occur before the call if both are in the same
23211 unit, and the size must be appropriately small. There are
23212 no other specific restrictions which cause subprograms to
23213 be incapable of being inlined.
23215 @node Restrictions on the Pragma INTERFACE
23216 @subsection Restrictions on Pragma @code{INTERFACE}
23219 The following restrictions on pragma @code{INTERFACE}
23220 are enforced by both HP Ada and GNAT:
23222 @item Languages accepted: Ada, Bliss, C, Fortran, Default.
23223 Default is the default on OpenVMS Alpha systems.
23225 @item Parameter passing: Language specifies default
23226 mechanisms but can be overridden with an @code{EXPORT} pragma.
23229 @item Ada: Use internal Ada rules.
23231 @item Bliss, C: Parameters must be mode @code{in}; cannot be
23232 record or task type. Result cannot be a string, an
23233 array, or a record.
23235 @item Fortran: Parameters cannot have a task type. Result cannot
23236 be a string, an array, or a record.
23241 GNAT is entirely upwards compatible with HP Ada, and in addition allows
23242 record parameters for all languages.
23244 @node Restrictions on the Pragma SYSTEM_NAME
23245 @subsection Restrictions on Pragma @code{SYSTEM_NAME}
23248 For HP Ada for OpenVMS Alpha, the enumeration literal
23249 for the type @code{NAME} is @code{OPENVMS_AXP}.
23250 In GNAT, the enumeration
23251 literal for the type @code{NAME} is @code{SYSTEM_NAME_GNAT}.
23253 @node Library of Predefined Units
23254 @section Library of Predefined Units
23257 A library of predefined units is provided as part of the
23258 HP Ada and GNAT implementations. HP Ada does not provide
23259 the package @code{MACHINE_CODE} but instead recommends importing
23262 The GNAT versions of the HP Ada Run-Time Library (@code{ADA$PREDEFINED:})
23263 units are taken from the OpenVMS Alpha version, not the OpenVMS VAX
23265 The HP Ada Predefined Library units are modified to remove post-Ada 83
23266 incompatibilities and to make them interoperable with GNAT
23267 (@pxref{Changes to DECLIB}, for details).
23268 The units are located in the @file{DECLIB} directory.
23270 The GNAT RTL is contained in
23271 the @file{ADALIB} directory, and
23272 the default search path is set up to find @code{DECLIB} units in preference
23273 to @code{ADALIB} units with the same name (@code{TEXT_IO},
23274 @code{SEQUENTIAL_IO}, and @code{DIRECT_IO}, for example).
23277 * Changes to DECLIB::
23280 @node Changes to DECLIB
23281 @subsection Changes to @code{DECLIB}
23284 The changes made to the HP Ada predefined library for GNAT and post-Ada 83
23285 compatibility are minor and include the following:
23288 @item Adjusting the location of pragmas and record representation
23289 clauses to obey Ada 95 (and thus Ada 2005) rules
23291 @item Adding the proper notation to generic formal parameters
23292 that take unconstrained types in instantiation
23294 @item Adding pragma @code{ELABORATE_BODY} to package specs
23295 that have package bodies not otherwise allowed
23297 @item Replacing occurrences of the identifier ``@code{PROTECTED}'' by
23298 ``@code{PROTECTD}''.
23299 Currently these are found only in the @code{STARLET} package spec.
23301 @item Changing @code{SYSTEM.ADDRESS} to @code{SYSTEM.SHORT_ADDRESS}
23302 where the address size is constrained to 32 bits.
23306 None of the above changes is visible to users.
23312 On OpenVMS Alpha, HP Ada provides the following strongly-typed bindings:
23315 @item Command Language Interpreter (CLI interface)
23317 @item DECtalk Run-Time Library (DTK interface)
23319 @item Librarian utility routines (LBR interface)
23321 @item General Purpose Run-Time Library (LIB interface)
23323 @item Math Run-Time Library (MTH interface)
23325 @item National Character Set Run-Time Library (NCS interface)
23327 @item Compiled Code Support Run-Time Library (OTS interface)
23329 @item Parallel Processing Run-Time Library (PPL interface)
23331 @item Screen Management Run-Time Library (SMG interface)
23333 @item Sort Run-Time Library (SOR interface)
23335 @item String Run-Time Library (STR interface)
23337 @item STARLET System Library
23340 @item X Window System Version 11R4 and 11R5 (X, XLIB interface)
23342 @item X Windows Toolkit (XT interface)
23344 @item X/Motif Version 1.1.3 and 1.2 (XM interface)
23348 GNAT provides implementations of these HP bindings in the @code{DECLIB}
23349 directory, on both the Alpha and I64 OpenVMS platforms.
23351 The X components of DECLIB compatibility package are located in a separate
23352 library, called XDECGNAT, which is not linked with by default; this library
23353 must be explicitly linked with any application that makes use of any X facilities,
23354 with a command similar to
23356 @code{GNAT MAKE USE_X /LINK /LIBRARY=XDECGNAT}
23358 The X/Motif bindings used to build @code{DECLIB} are whatever versions are
23360 HP Ada @file{ADA$PREDEFINED} directory with extension @file{.ADC}.
23361 A pragma @code{Linker_Options} has been added to packages @code{Xm},
23362 @code{Xt}, and @code{X_Lib}
23363 causing the default X/Motif sharable image libraries to be linked in. This
23364 is done via options files named @file{xm.opt}, @file{xt.opt}, and
23365 @file{x_lib.opt} (also located in the @file{DECLIB} directory).
23367 It may be necessary to edit these options files to update or correct the
23368 library names if, for example, the newer X/Motif bindings from
23369 @file{ADA$EXAMPLES}
23370 had been (previous to installing GNAT) copied and renamed to supersede the
23371 default @file{ADA$PREDEFINED} versions.
23374 * Shared Libraries and Options Files::
23375 * Interfaces to C::
23378 @node Shared Libraries and Options Files
23379 @subsection Shared Libraries and Options Files
23382 When using the HP Ada
23383 predefined X and Motif bindings, the linking with their sharable images is
23384 done automatically by @command{GNAT LINK}.
23385 When using other X and Motif bindings, you need
23386 to add the corresponding sharable images to the command line for
23387 @code{GNAT LINK}. When linking with shared libraries, or with
23388 @file{.OPT} files, you must
23389 also add them to the command line for @command{GNAT LINK}.
23391 A shared library to be used with GNAT is built in the same way as other
23392 libraries under VMS. The VMS Link command can be used in standard fashion.
23394 @node Interfaces to C
23395 @subsection Interfaces to C
23399 provides the following Ada types and operations:
23402 @item C types package (@code{C_TYPES})
23404 @item C strings (@code{C_TYPES.NULL_TERMINATED})
23406 @item Other_types (@code{SHORT_INT})
23410 Interfacing to C with GNAT, you can use the above approach
23411 described for HP Ada or the facilities of Annex B of
23412 the @cite{Ada Reference Manual} (packages @code{INTERFACES.C},
23413 @code{INTERFACES.C.STRINGS} and @code{INTERFACES.C.POINTERS}). For more
23414 information, see @ref{Interfacing to C,,, gnat_rm, GNAT Reference Manual}.
23416 The @option{-gnatF} qualifier forces default and explicit
23417 @code{External_Name} parameters in pragmas @code{Import} and @code{Export}
23418 to be uppercased for compatibility with the default behavior
23419 of HP C. The qualifier has no effect on @code{Link_Name} parameters.
23421 @node Main Program Definition
23422 @section Main Program Definition
23425 The following section discusses differences in the
23426 definition of main programs on HP Ada and GNAT.
23427 On HP Ada, main programs are defined to meet the
23428 following conditions:
23430 @item Procedure with no formal parameters (returns @code{0} upon
23433 @item Procedure with no formal parameters (returns @code{42} when
23434 an unhandled exception is raised)
23436 @item Function with no formal parameters whose returned value
23437 is of a discrete type
23439 @item Procedure with one @code{out} formal of a discrete type for
23440 which a specification of pragma @code{EXPORT_VALUED_PROCEDURE} is given.
23445 When declared with the pragma @code{EXPORT_VALUED_PROCEDURE},
23446 a main function or main procedure returns a discrete
23447 value whose size is less than 64 bits (32 on VAX systems),
23448 the value is zero- or sign-extended as appropriate.
23449 On GNAT, main programs are defined as follows:
23451 @item Must be a non-generic, parameterless subprogram that
23452 is either a procedure or function returning an Ada
23453 @code{STANDARD.INTEGER} (the predefined type)
23455 @item Cannot be a generic subprogram or an instantiation of a
23459 @node Implementation-Defined Attributes
23460 @section Implementation-Defined Attributes
23463 GNAT provides all HP Ada implementation-defined
23466 @node Compiler and Run-Time Interfacing
23467 @section Compiler and Run-Time Interfacing
23470 HP Ada provides the following qualifiers to pass options to the linker
23473 @item @option{/WAIT} and @option{/SUBMIT}
23475 @item @option{/COMMAND}
23477 @item @option{/@r{[}NO@r{]}MAP}
23479 @item @option{/OUTPUT=@var{file-spec}}
23481 @item @option{/@r{[}NO@r{]}DEBUG} and @option{/@r{[}NO@r{]}TRACEBACK}
23485 To pass options to the linker, GNAT provides the following
23489 @item @option{/EXECUTABLE=@var{exec-name}}
23491 @item @option{/VERBOSE}
23493 @item @option{/@r{[}NO@r{]}DEBUG} and @option{/@r{[}NO@r{]}TRACEBACK}
23497 For more information on these switches, see
23498 @ref{Switches for gnatlink}.
23499 In HP Ada, the command-line switch @option{/OPTIMIZE} is available
23500 to control optimization. HP Ada also supplies the
23503 @item @code{OPTIMIZE}
23505 @item @code{INLINE}
23507 @item @code{INLINE_GENERIC}
23509 @item @code{SUPPRESS_ALL}
23511 @item @code{PASSIVE}
23515 In GNAT, optimization is controlled strictly by command
23516 line parameters, as described in the corresponding section of this guide.
23517 The HP pragmas for control of optimization are
23518 recognized but ignored.
23520 Note that in GNAT, the default is optimization off, whereas in HP Ada
23521 the default is that optimization is turned on.
23523 @node Program Compilation and Library Management
23524 @section Program Compilation and Library Management
23527 HP Ada and GNAT provide a comparable set of commands to
23528 build programs. HP Ada also provides a program library,
23529 which is a concept that does not exist on GNAT. Instead,
23530 GNAT provides directories of sources that are compiled as
23533 The following table summarizes
23534 the HP Ada commands and provides
23535 equivalent GNAT commands. In this table, some GNAT
23536 equivalents reflect the fact that GNAT does not use the
23537 concept of a program library. Instead, it uses a model
23538 in which collections of source and object files are used
23539 in a manner consistent with other languages like C and
23540 Fortran. Therefore, standard system file commands are used
23541 to manipulate these elements. Those GNAT commands are marked with
23543 Note that, unlike HP Ada, none of the GNAT commands accepts wild cards.
23546 @multitable @columnfractions .35 .65
23548 @item @emph{HP Ada Command}
23549 @tab @emph{GNAT Equivalent / Description}
23551 @item @command{ADA}
23552 @tab @command{GNAT COMPILE}@*
23553 Invokes the compiler to compile one or more Ada source files.
23555 @item @command{ACS ATTACH}@*
23556 @tab [No equivalent]@*
23557 Switches control of terminal from current process running the program
23560 @item @command{ACS CHECK}
23561 @tab @command{GNAT MAKE /DEPENDENCY_LIST}@*
23562 Forms the execution closure of one
23563 or more compiled units and checks completeness and currency.
23565 @item @command{ACS COMPILE}
23566 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
23567 Forms the execution closure of one or
23568 more specified units, checks completeness and currency,
23569 identifies units that have revised source files, compiles same,
23570 and recompiles units that are or will become obsolete.
23571 Also completes incomplete generic instantiations.
23573 @item @command{ACS COPY FOREIGN}
23575 Copies a foreign object file into the program library as a
23578 @item @command{ACS COPY UNIT}
23580 Copies a compiled unit from one program library to another.
23582 @item @command{ACS CREATE LIBRARY}
23583 @tab Create /directory (*)@*
23584 Creates a program library.
23586 @item @command{ACS CREATE SUBLIBRARY}
23587 @tab Create /directory (*)@*
23588 Creates a program sublibrary.
23590 @item @command{ACS DELETE LIBRARY}
23592 Deletes a program library and its contents.
23594 @item @command{ACS DELETE SUBLIBRARY}
23596 Deletes a program sublibrary and its contents.
23598 @item @command{ACS DELETE UNIT}
23599 @tab Delete file (*)@*
23600 On OpenVMS systems, deletes one or more compiled units from
23601 the current program library.
23603 @item @command{ACS DIRECTORY}
23604 @tab Directory (*)@*
23605 On OpenVMS systems, lists units contained in the current
23608 @item @command{ACS ENTER FOREIGN}
23610 Allows the import of a foreign body as an Ada library
23611 spec and enters a reference to a pointer.
23613 @item @command{ACS ENTER UNIT}
23615 Enters a reference (pointer) from the current program library to
23616 a unit compiled into another program library.
23618 @item @command{ACS EXIT}
23619 @tab [No equivalent]@*
23620 Exits from the program library manager.
23622 @item @command{ACS EXPORT}
23624 Creates an object file that contains system-specific object code
23625 for one or more units. With GNAT, object files can simply be copied
23626 into the desired directory.
23628 @item @command{ACS EXTRACT SOURCE}
23630 Allows access to the copied source file for each Ada compilation unit
23632 @item @command{ACS HELP}
23633 @tab @command{HELP GNAT}@*
23634 Provides online help.
23636 @item @command{ACS LINK}
23637 @tab @command{GNAT LINK}@*
23638 Links an object file containing Ada units into an executable file.
23640 @item @command{ACS LOAD}
23642 Loads (partially compiles) Ada units into the program library.
23643 Allows loading a program from a collection of files into a library
23644 without knowing the relationship among units.
23646 @item @command{ACS MERGE}
23648 Merges into the current program library, one or more units from
23649 another library where they were modified.
23651 @item @command{ACS RECOMPILE}
23652 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
23653 Recompiles from external or copied source files any obsolete
23654 unit in the closure. Also, completes any incomplete generic
23657 @item @command{ACS REENTER}
23658 @tab @command{GNAT MAKE}@*
23659 Reenters current references to units compiled after last entered
23660 with the @command{ACS ENTER UNIT} command.
23662 @item @command{ACS SET LIBRARY}
23663 @tab Set default (*)@*
23664 Defines a program library to be the compilation context as well
23665 as the target library for compiler output and commands in general.
23667 @item @command{ACS SET PRAGMA}
23668 @tab Edit @file{gnat.adc} (*)@*
23669 Redefines specified values of the library characteristics
23670 @code{LONG_ FLOAT}, @code{MEMORY_SIZE}, @code{SYSTEM_NAME},
23671 and @code{Float_Representation}.
23673 @item @command{ACS SET SOURCE}
23674 @tab Define @code{ADA_INCLUDE_PATH} path (*)@*
23675 Defines the source file search list for the @command{ACS COMPILE} command.
23677 @item @command{ACS SHOW LIBRARY}
23678 @tab Directory (*)@*
23679 Lists information about one or more program libraries.
23681 @item @command{ACS SHOW PROGRAM}
23682 @tab [No equivalent]@*
23683 Lists information about the execution closure of one or
23684 more units in the program library.
23686 @item @command{ACS SHOW SOURCE}
23687 @tab Show logical @code{ADA_INCLUDE_PATH}@*
23688 Shows the source file search used when compiling units.
23690 @item @command{ACS SHOW VERSION}
23691 @tab Compile with @option{VERBOSE} option
23692 Displays the version number of the compiler and program library
23695 @item @command{ACS SPAWN}
23696 @tab [No equivalent]@*
23697 Creates a subprocess of the current process (same as @command{DCL SPAWN}
23700 @item @command{ACS VERIFY}
23701 @tab [No equivalent]@*
23702 Performs a series of consistency checks on a program library to
23703 determine whether the library structure and library files are in
23710 @section Input-Output
23713 On OpenVMS Alpha systems, HP Ada uses OpenVMS Record
23714 Management Services (RMS) to perform operations on
23718 HP Ada and GNAT predefine an identical set of input-
23719 output packages. To make the use of the
23720 generic @code{TEXT_IO} operations more convenient, HP Ada
23721 provides predefined library packages that instantiate the
23722 integer and floating-point operations for the predefined
23723 integer and floating-point types as shown in the following table.
23725 @multitable @columnfractions .45 .55
23726 @item @emph{Package Name} @tab Instantiation
23728 @item @code{INTEGER_TEXT_IO}
23729 @tab @code{INTEGER_IO(INTEGER)}
23731 @item @code{SHORT_INTEGER_TEXT_IO}
23732 @tab @code{INTEGER_IO(SHORT_INTEGER)}
23734 @item @code{SHORT_SHORT_INTEGER_TEXT_IO}
23735 @tab @code{INTEGER_IO(SHORT_SHORT_INTEGER)}
23737 @item @code{FLOAT_TEXT_IO}
23738 @tab @code{FLOAT_IO(FLOAT)}
23740 @item @code{LONG_FLOAT_TEXT_IO}
23741 @tab @code{FLOAT_IO(LONG_FLOAT)}
23745 The HP Ada predefined packages and their operations
23746 are implemented using OpenVMS Alpha files and input-output
23747 facilities. HP Ada supports asynchronous input-output on OpenVMS Alpha.
23748 Familiarity with the following is recommended:
23750 @item RMS file organizations and access methods
23752 @item OpenVMS file specifications and directories
23754 @item OpenVMS File Definition Language (FDL)
23758 GNAT provides I/O facilities that are completely
23759 compatible with HP Ada. The distribution includes the
23760 standard HP Ada versions of all I/O packages, operating
23761 in a manner compatible with HP Ada. In particular, the
23762 following packages are by default the HP Ada (Ada 83)
23763 versions of these packages rather than the renamings
23764 suggested in Annex J of the Ada Reference Manual:
23766 @item @code{TEXT_IO}
23768 @item @code{SEQUENTIAL_IO}
23770 @item @code{DIRECT_IO}
23774 The use of the standard child package syntax (for
23775 example, @code{ADA.TEXT_IO}) retrieves the post-Ada 83 versions of these
23777 GNAT provides HP-compatible predefined instantiations
23778 of the @code{TEXT_IO} packages, and also
23779 provides the standard predefined instantiations required
23780 by the @cite{Ada Reference Manual}.
23782 For further information on how GNAT interfaces to the file
23783 system or how I/O is implemented in programs written in
23784 mixed languages, see @ref{Implementation of the Standard I/O,,,
23785 gnat_rm, GNAT Reference Manual}.
23786 This chapter covers the following:
23788 @item Standard I/O packages
23790 @item @code{FORM} strings
23792 @item @code{ADA.DIRECT_IO}
23794 @item @code{ADA.SEQUENTIAL_IO}
23796 @item @code{ADA.TEXT_IO}
23798 @item Stream pointer positioning
23800 @item Reading and writing non-regular files
23802 @item @code{GET_IMMEDIATE}
23804 @item Treating @code{TEXT_IO} files as streams
23811 @node Implementation Limits
23812 @section Implementation Limits
23815 The following table lists implementation limits for HP Ada
23817 @multitable @columnfractions .60 .20 .20
23819 @item @emph{Compilation Parameter}
23824 @item In a subprogram or entry declaration, maximum number of
23825 formal parameters that are of an unconstrained record type
23830 @item Maximum identifier length (number of characters)
23835 @item Maximum number of characters in a source line
23840 @item Maximum collection size (number of bytes)
23845 @item Maximum number of discriminants for a record type
23850 @item Maximum number of formal parameters in an entry or
23851 subprogram declaration
23856 @item Maximum number of dimensions in an array type
23861 @item Maximum number of library units and subunits in a compilation.
23866 @item Maximum number of library units and subunits in an execution.
23871 @item Maximum number of objects declared with the pragma @code{COMMON_OBJECT}
23872 or @code{PSECT_OBJECT}
23877 @item Maximum number of enumeration literals in an enumeration type
23883 @item Maximum number of lines in a source file
23888 @item Maximum number of bits in any object
23893 @item Maximum size of the static portion of a stack frame (approximate)
23898 @node Tools and Utilities
23899 @section Tools and Utilities
23902 The following table lists some of the OpenVMS development tools
23903 available for HP Ada, and the corresponding tools for
23904 use with @value{EDITION} on Alpha and I64 platforms.
23905 Aside from the debugger, all the OpenVMS tools identified are part
23906 of the DECset package.
23909 @c Specify table in TeX since Texinfo does a poor job
23913 \settabs\+Language-Sensitive Editor\quad
23914 &Product with HP Ada\quad
23917 &\it Product with HP Ada
23918 & \it Product with @value{EDITION}\cr
23920 \+Code Management System
23924 \+Language-Sensitive Editor
23926 & emacs or HP LSE (Alpha)\cr
23936 & OpenVMS Debug (I64)\cr
23938 \+Source Code Analyzer /
23955 \+Coverage Analyzer
23959 \+Module Management
23961 & Not applicable\cr
23971 @c This is the Texinfo version of the table. It renders poorly in pdf, hence
23972 @c the TeX version above for the printed version
23974 @c @multitable @columnfractions .3 .4 .4
23975 @multitable {Source Code Analyzer /}{Tool with HP Ada}{Tool with @value{EDITION}}
23977 @tab @i{Tool with HP Ada}
23978 @tab @i{Tool with @value{EDITION}}
23979 @item Code Management@*System
23982 @item Language-Sensitive@*Editor
23984 @tab emacs or HP LSE (Alpha)
23993 @tab OpenVMS Debug (I64)
23994 @item Source Code Analyzer /@*Cross Referencer
23998 @tab HP Digital Test@*Manager (DTM)
24000 @item Performance and@*Coverage Analyzer
24003 @item Module Management@*System
24005 @tab Not applicable
24012 @c **************************************
24013 @node Platform-Specific Information for the Run-Time Libraries
24014 @appendix Platform-Specific Information for the Run-Time Libraries
24015 @cindex Tasking and threads libraries
24016 @cindex Threads libraries and tasking
24017 @cindex Run-time libraries (platform-specific information)
24020 The GNAT run-time implementation may vary with respect to both the
24021 underlying threads library and the exception handling scheme.
24022 For threads support, one or more of the following are supplied:
24024 @item @b{native threads library}, a binding to the thread package from
24025 the underlying operating system
24027 @item @b{pthreads library} (Sparc Solaris only), a binding to the Solaris
24028 POSIX thread package
24032 For exception handling, either or both of two models are supplied:
24034 @item @b{Zero-Cost Exceptions} (``ZCX''),@footnote{
24035 Most programs should experience a substantial speed improvement by
24036 being compiled with a ZCX run-time.
24037 This is especially true for
24038 tasking applications or applications with many exception handlers.}
24039 @cindex Zero-Cost Exceptions
24040 @cindex ZCX (Zero-Cost Exceptions)
24041 which uses binder-generated tables that
24042 are interrogated at run time to locate a handler
24044 @item @b{setjmp / longjmp} (``SJLJ''),
24045 @cindex setjmp/longjmp Exception Model
24046 @cindex SJLJ (setjmp/longjmp Exception Model)
24047 which uses dynamically-set data to establish
24048 the set of handlers
24052 This appendix summarizes which combinations of threads and exception support
24053 are supplied on various GNAT platforms.
24054 It then shows how to select a particular library either
24055 permanently or temporarily,
24056 explains the properties of (and tradeoffs among) the various threads
24057 libraries, and provides some additional
24058 information about several specific platforms.
24061 * Summary of Run-Time Configurations::
24062 * Specifying a Run-Time Library::
24063 * Choosing the Scheduling Policy::
24064 * Solaris-Specific Considerations::
24065 * Linux-Specific Considerations::
24066 * AIX-Specific Considerations::
24067 * RTX-Specific Considerations::
24068 * HP-UX-Specific Considerations::
24071 @node Summary of Run-Time Configurations
24072 @section Summary of Run-Time Configurations
24074 @multitable @columnfractions .30 .70
24075 @item @b{alpha-openvms}
24076 @item @code{@ @ }@i{rts-native (default)}
24077 @item @code{@ @ @ @ }Tasking @tab native VMS threads
24078 @item @code{@ @ @ @ }Exceptions @tab ZCX
24080 @item @code{@ @ }@i{rts-sjlj}
24081 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
24082 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24084 @item @b{ia64-hp_linux}
24085 @item @code{@ @ }@i{rts-native (default)}
24086 @item @code{@ @ @ @ }Tasking @tab pthread library
24087 @item @code{@ @ @ @ }Exceptions @tab ZCX
24089 @item @b{ia64-hpux}
24090 @item @code{@ @ }@i{rts-native (default)}
24091 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
24092 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24094 @item @b{ia64-openvms}
24095 @item @code{@ @ }@i{rts-native (default)}
24096 @item @code{@ @ @ @ }Tasking @tab native VMS threads
24097 @item @code{@ @ @ @ }Exceptions @tab ZCX
24099 @item @b{ia64-sgi_linux}
24100 @item @code{@ @ }@i{rts-native (default)}
24101 @item @code{@ @ @ @ }Tasking @tab pthread library
24102 @item @code{@ @ @ @ }Exceptions @tab ZCX
24105 @item @code{@ @ }@i{rts-native (default)}
24106 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
24107 @item @code{@ @ @ @ }Exceptions @tab ZCX
24109 @item @code{@ @ }@i{rts-sjlj}
24110 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
24111 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24114 @item @code{@ @ }@i{rts-native (default)}
24115 @item @code{@ @ @ @ }Tasking @tab native AIX threads
24116 @item @code{@ @ @ @ }Exceptions @tab ZCX
24118 @item @code{@ @ }@i{rts-sjlj}
24119 @item @code{@ @ @ @ }Tasking @tab native AIX threads
24120 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24122 @item @b{ppc-darwin}
24123 @item @code{@ @ }@i{rts-native (default)}
24124 @item @code{@ @ @ @ }Tasking @tab native MacOS threads
24125 @item @code{@ @ @ @ }Exceptions @tab ZCX
24127 @item @b{sparc-solaris} @tab
24128 @item @code{@ @ }@i{rts-native (default)}
24129 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24130 @item @code{@ @ @ @ }Exceptions @tab ZCX
24132 @item @code{@ @ }@i{rts-pthread}
24133 @item @code{@ @ @ @ }Tasking @tab pthread library
24134 @item @code{@ @ @ @ }Exceptions @tab ZCX
24136 @item @code{@ @ }@i{rts-sjlj}
24137 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24138 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24140 @item @b{sparc64-solaris} @tab
24141 @item @code{@ @ }@i{rts-native (default)}
24142 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24143 @item @code{@ @ @ @ }Exceptions @tab ZCX
24145 @item @b{x86-linux}
24146 @item @code{@ @ }@i{rts-native (default)}
24147 @item @code{@ @ @ @ }Tasking @tab pthread library
24148 @item @code{@ @ @ @ }Exceptions @tab ZCX
24150 @item @code{@ @ }@i{rts-sjlj}
24151 @item @code{@ @ @ @ }Tasking @tab pthread library
24152 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24155 @item @code{@ @ }@i{rts-native (default)}
24156 @item @code{@ @ @ @ }Tasking @tab native LynxOS threads
24157 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24159 @item @b{x86-solaris}
24160 @item @code{@ @ }@i{rts-native (default)}
24161 @item @code{@ @ @ @ }Tasking @tab native Solaris threads
24162 @item @code{@ @ @ @ }Exceptions @tab ZCX
24164 @item @code{@ @ }@i{rts-sjlj}
24165 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24166 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24168 @item @b{x86-windows}
24169 @item @code{@ @ }@i{rts-native (default)}
24170 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
24171 @item @code{@ @ @ @ }Exceptions @tab ZCX
24173 @item @code{@ @ }@i{rts-sjlj}
24174 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
24175 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24177 @item @b{x86-windows-rtx}
24178 @item @code{@ @ }@i{rts-rtx-rtss (default)}
24179 @item @code{@ @ @ @ }Tasking @tab RTX real-time subsystem RTSS threads (kernel mode)
24180 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24182 @item @code{@ @ }@i{rts-rtx-w32}
24183 @item @code{@ @ @ @ }Tasking @tab RTX Win32 threads (user mode)
24184 @item @code{@ @ @ @ }Exceptions @tab ZCX
24186 @item @b{x86_64-linux}
24187 @item @code{@ @ }@i{rts-native (default)}
24188 @item @code{@ @ @ @ }Tasking @tab pthread library
24189 @item @code{@ @ @ @ }Exceptions @tab ZCX
24191 @item @code{@ @ }@i{rts-sjlj}
24192 @item @code{@ @ @ @ }Tasking @tab pthread library
24193 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24197 @node Specifying a Run-Time Library
24198 @section Specifying a Run-Time Library
24201 The @file{adainclude} subdirectory containing the sources of the GNAT
24202 run-time library, and the @file{adalib} subdirectory containing the
24203 @file{ALI} files and the static and/or shared GNAT library, are located
24204 in the gcc target-dependent area:
24207 target=$prefix/lib/gcc/gcc-@i{dumpmachine}/gcc-@i{dumpversion}/
24211 As indicated above, on some platforms several run-time libraries are supplied.
24212 These libraries are installed in the target dependent area and
24213 contain a complete source and binary subdirectory. The detailed description
24214 below explains the differences between the different libraries in terms of
24215 their thread support.
24217 The default run-time library (when GNAT is installed) is @emph{rts-native}.
24218 This default run time is selected by the means of soft links.
24219 For example on x86-linux:
24225 +--- adainclude----------+
24227 +--- adalib-----------+ |
24229 +--- rts-native | |
24231 | +--- adainclude <---+
24233 | +--- adalib <----+
24244 If the @i{rts-sjlj} library is to be selected on a permanent basis,
24245 these soft links can be modified with the following commands:
24249 $ rm -f adainclude adalib
24250 $ ln -s rts-sjlj/adainclude adainclude
24251 $ ln -s rts-sjlj/adalib adalib
24255 Alternatively, you can specify @file{rts-sjlj/adainclude} in the file
24256 @file{$target/ada_source_path} and @file{rts-sjlj/adalib} in
24257 @file{$target/ada_object_path}.
24259 Selecting another run-time library temporarily can be
24260 achieved by using the @option{--RTS} switch, e.g., @option{--RTS=sjlj}
24261 @cindex @option{--RTS} option
24263 @node Choosing the Scheduling Policy
24264 @section Choosing the Scheduling Policy
24267 When using a POSIX threads implementation, you have a choice of several
24268 scheduling policies: @code{SCHED_FIFO},
24269 @cindex @code{SCHED_FIFO} scheduling policy
24271 @cindex @code{SCHED_RR} scheduling policy
24272 and @code{SCHED_OTHER}.
24273 @cindex @code{SCHED_OTHER} scheduling policy
24274 Typically, the default is @code{SCHED_OTHER}, while using @code{SCHED_FIFO}
24275 or @code{SCHED_RR} requires special (e.g., root) privileges.
24277 By default, GNAT uses the @code{SCHED_OTHER} policy. To specify
24279 @cindex @code{SCHED_FIFO} scheduling policy
24280 you can use one of the following:
24284 @code{pragma Time_Slice (0.0)}
24285 @cindex pragma Time_Slice
24287 the corresponding binder option @option{-T0}
24288 @cindex @option{-T0} option
24290 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
24291 @cindex pragma Task_Dispatching_Policy
24295 To specify @code{SCHED_RR},
24296 @cindex @code{SCHED_RR} scheduling policy
24297 you should use @code{pragma Time_Slice} with a
24298 value greater than @code{0.0}, or else use the corresponding @option{-T}
24301 @node Solaris-Specific Considerations
24302 @section Solaris-Specific Considerations
24303 @cindex Solaris Sparc threads libraries
24306 This section addresses some topics related to the various threads libraries
24310 * Solaris Threads Issues::
24313 @node Solaris Threads Issues
24314 @subsection Solaris Threads Issues
24317 GNAT under Solaris/Sparc 32 bits comes with an alternate tasking run-time
24318 library based on POSIX threads --- @emph{rts-pthread}.
24319 @cindex rts-pthread threads library
24320 This run-time library has the advantage of being mostly shared across all
24321 POSIX-compliant thread implementations, and it also provides under
24322 @w{Solaris 8} the @code{PTHREAD_PRIO_INHERIT}
24323 @cindex @code{PTHREAD_PRIO_INHERIT} policy (under rts-pthread)
24324 and @code{PTHREAD_PRIO_PROTECT}
24325 @cindex @code{PTHREAD_PRIO_PROTECT} policy (under rts-pthread)
24326 semantics that can be selected using the predefined pragma
24327 @code{Locking_Policy}
24328 @cindex pragma Locking_Policy (under rts-pthread)
24330 @code{Inheritance_Locking} and @code{Ceiling_Locking} as the policy.
24331 @cindex @code{Inheritance_Locking} (under rts-pthread)
24332 @cindex @code{Ceiling_Locking} (under rts-pthread)
24334 As explained above, the native run-time library is based on the Solaris thread
24335 library (@code{libthread}) and is the default library.
24337 When the Solaris threads library is used (this is the default), programs
24338 compiled with GNAT can automatically take advantage of
24339 and can thus execute on multiple processors.
24340 The user can alternatively specify a processor on which the program should run
24341 to emulate a single-processor system. The multiprocessor / uniprocessor choice
24343 setting the environment variable @env{GNAT_PROCESSOR}
24344 @cindex @env{GNAT_PROCESSOR} environment variable (on Sparc Solaris)
24345 to one of the following:
24349 Use the default configuration (run the program on all
24350 available processors) - this is the same as having @code{GNAT_PROCESSOR}
24354 Let the run-time implementation choose one processor and run the program on
24357 @item 0 .. Last_Proc
24358 Run the program on the specified processor.
24359 @code{Last_Proc} is equal to @code{_SC_NPROCESSORS_CONF - 1}
24360 (where @code{_SC_NPROCESSORS_CONF} is a system variable).
24363 @node Linux-Specific Considerations
24364 @section Linux-Specific Considerations
24365 @cindex Linux threads libraries
24368 On GNU/Linux without NPTL support (usually system with GNU C Library
24369 older than 2.3), the signal model is not POSIX compliant, which means
24370 that to send a signal to the process, you need to send the signal to all
24371 threads, e.g.@: by using @code{killpg()}.
24373 @node AIX-Specific Considerations
24374 @section AIX-Specific Considerations
24375 @cindex AIX resolver library
24378 On AIX, the resolver library initializes some internal structure on
24379 the first call to @code{get*by*} functions, which are used to implement
24380 @code{GNAT.Sockets.Get_Host_By_Name} and
24381 @code{GNAT.Sockets.Get_Host_By_Address}.
24382 If such initialization occurs within an Ada task, and the stack size for
24383 the task is the default size, a stack overflow may occur.
24385 To avoid this overflow, the user should either ensure that the first call
24386 to @code{GNAT.Sockets.Get_Host_By_Name} or
24387 @code{GNAT.Sockets.Get_Host_By_Addrss}
24388 occurs in the environment task, or use @code{pragma Storage_Size} to
24389 specify a sufficiently large size for the stack of the task that contains
24392 @node RTX-Specific Considerations
24393 @section RTX-Specific Considerations
24394 @cindex RTX libraries
24397 The Real-time Extension (RTX) to Windows is based on the Windows Win32
24398 API. Applications can be built to work in two different modes:
24402 Windows executables that run in Ring 3 to utilize memory protection
24403 (@emph{rts-rtx-w32}).
24406 Real-time subsystem (RTSS) executables that run in Ring 0, where
24407 performance can be optimized with RTSS applications taking precedent
24408 over all Windows applications (@emph{rts-rtx-rtss}). This mode requires
24409 the Microsoft linker to handle RTSS libraries.
24413 @node HP-UX-Specific Considerations
24414 @section HP-UX-Specific Considerations
24415 @cindex HP-UX Scheduling
24418 On HP-UX, appropriate privileges are required to change the scheduling
24419 parameters of a task. The calling process must have appropriate
24420 privileges or be a member of a group having @code{PRIV_RTSCHED} access to
24421 successfully change the scheduling parameters.
24423 By default, GNAT uses the @code{SCHED_HPUX} policy. To have access to the
24424 priority range 0-31 either the @code{FIFO_Within_Priorities} or the
24425 @code{Round_Robin_Within_Priorities} scheduling policies need to be set.
24427 To specify the @code{FIFO_Within_Priorities} scheduling policy you can use
24428 one of the following:
24432 @code{pragma Time_Slice (0.0)}
24433 @cindex pragma Time_Slice
24435 the corresponding binder option @option{-T0}
24436 @cindex @option{-T0} option
24438 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
24439 @cindex pragma Task_Dispatching_Policy
24443 To specify the @code{Round_Robin_Within_Priorities}, scheduling policy
24444 you should use @code{pragma Time_Slice} with a
24445 value greater than @code{0.0}, or use the corresponding @option{-T}
24446 binder option, or set the @code{pragma Task_Dispatching_Policy
24447 (Round_Robin_Within_Priorities)}.
24449 @c *******************************
24450 @node Example of Binder Output File
24451 @appendix Example of Binder Output File
24454 This Appendix displays the source code for @command{gnatbind}'s output
24455 file generated for a simple ``Hello World'' program.
24456 Comments have been added for clarification purposes.
24458 @smallexample @c adanocomment
24462 -- The package is called Ada_Main unless this name is actually used
24463 -- as a unit name in the partition, in which case some other unique
24467 package ada_main is
24469 Elab_Final_Code : Integer;
24470 pragma Import (C, Elab_Final_Code, "__gnat_inside_elab_final_code");
24472 -- The main program saves the parameters (argument count,
24473 -- argument values, environment pointer) in global variables
24474 -- for later access by other units including
24475 -- Ada.Command_Line.
24477 gnat_argc : Integer;
24478 gnat_argv : System.Address;
24479 gnat_envp : System.Address;
24481 -- The actual variables are stored in a library routine. This
24482 -- is useful for some shared library situations, where there
24483 -- are problems if variables are not in the library.
24485 pragma Import (C, gnat_argc);
24486 pragma Import (C, gnat_argv);
24487 pragma Import (C, gnat_envp);
24489 -- The exit status is similarly an external location
24491 gnat_exit_status : Integer;
24492 pragma Import (C, gnat_exit_status);
24494 GNAT_Version : constant String :=
24495 "GNAT Version: 6.0.0w (20061115)";
24496 pragma Export (C, GNAT_Version, "__gnat_version");
24498 -- This is the generated adafinal routine that performs
24499 -- finalization at the end of execution. In the case where
24500 -- Ada is the main program, this main program makes a call
24501 -- to adafinal at program termination.
24503 procedure adafinal;
24504 pragma Export (C, adafinal, "adafinal");
24506 -- This is the generated adainit routine that performs
24507 -- initialization at the start of execution. In the case
24508 -- where Ada is the main program, this main program makes
24509 -- a call to adainit at program startup.
24512 pragma Export (C, adainit, "adainit");
24514 -- This routine is called at the start of execution. It is
24515 -- a dummy routine that is used by the debugger to breakpoint
24516 -- at the start of execution.
24518 procedure Break_Start;
24519 pragma Import (C, Break_Start, "__gnat_break_start");
24521 -- This is the actual generated main program (it would be
24522 -- suppressed if the no main program switch were used). As
24523 -- required by standard system conventions, this program has
24524 -- the external name main.
24528 argv : System.Address;
24529 envp : System.Address)
24531 pragma Export (C, main, "main");
24533 -- The following set of constants give the version
24534 -- identification values for every unit in the bound
24535 -- partition. This identification is computed from all
24536 -- dependent semantic units, and corresponds to the
24537 -- string that would be returned by use of the
24538 -- Body_Version or Version attributes.
24540 type Version_32 is mod 2 ** 32;
24541 u00001 : constant Version_32 := 16#7880BEB3#;
24542 u00002 : constant Version_32 := 16#0D24CBD0#;
24543 u00003 : constant Version_32 := 16#3283DBEB#;
24544 u00004 : constant Version_32 := 16#2359F9ED#;
24545 u00005 : constant Version_32 := 16#664FB847#;
24546 u00006 : constant Version_32 := 16#68E803DF#;
24547 u00007 : constant Version_32 := 16#5572E604#;
24548 u00008 : constant Version_32 := 16#46B173D8#;
24549 u00009 : constant Version_32 := 16#156A40CF#;
24550 u00010 : constant Version_32 := 16#033DABE0#;
24551 u00011 : constant Version_32 := 16#6AB38FEA#;
24552 u00012 : constant Version_32 := 16#22B6217D#;
24553 u00013 : constant Version_32 := 16#68A22947#;
24554 u00014 : constant Version_32 := 16#18CC4A56#;
24555 u00015 : constant Version_32 := 16#08258E1B#;
24556 u00016 : constant Version_32 := 16#367D5222#;
24557 u00017 : constant Version_32 := 16#20C9ECA4#;
24558 u00018 : constant Version_32 := 16#50D32CB6#;
24559 u00019 : constant Version_32 := 16#39A8BB77#;
24560 u00020 : constant Version_32 := 16#5CF8FA2B#;
24561 u00021 : constant Version_32 := 16#2F1EB794#;
24562 u00022 : constant Version_32 := 16#31AB6444#;
24563 u00023 : constant Version_32 := 16#1574B6E9#;
24564 u00024 : constant Version_32 := 16#5109C189#;
24565 u00025 : constant Version_32 := 16#56D770CD#;
24566 u00026 : constant Version_32 := 16#02F9DE3D#;
24567 u00027 : constant Version_32 := 16#08AB6B2C#;
24568 u00028 : constant Version_32 := 16#3FA37670#;
24569 u00029 : constant Version_32 := 16#476457A0#;
24570 u00030 : constant Version_32 := 16#731E1B6E#;
24571 u00031 : constant Version_32 := 16#23C2E789#;
24572 u00032 : constant Version_32 := 16#0F1BD6A1#;
24573 u00033 : constant Version_32 := 16#7C25DE96#;
24574 u00034 : constant Version_32 := 16#39ADFFA2#;
24575 u00035 : constant Version_32 := 16#571DE3E7#;
24576 u00036 : constant Version_32 := 16#5EB646AB#;
24577 u00037 : constant Version_32 := 16#4249379B#;
24578 u00038 : constant Version_32 := 16#0357E00A#;
24579 u00039 : constant Version_32 := 16#3784FB72#;
24580 u00040 : constant Version_32 := 16#2E723019#;
24581 u00041 : constant Version_32 := 16#623358EA#;
24582 u00042 : constant Version_32 := 16#107F9465#;
24583 u00043 : constant Version_32 := 16#6843F68A#;
24584 u00044 : constant Version_32 := 16#63305874#;
24585 u00045 : constant Version_32 := 16#31E56CE1#;
24586 u00046 : constant Version_32 := 16#02917970#;
24587 u00047 : constant Version_32 := 16#6CCBA70E#;
24588 u00048 : constant Version_32 := 16#41CD4204#;
24589 u00049 : constant Version_32 := 16#572E3F58#;
24590 u00050 : constant Version_32 := 16#20729FF5#;
24591 u00051 : constant Version_32 := 16#1D4F93E8#;
24592 u00052 : constant Version_32 := 16#30B2EC3D#;
24593 u00053 : constant Version_32 := 16#34054F96#;
24594 u00054 : constant Version_32 := 16#5A199860#;
24595 u00055 : constant Version_32 := 16#0E7F912B#;
24596 u00056 : constant Version_32 := 16#5760634A#;
24597 u00057 : constant Version_32 := 16#5D851835#;
24599 -- The following Export pragmas export the version numbers
24600 -- with symbolic names ending in B (for body) or S
24601 -- (for spec) so that they can be located in a link. The
24602 -- information provided here is sufficient to track down
24603 -- the exact versions of units used in a given build.
24605 pragma Export (C, u00001, "helloB");
24606 pragma Export (C, u00002, "system__standard_libraryB");
24607 pragma Export (C, u00003, "system__standard_libraryS");
24608 pragma Export (C, u00004, "adaS");
24609 pragma Export (C, u00005, "ada__text_ioB");
24610 pragma Export (C, u00006, "ada__text_ioS");
24611 pragma Export (C, u00007, "ada__exceptionsB");
24612 pragma Export (C, u00008, "ada__exceptionsS");
24613 pragma Export (C, u00009, "gnatS");
24614 pragma Export (C, u00010, "gnat__heap_sort_aB");
24615 pragma Export (C, u00011, "gnat__heap_sort_aS");
24616 pragma Export (C, u00012, "systemS");
24617 pragma Export (C, u00013, "system__exception_tableB");
24618 pragma Export (C, u00014, "system__exception_tableS");
24619 pragma Export (C, u00015, "gnat__htableB");
24620 pragma Export (C, u00016, "gnat__htableS");
24621 pragma Export (C, u00017, "system__exceptionsS");
24622 pragma Export (C, u00018, "system__machine_state_operationsB");
24623 pragma Export (C, u00019, "system__machine_state_operationsS");
24624 pragma Export (C, u00020, "system__machine_codeS");
24625 pragma Export (C, u00021, "system__storage_elementsB");
24626 pragma Export (C, u00022, "system__storage_elementsS");
24627 pragma Export (C, u00023, "system__secondary_stackB");
24628 pragma Export (C, u00024, "system__secondary_stackS");
24629 pragma Export (C, u00025, "system__parametersB");
24630 pragma Export (C, u00026, "system__parametersS");
24631 pragma Export (C, u00027, "system__soft_linksB");
24632 pragma Export (C, u00028, "system__soft_linksS");
24633 pragma Export (C, u00029, "system__stack_checkingB");
24634 pragma Export (C, u00030, "system__stack_checkingS");
24635 pragma Export (C, u00031, "system__tracebackB");
24636 pragma Export (C, u00032, "system__tracebackS");
24637 pragma Export (C, u00033, "ada__streamsS");
24638 pragma Export (C, u00034, "ada__tagsB");
24639 pragma Export (C, u00035, "ada__tagsS");
24640 pragma Export (C, u00036, "system__string_opsB");
24641 pragma Export (C, u00037, "system__string_opsS");
24642 pragma Export (C, u00038, "interfacesS");
24643 pragma Export (C, u00039, "interfaces__c_streamsB");
24644 pragma Export (C, u00040, "interfaces__c_streamsS");
24645 pragma Export (C, u00041, "system__file_ioB");
24646 pragma Export (C, u00042, "system__file_ioS");
24647 pragma Export (C, u00043, "ada__finalizationB");
24648 pragma Export (C, u00044, "ada__finalizationS");
24649 pragma Export (C, u00045, "system__finalization_rootB");
24650 pragma Export (C, u00046, "system__finalization_rootS");
24651 pragma Export (C, u00047, "system__finalization_implementationB");
24652 pragma Export (C, u00048, "system__finalization_implementationS");
24653 pragma Export (C, u00049, "system__string_ops_concat_3B");
24654 pragma Export (C, u00050, "system__string_ops_concat_3S");
24655 pragma Export (C, u00051, "system__stream_attributesB");
24656 pragma Export (C, u00052, "system__stream_attributesS");
24657 pragma Export (C, u00053, "ada__io_exceptionsS");
24658 pragma Export (C, u00054, "system__unsigned_typesS");
24659 pragma Export (C, u00055, "system__file_control_blockS");
24660 pragma Export (C, u00056, "ada__finalization__list_controllerB");
24661 pragma Export (C, u00057, "ada__finalization__list_controllerS");
24663 -- BEGIN ELABORATION ORDER
24666 -- gnat.heap_sort_a (spec)
24667 -- gnat.heap_sort_a (body)
24668 -- gnat.htable (spec)
24669 -- gnat.htable (body)
24670 -- interfaces (spec)
24672 -- system.machine_code (spec)
24673 -- system.parameters (spec)
24674 -- system.parameters (body)
24675 -- interfaces.c_streams (spec)
24676 -- interfaces.c_streams (body)
24677 -- system.standard_library (spec)
24678 -- ada.exceptions (spec)
24679 -- system.exception_table (spec)
24680 -- system.exception_table (body)
24681 -- ada.io_exceptions (spec)
24682 -- system.exceptions (spec)
24683 -- system.storage_elements (spec)
24684 -- system.storage_elements (body)
24685 -- system.machine_state_operations (spec)
24686 -- system.machine_state_operations (body)
24687 -- system.secondary_stack (spec)
24688 -- system.stack_checking (spec)
24689 -- system.soft_links (spec)
24690 -- system.soft_links (body)
24691 -- system.stack_checking (body)
24692 -- system.secondary_stack (body)
24693 -- system.standard_library (body)
24694 -- system.string_ops (spec)
24695 -- system.string_ops (body)
24698 -- ada.streams (spec)
24699 -- system.finalization_root (spec)
24700 -- system.finalization_root (body)
24701 -- system.string_ops_concat_3 (spec)
24702 -- system.string_ops_concat_3 (body)
24703 -- system.traceback (spec)
24704 -- system.traceback (body)
24705 -- ada.exceptions (body)
24706 -- system.unsigned_types (spec)
24707 -- system.stream_attributes (spec)
24708 -- system.stream_attributes (body)
24709 -- system.finalization_implementation (spec)
24710 -- system.finalization_implementation (body)
24711 -- ada.finalization (spec)
24712 -- ada.finalization (body)
24713 -- ada.finalization.list_controller (spec)
24714 -- ada.finalization.list_controller (body)
24715 -- system.file_control_block (spec)
24716 -- system.file_io (spec)
24717 -- system.file_io (body)
24718 -- ada.text_io (spec)
24719 -- ada.text_io (body)
24721 -- END ELABORATION ORDER
24725 -- The following source file name pragmas allow the generated file
24726 -- names to be unique for different main programs. They are needed
24727 -- since the package name will always be Ada_Main.
24729 pragma Source_File_Name (ada_main, Spec_File_Name => "b~hello.ads");
24730 pragma Source_File_Name (ada_main, Body_File_Name => "b~hello.adb");
24732 -- Generated package body for Ada_Main starts here
24734 package body ada_main is
24736 -- The actual finalization is performed by calling the
24737 -- library routine in System.Standard_Library.Adafinal
24739 procedure Do_Finalize;
24740 pragma Import (C, Do_Finalize, "system__standard_library__adafinal");
24747 procedure adainit is
24749 -- These booleans are set to True once the associated unit has
24750 -- been elaborated. It is also used to avoid elaborating the
24751 -- same unit twice.
24754 pragma Import (Ada, E040, "interfaces__c_streams_E");
24757 pragma Import (Ada, E008, "ada__exceptions_E");
24760 pragma Import (Ada, E014, "system__exception_table_E");
24763 pragma Import (Ada, E053, "ada__io_exceptions_E");
24766 pragma Import (Ada, E017, "system__exceptions_E");
24769 pragma Import (Ada, E024, "system__secondary_stack_E");
24772 pragma Import (Ada, E030, "system__stack_checking_E");
24775 pragma Import (Ada, E028, "system__soft_links_E");
24778 pragma Import (Ada, E035, "ada__tags_E");
24781 pragma Import (Ada, E033, "ada__streams_E");
24784 pragma Import (Ada, E046, "system__finalization_root_E");
24787 pragma Import (Ada, E048, "system__finalization_implementation_E");
24790 pragma Import (Ada, E044, "ada__finalization_E");
24793 pragma Import (Ada, E057, "ada__finalization__list_controller_E");
24796 pragma Import (Ada, E055, "system__file_control_block_E");
24799 pragma Import (Ada, E042, "system__file_io_E");
24802 pragma Import (Ada, E006, "ada__text_io_E");
24804 -- Set_Globals is a library routine that stores away the
24805 -- value of the indicated set of global values in global
24806 -- variables within the library.
24808 procedure Set_Globals
24809 (Main_Priority : Integer;
24810 Time_Slice_Value : Integer;
24811 WC_Encoding : Character;
24812 Locking_Policy : Character;
24813 Queuing_Policy : Character;
24814 Task_Dispatching_Policy : Character;
24815 Adafinal : System.Address;
24816 Unreserve_All_Interrupts : Integer;
24817 Exception_Tracebacks : Integer);
24818 @findex __gnat_set_globals
24819 pragma Import (C, Set_Globals, "__gnat_set_globals");
24821 -- SDP_Table_Build is a library routine used to build the
24822 -- exception tables. See unit Ada.Exceptions in files
24823 -- a-except.ads/adb for full details of how zero cost
24824 -- exception handling works. This procedure, the call to
24825 -- it, and the two following tables are all omitted if the
24826 -- build is in longjmp/setjmp exception mode.
24828 @findex SDP_Table_Build
24829 @findex Zero Cost Exceptions
24830 procedure SDP_Table_Build
24831 (SDP_Addresses : System.Address;
24832 SDP_Count : Natural;
24833 Elab_Addresses : System.Address;
24834 Elab_Addr_Count : Natural);
24835 pragma Import (C, SDP_Table_Build, "__gnat_SDP_Table_Build");
24837 -- Table of Unit_Exception_Table addresses. Used for zero
24838 -- cost exception handling to build the top level table.
24840 ST : aliased constant array (1 .. 23) of System.Address := (
24842 Ada.Text_Io'UET_Address,
24843 Ada.Exceptions'UET_Address,
24844 Gnat.Heap_Sort_A'UET_Address,
24845 System.Exception_Table'UET_Address,
24846 System.Machine_State_Operations'UET_Address,
24847 System.Secondary_Stack'UET_Address,
24848 System.Parameters'UET_Address,
24849 System.Soft_Links'UET_Address,
24850 System.Stack_Checking'UET_Address,
24851 System.Traceback'UET_Address,
24852 Ada.Streams'UET_Address,
24853 Ada.Tags'UET_Address,
24854 System.String_Ops'UET_Address,
24855 Interfaces.C_Streams'UET_Address,
24856 System.File_Io'UET_Address,
24857 Ada.Finalization'UET_Address,
24858 System.Finalization_Root'UET_Address,
24859 System.Finalization_Implementation'UET_Address,
24860 System.String_Ops_Concat_3'UET_Address,
24861 System.Stream_Attributes'UET_Address,
24862 System.File_Control_Block'UET_Address,
24863 Ada.Finalization.List_Controller'UET_Address);
24865 -- Table of addresses of elaboration routines. Used for
24866 -- zero cost exception handling to make sure these
24867 -- addresses are included in the top level procedure
24870 EA : aliased constant array (1 .. 23) of System.Address := (
24871 adainit'Code_Address,
24872 Do_Finalize'Code_Address,
24873 Ada.Exceptions'Elab_Spec'Address,
24874 System.Exceptions'Elab_Spec'Address,
24875 Interfaces.C_Streams'Elab_Spec'Address,
24876 System.Exception_Table'Elab_Body'Address,
24877 Ada.Io_Exceptions'Elab_Spec'Address,
24878 System.Stack_Checking'Elab_Spec'Address,
24879 System.Soft_Links'Elab_Body'Address,
24880 System.Secondary_Stack'Elab_Body'Address,
24881 Ada.Tags'Elab_Spec'Address,
24882 Ada.Tags'Elab_Body'Address,
24883 Ada.Streams'Elab_Spec'Address,
24884 System.Finalization_Root'Elab_Spec'Address,
24885 Ada.Exceptions'Elab_Body'Address,
24886 System.Finalization_Implementation'Elab_Spec'Address,
24887 System.Finalization_Implementation'Elab_Body'Address,
24888 Ada.Finalization'Elab_Spec'Address,
24889 Ada.Finalization.List_Controller'Elab_Spec'Address,
24890 System.File_Control_Block'Elab_Spec'Address,
24891 System.File_Io'Elab_Body'Address,
24892 Ada.Text_Io'Elab_Spec'Address,
24893 Ada.Text_Io'Elab_Body'Address);
24895 -- Start of processing for adainit
24899 -- Call SDP_Table_Build to build the top level procedure
24900 -- table for zero cost exception handling (omitted in
24901 -- longjmp/setjmp mode).
24903 SDP_Table_Build (ST'Address, 23, EA'Address, 23);
24905 -- Call Set_Globals to record various information for
24906 -- this partition. The values are derived by the binder
24907 -- from information stored in the ali files by the compiler.
24909 @findex __gnat_set_globals
24911 (Main_Priority => -1,
24912 -- Priority of main program, -1 if no pragma Priority used
24914 Time_Slice_Value => -1,
24915 -- Time slice from Time_Slice pragma, -1 if none used
24917 WC_Encoding => 'b',
24918 -- Wide_Character encoding used, default is brackets
24920 Locking_Policy => ' ',
24921 -- Locking_Policy used, default of space means not
24922 -- specified, otherwise it is the first character of
24923 -- the policy name.
24925 Queuing_Policy => ' ',
24926 -- Queuing_Policy used, default of space means not
24927 -- specified, otherwise it is the first character of
24928 -- the policy name.
24930 Task_Dispatching_Policy => ' ',
24931 -- Task_Dispatching_Policy used, default of space means
24932 -- not specified, otherwise first character of the
24935 Adafinal => System.Null_Address,
24936 -- Address of Adafinal routine, not used anymore
24938 Unreserve_All_Interrupts => 0,
24939 -- Set true if pragma Unreserve_All_Interrupts was used
24941 Exception_Tracebacks => 0);
24942 -- Indicates if exception tracebacks are enabled
24944 Elab_Final_Code := 1;
24946 -- Now we have the elaboration calls for all units in the partition.
24947 -- The Elab_Spec and Elab_Body attributes generate references to the
24948 -- implicit elaboration procedures generated by the compiler for
24949 -- each unit that requires elaboration.
24952 Interfaces.C_Streams'Elab_Spec;
24956 Ada.Exceptions'Elab_Spec;
24959 System.Exception_Table'Elab_Body;
24963 Ada.Io_Exceptions'Elab_Spec;
24967 System.Exceptions'Elab_Spec;
24971 System.Stack_Checking'Elab_Spec;
24974 System.Soft_Links'Elab_Body;
24979 System.Secondary_Stack'Elab_Body;
24983 Ada.Tags'Elab_Spec;
24986 Ada.Tags'Elab_Body;
24990 Ada.Streams'Elab_Spec;
24994 System.Finalization_Root'Elab_Spec;
24998 Ada.Exceptions'Elab_Body;
25002 System.Finalization_Implementation'Elab_Spec;
25005 System.Finalization_Implementation'Elab_Body;
25009 Ada.Finalization'Elab_Spec;
25013 Ada.Finalization.List_Controller'Elab_Spec;
25017 System.File_Control_Block'Elab_Spec;
25021 System.File_Io'Elab_Body;
25025 Ada.Text_Io'Elab_Spec;
25028 Ada.Text_Io'Elab_Body;
25032 Elab_Final_Code := 0;
25040 procedure adafinal is
25049 -- main is actually a function, as in the ANSI C standard,
25050 -- defined to return the exit status. The three parameters
25051 -- are the argument count, argument values and environment
25054 @findex Main Program
25057 argv : System.Address;
25058 envp : System.Address)
25061 -- The initialize routine performs low level system
25062 -- initialization using a standard library routine which
25063 -- sets up signal handling and performs any other
25064 -- required setup. The routine can be found in file
25067 @findex __gnat_initialize
25068 procedure initialize;
25069 pragma Import (C, initialize, "__gnat_initialize");
25071 -- The finalize routine performs low level system
25072 -- finalization using a standard library routine. The
25073 -- routine is found in file a-final.c and in the standard
25074 -- distribution is a dummy routine that does nothing, so
25075 -- really this is a hook for special user finalization.
25077 @findex __gnat_finalize
25078 procedure finalize;
25079 pragma Import (C, finalize, "__gnat_finalize");
25081 -- We get to the main program of the partition by using
25082 -- pragma Import because if we try to with the unit and
25083 -- call it Ada style, then not only do we waste time
25084 -- recompiling it, but also, we don't really know the right
25085 -- switches (e.g.@: identifier character set) to be used
25088 procedure Ada_Main_Program;
25089 pragma Import (Ada, Ada_Main_Program, "_ada_hello");
25091 -- Start of processing for main
25094 -- Save global variables
25100 -- Call low level system initialization
25104 -- Call our generated Ada initialization routine
25108 -- This is the point at which we want the debugger to get
25113 -- Now we call the main program of the partition
25117 -- Perform Ada finalization
25121 -- Perform low level system finalization
25125 -- Return the proper exit status
25126 return (gnat_exit_status);
25129 -- This section is entirely comments, so it has no effect on the
25130 -- compilation of the Ada_Main package. It provides the list of
25131 -- object files and linker options, as well as some standard
25132 -- libraries needed for the link. The gnatlink utility parses
25133 -- this b~hello.adb file to read these comment lines to generate
25134 -- the appropriate command line arguments for the call to the
25135 -- system linker. The BEGIN/END lines are used for sentinels for
25136 -- this parsing operation.
25138 -- The exact file names will of course depend on the environment,
25139 -- host/target and location of files on the host system.
25141 @findex Object file list
25142 -- BEGIN Object file/option list
25145 -- -L/usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/
25146 -- /usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/libgnat.a
25147 -- END Object file/option list
25153 The Ada code in the above example is exactly what is generated by the
25154 binder. We have added comments to more clearly indicate the function
25155 of each part of the generated @code{Ada_Main} package.
25157 The code is standard Ada in all respects, and can be processed by any
25158 tools that handle Ada. In particular, it is possible to use the debugger
25159 in Ada mode to debug the generated @code{Ada_Main} package. For example,
25160 suppose that for reasons that you do not understand, your program is crashing
25161 during elaboration of the body of @code{Ada.Text_IO}. To locate this bug,
25162 you can place a breakpoint on the call:
25164 @smallexample @c ada
25165 Ada.Text_Io'Elab_Body;
25169 and trace the elaboration routine for this package to find out where
25170 the problem might be (more usually of course you would be debugging
25171 elaboration code in your own application).
25173 @node Elaboration Order Handling in GNAT
25174 @appendix Elaboration Order Handling in GNAT
25175 @cindex Order of elaboration
25176 @cindex Elaboration control
25179 * Elaboration Code::
25180 * Checking the Elaboration Order::
25181 * Controlling the Elaboration Order::
25182 * Controlling Elaboration in GNAT - Internal Calls::
25183 * Controlling Elaboration in GNAT - External Calls::
25184 * Default Behavior in GNAT - Ensuring Safety::
25185 * Treatment of Pragma Elaborate::
25186 * Elaboration Issues for Library Tasks::
25187 * Mixing Elaboration Models::
25188 * What to Do If the Default Elaboration Behavior Fails::
25189 * Elaboration for Indirect Calls::
25190 * Summary of Procedures for Elaboration Control::
25191 * Other Elaboration Order Considerations::
25192 * Determining the Chosen Elaboration Order::
25196 This chapter describes the handling of elaboration code in Ada and
25197 in GNAT, and discusses how the order of elaboration of program units can
25198 be controlled in GNAT, either automatically or with explicit programming
25201 @node Elaboration Code
25202 @section Elaboration Code
25205 Ada provides rather general mechanisms for executing code at elaboration
25206 time, that is to say before the main program starts executing. Such code arises
25210 @item Initializers for variables.
25211 Variables declared at the library level, in package specs or bodies, can
25212 require initialization that is performed at elaboration time, as in:
25213 @smallexample @c ada
25215 Sqrt_Half : Float := Sqrt (0.5);
25219 @item Package initialization code
25220 Code in a @code{BEGIN-END} section at the outer level of a package body is
25221 executed as part of the package body elaboration code.
25223 @item Library level task allocators
25224 Tasks that are declared using task allocators at the library level
25225 start executing immediately and hence can execute at elaboration time.
25229 Subprogram calls are possible in any of these contexts, which means that
25230 any arbitrary part of the program may be executed as part of the elaboration
25231 code. It is even possible to write a program which does all its work at
25232 elaboration time, with a null main program, although stylistically this
25233 would usually be considered an inappropriate way to structure
25236 An important concern arises in the context of elaboration code:
25237 we have to be sure that it is executed in an appropriate order. What we
25238 have is a series of elaboration code sections, potentially one section
25239 for each unit in the program. It is important that these execute
25240 in the correct order. Correctness here means that, taking the above
25241 example of the declaration of @code{Sqrt_Half},
25242 if some other piece of
25243 elaboration code references @code{Sqrt_Half},
25244 then it must run after the
25245 section of elaboration code that contains the declaration of
25248 There would never be any order of elaboration problem if we made a rule
25249 that whenever you @code{with} a unit, you must elaborate both the spec and body
25250 of that unit before elaborating the unit doing the @code{with}'ing:
25252 @smallexample @c ada
25256 package Unit_2 is @dots{}
25262 would require that both the body and spec of @code{Unit_1} be elaborated
25263 before the spec of @code{Unit_2}. However, a rule like that would be far too
25264 restrictive. In particular, it would make it impossible to have routines
25265 in separate packages that were mutually recursive.
25267 You might think that a clever enough compiler could look at the actual
25268 elaboration code and determine an appropriate correct order of elaboration,
25269 but in the general case, this is not possible. Consider the following
25272 In the body of @code{Unit_1}, we have a procedure @code{Func_1}
25274 the variable @code{Sqrt_1}, which is declared in the elaboration code
25275 of the body of @code{Unit_1}:
25277 @smallexample @c ada
25279 Sqrt_1 : Float := Sqrt (0.1);
25284 The elaboration code of the body of @code{Unit_1} also contains:
25286 @smallexample @c ada
25289 if expression_1 = 1 then
25290 Q := Unit_2.Func_2;
25297 @code{Unit_2} is exactly parallel,
25298 it has a procedure @code{Func_2} that references
25299 the variable @code{Sqrt_2}, which is declared in the elaboration code of
25300 the body @code{Unit_2}:
25302 @smallexample @c ada
25304 Sqrt_2 : Float := Sqrt (0.1);
25309 The elaboration code of the body of @code{Unit_2} also contains:
25311 @smallexample @c ada
25314 if expression_2 = 2 then
25315 Q := Unit_1.Func_1;
25322 Now the question is, which of the following orders of elaboration is
25347 If you carefully analyze the flow here, you will see that you cannot tell
25348 at compile time the answer to this question.
25349 If @code{expression_1} is not equal to 1,
25350 and @code{expression_2} is not equal to 2,
25351 then either order is acceptable, because neither of the function calls is
25352 executed. If both tests evaluate to true, then neither order is acceptable
25353 and in fact there is no correct order.
25355 If one of the two expressions is true, and the other is false, then one
25356 of the above orders is correct, and the other is incorrect. For example,
25357 if @code{expression_1} /= 1 and @code{expression_2} = 2,
25358 then the call to @code{Func_1}
25359 will occur, but not the call to @code{Func_2.}
25360 This means that it is essential
25361 to elaborate the body of @code{Unit_1} before
25362 the body of @code{Unit_2}, so the first
25363 order of elaboration is correct and the second is wrong.
25365 By making @code{expression_1} and @code{expression_2}
25366 depend on input data, or perhaps
25367 the time of day, we can make it impossible for the compiler or binder
25368 to figure out which of these expressions will be true, and hence it
25369 is impossible to guarantee a safe order of elaboration at run time.
25371 @node Checking the Elaboration Order
25372 @section Checking the Elaboration Order
25375 In some languages that involve the same kind of elaboration problems,
25376 e.g.@: Java and C++, the programmer is expected to worry about these
25377 ordering problems himself, and it is common to
25378 write a program in which an incorrect elaboration order gives
25379 surprising results, because it references variables before they
25381 Ada is designed to be a safe language, and a programmer-beware approach is
25382 clearly not sufficient. Consequently, the language provides three lines
25386 @item Standard rules
25387 Some standard rules restrict the possible choice of elaboration
25388 order. In particular, if you @code{with} a unit, then its spec is always
25389 elaborated before the unit doing the @code{with}. Similarly, a parent
25390 spec is always elaborated before the child spec, and finally
25391 a spec is always elaborated before its corresponding body.
25393 @item Dynamic elaboration checks
25394 @cindex Elaboration checks
25395 @cindex Checks, elaboration
25396 Dynamic checks are made at run time, so that if some entity is accessed
25397 before it is elaborated (typically by means of a subprogram call)
25398 then the exception (@code{Program_Error}) is raised.
25400 @item Elaboration control
25401 Facilities are provided for the programmer to specify the desired order
25405 Let's look at these facilities in more detail. First, the rules for
25406 dynamic checking. One possible rule would be simply to say that the
25407 exception is raised if you access a variable which has not yet been
25408 elaborated. The trouble with this approach is that it could require
25409 expensive checks on every variable reference. Instead Ada has two
25410 rules which are a little more restrictive, but easier to check, and
25414 @item Restrictions on calls
25415 A subprogram can only be called at elaboration time if its body
25416 has been elaborated. The rules for elaboration given above guarantee
25417 that the spec of the subprogram has been elaborated before the
25418 call, but not the body. If this rule is violated, then the
25419 exception @code{Program_Error} is raised.
25421 @item Restrictions on instantiations
25422 A generic unit can only be instantiated if the body of the generic
25423 unit has been elaborated. Again, the rules for elaboration given above
25424 guarantee that the spec of the generic unit has been elaborated
25425 before the instantiation, but not the body. If this rule is
25426 violated, then the exception @code{Program_Error} is raised.
25430 The idea is that if the body has been elaborated, then any variables
25431 it references must have been elaborated; by checking for the body being
25432 elaborated we guarantee that none of its references causes any
25433 trouble. As we noted above, this is a little too restrictive, because a
25434 subprogram that has no non-local references in its body may in fact be safe
25435 to call. However, it really would be unsafe to rely on this, because
25436 it would mean that the caller was aware of details of the implementation
25437 in the body. This goes against the basic tenets of Ada.
25439 A plausible implementation can be described as follows.
25440 A Boolean variable is associated with each subprogram
25441 and each generic unit. This variable is initialized to False, and is set to
25442 True at the point body is elaborated. Every call or instantiation checks the
25443 variable, and raises @code{Program_Error} if the variable is False.
25445 Note that one might think that it would be good enough to have one Boolean
25446 variable for each package, but that would not deal with cases of trying
25447 to call a body in the same package as the call
25448 that has not been elaborated yet.
25449 Of course a compiler may be able to do enough analysis to optimize away
25450 some of the Boolean variables as unnecessary, and @code{GNAT} indeed
25451 does such optimizations, but still the easiest conceptual model is to
25452 think of there being one variable per subprogram.
25454 @node Controlling the Elaboration Order
25455 @section Controlling the Elaboration Order
25458 In the previous section we discussed the rules in Ada which ensure
25459 that @code{Program_Error} is raised if an incorrect elaboration order is
25460 chosen. This prevents erroneous executions, but we need mechanisms to
25461 specify a correct execution and avoid the exception altogether.
25462 To achieve this, Ada provides a number of features for controlling
25463 the order of elaboration. We discuss these features in this section.
25465 First, there are several ways of indicating to the compiler that a given
25466 unit has no elaboration problems:
25469 @item packages that do not require a body
25470 A library package that does not require a body does not permit
25471 a body (this rule was introduced in Ada 95).
25472 Thus if we have a such a package, as in:
25474 @smallexample @c ada
25477 package Definitions is
25479 type m is new integer;
25481 type a is array (1 .. 10) of m;
25482 type b is array (1 .. 20) of m;
25490 A package that @code{with}'s @code{Definitions} may safely instantiate
25491 @code{Definitions.Subp} because the compiler can determine that there
25492 definitely is no package body to worry about in this case
25495 @cindex pragma Pure
25497 Places sufficient restrictions on a unit to guarantee that
25498 no call to any subprogram in the unit can result in an
25499 elaboration problem. This means that the compiler does not need
25500 to worry about the point of elaboration of such units, and in
25501 particular, does not need to check any calls to any subprograms
25504 @item pragma Preelaborate
25505 @findex Preelaborate
25506 @cindex pragma Preelaborate
25507 This pragma places slightly less stringent restrictions on a unit than
25509 but these restrictions are still sufficient to ensure that there
25510 are no elaboration problems with any calls to the unit.
25512 @item pragma Elaborate_Body
25513 @findex Elaborate_Body
25514 @cindex pragma Elaborate_Body
25515 This pragma requires that the body of a unit be elaborated immediately
25516 after its spec. Suppose a unit @code{A} has such a pragma,
25517 and unit @code{B} does
25518 a @code{with} of unit @code{A}. Recall that the standard rules require
25519 the spec of unit @code{A}
25520 to be elaborated before the @code{with}'ing unit; given the pragma in
25521 @code{A}, we also know that the body of @code{A}
25522 will be elaborated before @code{B}, so
25523 that calls to @code{A} are safe and do not need a check.
25528 unlike pragma @code{Pure} and pragma @code{Preelaborate},
25530 @code{Elaborate_Body} does not guarantee that the program is
25531 free of elaboration problems, because it may not be possible
25532 to satisfy the requested elaboration order.
25533 Let's go back to the example with @code{Unit_1} and @code{Unit_2}.
25535 marks @code{Unit_1} as @code{Elaborate_Body},
25536 and not @code{Unit_2,} then the order of
25537 elaboration will be:
25549 Now that means that the call to @code{Func_1} in @code{Unit_2}
25550 need not be checked,
25551 it must be safe. But the call to @code{Func_2} in
25552 @code{Unit_1} may still fail if
25553 @code{Expression_1} is equal to 1,
25554 and the programmer must still take
25555 responsibility for this not being the case.
25557 If all units carry a pragma @code{Elaborate_Body}, then all problems are
25558 eliminated, except for calls entirely within a body, which are
25559 in any case fully under programmer control. However, using the pragma
25560 everywhere is not always possible.
25561 In particular, for our @code{Unit_1}/@code{Unit_2} example, if
25562 we marked both of them as having pragma @code{Elaborate_Body}, then
25563 clearly there would be no possible elaboration order.
25565 The above pragmas allow a server to guarantee safe use by clients, and
25566 clearly this is the preferable approach. Consequently a good rule
25567 is to mark units as @code{Pure} or @code{Preelaborate} if possible,
25568 and if this is not possible,
25569 mark them as @code{Elaborate_Body} if possible.
25570 As we have seen, there are situations where neither of these
25571 three pragmas can be used.
25572 So we also provide methods for clients to control the
25573 order of elaboration of the servers on which they depend:
25576 @item pragma Elaborate (unit)
25578 @cindex pragma Elaborate
25579 This pragma is placed in the context clause, after a @code{with} clause,
25580 and it requires that the body of the named unit be elaborated before
25581 the unit in which the pragma occurs. The idea is to use this pragma
25582 if the current unit calls at elaboration time, directly or indirectly,
25583 some subprogram in the named unit.
25585 @item pragma Elaborate_All (unit)
25586 @findex Elaborate_All
25587 @cindex pragma Elaborate_All
25588 This is a stronger version of the Elaborate pragma. Consider the
25592 Unit A @code{with}'s unit B and calls B.Func in elab code
25593 Unit B @code{with}'s unit C, and B.Func calls C.Func
25597 Now if we put a pragma @code{Elaborate (B)}
25598 in unit @code{A}, this ensures that the
25599 body of @code{B} is elaborated before the call, but not the
25600 body of @code{C}, so
25601 the call to @code{C.Func} could still cause @code{Program_Error} to
25604 The effect of a pragma @code{Elaborate_All} is stronger, it requires
25605 not only that the body of the named unit be elaborated before the
25606 unit doing the @code{with}, but also the bodies of all units that the
25607 named unit uses, following @code{with} links transitively. For example,
25608 if we put a pragma @code{Elaborate_All (B)} in unit @code{A},
25610 not only that the body of @code{B} be elaborated before @code{A},
25612 body of @code{C}, because @code{B} @code{with}'s @code{C}.
25616 We are now in a position to give a usage rule in Ada for avoiding
25617 elaboration problems, at least if dynamic dispatching and access to
25618 subprogram values are not used. We will handle these cases separately
25621 The rule is simple. If a unit has elaboration code that can directly or
25622 indirectly make a call to a subprogram in a @code{with}'ed unit, or instantiate
25623 a generic package in a @code{with}'ed unit,
25624 then if the @code{with}'ed unit does not have
25625 pragma @code{Pure} or @code{Preelaborate}, then the client should have
25626 a pragma @code{Elaborate_All}
25627 for the @code{with}'ed unit. By following this rule a client is
25628 assured that calls can be made without risk of an exception.
25630 For generic subprogram instantiations, the rule can be relaxed to
25631 require only a pragma @code{Elaborate} since elaborating the body
25632 of a subprogram cannot cause any transitive elaboration (we are
25633 not calling the subprogram in this case, just elaborating its
25636 If this rule is not followed, then a program may be in one of four
25640 @item No order exists
25641 No order of elaboration exists which follows the rules, taking into
25642 account any @code{Elaborate}, @code{Elaborate_All},
25643 or @code{Elaborate_Body} pragmas. In
25644 this case, an Ada compiler must diagnose the situation at bind
25645 time, and refuse to build an executable program.
25647 @item One or more orders exist, all incorrect
25648 One or more acceptable elaboration orders exist, and all of them
25649 generate an elaboration order problem. In this case, the binder
25650 can build an executable program, but @code{Program_Error} will be raised
25651 when the program is run.
25653 @item Several orders exist, some right, some incorrect
25654 One or more acceptable elaboration orders exists, and some of them
25655 work, and some do not. The programmer has not controlled
25656 the order of elaboration, so the binder may or may not pick one of
25657 the correct orders, and the program may or may not raise an
25658 exception when it is run. This is the worst case, because it means
25659 that the program may fail when moved to another compiler, or even
25660 another version of the same compiler.
25662 @item One or more orders exists, all correct
25663 One ore more acceptable elaboration orders exist, and all of them
25664 work. In this case the program runs successfully. This state of
25665 affairs can be guaranteed by following the rule we gave above, but
25666 may be true even if the rule is not followed.
25670 Note that one additional advantage of following our rules on the use
25671 of @code{Elaborate} and @code{Elaborate_All}
25672 is that the program continues to stay in the ideal (all orders OK) state
25673 even if maintenance
25674 changes some bodies of some units. Conversely, if a program that does
25675 not follow this rule happens to be safe at some point, this state of affairs
25676 may deteriorate silently as a result of maintenance changes.
25678 You may have noticed that the above discussion did not mention
25679 the use of @code{Elaborate_Body}. This was a deliberate omission. If you
25680 @code{with} an @code{Elaborate_Body} unit, it still may be the case that
25681 code in the body makes calls to some other unit, so it is still necessary
25682 to use @code{Elaborate_All} on such units.
25684 @node Controlling Elaboration in GNAT - Internal Calls
25685 @section Controlling Elaboration in GNAT - Internal Calls
25688 In the case of internal calls, i.e., calls within a single package, the
25689 programmer has full control over the order of elaboration, and it is up
25690 to the programmer to elaborate declarations in an appropriate order. For
25693 @smallexample @c ada
25696 function One return Float;
25700 function One return Float is
25709 will obviously raise @code{Program_Error} at run time, because function
25710 One will be called before its body is elaborated. In this case GNAT will
25711 generate a warning that the call will raise @code{Program_Error}:
25717 2. function One return Float;
25719 4. Q : Float := One;
25721 >>> warning: cannot call "One" before body is elaborated
25722 >>> warning: Program_Error will be raised at run time
25725 6. function One return Float is
25738 Note that in this particular case, it is likely that the call is safe, because
25739 the function @code{One} does not access any global variables.
25740 Nevertheless in Ada, we do not want the validity of the check to depend on
25741 the contents of the body (think about the separate compilation case), so this
25742 is still wrong, as we discussed in the previous sections.
25744 The error is easily corrected by rearranging the declarations so that the
25745 body of @code{One} appears before the declaration containing the call
25746 (note that in Ada 95 and Ada 2005,
25747 declarations can appear in any order, so there is no restriction that
25748 would prevent this reordering, and if we write:
25750 @smallexample @c ada
25753 function One return Float;
25755 function One return Float is
25766 then all is well, no warning is generated, and no
25767 @code{Program_Error} exception
25769 Things are more complicated when a chain of subprograms is executed:
25771 @smallexample @c ada
25774 function A return Integer;
25775 function B return Integer;
25776 function C return Integer;
25778 function B return Integer is begin return A; end;
25779 function C return Integer is begin return B; end;
25783 function A return Integer is begin return 1; end;
25789 Now the call to @code{C}
25790 at elaboration time in the declaration of @code{X} is correct, because
25791 the body of @code{C} is already elaborated,
25792 and the call to @code{B} within the body of
25793 @code{C} is correct, but the call
25794 to @code{A} within the body of @code{B} is incorrect, because the body
25795 of @code{A} has not been elaborated, so @code{Program_Error}
25796 will be raised on the call to @code{A}.
25797 In this case GNAT will generate a
25798 warning that @code{Program_Error} may be
25799 raised at the point of the call. Let's look at the warning:
25805 2. function A return Integer;
25806 3. function B return Integer;
25807 4. function C return Integer;
25809 6. function B return Integer is begin return A; end;
25811 >>> warning: call to "A" before body is elaborated may
25812 raise Program_Error
25813 >>> warning: "B" called at line 7
25814 >>> warning: "C" called at line 9
25816 7. function C return Integer is begin return B; end;
25818 9. X : Integer := C;
25820 11. function A return Integer is begin return 1; end;
25830 Note that the message here says ``may raise'', instead of the direct case,
25831 where the message says ``will be raised''. That's because whether
25833 actually called depends in general on run-time flow of control.
25834 For example, if the body of @code{B} said
25836 @smallexample @c ada
25839 function B return Integer is
25841 if some-condition-depending-on-input-data then
25852 then we could not know until run time whether the incorrect call to A would
25853 actually occur, so @code{Program_Error} might
25854 or might not be raised. It is possible for a compiler to
25855 do a better job of analyzing bodies, to
25856 determine whether or not @code{Program_Error}
25857 might be raised, but it certainly
25858 couldn't do a perfect job (that would require solving the halting problem
25859 and is provably impossible), and because this is a warning anyway, it does
25860 not seem worth the effort to do the analysis. Cases in which it
25861 would be relevant are rare.
25863 In practice, warnings of either of the forms given
25864 above will usually correspond to
25865 real errors, and should be examined carefully and eliminated.
25866 In the rare case where a warning is bogus, it can be suppressed by any of
25867 the following methods:
25871 Compile with the @option{-gnatws} switch set
25874 Suppress @code{Elaboration_Check} for the called subprogram
25877 Use pragma @code{Warnings_Off} to turn warnings off for the call
25881 For the internal elaboration check case,
25882 GNAT by default generates the
25883 necessary run-time checks to ensure
25884 that @code{Program_Error} is raised if any
25885 call fails an elaboration check. Of course this can only happen if a
25886 warning has been issued as described above. The use of pragma
25887 @code{Suppress (Elaboration_Check)} may (but is not guaranteed to) suppress
25888 some of these checks, meaning that it may be possible (but is not
25889 guaranteed) for a program to be able to call a subprogram whose body
25890 is not yet elaborated, without raising a @code{Program_Error} exception.
25892 @node Controlling Elaboration in GNAT - External Calls
25893 @section Controlling Elaboration in GNAT - External Calls
25896 The previous section discussed the case in which the execution of a
25897 particular thread of elaboration code occurred entirely within a
25898 single unit. This is the easy case to handle, because a programmer
25899 has direct and total control over the order of elaboration, and
25900 furthermore, checks need only be generated in cases which are rare
25901 and which the compiler can easily detect.
25902 The situation is more complex when separate compilation is taken into account.
25903 Consider the following:
25905 @smallexample @c ada
25909 function Sqrt (Arg : Float) return Float;
25912 package body Math is
25913 function Sqrt (Arg : Float) return Float is
25922 X : Float := Math.Sqrt (0.5);
25935 where @code{Main} is the main program. When this program is executed, the
25936 elaboration code must first be executed, and one of the jobs of the
25937 binder is to determine the order in which the units of a program are
25938 to be elaborated. In this case we have four units: the spec and body
25940 the spec of @code{Stuff} and the body of @code{Main}).
25941 In what order should the four separate sections of elaboration code
25944 There are some restrictions in the order of elaboration that the binder
25945 can choose. In particular, if unit U has a @code{with}
25946 for a package @code{X}, then you
25947 are assured that the spec of @code{X}
25948 is elaborated before U , but you are
25949 not assured that the body of @code{X}
25950 is elaborated before U.
25951 This means that in the above case, the binder is allowed to choose the
25962 but that's not good, because now the call to @code{Math.Sqrt}
25963 that happens during
25964 the elaboration of the @code{Stuff}
25965 spec happens before the body of @code{Math.Sqrt} is
25966 elaborated, and hence causes @code{Program_Error} exception to be raised.
25967 At first glance, one might say that the binder is misbehaving, because
25968 obviously you want to elaborate the body of something you @code{with}
25970 that is not a general rule that can be followed in all cases. Consider
25972 @smallexample @c ada
25975 package X is @dots{}
25977 package Y is @dots{}
25980 package body Y is @dots{}
25983 package body X is @dots{}
25989 This is a common arrangement, and, apart from the order of elaboration
25990 problems that might arise in connection with elaboration code, this works fine.
25991 A rule that says that you must first elaborate the body of anything you
25992 @code{with} cannot work in this case:
25993 the body of @code{X} @code{with}'s @code{Y},
25994 which means you would have to
25995 elaborate the body of @code{Y} first, but that @code{with}'s @code{X},
25997 you have to elaborate the body of @code{X} first, but @dots{} and we have a
25998 loop that cannot be broken.
26000 It is true that the binder can in many cases guess an order of elaboration
26001 that is unlikely to cause a @code{Program_Error}
26002 exception to be raised, and it tries to do so (in the
26003 above example of @code{Math/Stuff/Spec}, the GNAT binder will
26005 elaborate the body of @code{Math} right after its spec, so all will be well).
26007 However, a program that blindly relies on the binder to be helpful can
26008 get into trouble, as we discussed in the previous sections, so
26010 provides a number of facilities for assisting the programmer in
26011 developing programs that are robust with respect to elaboration order.
26013 @node Default Behavior in GNAT - Ensuring Safety
26014 @section Default Behavior in GNAT - Ensuring Safety
26017 The default behavior in GNAT ensures elaboration safety. In its
26018 default mode GNAT implements the
26019 rule we previously described as the right approach. Let's restate it:
26023 @emph{If a unit has elaboration code that can directly or indirectly make a
26024 call to a subprogram in a @code{with}'ed unit, or instantiate a generic
26025 package in a @code{with}'ed unit, then if the @code{with}'ed unit
26026 does not have pragma @code{Pure} or
26027 @code{Preelaborate}, then the client should have an
26028 @code{Elaborate_All} pragma for the @code{with}'ed unit.}
26030 @emph{In the case of instantiating a generic subprogram, it is always
26031 sufficient to have only an @code{Elaborate} pragma for the
26032 @code{with}'ed unit.}
26036 By following this rule a client is assured that calls and instantiations
26037 can be made without risk of an exception.
26039 In this mode GNAT traces all calls that are potentially made from
26040 elaboration code, and puts in any missing implicit @code{Elaborate}
26041 and @code{Elaborate_All} pragmas.
26042 The advantage of this approach is that no elaboration problems
26043 are possible if the binder can find an elaboration order that is
26044 consistent with these implicit @code{Elaborate} and
26045 @code{Elaborate_All} pragmas. The
26046 disadvantage of this approach is that no such order may exist.
26048 If the binder does not generate any diagnostics, then it means that it has
26049 found an elaboration order that is guaranteed to be safe. However, the binder
26050 may still be relying on implicitly generated @code{Elaborate} and
26051 @code{Elaborate_All} pragmas so portability to other compilers than GNAT is not
26054 If it is important to guarantee portability, then the compilations should
26057 (info messages for elaboration prag mas) switch. This will cause info messages
26058 to be generated indicating the missing @code{Elaborate} and
26059 @code{Elaborate_All} pragmas.
26060 Consider the following source program:
26062 @smallexample @c ada
26067 m : integer := k.r;
26074 where it is clear that there
26075 should be a pragma @code{Elaborate_All}
26076 for unit @code{k}. An implicit pragma will be generated, and it is
26077 likely that the binder will be able to honor it. However, if you want
26078 to port this program to some other Ada compiler than GNAT.
26079 it is safer to include the pragma explicitly in the source. If this
26080 unit is compiled with the
26082 switch, then the compiler outputs an information message:
26089 3. m : integer := k.r;
26091 >>> info: call to "r" may raise Program_Error
26092 >>> info: missing pragma Elaborate_All for "k"
26100 and these messages can be used as a guide for supplying manually
26101 the missing pragmas. It is usually a bad idea to use this
26102 option during development. That's because it will tell you when
26103 you need to put in a pragma, but cannot tell you when it is time
26104 to take it out. So the use of pragma @code{Elaborate_All} may lead to
26105 unnecessary dependencies and even false circularities.
26107 This default mode is more restrictive than the Ada Reference
26108 Manual, and it is possible to construct programs which will compile
26109 using the dynamic model described there, but will run into a
26110 circularity using the safer static model we have described.
26112 Of course any Ada compiler must be able to operate in a mode
26113 consistent with the requirements of the Ada Reference Manual,
26114 and in particular must have the capability of implementing the
26115 standard dynamic model of elaboration with run-time checks.
26117 In GNAT, this standard mode can be achieved either by the use of
26118 the @option{-gnatE} switch on the compiler (@command{gcc} or
26119 @command{gnatmake}) command, or by the use of the configuration pragma:
26121 @smallexample @c ada
26122 pragma Elaboration_Checks (DYNAMIC);
26126 Either approach will cause the unit affected to be compiled using the
26127 standard dynamic run-time elaboration checks described in the Ada
26128 Reference Manual. The static model is generally preferable, since it
26129 is clearly safer to rely on compile and link time checks rather than
26130 run-time checks. However, in the case of legacy code, it may be
26131 difficult to meet the requirements of the static model. This
26132 issue is further discussed in
26133 @ref{What to Do If the Default Elaboration Behavior Fails}.
26135 Note that the static model provides a strict subset of the allowed
26136 behavior and programs of the Ada Reference Manual, so if you do
26137 adhere to the static model and no circularities exist,
26138 then you are assured that your program will
26139 work using the dynamic model, providing that you remove any
26140 pragma Elaborate statements from the source.
26142 @node Treatment of Pragma Elaborate
26143 @section Treatment of Pragma Elaborate
26144 @cindex Pragma Elaborate
26147 The use of @code{pragma Elaborate}
26148 should generally be avoided in Ada 95 and Ada 2005 programs,
26149 since there is no guarantee that transitive calls
26150 will be properly handled. Indeed at one point, this pragma was placed
26151 in Annex J (Obsolescent Features), on the grounds that it is never useful.
26153 Now that's a bit restrictive. In practice, the case in which
26154 @code{pragma Elaborate} is useful is when the caller knows that there
26155 are no transitive calls, or that the called unit contains all necessary
26156 transitive @code{pragma Elaborate} statements, and legacy code often
26157 contains such uses.
26159 Strictly speaking the static mode in GNAT should ignore such pragmas,
26160 since there is no assurance at compile time that the necessary safety
26161 conditions are met. In practice, this would cause GNAT to be incompatible
26162 with correctly written Ada 83 code that had all necessary
26163 @code{pragma Elaborate} statements in place. Consequently, we made the
26164 decision that GNAT in its default mode will believe that if it encounters
26165 a @code{pragma Elaborate} then the programmer knows what they are doing,
26166 and it will trust that no elaboration errors can occur.
26168 The result of this decision is two-fold. First to be safe using the
26169 static mode, you should remove all @code{pragma Elaborate} statements.
26170 Second, when fixing circularities in existing code, you can selectively
26171 use @code{pragma Elaborate} statements to convince the static mode of
26172 GNAT that it need not generate an implicit @code{pragma Elaborate_All}
26175 When using the static mode with @option{-gnatwl}, any use of
26176 @code{pragma Elaborate} will generate a warning about possible
26179 @node Elaboration Issues for Library Tasks
26180 @section Elaboration Issues for Library Tasks
26181 @cindex Library tasks, elaboration issues
26182 @cindex Elaboration of library tasks
26185 In this section we examine special elaboration issues that arise for
26186 programs that declare library level tasks.
26188 Generally the model of execution of an Ada program is that all units are
26189 elaborated, and then execution of the program starts. However, the
26190 declaration of library tasks definitely does not fit this model. The
26191 reason for this is that library tasks start as soon as they are declared
26192 (more precisely, as soon as the statement part of the enclosing package
26193 body is reached), that is to say before elaboration
26194 of the program is complete. This means that if such a task calls a
26195 subprogram, or an entry in another task, the callee may or may not be
26196 elaborated yet, and in the standard
26197 Reference Manual model of dynamic elaboration checks, you can even
26198 get timing dependent Program_Error exceptions, since there can be
26199 a race between the elaboration code and the task code.
26201 The static model of elaboration in GNAT seeks to avoid all such
26202 dynamic behavior, by being conservative, and the conservative
26203 approach in this particular case is to assume that all the code
26204 in a task body is potentially executed at elaboration time if
26205 a task is declared at the library level.
26207 This can definitely result in unexpected circularities. Consider
26208 the following example
26210 @smallexample @c ada
26216 type My_Int is new Integer;
26218 function Ident (M : My_Int) return My_Int;
26222 package body Decls is
26223 task body Lib_Task is
26229 function Ident (M : My_Int) return My_Int is
26237 procedure Put_Val (Arg : Decls.My_Int);
26241 package body Utils is
26242 procedure Put_Val (Arg : Decls.My_Int) is
26244 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
26251 Decls.Lib_Task.Start;
26256 If the above example is compiled in the default static elaboration
26257 mode, then a circularity occurs. The circularity comes from the call
26258 @code{Utils.Put_Val} in the task body of @code{Decls.Lib_Task}. Since
26259 this call occurs in elaboration code, we need an implicit pragma
26260 @code{Elaborate_All} for @code{Utils}. This means that not only must
26261 the spec and body of @code{Utils} be elaborated before the body
26262 of @code{Decls}, but also the spec and body of any unit that is
26263 @code{with'ed} by the body of @code{Utils} must also be elaborated before
26264 the body of @code{Decls}. This is the transitive implication of
26265 pragma @code{Elaborate_All} and it makes sense, because in general
26266 the body of @code{Put_Val} might have a call to something in a
26267 @code{with'ed} unit.
26269 In this case, the body of Utils (actually its spec) @code{with's}
26270 @code{Decls}. Unfortunately this means that the body of @code{Decls}
26271 must be elaborated before itself, in case there is a call from the
26272 body of @code{Utils}.
26274 Here is the exact chain of events we are worrying about:
26278 In the body of @code{Decls} a call is made from within the body of a library
26279 task to a subprogram in the package @code{Utils}. Since this call may
26280 occur at elaboration time (given that the task is activated at elaboration
26281 time), we have to assume the worst, i.e., that the
26282 call does happen at elaboration time.
26285 This means that the body and spec of @code{Util} must be elaborated before
26286 the body of @code{Decls} so that this call does not cause an access before
26290 Within the body of @code{Util}, specifically within the body of
26291 @code{Util.Put_Val} there may be calls to any unit @code{with}'ed
26295 One such @code{with}'ed package is package @code{Decls}, so there
26296 might be a call to a subprogram in @code{Decls} in @code{Put_Val}.
26297 In fact there is such a call in this example, but we would have to
26298 assume that there was such a call even if it were not there, since
26299 we are not supposed to write the body of @code{Decls} knowing what
26300 is in the body of @code{Utils}; certainly in the case of the
26301 static elaboration model, the compiler does not know what is in
26302 other bodies and must assume the worst.
26305 This means that the spec and body of @code{Decls} must also be
26306 elaborated before we elaborate the unit containing the call, but
26307 that unit is @code{Decls}! This means that the body of @code{Decls}
26308 must be elaborated before itself, and that's a circularity.
26312 Indeed, if you add an explicit pragma @code{Elaborate_All} for @code{Utils} in
26313 the body of @code{Decls} you will get a true Ada Reference Manual
26314 circularity that makes the program illegal.
26316 In practice, we have found that problems with the static model of
26317 elaboration in existing code often arise from library tasks, so
26318 we must address this particular situation.
26320 Note that if we compile and run the program above, using the dynamic model of
26321 elaboration (that is to say use the @option{-gnatE} switch),
26322 then it compiles, binds,
26323 links, and runs, printing the expected result of 2. Therefore in some sense
26324 the circularity here is only apparent, and we need to capture
26325 the properties of this program that distinguish it from other library-level
26326 tasks that have real elaboration problems.
26328 We have four possible answers to this question:
26333 Use the dynamic model of elaboration.
26335 If we use the @option{-gnatE} switch, then as noted above, the program works.
26336 Why is this? If we examine the task body, it is apparent that the task cannot
26338 @code{accept} statement until after elaboration has been completed, because
26339 the corresponding entry call comes from the main program, not earlier.
26340 This is why the dynamic model works here. But that's really giving
26341 up on a precise analysis, and we prefer to take this approach only if we cannot
26343 problem in any other manner. So let us examine two ways to reorganize
26344 the program to avoid the potential elaboration problem.
26347 Split library tasks into separate packages.
26349 Write separate packages, so that library tasks are isolated from
26350 other declarations as much as possible. Let us look at a variation on
26353 @smallexample @c ada
26361 package body Decls1 is
26362 task body Lib_Task is
26370 type My_Int is new Integer;
26371 function Ident (M : My_Int) return My_Int;
26375 package body Decls2 is
26376 function Ident (M : My_Int) return My_Int is
26384 procedure Put_Val (Arg : Decls2.My_Int);
26388 package body Utils is
26389 procedure Put_Val (Arg : Decls2.My_Int) is
26391 Text_IO.Put_Line (Decls2.My_Int'Image (Decls2.Ident (Arg)));
26398 Decls1.Lib_Task.Start;
26403 All we have done is to split @code{Decls} into two packages, one
26404 containing the library task, and one containing everything else. Now
26405 there is no cycle, and the program compiles, binds, links and executes
26406 using the default static model of elaboration.
26409 Declare separate task types.
26411 A significant part of the problem arises because of the use of the
26412 single task declaration form. This means that the elaboration of
26413 the task type, and the elaboration of the task itself (i.e.@: the
26414 creation of the task) happen at the same time. A good rule
26415 of style in Ada is to always create explicit task types. By
26416 following the additional step of placing task objects in separate
26417 packages from the task type declaration, many elaboration problems
26418 are avoided. Here is another modified example of the example program:
26420 @smallexample @c ada
26422 task type Lib_Task_Type is
26426 type My_Int is new Integer;
26428 function Ident (M : My_Int) return My_Int;
26432 package body Decls is
26433 task body Lib_Task_Type is
26439 function Ident (M : My_Int) return My_Int is
26447 procedure Put_Val (Arg : Decls.My_Int);
26451 package body Utils is
26452 procedure Put_Val (Arg : Decls.My_Int) is
26454 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
26460 Lib_Task : Decls.Lib_Task_Type;
26466 Declst.Lib_Task.Start;
26471 What we have done here is to replace the @code{task} declaration in
26472 package @code{Decls} with a @code{task type} declaration. Then we
26473 introduce a separate package @code{Declst} to contain the actual
26474 task object. This separates the elaboration issues for
26475 the @code{task type}
26476 declaration, which causes no trouble, from the elaboration issues
26477 of the task object, which is also unproblematic, since it is now independent
26478 of the elaboration of @code{Utils}.
26479 This separation of concerns also corresponds to
26480 a generally sound engineering principle of separating declarations
26481 from instances. This version of the program also compiles, binds, links,
26482 and executes, generating the expected output.
26485 Use No_Entry_Calls_In_Elaboration_Code restriction.
26486 @cindex No_Entry_Calls_In_Elaboration_Code
26488 The previous two approaches described how a program can be restructured
26489 to avoid the special problems caused by library task bodies. in practice,
26490 however, such restructuring may be difficult to apply to existing legacy code,
26491 so we must consider solutions that do not require massive rewriting.
26493 Let us consider more carefully why our original sample program works
26494 under the dynamic model of elaboration. The reason is that the code
26495 in the task body blocks immediately on the @code{accept}
26496 statement. Now of course there is nothing to prohibit elaboration
26497 code from making entry calls (for example from another library level task),
26498 so we cannot tell in isolation that
26499 the task will not execute the accept statement during elaboration.
26501 However, in practice it is very unusual to see elaboration code
26502 make any entry calls, and the pattern of tasks starting
26503 at elaboration time and then immediately blocking on @code{accept} or
26504 @code{select} statements is very common. What this means is that
26505 the compiler is being too pessimistic when it analyzes the
26506 whole package body as though it might be executed at elaboration
26509 If we know that the elaboration code contains no entry calls, (a very safe
26510 assumption most of the time, that could almost be made the default
26511 behavior), then we can compile all units of the program under control
26512 of the following configuration pragma:
26515 pragma Restrictions (No_Entry_Calls_In_Elaboration_Code);
26519 This pragma can be placed in the @file{gnat.adc} file in the usual
26520 manner. If we take our original unmodified program and compile it
26521 in the presence of a @file{gnat.adc} containing the above pragma,
26522 then once again, we can compile, bind, link, and execute, obtaining
26523 the expected result. In the presence of this pragma, the compiler does
26524 not trace calls in a task body, that appear after the first @code{accept}
26525 or @code{select} statement, and therefore does not report a potential
26526 circularity in the original program.
26528 The compiler will check to the extent it can that the above
26529 restriction is not violated, but it is not always possible to do a
26530 complete check at compile time, so it is important to use this
26531 pragma only if the stated restriction is in fact met, that is to say
26532 no task receives an entry call before elaboration of all units is completed.
26536 @node Mixing Elaboration Models
26537 @section Mixing Elaboration Models
26539 So far, we have assumed that the entire program is either compiled
26540 using the dynamic model or static model, ensuring consistency. It
26541 is possible to mix the two models, but rules have to be followed
26542 if this mixing is done to ensure that elaboration checks are not
26545 The basic rule is that @emph{a unit compiled with the static model cannot
26546 be @code{with'ed} by a unit compiled with the dynamic model}. The
26547 reason for this is that in the static model, a unit assumes that
26548 its clients guarantee to use (the equivalent of) pragma
26549 @code{Elaborate_All} so that no elaboration checks are required
26550 in inner subprograms, and this assumption is violated if the
26551 client is compiled with dynamic checks.
26553 The precise rule is as follows. A unit that is compiled with dynamic
26554 checks can only @code{with} a unit that meets at least one of the
26555 following criteria:
26560 The @code{with'ed} unit is itself compiled with dynamic elaboration
26561 checks (that is with the @option{-gnatE} switch.
26564 The @code{with'ed} unit is an internal GNAT implementation unit from
26565 the System, Interfaces, Ada, or GNAT hierarchies.
26568 The @code{with'ed} unit has pragma Preelaborate or pragma Pure.
26571 The @code{with'ing} unit (that is the client) has an explicit pragma
26572 @code{Elaborate_All} for the @code{with'ed} unit.
26577 If this rule is violated, that is if a unit with dynamic elaboration
26578 checks @code{with's} a unit that does not meet one of the above four
26579 criteria, then the binder (@code{gnatbind}) will issue a warning
26580 similar to that in the following example:
26583 warning: "x.ads" has dynamic elaboration checks and with's
26584 warning: "y.ads" which has static elaboration checks
26588 These warnings indicate that the rule has been violated, and that as a result
26589 elaboration checks may be missed in the resulting executable file.
26590 This warning may be suppressed using the @option{-ws} binder switch
26591 in the usual manner.
26593 One useful application of this mixing rule is in the case of a subsystem
26594 which does not itself @code{with} units from the remainder of the
26595 application. In this case, the entire subsystem can be compiled with
26596 dynamic checks to resolve a circularity in the subsystem, while
26597 allowing the main application that uses this subsystem to be compiled
26598 using the more reliable default static model.
26600 @node What to Do If the Default Elaboration Behavior Fails
26601 @section What to Do If the Default Elaboration Behavior Fails
26604 If the binder cannot find an acceptable order, it outputs detailed
26605 diagnostics. For example:
26611 error: elaboration circularity detected
26612 info: "proc (body)" must be elaborated before "pack (body)"
26613 info: reason: Elaborate_All probably needed in unit "pack (body)"
26614 info: recompile "pack (body)" with -gnatel
26615 info: for full details
26616 info: "proc (body)"
26617 info: is needed by its spec:
26618 info: "proc (spec)"
26619 info: which is withed by:
26620 info: "pack (body)"
26621 info: "pack (body)" must be elaborated before "proc (body)"
26622 info: reason: pragma Elaborate in unit "proc (body)"
26628 In this case we have a cycle that the binder cannot break. On the one
26629 hand, there is an explicit pragma Elaborate in @code{proc} for
26630 @code{pack}. This means that the body of @code{pack} must be elaborated
26631 before the body of @code{proc}. On the other hand, there is elaboration
26632 code in @code{pack} that calls a subprogram in @code{proc}. This means
26633 that for maximum safety, there should really be a pragma
26634 Elaborate_All in @code{pack} for @code{proc} which would require that
26635 the body of @code{proc} be elaborated before the body of
26636 @code{pack}. Clearly both requirements cannot be satisfied.
26637 Faced with a circularity of this kind, you have three different options.
26640 @item Fix the program
26641 The most desirable option from the point of view of long-term maintenance
26642 is to rearrange the program so that the elaboration problems are avoided.
26643 One useful technique is to place the elaboration code into separate
26644 child packages. Another is to move some of the initialization code to
26645 explicitly called subprograms, where the program controls the order
26646 of initialization explicitly. Although this is the most desirable option,
26647 it may be impractical and involve too much modification, especially in
26648 the case of complex legacy code.
26650 @item Perform dynamic checks
26651 If the compilations are done using the
26653 (dynamic elaboration check) switch, then GNAT behaves in a quite different
26654 manner. Dynamic checks are generated for all calls that could possibly result
26655 in raising an exception. With this switch, the compiler does not generate
26656 implicit @code{Elaborate} or @code{Elaborate_All} pragmas. The behavior then is
26657 exactly as specified in the @cite{Ada Reference Manual}.
26658 The binder will generate
26659 an executable program that may or may not raise @code{Program_Error}, and then
26660 it is the programmer's job to ensure that it does not raise an exception. Note
26661 that it is important to compile all units with the switch, it cannot be used
26664 @item Suppress checks
26665 The drawback of dynamic checks is that they generate a
26666 significant overhead at run time, both in space and time. If you
26667 are absolutely sure that your program cannot raise any elaboration
26668 exceptions, and you still want to use the dynamic elaboration model,
26669 then you can use the configuration pragma
26670 @code{Suppress (Elaboration_Check)} to suppress all such checks. For
26671 example this pragma could be placed in the @file{gnat.adc} file.
26673 @item Suppress checks selectively
26674 When you know that certain calls or instantiations in elaboration code cannot
26675 possibly lead to an elaboration error, and the binder nevertheless complains
26676 about implicit @code{Elaborate} and @code{Elaborate_All} pragmas that lead to
26677 elaboration circularities, it is possible to remove those warnings locally and
26678 obtain a program that will bind. Clearly this can be unsafe, and it is the
26679 responsibility of the programmer to make sure that the resulting program has no
26680 elaboration anomalies. The pragma @code{Suppress (Elaboration_Check)} can be
26681 used with different granularity to suppress warnings and break elaboration
26686 Place the pragma that names the called subprogram in the declarative part
26687 that contains the call.
26690 Place the pragma in the declarative part, without naming an entity. This
26691 disables warnings on all calls in the corresponding declarative region.
26694 Place the pragma in the package spec that declares the called subprogram,
26695 and name the subprogram. This disables warnings on all elaboration calls to
26699 Place the pragma in the package spec that declares the called subprogram,
26700 without naming any entity. This disables warnings on all elaboration calls to
26701 all subprograms declared in this spec.
26703 @item Use Pragma Elaborate
26704 As previously described in section @xref{Treatment of Pragma Elaborate},
26705 GNAT in static mode assumes that a @code{pragma} Elaborate indicates correctly
26706 that no elaboration checks are required on calls to the designated unit.
26707 There may be cases in which the caller knows that no transitive calls
26708 can occur, so that a @code{pragma Elaborate} will be sufficient in a
26709 case where @code{pragma Elaborate_All} would cause a circularity.
26713 These five cases are listed in order of decreasing safety, and therefore
26714 require increasing programmer care in their application. Consider the
26717 @smallexample @c adanocomment
26719 function F1 return Integer;
26724 function F2 return Integer;
26725 function Pure (x : integer) return integer;
26726 -- pragma Suppress (Elaboration_Check, On => Pure); -- (3)
26727 -- pragma Suppress (Elaboration_Check); -- (4)
26731 package body Pack1 is
26732 function F1 return Integer is
26736 Val : integer := Pack2.Pure (11); -- Elab. call (1)
26739 -- pragma Suppress(Elaboration_Check, Pack2.F2); -- (1)
26740 -- pragma Suppress(Elaboration_Check); -- (2)
26742 X1 := Pack2.F2 + 1; -- Elab. call (2)
26747 package body Pack2 is
26748 function F2 return Integer is
26752 function Pure (x : integer) return integer is
26754 return x ** 3 - 3 * x;
26758 with Pack1, Ada.Text_IO;
26761 Ada.Text_IO.Put_Line(Pack1.X1'Img); -- 101
26764 In the absence of any pragmas, an attempt to bind this program produces
26765 the following diagnostics:
26771 error: elaboration circularity detected
26772 info: "pack1 (body)" must be elaborated before "pack1 (body)"
26773 info: reason: Elaborate_All probably needed in unit "pack1 (body)"
26774 info: recompile "pack1 (body)" with -gnatel for full details
26775 info: "pack1 (body)"
26776 info: must be elaborated along with its spec:
26777 info: "pack1 (spec)"
26778 info: which is withed by:
26779 info: "pack2 (body)"
26780 info: which must be elaborated along with its spec:
26781 info: "pack2 (spec)"
26782 info: which is withed by:
26783 info: "pack1 (body)"
26786 The sources of the circularity are the two calls to @code{Pack2.Pure} and
26787 @code{Pack2.F2} in the body of @code{Pack1}. We can see that the call to
26788 F2 is safe, even though F2 calls F1, because the call appears after the
26789 elaboration of the body of F1. Therefore the pragma (1) is safe, and will
26790 remove the warning on the call. It is also possible to use pragma (2)
26791 because there are no other potentially unsafe calls in the block.
26794 The call to @code{Pure} is safe because this function does not depend on the
26795 state of @code{Pack2}. Therefore any call to this function is safe, and it
26796 is correct to place pragma (3) in the corresponding package spec.
26799 Finally, we could place pragma (4) in the spec of @code{Pack2} to disable
26800 warnings on all calls to functions declared therein. Note that this is not
26801 necessarily safe, and requires more detailed examination of the subprogram
26802 bodies involved. In particular, a call to @code{F2} requires that @code{F1}
26803 be already elaborated.
26807 It is hard to generalize on which of these four approaches should be
26808 taken. Obviously if it is possible to fix the program so that the default
26809 treatment works, this is preferable, but this may not always be practical.
26810 It is certainly simple enough to use
26812 but the danger in this case is that, even if the GNAT binder
26813 finds a correct elaboration order, it may not always do so,
26814 and certainly a binder from another Ada compiler might not. A
26815 combination of testing and analysis (for which the
26816 information messages generated with the
26818 switch can be useful) must be used to ensure that the program is free
26819 of errors. One switch that is useful in this testing is the
26820 @option{^-p (pessimistic elaboration order)^/PESSIMISTIC_ELABORATION_ORDER^}
26823 Normally the binder tries to find an order that has the best chance
26824 of avoiding elaboration problems. However, if this switch is used, the binder
26825 plays a devil's advocate role, and tries to choose the order that
26826 has the best chance of failing. If your program works even with this
26827 switch, then it has a better chance of being error free, but this is still
26830 For an example of this approach in action, consider the C-tests (executable
26831 tests) from the ACVC suite. If these are compiled and run with the default
26832 treatment, then all but one of them succeed without generating any error
26833 diagnostics from the binder. However, there is one test that fails, and
26834 this is not surprising, because the whole point of this test is to ensure
26835 that the compiler can handle cases where it is impossible to determine
26836 a correct order statically, and it checks that an exception is indeed
26837 raised at run time.
26839 This one test must be compiled and run using the
26841 switch, and then it passes. Alternatively, the entire suite can
26842 be run using this switch. It is never wrong to run with the dynamic
26843 elaboration switch if your code is correct, and we assume that the
26844 C-tests are indeed correct (it is less efficient, but efficiency is
26845 not a factor in running the ACVC tests.)
26847 @node Elaboration for Indirect Calls
26848 @section Elaboration for Indirect Calls
26849 @cindex Dispatching calls
26850 @cindex Indirect calls
26853 In rare cases, the static elaboration model fails to prevent
26854 dispatching calls to not-yet-elaborated subprograms. In such cases, we
26855 fall back to run-time checks; premature calls to any primitive
26856 operation of a tagged type before the body of the operation has been
26857 elaborated will raise @code{Program_Error}.
26859 Access-to-subprogram types, however, are handled conservatively, and
26860 do not require run-time checks. This was not true in earlier versions
26861 of the compiler; you can use the @option{-gnatd.U} debug switch to
26862 revert to the old behavior if the new conservative behavior causes
26863 elaboration cycles. Here, ``conservative'' means that if you do
26864 @code{P'Access} during elaboration, the compiler will assume that you
26865 might call @code{P} indirectly during elaboration, so it adds an
26866 implicit @code{pragma Elaborate_All} on the library unit containing
26867 @code{P}. The @option{-gnatd.U} switch is safe if you know there are
26868 no such calls. If the program worked before, it will continue to work
26869 with @option{-gnatd.U}. But beware that code modifications such as
26870 adding an indirect call can cause erroneous behavior in the presence
26871 of @option{-gnatd.U}.
26873 @node Summary of Procedures for Elaboration Control
26874 @section Summary of Procedures for Elaboration Control
26875 @cindex Elaboration control
26878 First, compile your program with the default options, using none of
26879 the special elaboration control switches. If the binder successfully
26880 binds your program, then you can be confident that, apart from issues
26881 raised by the use of access-to-subprogram types and dynamic dispatching,
26882 the program is free of elaboration errors. If it is important that the
26883 program be portable to other compilers than GNAT, then use the
26885 switch to generate messages about missing @code{Elaborate} or
26886 @code{Elaborate_All} pragmas, and supply the missing pragmas.
26888 If the program fails to bind using the default static elaboration
26889 handling, then you can fix the program to eliminate the binder
26890 message, or recompile the entire program with the
26891 @option{-gnatE} switch to generate dynamic elaboration checks,
26892 and, if you are sure there really are no elaboration problems,
26893 use a global pragma @code{Suppress (Elaboration_Check)}.
26895 @node Other Elaboration Order Considerations
26896 @section Other Elaboration Order Considerations
26898 This section has been entirely concerned with the issue of finding a valid
26899 elaboration order, as defined by the Ada Reference Manual. In a case
26900 where several elaboration orders are valid, the task is to find one
26901 of the possible valid elaboration orders (and the static model in GNAT
26902 will ensure that this is achieved).
26904 The purpose of the elaboration rules in the Ada Reference Manual is to
26905 make sure that no entity is accessed before it has been elaborated. For
26906 a subprogram, this means that the spec and body must have been elaborated
26907 before the subprogram is called. For an object, this means that the object
26908 must have been elaborated before its value is read or written. A violation
26909 of either of these two requirements is an access before elaboration order,
26910 and this section has been all about avoiding such errors.
26912 In the case where more than one order of elaboration is possible, in the
26913 sense that access before elaboration errors are avoided, then any one of
26914 the orders is ``correct'' in the sense that it meets the requirements of
26915 the Ada Reference Manual, and no such error occurs.
26917 However, it may be the case for a given program, that there are
26918 constraints on the order of elaboration that come not from consideration
26919 of avoiding elaboration errors, but rather from extra-lingual logic
26920 requirements. Consider this example:
26922 @smallexample @c ada
26923 with Init_Constants;
26924 package Constants is
26929 package Init_Constants is
26930 procedure P; -- require a body
26931 end Init_Constants;
26934 package body Init_Constants is
26935 procedure P is begin null; end;
26939 end Init_Constants;
26943 Z : Integer := Constants.X + Constants.Y;
26947 with Text_IO; use Text_IO;
26950 Put_Line (Calc.Z'Img);
26955 In this example, there is more than one valid order of elaboration. For
26956 example both the following are correct orders:
26959 Init_Constants spec
26962 Init_Constants body
26967 Init_Constants spec
26968 Init_Constants body
26975 There is no language rule to prefer one or the other, both are correct
26976 from an order of elaboration point of view. But the programmatic effects
26977 of the two orders are very different. In the first, the elaboration routine
26978 of @code{Calc} initializes @code{Z} to zero, and then the main program
26979 runs with this value of zero. But in the second order, the elaboration
26980 routine of @code{Calc} runs after the body of Init_Constants has set
26981 @code{X} and @code{Y} and thus @code{Z} is set to 7 before @code{Main}
26984 One could perhaps by applying pretty clever non-artificial intelligence
26985 to the situation guess that it is more likely that the second order of
26986 elaboration is the one desired, but there is no formal linguistic reason
26987 to prefer one over the other. In fact in this particular case, GNAT will
26988 prefer the second order, because of the rule that bodies are elaborated
26989 as soon as possible, but it's just luck that this is what was wanted
26990 (if indeed the second order was preferred).
26992 If the program cares about the order of elaboration routines in a case like
26993 this, it is important to specify the order required. In this particular
26994 case, that could have been achieved by adding to the spec of Calc:
26996 @smallexample @c ada
26997 pragma Elaborate_All (Constants);
27001 which requires that the body (if any) and spec of @code{Constants},
27002 as well as the body and spec of any unit @code{with}'ed by
27003 @code{Constants} be elaborated before @code{Calc} is elaborated.
27005 Clearly no automatic method can always guess which alternative you require,
27006 and if you are working with legacy code that had constraints of this kind
27007 which were not properly specified by adding @code{Elaborate} or
27008 @code{Elaborate_All} pragmas, then indeed it is possible that two different
27009 compilers can choose different orders.
27011 However, GNAT does attempt to diagnose the common situation where there
27012 are uninitialized variables in the visible part of a package spec, and the
27013 corresponding package body has an elaboration block that directly or
27014 indirectly initialized one or more of these variables. This is the situation
27015 in which a pragma Elaborate_Body is usually desirable, and GNAT will generate
27016 a warning that suggests this addition if it detects this situation.
27018 The @code{gnatbind}
27019 @option{^-p^/PESSIMISTIC_ELABORATION^} switch may be useful in smoking
27020 out problems. This switch causes bodies to be elaborated as late as possible
27021 instead of as early as possible. In the example above, it would have forced
27022 the choice of the first elaboration order. If you get different results
27023 when using this switch, and particularly if one set of results is right,
27024 and one is wrong as far as you are concerned, it shows that you have some
27025 missing @code{Elaborate} pragmas. For the example above, we have the
27029 gnatmake -f -q main
27032 gnatmake -f -q main -bargs -p
27038 It is of course quite unlikely that both these results are correct, so
27039 it is up to you in a case like this to investigate the source of the
27040 difference, by looking at the two elaboration orders that are chosen,
27041 and figuring out which is correct, and then adding the necessary
27042 @code{Elaborate} or @code{Elaborate_All} pragmas to ensure the desired order.
27044 @node Determining the Chosen Elaboration Order
27045 @section Determining the Chosen Elaboration Order
27048 To see the elaboration order that the binder chooses, you can look at
27049 the last part of the b~xxx.adb binder output file. Here is an example:
27051 @smallexample @c ada
27052 System.Soft_Links'Elab_Body;
27054 System.Secondary_Stack'Elab_Body;
27056 System.Exception_Table'Elab_Body;
27058 Ada.Io_Exceptions'Elab_Spec;
27060 Ada.Tags'Elab_Spec;
27061 Ada.Streams'Elab_Spec;
27063 Interfaces.C'Elab_Spec;
27065 System.Finalization_Root'Elab_Spec;
27067 System.Os_Lib'Elab_Body;
27069 System.Finalization_Implementation'Elab_Spec;
27070 System.Finalization_Implementation'Elab_Body;
27072 Ada.Finalization'Elab_Spec;
27074 Ada.Finalization.List_Controller'Elab_Spec;
27076 System.File_Control_Block'Elab_Spec;
27078 System.File_Io'Elab_Body;
27080 Ada.Tags'Elab_Body;
27082 Ada.Text_Io'Elab_Spec;
27083 Ada.Text_Io'Elab_Body;
27088 Here Elab_Spec elaborates the spec
27089 and Elab_Body elaborates the body. The assignments to the Exx flags
27090 flag that the corresponding body is now elaborated.
27092 You can also ask the binder to generate a more
27093 readable list of the elaboration order using the
27094 @code{-l} switch when invoking the binder. Here is
27095 an example of the output generated by this switch:
27101 system.case_util (spec)
27102 system.case_util (body)
27103 system.concat_2 (spec)
27104 system.concat_2 (body)
27105 system.concat_3 (spec)
27106 system.concat_3 (body)
27107 system.htable (spec)
27108 system.parameters (spec)
27109 system.parameters (body)
27111 interfaces.c_streams (spec)
27112 interfaces.c_streams (body)
27113 system.restrictions (spec)
27114 system.restrictions (body)
27115 system.standard_library (spec)
27116 system.exceptions (spec)
27117 system.exceptions (body)
27118 system.storage_elements (spec)
27119 system.storage_elements (body)
27120 system.secondary_stack (spec)
27121 system.stack_checking (spec)
27122 system.stack_checking (body)
27123 system.string_hash (spec)
27124 system.string_hash (body)
27125 system.htable (body)
27126 system.strings (spec)
27127 system.strings (body)
27128 system.traceback (spec)
27129 system.traceback (body)
27130 system.traceback_entries (spec)
27131 system.traceback_entries (body)
27132 ada.exceptions (spec)
27133 ada.exceptions.last_chance_handler (spec)
27134 system.soft_links (spec)
27135 system.soft_links (body)
27136 ada.exceptions.last_chance_handler (body)
27137 system.secondary_stack (body)
27138 system.exception_table (spec)
27139 system.exception_table (body)
27140 ada.io_exceptions (spec)
27143 interfaces.c (spec)
27144 interfaces.c (body)
27145 system.finalization_root (spec)
27146 system.finalization_root (body)
27147 system.memory (spec)
27148 system.memory (body)
27149 system.standard_library (body)
27150 system.os_lib (spec)
27151 system.os_lib (body)
27152 system.unsigned_types (spec)
27153 system.stream_attributes (spec)
27154 system.stream_attributes (body)
27155 system.finalization_implementation (spec)
27156 system.finalization_implementation (body)
27157 ada.finalization (spec)
27158 ada.finalization (body)
27159 ada.finalization.list_controller (spec)
27160 ada.finalization.list_controller (body)
27161 system.file_control_block (spec)
27162 system.file_io (spec)
27163 system.file_io (body)
27164 system.val_uns (spec)
27165 system.val_util (spec)
27166 system.val_util (body)
27167 system.val_uns (body)
27168 system.wch_con (spec)
27169 system.wch_con (body)
27170 system.wch_cnv (spec)
27171 system.wch_jis (spec)
27172 system.wch_jis (body)
27173 system.wch_cnv (body)
27174 system.wch_stw (spec)
27175 system.wch_stw (body)
27177 ada.exceptions (body)
27184 @c **********************************
27185 @node Overflow Check Handling in GNAT
27186 @appendix Overflow Check Handling in GNAT
27187 @cindex Overflow checks
27188 @cindex Checks (overflow)
27189 @c **********************************
27193 * Overflow Checking Modes in GNAT::
27194 * Specifying the Desired Mode::
27195 * Default Settings::
27196 * Implementation Notes::
27201 @section Background
27204 Overflow checks are checks that the compiler may make to ensure
27205 that intermediate results are not out of range. For example:
27207 @smallexample @c ada
27214 if @code{A} has the value @code{Integer'Last}, then the addition may cause
27215 overflow since the result is out of range of the type @code{Integer}.
27216 In this case @code{Constraint_Error} will be raised if checks are
27219 A trickier situation arises in examples like the following:
27221 @smallexample @c ada
27228 where @code{A} is @code{Integer'Last} and @code{C} is @code{-1}.
27229 Now the final result of the expression on the right hand side is
27230 @code{Integer'Last} which is in range, but the question arises whether the
27231 intermediate addition of @code{(A + 1)} raises an overflow error.
27233 The (perhaps surprising) answer is that the Ada language
27234 definition does not answer this question. Instead it leaves
27235 it up to the implementation to do one of two things if overflow
27236 checks are enabled.
27240 raise an exception (@code{Constraint_Error}), or
27243 yield the correct mathematical result which is then used in
27244 subsequent operations.
27248 If the compiler chooses the first approach, then the assignment of this
27249 example will indeed raise @code{Constraint_Error} if overflow checking is
27250 enabled, or result in erroneous execution if overflow checks are suppressed.
27252 But if the compiler
27253 chooses the second approach, then it can perform both additions yielding
27254 the correct mathematical result, which is in range, so no exception
27255 will be raised, and the right result is obtained, regardless of whether
27256 overflow checks are suppressed.
27258 Note that in the first example an
27259 exception will be raised in either case, since if the compiler
27260 gives the correct mathematical result for the addition, it will
27261 be out of range of the target type of the assignment, and thus
27262 fails the range check.
27264 This lack of specified behavior in the handling of overflow for
27265 intermediate results is a source of non-portability, and can thus
27266 be problematic when programs are ported. Most typically this arises
27267 in a situation where the original compiler did not raise an exception,
27268 and then the application is moved to a compiler where the check is
27269 performed on the intermediate result and an unexpected exception is
27272 Furthermore, when using Ada 2012's preconditions and other
27273 assertion forms, another issue arises. Consider:
27275 @smallexample @c ada
27276 procedure P (A, B : Integer) with
27277 Pre => A + B <= Integer'Last;
27281 One often wants to regard arithmetic in a context like this from
27282 a mathematical point of view. So for example, if the two actual parameters
27283 for a call to @code{P} are both @code{Integer'Last}, then
27284 the precondition should be regarded as False. If we are executing
27285 in a mode with run-time checks enabled for preconditions, then we would
27286 like this precondition to fail, rather than raising an exception
27287 because of the intermediate overflow.
27289 However, the language definition leaves the specification of
27290 whether the above condition fails (raising @code{Assert_Error}) or
27291 causes an intermediate overflow (raising @code{Constraint_Error})
27292 up to the implementation.
27294 The situation is worse in a case such as the following:
27296 @smallexample @c ada
27297 procedure Q (A, B, C : Integer) with
27298 Pre => A + B + C <= Integer'Last;
27304 @smallexample @c ada
27305 Q (A => Integer'Last, B => 1, C => -1);
27309 From a mathematical point of view the precondition
27310 is True, but at run time we may (but are not guaranteed to) get an
27311 exception raised because of the intermediate overflow (and we really
27312 would prefer this precondition to be considered True at run time).
27314 @node Overflow Checking Modes in GNAT
27315 @section Overflow Checking Modes in GNAT
27318 To deal with the portability issue, and with the problem of
27319 mathematical versus run-time interpretation of the expressions in
27320 assertions, GNAT provides comprehensive control over the handling
27321 of intermediate overflow. GNAT can operate in three modes, and
27322 furthemore, permits separate selection of operating modes for
27323 the expressions within assertions (here the term ``assertions''
27324 is used in the technical sense, which includes preconditions and so forth)
27325 and for expressions appearing outside assertions.
27327 The three modes are:
27330 @item @i{Use base type for intermediate operations} (@code{STRICT})
27332 In this mode, all intermediate results for predefined arithmetic
27333 operators are computed using the base type, and the result must
27334 be in range of the base type. If this is not the
27335 case then either an exception is raised (if overflow checks are
27336 enabled) or the execution is erroneous (if overflow checks are suppressed).
27337 This is the normal default mode.
27339 @item @i{Most intermediate overflows avoided} (@code{MINIMIZED})
27341 In this mode, the compiler attempts to avoid intermediate overflows by
27342 using a larger integer type, typically @code{Long_Long_Integer},
27343 as the type in which arithmetic is
27344 performed for predefined arithmetic operators. This may be slightly more
27346 run time (compared to suppressing intermediate overflow checks), though
27347 the cost is negligible on modern 64-bit machines. For the examples given
27348 earlier, no intermediate overflows would have resulted in exceptions,
27349 since the intermediate results are all in the range of
27350 @code{Long_Long_Integer} (typically 64-bits on nearly all implementations
27351 of GNAT). In addition, if checks are enabled, this reduces the number of
27352 checks that must be made, so this choice may actually result in an
27353 improvement in space and time behavior.
27355 However, there are cases where @code{Long_Long_Integer} is not large
27356 enough, consider the following example:
27358 @smallexample @c ada
27359 procedure R (A, B, C, D : Integer) with
27360 Pre => (A**2 * B**2) / (C**2 * D**2) <= 10;
27363 where @code{A} = @code{B} = @code{C} = @code{D} = @code{Integer'Last}.
27364 Now the intermediate results are
27365 out of the range of @code{Long_Long_Integer} even though the final result
27366 is in range and the precondition is True (from a mathematical point
27367 of view). In such a case, operating in this mode, an overflow occurs
27368 for the intermediate computation (which is why this mode
27369 says @i{most} intermediate overflows are avoided). In this case,
27370 an exception is raised if overflow checks are enabled, and the
27371 execution is erroneous if overflow checks are suppressed.
27373 @item @i{All intermediate overflows avoided} (@code{ELIMINATED})
27375 In this mode, the compiler avoids all intermediate overflows
27376 by using arbitrary precision arithmetic as required. In this
27377 mode, the above example with @code{A**2 * B**2} would
27378 not cause intermediate overflow, because the intermediate result
27379 would be evaluated using sufficient precision, and the result
27380 of evaluating the precondition would be True.
27382 This mode has the advantage of avoiding any intermediate
27383 overflows, but at the expense of significant run-time overhead,
27384 including the use of a library (included automatically in this
27385 mode) for multiple-precision arithmetic.
27387 This mode provides cleaner semantics for assertions, since now
27388 the run-time behavior emulates true arithmetic behavior for the
27389 predefined arithmetic operators, meaning that there is never a
27390 conflict between the mathematical view of the assertion, and its
27393 Note that in this mode, the behavior is unaffected by whether or
27394 not overflow checks are suppressed, since overflow does not occur.
27395 It is possible for gigantic intermediate expressions to raise
27396 @code{Storage_Error} as a result of attempting to compute the
27397 results of such expressions (e.g. @code{Integer'Last ** Integer'Last})
27398 but overflow is impossible.
27404 Note that these modes apply only to the evaluation of predefined
27405 arithmetic, membership, and comparison operators for signed integer
27408 For fixed-point arithmetic, checks can be suppressed. But if checks
27410 then fixed-point values are always checked for overflow against the
27411 base type for intermediate expressions (that is such checks always
27412 operate in the equivalent of @code{STRICT} mode).
27414 For floating-point, on nearly all architectures, @code{Machine_Overflows}
27415 is False, and IEEE infinities are generated, so overflow exceptions
27416 are never raised. If you want to avoid infinities, and check that
27417 final results of expressions are in range, then you can declare a
27418 constrained floating-point type, and range checks will be carried
27419 out in the normal manner (with infinite values always failing all
27423 @c -------------------------
27424 @node Specifying the Desired Mode
27425 @section Specifying the Desired Mode
27428 The desired mode of for handling intermediate overflow can be specified using
27429 either the @code{Overflow_Mode} pragma or an equivalent compiler switch.
27430 The pragma has the form
27431 @cindex pragma @code{Overflow_Mode}
27433 @smallexample @c ada
27434 pragma Overflow_Mode ([General =>] MODE [, [Assertions =>] MODE]);
27438 where @code{MODE} is one of
27441 @item @code{STRICT}: intermediate overflows checked (using base type)
27442 @item @code{MINIMIZED}: minimize intermediate overflows
27443 @item @code{ELIMINATED}: eliminate intermediate overflows
27447 The case is ignored, so @code{MINIMIZED}, @code{Minimized} and
27448 @code{minimized} all have the same effect.
27450 If only the @code{General} parameter is present, then the given @code{MODE}
27452 to expressions both within and outside assertions. If both arguments
27453 are present, then @code{General} applies to expressions outside assertions,
27454 and @code{Assertions} applies to expressions within assertions. For example:
27456 @smallexample @c ada
27457 pragma Overflow_Mode
27458 (General => Minimized, Assertions => Eliminated);
27462 specifies that general expressions outside assertions be evaluated
27463 in ``minimize intermediate overflows'' mode, and expressions within
27464 assertions be evaluated in ``eliminate intermediate overflows'' mode.
27465 This is often a reasonable choice, avoiding excessive overhead
27466 outside assertions, but assuring a high degree of portability
27467 when importing code from another compiler, while incurring
27468 the extra overhead for assertion expressions to ensure that
27469 the behavior at run time matches the expected mathematical
27472 The @code{Overflow_Mode} pragma has the same scoping and placement
27473 rules as pragma @code{Suppress}, so it can occur either as a
27474 configuration pragma, specifying a default for the whole
27475 program, or in a declarative scope, where it applies to the
27476 remaining declarations and statements in that scope.
27478 Note that pragma @code{Overflow_Mode} does not affect whether
27479 overflow checks are enabled or suppressed. It only controls the
27480 method used to compute intermediate values. To control whether
27481 overflow checking is enabled or suppressed, use pragma @code{Suppress}
27482 or @code{Unsuppress} in the usual manner
27484 Additionally, a compiler switch @option{-gnato?} or @option{-gnato??}
27485 can be used to control the checking mode default (which can be subsequently
27486 overridden using pragmas).
27487 @cindex @option{-gnato?} (gcc)
27488 @cindex @option{-gnato??} (gcc)
27490 Here `@code{?}' is one of the digits `@code{1}' through `@code{3}':
27494 use base type for intermediate operations (@code{STRICT})
27496 minimize intermediate overflows (@code{MINIMIZED})
27498 eliminate intermediate overflows (@code{ELIMINATED})
27502 As with the pragma, if only one digit appears then it applies to all
27503 cases; if two digits are given, then the first applies outside
27504 assertions, and the second within assertions. Thus the equivalent
27505 of the example pragma above would be
27506 @option{^-gnato23^/OVERFLOW_CHECKS=23^}.
27508 If no digits follow the @option{-gnato}, then it is equivalent to
27509 @option{^-gnato11^/OVERFLOW_CHECKS=11^},
27510 causing all intermediate operations to be computed using the base
27511 type (@code{STRICT} mode).
27513 In addition to setting the mode used for computation of intermediate
27514 results, the @code{-gnato} switch also enables overflow checking (which
27515 is suppressed by default). It thus combines the effect of using
27516 a pragma @code{Overflow_Mode} and pragma @code{Unsuppress}.
27519 @c -------------------------
27520 @node Default Settings
27521 @section Default Settings
27523 The default mode for overflow checks is
27530 which causes all computations both inside and outside assertions to use
27531 the base type. In addition overflow checks are suppressed.
27533 This retains compatibility with previous versions of
27534 GNAT which suppressed overflow checks by default and always
27535 used the base type for computation of intermediate results.
27537 The switch @option{-gnato} (with no digits following) is equivalent to
27538 @cindex @option{-gnato} (gcc)
27545 which causes overflow checking of all intermediate overflows
27546 both inside and outside assertions against the base type.
27547 This provides compatibility
27548 with this switch as implemented in previous versions of GNAT.
27550 The pragma @code{Suppress (Overflow_Check)} disables overflow
27551 checking, but it has no effect on the method used for computing
27552 intermediate results.
27554 The pragma @code{Unsuppress (Overflow_Check)} enables overflow
27555 checking, but it has no effect on the method used for computing
27556 intermediate results.
27558 @c -------------------------
27559 @node Implementation Notes
27560 @section Implementation Notes
27562 In practice on typical 64-bit machines, the @code{MINIMIZED} mode is
27563 reasonably efficient, and can be generally used. It also helps
27564 to ensure compatibility with code imported from some other
27567 Setting all intermediate overflows checking (@code{CHECKED} mode)
27568 makes sense if you want to
27569 make sure that your code is compatible with any other possible
27570 Ada implementation. This may be useful in ensuring portability
27571 for code that is to be exported to some other compiler than GNAT.
27574 The Ada standard allows the reassociation of expressions at
27575 the same precedence level if no parentheses are present. For
27576 example, @w{@code{A+B+C}} parses as though it were @w{@code{(A+B)+C}}, but
27577 the compiler can reintepret this as @w{@code{A+(B+C)}}, possibly
27578 introducing or eliminating an overflow exception. The GNAT
27579 compiler never takes advantage of this freedom, and the
27580 expression @w{@code{A+B+C}} will be evaluated as @w{@code{(A+B)+C}}.
27581 If you need the other order, you can write the parentheses
27582 explicitly @w{@code{A+(B+C)}} and GNAT will respect this order.
27584 The use of @code{ELIMINATED} mode will cause the compiler to
27585 automatically include an appropriate arbitrary precision
27586 integer arithmetic package. The compiler will make calls
27587 to this package, though only in cases where it cannot be
27588 sure that @code{Long_Long_Integer} is sufficient to guard against
27589 intermediate overflows. This package does not use dynamic
27590 alllocation, but it does use the secondary stack, so an
27591 appropriate secondary stack package must be present (this
27592 is always true for standard full Ada, but may require
27593 specific steps for restricted run times such as ZFP).
27595 Although @code{ELIMINATED} mode causes expressions to use arbitrary
27596 precision arithmetic, avoiding overflow, the final result
27597 must be in an appropriate range. This is true even if the
27598 final result is of type @code{[Long_[Long_]]Integer'Base}, which
27599 still has the same bounds as its associated constrained
27602 Currently, the @code{ELIMINATED} mode is only available on target
27603 platforms for which @code{Long_Long_Integer} is 64-bits (nearly all GNAT
27606 @c *******************************
27607 @node Conditional Compilation
27608 @appendix Conditional Compilation
27609 @c *******************************
27610 @cindex Conditional compilation
27613 It is often necessary to arrange for a single source program
27614 to serve multiple purposes, where it is compiled in different
27615 ways to achieve these different goals. Some examples of the
27616 need for this feature are
27619 @item Adapting a program to a different hardware environment
27620 @item Adapting a program to a different target architecture
27621 @item Turning debugging features on and off
27622 @item Arranging for a program to compile with different compilers
27626 In C, or C++, the typical approach would be to use the preprocessor
27627 that is defined as part of the language. The Ada language does not
27628 contain such a feature. This is not an oversight, but rather a very
27629 deliberate design decision, based on the experience that overuse of
27630 the preprocessing features in C and C++ can result in programs that
27631 are extremely difficult to maintain. For example, if we have ten
27632 switches that can be on or off, this means that there are a thousand
27633 separate programs, any one of which might not even be syntactically
27634 correct, and even if syntactically correct, the resulting program
27635 might not work correctly. Testing all combinations can quickly become
27638 Nevertheless, the need to tailor programs certainly exists, and in
27639 this Appendix we will discuss how this can
27640 be achieved using Ada in general, and GNAT in particular.
27643 * Use of Boolean Constants::
27644 * Debugging - A Special Case::
27645 * Conditionalizing Declarations::
27646 * Use of Alternative Implementations::
27650 @node Use of Boolean Constants
27651 @section Use of Boolean Constants
27654 In the case where the difference is simply which code
27655 sequence is executed, the cleanest solution is to use Boolean
27656 constants to control which code is executed.
27658 @smallexample @c ada
27660 FP_Initialize_Required : constant Boolean := True;
27662 if FP_Initialize_Required then
27669 Not only will the code inside the @code{if} statement not be executed if
27670 the constant Boolean is @code{False}, but it will also be completely
27671 deleted from the program.
27672 However, the code is only deleted after the @code{if} statement
27673 has been checked for syntactic and semantic correctness.
27674 (In contrast, with preprocessors the code is deleted before the
27675 compiler ever gets to see it, so it is not checked until the switch
27677 @cindex Preprocessors (contrasted with conditional compilation)
27679 Typically the Boolean constants will be in a separate package,
27682 @smallexample @c ada
27685 FP_Initialize_Required : constant Boolean := True;
27686 Reset_Available : constant Boolean := False;
27693 The @code{Config} package exists in multiple forms for the various targets,
27694 with an appropriate script selecting the version of @code{Config} needed.
27695 Then any other unit requiring conditional compilation can do a @code{with}
27696 of @code{Config} to make the constants visible.
27699 @node Debugging - A Special Case
27700 @section Debugging - A Special Case
27703 A common use of conditional code is to execute statements (for example
27704 dynamic checks, or output of intermediate results) under control of a
27705 debug switch, so that the debugging behavior can be turned on and off.
27706 This can be done using a Boolean constant to control whether the code
27709 @smallexample @c ada
27712 Put_Line ("got to the first stage!");
27720 @smallexample @c ada
27722 if Debugging and then Temperature > 999.0 then
27723 raise Temperature_Crazy;
27729 Since this is a common case, there are special features to deal with
27730 this in a convenient manner. For the case of tests, Ada 2005 has added
27731 a pragma @code{Assert} that can be used for such tests. This pragma is modeled
27732 @cindex pragma @code{Assert}
27733 on the @code{Assert} pragma that has always been available in GNAT, so this
27734 feature may be used with GNAT even if you are not using Ada 2005 features.
27735 The use of pragma @code{Assert} is described in
27736 @ref{Pragma Assert,,, gnat_rm, GNAT Reference Manual}, but as an
27737 example, the last test could be written:
27739 @smallexample @c ada
27740 pragma Assert (Temperature <= 999.0, "Temperature Crazy");
27746 @smallexample @c ada
27747 pragma Assert (Temperature <= 999.0);
27751 In both cases, if assertions are active and the temperature is excessive,
27752 the exception @code{Assert_Failure} will be raised, with the given string in
27753 the first case or a string indicating the location of the pragma in the second
27754 case used as the exception message.
27756 You can turn assertions on and off by using the @code{Assertion_Policy}
27758 @cindex pragma @code{Assertion_Policy}
27759 This is an Ada 2005 pragma which is implemented in all modes by
27760 GNAT, but only in the latest versions of GNAT which include Ada 2005
27761 capability. Alternatively, you can use the @option{-gnata} switch
27762 @cindex @option{-gnata} switch
27763 to enable assertions from the command line (this is recognized by all versions
27766 For the example above with the @code{Put_Line}, the GNAT-specific pragma
27767 @code{Debug} can be used:
27768 @cindex pragma @code{Debug}
27770 @smallexample @c ada
27771 pragma Debug (Put_Line ("got to the first stage!"));
27775 If debug pragmas are enabled, the argument, which must be of the form of
27776 a procedure call, is executed (in this case, @code{Put_Line} will be called).
27777 Only one call can be present, but of course a special debugging procedure
27778 containing any code you like can be included in the program and then
27779 called in a pragma @code{Debug} argument as needed.
27781 One advantage of pragma @code{Debug} over the @code{if Debugging then}
27782 construct is that pragma @code{Debug} can appear in declarative contexts,
27783 such as at the very beginning of a procedure, before local declarations have
27786 Debug pragmas are enabled using either the @option{-gnata} switch that also
27787 controls assertions, or with a separate Debug_Policy pragma.
27788 @cindex pragma @code{Debug_Policy}
27789 The latter pragma is new in the Ada 2005 versions of GNAT (but it can be used
27790 in Ada 95 and Ada 83 programs as well), and is analogous to
27791 pragma @code{Assertion_Policy} to control assertions.
27793 @code{Assertion_Policy} and @code{Debug_Policy} are configuration pragmas,
27794 and thus they can appear in @file{gnat.adc} if you are not using a
27795 project file, or in the file designated to contain configuration pragmas
27797 They then apply to all subsequent compilations. In practice the use of
27798 the @option{-gnata} switch is often the most convenient method of controlling
27799 the status of these pragmas.
27801 Note that a pragma is not a statement, so in contexts where a statement
27802 sequence is required, you can't just write a pragma on its own. You have
27803 to add a @code{null} statement.
27805 @smallexample @c ada
27808 @dots{} -- some statements
27810 pragma Assert (Num_Cases < 10);
27817 @node Conditionalizing Declarations
27818 @section Conditionalizing Declarations
27821 In some cases, it may be necessary to conditionalize declarations to meet
27822 different requirements. For example we might want a bit string whose length
27823 is set to meet some hardware message requirement.
27825 In some cases, it may be possible to do this using declare blocks controlled
27826 by conditional constants:
27828 @smallexample @c ada
27830 if Small_Machine then
27832 X : Bit_String (1 .. 10);
27838 X : Large_Bit_String (1 .. 1000);
27847 Note that in this approach, both declarations are analyzed by the
27848 compiler so this can only be used where both declarations are legal,
27849 even though one of them will not be used.
27851 Another approach is to define integer constants, e.g.@: @code{Bits_Per_Word},
27852 or Boolean constants, e.g.@: @code{Little_Endian}, and then write declarations
27853 that are parameterized by these constants. For example
27855 @smallexample @c ada
27858 Field1 at 0 range Boolean'Pos (Little_Endian) * 10 .. Bits_Per_Word;
27864 If @code{Bits_Per_Word} is set to 32, this generates either
27866 @smallexample @c ada
27869 Field1 at 0 range 0 .. 32;
27875 for the big endian case, or
27877 @smallexample @c ada
27880 Field1 at 0 range 10 .. 32;
27886 for the little endian case. Since a powerful subset of Ada expression
27887 notation is usable for creating static constants, clever use of this
27888 feature can often solve quite difficult problems in conditionalizing
27889 compilation (note incidentally that in Ada 95, the little endian
27890 constant was introduced as @code{System.Default_Bit_Order}, so you do not
27891 need to define this one yourself).
27894 @node Use of Alternative Implementations
27895 @section Use of Alternative Implementations
27898 In some cases, none of the approaches described above are adequate. This
27899 can occur for example if the set of declarations required is radically
27900 different for two different configurations.
27902 In this situation, the official Ada way of dealing with conditionalizing
27903 such code is to write separate units for the different cases. As long as
27904 this does not result in excessive duplication of code, this can be done
27905 without creating maintenance problems. The approach is to share common
27906 code as far as possible, and then isolate the code and declarations
27907 that are different. Subunits are often a convenient method for breaking
27908 out a piece of a unit that is to be conditionalized, with separate files
27909 for different versions of the subunit for different targets, where the
27910 build script selects the right one to give to the compiler.
27911 @cindex Subunits (and conditional compilation)
27913 As an example, consider a situation where a new feature in Ada 2005
27914 allows something to be done in a really nice way. But your code must be able
27915 to compile with an Ada 95 compiler. Conceptually you want to say:
27917 @smallexample @c ada
27920 @dots{} neat Ada 2005 code
27922 @dots{} not quite as neat Ada 95 code
27928 where @code{Ada_2005} is a Boolean constant.
27930 But this won't work when @code{Ada_2005} is set to @code{False},
27931 since the @code{then} clause will be illegal for an Ada 95 compiler.
27932 (Recall that although such unreachable code would eventually be deleted
27933 by the compiler, it still needs to be legal. If it uses features
27934 introduced in Ada 2005, it will be illegal in Ada 95.)
27936 So instead we write
27938 @smallexample @c ada
27939 procedure Insert is separate;
27943 Then we have two files for the subunit @code{Insert}, with the two sets of
27945 If the package containing this is called @code{File_Queries}, then we might
27949 @item @file{file_queries-insert-2005.adb}
27950 @item @file{file_queries-insert-95.adb}
27954 and the build script renames the appropriate file to
27957 file_queries-insert.adb
27961 and then carries out the compilation.
27963 This can also be done with project files' naming schemes. For example:
27965 @smallexample @c project
27966 For Body ("File_Queries.Insert") use "file_queries-insert-2005.ada";
27970 Note also that with project files it is desirable to use a different extension
27971 than @file{ads} / @file{adb} for alternative versions. Otherwise a naming
27972 conflict may arise through another commonly used feature: to declare as part
27973 of the project a set of directories containing all the sources obeying the
27974 default naming scheme.
27976 The use of alternative units is certainly feasible in all situations,
27977 and for example the Ada part of the GNAT run-time is conditionalized
27978 based on the target architecture using this approach. As a specific example,
27979 consider the implementation of the AST feature in VMS. There is one
27987 which is the same for all architectures, and three bodies:
27991 used for all non-VMS operating systems
27992 @item s-asthan-vms-alpha.adb
27993 used for VMS on the Alpha
27994 @item s-asthan-vms-ia64.adb
27995 used for VMS on the ia64
27999 The dummy version @file{s-asthan.adb} simply raises exceptions noting that
28000 this operating system feature is not available, and the two remaining
28001 versions interface with the corresponding versions of VMS to provide
28002 VMS-compatible AST handling. The GNAT build script knows the architecture
28003 and operating system, and automatically selects the right version,
28004 renaming it if necessary to @file{s-asthan.adb} before the run-time build.
28006 Another style for arranging alternative implementations is through Ada's
28007 access-to-subprogram facility.
28008 In case some functionality is to be conditionally included,
28009 you can declare an access-to-procedure variable @code{Ref} that is initialized
28010 to designate a ``do nothing'' procedure, and then invoke @code{Ref.all}
28012 In some library package, set @code{Ref} to @code{Proc'Access} for some
28013 procedure @code{Proc} that performs the relevant processing.
28014 The initialization only occurs if the library package is included in the
28016 The same idea can also be implemented using tagged types and dispatching
28020 @node Preprocessing
28021 @section Preprocessing
28022 @cindex Preprocessing
28025 Although it is quite possible to conditionalize code without the use of
28026 C-style preprocessing, as described earlier in this section, it is
28027 nevertheless convenient in some cases to use the C approach. Moreover,
28028 older Ada compilers have often provided some preprocessing capability,
28029 so legacy code may depend on this approach, even though it is not
28032 To accommodate such use, GNAT provides a preprocessor (modeled to a large
28033 extent on the various preprocessors that have been used
28034 with legacy code on other compilers, to enable easier transition).
28036 The preprocessor may be used in two separate modes. It can be used quite
28037 separately from the compiler, to generate a separate output source file
28038 that is then fed to the compiler as a separate step. This is the
28039 @code{gnatprep} utility, whose use is fully described in
28040 @ref{Preprocessing with gnatprep}.
28041 @cindex @code{gnatprep}
28043 The preprocessing language allows such constructs as
28047 #if DEBUG or else (PRIORITY > 4) then
28048 bunch of declarations
28050 completely different bunch of declarations
28056 The values of the symbols @code{DEBUG} and @code{PRIORITY} can be
28057 defined either on the command line or in a separate file.
28059 The other way of running the preprocessor is even closer to the C style and
28060 often more convenient. In this approach the preprocessing is integrated into
28061 the compilation process. The compiler is fed the preprocessor input which
28062 includes @code{#if} lines etc, and then the compiler carries out the
28063 preprocessing internally and processes the resulting output.
28064 For more details on this approach, see @ref{Integrated Preprocessing}.
28067 @c *******************************
28068 @node Inline Assembler
28069 @appendix Inline Assembler
28070 @c *******************************
28073 If you need to write low-level software that interacts directly
28074 with the hardware, Ada provides two ways to incorporate assembly
28075 language code into your program. First, you can import and invoke
28076 external routines written in assembly language, an Ada feature fully
28077 supported by GNAT@. However, for small sections of code it may be simpler
28078 or more efficient to include assembly language statements directly
28079 in your Ada source program, using the facilities of the implementation-defined
28080 package @code{System.Machine_Code}, which incorporates the gcc
28081 Inline Assembler. The Inline Assembler approach offers a number of advantages,
28082 including the following:
28085 @item No need to use non-Ada tools
28086 @item Consistent interface over different targets
28087 @item Automatic usage of the proper calling conventions
28088 @item Access to Ada constants and variables
28089 @item Definition of intrinsic routines
28090 @item Possibility of inlining a subprogram comprising assembler code
28091 @item Code optimizer can take Inline Assembler code into account
28094 This chapter presents a series of examples to show you how to use
28095 the Inline Assembler. Although it focuses on the Intel x86,
28096 the general approach applies also to other processors.
28097 It is assumed that you are familiar with Ada
28098 and with assembly language programming.
28101 * Basic Assembler Syntax::
28102 * A Simple Example of Inline Assembler::
28103 * Output Variables in Inline Assembler::
28104 * Input Variables in Inline Assembler::
28105 * Inlining Inline Assembler Code::
28106 * Other Asm Functionality::
28109 @c ---------------------------------------------------------------------------
28110 @node Basic Assembler Syntax
28111 @section Basic Assembler Syntax
28114 The assembler used by GNAT and gcc is based not on the Intel assembly
28115 language, but rather on a language that descends from the AT&T Unix
28116 assembler @emph{as} (and which is often referred to as ``AT&T syntax'').
28117 The following table summarizes the main features of @emph{as} syntax
28118 and points out the differences from the Intel conventions.
28119 See the gcc @emph{as} and @emph{gas} (an @emph{as} macro
28120 pre-processor) documentation for further information.
28123 @item Register names
28124 gcc / @emph{as}: Prefix with ``%''; for example @code{%eax}
28126 Intel: No extra punctuation; for example @code{eax}
28128 @item Immediate operand
28129 gcc / @emph{as}: Prefix with ``$''; for example @code{$4}
28131 Intel: No extra punctuation; for example @code{4}
28134 gcc / @emph{as}: Prefix with ``$''; for example @code{$loc}
28136 Intel: No extra punctuation; for example @code{loc}
28138 @item Memory contents
28139 gcc / @emph{as}: No extra punctuation; for example @code{loc}
28141 Intel: Square brackets; for example @code{[loc]}
28143 @item Register contents
28144 gcc / @emph{as}: Parentheses; for example @code{(%eax)}
28146 Intel: Square brackets; for example @code{[eax]}
28148 @item Hexadecimal numbers
28149 gcc / @emph{as}: Leading ``0x'' (C language syntax); for example @code{0xA0}
28151 Intel: Trailing ``h''; for example @code{A0h}
28154 gcc / @emph{as}: Explicit in op code; for example @code{movw} to move
28157 Intel: Implicit, deduced by assembler; for example @code{mov}
28159 @item Instruction repetition
28160 gcc / @emph{as}: Split into two lines; for example
28166 Intel: Keep on one line; for example @code{rep stosl}
28168 @item Order of operands
28169 gcc / @emph{as}: Source first; for example @code{movw $4, %eax}
28171 Intel: Destination first; for example @code{mov eax, 4}
28174 @c ---------------------------------------------------------------------------
28175 @node A Simple Example of Inline Assembler
28176 @section A Simple Example of Inline Assembler
28179 The following example will generate a single assembly language statement,
28180 @code{nop}, which does nothing. Despite its lack of run-time effect,
28181 the example will be useful in illustrating the basics of
28182 the Inline Assembler facility.
28184 @smallexample @c ada
28186 with System.Machine_Code; use System.Machine_Code;
28187 procedure Nothing is
28194 @code{Asm} is a procedure declared in package @code{System.Machine_Code};
28195 here it takes one parameter, a @emph{template string} that must be a static
28196 expression and that will form the generated instruction.
28197 @code{Asm} may be regarded as a compile-time procedure that parses
28198 the template string and additional parameters (none here),
28199 from which it generates a sequence of assembly language instructions.
28201 The examples in this chapter will illustrate several of the forms
28202 for invoking @code{Asm}; a complete specification of the syntax
28203 is found in @ref{Machine Code Insertions,,, gnat_rm, GNAT Reference
28206 Under the standard GNAT conventions, the @code{Nothing} procedure
28207 should be in a file named @file{nothing.adb}.
28208 You can build the executable in the usual way:
28212 However, the interesting aspect of this example is not its run-time behavior
28213 but rather the generated assembly code.
28214 To see this output, invoke the compiler as follows:
28216 gcc -c -S -fomit-frame-pointer -gnatp @file{nothing.adb}
28218 where the options are:
28222 compile only (no bind or link)
28224 generate assembler listing
28225 @item -fomit-frame-pointer
28226 do not set up separate stack frames
28228 do not add runtime checks
28231 This gives a human-readable assembler version of the code. The resulting
28232 file will have the same name as the Ada source file, but with a @code{.s}
28233 extension. In our example, the file @file{nothing.s} has the following
28238 .file "nothing.adb"
28240 ___gnu_compiled_ada:
28243 .globl __ada_nothing
28255 The assembly code you included is clearly indicated by
28256 the compiler, between the @code{#APP} and @code{#NO_APP}
28257 delimiters. The character before the 'APP' and 'NOAPP'
28258 can differ on different targets. For example, GNU/Linux uses '#APP' while
28259 on NT you will see '/APP'.
28261 If you make a mistake in your assembler code (such as using the
28262 wrong size modifier, or using a wrong operand for the instruction) GNAT
28263 will report this error in a temporary file, which will be deleted when
28264 the compilation is finished. Generating an assembler file will help
28265 in such cases, since you can assemble this file separately using the
28266 @emph{as} assembler that comes with gcc.
28268 Assembling the file using the command
28271 as @file{nothing.s}
28274 will give you error messages whose lines correspond to the assembler
28275 input file, so you can easily find and correct any mistakes you made.
28276 If there are no errors, @emph{as} will generate an object file
28277 @file{nothing.out}.
28279 @c ---------------------------------------------------------------------------
28280 @node Output Variables in Inline Assembler
28281 @section Output Variables in Inline Assembler
28284 The examples in this section, showing how to access the processor flags,
28285 illustrate how to specify the destination operands for assembly language
28288 @smallexample @c ada
28290 with Interfaces; use Interfaces;
28291 with Ada.Text_IO; use Ada.Text_IO;
28292 with System.Machine_Code; use System.Machine_Code;
28293 procedure Get_Flags is
28294 Flags : Unsigned_32;
28297 Asm ("pushfl" & LF & HT & -- push flags on stack
28298 "popl %%eax" & LF & HT & -- load eax with flags
28299 "movl %%eax, %0", -- store flags in variable
28300 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
28301 Put_Line ("Flags register:" & Flags'Img);
28306 In order to have a nicely aligned assembly listing, we have separated
28307 multiple assembler statements in the Asm template string with linefeed
28308 (ASCII.LF) and horizontal tab (ASCII.HT) characters.
28309 The resulting section of the assembly output file is:
28316 movl %eax, -40(%ebp)
28321 It would have been legal to write the Asm invocation as:
28324 Asm ("pushfl popl %%eax movl %%eax, %0")
28327 but in the generated assembler file, this would come out as:
28331 pushfl popl %eax movl %eax, -40(%ebp)
28335 which is not so convenient for the human reader.
28337 We use Ada comments
28338 at the end of each line to explain what the assembler instructions
28339 actually do. This is a useful convention.
28341 When writing Inline Assembler instructions, you need to precede each register
28342 and variable name with a percent sign. Since the assembler already requires
28343 a percent sign at the beginning of a register name, you need two consecutive
28344 percent signs for such names in the Asm template string, thus @code{%%eax}.
28345 In the generated assembly code, one of the percent signs will be stripped off.
28347 Names such as @code{%0}, @code{%1}, @code{%2}, etc., denote input or output
28348 variables: operands you later define using @code{Input} or @code{Output}
28349 parameters to @code{Asm}.
28350 An output variable is illustrated in
28351 the third statement in the Asm template string:
28355 The intent is to store the contents of the eax register in a variable that can
28356 be accessed in Ada. Simply writing @code{movl %%eax, Flags} would not
28357 necessarily work, since the compiler might optimize by using a register
28358 to hold Flags, and the expansion of the @code{movl} instruction would not be
28359 aware of this optimization. The solution is not to store the result directly
28360 but rather to advise the compiler to choose the correct operand form;
28361 that is the purpose of the @code{%0} output variable.
28363 Information about the output variable is supplied in the @code{Outputs}
28364 parameter to @code{Asm}:
28366 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
28369 The output is defined by the @code{Asm_Output} attribute of the target type;
28370 the general format is
28372 Type'Asm_Output (constraint_string, variable_name)
28375 The constraint string directs the compiler how
28376 to store/access the associated variable. In the example
28378 Unsigned_32'Asm_Output ("=m", Flags);
28380 the @code{"m"} (memory) constraint tells the compiler that the variable
28381 @code{Flags} should be stored in a memory variable, thus preventing
28382 the optimizer from keeping it in a register. In contrast,
28384 Unsigned_32'Asm_Output ("=r", Flags);
28386 uses the @code{"r"} (register) constraint, telling the compiler to
28387 store the variable in a register.
28389 If the constraint is preceded by the equal character (@strong{=}), it tells
28390 the compiler that the variable will be used to store data into it.
28392 In the @code{Get_Flags} example, we used the @code{"g"} (global) constraint,
28393 allowing the optimizer to choose whatever it deems best.
28395 There are a fairly large number of constraints, but the ones that are
28396 most useful (for the Intel x86 processor) are the following:
28402 global (i.e.@: can be stored anywhere)
28420 use one of eax, ebx, ecx or edx
28422 use one of eax, ebx, ecx, edx, esi or edi
28425 The full set of constraints is described in the gcc and @emph{as}
28426 documentation; note that it is possible to combine certain constraints
28427 in one constraint string.
28429 You specify the association of an output variable with an assembler operand
28430 through the @code{%}@emph{n} notation, where @emph{n} is a non-negative
28432 @smallexample @c ada
28434 Asm ("pushfl" & LF & HT & -- push flags on stack
28435 "popl %%eax" & LF & HT & -- load eax with flags
28436 "movl %%eax, %0", -- store flags in variable
28437 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
28441 @code{%0} will be replaced in the expanded code by the appropriate operand,
28443 the compiler decided for the @code{Flags} variable.
28445 In general, you may have any number of output variables:
28448 Count the operands starting at 0; thus @code{%0}, @code{%1}, etc.
28450 Specify the @code{Outputs} parameter as a parenthesized comma-separated list
28451 of @code{Asm_Output} attributes
28455 @smallexample @c ada
28457 Asm ("movl %%eax, %0" & LF & HT &
28458 "movl %%ebx, %1" & LF & HT &
28460 Outputs => (Unsigned_32'Asm_Output ("=g", Var_A), -- %0 = Var_A
28461 Unsigned_32'Asm_Output ("=g", Var_B), -- %1 = Var_B
28462 Unsigned_32'Asm_Output ("=g", Var_C))); -- %2 = Var_C
28466 where @code{Var_A}, @code{Var_B}, and @code{Var_C} are variables
28467 in the Ada program.
28469 As a variation on the @code{Get_Flags} example, we can use the constraints
28470 string to direct the compiler to store the eax register into the @code{Flags}
28471 variable, instead of including the store instruction explicitly in the
28472 @code{Asm} template string:
28474 @smallexample @c ada
28476 with Interfaces; use Interfaces;
28477 with Ada.Text_IO; use Ada.Text_IO;
28478 with System.Machine_Code; use System.Machine_Code;
28479 procedure Get_Flags_2 is
28480 Flags : Unsigned_32;
28483 Asm ("pushfl" & LF & HT & -- push flags on stack
28484 "popl %%eax", -- save flags in eax
28485 Outputs => Unsigned_32'Asm_Output ("=a", Flags));
28486 Put_Line ("Flags register:" & Flags'Img);
28492 The @code{"a"} constraint tells the compiler that the @code{Flags}
28493 variable will come from the eax register. Here is the resulting code:
28501 movl %eax,-40(%ebp)
28506 The compiler generated the store of eax into Flags after
28507 expanding the assembler code.
28509 Actually, there was no need to pop the flags into the eax register;
28510 more simply, we could just pop the flags directly into the program variable:
28512 @smallexample @c ada
28514 with Interfaces; use Interfaces;
28515 with Ada.Text_IO; use Ada.Text_IO;
28516 with System.Machine_Code; use System.Machine_Code;
28517 procedure Get_Flags_3 is
28518 Flags : Unsigned_32;
28521 Asm ("pushfl" & LF & HT & -- push flags on stack
28522 "pop %0", -- save flags in Flags
28523 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
28524 Put_Line ("Flags register:" & Flags'Img);
28529 @c ---------------------------------------------------------------------------
28530 @node Input Variables in Inline Assembler
28531 @section Input Variables in Inline Assembler
28534 The example in this section illustrates how to specify the source operands
28535 for assembly language statements.
28536 The program simply increments its input value by 1:
28538 @smallexample @c ada
28540 with Interfaces; use Interfaces;
28541 with Ada.Text_IO; use Ada.Text_IO;
28542 with System.Machine_Code; use System.Machine_Code;
28543 procedure Increment is
28545 function Incr (Value : Unsigned_32) return Unsigned_32 is
28546 Result : Unsigned_32;
28549 Outputs => Unsigned_32'Asm_Output ("=a", Result),
28550 Inputs => Unsigned_32'Asm_Input ("a", Value));
28554 Value : Unsigned_32;
28558 Put_Line ("Value before is" & Value'Img);
28559 Value := Incr (Value);
28560 Put_Line ("Value after is" & Value'Img);
28565 The @code{Outputs} parameter to @code{Asm} specifies
28566 that the result will be in the eax register and that it is to be stored
28567 in the @code{Result} variable.
28569 The @code{Inputs} parameter looks much like the @code{Outputs} parameter,
28570 but with an @code{Asm_Input} attribute.
28571 The @code{"="} constraint, indicating an output value, is not present.
28573 You can have multiple input variables, in the same way that you can have more
28574 than one output variable.
28576 The parameter count (%0, %1) etc, still starts at the first output statement,
28577 and continues with the input statements.
28579 Just as the @code{Outputs} parameter causes the register to be stored into the
28580 target variable after execution of the assembler statements, so does the
28581 @code{Inputs} parameter cause its variable to be loaded into the register
28582 before execution of the assembler statements.
28584 Thus the effect of the @code{Asm} invocation is:
28586 @item load the 32-bit value of @code{Value} into eax
28587 @item execute the @code{incl %eax} instruction
28588 @item store the contents of eax into the @code{Result} variable
28591 The resulting assembler file (with @option{-O2} optimization) contains:
28594 _increment__incr.1:
28607 @c ---------------------------------------------------------------------------
28608 @node Inlining Inline Assembler Code
28609 @section Inlining Inline Assembler Code
28612 For a short subprogram such as the @code{Incr} function in the previous
28613 section, the overhead of the call and return (creating / deleting the stack
28614 frame) can be significant, compared to the amount of code in the subprogram
28615 body. A solution is to apply Ada's @code{Inline} pragma to the subprogram,
28616 which directs the compiler to expand invocations of the subprogram at the
28617 point(s) of call, instead of setting up a stack frame for out-of-line calls.
28618 Here is the resulting program:
28620 @smallexample @c ada
28622 with Interfaces; use Interfaces;
28623 with Ada.Text_IO; use Ada.Text_IO;
28624 with System.Machine_Code; use System.Machine_Code;
28625 procedure Increment_2 is
28627 function Incr (Value : Unsigned_32) return Unsigned_32 is
28628 Result : Unsigned_32;
28631 Outputs => Unsigned_32'Asm_Output ("=a", Result),
28632 Inputs => Unsigned_32'Asm_Input ("a", Value));
28635 pragma Inline (Increment);
28637 Value : Unsigned_32;
28641 Put_Line ("Value before is" & Value'Img);
28642 Value := Increment (Value);
28643 Put_Line ("Value after is" & Value'Img);
28648 Compile the program with both optimization (@option{-O2}) and inlining
28649 (@option{-gnatn}) enabled.
28651 The @code{Incr} function is still compiled as usual, but at the
28652 point in @code{Increment} where our function used to be called:
28657 call _increment__incr.1
28662 the code for the function body directly appears:
28675 thus saving the overhead of stack frame setup and an out-of-line call.
28677 @c ---------------------------------------------------------------------------
28678 @node Other Asm Functionality
28679 @section Other @code{Asm} Functionality
28682 This section describes two important parameters to the @code{Asm}
28683 procedure: @code{Clobber}, which identifies register usage;
28684 and @code{Volatile}, which inhibits unwanted optimizations.
28687 * The Clobber Parameter::
28688 * The Volatile Parameter::
28691 @c ---------------------------------------------------------------------------
28692 @node The Clobber Parameter
28693 @subsection The @code{Clobber} Parameter
28696 One of the dangers of intermixing assembly language and a compiled language
28697 such as Ada is that the compiler needs to be aware of which registers are
28698 being used by the assembly code. In some cases, such as the earlier examples,
28699 the constraint string is sufficient to indicate register usage (e.g.,
28701 the eax register). But more generally, the compiler needs an explicit
28702 identification of the registers that are used by the Inline Assembly
28705 Using a register that the compiler doesn't know about
28706 could be a side effect of an instruction (like @code{mull}
28707 storing its result in both eax and edx).
28708 It can also arise from explicit register usage in your
28709 assembly code; for example:
28712 Asm ("movl %0, %%ebx" & LF & HT &
28714 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
28715 Inputs => Unsigned_32'Asm_Input ("g", Var_In));
28719 where the compiler (since it does not analyze the @code{Asm} template string)
28720 does not know you are using the ebx register.
28722 In such cases you need to supply the @code{Clobber} parameter to @code{Asm},
28723 to identify the registers that will be used by your assembly code:
28727 Asm ("movl %0, %%ebx" & LF & HT &
28729 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
28730 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
28735 The Clobber parameter is a static string expression specifying the
28736 register(s) you are using. Note that register names are @emph{not} prefixed
28737 by a percent sign. Also, if more than one register is used then their names
28738 are separated by commas; e.g., @code{"eax, ebx"}
28740 The @code{Clobber} parameter has several additional uses:
28742 @item Use ``register'' name @code{cc} to indicate that flags might have changed
28743 @item Use ``register'' name @code{memory} if you changed a memory location
28746 @c ---------------------------------------------------------------------------
28747 @node The Volatile Parameter
28748 @subsection The @code{Volatile} Parameter
28749 @cindex Volatile parameter
28752 Compiler optimizations in the presence of Inline Assembler may sometimes have
28753 unwanted effects. For example, when an @code{Asm} invocation with an input
28754 variable is inside a loop, the compiler might move the loading of the input
28755 variable outside the loop, regarding it as a one-time initialization.
28757 If this effect is not desired, you can disable such optimizations by setting
28758 the @code{Volatile} parameter to @code{True}; for example:
28760 @smallexample @c ada
28762 Asm ("movl %0, %%ebx" & LF & HT &
28764 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
28765 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
28771 By default, @code{Volatile} is set to @code{False} unless there is no
28772 @code{Outputs} parameter.
28774 Although setting @code{Volatile} to @code{True} prevents unwanted
28775 optimizations, it will also disable other optimizations that might be
28776 important for efficiency. In general, you should set @code{Volatile}
28777 to @code{True} only if the compiler's optimizations have created
28779 @c END OF INLINE ASSEMBLER CHAPTER
28780 @c ===============================
28782 @c ***********************************
28783 @c * Compatibility and Porting Guide *
28784 @c ***********************************
28785 @node Compatibility and Porting Guide
28786 @appendix Compatibility and Porting Guide
28789 This chapter describes the compatibility issues that may arise between
28790 GNAT and other Ada compilation systems (including those for Ada 83),
28791 and shows how GNAT can expedite porting
28792 applications developed in other Ada environments.
28795 * Compatibility with Ada 83::
28796 * Compatibility between Ada 95 and Ada 2005::
28797 * Implementation-dependent characteristics::
28798 * Compatibility with Other Ada Systems::
28799 * Representation Clauses::
28801 @c Brief section is only in non-VMS version
28802 @c Full chapter is in VMS version
28803 * Compatibility with HP Ada 83::
28806 * Transitioning to 64-Bit GNAT for OpenVMS::
28810 @node Compatibility with Ada 83
28811 @section Compatibility with Ada 83
28812 @cindex Compatibility (between Ada 83 and Ada 95 / Ada 2005)
28815 Ada 95 and Ada 2005 are highly upwards compatible with Ada 83. In
28816 particular, the design intention was that the difficulties associated
28817 with moving from Ada 83 to Ada 95 or Ada 2005 should be no greater than those
28818 that occur when moving from one Ada 83 system to another.
28820 However, there are a number of points at which there are minor
28821 incompatibilities. The @cite{Ada 95 Annotated Reference Manual} contains
28822 full details of these issues,
28823 and should be consulted for a complete treatment.
28825 following subsections treat the most likely issues to be encountered.
28828 * Legal Ada 83 programs that are illegal in Ada 95::
28829 * More deterministic semantics::
28830 * Changed semantics::
28831 * Other language compatibility issues::
28834 @node Legal Ada 83 programs that are illegal in Ada 95
28835 @subsection Legal Ada 83 programs that are illegal in Ada 95
28837 Some legal Ada 83 programs are illegal (i.e., they will fail to compile) in
28838 Ada 95 and thus also in Ada 2005:
28841 @item Character literals
28842 Some uses of character literals are ambiguous. Since Ada 95 has introduced
28843 @code{Wide_Character} as a new predefined character type, some uses of
28844 character literals that were legal in Ada 83 are illegal in Ada 95.
28846 @smallexample @c ada
28847 for Char in 'A' .. 'Z' loop @dots{} end loop;
28851 The problem is that @code{'A'} and @code{'Z'} could be from either
28852 @code{Character} or @code{Wide_Character}. The simplest correction
28853 is to make the type explicit; e.g.:
28854 @smallexample @c ada
28855 for Char in Character range 'A' .. 'Z' loop @dots{} end loop;
28858 @item New reserved words
28859 The identifiers @code{abstract}, @code{aliased}, @code{protected},
28860 @code{requeue}, @code{tagged}, and @code{until} are reserved in Ada 95.
28861 Existing Ada 83 code using any of these identifiers must be edited to
28862 use some alternative name.
28864 @item Freezing rules
28865 The rules in Ada 95 are slightly different with regard to the point at
28866 which entities are frozen, and representation pragmas and clauses are
28867 not permitted past the freeze point. This shows up most typically in
28868 the form of an error message complaining that a representation item
28869 appears too late, and the appropriate corrective action is to move
28870 the item nearer to the declaration of the entity to which it refers.
28872 A particular case is that representation pragmas
28875 extended HP Ada 83 compatibility pragmas such as @code{Export_Procedure})
28877 cannot be applied to a subprogram body. If necessary, a separate subprogram
28878 declaration must be introduced to which the pragma can be applied.
28880 @item Optional bodies for library packages
28881 In Ada 83, a package that did not require a package body was nevertheless
28882 allowed to have one. This lead to certain surprises in compiling large
28883 systems (situations in which the body could be unexpectedly ignored by the
28884 binder). In Ada 95, if a package does not require a body then it is not
28885 permitted to have a body. To fix this problem, simply remove a redundant
28886 body if it is empty, or, if it is non-empty, introduce a dummy declaration
28887 into the spec that makes the body required. One approach is to add a private
28888 part to the package declaration (if necessary), and define a parameterless
28889 procedure called @code{Requires_Body}, which must then be given a dummy
28890 procedure body in the package body, which then becomes required.
28891 Another approach (assuming that this does not introduce elaboration
28892 circularities) is to add an @code{Elaborate_Body} pragma to the package spec,
28893 since one effect of this pragma is to require the presence of a package body.
28895 @item @code{Numeric_Error} is now the same as @code{Constraint_Error}
28896 In Ada 95, the exception @code{Numeric_Error} is a renaming of
28897 @code{Constraint_Error}.
28898 This means that it is illegal to have separate exception handlers for
28899 the two exceptions. The fix is simply to remove the handler for the
28900 @code{Numeric_Error} case (since even in Ada 83, a compiler was free to raise
28901 @code{Constraint_Error} in place of @code{Numeric_Error} in all cases).
28903 @item Indefinite subtypes in generics
28904 In Ada 83, it was permissible to pass an indefinite type (e.g.@: @code{String})
28905 as the actual for a generic formal private type, but then the instantiation
28906 would be illegal if there were any instances of declarations of variables
28907 of this type in the generic body. In Ada 95, to avoid this clear violation
28908 of the methodological principle known as the ``contract model'',
28909 the generic declaration explicitly indicates whether
28910 or not such instantiations are permitted. If a generic formal parameter
28911 has explicit unknown discriminants, indicated by using @code{(<>)} after the
28912 subtype name, then it can be instantiated with indefinite types, but no
28913 stand-alone variables can be declared of this type. Any attempt to declare
28914 such a variable will result in an illegality at the time the generic is
28915 declared. If the @code{(<>)} notation is not used, then it is illegal
28916 to instantiate the generic with an indefinite type.
28917 This is the potential incompatibility issue when porting Ada 83 code to Ada 95.
28918 It will show up as a compile time error, and
28919 the fix is usually simply to add the @code{(<>)} to the generic declaration.
28922 @node More deterministic semantics
28923 @subsection More deterministic semantics
28927 Conversions from real types to integer types round away from 0. In Ada 83
28928 the conversion Integer(2.5) could deliver either 2 or 3 as its value. This
28929 implementation freedom was intended to support unbiased rounding in
28930 statistical applications, but in practice it interfered with portability.
28931 In Ada 95 the conversion semantics are unambiguous, and rounding away from 0
28932 is required. Numeric code may be affected by this change in semantics.
28933 Note, though, that this issue is no worse than already existed in Ada 83
28934 when porting code from one vendor to another.
28937 The Real-Time Annex introduces a set of policies that define the behavior of
28938 features that were implementation dependent in Ada 83, such as the order in
28939 which open select branches are executed.
28942 @node Changed semantics
28943 @subsection Changed semantics
28946 The worst kind of incompatibility is one where a program that is legal in
28947 Ada 83 is also legal in Ada 95 but can have an effect in Ada 95 that was not
28948 possible in Ada 83. Fortunately this is extremely rare, but the one
28949 situation that you should be alert to is the change in the predefined type
28950 @code{Character} from 7-bit ASCII to 8-bit Latin-1.
28953 @item Range of type @code{Character}
28954 The range of @code{Standard.Character} is now the full 256 characters
28955 of Latin-1, whereas in most Ada 83 implementations it was restricted
28956 to 128 characters. Although some of the effects of
28957 this change will be manifest in compile-time rejection of legal
28958 Ada 83 programs it is possible for a working Ada 83 program to have
28959 a different effect in Ada 95, one that was not permitted in Ada 83.
28960 As an example, the expression
28961 @code{Character'Pos(Character'Last)} returned @code{127} in Ada 83 and now
28962 delivers @code{255} as its value.
28963 In general, you should look at the logic of any
28964 character-processing Ada 83 program and see whether it needs to be adapted
28965 to work correctly with Latin-1. Note that the predefined Ada 95 API has a
28966 character handling package that may be relevant if code needs to be adapted
28967 to account for the additional Latin-1 elements.
28968 The desirable fix is to
28969 modify the program to accommodate the full character set, but in some cases
28970 it may be convenient to define a subtype or derived type of Character that
28971 covers only the restricted range.
28975 @node Other language compatibility issues
28976 @subsection Other language compatibility issues
28979 @item @option{-gnat83} switch
28980 All implementations of GNAT provide a switch that causes GNAT to operate
28981 in Ada 83 mode. In this mode, some but not all compatibility problems
28982 of the type described above are handled automatically. For example, the
28983 new reserved words introduced in Ada 95 and Ada 2005 are treated simply
28984 as identifiers as in Ada 83.
28986 in practice, it is usually advisable to make the necessary modifications
28987 to the program to remove the need for using this switch.
28988 See @ref{Compiling Different Versions of Ada}.
28990 @item Support for removed Ada 83 pragmas and attributes
28991 A number of pragmas and attributes from Ada 83 were removed from Ada 95,
28992 generally because they were replaced by other mechanisms. Ada 95 and Ada 2005
28993 compilers are allowed, but not required, to implement these missing
28994 elements. In contrast with some other compilers, GNAT implements all
28995 such pragmas and attributes, eliminating this compatibility concern. These
28996 include @code{pragma Interface} and the floating point type attributes
28997 (@code{Emax}, @code{Mantissa}, etc.), among other items.
29001 @node Compatibility between Ada 95 and Ada 2005
29002 @section Compatibility between Ada 95 and Ada 2005
29003 @cindex Compatibility between Ada 95 and Ada 2005
29006 Although Ada 2005 was designed to be upwards compatible with Ada 95, there are
29007 a number of incompatibilities. Several are enumerated below;
29008 for a complete description please see the
29009 Annotated Ada 2005 Reference Manual, or section 9.1.1 in
29010 @cite{Rationale for Ada 2005}.
29013 @item New reserved words.
29014 The words @code{interface}, @code{overriding} and @code{synchronized} are
29015 reserved in Ada 2005.
29016 A pre-Ada 2005 program that uses any of these as an identifier will be
29019 @item New declarations in predefined packages.
29020 A number of packages in the predefined environment contain new declarations:
29021 @code{Ada.Exceptions}, @code{Ada.Real_Time}, @code{Ada.Strings},
29022 @code{Ada.Strings.Fixed}, @code{Ada.Strings.Bounded},
29023 @code{Ada.Strings.Unbounded}, @code{Ada.Strings.Wide_Fixed},
29024 @code{Ada.Strings.Wide_Bounded}, @code{Ada.Strings.Wide_Unbounded},
29025 @code{Ada.Tags}, @code{Ada.Text_IO}, and @code{Interfaces.C}.
29026 If an Ada 95 program does a @code{with} and @code{use} of any of these
29027 packages, the new declarations may cause name clashes.
29029 @item Access parameters.
29030 A nondispatching subprogram with an access parameter cannot be renamed
29031 as a dispatching operation. This was permitted in Ada 95.
29033 @item Access types, discriminants, and constraints.
29034 Rule changes in this area have led to some incompatibilities; for example,
29035 constrained subtypes of some access types are not permitted in Ada 2005.
29037 @item Aggregates for limited types.
29038 The allowance of aggregates for limited types in Ada 2005 raises the
29039 possibility of ambiguities in legal Ada 95 programs, since additional types
29040 now need to be considered in expression resolution.
29042 @item Fixed-point multiplication and division.
29043 Certain expressions involving ``*'' or ``/'' for a fixed-point type, which
29044 were legal in Ada 95 and invoked the predefined versions of these operations,
29046 The ambiguity may be resolved either by applying a type conversion to the
29047 expression, or by explicitly invoking the operation from package
29050 @item Return-by-reference types.
29051 The Ada 95 return-by-reference mechanism has been removed. Instead, the user
29052 can declare a function returning a value from an anonymous access type.
29056 @node Implementation-dependent characteristics
29057 @section Implementation-dependent characteristics
29059 Although the Ada language defines the semantics of each construct as
29060 precisely as practical, in some situations (for example for reasons of
29061 efficiency, or where the effect is heavily dependent on the host or target
29062 platform) the implementation is allowed some freedom. In porting Ada 83
29063 code to GNAT, you need to be aware of whether / how the existing code
29064 exercised such implementation dependencies. Such characteristics fall into
29065 several categories, and GNAT offers specific support in assisting the
29066 transition from certain Ada 83 compilers.
29069 * Implementation-defined pragmas::
29070 * Implementation-defined attributes::
29072 * Elaboration order::
29073 * Target-specific aspects::
29076 @node Implementation-defined pragmas
29077 @subsection Implementation-defined pragmas
29080 Ada compilers are allowed to supplement the language-defined pragmas, and
29081 these are a potential source of non-portability. All GNAT-defined pragmas
29082 are described in @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT
29083 Reference Manual}, and these include several that are specifically
29084 intended to correspond to other vendors' Ada 83 pragmas.
29085 For migrating from VADS, the pragma @code{Use_VADS_Size} may be useful.
29086 For compatibility with HP Ada 83, GNAT supplies the pragmas
29087 @code{Extend_System}, @code{Ident}, @code{Inline_Generic},
29088 @code{Interface_Name}, @code{Passive}, @code{Suppress_All},
29089 and @code{Volatile}.
29090 Other relevant pragmas include @code{External} and @code{Link_With}.
29091 Some vendor-specific
29092 Ada 83 pragmas (@code{Share_Generic}, @code{Subtitle}, and @code{Title}) are
29094 avoiding compiler rejection of units that contain such pragmas; they are not
29095 relevant in a GNAT context and hence are not otherwise implemented.
29097 @node Implementation-defined attributes
29098 @subsection Implementation-defined attributes
29100 Analogous to pragmas, the set of attributes may be extended by an
29101 implementation. All GNAT-defined attributes are described in
29102 @ref{Implementation Defined Attributes,,, gnat_rm, GNAT Reference
29103 Manual}, and these include several that are specifically intended
29104 to correspond to other vendors' Ada 83 attributes. For migrating from VADS,
29105 the attribute @code{VADS_Size} may be useful. For compatibility with HP
29106 Ada 83, GNAT supplies the attributes @code{Bit}, @code{Machine_Size} and
29110 @subsection Libraries
29112 Vendors may supply libraries to supplement the standard Ada API. If Ada 83
29113 code uses vendor-specific libraries then there are several ways to manage
29114 this in Ada 95 or Ada 2005:
29117 If the source code for the libraries (specs and bodies) are
29118 available, then the libraries can be migrated in the same way as the
29121 If the source code for the specs but not the bodies are
29122 available, then you can reimplement the bodies.
29124 Some features introduced by Ada 95 obviate the need for library support. For
29125 example most Ada 83 vendors supplied a package for unsigned integers. The
29126 Ada 95 modular type feature is the preferred way to handle this need, so
29127 instead of migrating or reimplementing the unsigned integer package it may
29128 be preferable to retrofit the application using modular types.
29131 @node Elaboration order
29132 @subsection Elaboration order
29134 The implementation can choose any elaboration order consistent with the unit
29135 dependency relationship. This freedom means that some orders can result in
29136 Program_Error being raised due to an ``Access Before Elaboration'': an attempt
29137 to invoke a subprogram its body has been elaborated, or to instantiate a
29138 generic before the generic body has been elaborated. By default GNAT
29139 attempts to choose a safe order (one that will not encounter access before
29140 elaboration problems) by implicitly inserting @code{Elaborate} or
29141 @code{Elaborate_All} pragmas where
29142 needed. However, this can lead to the creation of elaboration circularities
29143 and a resulting rejection of the program by gnatbind. This issue is
29144 thoroughly described in @ref{Elaboration Order Handling in GNAT}.
29145 In brief, there are several
29146 ways to deal with this situation:
29150 Modify the program to eliminate the circularities, e.g.@: by moving
29151 elaboration-time code into explicitly-invoked procedures
29153 Constrain the elaboration order by including explicit @code{Elaborate_Body} or
29154 @code{Elaborate} pragmas, and then inhibit the generation of implicit
29155 @code{Elaborate_All}
29156 pragmas either globally (as an effect of the @option{-gnatE} switch) or locally
29157 (by selectively suppressing elaboration checks via pragma
29158 @code{Suppress(Elaboration_Check)} when it is safe to do so).
29161 @node Target-specific aspects
29162 @subsection Target-specific aspects
29164 Low-level applications need to deal with machine addresses, data
29165 representations, interfacing with assembler code, and similar issues. If
29166 such an Ada 83 application is being ported to different target hardware (for
29167 example where the byte endianness has changed) then you will need to
29168 carefully examine the program logic; the porting effort will heavily depend
29169 on the robustness of the original design. Moreover, Ada 95 (and thus
29170 Ada 2005) are sometimes
29171 incompatible with typical Ada 83 compiler practices regarding implicit
29172 packing, the meaning of the Size attribute, and the size of access values.
29173 GNAT's approach to these issues is described in @ref{Representation Clauses}.
29175 @node Compatibility with Other Ada Systems
29176 @section Compatibility with Other Ada Systems
29179 If programs avoid the use of implementation dependent and
29180 implementation defined features, as documented in the @cite{Ada
29181 Reference Manual}, there should be a high degree of portability between
29182 GNAT and other Ada systems. The following are specific items which
29183 have proved troublesome in moving Ada 95 programs from GNAT to other Ada 95
29184 compilers, but do not affect porting code to GNAT@.
29185 (As of @value{NOW}, GNAT is the only compiler available for Ada 2005;
29186 the following issues may or may not arise for Ada 2005 programs
29187 when other compilers appear.)
29190 @item Ada 83 Pragmas and Attributes
29191 Ada 95 compilers are allowed, but not required, to implement the missing
29192 Ada 83 pragmas and attributes that are no longer defined in Ada 95.
29193 GNAT implements all such pragmas and attributes, eliminating this as
29194 a compatibility concern, but some other Ada 95 compilers reject these
29195 pragmas and attributes.
29197 @item Specialized Needs Annexes
29198 GNAT implements the full set of special needs annexes. At the
29199 current time, it is the only Ada 95 compiler to do so. This means that
29200 programs making use of these features may not be portable to other Ada
29201 95 compilation systems.
29203 @item Representation Clauses
29204 Some other Ada 95 compilers implement only the minimal set of
29205 representation clauses required by the Ada 95 reference manual. GNAT goes
29206 far beyond this minimal set, as described in the next section.
29209 @node Representation Clauses
29210 @section Representation Clauses
29213 The Ada 83 reference manual was quite vague in describing both the minimal
29214 required implementation of representation clauses, and also their precise
29215 effects. Ada 95 (and thus also Ada 2005) are much more explicit, but the
29216 minimal set of capabilities required is still quite limited.
29218 GNAT implements the full required set of capabilities in
29219 Ada 95 and Ada 2005, but also goes much further, and in particular
29220 an effort has been made to be compatible with existing Ada 83 usage to the
29221 greatest extent possible.
29223 A few cases exist in which Ada 83 compiler behavior is incompatible with
29224 the requirements in Ada 95 (and thus also Ada 2005). These are instances of
29225 intentional or accidental dependence on specific implementation dependent
29226 characteristics of these Ada 83 compilers. The following is a list of
29227 the cases most likely to arise in existing Ada 83 code.
29230 @item Implicit Packing
29231 Some Ada 83 compilers allowed a Size specification to cause implicit
29232 packing of an array or record. This could cause expensive implicit
29233 conversions for change of representation in the presence of derived
29234 types, and the Ada design intends to avoid this possibility.
29235 Subsequent AI's were issued to make it clear that such implicit
29236 change of representation in response to a Size clause is inadvisable,
29237 and this recommendation is represented explicitly in the Ada 95 (and Ada 2005)
29238 Reference Manuals as implementation advice that is followed by GNAT@.
29239 The problem will show up as an error
29240 message rejecting the size clause. The fix is simply to provide
29241 the explicit pragma @code{Pack}, or for more fine tuned control, provide
29242 a Component_Size clause.
29244 @item Meaning of Size Attribute
29245 The Size attribute in Ada 95 (and Ada 2005) for discrete types is defined as
29246 the minimal number of bits required to hold values of the type. For example,
29247 on a 32-bit machine, the size of @code{Natural} will typically be 31 and not
29248 32 (since no sign bit is required). Some Ada 83 compilers gave 31, and
29249 some 32 in this situation. This problem will usually show up as a compile
29250 time error, but not always. It is a good idea to check all uses of the
29251 'Size attribute when porting Ada 83 code. The GNAT specific attribute
29252 Object_Size can provide a useful way of duplicating the behavior of
29253 some Ada 83 compiler systems.
29255 @item Size of Access Types
29256 A common assumption in Ada 83 code is that an access type is in fact a pointer,
29257 and that therefore it will be the same size as a System.Address value. This
29258 assumption is true for GNAT in most cases with one exception. For the case of
29259 a pointer to an unconstrained array type (where the bounds may vary from one
29260 value of the access type to another), the default is to use a ``fat pointer'',
29261 which is represented as two separate pointers, one to the bounds, and one to
29262 the array. This representation has a number of advantages, including improved
29263 efficiency. However, it may cause some difficulties in porting existing Ada 83
29264 code which makes the assumption that, for example, pointers fit in 32 bits on
29265 a machine with 32-bit addressing.
29267 To get around this problem, GNAT also permits the use of ``thin pointers'' for
29268 access types in this case (where the designated type is an unconstrained array
29269 type). These thin pointers are indeed the same size as a System.Address value.
29270 To specify a thin pointer, use a size clause for the type, for example:
29272 @smallexample @c ada
29273 type X is access all String;
29274 for X'Size use Standard'Address_Size;
29278 which will cause the type X to be represented using a single pointer.
29279 When using this representation, the bounds are right behind the array.
29280 This representation is slightly less efficient, and does not allow quite
29281 such flexibility in the use of foreign pointers or in using the
29282 Unrestricted_Access attribute to create pointers to non-aliased objects.
29283 But for any standard portable use of the access type it will work in
29284 a functionally correct manner and allow porting of existing code.
29285 Note that another way of forcing a thin pointer representation
29286 is to use a component size clause for the element size in an array,
29287 or a record representation clause for an access field in a record.
29289 See the documentation of Unrestricted_Access in the GNAT RM for a
29290 full discussion of possible problems using this attribute in conjunction
29291 with thin pointers.
29295 @c This brief section is only in the non-VMS version
29296 @c The complete chapter on HP Ada is in the VMS version
29297 @node Compatibility with HP Ada 83
29298 @section Compatibility with HP Ada 83
29301 The VMS version of GNAT fully implements all the pragmas and attributes
29302 provided by HP Ada 83, as well as providing the standard HP Ada 83
29303 libraries, including Starlet. In addition, data layouts and parameter
29304 passing conventions are highly compatible. This means that porting
29305 existing HP Ada 83 code to GNAT in VMS systems should be easier than
29306 most other porting efforts. The following are some of the most
29307 significant differences between GNAT and HP Ada 83.
29310 @item Default floating-point representation
29311 In GNAT, the default floating-point format is IEEE, whereas in HP Ada 83,
29312 it is VMS format. GNAT does implement the necessary pragmas
29313 (Long_Float, Float_Representation) for changing this default.
29316 The package System in GNAT exactly corresponds to the definition in the
29317 Ada 95 reference manual, which means that it excludes many of the
29318 HP Ada 83 extensions. However, a separate package Aux_DEC is provided
29319 that contains the additional definitions, and a special pragma,
29320 Extend_System allows this package to be treated transparently as an
29321 extension of package System.
29324 The definitions provided by Aux_DEC are exactly compatible with those
29325 in the HP Ada 83 version of System, with one exception.
29326 HP Ada provides the following declarations:
29328 @smallexample @c ada
29329 TO_ADDRESS (INTEGER)
29330 TO_ADDRESS (UNSIGNED_LONGWORD)
29331 TO_ADDRESS (@i{universal_integer})
29335 The version of TO_ADDRESS taking a @i{universal integer} argument is in fact
29336 an extension to Ada 83 not strictly compatible with the reference manual.
29337 In GNAT, we are constrained to be exactly compatible with the standard,
29338 and this means we cannot provide this capability. In HP Ada 83, the
29339 point of this definition is to deal with a call like:
29341 @smallexample @c ada
29342 TO_ADDRESS (16#12777#);
29346 Normally, according to the Ada 83 standard, one would expect this to be
29347 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
29348 of TO_ADDRESS@. However, in HP Ada 83, there is no ambiguity, since the
29349 definition using @i{universal_integer} takes precedence.
29351 In GNAT, since the version with @i{universal_integer} cannot be supplied, it
29352 is not possible to be 100% compatible. Since there are many programs using
29353 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
29354 to change the name of the function in the UNSIGNED_LONGWORD case, so the
29355 declarations provided in the GNAT version of AUX_Dec are:
29357 @smallexample @c ada
29358 function To_Address (X : Integer) return Address;
29359 pragma Pure_Function (To_Address);
29361 function To_Address_Long (X : Unsigned_Longword)
29363 pragma Pure_Function (To_Address_Long);
29367 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
29368 change the name to TO_ADDRESS_LONG@.
29370 @item Task_Id values
29371 The Task_Id values assigned will be different in the two systems, and GNAT
29372 does not provide a specified value for the Task_Id of the environment task,
29373 which in GNAT is treated like any other declared task.
29377 For full details on these and other less significant compatibility issues,
29378 see appendix E of the HP publication entitled @cite{HP Ada, Technical
29379 Overview and Comparison on HP Platforms}.
29381 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
29382 attributes are recognized, although only a subset of them can sensibly
29383 be implemented. The description of pragmas in @ref{Implementation
29384 Defined Pragmas,,, gnat_rm, GNAT Reference Manual}
29385 indicates whether or not they are applicable to non-VMS systems.
29389 @node Transitioning to 64-Bit GNAT for OpenVMS
29390 @section Transitioning to 64-Bit @value{EDITION} for OpenVMS
29393 This section is meant to assist users of pre-2006 @value{EDITION}
29394 for Alpha OpenVMS who are transitioning to 64-bit @value{EDITION},
29395 the version of the GNAT technology supplied in 2006 and later for
29396 OpenVMS on both Alpha and I64.
29399 * Introduction to transitioning::
29400 * Migration of 32 bit code::
29401 * Taking advantage of 64 bit addressing::
29402 * Technical details::
29405 @node Introduction to transitioning
29406 @subsection Introduction
29409 64-bit @value{EDITION} for Open VMS has been designed to meet
29414 Providing a full conforming implementation of Ada 95 and Ada 2005
29417 Allowing maximum backward compatibility, thus easing migration of existing
29421 Supplying a path for exploiting the full 64-bit address range
29425 Ada's strong typing semantics has made it
29426 impractical to have different 32-bit and 64-bit modes. As soon as
29427 one object could possibly be outside the 32-bit address space, this
29428 would make it necessary for the @code{System.Address} type to be 64 bits.
29429 In particular, this would cause inconsistencies if 32-bit code is
29430 called from 64-bit code that raises an exception.
29432 This issue has been resolved by always using 64-bit addressing
29433 at the system level, but allowing for automatic conversions between
29434 32-bit and 64-bit addresses where required. Thus users who
29435 do not currently require 64-bit addressing capabilities, can
29436 recompile their code with only minimal changes (and indeed
29437 if the code is written in portable Ada, with no assumptions about
29438 the size of the @code{Address} type, then no changes at all are necessary).
29440 this approach provides a simple, gradual upgrade path to future
29441 use of larger memories than available for 32-bit systems.
29442 Also, newly written applications or libraries will by default
29443 be fully compatible with future systems exploiting 64-bit
29444 addressing capabilities.
29446 @ref{Migration of 32 bit code}, will focus on porting applications
29447 that do not require more than 2 GB of
29448 addressable memory. This code will be referred to as
29449 @emph{32-bit code}.
29450 For applications intending to exploit the full 64-bit address space,
29451 @ref{Taking advantage of 64 bit addressing},
29452 will consider further changes that may be required.
29453 Such code will be referred to below as @emph{64-bit code}.
29455 @node Migration of 32 bit code
29456 @subsection Migration of 32-bit code
29460 * Access types and 32/64-bit allocation::
29461 * Unchecked conversions::
29462 * Predefined constants::
29463 * Interfacing with C::
29464 * 32/64-bit descriptors::
29465 * Experience with source compatibility::
29468 @node Address types
29469 @subsubsection Address types
29472 To solve the problem of mixing 64-bit and 32-bit addressing,
29473 while maintaining maximum backward compatibility, the following
29474 approach has been taken:
29478 @code{System.Address} always has a size of 64 bits
29479 @cindex @code{System.Address} size
29480 @cindex @code{Address} size
29483 @code{System.Short_Address} is a 32-bit subtype of @code{System.Address}
29484 @cindex @code{System.Short_Address} size
29485 @cindex @code{Short_Address} size
29489 Since @code{System.Short_Address} is a subtype of @code{System.Address},
29490 a @code{Short_Address}
29491 may be used where an @code{Address} is required, and vice versa, without
29492 needing explicit type conversions.
29493 By virtue of the Open VMS parameter passing conventions,
29495 and exported subprograms that have 32-bit address parameters are
29496 compatible with those that have 64-bit address parameters.
29497 (See @ref{Making code 64 bit clean} for details.)
29499 The areas that may need attention are those where record types have
29500 been defined that contain components of the type @code{System.Address}, and
29501 where objects of this type are passed to code expecting a record layout with
29504 Different compilers on different platforms cannot be
29505 expected to represent the same type in the same way,
29506 since alignment constraints
29507 and other system-dependent properties affect the compiler's decision.
29508 For that reason, Ada code
29509 generally uses representation clauses to specify the expected
29510 layout where required.
29512 If such a representation clause uses 32 bits for a component having
29513 the type @code{System.Address}, 64-bit @value{EDITION} for OpenVMS
29514 will detect that error and produce a specific diagnostic message.
29515 The developer should then determine whether the representation
29516 should be 64 bits or not and make either of two changes:
29517 change the size to 64 bits and leave the type as @code{System.Address}, or
29518 leave the size as 32 bits and change the type to @code{System.Short_Address}.
29519 Since @code{Short_Address} is a subtype of @code{Address}, no changes are
29520 required in any code setting or accessing the field; the compiler will
29521 automatically perform any needed conversions between address
29524 @node Access types and 32/64-bit allocation
29525 @subsubsection Access types and 32/64-bit allocation
29526 @cindex 32-bit allocation
29527 @cindex 64-bit allocation
29530 By default, objects designated by access values are always allocated in
29531 the 64-bit address space, and access values themselves are represented
29532 in 64 bits. If these defaults are not appropriate, and 32-bit allocation
29533 is required (for example if the address of an allocated object is assigned
29534 to a @code{Short_Address} variable), then several alternatives are available:
29538 A pool-specific access type (ie, an @w{Ada 83} access type, whose
29539 definition is @code{access T} versus @code{access all T} or
29540 @code{access constant T}), may be declared with a @code{'Size} representation
29541 clause that establishes the size as 32 bits.
29542 In such circumstances allocations for that type will
29543 be from the 32-bit heap. Such a clause is not permitted
29544 for a general access type (declared with @code{access all} or
29545 @code{access constant}) as values of such types must be able to refer
29546 to any object of the designated type, including objects residing outside
29547 the 32-bit address range. Existing @w{Ada 83} code will not contain such
29548 type definitions, however, since general access types were introduced
29552 Switches for @command{GNAT BIND} control whether the internal GNAT
29553 allocation routine @code{__gnat_malloc} uses 64-bit or 32-bit allocations.
29554 @cindex @code{__gnat_malloc}
29555 The switches are respectively @option{-H64} (the default) and
29557 @cindex @option{-H32} (@command{gnatbind})
29558 @cindex @option{-H64} (@command{gnatbind})
29561 The environment variable (logical name) @code{GNAT$NO_MALLOC_64}
29562 @cindex @code{GNAT$NO_MALLOC_64} environment variable
29563 may be used to force @code{__gnat_malloc} to use 32-bit allocation.
29564 If this variable is left
29565 undefined, or defined as @code{"DISABLE"}, @code{"FALSE"}, or @code{"0"},
29566 then the default (64-bit) allocation is used.
29567 If defined as @code{"ENABLE"}, @code{"TRUE"}, or @code{"1"},
29568 then 32-bit allocation is used. The gnatbind qualifiers described above
29569 override this logical name.
29572 A ^gcc switch^gcc switch^ for OpenVMS, @option{-mno-malloc64}, operates
29573 @cindex @option{-mno-malloc64} (^gcc^gcc^)
29574 at a low level to convert explicit calls to @code{malloc} and related
29575 functions from the C run-time library so that they perform allocations
29576 in the 32-bit heap.
29577 Since all internal allocations from GNAT use @code{__gnat_malloc},
29578 this switch is not required unless the program makes explicit calls on
29579 @code{malloc} (or related functions) from interfaced C code.
29583 @node Unchecked conversions
29584 @subsubsection Unchecked conversions
29587 In the case of an @code{Unchecked_Conversion} where the source type is a
29588 64-bit access type or the type @code{System.Address}, and the target
29589 type is a 32-bit type, the compiler will generate a warning.
29590 Even though the generated code will still perform the required
29591 conversions, it is highly recommended in these cases to use
29592 respectively a 32-bit access type or @code{System.Short_Address}
29593 as the source type.
29595 @node Predefined constants
29596 @subsubsection Predefined constants
29599 The following table shows the correspondence between pre-2006 versions of
29600 @value{EDITION} on Alpha OpenVMS (``Old'') and 64-bit @value{EDITION}
29603 @multitable {@code{System.Short_Memory_Size}} {2**32} {2**64}
29604 @item @b{Constant} @tab @b{Old} @tab @b{New}
29605 @item @code{System.Word_Size} @tab 32 @tab 64
29606 @item @code{System.Memory_Size} @tab 2**32 @tab 2**64
29607 @item @code{System.Short_Memory_Size} @tab 2**32 @tab 2**32
29608 @item @code{System.Address_Size} @tab 32 @tab 64
29612 If you need to refer to the specific
29613 memory size of a 32-bit implementation, instead of the
29614 actual memory size, use @code{System.Short_Memory_Size}
29615 rather than @code{System.Memory_Size}.
29616 Similarly, references to @code{System.Address_Size} may need
29617 to be replaced by @code{System.Short_Address'Size}.
29618 The program @command{gnatfind} may be useful for locating
29619 references to the above constants, so that you can verify that they
29622 @node Interfacing with C
29623 @subsubsection Interfacing with C
29626 In order to minimize the impact of the transition to 64-bit addresses on
29627 legacy programs, some fundamental types in the @code{Interfaces.C}
29628 package hierarchy continue to be represented in 32 bits.
29629 These types are: @code{ptrdiff_t}, @code{size_t}, and @code{chars_ptr}.
29630 This eases integration with the default HP C layout choices, for example
29631 as found in the system routines in @code{DECC$SHR.EXE}.
29632 Because of this implementation choice, the type fully compatible with
29633 @code{chars_ptr} is now @code{Short_Address} and not @code{Address}.
29634 Depending on the context the compiler will issue a
29635 warning or an error when type @code{Address} is used, alerting the user to a
29636 potential problem. Otherwise 32-bit programs that use
29637 @code{Interfaces.C} should normally not require code modifications
29639 The other issue arising with C interfacing concerns pragma @code{Convention}.
29640 For VMS 64-bit systems, there is an issue of the appropriate default size
29641 of C convention pointers in the absence of an explicit size clause. The HP
29642 C compiler can choose either 32 or 64 bits depending on compiler options.
29643 GNAT chooses 32-bits rather than 64-bits in the default case where no size
29644 clause is given. This proves a better choice for porting 32-bit legacy
29645 applications. In order to have a 64-bit representation, it is necessary to
29646 specify a size representation clause. For example:
29648 @smallexample @c ada
29649 type int_star is access Interfaces.C.int;
29650 pragma Convention(C, int_star);
29651 for int_star'Size use 64; -- Necessary to get 64 and not 32 bits
29654 @node 32/64-bit descriptors
29655 @subsubsection 32/64-bit descriptors
29658 By default, GNAT uses a 64-bit descriptor mechanism. For an imported
29659 subprogram (i.e., a subprogram identified by pragma @code{Import_Function},
29660 @code{Import_Procedure}, or @code{Import_Valued_Procedure}) that specifies
29661 @code{Short_Descriptor} as its mechanism, a 32-bit descriptor is used.
29662 @cindex @code{Short_Descriptor} mechanism for imported subprograms
29664 If the configuration pragma @code{Short_Descriptors} is supplied, then
29665 all descriptors will be 32 bits.
29666 @cindex pragma @code{Short_Descriptors}
29668 @node Experience with source compatibility
29669 @subsubsection Experience with source compatibility
29672 The Security Server and STARLET on I64 provide an interesting ``test case''
29673 for source compatibility issues, since it is in such system code
29674 where assumptions about @code{Address} size might be expected to occur.
29675 Indeed, there were a small number of occasions in the Security Server
29676 file @file{jibdef.ads}
29677 where a representation clause for a record type specified
29678 32 bits for a component of type @code{Address}.
29679 All of these errors were detected by the compiler.
29680 The repair was obvious and immediate; to simply replace @code{Address} by
29681 @code{Short_Address}.
29683 In the case of STARLET, there were several record types that should
29684 have had representation clauses but did not. In these record types
29685 there was an implicit assumption that an @code{Address} value occupied
29687 These compiled without error, but their usage resulted in run-time error
29688 returns from STARLET system calls.
29689 Future GNAT technology enhancements may include a tool that detects and flags
29690 these sorts of potential source code porting problems.
29692 @c ****************************************
29693 @node Taking advantage of 64 bit addressing
29694 @subsection Taking advantage of 64-bit addressing
29697 * Making code 64 bit clean::
29698 * Allocating memory from the 64 bit storage pool::
29699 * Restrictions on use of 64 bit objects::
29700 * STARLET and other predefined libraries::
29703 @node Making code 64 bit clean
29704 @subsubsection Making code 64-bit clean
29707 In order to prevent problems that may occur when (parts of) a
29708 system start using memory outside the 32-bit address range,
29709 we recommend some additional guidelines:
29713 For imported subprograms that take parameters of the
29714 type @code{System.Address}, ensure that these subprograms can
29715 indeed handle 64-bit addresses. If not, or when in doubt,
29716 change the subprogram declaration to specify
29717 @code{System.Short_Address} instead.
29720 Resolve all warnings related to size mismatches in
29721 unchecked conversions. Failing to do so causes
29722 erroneous execution if the source object is outside
29723 the 32-bit address space.
29726 (optional) Explicitly use the 32-bit storage pool
29727 for access types used in a 32-bit context, or use
29728 generic access types where possible
29729 (@pxref{Restrictions on use of 64 bit objects}).
29733 If these rules are followed, the compiler will automatically insert
29734 any necessary checks to ensure that no addresses or access values
29735 passed to 32-bit code ever refer to objects outside the 32-bit
29737 Any attempt to do this will raise @code{Constraint_Error}.
29739 @node Allocating memory from the 64 bit storage pool
29740 @subsubsection Allocating memory from the 64-bit storage pool
29743 By default, all allocations -- for both pool-specific and general
29744 access types -- use the 64-bit storage pool. To override
29745 this default, for an individual access type or globally, see
29746 @ref{Access types and 32/64-bit allocation}.
29748 @node Restrictions on use of 64 bit objects
29749 @subsubsection Restrictions on use of 64-bit objects
29752 Taking the address of an object allocated from a 64-bit storage pool,
29753 and then passing this address to a subprogram expecting
29754 @code{System.Short_Address},
29755 or assigning it to a variable of type @code{Short_Address}, will cause
29756 @code{Constraint_Error} to be raised. In case the code is not 64-bit clean
29757 (@pxref{Making code 64 bit clean}), or checks are suppressed,
29758 no exception is raised and execution
29759 will become erroneous.
29761 @node STARLET and other predefined libraries
29762 @subsubsection STARLET and other predefined libraries
29765 All code that comes as part of GNAT is 64-bit clean, but the
29766 restrictions given in @ref{Restrictions on use of 64 bit objects},
29767 still apply. Look at the package
29768 specs to see in which contexts objects allocated
29769 in 64-bit address space are acceptable.
29771 @node Technical details
29772 @subsection Technical details
29775 64-bit @value{EDITION} for Open VMS takes advantage of the freedom given in the
29776 Ada standard with respect to the type of @code{System.Address}. Previous
29777 versions of @value{EDITION} have defined this type as private and implemented it as a
29780 In order to allow defining @code{System.Short_Address} as a proper subtype,
29781 and to match the implicit sign extension in parameter passing,
29782 in 64-bit @value{EDITION} for Open VMS, @code{System.Address} is defined as a
29783 visible (i.e., non-private) integer type.
29784 Standard operations on the type, such as the binary operators ``+'', ``-'',
29785 etc., that take @code{Address} operands and return an @code{Address} result,
29786 have been hidden by declaring these
29787 @code{abstract}, a feature introduced in Ada 95 that helps avoid the potential
29788 ambiguities that would otherwise result from overloading.
29789 (Note that, although @code{Address} is a visible integer type,
29790 good programming practice dictates against exploiting the type's
29791 integer properties such as literals, since this will compromise
29794 Defining @code{Address} as a visible integer type helps achieve
29795 maximum compatibility for existing Ada code,
29796 without sacrificing the capabilities of the 64-bit architecture.
29799 @c ************************************************
29800 @node Microsoft Windows Topics
29801 @appendix Microsoft Windows Topics
29807 This chapter describes topics that are specific to the Microsoft Windows
29808 platforms (NT, 2000, and XP Professional).
29811 @ifclear FSFEDITION
29812 * Installing from the Command Line::
29814 * Using GNAT on Windows::
29815 * Using a network installation of GNAT::
29816 * CONSOLE and WINDOWS subsystems::
29817 * Temporary Files::
29818 * Mixed-Language Programming on Windows::
29819 * Windows Calling Conventions::
29820 * Introduction to Dynamic Link Libraries (DLLs)::
29821 * Using DLLs with GNAT::
29822 * Building DLLs with GNAT Project files::
29823 * Building DLLs with GNAT::
29824 * Building DLLs with gnatdll::
29825 * GNAT and Windows Resources::
29826 * Debugging a DLL::
29827 * Setting Stack Size from gnatlink::
29828 * Setting Heap Size from gnatlink::
29831 @ifclear FSFEDITION
29832 @node Installing from the Command Line
29833 @section Installing from the Command Line
29834 @cindex Batch installation
29835 @cindex Silent installation
29836 @cindex Unassisted installation
29839 By default the @value{EDITION} installers display a GUI that prompts the user
29840 to enter installation path and similar information, and guide him through the
29841 installation process. It is also possible to perform silent installations
29842 using the command-line interface.
29844 In order to install one of the @value{EDITION} installers from the command
29845 line you should pass parameter @code{/S} (and, optionally,
29846 @code{/D=<directory>}) as command-line arguments.
29849 For example, for an unattended installation of
29850 @value{EDITION} 7.0.2 into the default directory
29851 @code{C:\GNATPRO\7.0.2} you would run:
29854 gnatpro-7.0.2-i686-pc-mingw32-bin.exe /S
29857 To install into a custom directory, say, @code{C:\TOOLS\GNATPRO\7.0.2}:
29860 gnatpro-7.0.2-i686-pc-mingw32-bin /S /D=C:\TOOLS\GNATPRO\7.0.2
29865 For example, for an unattended installation of
29866 @value{EDITION} 2012 into @code{C:\GNAT\2012}:
29869 gnat-gpl-2012-i686-pc-mingw32-bin /S /D=C:\GNAT\2012
29873 You can use the same syntax for all installers.
29875 Note that unattended installations don't modify system path, nor create file
29876 associations, so such activities need to be done by hand.
29879 @node Using GNAT on Windows
29880 @section Using GNAT on Windows
29883 One of the strengths of the GNAT technology is that its tool set
29884 (@command{gcc}, @command{gnatbind}, @command{gnatlink}, @command{gnatmake}, the
29885 @code{gdb} debugger, etc.) is used in the same way regardless of the
29888 On Windows this tool set is complemented by a number of Microsoft-specific
29889 tools that have been provided to facilitate interoperability with Windows
29890 when this is required. With these tools:
29895 You can build applications using the @code{CONSOLE} or @code{WINDOWS}
29899 You can use any Dynamically Linked Library (DLL) in your Ada code (both
29900 relocatable and non-relocatable DLLs are supported).
29903 You can build Ada DLLs for use in other applications. These applications
29904 can be written in a language other than Ada (e.g., C, C++, etc). Again both
29905 relocatable and non-relocatable Ada DLLs are supported.
29908 You can include Windows resources in your Ada application.
29911 You can use or create COM/DCOM objects.
29915 Immediately below are listed all known general GNAT-for-Windows restrictions.
29916 Other restrictions about specific features like Windows Resources and DLLs
29917 are listed in separate sections below.
29922 It is not possible to use @code{GetLastError} and @code{SetLastError}
29923 when tasking, protected records, or exceptions are used. In these
29924 cases, in order to implement Ada semantics, the GNAT run-time system
29925 calls certain Win32 routines that set the last error variable to 0 upon
29926 success. It should be possible to use @code{GetLastError} and
29927 @code{SetLastError} when tasking, protected record, and exception
29928 features are not used, but it is not guaranteed to work.
29931 It is not possible to link against Microsoft C++ libraries except for
29932 import libraries. Interfacing must be done by the mean of DLLs.
29935 It is possible to link against Microsoft C libraries. Yet the preferred
29936 solution is to use C/C++ compiler that comes with @value{EDITION}, since it
29937 doesn't require having two different development environments and makes the
29938 inter-language debugging experience smoother.
29941 When the compilation environment is located on FAT32 drives, users may
29942 experience recompilations of the source files that have not changed if
29943 Daylight Saving Time (DST) state has changed since the last time files
29944 were compiled. NTFS drives do not have this problem.
29947 No components of the GNAT toolset use any entries in the Windows
29948 registry. The only entries that can be created are file associations and
29949 PATH settings, provided the user has chosen to create them at installation
29950 time, as well as some minimal book-keeping information needed to correctly
29951 uninstall or integrate different GNAT products.
29954 @node Using a network installation of GNAT
29955 @section Using a network installation of GNAT
29958 Make sure the system on which GNAT is installed is accessible from the
29959 current machine, i.e., the install location is shared over the network.
29960 Shared resources are accessed on Windows by means of UNC paths, which
29961 have the format @code{\\server\sharename\path}
29963 In order to use such a network installation, simply add the UNC path of the
29964 @file{bin} directory of your GNAT installation in front of your PATH. For
29965 example, if GNAT is installed in @file{\GNAT} directory of a share location
29966 called @file{c-drive} on a machine @file{LOKI}, the following command will
29969 @code{@ @ @ path \\loki\c-drive\gnat\bin;%path%}
29971 Be aware that every compilation using the network installation results in the
29972 transfer of large amounts of data across the network and will likely cause
29973 serious performance penalty.
29975 @node CONSOLE and WINDOWS subsystems
29976 @section CONSOLE and WINDOWS subsystems
29977 @cindex CONSOLE Subsystem
29978 @cindex WINDOWS Subsystem
29982 There are two main subsystems under Windows. The @code{CONSOLE} subsystem
29983 (which is the default subsystem) will always create a console when
29984 launching the application. This is not something desirable when the
29985 application has a Windows GUI. To get rid of this console the
29986 application must be using the @code{WINDOWS} subsystem. To do so
29987 the @option{-mwindows} linker option must be specified.
29990 $ gnatmake winprog -largs -mwindows
29993 @node Temporary Files
29994 @section Temporary Files
29995 @cindex Temporary files
29998 It is possible to control where temporary files gets created by setting
29999 the @env{TMP} environment variable. The file will be created:
30002 @item Under the directory pointed to by the @env{TMP} environment variable if
30003 this directory exists.
30005 @item Under @file{c:\temp}, if the @env{TMP} environment variable is not
30006 set (or not pointing to a directory) and if this directory exists.
30008 @item Under the current working directory otherwise.
30012 This allows you to determine exactly where the temporary
30013 file will be created. This is particularly useful in networked
30014 environments where you may not have write access to some
30017 @node Mixed-Language Programming on Windows
30018 @section Mixed-Language Programming on Windows
30021 Developing pure Ada applications on Windows is no different than on
30022 other GNAT-supported platforms. However, when developing or porting an
30023 application that contains a mix of Ada and C/C++, the choice of your
30024 Windows C/C++ development environment conditions your overall
30025 interoperability strategy.
30027 If you use @command{gcc} or Microsoft C to compile the non-Ada part of
30028 your application, there are no Windows-specific restrictions that
30029 affect the overall interoperability with your Ada code. If you do want
30030 to use the Microsoft tools for your C++ code, you have two choices:
30034 Encapsulate your C++ code in a DLL to be linked with your Ada
30035 application. In this case, use the Microsoft or whatever environment to
30036 build the DLL and use GNAT to build your executable
30037 (@pxref{Using DLLs with GNAT}).
30040 Or you can encapsulate your Ada code in a DLL to be linked with the
30041 other part of your application. In this case, use GNAT to build the DLL
30042 (@pxref{Building DLLs with GNAT Project files}) and use the Microsoft
30043 or whatever environment to build your executable.
30046 In addition to the description about C main in
30047 @pxref{Mixed Language Programming} section, if the C main uses a
30048 stand-alone library it is required on x86-windows to
30049 setup the SEH context. For this the C main must looks like this:
30053 extern void adainit (void);
30054 extern void adafinal (void);
30055 extern void __gnat_initialize(void*);
30056 extern void call_to_ada (void);
30058 int main (int argc, char *argv[])
30062 /* Initialize the SEH context */
30063 __gnat_initialize (&SEH);
30067 /* Then call Ada services in the stand-alone library */
30075 Note that this is not needed on x86_64-windows where the Windows
30076 native SEH support is used.
30078 @node Windows Calling Conventions
30079 @section Windows Calling Conventions
30083 This section pertain only to Win32. On Win64 there is a single native
30084 calling convention. All convention specifiers are ignored on this
30088 * C Calling Convention::
30089 * Stdcall Calling Convention::
30090 * Win32 Calling Convention::
30091 * DLL Calling Convention::
30095 When a subprogram @code{F} (caller) calls a subprogram @code{G}
30096 (callee), there are several ways to push @code{G}'s parameters on the
30097 stack and there are several possible scenarios to clean up the stack
30098 upon @code{G}'s return. A calling convention is an agreed upon software
30099 protocol whereby the responsibilities between the caller (@code{F}) and
30100 the callee (@code{G}) are clearly defined. Several calling conventions
30101 are available for Windows:
30105 @code{C} (Microsoft defined)
30108 @code{Stdcall} (Microsoft defined)
30111 @code{Win32} (GNAT specific)
30114 @code{DLL} (GNAT specific)
30117 @node C Calling Convention
30118 @subsection @code{C} Calling Convention
30121 This is the default calling convention used when interfacing to C/C++
30122 routines compiled with either @command{gcc} or Microsoft Visual C++.
30124 In the @code{C} calling convention subprogram parameters are pushed on the
30125 stack by the caller from right to left. The caller itself is in charge of
30126 cleaning up the stack after the call. In addition, the name of a routine
30127 with @code{C} calling convention is mangled by adding a leading underscore.
30129 The name to use on the Ada side when importing (or exporting) a routine
30130 with @code{C} calling convention is the name of the routine. For
30131 instance the C function:
30134 int get_val (long);
30138 should be imported from Ada as follows:
30140 @smallexample @c ada
30142 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
30143 pragma Import (C, Get_Val, External_Name => "get_val");
30148 Note that in this particular case the @code{External_Name} parameter could
30149 have been omitted since, when missing, this parameter is taken to be the
30150 name of the Ada entity in lower case. When the @code{Link_Name} parameter
30151 is missing, as in the above example, this parameter is set to be the
30152 @code{External_Name} with a leading underscore.
30154 When importing a variable defined in C, you should always use the @code{C}
30155 calling convention unless the object containing the variable is part of a
30156 DLL (in which case you should use the @code{Stdcall} calling
30157 convention, @pxref{Stdcall Calling Convention}).
30159 @node Stdcall Calling Convention
30160 @subsection @code{Stdcall} Calling Convention
30163 This convention, which was the calling convention used for Pascal
30164 programs, is used by Microsoft for all the routines in the Win32 API for
30165 efficiency reasons. It must be used to import any routine for which this
30166 convention was specified.
30168 In the @code{Stdcall} calling convention subprogram parameters are pushed
30169 on the stack by the caller from right to left. The callee (and not the
30170 caller) is in charge of cleaning the stack on routine exit. In addition,
30171 the name of a routine with @code{Stdcall} calling convention is mangled by
30172 adding a leading underscore (as for the @code{C} calling convention) and a
30173 trailing @code{@@}@code{@var{nn}}, where @var{nn} is the overall size (in
30174 bytes) of the parameters passed to the routine.
30176 The name to use on the Ada side when importing a C routine with a
30177 @code{Stdcall} calling convention is the name of the C routine. The leading
30178 underscore and trailing @code{@@}@code{@var{nn}} are added automatically by
30179 the compiler. For instance the Win32 function:
30182 @b{APIENTRY} int get_val (long);
30186 should be imported from Ada as follows:
30188 @smallexample @c ada
30190 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
30191 pragma Import (Stdcall, Get_Val);
30192 -- On the x86 a long is 4 bytes, so the Link_Name is "_get_val@@4"
30197 As for the @code{C} calling convention, when the @code{External_Name}
30198 parameter is missing, it is taken to be the name of the Ada entity in lower
30199 case. If instead of writing the above import pragma you write:
30201 @smallexample @c ada
30203 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
30204 pragma Import (Stdcall, Get_Val, External_Name => "retrieve_val");
30209 then the imported routine is @code{_retrieve_val@@4}. However, if instead
30210 of specifying the @code{External_Name} parameter you specify the
30211 @code{Link_Name} as in the following example:
30213 @smallexample @c ada
30215 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
30216 pragma Import (Stdcall, Get_Val, Link_Name => "retrieve_val");
30221 then the imported routine is @code{retrieve_val}, that is, there is no
30222 decoration at all. No leading underscore and no Stdcall suffix
30223 @code{@@}@code{@var{nn}}.
30226 This is especially important as in some special cases a DLL's entry
30227 point name lacks a trailing @code{@@}@code{@var{nn}} while the exported
30228 name generated for a call has it.
30231 It is also possible to import variables defined in a DLL by using an
30232 import pragma for a variable. As an example, if a DLL contains a
30233 variable defined as:
30240 then, to access this variable from Ada you should write:
30242 @smallexample @c ada
30244 My_Var : Interfaces.C.int;
30245 pragma Import (Stdcall, My_Var);
30250 Note that to ease building cross-platform bindings this convention
30251 will be handled as a @code{C} calling convention on non-Windows platforms.
30253 @node Win32 Calling Convention
30254 @subsection @code{Win32} Calling Convention
30257 This convention, which is GNAT-specific is fully equivalent to the
30258 @code{Stdcall} calling convention described above.
30260 @node DLL Calling Convention
30261 @subsection @code{DLL} Calling Convention
30264 This convention, which is GNAT-specific is fully equivalent to the
30265 @code{Stdcall} calling convention described above.
30267 @node Introduction to Dynamic Link Libraries (DLLs)
30268 @section Introduction to Dynamic Link Libraries (DLLs)
30272 A Dynamically Linked Library (DLL) is a library that can be shared by
30273 several applications running under Windows. A DLL can contain any number of
30274 routines and variables.
30276 One advantage of DLLs is that you can change and enhance them without
30277 forcing all the applications that depend on them to be relinked or
30278 recompiled. However, you should be aware than all calls to DLL routines are
30279 slower since, as you will understand below, such calls are indirect.
30281 To illustrate the remainder of this section, suppose that an application
30282 wants to use the services of a DLL @file{API.dll}. To use the services
30283 provided by @file{API.dll} you must statically link against the DLL or
30284 an import library which contains a jump table with an entry for each
30285 routine and variable exported by the DLL. In the Microsoft world this
30286 import library is called @file{API.lib}. When using GNAT this import
30287 library is called either @file{libAPI.dll.a}, @file{libapi.dll.a},
30288 @file{libAPI.a} or @file{libapi.a} (names are case insensitive).
30290 After you have linked your application with the DLL or the import library
30291 and you run your application, here is what happens:
30295 Your application is loaded into memory.
30298 The DLL @file{API.dll} is mapped into the address space of your
30299 application. This means that:
30303 The DLL will use the stack of the calling thread.
30306 The DLL will use the virtual address space of the calling process.
30309 The DLL will allocate memory from the virtual address space of the calling
30313 Handles (pointers) can be safely exchanged between routines in the DLL
30314 routines and routines in the application using the DLL.
30318 The entries in the jump table (from the import library @file{libAPI.dll.a}
30319 or @file{API.lib} or automatically created when linking against a DLL)
30320 which is part of your application are initialized with the addresses
30321 of the routines and variables in @file{API.dll}.
30324 If present in @file{API.dll}, routines @code{DllMain} or
30325 @code{DllMainCRTStartup} are invoked. These routines typically contain
30326 the initialization code needed for the well-being of the routines and
30327 variables exported by the DLL.
30331 There is an additional point which is worth mentioning. In the Windows
30332 world there are two kind of DLLs: relocatable and non-relocatable
30333 DLLs. Non-relocatable DLLs can only be loaded at a very specific address
30334 in the target application address space. If the addresses of two
30335 non-relocatable DLLs overlap and these happen to be used by the same
30336 application, a conflict will occur and the application will run
30337 incorrectly. Hence, when possible, it is always preferable to use and
30338 build relocatable DLLs. Both relocatable and non-relocatable DLLs are
30339 supported by GNAT. Note that the @option{-s} linker option (see GNU Linker
30340 User's Guide) removes the debugging symbols from the DLL but the DLL can
30341 still be relocated.
30343 As a side note, an interesting difference between Microsoft DLLs and
30344 Unix shared libraries, is the fact that on most Unix systems all public
30345 routines are exported by default in a Unix shared library, while under
30346 Windows it is possible (but not required) to list exported routines in
30347 a definition file (@pxref{The Definition File}).
30349 @node Using DLLs with GNAT
30350 @section Using DLLs with GNAT
30353 * Creating an Ada Spec for the DLL Services::
30354 * Creating an Import Library::
30358 To use the services of a DLL, say @file{API.dll}, in your Ada application
30363 The Ada spec for the routines and/or variables you want to access in
30364 @file{API.dll}. If not available this Ada spec must be built from the C/C++
30365 header files provided with the DLL.
30368 The import library (@file{libAPI.dll.a} or @file{API.lib}). As previously
30369 mentioned an import library is a statically linked library containing the
30370 import table which will be filled at load time to point to the actual
30371 @file{API.dll} routines. Sometimes you don't have an import library for the
30372 DLL you want to use. The following sections will explain how to build
30373 one. Note that this is optional.
30376 The actual DLL, @file{API.dll}.
30380 Once you have all the above, to compile an Ada application that uses the
30381 services of @file{API.dll} and whose main subprogram is @code{My_Ada_App},
30382 you simply issue the command
30385 $ gnatmake my_ada_app -largs -lAPI
30389 The argument @option{-largs -lAPI} at the end of the @command{gnatmake} command
30390 tells the GNAT linker to look for an import library. The linker will
30391 look for a library name in this specific order:
30394 @item @file{libAPI.dll.a}
30395 @item @file{API.dll.a}
30396 @item @file{libAPI.a}
30397 @item @file{API.lib}
30398 @item @file{libAPI.dll}
30399 @item @file{API.dll}
30402 The first three are the GNU style import libraries. The third is the
30403 Microsoft style import libraries. The last two are the actual DLL names.
30405 Note that if the Ada package spec for @file{API.dll} contains the
30408 @smallexample @c ada
30409 pragma Linker_Options ("-lAPI");
30413 you do not have to add @option{-largs -lAPI} at the end of the
30414 @command{gnatmake} command.
30416 If any one of the items above is missing you will have to create it
30417 yourself. The following sections explain how to do so using as an
30418 example a fictitious DLL called @file{API.dll}.
30420 @node Creating an Ada Spec for the DLL Services
30421 @subsection Creating an Ada Spec for the DLL Services
30424 A DLL typically comes with a C/C++ header file which provides the
30425 definitions of the routines and variables exported by the DLL. The Ada
30426 equivalent of this header file is a package spec that contains definitions
30427 for the imported entities. If the DLL you intend to use does not come with
30428 an Ada spec you have to generate one such spec yourself. For example if
30429 the header file of @file{API.dll} is a file @file{api.h} containing the
30430 following two definitions:
30442 then the equivalent Ada spec could be:
30444 @smallexample @c ada
30447 with Interfaces.C.Strings;
30452 function Get (Str : C.Strings.Chars_Ptr) return C.int;
30455 pragma Import (C, Get);
30456 pragma Import (DLL, Some_Var);
30462 @node Creating an Import Library
30463 @subsection Creating an Import Library
30464 @cindex Import library
30467 * The Definition File::
30468 * GNAT-Style Import Library::
30469 * Microsoft-Style Import Library::
30473 If a Microsoft-style import library @file{API.lib} or a GNAT-style
30474 import library @file{libAPI.dll.a} or @file{libAPI.a} is available
30475 with @file{API.dll} you can skip this section. You can also skip this
30476 section if @file{API.dll} or @file{libAPI.dll} is built with GNU tools
30477 as in this case it is possible to link directly against the
30478 DLL. Otherwise read on.
30480 @node The Definition File
30481 @subsubsection The Definition File
30482 @cindex Definition file
30486 As previously mentioned, and unlike Unix systems, the list of symbols
30487 that are exported from a DLL must be provided explicitly in Windows.
30488 The main goal of a definition file is precisely that: list the symbols
30489 exported by a DLL. A definition file (usually a file with a @code{.def}
30490 suffix) has the following structure:
30495 @r{[}LIBRARY @var{name}@r{]}
30496 @r{[}DESCRIPTION @var{string}@r{]}
30506 @item LIBRARY @var{name}
30507 This section, which is optional, gives the name of the DLL.
30509 @item DESCRIPTION @var{string}
30510 This section, which is optional, gives a description string that will be
30511 embedded in the import library.
30514 This section gives the list of exported symbols (procedures, functions or
30515 variables). For instance in the case of @file{API.dll} the @code{EXPORTS}
30516 section of @file{API.def} looks like:
30530 Note that you must specify the correct suffix (@code{@@}@code{@var{nn}})
30531 (@pxref{Windows Calling Conventions}) for a Stdcall
30532 calling convention function in the exported symbols list.
30535 There can actually be other sections in a definition file, but these
30536 sections are not relevant to the discussion at hand.
30538 @node GNAT-Style Import Library
30539 @subsubsection GNAT-Style Import Library
30542 To create a static import library from @file{API.dll} with the GNAT tools
30543 you should proceed as follows:
30547 Create the definition file @file{API.def} (@pxref{The Definition File}).
30548 For that use the @code{dll2def} tool as follows:
30551 $ dll2def API.dll > API.def
30555 @code{dll2def} is a very simple tool: it takes as input a DLL and prints
30556 to standard output the list of entry points in the DLL. Note that if
30557 some routines in the DLL have the @code{Stdcall} convention
30558 (@pxref{Windows Calling Conventions}) with stripped @code{@@}@var{nn}
30559 suffix then you'll have to edit @file{api.def} to add it, and specify
30560 @option{-k} to @command{gnatdll} when creating the import library.
30563 Here are some hints to find the right @code{@@}@var{nn} suffix.
30567 If you have the Microsoft import library (.lib), it is possible to get
30568 the right symbols by using Microsoft @code{dumpbin} tool (see the
30569 corresponding Microsoft documentation for further details).
30572 $ dumpbin /exports api.lib
30576 If you have a message about a missing symbol at link time the compiler
30577 tells you what symbol is expected. You just have to go back to the
30578 definition file and add the right suffix.
30582 Build the import library @code{libAPI.dll.a}, using @code{gnatdll}
30583 (@pxref{Using gnatdll}) as follows:
30586 $ gnatdll -e API.def -d API.dll
30590 @code{gnatdll} takes as input a definition file @file{API.def} and the
30591 name of the DLL containing the services listed in the definition file
30592 @file{API.dll}. The name of the static import library generated is
30593 computed from the name of the definition file as follows: if the
30594 definition file name is @var{xyz}@code{.def}, the import library name will
30595 be @code{lib}@var{xyz}@code{.a}. Note that in the previous example option
30596 @option{-e} could have been removed because the name of the definition
30597 file (before the ``@code{.def}'' suffix) is the same as the name of the
30598 DLL (@pxref{Using gnatdll} for more information about @code{gnatdll}).
30601 @node Microsoft-Style Import Library
30602 @subsubsection Microsoft-Style Import Library
30605 With GNAT you can either use a GNAT-style or Microsoft-style import
30606 library. A Microsoft import library is needed only if you plan to make an
30607 Ada DLL available to applications developed with Microsoft
30608 tools (@pxref{Mixed-Language Programming on Windows}).
30610 To create a Microsoft-style import library for @file{API.dll} you
30611 should proceed as follows:
30615 Create the definition file @file{API.def} from the DLL. For this use either
30616 the @code{dll2def} tool as described above or the Microsoft @code{dumpbin}
30617 tool (see the corresponding Microsoft documentation for further details).
30620 Build the actual import library using Microsoft's @code{lib} utility:
30623 $ lib -machine:IX86 -def:API.def -out:API.lib
30627 If you use the above command the definition file @file{API.def} must
30628 contain a line giving the name of the DLL:
30635 See the Microsoft documentation for further details about the usage of
30639 @node Building DLLs with GNAT Project files
30640 @section Building DLLs with GNAT Project files
30641 @cindex DLLs, building
30644 There is nothing specific to Windows in the build process.
30645 @pxref{Library Projects}.
30648 Due to a system limitation, it is not possible under Windows to create threads
30649 when inside the @code{DllMain} routine which is used for auto-initialization
30650 of shared libraries, so it is not possible to have library level tasks in SALs.
30652 @node Building DLLs with GNAT
30653 @section Building DLLs with GNAT
30654 @cindex DLLs, building
30657 This section explain how to build DLLs using the GNAT built-in DLL
30658 support. With the following procedure it is straight forward to build
30659 and use DLLs with GNAT.
30663 @item building object files
30665 The first step is to build all objects files that are to be included
30666 into the DLL. This is done by using the standard @command{gnatmake} tool.
30668 @item building the DLL
30670 To build the DLL you must use @command{gcc}'s @option{-shared} and
30671 @option{-shared-libgcc} options. It is quite simple to use this method:
30674 $ gcc -shared -shared-libgcc -o api.dll obj1.o obj2.o @dots{}
30677 It is important to note that in this case all symbols found in the
30678 object files are automatically exported. It is possible to restrict
30679 the set of symbols to export by passing to @command{gcc} a definition
30680 file, @pxref{The Definition File}. For example:
30683 $ gcc -shared -shared-libgcc -o api.dll api.def obj1.o obj2.o @dots{}
30686 If you use a definition file you must export the elaboration procedures
30687 for every package that required one. Elaboration procedures are named
30688 using the package name followed by "_E".
30690 @item preparing DLL to be used
30692 For the DLL to be used by client programs the bodies must be hidden
30693 from it and the .ali set with read-only attribute. This is very important
30694 otherwise GNAT will recompile all packages and will not actually use
30695 the code in the DLL. For example:
30699 $ copy *.ads *.ali api.dll apilib
30700 $ attrib +R apilib\*.ali
30705 At this point it is possible to use the DLL by directly linking
30706 against it. Note that you must use the GNAT shared runtime when using
30707 GNAT shared libraries. This is achieved by using @option{-shared} binder's
30711 $ gnatmake main -Iapilib -bargs -shared -largs -Lapilib -lAPI
30714 @node Building DLLs with gnatdll
30715 @section Building DLLs with gnatdll
30716 @cindex DLLs, building
30719 * Limitations When Using Ada DLLs from Ada::
30720 * Exporting Ada Entities::
30721 * Ada DLLs and Elaboration::
30722 * Ada DLLs and Finalization::
30723 * Creating a Spec for Ada DLLs::
30724 * Creating the Definition File::
30729 Note that it is preferred to use GNAT Project files
30730 (@pxref{Building DLLs with GNAT Project files}) or the built-in GNAT
30731 DLL support (@pxref{Building DLLs with GNAT}) or to build DLLs.
30733 This section explains how to build DLLs containing Ada code using
30734 @code{gnatdll}. These DLLs will be referred to as Ada DLLs in the
30735 remainder of this section.
30737 The steps required to build an Ada DLL that is to be used by Ada as well as
30738 non-Ada applications are as follows:
30742 You need to mark each Ada @i{entity} exported by the DLL with a @code{C} or
30743 @code{Stdcall} calling convention to avoid any Ada name mangling for the
30744 entities exported by the DLL (@pxref{Exporting Ada Entities}). You can
30745 skip this step if you plan to use the Ada DLL only from Ada applications.
30748 Your Ada code must export an initialization routine which calls the routine
30749 @code{adainit} generated by @command{gnatbind} to perform the elaboration of
30750 the Ada code in the DLL (@pxref{Ada DLLs and Elaboration}). The initialization
30751 routine exported by the Ada DLL must be invoked by the clients of the DLL
30752 to initialize the DLL.
30755 When useful, the DLL should also export a finalization routine which calls
30756 routine @code{adafinal} generated by @command{gnatbind} to perform the
30757 finalization of the Ada code in the DLL (@pxref{Ada DLLs and Finalization}).
30758 The finalization routine exported by the Ada DLL must be invoked by the
30759 clients of the DLL when the DLL services are no further needed.
30762 You must provide a spec for the services exported by the Ada DLL in each
30763 of the programming languages to which you plan to make the DLL available.
30766 You must provide a definition file listing the exported entities
30767 (@pxref{The Definition File}).
30770 Finally you must use @code{gnatdll} to produce the DLL and the import
30771 library (@pxref{Using gnatdll}).
30775 Note that a relocatable DLL stripped using the @code{strip}
30776 binutils tool will not be relocatable anymore. To build a DLL without
30777 debug information pass @code{-largs -s} to @code{gnatdll}. This
30778 restriction does not apply to a DLL built using a Library Project.
30779 @pxref{Library Projects}.
30781 @node Limitations When Using Ada DLLs from Ada
30782 @subsection Limitations When Using Ada DLLs from Ada
30785 When using Ada DLLs from Ada applications there is a limitation users
30786 should be aware of. Because on Windows the GNAT run time is not in a DLL of
30787 its own, each Ada DLL includes a part of the GNAT run time. Specifically,
30788 each Ada DLL includes the services of the GNAT run time that are necessary
30789 to the Ada code inside the DLL. As a result, when an Ada program uses an
30790 Ada DLL there are two independent GNAT run times: one in the Ada DLL and
30791 one in the main program.
30793 It is therefore not possible to exchange GNAT run-time objects between the
30794 Ada DLL and the main Ada program. Example of GNAT run-time objects are file
30795 handles (e.g.@: @code{Text_IO.File_Type}), tasks types, protected objects
30798 It is completely safe to exchange plain elementary, array or record types,
30799 Windows object handles, etc.
30801 @node Exporting Ada Entities
30802 @subsection Exporting Ada Entities
30803 @cindex Export table
30806 Building a DLL is a way to encapsulate a set of services usable from any
30807 application. As a result, the Ada entities exported by a DLL should be
30808 exported with the @code{C} or @code{Stdcall} calling conventions to avoid
30809 any Ada name mangling. As an example here is an Ada package
30810 @code{API}, spec and body, exporting two procedures, a function, and a
30813 @smallexample @c ada
30816 with Interfaces.C; use Interfaces;
30818 Count : C.int := 0;
30819 function Factorial (Val : C.int) return C.int;
30821 procedure Initialize_API;
30822 procedure Finalize_API;
30823 -- Initialization & Finalization routines. More in the next section.
30825 pragma Export (C, Initialize_API);
30826 pragma Export (C, Finalize_API);
30827 pragma Export (C, Count);
30828 pragma Export (C, Factorial);
30834 @smallexample @c ada
30837 package body API is
30838 function Factorial (Val : C.int) return C.int is
30841 Count := Count + 1;
30842 for K in 1 .. Val loop
30848 procedure Initialize_API is
30850 pragma Import (C, Adainit);
30853 end Initialize_API;
30855 procedure Finalize_API is
30856 procedure Adafinal;
30857 pragma Import (C, Adafinal);
30867 If the Ada DLL you are building will only be used by Ada applications
30868 you do not have to export Ada entities with a @code{C} or @code{Stdcall}
30869 convention. As an example, the previous package could be written as
30872 @smallexample @c ada
30876 Count : Integer := 0;
30877 function Factorial (Val : Integer) return Integer;
30879 procedure Initialize_API;
30880 procedure Finalize_API;
30881 -- Initialization and Finalization routines.
30887 @smallexample @c ada
30890 package body API is
30891 function Factorial (Val : Integer) return Integer is
30892 Fact : Integer := 1;
30894 Count := Count + 1;
30895 for K in 1 .. Val loop
30902 -- The remainder of this package body is unchanged.
30909 Note that if you do not export the Ada entities with a @code{C} or
30910 @code{Stdcall} convention you will have to provide the mangled Ada names
30911 in the definition file of the Ada DLL
30912 (@pxref{Creating the Definition File}).
30914 @node Ada DLLs and Elaboration
30915 @subsection Ada DLLs and Elaboration
30916 @cindex DLLs and elaboration
30919 The DLL that you are building contains your Ada code as well as all the
30920 routines in the Ada library that are needed by it. The first thing a
30921 user of your DLL must do is elaborate the Ada code
30922 (@pxref{Elaboration Order Handling in GNAT}).
30924 To achieve this you must export an initialization routine
30925 (@code{Initialize_API} in the previous example), which must be invoked
30926 before using any of the DLL services. This elaboration routine must call
30927 the Ada elaboration routine @code{adainit} generated by the GNAT binder
30928 (@pxref{Binding with Non-Ada Main Programs}). See the body of
30929 @code{Initialize_Api} for an example. Note that the GNAT binder is
30930 automatically invoked during the DLL build process by the @code{gnatdll}
30931 tool (@pxref{Using gnatdll}).
30933 When a DLL is loaded, Windows systematically invokes a routine called
30934 @code{DllMain}. It would therefore be possible to call @code{adainit}
30935 directly from @code{DllMain} without having to provide an explicit
30936 initialization routine. Unfortunately, it is not possible to call
30937 @code{adainit} from the @code{DllMain} if your program has library level
30938 tasks because access to the @code{DllMain} entry point is serialized by
30939 the system (that is, only a single thread can execute ``through'' it at a
30940 time), which means that the GNAT run time will deadlock waiting for the
30941 newly created task to complete its initialization.
30943 @node Ada DLLs and Finalization
30944 @subsection Ada DLLs and Finalization
30945 @cindex DLLs and finalization
30948 When the services of an Ada DLL are no longer needed, the client code should
30949 invoke the DLL finalization routine, if available. The DLL finalization
30950 routine is in charge of releasing all resources acquired by the DLL. In the
30951 case of the Ada code contained in the DLL, this is achieved by calling
30952 routine @code{adafinal} generated by the GNAT binder
30953 (@pxref{Binding with Non-Ada Main Programs}).
30954 See the body of @code{Finalize_Api} for an
30955 example. As already pointed out the GNAT binder is automatically invoked
30956 during the DLL build process by the @code{gnatdll} tool
30957 (@pxref{Using gnatdll}).
30959 @node Creating a Spec for Ada DLLs
30960 @subsection Creating a Spec for Ada DLLs
30963 To use the services exported by the Ada DLL from another programming
30964 language (e.g.@: C), you have to translate the specs of the exported Ada
30965 entities in that language. For instance in the case of @code{API.dll},
30966 the corresponding C header file could look like:
30971 extern int *_imp__count;
30972 #define count (*_imp__count)
30973 int factorial (int);
30979 It is important to understand that when building an Ada DLL to be used by
30980 other Ada applications, you need two different specs for the packages
30981 contained in the DLL: one for building the DLL and the other for using
30982 the DLL. This is because the @code{DLL} calling convention is needed to
30983 use a variable defined in a DLL, but when building the DLL, the variable
30984 must have either the @code{Ada} or @code{C} calling convention. As an
30985 example consider a DLL comprising the following package @code{API}:
30987 @smallexample @c ada
30991 Count : Integer := 0;
30993 -- Remainder of the package omitted.
31000 After producing a DLL containing package @code{API}, the spec that
31001 must be used to import @code{API.Count} from Ada code outside of the
31004 @smallexample @c ada
31009 pragma Import (DLL, Count);
31015 @node Creating the Definition File
31016 @subsection Creating the Definition File
31019 The definition file is the last file needed to build the DLL. It lists
31020 the exported symbols. As an example, the definition file for a DLL
31021 containing only package @code{API} (where all the entities are exported
31022 with a @code{C} calling convention) is:
31037 If the @code{C} calling convention is missing from package @code{API},
31038 then the definition file contains the mangled Ada names of the above
31039 entities, which in this case are:
31048 api__initialize_api
31053 @node Using gnatdll
31054 @subsection Using @code{gnatdll}
31058 * gnatdll Example::
31059 * gnatdll behind the Scenes::
31064 @code{gnatdll} is a tool to automate the DLL build process once all the Ada
31065 and non-Ada sources that make up your DLL have been compiled.
31066 @code{gnatdll} is actually in charge of two distinct tasks: build the
31067 static import library for the DLL and the actual DLL. The form of the
31068 @code{gnatdll} command is
31072 @c $ gnatdll @ovar{switches} @var{list-of-files} @r{[}-largs @var{opts}@r{]}
31073 @c Expanding @ovar macro inline (explanation in macro def comments)
31074 $ gnatdll @r{[}@var{switches}@r{]} @var{list-of-files} @r{[}-largs @var{opts}@r{]}
31079 where @var{list-of-files} is a list of ALI and object files. The object
31080 file list must be the exact list of objects corresponding to the non-Ada
31081 sources whose services are to be included in the DLL. The ALI file list
31082 must be the exact list of ALI files for the corresponding Ada sources
31083 whose services are to be included in the DLL. If @var{list-of-files} is
31084 missing, only the static import library is generated.
31087 You may specify any of the following switches to @code{gnatdll}:
31090 @c @item -a@ovar{address}
31091 @c Expanding @ovar macro inline (explanation in macro def comments)
31092 @item -a@r{[}@var{address}@r{]}
31093 @cindex @option{-a} (@code{gnatdll})
31094 Build a non-relocatable DLL at @var{address}. If @var{address} is not
31095 specified the default address @var{0x11000000} will be used. By default,
31096 when this switch is missing, @code{gnatdll} builds relocatable DLL. We
31097 advise the reader to build relocatable DLL.
31099 @item -b @var{address}
31100 @cindex @option{-b} (@code{gnatdll})
31101 Set the relocatable DLL base address. By default the address is
31104 @item -bargs @var{opts}
31105 @cindex @option{-bargs} (@code{gnatdll})
31106 Binder options. Pass @var{opts} to the binder.
31108 @item -d @var{dllfile}
31109 @cindex @option{-d} (@code{gnatdll})
31110 @var{dllfile} is the name of the DLL. This switch must be present for
31111 @code{gnatdll} to do anything. The name of the generated import library is
31112 obtained algorithmically from @var{dllfile} as shown in the following
31113 example: if @var{dllfile} is @code{xyz.dll}, the import library name is
31114 @code{libxyz.dll.a}. The name of the definition file to use (if not specified
31115 by option @option{-e}) is obtained algorithmically from @var{dllfile}
31116 as shown in the following example:
31117 if @var{dllfile} is @code{xyz.dll}, the definition
31118 file used is @code{xyz.def}.
31120 @item -e @var{deffile}
31121 @cindex @option{-e} (@code{gnatdll})
31122 @var{deffile} is the name of the definition file.
31125 @cindex @option{-g} (@code{gnatdll})
31126 Generate debugging information. This information is stored in the object
31127 file and copied from there to the final DLL file by the linker,
31128 where it can be read by the debugger. You must use the
31129 @option{-g} switch if you plan on using the debugger or the symbolic
31133 @cindex @option{-h} (@code{gnatdll})
31134 Help mode. Displays @code{gnatdll} switch usage information.
31137 @cindex @option{-I} (@code{gnatdll})
31138 Direct @code{gnatdll} to search the @var{dir} directory for source and
31139 object files needed to build the DLL.
31140 (@pxref{Search Paths and the Run-Time Library (RTL)}).
31143 @cindex @option{-k} (@code{gnatdll})
31144 Removes the @code{@@}@var{nn} suffix from the import library's exported
31145 names, but keeps them for the link names. You must specify this
31146 option if you want to use a @code{Stdcall} function in a DLL for which
31147 the @code{@@}@var{nn} suffix has been removed. This is the case for most
31148 of the Windows NT DLL for example. This option has no effect when
31149 @option{-n} option is specified.
31151 @item -l @var{file}
31152 @cindex @option{-l} (@code{gnatdll})
31153 The list of ALI and object files used to build the DLL are listed in
31154 @var{file}, instead of being given in the command line. Each line in
31155 @var{file} contains the name of an ALI or object file.
31158 @cindex @option{-n} (@code{gnatdll})
31159 No Import. Do not create the import library.
31162 @cindex @option{-q} (@code{gnatdll})
31163 Quiet mode. Do not display unnecessary messages.
31166 @cindex @option{-v} (@code{gnatdll})
31167 Verbose mode. Display extra information.
31169 @item -largs @var{opts}
31170 @cindex @option{-largs} (@code{gnatdll})
31171 Linker options. Pass @var{opts} to the linker.
31174 @node gnatdll Example
31175 @subsubsection @code{gnatdll} Example
31178 As an example the command to build a relocatable DLL from @file{api.adb}
31179 once @file{api.adb} has been compiled and @file{api.def} created is
31182 $ gnatdll -d api.dll api.ali
31186 The above command creates two files: @file{libapi.dll.a} (the import
31187 library) and @file{api.dll} (the actual DLL). If you want to create
31188 only the DLL, just type:
31191 $ gnatdll -d api.dll -n api.ali
31195 Alternatively if you want to create just the import library, type:
31198 $ gnatdll -d api.dll
31201 @node gnatdll behind the Scenes
31202 @subsubsection @code{gnatdll} behind the Scenes
31205 This section details the steps involved in creating a DLL. @code{gnatdll}
31206 does these steps for you. Unless you are interested in understanding what
31207 goes on behind the scenes, you should skip this section.
31209 We use the previous example of a DLL containing the Ada package @code{API},
31210 to illustrate the steps necessary to build a DLL. The starting point is a
31211 set of objects that will make up the DLL and the corresponding ALI
31212 files. In the case of this example this means that @file{api.o} and
31213 @file{api.ali} are available. To build a relocatable DLL, @code{gnatdll} does
31218 @code{gnatdll} builds the base file (@file{api.base}). A base file gives
31219 the information necessary to generate relocation information for the
31225 $ gnatlink api -o api.jnk -mdll -Wl,--base-file,api.base
31230 In addition to the base file, the @command{gnatlink} command generates an
31231 output file @file{api.jnk} which can be discarded. The @option{-mdll} switch
31232 asks @command{gnatlink} to generate the routines @code{DllMain} and
31233 @code{DllMainCRTStartup} that are called by the Windows loader when the DLL
31234 is loaded into memory.
31237 @code{gnatdll} uses @code{dlltool} (@pxref{Using dlltool}) to build the
31238 export table (@file{api.exp}). The export table contains the relocation
31239 information in a form which can be used during the final link to ensure
31240 that the Windows loader is able to place the DLL anywhere in memory.
31244 $ dlltool --dllname api.dll --def api.def --base-file api.base \
31245 --output-exp api.exp
31250 @code{gnatdll} builds the base file using the new export table. Note that
31251 @command{gnatbind} must be called once again since the binder generated file
31252 has been deleted during the previous call to @command{gnatlink}.
31257 $ gnatlink api -o api.jnk api.exp -mdll
31258 -Wl,--base-file,api.base
31263 @code{gnatdll} builds the new export table using the new base file and
31264 generates the DLL import library @file{libAPI.dll.a}.
31268 $ dlltool --dllname api.dll --def api.def --base-file api.base \
31269 --output-exp api.exp --output-lib libAPI.a
31274 Finally @code{gnatdll} builds the relocatable DLL using the final export
31280 $ gnatlink api api.exp -o api.dll -mdll
31285 @node Using dlltool
31286 @subsubsection Using @code{dlltool}
31289 @code{dlltool} is the low-level tool used by @code{gnatdll} to build
31290 DLLs and static import libraries. This section summarizes the most
31291 common @code{dlltool} switches. The form of the @code{dlltool} command
31295 @c $ dlltool @ovar{switches}
31296 @c Expanding @ovar macro inline (explanation in macro def comments)
31297 $ dlltool @r{[}@var{switches}@r{]}
31301 @code{dlltool} switches include:
31304 @item --base-file @var{basefile}
31305 @cindex @option{--base-file} (@command{dlltool})
31306 Read the base file @var{basefile} generated by the linker. This switch
31307 is used to create a relocatable DLL.
31309 @item --def @var{deffile}
31310 @cindex @option{--def} (@command{dlltool})
31311 Read the definition file.
31313 @item --dllname @var{name}
31314 @cindex @option{--dllname} (@command{dlltool})
31315 Gives the name of the DLL. This switch is used to embed the name of the
31316 DLL in the static import library generated by @code{dlltool} with switch
31317 @option{--output-lib}.
31320 @cindex @option{-k} (@command{dlltool})
31321 Kill @code{@@}@var{nn} from exported names
31322 (@pxref{Windows Calling Conventions}
31323 for a discussion about @code{Stdcall}-style symbols.
31326 @cindex @option{--help} (@command{dlltool})
31327 Prints the @code{dlltool} switches with a concise description.
31329 @item --output-exp @var{exportfile}
31330 @cindex @option{--output-exp} (@command{dlltool})
31331 Generate an export file @var{exportfile}. The export file contains the
31332 export table (list of symbols in the DLL) and is used to create the DLL.
31334 @item --output-lib @var{libfile}
31335 @cindex @option{--output-lib} (@command{dlltool})
31336 Generate a static import library @var{libfile}.
31339 @cindex @option{-v} (@command{dlltool})
31342 @item --as @var{assembler-name}
31343 @cindex @option{--as} (@command{dlltool})
31344 Use @var{assembler-name} as the assembler. The default is @code{as}.
31347 @node GNAT and Windows Resources
31348 @section GNAT and Windows Resources
31349 @cindex Resources, windows
31352 * Building Resources::
31353 * Compiling Resources::
31354 * Using Resources::
31358 Resources are an easy way to add Windows specific objects to your
31359 application. The objects that can be added as resources include:
31368 @item string tables
31378 @item version information
31381 For example, a version information resource can be defined as follow and
31382 embedded into an executable or DLL:
31384 A version information resource can be used to embed information into an
31385 executable or a DLL. These information can be viewed using the file properties
31386 from the Windows Explorer. Here is an example of a version information
31392 FILEVERSION 1,0,0,0
31393 PRODUCTVERSION 1,0,0,0
31395 BLOCK "StringFileInfo"
31399 VALUE "CompanyName", "My Company Name"
31400 VALUE "FileDescription", "My application"
31401 VALUE "FileVersion", "1.0"
31402 VALUE "InternalName", "my_app"
31403 VALUE "LegalCopyright", "My Name"
31404 VALUE "OriginalFilename", "my_app.exe"
31405 VALUE "ProductName", "My App"
31406 VALUE "ProductVersion", "1.0"
31410 BLOCK "VarFileInfo"
31412 VALUE "Translation", 0x809, 1252
31418 The value @code{0809} (langID) is for the U.K English language and
31419 @code{04E4} (charsetID), which is equal to @code{1252} decimal, for
31423 This section explains how to build, compile and use resources. Note that this
31424 section does not cover all resource objects, for a complete description see
31425 the corresponding Microsoft documentation.
31427 @node Building Resources
31428 @subsection Building Resources
31429 @cindex Resources, building
31432 A resource file is an ASCII file. By convention resource files have an
31433 @file{.rc} extension.
31434 The easiest way to build a resource file is to use Microsoft tools
31435 such as @code{imagedit.exe} to build bitmaps, icons and cursors and
31436 @code{dlgedit.exe} to build dialogs.
31437 It is always possible to build an @file{.rc} file yourself by writing a
31440 It is not our objective to explain how to write a resource file. A
31441 complete description of the resource script language can be found in the
31442 Microsoft documentation.
31444 @node Compiling Resources
31445 @subsection Compiling Resources
31448 @cindex Resources, compiling
31451 This section describes how to build a GNAT-compatible (COFF) object file
31452 containing the resources. This is done using the Resource Compiler
31453 @code{windres} as follows:
31456 $ windres -i myres.rc -o myres.o
31460 By default @code{windres} will run @command{gcc} to preprocess the @file{.rc}
31461 file. You can specify an alternate preprocessor (usually named
31462 @file{cpp.exe}) using the @code{windres} @option{--preprocessor}
31463 parameter. A list of all possible options may be obtained by entering
31464 the command @code{windres} @option{--help}.
31466 It is also possible to use the Microsoft resource compiler @code{rc.exe}
31467 to produce a @file{.res} file (binary resource file). See the
31468 corresponding Microsoft documentation for further details. In this case
31469 you need to use @code{windres} to translate the @file{.res} file to a
31470 GNAT-compatible object file as follows:
31473 $ windres -i myres.res -o myres.o
31476 @node Using Resources
31477 @subsection Using Resources
31478 @cindex Resources, using
31481 To include the resource file in your program just add the
31482 GNAT-compatible object file for the resource(s) to the linker
31483 arguments. With @command{gnatmake} this is done by using the @option{-largs}
31487 $ gnatmake myprog -largs myres.o
31490 @node Debugging a DLL
31491 @section Debugging a DLL
31492 @cindex DLL debugging
31495 * Program and DLL Both Built with GCC/GNAT::
31496 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
31500 Debugging a DLL is similar to debugging a standard program. But
31501 we have to deal with two different executable parts: the DLL and the
31502 program that uses it. We have the following four possibilities:
31506 The program and the DLL are built with @code{GCC/GNAT}.
31508 The program is built with foreign tools and the DLL is built with
31511 The program is built with @code{GCC/GNAT} and the DLL is built with
31516 In this section we address only cases one and two above.
31517 There is no point in trying to debug
31518 a DLL with @code{GNU/GDB}, if there is no GDB-compatible debugging
31519 information in it. To do so you must use a debugger compatible with the
31520 tools suite used to build the DLL.
31522 @node Program and DLL Both Built with GCC/GNAT
31523 @subsection Program and DLL Both Built with GCC/GNAT
31526 This is the simplest case. Both the DLL and the program have @code{GDB}
31527 compatible debugging information. It is then possible to break anywhere in
31528 the process. Let's suppose here that the main procedure is named
31529 @code{ada_main} and that in the DLL there is an entry point named
31533 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) and
31534 program must have been built with the debugging information (see GNAT -g
31535 switch). Here are the step-by-step instructions for debugging it:
31538 @item Launch @code{GDB} on the main program.
31544 @item Start the program and stop at the beginning of the main procedure
31551 This step is required to be able to set a breakpoint inside the DLL. As long
31552 as the program is not run, the DLL is not loaded. This has the
31553 consequence that the DLL debugging information is also not loaded, so it is not
31554 possible to set a breakpoint in the DLL.
31556 @item Set a breakpoint inside the DLL
31559 (gdb) break ada_dll
31566 At this stage a breakpoint is set inside the DLL. From there on
31567 you can use the standard approach to debug the whole program
31568 (@pxref{Running and Debugging Ada Programs}).
31571 @c This used to work, probably because the DLLs were non-relocatable
31572 @c keep this section around until the problem is sorted out.
31574 To break on the @code{DllMain} routine it is not possible to follow
31575 the procedure above. At the time the program stop on @code{ada_main}
31576 the @code{DllMain} routine as already been called. Either you can use
31577 the procedure below @pxref{Debugging the DLL Directly} or this procedure:
31580 @item Launch @code{GDB} on the main program.
31586 @item Load DLL symbols
31589 (gdb) add-sym api.dll
31592 @item Set a breakpoint inside the DLL
31595 (gdb) break ada_dll.adb:45
31598 Note that at this point it is not possible to break using the routine symbol
31599 directly as the program is not yet running. The solution is to break
31600 on the proper line (break in @file{ada_dll.adb} line 45).
31602 @item Start the program
31611 @node Program Built with Foreign Tools and DLL Built with GCC/GNAT
31612 @subsection Program Built with Foreign Tools and DLL Built with GCC/GNAT
31615 * Debugging the DLL Directly::
31616 * Attaching to a Running Process::
31620 In this case things are slightly more complex because it is not possible to
31621 start the main program and then break at the beginning to load the DLL and the
31622 associated DLL debugging information. It is not possible to break at the
31623 beginning of the program because there is no @code{GDB} debugging information,
31624 and therefore there is no direct way of getting initial control. This
31625 section addresses this issue by describing some methods that can be used
31626 to break somewhere in the DLL to debug it.
31629 First suppose that the main procedure is named @code{main} (this is for
31630 example some C code built with Microsoft Visual C) and that there is a
31631 DLL named @code{test.dll} containing an Ada entry point named
31635 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) must have
31636 been built with debugging information (see GNAT -g option).
31638 @node Debugging the DLL Directly
31639 @subsubsection Debugging the DLL Directly
31643 Find out the executable starting address
31646 $ objdump --file-header main.exe
31649 The starting address is reported on the last line. For example:
31652 main.exe: file format pei-i386
31653 architecture: i386, flags 0x0000010a:
31654 EXEC_P, HAS_DEBUG, D_PAGED
31655 start address 0x00401010
31659 Launch the debugger on the executable.
31666 Set a breakpoint at the starting address, and launch the program.
31669 $ (gdb) break *0x00401010
31673 The program will stop at the given address.
31676 Set a breakpoint on a DLL subroutine.
31679 (gdb) break ada_dll.adb:45
31682 Or if you want to break using a symbol on the DLL, you need first to
31683 select the Ada language (language used by the DLL).
31686 (gdb) set language ada
31687 (gdb) break ada_dll
31691 Continue the program.
31698 This will run the program until it reaches the breakpoint that has been
31699 set. From that point you can use the standard way to debug a program
31700 as described in (@pxref{Running and Debugging Ada Programs}).
31705 It is also possible to debug the DLL by attaching to a running process.
31707 @node Attaching to a Running Process
31708 @subsubsection Attaching to a Running Process
31709 @cindex DLL debugging, attach to process
31712 With @code{GDB} it is always possible to debug a running process by
31713 attaching to it. It is possible to debug a DLL this way. The limitation
31714 of this approach is that the DLL must run long enough to perform the
31715 attach operation. It may be useful for instance to insert a time wasting
31716 loop in the code of the DLL to meet this criterion.
31720 @item Launch the main program @file{main.exe}.
31726 @item Use the Windows @i{Task Manager} to find the process ID. Let's say
31727 that the process PID for @file{main.exe} is 208.
31735 @item Attach to the running process to be debugged.
31741 @item Load the process debugging information.
31744 (gdb) symbol-file main.exe
31747 @item Break somewhere in the DLL.
31750 (gdb) break ada_dll
31753 @item Continue process execution.
31762 This last step will resume the process execution, and stop at
31763 the breakpoint we have set. From there you can use the standard
31764 approach to debug a program as described in
31765 (@pxref{Running and Debugging Ada Programs}).
31767 @node Setting Stack Size from gnatlink
31768 @section Setting Stack Size from @command{gnatlink}
31771 It is possible to specify the program stack size at link time. On modern
31772 versions of Windows, starting with XP, this is mostly useful to set the size of
31773 the main stack (environment task). The other task stacks are set with pragma
31774 Storage_Size or with the @command{gnatbind -d} command.
31776 Since older versions of Windows (2000, NT4, etc.) do not allow setting the
31777 reserve size of individual tasks, the link-time stack size applies to all
31778 tasks, and pragma Storage_Size has no effect.
31779 In particular, Stack Overflow checks are made against this
31780 link-time specified size.
31782 This setting can be done with
31783 @command{gnatlink} using either:
31787 @item using @option{-Xlinker} linker option
31790 $ gnatlink hello -Xlinker --stack=0x10000,0x1000
31793 This sets the stack reserve size to 0x10000 bytes and the stack commit
31794 size to 0x1000 bytes.
31796 @item using @option{-Wl} linker option
31799 $ gnatlink hello -Wl,--stack=0x1000000
31802 This sets the stack reserve size to 0x1000000 bytes. Note that with
31803 @option{-Wl} option it is not possible to set the stack commit size
31804 because the coma is a separator for this option.
31808 @node Setting Heap Size from gnatlink
31809 @section Setting Heap Size from @command{gnatlink}
31812 Under Windows systems, it is possible to specify the program heap size from
31813 @command{gnatlink} using either:
31817 @item using @option{-Xlinker} linker option
31820 $ gnatlink hello -Xlinker --heap=0x10000,0x1000
31823 This sets the heap reserve size to 0x10000 bytes and the heap commit
31824 size to 0x1000 bytes.
31826 @item using @option{-Wl} linker option
31829 $ gnatlink hello -Wl,--heap=0x1000000
31832 This sets the heap reserve size to 0x1000000 bytes. Note that with
31833 @option{-Wl} option it is not possible to set the heap commit size
31834 because the coma is a separator for this option.
31838 @node Mac OS Topics
31839 @appendix Mac OS Topics
31843 This chapter describes topics that are specific to Apple's OS X
31847 * Codesigning the Debugger::
31850 @node Codesigning the Debugger
31851 @section Codesigning the Debugger
31854 The Darwin Kernel requires the debugger to have special permissions
31855 before it is allowed to control other processes. These permissions
31856 are granted by codesigning the GDB executable. Without these
31857 permissions, the debugger will report error messages such as:
31860 Starting program: /x/y/foo
31861 Unable to find Mach task port for process-id 28885: (os/kern) failure (0x5).
31862 (please check gdb is codesigned - see taskgated(8))
31865 Codesigning requires a certificate. The following procedure explains
31869 @item Start the Keychain Access application (in
31870 /Applications/Utilities/Keychain Access.app)
31872 @item Select the Keychain Access -> Certificate Assistant ->
31873 Create a Certificate... menu
31878 @item Choose a name for the new certificate (this procedure will use
31879 "gdb-cert" as an example)
31881 @item Set "Identity Type" to "Self Signed Root"
31883 @item Set "Certificate Type" to "Code Signing"
31885 @item Activate the "Let me override defaults" option
31889 @item Click several times on "Continue" until the "Specify a Location
31890 For The Certificate" screen appears, then set "Keychain" to "System"
31892 @item Click on "Continue" until the certificate is created
31894 @item Finally, in the view, double-click on the new certificate,
31895 and set "When using this certificate" to "Always Trust"
31897 @item Exit the Keychain Access application and restart the computer
31898 (this is unfortunately required)
31902 Once a certificate has been created, the debugger can be codesigned
31903 as follow. In a Terminal, run the following command...
31906 codesign -f -s "gdb-cert" <gnat_install_prefix>/bin/gdb
31909 ... where "gdb-cert" should be replaced by the actual certificate
31910 name chosen above, and <gnat_install_prefix> should be replaced by
31911 the location where you installed GNAT.
31913 @c **********************************
31914 @c * GNU Free Documentation License *
31915 @c **********************************
31917 @c GNU Free Documentation License
31925 @c Put table of contents at end, otherwise it precedes the "title page" in
31926 @c the .txt version
31927 @c Edit the pdf file to move the contents to the beginning, after the title