1 \input texinfo @c -*-texinfo-*-
3 @setfilename gnat_ugn.info
4 @documentencoding UTF-8
6 @*Generated by Sphinx 1.4.6.@*
8 @settitle GNAT User's Guide for Native Platforms
13 @dircategory GNU Ada Tools
15 * gnat_ugn: (gnat_ugn.info). gnat_ugn
18 @definfoenclose strong,`,'
19 @definfoenclose emph,`,'
24 GNAT User's Guide for Native Platforms , Jun 21, 2019
28 Copyright @copyright{} 2008-2019, Free Software Foundation
34 @title GNAT User's Guide for Native Platforms
39 @c %** start of user preamble
41 @c %** end of user preamble
45 @top GNAT User's Guide for Native Platforms
50 @anchor{gnat_ugn doc}@anchor{0}
51 @emph{GNAT, The GNU Ada Development Environment}
54 @include gcc-common.texi
55 GCC version @value{version-GCC}@*
58 Permission is granted to copy, distribute and/or modify this document
59 under the terms of the GNU Free Documentation License, Version 1.3 or
60 any later version published by the Free Software Foundation; with no
61 Invariant Sections, with the Front-Cover Texts being
62 "GNAT User's Guide for Native Platforms",
63 and with no Back-Cover Texts. A copy of the license is
64 included in the section entitled @ref{1,,GNU Free Documentation License}.
68 * Getting Started with GNAT::
69 * The GNAT Compilation Model::
70 * Building Executable Programs with GNAT::
71 * GNAT Utility Programs::
72 * GNAT and Program Execution::
73 * Platform-Specific Information::
74 * Example of Binder Output File::
75 * Elaboration Order Handling in GNAT::
77 * GNU Free Documentation License::
81 --- The Detailed Node Listing ---
85 * What This Guide Contains::
86 * What You Should Know before Reading This Guide::
87 * Related Information::
88 * A Note to Readers of Previous Versions of the Manual::
91 Getting Started with GNAT
94 * Running a Simple Ada Program::
95 * Running a Program with Multiple Units::
96 * Using the gnatmake Utility::
98 The GNAT Compilation Model
100 * Source Representation::
101 * Foreign Language Representation::
102 * File Naming Topics and Utilities::
103 * Configuration Pragmas::
104 * Generating Object Files::
105 * Source Dependencies::
106 * The Ada Library Information Files::
107 * Binding an Ada Program::
108 * GNAT and Libraries::
109 * Conditional Compilation::
110 * Mixed Language Programming::
111 * GNAT and Other Compilation Models::
112 * Using GNAT Files with External Tools::
114 Foreign Language Representation
117 * Other 8-Bit Codes::
118 * Wide_Character Encodings::
119 * Wide_Wide_Character Encodings::
121 File Naming Topics and Utilities
123 * File Naming Rules::
124 * Using Other File Names::
125 * Alternative File Naming Schemes::
126 * Handling Arbitrary File Naming Conventions with gnatname::
127 * File Name Krunching with gnatkr::
128 * Renaming Files with gnatchop::
130 Handling Arbitrary File Naming Conventions with gnatname
132 * Arbitrary File Naming Conventions::
134 * Switches for gnatname::
135 * Examples of gnatname Usage::
137 File Name Krunching with gnatkr
142 * Examples of gnatkr Usage::
144 Renaming Files with gnatchop
146 * Handling Files with Multiple Units::
147 * Operating gnatchop in Compilation Mode::
148 * Command Line for gnatchop::
149 * Switches for gnatchop::
150 * Examples of gnatchop Usage::
152 Configuration Pragmas
154 * Handling of Configuration Pragmas::
155 * The Configuration Pragmas Files::
159 * Introduction to Libraries in GNAT::
160 * General Ada Libraries::
161 * Stand-alone Ada Libraries::
162 * Rebuilding the GNAT Run-Time Library::
164 General Ada Libraries
166 * Building a library::
167 * Installing a library::
170 Stand-alone Ada Libraries
172 * Introduction to Stand-alone Libraries::
173 * Building a Stand-alone Library::
174 * Creating a Stand-alone Library to be used in a non-Ada context::
175 * Restrictions in Stand-alone Libraries::
177 Conditional Compilation
179 * Modeling Conditional Compilation in Ada::
180 * Preprocessing with gnatprep::
181 * Integrated Preprocessing::
183 Modeling Conditional Compilation in Ada
185 * Use of Boolean Constants::
186 * Debugging - A Special Case::
187 * Conditionalizing Declarations::
188 * Use of Alternative Implementations::
191 Preprocessing with gnatprep
193 * Preprocessing Symbols::
195 * Switches for gnatprep::
196 * Form of Definitions File::
197 * Form of Input Text for gnatprep::
199 Mixed Language Programming
202 * Calling Conventions::
203 * Building Mixed Ada and C++ Programs::
204 * Generating Ada Bindings for C and C++ headers::
205 * Generating C Headers for Ada Specifications::
207 Building Mixed Ada and C++ Programs
209 * Interfacing to C++::
210 * Linking a Mixed C++ & Ada Program::
212 * Interfacing with C++ constructors::
213 * Interfacing with C++ at the Class Level::
215 Generating Ada Bindings for C and C++ headers
217 * Running the Binding Generator::
218 * Generating Bindings for C++ Headers::
221 Generating C Headers for Ada Specifications
223 * Running the C Header Generator::
225 GNAT and Other Compilation Models
227 * Comparison between GNAT and C/C++ Compilation Models::
228 * Comparison between GNAT and Conventional Ada Library Models::
230 Using GNAT Files with External Tools
232 * Using Other Utility Programs with GNAT::
233 * The External Symbol Naming Scheme of GNAT::
235 Building Executable Programs with GNAT
237 * Building with gnatmake::
238 * Compiling with gcc::
239 * Compiler Switches::
241 * Binding with gnatbind::
242 * Linking with gnatlink::
243 * Using the GNU make Utility::
245 Building with gnatmake
248 * Switches for gnatmake::
249 * Mode Switches for gnatmake::
250 * Notes on the Command Line::
251 * How gnatmake Works::
252 * Examples of gnatmake Usage::
256 * Compiling Programs::
257 * Search Paths and the Run-Time Library (RTL): Search Paths and the Run-Time Library RTL.
258 * Order of Compilation Issues::
263 * Alphabetical List of All Switches::
264 * Output and Error Message Control::
265 * Warning Message Control::
266 * Debugging and Assertion Control::
267 * Validity Checking::
270 * Using gcc for Syntax Checking::
271 * Using gcc for Semantic Checking::
272 * Compiling Different Versions of Ada::
273 * Character Set Control::
274 * File Naming Control::
275 * Subprogram Inlining Control::
276 * Auxiliary Output Control::
277 * Debugging Control::
278 * Exception Handling Control::
279 * Units to Sources Mapping Files::
280 * Code Generation Control::
282 Binding with gnatbind
285 * Switches for gnatbind::
286 * Command-Line Access::
287 * Search Paths for gnatbind::
288 * Examples of gnatbind Usage::
290 Switches for gnatbind
292 * Consistency-Checking Modes::
293 * Binder Error Message Control::
294 * Elaboration Control::
296 * Dynamic Allocation Control::
297 * Binding with Non-Ada Main Programs::
298 * Binding Programs with No Main Subprogram::
300 Linking with gnatlink
303 * Switches for gnatlink::
305 Using the GNU make Utility
307 * Using gnatmake in a Makefile::
308 * Automatically Creating a List of Directories::
309 * Generating the Command Line Switches::
310 * Overcoming Command Line Length Limits::
312 GNAT Utility Programs
314 * The File Cleanup Utility gnatclean::
315 * The GNAT Library Browser gnatls::
316 * The Cross-Referencing Tools gnatxref and gnatfind::
317 * The Ada to HTML Converter gnathtml::
319 The File Cleanup Utility gnatclean
321 * Running gnatclean::
322 * Switches for gnatclean::
324 The GNAT Library Browser gnatls
327 * Switches for gnatls::
328 * Example of gnatls Usage::
330 The Cross-Referencing Tools gnatxref and gnatfind
332 * gnatxref Switches::
333 * gnatfind Switches::
334 * Configuration Files for gnatxref and gnatfind::
335 * Regular Expressions in gnatfind and gnatxref::
336 * Examples of gnatxref Usage::
337 * Examples of gnatfind Usage::
339 Examples of gnatxref Usage
342 * Using gnatxref with vi::
344 The Ada to HTML Converter gnathtml
346 * Invoking gnathtml::
347 * Installing gnathtml::
349 GNAT and Program Execution
351 * Running and Debugging Ada Programs::
353 * Improving Performance::
354 * Overflow Check Handling in GNAT::
355 * Performing Dimensionality Analysis in GNAT::
356 * Stack Related Facilities::
357 * Memory Management Issues::
359 Running and Debugging Ada Programs
361 * The GNAT Debugger GDB::
363 * Introduction to GDB Commands::
364 * Using Ada Expressions::
365 * Calling User-Defined Subprograms::
366 * Using the next Command in a Function::
367 * Stopping When Ada Exceptions Are Raised::
369 * Debugging Generic Units::
370 * Remote Debugging with gdbserver::
371 * GNAT Abnormal Termination or Failure to Terminate::
372 * Naming Conventions for GNAT Source Files::
373 * Getting Internal Debugging Information::
375 * Pretty-Printers for the GNAT runtime::
379 * Non-Symbolic Traceback::
380 * Symbolic Traceback::
384 * Profiling an Ada Program with gprof::
386 Profiling an Ada Program with gprof
388 * Compilation for profiling::
389 * Program execution::
391 * Interpretation of profiling results::
393 Improving Performance
395 * Performance Considerations::
396 * Text_IO Suggestions::
397 * Reducing Size of Executables with Unused Subprogram/Data Elimination::
399 Performance Considerations
401 * Controlling Run-Time Checks::
402 * Use of Restrictions::
403 * Optimization Levels::
404 * Debugging Optimized Code::
405 * Inlining of Subprograms::
406 * Floating_Point_Operations::
407 * Vectorization of loops::
408 * Other Optimization Switches::
409 * Optimization and Strict Aliasing::
410 * Aliased Variables and Optimization::
411 * Atomic Variables and Optimization::
412 * Passive Task Optimization::
414 Reducing Size of Executables with Unused Subprogram/Data Elimination
416 * About unused subprogram/data elimination::
417 * Compilation options::
418 * Example of unused subprogram/data elimination::
420 Overflow Check Handling in GNAT
423 * Management of Overflows in GNAT::
424 * Specifying the Desired Mode::
426 * Implementation Notes::
428 Stack Related Facilities
430 * Stack Overflow Checking::
431 * Static Stack Usage Analysis::
432 * Dynamic Stack Usage Analysis::
434 Memory Management Issues
436 * Some Useful Memory Pools::
437 * The GNAT Debug Pool Facility::
439 Platform-Specific Information
441 * Run-Time Libraries::
442 * Specifying a Run-Time Library::
444 * Microsoft Windows Topics::
449 * Summary of Run-Time Configurations::
451 Specifying a Run-Time Library
453 * Choosing the Scheduling Policy::
457 * Required Packages on GNU/Linux::
459 Microsoft Windows Topics
461 * Using GNAT on Windows::
462 * Using a network installation of GNAT::
463 * CONSOLE and WINDOWS subsystems::
465 * Disabling Command Line Argument Expansion::
466 * Mixed-Language Programming on Windows::
467 * Windows Specific Add-Ons::
469 Mixed-Language Programming on Windows
471 * Windows Calling Conventions::
472 * Introduction to Dynamic Link Libraries (DLLs): Introduction to Dynamic Link Libraries DLLs.
473 * Using DLLs with GNAT::
474 * Building DLLs with GNAT Project files::
475 * Building DLLs with GNAT::
476 * Building DLLs with gnatdll::
477 * Ada DLLs and Finalization::
478 * Creating a Spec for Ada DLLs::
479 * GNAT and Windows Resources::
480 * Using GNAT DLLs from Microsoft Visual Studio Applications::
482 * Setting Stack Size from gnatlink::
483 * Setting Heap Size from gnatlink::
485 Windows Calling Conventions
487 * C Calling Convention::
488 * Stdcall Calling Convention::
489 * Win32 Calling Convention::
490 * DLL Calling Convention::
494 * Creating an Ada Spec for the DLL Services::
495 * Creating an Import Library::
497 Building DLLs with gnatdll
499 * Limitations When Using Ada DLLs from Ada::
500 * Exporting Ada Entities::
501 * Ada DLLs and Elaboration::
503 Creating a Spec for Ada DLLs
505 * Creating the Definition File::
508 GNAT and Windows Resources
510 * Building Resources::
511 * Compiling Resources::
516 * Program and DLL Both Built with GCC/GNAT::
517 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
519 Windows Specific Add-Ons
526 * Codesigning the Debugger::
528 Elaboration Order Handling in GNAT
531 * Elaboration Order::
532 * Checking the Elaboration Order::
533 * Controlling the Elaboration Order in Ada::
534 * Controlling the Elaboration Order in GNAT::
535 * Common Elaboration-model Traits::
536 * Dynamic Elaboration Model in GNAT::
537 * Static Elaboration Model in GNAT::
538 * SPARK Elaboration Model in GNAT::
539 * Legacy Elaboration Model in GNAT::
540 * Mixing Elaboration Models::
541 * Elaboration Circularities::
542 * Resolving Elaboration Circularities::
543 * Resolving Task Issues::
544 * Elaboration-related Compiler Switches::
545 * Summary of Procedures for Elaboration Control::
546 * Inspecting the Chosen Elaboration Order::
550 * Basic Assembler Syntax::
551 * A Simple Example of Inline Assembler::
552 * Output Variables in Inline Assembler::
553 * Input Variables in Inline Assembler::
554 * Inlining Inline Assembler Code::
555 * Other Asm Functionality::
557 Other Asm Functionality
559 * The Clobber Parameter::
560 * The Volatile Parameter::
565 @node About This Guide,Getting Started with GNAT,Top,Top
566 @anchor{gnat_ugn/about_this_guide about-this-guide}@anchor{2}@anchor{gnat_ugn/about_this_guide doc}@anchor{3}@anchor{gnat_ugn/about_this_guide gnat-user-s-guide-for-native-platforms}@anchor{4}@anchor{gnat_ugn/about_this_guide id1}@anchor{5}
567 @chapter About This Guide
571 This guide describes the use of GNAT,
572 a compiler and software development
573 toolset for the full Ada programming language.
574 It documents the features of the compiler and tools, and explains
575 how to use them to build Ada applications.
577 GNAT implements Ada 95, Ada 2005 and Ada 2012, and it may also be
578 invoked in Ada 83 compatibility mode.
579 By default, GNAT assumes Ada 2012, but you can override with a
580 compiler switch (@ref{6,,Compiling Different Versions of Ada})
581 to explicitly specify the language version.
582 Throughout this manual, references to 'Ada' without a year suffix
583 apply to all Ada 95/2005/2012 versions of the language.
586 * What This Guide Contains::
587 * What You Should Know before Reading This Guide::
588 * Related Information::
589 * A Note to Readers of Previous Versions of the Manual::
594 @node What This Guide Contains,What You Should Know before Reading This Guide,,About This Guide
595 @anchor{gnat_ugn/about_this_guide what-this-guide-contains}@anchor{7}
596 @section What This Guide Contains
599 This guide contains the following chapters:
605 @ref{8,,Getting Started with GNAT} describes how to get started compiling
606 and running Ada programs with the GNAT Ada programming environment.
609 @ref{9,,The GNAT Compilation Model} describes the compilation model used
613 @ref{a,,Building Executable Programs with GNAT} describes how to use the
614 main GNAT tools to build executable programs, and it also gives examples of
615 using the GNU make utility with GNAT.
618 @ref{b,,GNAT Utility Programs} explains the various utility programs that
619 are included in the GNAT environment
622 @ref{c,,GNAT and Program Execution} covers a number of topics related to
623 running, debugging, and tuning the performace of programs developed
627 Appendices cover several additional topics:
633 @ref{d,,Platform-Specific Information} describes the different run-time
634 library implementations and also presents information on how to use
635 GNAT on several specific platforms
638 @ref{e,,Example of Binder Output File} shows the source code for the binder
639 output file for a sample program.
642 @ref{f,,Elaboration Order Handling in GNAT} describes how GNAT helps
643 you deal with elaboration order issues.
646 @ref{10,,Inline Assembler} shows how to use the inline assembly facility
650 @node What You Should Know before Reading This Guide,Related Information,What This Guide Contains,About This Guide
651 @anchor{gnat_ugn/about_this_guide what-you-should-know-before-reading-this-guide}@anchor{11}
652 @section What You Should Know before Reading This Guide
655 @geindex Ada 95 Language Reference Manual
657 @geindex Ada 2005 Language Reference Manual
659 This guide assumes a basic familiarity with the Ada 95 language, as
660 described in the International Standard ANSI/ISO/IEC-8652:1995, January
662 It does not require knowledge of the features introduced by Ada 2005
664 Reference manuals for Ada 95, Ada 2005, and Ada 2012 are included in
665 the GNAT documentation package.
667 @node Related Information,A Note to Readers of Previous Versions of the Manual,What You Should Know before Reading This Guide,About This Guide
668 @anchor{gnat_ugn/about_this_guide related-information}@anchor{12}
669 @section Related Information
672 For further information about Ada and related tools, please refer to the
679 @cite{Ada 95 Reference Manual}, @cite{Ada 2005 Reference Manual}, and
680 @cite{Ada 2012 Reference Manual}, which contain reference
681 material for the several revisions of the Ada language standard.
684 @cite{GNAT Reference_Manual}, which contains all reference material for the GNAT
685 implementation of Ada.
688 @cite{Using the GNAT Programming Studio}, which describes the GPS
689 Integrated Development Environment.
692 @cite{GNAT Programming Studio Tutorial}, which introduces the
693 main GPS features through examples.
696 @cite{Debugging with GDB},
697 for all details on the use of the GNU source-level debugger.
700 @cite{GNU Emacs Manual},
701 for full information on the extensible editor and programming
705 @node A Note to Readers of Previous Versions of the Manual,Conventions,Related Information,About This Guide
706 @anchor{gnat_ugn/about_this_guide a-note-to-readers-of-previous-versions-of-the-manual}@anchor{13}
707 @section A Note to Readers of Previous Versions of the Manual
710 In early 2015 the GNAT manuals were transitioned to the
711 reStructuredText (rst) / Sphinx documentation generator technology.
712 During that process the @cite{GNAT User's Guide} was reorganized
713 so that related topics would be described together in the same chapter
714 or appendix. Here's a summary of the major changes realized in
715 the new document structure.
721 @ref{9,,The GNAT Compilation Model} has been extended so that it now covers
722 the following material:
728 The @code{gnatname}, @code{gnatkr}, and @code{gnatchop} tools
731 @ref{14,,Configuration Pragmas}
734 @ref{15,,GNAT and Libraries}
737 @ref{16,,Conditional Compilation} including @ref{17,,Preprocessing with gnatprep}
738 and @ref{18,,Integrated Preprocessing}
741 @ref{19,,Generating Ada Bindings for C and C++ headers}
744 @ref{1a,,Using GNAT Files with External Tools}
748 @ref{a,,Building Executable Programs with GNAT} is a new chapter consolidating
749 the following content:
755 @ref{1b,,Building with gnatmake}
758 @ref{1c,,Compiling with gcc}
761 @ref{1d,,Binding with gnatbind}
764 @ref{1e,,Linking with gnatlink}
767 @ref{1f,,Using the GNU make Utility}
771 @ref{b,,GNAT Utility Programs} is a new chapter consolidating the information about several
779 @ref{20,,The File Cleanup Utility gnatclean}
782 @ref{21,,The GNAT Library Browser gnatls}
785 @ref{22,,The Cross-Referencing Tools gnatxref and gnatfind}
788 @ref{23,,The Ada to HTML Converter gnathtml}
792 @ref{c,,GNAT and Program Execution} is a new chapter consolidating the following:
798 @ref{24,,Running and Debugging Ada Programs}
804 @ref{26,,Improving Performance}
807 @ref{27,,Overflow Check Handling in GNAT}
810 @ref{28,,Performing Dimensionality Analysis in GNAT}
813 @ref{29,,Stack Related Facilities}
816 @ref{2a,,Memory Management Issues}
820 @ref{d,,Platform-Specific Information} is a new appendix consolidating the following:
826 @ref{2b,,Run-Time Libraries}
829 @ref{2c,,Microsoft Windows Topics}
832 @ref{2d,,Mac OS Topics}
836 The @emph{Compatibility and Porting Guide} appendix has been moved to the
837 @cite{GNAT Reference Manual}. It now includes a section
838 @emph{Writing Portable Fixed-Point Declarations} which was previously
839 a separate chapter in the @cite{GNAT User's Guide}.
842 @node Conventions,,A Note to Readers of Previous Versions of the Manual,About This Guide
843 @anchor{gnat_ugn/about_this_guide conventions}@anchor{2e}
848 @geindex typographical
850 @geindex Typographical conventions
852 Following are examples of the typographical and graphic conventions used
859 @code{Functions}, @code{utility program names}, @code{standard names},
875 [optional information or parameters]
878 Examples are described by text
881 and then shown this way.
885 Commands that are entered by the user are shown as preceded by a prompt string
886 comprising the @code{$} character followed by a space.
889 Full file names are shown with the '/' character
890 as the directory separator; e.g., @code{parent-dir/subdir/myfile.adb}.
891 If you are using GNAT on a Windows platform, please note that
892 the '\' character should be used instead.
895 @node Getting Started with GNAT,The GNAT Compilation Model,About This Guide,Top
896 @anchor{gnat_ugn/getting_started_with_gnat getting-started-with-gnat}@anchor{8}@anchor{gnat_ugn/getting_started_with_gnat doc}@anchor{2f}@anchor{gnat_ugn/getting_started_with_gnat id1}@anchor{30}
897 @chapter Getting Started with GNAT
900 This chapter describes how to use GNAT's command line interface to build
901 executable Ada programs.
902 On most platforms a visually oriented Integrated Development Environment
903 is also available, the GNAT Programming Studio (GPS).
904 GPS offers a graphical "look and feel", support for development in
905 other programming languages, comprehensive browsing features, and
906 many other capabilities.
907 For information on GPS please refer to
908 @cite{Using the GNAT Programming Studio}.
912 * Running a Simple Ada Program::
913 * Running a Program with Multiple Units::
914 * Using the gnatmake Utility::
918 @node Running GNAT,Running a Simple Ada Program,,Getting Started with GNAT
919 @anchor{gnat_ugn/getting_started_with_gnat running-gnat}@anchor{31}@anchor{gnat_ugn/getting_started_with_gnat id2}@anchor{32}
920 @section Running GNAT
923 Three steps are needed to create an executable file from an Ada source
930 The source file(s) must be compiled.
933 The file(s) must be bound using the GNAT binder.
936 All appropriate object files must be linked to produce an executable.
939 All three steps are most commonly handled by using the @code{gnatmake}
940 utility program that, given the name of the main program, automatically
941 performs the necessary compilation, binding and linking steps.
943 @node Running a Simple Ada Program,Running a Program with Multiple Units,Running GNAT,Getting Started with GNAT
944 @anchor{gnat_ugn/getting_started_with_gnat running-a-simple-ada-program}@anchor{33}@anchor{gnat_ugn/getting_started_with_gnat id3}@anchor{34}
945 @section Running a Simple Ada Program
948 Any text editor may be used to prepare an Ada program.
949 (If Emacs is used, the optional Ada mode may be helpful in laying out the
951 The program text is a normal text file. We will assume in our initial
952 example that you have used your editor to prepare the following
953 standard format text file:
956 with Ada.Text_IO; use Ada.Text_IO;
959 Put_Line ("Hello WORLD!");
963 This file should be named @code{hello.adb}.
964 With the normal default file naming conventions, GNAT requires
966 contain a single compilation unit whose file name is the
968 with periods replaced by hyphens; the
969 extension is @code{ads} for a
970 spec and @code{adb} for a body.
971 You can override this default file naming convention by use of the
972 special pragma @code{Source_File_Name} (for further information please
973 see @ref{35,,Using Other File Names}).
974 Alternatively, if you want to rename your files according to this default
975 convention, which is probably more convenient if you will be using GNAT
976 for all your compilations, then the @code{gnatchop} utility
977 can be used to generate correctly-named source files
978 (see @ref{36,,Renaming Files with gnatchop}).
980 You can compile the program using the following command (@code{$} is used
981 as the command prompt in the examples in this document):
987 @code{gcc} is the command used to run the compiler. This compiler is
988 capable of compiling programs in several languages, including Ada and
989 C. It assumes that you have given it an Ada program if the file extension is
990 either @code{.ads} or @code{.adb}, and it will then call
991 the GNAT compiler to compile the specified file.
993 The @code{-c} switch is required. It tells @code{gcc} to only do a
994 compilation. (For C programs, @code{gcc} can also do linking, but this
995 capability is not used directly for Ada programs, so the @code{-c}
996 switch must always be present.)
998 This compile command generates a file
999 @code{hello.o}, which is the object
1000 file corresponding to your Ada program. It also generates
1001 an 'Ada Library Information' file @code{hello.ali},
1002 which contains additional information used to check
1003 that an Ada program is consistent.
1004 To build an executable file,
1005 use @code{gnatbind} to bind the program
1006 and @code{gnatlink} to link it. The
1007 argument to both @code{gnatbind} and @code{gnatlink} is the name of the
1008 @code{ALI} file, but the default extension of @code{.ali} can
1009 be omitted. This means that in the most common case, the argument
1010 is simply the name of the main program:
1017 A simpler method of carrying out these steps is to use @code{gnatmake},
1018 a master program that invokes all the required
1019 compilation, binding and linking tools in the correct order. In particular,
1020 @code{gnatmake} automatically recompiles any sources that have been
1021 modified since they were last compiled, or sources that depend
1022 on such modified sources, so that 'version skew' is avoided.
1024 @geindex Version skew (avoided by `@w{`}gnatmake`@w{`})
1027 $ gnatmake hello.adb
1030 The result is an executable program called @code{hello}, which can be
1037 assuming that the current directory is on the search path
1038 for executable programs.
1040 and, if all has gone well, you will see:
1046 appear in response to this command.
1048 @node Running a Program with Multiple Units,Using the gnatmake Utility,Running a Simple Ada Program,Getting Started with GNAT
1049 @anchor{gnat_ugn/getting_started_with_gnat id4}@anchor{37}@anchor{gnat_ugn/getting_started_with_gnat running-a-program-with-multiple-units}@anchor{38}
1050 @section Running a Program with Multiple Units
1053 Consider a slightly more complicated example that has three files: a
1054 main program, and the spec and body of a package:
1057 package Greetings is
1062 with Ada.Text_IO; use Ada.Text_IO;
1063 package body Greetings is
1066 Put_Line ("Hello WORLD!");
1069 procedure Goodbye is
1071 Put_Line ("Goodbye WORLD!");
1083 Following the one-unit-per-file rule, place this program in the
1084 following three separate files:
1089 @item @emph{greetings.ads}
1091 spec of package @code{Greetings}
1093 @item @emph{greetings.adb}
1095 body of package @code{Greetings}
1097 @item @emph{gmain.adb}
1099 body of main program
1102 To build an executable version of
1103 this program, we could use four separate steps to compile, bind, and link
1104 the program, as follows:
1108 $ gcc -c greetings.adb
1113 Note that there is no required order of compilation when using GNAT.
1114 In particular it is perfectly fine to compile the main program first.
1115 Also, it is not necessary to compile package specs in the case where
1116 there is an accompanying body; you only need to compile the body. If you want
1117 to submit these files to the compiler for semantic checking and not code
1118 generation, then use the @code{-gnatc} switch:
1121 $ gcc -c greetings.ads -gnatc
1124 Although the compilation can be done in separate steps as in the
1125 above example, in practice it is almost always more convenient
1126 to use the @code{gnatmake} tool. All you need to know in this case
1127 is the name of the main program's source file. The effect of the above four
1128 commands can be achieved with a single one:
1131 $ gnatmake gmain.adb
1134 In the next section we discuss the advantages of using @code{gnatmake} in
1137 @node Using the gnatmake Utility,,Running a Program with Multiple Units,Getting Started with GNAT
1138 @anchor{gnat_ugn/getting_started_with_gnat using-the-gnatmake-utility}@anchor{39}@anchor{gnat_ugn/getting_started_with_gnat id5}@anchor{3a}
1139 @section Using the @code{gnatmake} Utility
1142 If you work on a program by compiling single components at a time using
1143 @code{gcc}, you typically keep track of the units you modify. In order to
1144 build a consistent system, you compile not only these units, but also any
1145 units that depend on the units you have modified.
1146 For example, in the preceding case,
1147 if you edit @code{gmain.adb}, you only need to recompile that file. But if
1148 you edit @code{greetings.ads}, you must recompile both
1149 @code{greetings.adb} and @code{gmain.adb}, because both files contain
1150 units that depend on @code{greetings.ads}.
1152 @code{gnatbind} will warn you if you forget one of these compilation
1153 steps, so that it is impossible to generate an inconsistent program as a
1154 result of forgetting to do a compilation. Nevertheless it is tedious and
1155 error-prone to keep track of dependencies among units.
1156 One approach to handle the dependency-bookkeeping is to use a
1157 makefile. However, makefiles present maintenance problems of their own:
1158 if the dependencies change as you change the program, you must make
1159 sure that the makefile is kept up-to-date manually, which is also an
1160 error-prone process.
1162 The @code{gnatmake} utility takes care of these details automatically.
1163 Invoke it using either one of the following forms:
1166 $ gnatmake gmain.adb
1170 The argument is the name of the file containing the main program;
1171 you may omit the extension. @code{gnatmake}
1172 examines the environment, automatically recompiles any files that need
1173 recompiling, and binds and links the resulting set of object files,
1174 generating the executable file, @code{gmain}.
1175 In a large program, it
1176 can be extremely helpful to use @code{gnatmake}, because working out by hand
1177 what needs to be recompiled can be difficult.
1179 Note that @code{gnatmake} takes into account all the Ada rules that
1180 establish dependencies among units. These include dependencies that result
1181 from inlining subprogram bodies, and from
1182 generic instantiation. Unlike some other
1183 Ada make tools, @code{gnatmake} does not rely on the dependencies that were
1184 found by the compiler on a previous compilation, which may possibly
1185 be wrong when sources change. @code{gnatmake} determines the exact set of
1186 dependencies from scratch each time it is run.
1188 @c -- Example: A |withing| unit has a |with| clause, it |withs| a |withed| unit
1190 @node The GNAT Compilation Model,Building Executable Programs with GNAT,Getting Started with GNAT,Top
1191 @anchor{gnat_ugn/the_gnat_compilation_model doc}@anchor{3b}@anchor{gnat_ugn/the_gnat_compilation_model the-gnat-compilation-model}@anchor{9}@anchor{gnat_ugn/the_gnat_compilation_model id1}@anchor{3c}
1192 @chapter The GNAT Compilation Model
1195 @geindex GNAT compilation model
1197 @geindex Compilation model
1199 This chapter describes the compilation model used by GNAT. Although
1200 similar to that used by other languages such as C and C++, this model
1201 is substantially different from the traditional Ada compilation models,
1202 which are based on a centralized program library. The chapter covers
1203 the following material:
1209 Topics related to source file makeup and naming
1215 @ref{3d,,Source Representation}
1218 @ref{3e,,Foreign Language Representation}
1221 @ref{3f,,File Naming Topics and Utilities}
1225 @ref{14,,Configuration Pragmas}
1228 @ref{40,,Generating Object Files}
1231 @ref{41,,Source Dependencies}
1234 @ref{42,,The Ada Library Information Files}
1237 @ref{43,,Binding an Ada Program}
1240 @ref{15,,GNAT and Libraries}
1243 @ref{16,,Conditional Compilation}
1246 @ref{44,,Mixed Language Programming}
1249 @ref{45,,GNAT and Other Compilation Models}
1252 @ref{1a,,Using GNAT Files with External Tools}
1256 * Source Representation::
1257 * Foreign Language Representation::
1258 * File Naming Topics and Utilities::
1259 * Configuration Pragmas::
1260 * Generating Object Files::
1261 * Source Dependencies::
1262 * The Ada Library Information Files::
1263 * Binding an Ada Program::
1264 * GNAT and Libraries::
1265 * Conditional Compilation::
1266 * Mixed Language Programming::
1267 * GNAT and Other Compilation Models::
1268 * Using GNAT Files with External Tools::
1272 @node Source Representation,Foreign Language Representation,,The GNAT Compilation Model
1273 @anchor{gnat_ugn/the_gnat_compilation_model source-representation}@anchor{3d}@anchor{gnat_ugn/the_gnat_compilation_model id2}@anchor{46}
1274 @section Source Representation
1285 Ada source programs are represented in standard text files, using
1286 Latin-1 coding. Latin-1 is an 8-bit code that includes the familiar
1287 7-bit ASCII set, plus additional characters used for
1288 representing foreign languages (see @ref{3e,,Foreign Language Representation}
1289 for support of non-USA character sets). The format effector characters
1290 are represented using their standard ASCII encodings, as follows:
1295 @multitable {xxxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxx} {xxxxxxxxxxxxx}
1372 Source files are in standard text file format. In addition, GNAT will
1373 recognize a wide variety of stream formats, in which the end of
1374 physical lines is marked by any of the following sequences:
1375 @code{LF}, @code{CR}, @code{CR-LF}, or @code{LF-CR}. This is useful
1376 in accommodating files that are imported from other operating systems.
1378 @geindex End of source file; Source file@comma{} end
1380 @geindex SUB (control character)
1382 The end of a source file is normally represented by the physical end of
1383 file. However, the control character @code{16#1A#} (@code{SUB}) is also
1384 recognized as signalling the end of the source file. Again, this is
1385 provided for compatibility with other operating systems where this
1386 code is used to represent the end of file.
1388 @geindex spec (definition)
1389 @geindex compilation (definition)
1391 Each file contains a single Ada compilation unit, including any pragmas
1392 associated with the unit. For example, this means you must place a
1393 package declaration (a package @emph{spec}) and the corresponding body in
1394 separate files. An Ada @emph{compilation} (which is a sequence of
1395 compilation units) is represented using a sequence of files. Similarly,
1396 you will place each subunit or child unit in a separate file.
1398 @node Foreign Language Representation,File Naming Topics and Utilities,Source Representation,The GNAT Compilation Model
1399 @anchor{gnat_ugn/the_gnat_compilation_model foreign-language-representation}@anchor{3e}@anchor{gnat_ugn/the_gnat_compilation_model id3}@anchor{47}
1400 @section Foreign Language Representation
1403 GNAT supports the standard character sets defined in Ada as well as
1404 several other non-standard character sets for use in localized versions
1405 of the compiler (@ref{48,,Character Set Control}).
1409 * Other 8-Bit Codes::
1410 * Wide_Character Encodings::
1411 * Wide_Wide_Character Encodings::
1415 @node Latin-1,Other 8-Bit Codes,,Foreign Language Representation
1416 @anchor{gnat_ugn/the_gnat_compilation_model id4}@anchor{49}@anchor{gnat_ugn/the_gnat_compilation_model latin-1}@anchor{4a}
1422 The basic character set is Latin-1. This character set is defined by ISO
1423 standard 8859, part 1. The lower half (character codes @code{16#00#}
1424 ... @code{16#7F#)} is identical to standard ASCII coding, but the upper
1425 half is used to represent additional characters. These include extended letters
1426 used by European languages, such as French accents, the vowels with umlauts
1427 used in German, and the extra letter A-ring used in Swedish.
1429 @geindex Ada.Characters.Latin_1
1431 For a complete list of Latin-1 codes and their encodings, see the source
1432 file of library unit @code{Ada.Characters.Latin_1} in file
1433 @code{a-chlat1.ads}.
1434 You may use any of these extended characters freely in character or
1435 string literals. In addition, the extended characters that represent
1436 letters can be used in identifiers.
1438 @node Other 8-Bit Codes,Wide_Character Encodings,Latin-1,Foreign Language Representation
1439 @anchor{gnat_ugn/the_gnat_compilation_model other-8-bit-codes}@anchor{4b}@anchor{gnat_ugn/the_gnat_compilation_model id5}@anchor{4c}
1440 @subsection Other 8-Bit Codes
1443 GNAT also supports several other 8-bit coding schemes:
1452 @item @emph{ISO 8859-2 (Latin-2)}
1454 Latin-2 letters allowed in identifiers, with uppercase and lowercase
1465 @item @emph{ISO 8859-3 (Latin-3)}
1467 Latin-3 letters allowed in identifiers, with uppercase and lowercase
1478 @item @emph{ISO 8859-4 (Latin-4)}
1480 Latin-4 letters allowed in identifiers, with uppercase and lowercase
1491 @item @emph{ISO 8859-5 (Cyrillic)}
1493 ISO 8859-5 letters (Cyrillic) allowed in identifiers, with uppercase and
1494 lowercase equivalence.
1497 @geindex ISO 8859-15
1504 @item @emph{ISO 8859-15 (Latin-9)}
1506 ISO 8859-15 (Latin-9) letters allowed in identifiers, with uppercase and
1507 lowercase equivalence
1510 @geindex code page 437 (IBM PC)
1515 @item @emph{IBM PC (code page 437)}
1517 This code page is the normal default for PCs in the U.S. It corresponds
1518 to the original IBM PC character set. This set has some, but not all, of
1519 the extended Latin-1 letters, but these letters do not have the same
1520 encoding as Latin-1. In this mode, these letters are allowed in
1521 identifiers with uppercase and lowercase equivalence.
1524 @geindex code page 850 (IBM PC)
1529 @item @emph{IBM PC (code page 850)}
1531 This code page is a modification of 437 extended to include all the
1532 Latin-1 letters, but still not with the usual Latin-1 encoding. In this
1533 mode, all these letters are allowed in identifiers with uppercase and
1534 lowercase equivalence.
1536 @item @emph{Full Upper 8-bit}
1538 Any character in the range 80-FF allowed in identifiers, and all are
1539 considered distinct. In other words, there are no uppercase and lowercase
1540 equivalences in this range. This is useful in conjunction with
1541 certain encoding schemes used for some foreign character sets (e.g.,
1542 the typical method of representing Chinese characters on the PC).
1544 @item @emph{No Upper-Half}
1546 No upper-half characters in the range 80-FF are allowed in identifiers.
1547 This gives Ada 83 compatibility for identifier names.
1550 For precise data on the encodings permitted, and the uppercase and lowercase
1551 equivalences that are recognized, see the file @code{csets.adb} in
1552 the GNAT compiler sources. You will need to obtain a full source release
1553 of GNAT to obtain this file.
1555 @node Wide_Character Encodings,Wide_Wide_Character Encodings,Other 8-Bit Codes,Foreign Language Representation
1556 @anchor{gnat_ugn/the_gnat_compilation_model id6}@anchor{4d}@anchor{gnat_ugn/the_gnat_compilation_model wide-character-encodings}@anchor{4e}
1557 @subsection Wide_Character Encodings
1560 GNAT allows wide character codes to appear in character and string
1561 literals, and also optionally in identifiers, by means of the following
1562 possible encoding schemes:
1567 @item @emph{Hex Coding}
1569 In this encoding, a wide character is represented by the following five
1576 where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1577 characters (using uppercase letters) of the wide character code. For
1578 example, ESC A345 is used to represent the wide character with code
1580 This scheme is compatible with use of the full Wide_Character set.
1582 @item @emph{Upper-Half Coding}
1584 @geindex Upper-Half Coding
1586 The wide character with encoding @code{16#abcd#} where the upper bit is on
1587 (in other words, 'a' is in the range 8-F) is represented as two bytes,
1588 @code{16#ab#} and @code{16#cd#}. The second byte cannot be a format control
1589 character, but is not required to be in the upper half. This method can
1590 be also used for shift-JIS or EUC, where the internal coding matches the
1593 @item @emph{Shift JIS Coding}
1595 @geindex Shift JIS Coding
1597 A wide character is represented by a two-character sequence,
1599 @code{16#cd#}, with the restrictions described for upper-half encoding as
1600 described above. The internal character code is the corresponding JIS
1601 character according to the standard algorithm for Shift-JIS
1602 conversion. Only characters defined in the JIS code set table can be
1603 used with this encoding method.
1605 @item @emph{EUC Coding}
1609 A wide character is represented by a two-character sequence
1611 @code{16#cd#}, with both characters being in the upper half. The internal
1612 character code is the corresponding JIS character according to the EUC
1613 encoding algorithm. Only characters defined in the JIS code set table
1614 can be used with this encoding method.
1616 @item @emph{UTF-8 Coding}
1618 A wide character is represented using
1619 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
1620 10646-1/Am.2. Depending on the character value, the representation
1621 is a one, two, or three byte sequence:
1624 16#0000#-16#007f#: 2#0xxxxxxx#
1625 16#0080#-16#07ff#: 2#110xxxxx# 2#10xxxxxx#
1626 16#0800#-16#ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
1629 where the @code{xxx} bits correspond to the left-padded bits of the
1630 16-bit character value. Note that all lower half ASCII characters
1631 are represented as ASCII bytes and all upper half characters and
1632 other wide characters are represented as sequences of upper-half
1633 (The full UTF-8 scheme allows for encoding 31-bit characters as
1634 6-byte sequences, and in the following section on wide wide
1635 characters, the use of these sequences is documented).
1637 @item @emph{Brackets Coding}
1639 In this encoding, a wide character is represented by the following eight
1646 where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1647 characters (using uppercase letters) of the wide character code. For
1648 example, ['A345'] is used to represent the wide character with code
1649 @code{16#A345#}. It is also possible (though not required) to use the
1650 Brackets coding for upper half characters. For example, the code
1651 @code{16#A3#} can be represented as @code{['A3']}.
1653 This scheme is compatible with use of the full Wide_Character set,
1654 and is also the method used for wide character encoding in some standard
1655 ACATS (Ada Conformity Assessment Test Suite) test suite distributions.
1660 Some of these coding schemes do not permit the full use of the
1661 Ada character set. For example, neither Shift JIS nor EUC allow the
1662 use of the upper half of the Latin-1 set.
1666 @node Wide_Wide_Character Encodings,,Wide_Character Encodings,Foreign Language Representation
1667 @anchor{gnat_ugn/the_gnat_compilation_model id7}@anchor{4f}@anchor{gnat_ugn/the_gnat_compilation_model wide-wide-character-encodings}@anchor{50}
1668 @subsection Wide_Wide_Character Encodings
1671 GNAT allows wide wide character codes to appear in character and string
1672 literals, and also optionally in identifiers, by means of the following
1673 possible encoding schemes:
1678 @item @emph{UTF-8 Coding}
1680 A wide character is represented using
1681 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
1682 10646-1/Am.2. Depending on the character value, the representation
1683 of character codes with values greater than 16#FFFF# is a
1684 is a four, five, or six byte sequence:
1687 16#01_0000#-16#10_FFFF#: 11110xxx 10xxxxxx 10xxxxxx
1689 16#0020_0000#-16#03FF_FFFF#: 111110xx 10xxxxxx 10xxxxxx
1691 16#0400_0000#-16#7FFF_FFFF#: 1111110x 10xxxxxx 10xxxxxx
1692 10xxxxxx 10xxxxxx 10xxxxxx
1695 where the @code{xxx} bits correspond to the left-padded bits of the
1696 32-bit character value.
1698 @item @emph{Brackets Coding}
1700 In this encoding, a wide wide character is represented by the following ten or
1701 twelve byte character sequence:
1705 [ " a b c d e f g h " ]
1708 where @code{a-h} are the six or eight hexadecimal
1709 characters (using uppercase letters) of the wide wide character code. For
1710 example, ["1F4567"] is used to represent the wide wide character with code
1711 @code{16#001F_4567#}.
1713 This scheme is compatible with use of the full Wide_Wide_Character set,
1714 and is also the method used for wide wide character encoding in some standard
1715 ACATS (Ada Conformity Assessment Test Suite) test suite distributions.
1718 @node File Naming Topics and Utilities,Configuration Pragmas,Foreign Language Representation,The GNAT Compilation Model
1719 @anchor{gnat_ugn/the_gnat_compilation_model id8}@anchor{51}@anchor{gnat_ugn/the_gnat_compilation_model file-naming-topics-and-utilities}@anchor{3f}
1720 @section File Naming Topics and Utilities
1723 GNAT has a default file naming scheme and also provides the user with
1724 a high degree of control over how the names and extensions of the
1725 source files correspond to the Ada compilation units that they contain.
1728 * File Naming Rules::
1729 * Using Other File Names::
1730 * Alternative File Naming Schemes::
1731 * Handling Arbitrary File Naming Conventions with gnatname::
1732 * File Name Krunching with gnatkr::
1733 * Renaming Files with gnatchop::
1737 @node File Naming Rules,Using Other File Names,,File Naming Topics and Utilities
1738 @anchor{gnat_ugn/the_gnat_compilation_model file-naming-rules}@anchor{52}@anchor{gnat_ugn/the_gnat_compilation_model id9}@anchor{53}
1739 @subsection File Naming Rules
1742 The default file name is determined by the name of the unit that the
1743 file contains. The name is formed by taking the full expanded name of
1744 the unit and replacing the separating dots with hyphens and using
1745 lowercase for all letters.
1747 An exception arises if the file name generated by the above rules starts
1748 with one of the characters
1749 @code{a}, @code{g}, @code{i}, or @code{s}, and the second character is a
1750 minus. In this case, the character tilde is used in place
1751 of the minus. The reason for this special rule is to avoid clashes with
1752 the standard names for child units of the packages System, Ada,
1753 Interfaces, and GNAT, which use the prefixes
1754 @code{s-}, @code{a-}, @code{i-}, and @code{g-},
1757 The file extension is @code{.ads} for a spec and
1758 @code{.adb} for a body. The following table shows some
1759 examples of these rules.
1764 @multitable {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
1771 Ada Compilation Unit
1791 @code{arith_functions.ads}
1795 Arith_Functions (package spec)
1799 @code{arith_functions.adb}
1803 Arith_Functions (package body)
1807 @code{func-spec.ads}
1811 Func.Spec (child package spec)
1815 @code{func-spec.adb}
1819 Func.Spec (child package body)
1827 Sub (subunit of Main)
1835 A.Bad (child package body)
1841 Following these rules can result in excessively long
1842 file names if corresponding
1843 unit names are long (for example, if child units or subunits are
1844 heavily nested). An option is available to shorten such long file names
1845 (called file name 'krunching'). This may be particularly useful when
1846 programs being developed with GNAT are to be used on operating systems
1847 with limited file name lengths. @ref{54,,Using gnatkr}.
1849 Of course, no file shortening algorithm can guarantee uniqueness over
1850 all possible unit names; if file name krunching is used, it is your
1851 responsibility to ensure no name clashes occur. Alternatively you
1852 can specify the exact file names that you want used, as described
1853 in the next section. Finally, if your Ada programs are migrating from a
1854 compiler with a different naming convention, you can use the gnatchop
1855 utility to produce source files that follow the GNAT naming conventions.
1856 (For details see @ref{36,,Renaming Files with gnatchop}.)
1858 Note: in the case of Windows or Mac OS operating systems, case is not
1859 significant. So for example on Windows if the canonical name is
1860 @code{main-sub.adb}, you can use the file name @code{Main-Sub.adb} instead.
1861 However, case is significant for other operating systems, so for example,
1862 if you want to use other than canonically cased file names on a Unix system,
1863 you need to follow the procedures described in the next section.
1865 @node Using Other File Names,Alternative File Naming Schemes,File Naming Rules,File Naming Topics and Utilities
1866 @anchor{gnat_ugn/the_gnat_compilation_model id10}@anchor{55}@anchor{gnat_ugn/the_gnat_compilation_model using-other-file-names}@anchor{35}
1867 @subsection Using Other File Names
1872 In the previous section, we have described the default rules used by
1873 GNAT to determine the file name in which a given unit resides. It is
1874 often convenient to follow these default rules, and if you follow them,
1875 the compiler knows without being explicitly told where to find all
1878 @geindex Source_File_Name pragma
1880 However, in some cases, particularly when a program is imported from
1881 another Ada compiler environment, it may be more convenient for the
1882 programmer to specify which file names contain which units. GNAT allows
1883 arbitrary file names to be used by means of the Source_File_Name pragma.
1884 The form of this pragma is as shown in the following examples:
1887 pragma Source_File_Name (My_Utilities.Stacks,
1888 Spec_File_Name => "myutilst_a.ada");
1889 pragma Source_File_name (My_Utilities.Stacks,
1890 Body_File_Name => "myutilst.ada");
1893 As shown in this example, the first argument for the pragma is the unit
1894 name (in this example a child unit). The second argument has the form
1895 of a named association. The identifier
1896 indicates whether the file name is for a spec or a body;
1897 the file name itself is given by a string literal.
1899 The source file name pragma is a configuration pragma, which means that
1900 normally it will be placed in the @code{gnat.adc}
1901 file used to hold configuration
1902 pragmas that apply to a complete compilation environment.
1903 For more details on how the @code{gnat.adc} file is created and used
1904 see @ref{56,,Handling of Configuration Pragmas}.
1908 GNAT allows completely arbitrary file names to be specified using the
1909 source file name pragma. However, if the file name specified has an
1910 extension other than @code{.ads} or @code{.adb} it is necessary to use
1911 a special syntax when compiling the file. The name in this case must be
1912 preceded by the special sequence @code{-x} followed by a space and the name
1913 of the language, here @code{ada}, as in:
1916 $ gcc -c -x ada peculiar_file_name.sim
1919 @code{gnatmake} handles non-standard file names in the usual manner (the
1920 non-standard file name for the main program is simply used as the
1921 argument to gnatmake). Note that if the extension is also non-standard,
1922 then it must be included in the @code{gnatmake} command, it may not
1925 @node Alternative File Naming Schemes,Handling Arbitrary File Naming Conventions with gnatname,Using Other File Names,File Naming Topics and Utilities
1926 @anchor{gnat_ugn/the_gnat_compilation_model id11}@anchor{57}@anchor{gnat_ugn/the_gnat_compilation_model alternative-file-naming-schemes}@anchor{58}
1927 @subsection Alternative File Naming Schemes
1930 @geindex File naming schemes
1931 @geindex alternative
1935 The previous section described the use of the @code{Source_File_Name}
1936 pragma to allow arbitrary names to be assigned to individual source files.
1937 However, this approach requires one pragma for each file, and especially in
1938 large systems can result in very long @code{gnat.adc} files, and also create
1939 a maintenance problem.
1941 @geindex Source_File_Name pragma
1943 GNAT also provides a facility for specifying systematic file naming schemes
1944 other than the standard default naming scheme previously described. An
1945 alternative scheme for naming is specified by the use of
1946 @code{Source_File_Name} pragmas having the following format:
1949 pragma Source_File_Name (
1950 Spec_File_Name => FILE_NAME_PATTERN
1951 [ , Casing => CASING_SPEC]
1952 [ , Dot_Replacement => STRING_LITERAL ] );
1954 pragma Source_File_Name (
1955 Body_File_Name => FILE_NAME_PATTERN
1956 [ , Casing => CASING_SPEC ]
1957 [ , Dot_Replacement => STRING_LITERAL ] ) ;
1959 pragma Source_File_Name (
1960 Subunit_File_Name => FILE_NAME_PATTERN
1961 [ , Casing => CASING_SPEC ]
1962 [ , Dot_Replacement => STRING_LITERAL ] ) ;
1964 FILE_NAME_PATTERN ::= STRING_LITERAL
1965 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
1968 The @code{FILE_NAME_PATTERN} string shows how the file name is constructed.
1969 It contains a single asterisk character, and the unit name is substituted
1970 systematically for this asterisk. The optional parameter
1971 @code{Casing} indicates
1972 whether the unit name is to be all upper-case letters, all lower-case letters,
1973 or mixed-case. If no
1974 @code{Casing} parameter is used, then the default is all
1977 The optional @code{Dot_Replacement} string is used to replace any periods
1978 that occur in subunit or child unit names. If no @code{Dot_Replacement}
1979 argument is used then separating dots appear unchanged in the resulting
1981 Although the above syntax indicates that the
1982 @code{Casing} argument must appear
1983 before the @code{Dot_Replacement} argument, but it
1984 is also permissible to write these arguments in the opposite order.
1986 As indicated, it is possible to specify different naming schemes for
1987 bodies, specs, and subunits. Quite often the rule for subunits is the
1988 same as the rule for bodies, in which case, there is no need to give
1989 a separate @code{Subunit_File_Name} rule, and in this case the
1990 @code{Body_File_name} rule is used for subunits as well.
1992 The separate rule for subunits can also be used to implement the rather
1993 unusual case of a compilation environment (e.g., a single directory) which
1994 contains a subunit and a child unit with the same unit name. Although
1995 both units cannot appear in the same partition, the Ada Reference Manual
1996 allows (but does not require) the possibility of the two units coexisting
1997 in the same environment.
1999 The file name translation works in the following steps:
2005 If there is a specific @code{Source_File_Name} pragma for the given unit,
2006 then this is always used, and any general pattern rules are ignored.
2009 If there is a pattern type @code{Source_File_Name} pragma that applies to
2010 the unit, then the resulting file name will be used if the file exists. If
2011 more than one pattern matches, the latest one will be tried first, and the
2012 first attempt resulting in a reference to a file that exists will be used.
2015 If no pattern type @code{Source_File_Name} pragma that applies to the unit
2016 for which the corresponding file exists, then the standard GNAT default
2017 naming rules are used.
2020 As an example of the use of this mechanism, consider a commonly used scheme
2021 in which file names are all lower case, with separating periods copied
2022 unchanged to the resulting file name, and specs end with @code{.1.ada}, and
2023 bodies end with @code{.2.ada}. GNAT will follow this scheme if the following
2027 pragma Source_File_Name
2028 (Spec_File_Name => ".1.ada");
2029 pragma Source_File_Name
2030 (Body_File_Name => ".2.ada");
2033 The default GNAT scheme is actually implemented by providing the following
2034 default pragmas internally:
2037 pragma Source_File_Name
2038 (Spec_File_Name => ".ads", Dot_Replacement => "-");
2039 pragma Source_File_Name
2040 (Body_File_Name => ".adb", Dot_Replacement => "-");
2043 Our final example implements a scheme typically used with one of the
2044 Ada 83 compilers, where the separator character for subunits was '__'
2045 (two underscores), specs were identified by adding @code{_.ADA}, bodies
2046 by adding @code{.ADA}, and subunits by
2047 adding @code{.SEP}. All file names were
2048 upper case. Child units were not present of course since this was an
2049 Ada 83 compiler, but it seems reasonable to extend this scheme to use
2050 the same double underscore separator for child units.
2053 pragma Source_File_Name
2054 (Spec_File_Name => "_.ADA",
2055 Dot_Replacement => "__",
2056 Casing = Uppercase);
2057 pragma Source_File_Name
2058 (Body_File_Name => ".ADA",
2059 Dot_Replacement => "__",
2060 Casing = Uppercase);
2061 pragma Source_File_Name
2062 (Subunit_File_Name => ".SEP",
2063 Dot_Replacement => "__",
2064 Casing = Uppercase);
2069 @node Handling Arbitrary File Naming Conventions with gnatname,File Name Krunching with gnatkr,Alternative File Naming Schemes,File Naming Topics and Utilities
2070 @anchor{gnat_ugn/the_gnat_compilation_model handling-arbitrary-file-naming-conventions-with-gnatname}@anchor{59}@anchor{gnat_ugn/the_gnat_compilation_model id12}@anchor{5a}
2071 @subsection Handling Arbitrary File Naming Conventions with @code{gnatname}
2074 @geindex File Naming Conventions
2077 * Arbitrary File Naming Conventions::
2078 * Running gnatname::
2079 * Switches for gnatname::
2080 * Examples of gnatname Usage::
2084 @node Arbitrary File Naming Conventions,Running gnatname,,Handling Arbitrary File Naming Conventions with gnatname
2085 @anchor{gnat_ugn/the_gnat_compilation_model arbitrary-file-naming-conventions}@anchor{5b}@anchor{gnat_ugn/the_gnat_compilation_model id13}@anchor{5c}
2086 @subsubsection Arbitrary File Naming Conventions
2089 The GNAT compiler must be able to know the source file name of a compilation
2090 unit. When using the standard GNAT default file naming conventions
2091 (@code{.ads} for specs, @code{.adb} for bodies), the GNAT compiler
2092 does not need additional information.
2094 When the source file names do not follow the standard GNAT default file naming
2095 conventions, the GNAT compiler must be given additional information through
2096 a configuration pragmas file (@ref{14,,Configuration Pragmas})
2098 When the non-standard file naming conventions are well-defined,
2099 a small number of pragmas @code{Source_File_Name} specifying a naming pattern
2100 (@ref{58,,Alternative File Naming Schemes}) may be sufficient. However,
2101 if the file naming conventions are irregular or arbitrary, a number
2102 of pragma @code{Source_File_Name} for individual compilation units
2104 To help maintain the correspondence between compilation unit names and
2105 source file names within the compiler,
2106 GNAT provides a tool @code{gnatname} to generate the required pragmas for a
2109 @node Running gnatname,Switches for gnatname,Arbitrary File Naming Conventions,Handling Arbitrary File Naming Conventions with gnatname
2110 @anchor{gnat_ugn/the_gnat_compilation_model running-gnatname}@anchor{5d}@anchor{gnat_ugn/the_gnat_compilation_model id14}@anchor{5e}
2111 @subsubsection Running @code{gnatname}
2114 The usual form of the @code{gnatname} command is:
2117 $ gnatname [ switches ] naming_pattern [ naming_patterns ]
2118 [--and [ switches ] naming_pattern [ naming_patterns ]]
2121 All of the arguments are optional. If invoked without any argument,
2122 @code{gnatname} will display its usage.
2124 When used with at least one naming pattern, @code{gnatname} will attempt to
2125 find all the compilation units in files that follow at least one of the
2126 naming patterns. To find these compilation units,
2127 @code{gnatname} will use the GNAT compiler in syntax-check-only mode on all
2130 One or several Naming Patterns may be given as arguments to @code{gnatname}.
2131 Each Naming Pattern is enclosed between double quotes (or single
2133 A Naming Pattern is a regular expression similar to the wildcard patterns
2134 used in file names by the Unix shells or the DOS prompt.
2136 @code{gnatname} may be called with several sections of directories/patterns.
2137 Sections are separated by the switch @code{--and}. In each section, there must be
2138 at least one pattern. If no directory is specified in a section, the current
2139 directory (or the project directory if @code{-P} is used) is implied.
2140 The options other that the directory switches and the patterns apply globally
2141 even if they are in different sections.
2143 Examples of Naming Patterns are:
2151 For a more complete description of the syntax of Naming Patterns,
2152 see the second kind of regular expressions described in @code{g-regexp.ads}
2153 (the 'Glob' regular expressions).
2155 When invoked without the switch @code{-P}, @code{gnatname} will create a
2156 configuration pragmas file @code{gnat.adc} in the current working directory,
2157 with pragmas @code{Source_File_Name} for each file that contains a valid Ada
2160 @node Switches for gnatname,Examples of gnatname Usage,Running gnatname,Handling Arbitrary File Naming Conventions with gnatname
2161 @anchor{gnat_ugn/the_gnat_compilation_model id15}@anchor{5f}@anchor{gnat_ugn/the_gnat_compilation_model switches-for-gnatname}@anchor{60}
2162 @subsubsection Switches for @code{gnatname}
2165 Switches for @code{gnatname} must precede any specified Naming Pattern.
2167 You may specify any of the following switches to @code{gnatname}:
2169 @geindex --version (gnatname)
2174 @item @code{--version}
2176 Display Copyright and version, then exit disregarding all other options.
2179 @geindex --help (gnatname)
2186 If @code{--version} was not used, display usage, then exit disregarding
2189 @item @code{--subdirs=@emph{dir}}
2191 Real object, library or exec directories are subdirectories <dir> of the
2194 @item @code{--no-backup}
2196 Do not create a backup copy of an existing project file.
2200 Start another section of directories/patterns.
2203 @geindex -c (gnatname)
2208 @item @code{-c@emph{filename}}
2210 Create a configuration pragmas file @code{filename} (instead of the default
2212 There may be zero, one or more space between @code{-c} and
2214 @code{filename} may include directory information. @code{filename} must be
2215 writable. There may be only one switch @code{-c}.
2216 When a switch @code{-c} is
2217 specified, no switch @code{-P} may be specified (see below).
2220 @geindex -d (gnatname)
2225 @item @code{-d@emph{dir}}
2227 Look for source files in directory @code{dir}. There may be zero, one or more
2228 spaces between @code{-d} and @code{dir}.
2229 @code{dir} may end with @code{/**}, that is it may be of the form
2230 @code{root_dir/**}. In this case, the directory @code{root_dir} and all of its
2231 subdirectories, recursively, have to be searched for sources.
2232 When a switch @code{-d}
2233 is specified, the current working directory will not be searched for source
2234 files, unless it is explicitly specified with a @code{-d}
2235 or @code{-D} switch.
2236 Several switches @code{-d} may be specified.
2237 If @code{dir} is a relative path, it is relative to the directory of
2238 the configuration pragmas file specified with switch
2240 or to the directory of the project file specified with switch
2242 if neither switch @code{-c}
2243 nor switch @code{-P} are specified, it is relative to the
2244 current working directory. The directory
2245 specified with switch @code{-d} must exist and be readable.
2248 @geindex -D (gnatname)
2253 @item @code{-D@emph{filename}}
2255 Look for source files in all directories listed in text file @code{filename}.
2256 There may be zero, one or more spaces between @code{-D}
2257 and @code{filename}.
2258 @code{filename} must be an existing, readable text file.
2259 Each nonempty line in @code{filename} must be a directory.
2260 Specifying switch @code{-D} is equivalent to specifying as many
2261 switches @code{-d} as there are nonempty lines in
2266 Follow symbolic links when processing project files.
2268 @geindex -f (gnatname)
2270 @item @code{-f@emph{pattern}}
2272 Foreign patterns. Using this switch, it is possible to add sources of languages
2273 other than Ada to the list of sources of a project file.
2274 It is only useful if a -P switch is used.
2278 gnatname -Pprj -f"*.c" "*.ada"
2281 will look for Ada units in all files with the @code{.ada} extension,
2282 and will add to the list of file for project @code{prj.gpr} the C files
2283 with extension @code{.c}.
2285 @geindex -h (gnatname)
2289 Output usage (help) information. The output is written to @code{stdout}.
2291 @geindex -P (gnatname)
2293 @item @code{-P@emph{proj}}
2295 Create or update project file @code{proj}. There may be zero, one or more space
2296 between @code{-P} and @code{proj}. @code{proj} may include directory
2297 information. @code{proj} must be writable.
2298 There may be only one switch @code{-P}.
2299 When a switch @code{-P} is specified,
2300 no switch @code{-c} may be specified.
2301 On all platforms, except on VMS, when @code{gnatname} is invoked for an
2302 existing project file <proj>.gpr, a backup copy of the project file is created
2303 in the project directory with file name <proj>.gpr.saved_x. 'x' is the first
2304 non negative number that makes this backup copy a new file.
2306 @geindex -v (gnatname)
2310 Verbose mode. Output detailed explanation of behavior to @code{stdout}.
2311 This includes name of the file written, the name of the directories to search
2312 and, for each file in those directories whose name matches at least one of
2313 the Naming Patterns, an indication of whether the file contains a unit,
2314 and if so the name of the unit.
2317 @geindex -v -v (gnatname)
2324 Very Verbose mode. In addition to the output produced in verbose mode,
2325 for each file in the searched directories whose name matches none of
2326 the Naming Patterns, an indication is given that there is no match.
2328 @geindex -x (gnatname)
2330 @item @code{-x@emph{pattern}}
2332 Excluded patterns. Using this switch, it is possible to exclude some files
2333 that would match the name patterns. For example,
2336 gnatname -x "*_nt.ada" "*.ada"
2339 will look for Ada units in all files with the @code{.ada} extension,
2340 except those whose names end with @code{_nt.ada}.
2343 @node Examples of gnatname Usage,,Switches for gnatname,Handling Arbitrary File Naming Conventions with gnatname
2344 @anchor{gnat_ugn/the_gnat_compilation_model examples-of-gnatname-usage}@anchor{61}@anchor{gnat_ugn/the_gnat_compilation_model id16}@anchor{62}
2345 @subsubsection Examples of @code{gnatname} Usage
2349 $ gnatname -c /home/me/names.adc -d sources "[a-z]*.ada*"
2352 In this example, the directory @code{/home/me} must already exist
2353 and be writable. In addition, the directory
2354 @code{/home/me/sources} (specified by
2355 @code{-d sources}) must exist and be readable.
2357 Note the optional spaces after @code{-c} and @code{-d}.
2360 $ gnatname -P/home/me/proj -x "*_nt_body.ada"
2361 -dsources -dsources/plus -Dcommon_dirs.txt "body_*" "spec_*"
2364 Note that several switches @code{-d} may be used,
2365 even in conjunction with one or several switches
2366 @code{-D}. Several Naming Patterns and one excluded pattern
2367 are used in this example.
2369 @node File Name Krunching with gnatkr,Renaming Files with gnatchop,Handling Arbitrary File Naming Conventions with gnatname,File Naming Topics and Utilities
2370 @anchor{gnat_ugn/the_gnat_compilation_model file-name-krunching-with-gnatkr}@anchor{63}@anchor{gnat_ugn/the_gnat_compilation_model id17}@anchor{64}
2371 @subsection File Name Krunching with @code{gnatkr}
2376 This section discusses the method used by the compiler to shorten
2377 the default file names chosen for Ada units so that they do not
2378 exceed the maximum length permitted. It also describes the
2379 @code{gnatkr} utility that can be used to determine the result of
2380 applying this shortening.
2385 * Krunching Method::
2386 * Examples of gnatkr Usage::
2390 @node About gnatkr,Using gnatkr,,File Name Krunching with gnatkr
2391 @anchor{gnat_ugn/the_gnat_compilation_model id18}@anchor{65}@anchor{gnat_ugn/the_gnat_compilation_model about-gnatkr}@anchor{66}
2392 @subsubsection About @code{gnatkr}
2395 The default file naming rule in GNAT
2396 is that the file name must be derived from
2397 the unit name. The exact default rule is as follows:
2403 Take the unit name and replace all dots by hyphens.
2406 If such a replacement occurs in the
2407 second character position of a name, and the first character is
2408 @code{a}, @code{g}, @code{s}, or @code{i},
2409 then replace the dot by the character
2413 The reason for this exception is to avoid clashes
2414 with the standard names for children of System, Ada, Interfaces,
2415 and GNAT, which use the prefixes
2416 @code{s-}, @code{a-}, @code{i-}, and @code{g-},
2420 The @code{-gnatk@emph{nn}}
2421 switch of the compiler activates a 'krunching'
2422 circuit that limits file names to nn characters (where nn is a decimal
2425 The @code{gnatkr} utility can be used to determine the krunched name for
2426 a given file, when krunched to a specified maximum length.
2428 @node Using gnatkr,Krunching Method,About gnatkr,File Name Krunching with gnatkr
2429 @anchor{gnat_ugn/the_gnat_compilation_model id19}@anchor{67}@anchor{gnat_ugn/the_gnat_compilation_model using-gnatkr}@anchor{54}
2430 @subsubsection Using @code{gnatkr}
2433 The @code{gnatkr} command has the form:
2436 $ gnatkr name [ length ]
2439 @code{name} is the uncrunched file name, derived from the name of the unit
2440 in the standard manner described in the previous section (i.e., in particular
2441 all dots are replaced by hyphens). The file name may or may not have an
2442 extension (defined as a suffix of the form period followed by arbitrary
2443 characters other than period). If an extension is present then it will
2444 be preserved in the output. For example, when krunching @code{hellofile.ads}
2445 to eight characters, the result will be hellofil.ads.
2447 Note: for compatibility with previous versions of @code{gnatkr} dots may
2448 appear in the name instead of hyphens, but the last dot will always be
2449 taken as the start of an extension. So if @code{gnatkr} is given an argument
2450 such as @code{Hello.World.adb} it will be treated exactly as if the first
2451 period had been a hyphen, and for example krunching to eight characters
2452 gives the result @code{hellworl.adb}.
2454 Note that the result is always all lower case.
2455 Characters of the other case are folded as required.
2457 @code{length} represents the length of the krunched name. The default
2458 when no argument is given is 8 characters. A length of zero stands for
2459 unlimited, in other words do not chop except for system files where the
2460 implied crunching length is always eight characters.
2462 The output is the krunched name. The output has an extension only if the
2463 original argument was a file name with an extension.
2465 @node Krunching Method,Examples of gnatkr Usage,Using gnatkr,File Name Krunching with gnatkr
2466 @anchor{gnat_ugn/the_gnat_compilation_model id20}@anchor{68}@anchor{gnat_ugn/the_gnat_compilation_model krunching-method}@anchor{69}
2467 @subsubsection Krunching Method
2470 The initial file name is determined by the name of the unit that the file
2471 contains. The name is formed by taking the full expanded name of the
2472 unit and replacing the separating dots with hyphens and
2474 for all letters, except that a hyphen in the second character position is
2475 replaced by a tilde if the first character is
2476 @code{a}, @code{i}, @code{g}, or @code{s}.
2477 The extension is @code{.ads} for a
2478 spec and @code{.adb} for a body.
2479 Krunching does not affect the extension, but the file name is shortened to
2480 the specified length by following these rules:
2486 The name is divided into segments separated by hyphens, tildes or
2487 underscores and all hyphens, tildes, and underscores are
2488 eliminated. If this leaves the name short enough, we are done.
2491 If the name is too long, the longest segment is located (left-most
2492 if there are two of equal length), and shortened by dropping
2493 its last character. This is repeated until the name is short enough.
2495 As an example, consider the krunching of @code{our-strings-wide_fixed.adb}
2496 to fit the name into 8 characters as required by some operating systems:
2499 our-strings-wide_fixed 22
2500 our strings wide fixed 19
2501 our string wide fixed 18
2502 our strin wide fixed 17
2503 our stri wide fixed 16
2504 our stri wide fixe 15
2505 our str wide fixe 14
2512 Final file name: oustwifi.adb
2516 The file names for all predefined units are always krunched to eight
2517 characters. The krunching of these predefined units uses the following
2518 special prefix replacements:
2521 @multitable {xxxxxxxxxxxxxxxxxxxxxxx} {xxxxxxxxxxxxxxxx}
2565 These system files have a hyphen in the second character position. That
2566 is why normal user files replace such a character with a
2567 tilde, to avoid confusion with system file names.
2569 As an example of this special rule, consider
2570 @code{ada-strings-wide_fixed.adb}, which gets krunched as follows:
2573 ada-strings-wide_fixed 22
2574 a- strings wide fixed 18
2575 a- string wide fixed 17
2576 a- strin wide fixed 16
2577 a- stri wide fixed 15
2578 a- stri wide fixe 14
2585 Final file name: a-stwifi.adb
2589 Of course no file shortening algorithm can guarantee uniqueness over all
2590 possible unit names, and if file name krunching is used then it is your
2591 responsibility to ensure that no name clashes occur. The utility
2592 program @code{gnatkr} is supplied for conveniently determining the
2593 krunched name of a file.
2595 @node Examples of gnatkr Usage,,Krunching Method,File Name Krunching with gnatkr
2596 @anchor{gnat_ugn/the_gnat_compilation_model id21}@anchor{6a}@anchor{gnat_ugn/the_gnat_compilation_model examples-of-gnatkr-usage}@anchor{6b}
2597 @subsubsection Examples of @code{gnatkr} Usage
2601 $ gnatkr very_long_unit_name.ads --> velounna.ads
2602 $ gnatkr grandparent-parent-child.ads --> grparchi.ads
2603 $ gnatkr Grandparent.Parent.Child.ads --> grparchi.ads
2604 $ gnatkr grandparent-parent-child --> grparchi
2605 $ gnatkr very_long_unit_name.ads/count=6 --> vlunna.ads
2606 $ gnatkr very_long_unit_name.ads/count=0 --> very_long_unit_name.ads
2609 @node Renaming Files with gnatchop,,File Name Krunching with gnatkr,File Naming Topics and Utilities
2610 @anchor{gnat_ugn/the_gnat_compilation_model id22}@anchor{6c}@anchor{gnat_ugn/the_gnat_compilation_model renaming-files-with-gnatchop}@anchor{36}
2611 @subsection Renaming Files with @code{gnatchop}
2616 This section discusses how to handle files with multiple units by using
2617 the @code{gnatchop} utility. This utility is also useful in renaming
2618 files to meet the standard GNAT default file naming conventions.
2621 * Handling Files with Multiple Units::
2622 * Operating gnatchop in Compilation Mode::
2623 * Command Line for gnatchop::
2624 * Switches for gnatchop::
2625 * Examples of gnatchop Usage::
2629 @node Handling Files with Multiple Units,Operating gnatchop in Compilation Mode,,Renaming Files with gnatchop
2630 @anchor{gnat_ugn/the_gnat_compilation_model id23}@anchor{6d}@anchor{gnat_ugn/the_gnat_compilation_model handling-files-with-multiple-units}@anchor{6e}
2631 @subsubsection Handling Files with Multiple Units
2634 The basic compilation model of GNAT requires that a file submitted to the
2635 compiler have only one unit and there be a strict correspondence
2636 between the file name and the unit name.
2638 The @code{gnatchop} utility allows both of these rules to be relaxed,
2639 allowing GNAT to process files which contain multiple compilation units
2640 and files with arbitrary file names. @code{gnatchop}
2641 reads the specified file and generates one or more output files,
2642 containing one unit per file. The unit and the file name correspond,
2643 as required by GNAT.
2645 If you want to permanently restructure a set of 'foreign' files so that
2646 they match the GNAT rules, and do the remaining development using the
2647 GNAT structure, you can simply use @code{gnatchop} once, generate the
2648 new set of files and work with them from that point on.
2650 Alternatively, if you want to keep your files in the 'foreign' format,
2651 perhaps to maintain compatibility with some other Ada compilation
2652 system, you can set up a procedure where you use @code{gnatchop} each
2653 time you compile, regarding the source files that it writes as temporary
2654 files that you throw away.
2656 Note that if your file containing multiple units starts with a byte order
2657 mark (BOM) specifying UTF-8 encoding, then the files generated by gnatchop
2658 will each start with a copy of this BOM, meaning that they can be compiled
2659 automatically in UTF-8 mode without needing to specify an explicit encoding.
2661 @node Operating gnatchop in Compilation Mode,Command Line for gnatchop,Handling Files with Multiple Units,Renaming Files with gnatchop
2662 @anchor{gnat_ugn/the_gnat_compilation_model operating-gnatchop-in-compilation-mode}@anchor{6f}@anchor{gnat_ugn/the_gnat_compilation_model id24}@anchor{70}
2663 @subsubsection Operating gnatchop in Compilation Mode
2666 The basic function of @code{gnatchop} is to take a file with multiple units
2667 and split it into separate files. The boundary between files is reasonably
2668 clear, except for the issue of comments and pragmas. In default mode, the
2669 rule is that any pragmas between units belong to the previous unit, except
2670 that configuration pragmas always belong to the following unit. Any comments
2671 belong to the following unit. These rules
2672 almost always result in the right choice of
2673 the split point without needing to mark it explicitly and most users will
2674 find this default to be what they want. In this default mode it is incorrect to
2675 submit a file containing only configuration pragmas, or one that ends in
2676 configuration pragmas, to @code{gnatchop}.
2678 However, using a special option to activate 'compilation mode',
2680 can perform another function, which is to provide exactly the semantics
2681 required by the RM for handling of configuration pragmas in a compilation.
2682 In the absence of configuration pragmas (at the main file level), this
2683 option has no effect, but it causes such configuration pragmas to be handled
2684 in a quite different manner.
2686 First, in compilation mode, if @code{gnatchop} is given a file that consists of
2687 only configuration pragmas, then this file is appended to the
2688 @code{gnat.adc} file in the current directory. This behavior provides
2689 the required behavior described in the RM for the actions to be taken
2690 on submitting such a file to the compiler, namely that these pragmas
2691 should apply to all subsequent compilations in the same compilation
2692 environment. Using GNAT, the current directory, possibly containing a
2693 @code{gnat.adc} file is the representation
2694 of a compilation environment. For more information on the
2695 @code{gnat.adc} file, see @ref{56,,Handling of Configuration Pragmas}.
2697 Second, in compilation mode, if @code{gnatchop}
2698 is given a file that starts with
2699 configuration pragmas, and contains one or more units, then these
2700 configuration pragmas are prepended to each of the chopped files. This
2701 behavior provides the required behavior described in the RM for the
2702 actions to be taken on compiling such a file, namely that the pragmas
2703 apply to all units in the compilation, but not to subsequently compiled
2706 Finally, if configuration pragmas appear between units, they are appended
2707 to the previous unit. This results in the previous unit being illegal,
2708 since the compiler does not accept configuration pragmas that follow
2709 a unit. This provides the required RM behavior that forbids configuration
2710 pragmas other than those preceding the first compilation unit of a
2713 For most purposes, @code{gnatchop} will be used in default mode. The
2714 compilation mode described above is used only if you need exactly
2715 accurate behavior with respect to compilations, and you have files
2716 that contain multiple units and configuration pragmas. In this
2717 circumstance the use of @code{gnatchop} with the compilation mode
2718 switch provides the required behavior, and is for example the mode
2719 in which GNAT processes the ACVC tests.
2721 @node Command Line for gnatchop,Switches for gnatchop,Operating gnatchop in Compilation Mode,Renaming Files with gnatchop
2722 @anchor{gnat_ugn/the_gnat_compilation_model id25}@anchor{71}@anchor{gnat_ugn/the_gnat_compilation_model command-line-for-gnatchop}@anchor{72}
2723 @subsubsection Command Line for @code{gnatchop}
2726 The @code{gnatchop} command has the form:
2729 $ gnatchop switches file_name [file_name ...]
2733 The only required argument is the file name of the file to be chopped.
2734 There are no restrictions on the form of this file name. The file itself
2735 contains one or more Ada units, in normal GNAT format, concatenated
2736 together. As shown, more than one file may be presented to be chopped.
2738 When run in default mode, @code{gnatchop} generates one output file in
2739 the current directory for each unit in each of the files.
2741 @code{directory}, if specified, gives the name of the directory to which
2742 the output files will be written. If it is not specified, all files are
2743 written to the current directory.
2745 For example, given a
2746 file called @code{hellofiles} containing
2751 with Ada.Text_IO; use Ada.Text_IO;
2761 $ gnatchop hellofiles
2764 generates two files in the current directory, one called
2765 @code{hello.ads} containing the single line that is the procedure spec,
2766 and the other called @code{hello.adb} containing the remaining text. The
2767 original file is not affected. The generated files can be compiled in
2770 When gnatchop is invoked on a file that is empty or that contains only empty
2771 lines and/or comments, gnatchop will not fail, but will not produce any
2774 For example, given a
2775 file called @code{toto.txt} containing
2787 will not produce any new file and will result in the following warnings:
2790 toto.txt:1:01: warning: empty file, contains no compilation units
2791 no compilation units found
2792 no source files written
2795 @node Switches for gnatchop,Examples of gnatchop Usage,Command Line for gnatchop,Renaming Files with gnatchop
2796 @anchor{gnat_ugn/the_gnat_compilation_model switches-for-gnatchop}@anchor{73}@anchor{gnat_ugn/the_gnat_compilation_model id26}@anchor{74}
2797 @subsubsection Switches for @code{gnatchop}
2800 @code{gnatchop} recognizes the following switches:
2802 @geindex --version (gnatchop)
2807 @item @code{--version}
2809 Display Copyright and version, then exit disregarding all other options.
2812 @geindex --help (gnatchop)
2819 If @code{--version} was not used, display usage, then exit disregarding
2823 @geindex -c (gnatchop)
2830 Causes @code{gnatchop} to operate in compilation mode, in which
2831 configuration pragmas are handled according to strict RM rules. See
2832 previous section for a full description of this mode.
2834 @item @code{-gnat@emph{xxx}}
2836 This passes the given @code{-gnat@emph{xxx}} switch to @code{gnat} which is
2837 used to parse the given file. Not all @emph{xxx} options make sense,
2838 but for example, the use of @code{-gnati2} allows @code{gnatchop} to
2839 process a source file that uses Latin-2 coding for identifiers.
2843 Causes @code{gnatchop} to generate a brief help summary to the standard
2844 output file showing usage information.
2847 @geindex -k (gnatchop)
2852 @item @code{-k@emph{mm}}
2854 Limit generated file names to the specified number @code{mm}
2856 This is useful if the
2857 resulting set of files is required to be interoperable with systems
2858 which limit the length of file names.
2859 No space is allowed between the @code{-k} and the numeric value. The numeric
2860 value may be omitted in which case a default of @code{-k8},
2862 with DOS-like file systems, is used. If no @code{-k} switch
2864 there is no limit on the length of file names.
2867 @geindex -p (gnatchop)
2874 Causes the file modification time stamp of the input file to be
2875 preserved and used for the time stamp of the output file(s). This may be
2876 useful for preserving coherency of time stamps in an environment where
2877 @code{gnatchop} is used as part of a standard build process.
2880 @geindex -q (gnatchop)
2887 Causes output of informational messages indicating the set of generated
2888 files to be suppressed. Warnings and error messages are unaffected.
2891 @geindex -r (gnatchop)
2893 @geindex Source_Reference pragmas
2900 Generate @code{Source_Reference} pragmas. Use this switch if the output
2901 files are regarded as temporary and development is to be done in terms
2902 of the original unchopped file. This switch causes
2903 @code{Source_Reference} pragmas to be inserted into each of the
2904 generated files to refers back to the original file name and line number.
2905 The result is that all error messages refer back to the original
2907 In addition, the debugging information placed into the object file (when
2908 the @code{-g} switch of @code{gcc} or @code{gnatmake} is
2910 also refers back to this original file so that tools like profilers and
2911 debuggers will give information in terms of the original unchopped file.
2913 If the original file to be chopped itself contains
2914 a @code{Source_Reference}
2915 pragma referencing a third file, then gnatchop respects
2916 this pragma, and the generated @code{Source_Reference} pragmas
2917 in the chopped file refer to the original file, with appropriate
2918 line numbers. This is particularly useful when @code{gnatchop}
2919 is used in conjunction with @code{gnatprep} to compile files that
2920 contain preprocessing statements and multiple units.
2923 @geindex -v (gnatchop)
2930 Causes @code{gnatchop} to operate in verbose mode. The version
2931 number and copyright notice are output, as well as exact copies of
2932 the gnat1 commands spawned to obtain the chop control information.
2935 @geindex -w (gnatchop)
2942 Overwrite existing file names. Normally @code{gnatchop} regards it as a
2943 fatal error if there is already a file with the same name as a
2944 file it would otherwise output, in other words if the files to be
2945 chopped contain duplicated units. This switch bypasses this
2946 check, and causes all but the last instance of such duplicated
2947 units to be skipped.
2950 @geindex --GCC= (gnatchop)
2955 @item @code{--GCC=@emph{xxxx}}
2957 Specify the path of the GNAT parser to be used. When this switch is used,
2958 no attempt is made to add the prefix to the GNAT parser executable.
2961 @node Examples of gnatchop Usage,,Switches for gnatchop,Renaming Files with gnatchop
2962 @anchor{gnat_ugn/the_gnat_compilation_model id27}@anchor{75}@anchor{gnat_ugn/the_gnat_compilation_model examples-of-gnatchop-usage}@anchor{76}
2963 @subsubsection Examples of @code{gnatchop} Usage
2967 $ gnatchop -w hello_s.ada prerelease/files
2970 Chops the source file @code{hello_s.ada}. The output files will be
2971 placed in the directory @code{prerelease/files},
2973 files with matching names in that directory (no files in the current
2974 directory are modified).
2980 Chops the source file @code{archive}
2981 into the current directory. One
2982 useful application of @code{gnatchop} is in sending sets of sources
2983 around, for example in email messages. The required sources are simply
2984 concatenated (for example, using a Unix @code{cat}
2986 @code{gnatchop} is used at the other end to reconstitute the original
2990 $ gnatchop file1 file2 file3 direc
2993 Chops all units in files @code{file1}, @code{file2}, @code{file3}, placing
2994 the resulting files in the directory @code{direc}. Note that if any units
2995 occur more than once anywhere within this set of files, an error message
2996 is generated, and no files are written. To override this check, use the
2998 in which case the last occurrence in the last file will
2999 be the one that is output, and earlier duplicate occurrences for a given
3000 unit will be skipped.
3002 @node Configuration Pragmas,Generating Object Files,File Naming Topics and Utilities,The GNAT Compilation Model
3003 @anchor{gnat_ugn/the_gnat_compilation_model id28}@anchor{77}@anchor{gnat_ugn/the_gnat_compilation_model configuration-pragmas}@anchor{14}
3004 @section Configuration Pragmas
3007 @geindex Configuration pragmas
3010 @geindex configuration
3012 Configuration pragmas include those pragmas described as
3013 such in the Ada Reference Manual, as well as
3014 implementation-dependent pragmas that are configuration pragmas.
3015 See the @code{Implementation_Defined_Pragmas} chapter in the
3016 @cite{GNAT_Reference_Manual} for details on these
3017 additional GNAT-specific configuration pragmas.
3018 Most notably, the pragma @code{Source_File_Name}, which allows
3019 specifying non-default names for source files, is a configuration
3020 pragma. The following is a complete list of configuration pragmas
3030 Allow_Integer_Address
3033 Assume_No_Invalid_Values
3035 Check_Float_Overflow
3039 Compile_Time_Warning
3041 Compiler_Unit_Warning
3043 Convention_Identifier
3046 Default_Scalar_Storage_Order
3047 Default_Storage_Pool
3048 Disable_Atomic_Synchronization
3052 Enable_Atomic_Synchronization
3055 External_Name_Casing
3064 No_Component_Reordering
3065 No_Heap_Finalization
3071 Overriding_Renamings
3072 Partition_Elaboration_Policy
3075 Prefix_Exception_Messages
3076 Priority_Specific_Dispatching
3079 Propagate_Exceptions
3086 Restrictions_Warnings
3088 Short_Circuit_And_Or
3091 Source_File_Name_Project
3095 Suppress_Exception_Locations
3096 Task_Dispatching_Policy
3097 Unevaluated_Use_Of_Old
3104 Wide_Character_Encoding
3108 * Handling of Configuration Pragmas::
3109 * The Configuration Pragmas Files::
3113 @node Handling of Configuration Pragmas,The Configuration Pragmas Files,,Configuration Pragmas
3114 @anchor{gnat_ugn/the_gnat_compilation_model id29}@anchor{78}@anchor{gnat_ugn/the_gnat_compilation_model handling-of-configuration-pragmas}@anchor{56}
3115 @subsection Handling of Configuration Pragmas
3118 Configuration pragmas may either appear at the start of a compilation
3119 unit, or they can appear in a configuration pragma file to apply to
3120 all compilations performed in a given compilation environment.
3122 GNAT also provides the @code{gnatchop} utility to provide an automatic
3123 way to handle configuration pragmas following the semantics for
3124 compilations (that is, files with multiple units), described in the RM.
3125 See @ref{6f,,Operating gnatchop in Compilation Mode} for details.
3126 However, for most purposes, it will be more convenient to edit the
3127 @code{gnat.adc} file that contains configuration pragmas directly,
3128 as described in the following section.
3130 In the case of @code{Restrictions} pragmas appearing as configuration
3131 pragmas in individual compilation units, the exact handling depends on
3132 the type of restriction.
3134 Restrictions that require partition-wide consistency (like
3135 @code{No_Tasking}) are
3136 recognized wherever they appear
3137 and can be freely inherited, e.g. from a @emph{with}ed unit to the @emph{with}ing
3138 unit. This makes sense since the binder will in any case insist on seeing
3139 consistent use, so any unit not conforming to any restrictions that are
3140 anywhere in the partition will be rejected, and you might as well find
3141 that out at compile time rather than at bind time.
3143 For restrictions that do not require partition-wide consistency, e.g.
3144 SPARK or No_Implementation_Attributes, in general the restriction applies
3145 only to the unit in which the pragma appears, and not to any other units.
3147 The exception is No_Elaboration_Code which always applies to the entire
3148 object file from a compilation, i.e. to the body, spec, and all subunits.
3149 This restriction can be specified in a configuration pragma file, or it
3150 can be on the body and/or the spec (in eithe case it applies to all the
3151 relevant units). It can appear on a subunit only if it has previously
3152 appeared in the body of spec.
3154 @node The Configuration Pragmas Files,,Handling of Configuration Pragmas,Configuration Pragmas
3155 @anchor{gnat_ugn/the_gnat_compilation_model the-configuration-pragmas-files}@anchor{79}@anchor{gnat_ugn/the_gnat_compilation_model id30}@anchor{7a}
3156 @subsection The Configuration Pragmas Files
3161 In GNAT a compilation environment is defined by the current
3162 directory at the time that a compile command is given. This current
3163 directory is searched for a file whose name is @code{gnat.adc}. If
3164 this file is present, it is expected to contain one or more
3165 configuration pragmas that will be applied to the current compilation.
3166 However, if the switch @code{-gnatA} is used, @code{gnat.adc} is not
3167 considered. When taken into account, @code{gnat.adc} is added to the
3168 dependencies, so that if @code{gnat.adc} is modified later, an invocation of
3169 @code{gnatmake} will recompile the source.
3171 Configuration pragmas may be entered into the @code{gnat.adc} file
3172 either by running @code{gnatchop} on a source file that consists only of
3173 configuration pragmas, or more conveniently by direct editing of the
3174 @code{gnat.adc} file, which is a standard format source file.
3176 Besides @code{gnat.adc}, additional files containing configuration
3177 pragmas may be applied to the current compilation using the switch
3178 @code{-gnatec=@emph{path}} where @code{path} must designate an existing file that
3179 contains only configuration pragmas. These configuration pragmas are
3180 in addition to those found in @code{gnat.adc} (provided @code{gnat.adc}
3181 is present and switch @code{-gnatA} is not used).
3183 It is allowable to specify several switches @code{-gnatec=}, all of which
3184 will be taken into account.
3186 Files containing configuration pragmas specified with switches
3187 @code{-gnatec=} are added to the dependencies, unless they are
3188 temporary files. A file is considered temporary if its name ends in
3189 @code{.tmp} or @code{.TMP}. Certain tools follow this naming
3190 convention because they pass information to @code{gcc} via
3191 temporary files that are immediately deleted; it doesn't make sense to
3192 depend on a file that no longer exists. Such tools include
3193 @code{gprbuild}, @code{gnatmake}, and @code{gnatcheck}.
3195 If you are using project file, a separate mechanism is provided using
3199 @c See :ref:`Specifying_Configuration_Pragmas` for more details.
3201 @node Generating Object Files,Source Dependencies,Configuration Pragmas,The GNAT Compilation Model
3202 @anchor{gnat_ugn/the_gnat_compilation_model generating-object-files}@anchor{40}@anchor{gnat_ugn/the_gnat_compilation_model id31}@anchor{7b}
3203 @section Generating Object Files
3206 An Ada program consists of a set of source files, and the first step in
3207 compiling the program is to generate the corresponding object files.
3208 These are generated by compiling a subset of these source files.
3209 The files you need to compile are the following:
3215 If a package spec has no body, compile the package spec to produce the
3216 object file for the package.
3219 If a package has both a spec and a body, compile the body to produce the
3220 object file for the package. The source file for the package spec need
3221 not be compiled in this case because there is only one object file, which
3222 contains the code for both the spec and body of the package.
3225 For a subprogram, compile the subprogram body to produce the object file
3226 for the subprogram. The spec, if one is present, is as usual in a
3227 separate file, and need not be compiled.
3236 In the case of subunits, only compile the parent unit. A single object
3237 file is generated for the entire subunit tree, which includes all the
3241 Compile child units independently of their parent units
3242 (though, of course, the spec of all the ancestor unit must be present in order
3243 to compile a child unit).
3248 Compile generic units in the same manner as any other units. The object
3249 files in this case are small dummy files that contain at most the
3250 flag used for elaboration checking. This is because GNAT always handles generic
3251 instantiation by means of macro expansion. However, it is still necessary to
3252 compile generic units, for dependency checking and elaboration purposes.
3255 The preceding rules describe the set of files that must be compiled to
3256 generate the object files for a program. Each object file has the same
3257 name as the corresponding source file, except that the extension is
3260 You may wish to compile other files for the purpose of checking their
3261 syntactic and semantic correctness. For example, in the case where a
3262 package has a separate spec and body, you would not normally compile the
3263 spec. However, it is convenient in practice to compile the spec to make
3264 sure it is error-free before compiling clients of this spec, because such
3265 compilations will fail if there is an error in the spec.
3267 GNAT provides an option for compiling such files purely for the
3268 purposes of checking correctness; such compilations are not required as
3269 part of the process of building a program. To compile a file in this
3270 checking mode, use the @code{-gnatc} switch.
3272 @node Source Dependencies,The Ada Library Information Files,Generating Object Files,The GNAT Compilation Model
3273 @anchor{gnat_ugn/the_gnat_compilation_model id32}@anchor{7c}@anchor{gnat_ugn/the_gnat_compilation_model source-dependencies}@anchor{41}
3274 @section Source Dependencies
3277 A given object file clearly depends on the source file which is compiled
3278 to produce it. Here we are using "depends" in the sense of a typical
3279 @code{make} utility; in other words, an object file depends on a source
3280 file if changes to the source file require the object file to be
3282 In addition to this basic dependency, a given object may depend on
3283 additional source files as follows:
3289 If a file being compiled @emph{with}s a unit @code{X}, the object file
3290 depends on the file containing the spec of unit @code{X}. This includes
3291 files that are @emph{with}ed implicitly either because they are parents
3292 of @emph{with}ed child units or they are run-time units required by the
3293 language constructs used in a particular unit.
3296 If a file being compiled instantiates a library level generic unit, the
3297 object file depends on both the spec and body files for this generic
3301 If a file being compiled instantiates a generic unit defined within a
3302 package, the object file depends on the body file for the package as
3303 well as the spec file.
3308 @geindex -gnatn switch
3314 If a file being compiled contains a call to a subprogram for which
3315 pragma @code{Inline} applies and inlining is activated with the
3316 @code{-gnatn} switch, the object file depends on the file containing the
3317 body of this subprogram as well as on the file containing the spec. Note
3318 that for inlining to actually occur as a result of the use of this switch,
3319 it is necessary to compile in optimizing mode.
3321 @geindex -gnatN switch
3323 The use of @code{-gnatN} activates inlining optimization
3324 that is performed by the front end of the compiler. This inlining does
3325 not require that the code generation be optimized. Like @code{-gnatn},
3326 the use of this switch generates additional dependencies.
3328 When using a gcc-based back end (in practice this means using any version
3329 of GNAT other than for the JVM, .NET or GNAAMP platforms), then the use of
3330 @code{-gnatN} is deprecated, and the use of @code{-gnatn} is preferred.
3331 Historically front end inlining was more extensive than the gcc back end
3332 inlining, but that is no longer the case.
3335 If an object file @code{O} depends on the proper body of a subunit through
3336 inlining or instantiation, it depends on the parent unit of the subunit.
3337 This means that any modification of the parent unit or one of its subunits
3338 affects the compilation of @code{O}.
3341 The object file for a parent unit depends on all its subunit body files.
3344 The previous two rules meant that for purposes of computing dependencies and
3345 recompilation, a body and all its subunits are treated as an indivisible whole.
3347 These rules are applied transitively: if unit @code{A} @emph{with}s
3348 unit @code{B}, whose elaboration calls an inlined procedure in package
3349 @code{C}, the object file for unit @code{A} will depend on the body of
3350 @code{C}, in file @code{c.adb}.
3352 The set of dependent files described by these rules includes all the
3353 files on which the unit is semantically dependent, as dictated by the
3354 Ada language standard. However, it is a superset of what the
3355 standard describes, because it includes generic, inline, and subunit
3358 An object file must be recreated by recompiling the corresponding source
3359 file if any of the source files on which it depends are modified. For
3360 example, if the @code{make} utility is used to control compilation,
3361 the rule for an Ada object file must mention all the source files on
3362 which the object file depends, according to the above definition.
3363 The determination of the necessary
3364 recompilations is done automatically when one uses @code{gnatmake}.
3367 @node The Ada Library Information Files,Binding an Ada Program,Source Dependencies,The GNAT Compilation Model
3368 @anchor{gnat_ugn/the_gnat_compilation_model id33}@anchor{7d}@anchor{gnat_ugn/the_gnat_compilation_model the-ada-library-information-files}@anchor{42}
3369 @section The Ada Library Information Files
3372 @geindex Ada Library Information files
3376 Each compilation actually generates two output files. The first of these
3377 is the normal object file that has a @code{.o} extension. The second is a
3378 text file containing full dependency information. It has the same
3379 name as the source file, but an @code{.ali} extension.
3380 This file is known as the Ada Library Information (@code{ALI}) file.
3381 The following information is contained in the @code{ALI} file.
3387 Version information (indicates which version of GNAT was used to compile
3388 the unit(s) in question)
3391 Main program information (including priority and time slice settings,
3392 as well as the wide character encoding used during compilation).
3395 List of arguments used in the @code{gcc} command for the compilation
3398 Attributes of the unit, including configuration pragmas used, an indication
3399 of whether the compilation was successful, exception model used etc.
3402 A list of relevant restrictions applying to the unit (used for consistency)
3406 Categorization information (e.g., use of pragma @code{Pure}).
3409 Information on all @emph{with}ed units, including presence of
3410 @code{Elaborate} or @code{Elaborate_All} pragmas.
3413 Information from any @code{Linker_Options} pragmas used in the unit
3416 Information on the use of @code{Body_Version} or @code{Version}
3417 attributes in the unit.
3420 Dependency information. This is a list of files, together with
3421 time stamp and checksum information. These are files on which
3422 the unit depends in the sense that recompilation is required
3423 if any of these units are modified.
3426 Cross-reference data. Contains information on all entities referenced
3427 in the unit. Used by tools like @code{gnatxref} and @code{gnatfind} to
3428 provide cross-reference information.
3431 For a full detailed description of the format of the @code{ALI} file,
3432 see the source of the body of unit @code{Lib.Writ}, contained in file
3433 @code{lib-writ.adb} in the GNAT compiler sources.
3435 @node Binding an Ada Program,GNAT and Libraries,The Ada Library Information Files,The GNAT Compilation Model
3436 @anchor{gnat_ugn/the_gnat_compilation_model id34}@anchor{7e}@anchor{gnat_ugn/the_gnat_compilation_model binding-an-ada-program}@anchor{43}
3437 @section Binding an Ada Program
3440 When using languages such as C and C++, once the source files have been
3441 compiled the only remaining step in building an executable program
3442 is linking the object modules together. This means that it is possible to
3443 link an inconsistent version of a program, in which two units have
3444 included different versions of the same header.
3446 The rules of Ada do not permit such an inconsistent program to be built.
3447 For example, if two clients have different versions of the same package,
3448 it is illegal to build a program containing these two clients.
3449 These rules are enforced by the GNAT binder, which also determines an
3450 elaboration order consistent with the Ada rules.
3452 The GNAT binder is run after all the object files for a program have
3453 been created. It is given the name of the main program unit, and from
3454 this it determines the set of units required by the program, by reading the
3455 corresponding ALI files. It generates error messages if the program is
3456 inconsistent or if no valid order of elaboration exists.
3458 If no errors are detected, the binder produces a main program, in Ada by
3459 default, that contains calls to the elaboration procedures of those
3460 compilation unit that require them, followed by
3461 a call to the main program. This Ada program is compiled to generate the
3462 object file for the main program. The name of
3463 the Ada file is @code{b~xxx}.adb` (with the corresponding spec
3464 @code{b~xxx}.ads`) where @code{xxx} is the name of the
3467 Finally, the linker is used to build the resulting executable program,
3468 using the object from the main program from the bind step as well as the
3469 object files for the Ada units of the program.
3471 @node GNAT and Libraries,Conditional Compilation,Binding an Ada Program,The GNAT Compilation Model
3472 @anchor{gnat_ugn/the_gnat_compilation_model gnat-and-libraries}@anchor{15}@anchor{gnat_ugn/the_gnat_compilation_model id35}@anchor{7f}
3473 @section GNAT and Libraries
3476 @geindex Library building and using
3478 This section describes how to build and use libraries with GNAT, and also shows
3479 how to recompile the GNAT run-time library. You should be familiar with the
3480 Project Manager facility (see the @emph{GNAT_Project_Manager} chapter of the
3481 @emph{GPRbuild User's Guide}) before reading this chapter.
3484 * Introduction to Libraries in GNAT::
3485 * General Ada Libraries::
3486 * Stand-alone Ada Libraries::
3487 * Rebuilding the GNAT Run-Time Library::
3491 @node Introduction to Libraries in GNAT,General Ada Libraries,,GNAT and Libraries
3492 @anchor{gnat_ugn/the_gnat_compilation_model introduction-to-libraries-in-gnat}@anchor{80}@anchor{gnat_ugn/the_gnat_compilation_model id36}@anchor{81}
3493 @subsection Introduction to Libraries in GNAT
3496 A library is, conceptually, a collection of objects which does not have its
3497 own main thread of execution, but rather provides certain services to the
3498 applications that use it. A library can be either statically linked with the
3499 application, in which case its code is directly included in the application,
3500 or, on platforms that support it, be dynamically linked, in which case
3501 its code is shared by all applications making use of this library.
3503 GNAT supports both types of libraries.
3504 In the static case, the compiled code can be provided in different ways. The
3505 simplest approach is to provide directly the set of objects resulting from
3506 compilation of the library source files. Alternatively, you can group the
3507 objects into an archive using whatever commands are provided by the operating
3508 system. For the latter case, the objects are grouped into a shared library.
3510 In the GNAT environment, a library has three types of components:
3519 @code{ALI} files (see @ref{42,,The Ada Library Information Files}), and
3522 Object files, an archive or a shared library.
3525 A GNAT library may expose all its source files, which is useful for
3526 documentation purposes. Alternatively, it may expose only the units needed by
3527 an external user to make use of the library. That is to say, the specs
3528 reflecting the library services along with all the units needed to compile
3529 those specs, which can include generic bodies or any body implementing an
3530 inlined routine. In the case of @emph{stand-alone libraries} those exposed
3531 units are called @emph{interface units} (@ref{82,,Stand-alone Ada Libraries}).
3533 All compilation units comprising an application, including those in a library,
3534 need to be elaborated in an order partially defined by Ada's semantics. GNAT
3535 computes the elaboration order from the @code{ALI} files and this is why they
3536 constitute a mandatory part of GNAT libraries.
3537 @emph{Stand-alone libraries} are the exception to this rule because a specific
3538 library elaboration routine is produced independently of the application(s)
3541 @node General Ada Libraries,Stand-alone Ada Libraries,Introduction to Libraries in GNAT,GNAT and Libraries
3542 @anchor{gnat_ugn/the_gnat_compilation_model general-ada-libraries}@anchor{83}@anchor{gnat_ugn/the_gnat_compilation_model id37}@anchor{84}
3543 @subsection General Ada Libraries
3547 * Building a library::
3548 * Installing a library::
3553 @node Building a library,Installing a library,,General Ada Libraries
3554 @anchor{gnat_ugn/the_gnat_compilation_model building-a-library}@anchor{85}@anchor{gnat_ugn/the_gnat_compilation_model id38}@anchor{86}
3555 @subsubsection Building a library
3558 The easiest way to build a library is to use the Project Manager,
3559 which supports a special type of project called a @emph{Library Project}
3560 (see the @emph{Library Projects} section in the @emph{GNAT Project Manager}
3561 chapter of the @emph{GPRbuild User's Guide}).
3563 A project is considered a library project, when two project-level attributes
3564 are defined in it: @code{Library_Name} and @code{Library_Dir}. In order to
3565 control different aspects of library configuration, additional optional
3566 project-level attributes can be specified:
3575 @item @code{Library_Kind}
3577 This attribute controls whether the library is to be static or dynamic
3584 @item @code{Library_Version}
3586 This attribute specifies the library version; this value is used
3587 during dynamic linking of shared libraries to determine if the currently
3588 installed versions of the binaries are compatible.
3592 @code{Library_Options}
3598 @item @code{Library_GCC}
3600 These attributes specify additional low-level options to be used during
3601 library generation, and redefine the actual application used to generate
3606 The GNAT Project Manager takes full care of the library maintenance task,
3607 including recompilation of the source files for which objects do not exist
3608 or are not up to date, assembly of the library archive, and installation of
3609 the library (i.e., copying associated source, object and @code{ALI} files
3610 to the specified location).
3612 Here is a simple library project file:
3616 for Source_Dirs use ("src1", "src2");
3617 for Object_Dir use "obj";
3618 for Library_Name use "mylib";
3619 for Library_Dir use "lib";
3620 for Library_Kind use "dynamic";
3624 and the compilation command to build and install the library:
3630 It is not entirely trivial to perform manually all the steps required to
3631 produce a library. We recommend that you use the GNAT Project Manager
3632 for this task. In special cases where this is not desired, the necessary
3633 steps are discussed below.
3635 There are various possibilities for compiling the units that make up the
3636 library: for example with a Makefile (@ref{1f,,Using the GNU make Utility}) or
3637 with a conventional script. For simple libraries, it is also possible to create
3638 a dummy main program which depends upon all the packages that comprise the
3639 interface of the library. This dummy main program can then be given to
3640 @code{gnatmake}, which will ensure that all necessary objects are built.
3642 After this task is accomplished, you should follow the standard procedure
3643 of the underlying operating system to produce the static or shared library.
3645 Here is an example of such a dummy program:
3648 with My_Lib.Service1;
3649 with My_Lib.Service2;
3650 with My_Lib.Service3;
3651 procedure My_Lib_Dummy is
3657 Here are the generic commands that will build an archive or a shared library.
3660 # compiling the library
3661 $ gnatmake -c my_lib_dummy.adb
3663 # we don't need the dummy object itself
3664 $ rm my_lib_dummy.o my_lib_dummy.ali
3666 # create an archive with the remaining objects
3667 $ ar rc libmy_lib.a *.o
3668 # some systems may require "ranlib" to be run as well
3670 # or create a shared library
3671 $ gcc -shared -o libmy_lib.so *.o
3672 # some systems may require the code to have been compiled with -fPIC
3674 # remove the object files that are now in the library
3677 # Make the ALI files read-only so that gnatmake will not try to
3678 # regenerate the objects that are in the library
3682 Please note that the library must have a name of the form @code{lib@emph{xxx}.a}
3683 or @code{lib@emph{xxx}.so} (or @code{lib@emph{xxx}.dll} on Windows) in order to
3684 be accessed by the directive @code{-l@emph{xxx}} at link time.
3686 @node Installing a library,Using a library,Building a library,General Ada Libraries
3687 @anchor{gnat_ugn/the_gnat_compilation_model installing-a-library}@anchor{87}@anchor{gnat_ugn/the_gnat_compilation_model id39}@anchor{88}
3688 @subsubsection Installing a library
3691 @geindex ADA_PROJECT_PATH
3693 @geindex GPR_PROJECT_PATH
3695 If you use project files, library installation is part of the library build
3696 process (see the @emph{Installing a Library with Project Files} section of the
3697 @emph{GNAT Project Manager} chapter of the @emph{GPRbuild User's Guide}).
3699 When project files are not an option, it is also possible, but not recommended,
3700 to install the library so that the sources needed to use the library are on the
3701 Ada source path and the ALI files & libraries be on the Ada Object path (see
3702 @ref{89,,Search Paths and the Run-Time Library (RTL)}. Alternatively, the system
3703 administrator can place general-purpose libraries in the default compiler
3704 paths, by specifying the libraries' location in the configuration files
3705 @code{ada_source_path} and @code{ada_object_path}. These configuration files
3706 must be located in the GNAT installation tree at the same place as the gcc spec
3707 file. The location of the gcc spec file can be determined as follows:
3713 The configuration files mentioned above have a simple format: each line
3714 must contain one unique directory name.
3715 Those names are added to the corresponding path
3716 in their order of appearance in the file. The names can be either absolute
3717 or relative; in the latter case, they are relative to where theses files
3720 The files @code{ada_source_path} and @code{ada_object_path} might not be
3722 GNAT installation, in which case, GNAT will look for its run-time library in
3723 the directories @code{adainclude} (for the sources) and @code{adalib} (for the
3724 objects and @code{ALI} files). When the files exist, the compiler does not
3725 look in @code{adainclude} and @code{adalib}, and thus the
3726 @code{ada_source_path} file
3727 must contain the location for the GNAT run-time sources (which can simply
3728 be @code{adainclude}). In the same way, the @code{ada_object_path} file must
3729 contain the location for the GNAT run-time objects (which can simply
3732 You can also specify a new default path to the run-time library at compilation
3733 time with the switch @code{--RTS=rts-path}. You can thus choose / change
3734 the run-time library you want your program to be compiled with. This switch is
3735 recognized by @code{gcc}, @code{gnatmake}, @code{gnatbind},
3736 @code{gnatls}, @code{gnatfind} and @code{gnatxref}.
3738 It is possible to install a library before or after the standard GNAT
3739 library, by reordering the lines in the configuration files. In general, a
3740 library must be installed before the GNAT library if it redefines
3743 @node Using a library,,Installing a library,General Ada Libraries
3744 @anchor{gnat_ugn/the_gnat_compilation_model using-a-library}@anchor{8a}@anchor{gnat_ugn/the_gnat_compilation_model id40}@anchor{8b}
3745 @subsubsection Using a library
3748 Once again, the project facility greatly simplifies the use of
3749 libraries. In this context, using a library is just a matter of adding a
3750 @emph{with} clause in the user project. For instance, to make use of the
3751 library @code{My_Lib} shown in examples in earlier sections, you can
3761 Even if you have a third-party, non-Ada library, you can still use GNAT's
3762 Project Manager facility to provide a wrapper for it. For example, the
3763 following project, when @emph{with}ed by your main project, will link with the
3764 third-party library @code{liba.a}:
3768 for Externally_Built use "true";
3769 for Source_Files use ();
3770 for Library_Dir use "lib";
3771 for Library_Name use "a";
3772 for Library_Kind use "static";
3776 This is an alternative to the use of @code{pragma Linker_Options}. It is
3777 especially interesting in the context of systems with several interdependent
3778 static libraries where finding a proper linker order is not easy and best be
3779 left to the tools having visibility over project dependence information.
3781 In order to use an Ada library manually, you need to make sure that this
3782 library is on both your source and object path
3783 (see @ref{89,,Search Paths and the Run-Time Library (RTL)}
3784 and @ref{8c,,Search Paths for gnatbind}). Furthermore, when the objects are grouped
3785 in an archive or a shared library, you need to specify the desired
3786 library at link time.
3788 For example, you can use the library @code{mylib} installed in
3789 @code{/dir/my_lib_src} and @code{/dir/my_lib_obj} with the following commands:
3792 $ gnatmake -aI/dir/my_lib_src -aO/dir/my_lib_obj my_appl \\
3796 This can be expressed more simply:
3802 when the following conditions are met:
3808 @code{/dir/my_lib_src} has been added by the user to the environment
3810 @geindex ADA_INCLUDE_PATH
3811 @geindex environment variable; ADA_INCLUDE_PATH
3812 @code{ADA_INCLUDE_PATH}, or by the administrator to the file
3813 @code{ada_source_path}
3816 @code{/dir/my_lib_obj} has been added by the user to the environment
3818 @geindex ADA_OBJECTS_PATH
3819 @geindex environment variable; ADA_OBJECTS_PATH
3820 @code{ADA_OBJECTS_PATH}, or by the administrator to the file
3821 @code{ada_object_path}
3824 a pragma @code{Linker_Options} has been added to one of the sources.
3828 pragma Linker_Options ("-lmy_lib");
3832 Note that you may also load a library dynamically at
3833 run time given its filename, as illustrated in the GNAT @code{plugins} example
3834 in the directory @code{share/examples/gnat/plugins} within the GNAT
3837 @node Stand-alone Ada Libraries,Rebuilding the GNAT Run-Time Library,General Ada Libraries,GNAT and Libraries
3838 @anchor{gnat_ugn/the_gnat_compilation_model stand-alone-ada-libraries}@anchor{82}@anchor{gnat_ugn/the_gnat_compilation_model id41}@anchor{8d}
3839 @subsection Stand-alone Ada Libraries
3842 @geindex Stand-alone libraries
3845 * Introduction to Stand-alone Libraries::
3846 * Building a Stand-alone Library::
3847 * Creating a Stand-alone Library to be used in a non-Ada context::
3848 * Restrictions in Stand-alone Libraries::
3852 @node Introduction to Stand-alone Libraries,Building a Stand-alone Library,,Stand-alone Ada Libraries
3853 @anchor{gnat_ugn/the_gnat_compilation_model introduction-to-stand-alone-libraries}@anchor{8e}@anchor{gnat_ugn/the_gnat_compilation_model id42}@anchor{8f}
3854 @subsubsection Introduction to Stand-alone Libraries
3857 A Stand-alone Library (abbreviated 'SAL') is a library that contains the
3859 elaborate the Ada units that are included in the library. In contrast with
3860 an ordinary library, which consists of all sources, objects and @code{ALI}
3862 library, a SAL may specify a restricted subset of compilation units
3863 to serve as a library interface. In this case, the fully
3864 self-sufficient set of files will normally consist of an objects
3865 archive, the sources of interface units' specs, and the @code{ALI}
3866 files of interface units.
3867 If an interface spec contains a generic unit or an inlined subprogram,
3869 source must also be provided; if the units that must be provided in the source
3870 form depend on other units, the source and @code{ALI} files of those must
3873 The main purpose of a SAL is to minimize the recompilation overhead of client
3874 applications when a new version of the library is installed. Specifically,
3875 if the interface sources have not changed, client applications do not need to
3876 be recompiled. If, furthermore, a SAL is provided in the shared form and its
3877 version, controlled by @code{Library_Version} attribute, is not changed,
3878 then the clients do not need to be relinked.
3880 SALs also allow the library providers to minimize the amount of library source
3881 text exposed to the clients. Such 'information hiding' might be useful or
3882 necessary for various reasons.
3884 Stand-alone libraries are also well suited to be used in an executable whose
3885 main routine is not written in Ada.
3887 @node Building a Stand-alone Library,Creating a Stand-alone Library to be used in a non-Ada context,Introduction to Stand-alone Libraries,Stand-alone Ada Libraries
3888 @anchor{gnat_ugn/the_gnat_compilation_model id43}@anchor{90}@anchor{gnat_ugn/the_gnat_compilation_model building-a-stand-alone-library}@anchor{91}
3889 @subsubsection Building a Stand-alone Library
3892 GNAT's Project facility provides a simple way of building and installing
3893 stand-alone libraries; see the @emph{Stand-alone Library Projects} section
3894 in the @emph{GNAT Project Manager} chapter of the @emph{GPRbuild User's Guide}.
3895 To be a Stand-alone Library Project, in addition to the two attributes
3896 that make a project a Library Project (@code{Library_Name} and
3897 @code{Library_Dir}; see the @emph{Library Projects} section in the
3898 @emph{GNAT Project Manager} chapter of the @emph{GPRbuild User's Guide}),
3899 the attribute @code{Library_Interface} must be defined. For example:
3902 for Library_Dir use "lib_dir";
3903 for Library_Name use "dummy";
3904 for Library_Interface use ("int1", "int1.child");
3907 Attribute @code{Library_Interface} has a non-empty string list value,
3908 each string in the list designating a unit contained in an immediate source
3909 of the project file.
3911 When a Stand-alone Library is built, first the binder is invoked to build
3912 a package whose name depends on the library name
3913 (@code{b~dummy.ads/b} in the example above).
3914 This binder-generated package includes initialization and
3915 finalization procedures whose
3916 names depend on the library name (@code{dummyinit} and @code{dummyfinal}
3918 above). The object corresponding to this package is included in the library.
3920 You must ensure timely (e.g., prior to any use of interfaces in the SAL)
3921 calling of these procedures if a static SAL is built, or if a shared SAL
3923 with the project-level attribute @code{Library_Auto_Init} set to
3926 For a Stand-Alone Library, only the @code{ALI} files of the Interface Units
3927 (those that are listed in attribute @code{Library_Interface}) are copied to
3928 the Library Directory. As a consequence, only the Interface Units may be
3929 imported from Ada units outside of the library. If other units are imported,
3930 the binding phase will fail.
3932 It is also possible to build an encapsulated library where not only
3933 the code to elaborate and finalize the library is embedded but also
3934 ensuring that the library is linked only against static
3935 libraries. So an encapsulated library only depends on system
3936 libraries, all other code, including the GNAT runtime, is embedded. To
3937 build an encapsulated library the attribute
3938 @code{Library_Standalone} must be set to @code{encapsulated}:
3941 for Library_Dir use "lib_dir";
3942 for Library_Name use "dummy";
3943 for Library_Kind use "dynamic";
3944 for Library_Interface use ("int1", "int1.child");
3945 for Library_Standalone use "encapsulated";
3948 The default value for this attribute is @code{standard} in which case
3949 a stand-alone library is built.
3951 The attribute @code{Library_Src_Dir} may be specified for a
3952 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
3953 single string value. Its value must be the path (absolute or relative to the
3954 project directory) of an existing directory. This directory cannot be the
3955 object directory or one of the source directories, but it can be the same as
3956 the library directory. The sources of the Interface
3957 Units of the library that are needed by an Ada client of the library will be
3958 copied to the designated directory, called the Interface Copy directory.
3959 These sources include the specs of the Interface Units, but they may also
3960 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
3961 are used, or when there is a generic unit in the spec. Before the sources
3962 are copied to the Interface Copy directory, an attempt is made to delete all
3963 files in the Interface Copy directory.
3965 Building stand-alone libraries by hand is somewhat tedious, but for those
3966 occasions when it is necessary here are the steps that you need to perform:
3972 Compile all library sources.
3975 Invoke the binder with the switch @code{-n} (No Ada main program),
3976 with all the @code{ALI} files of the interfaces, and
3977 with the switch @code{-L} to give specific names to the @code{init}
3978 and @code{final} procedures. For example:
3981 $ gnatbind -n int1.ali int2.ali -Lsal1
3985 Compile the binder generated file:
3992 Link the dynamic library with all the necessary object files,
3993 indicating to the linker the names of the @code{init} (and possibly
3994 @code{final}) procedures for automatic initialization (and finalization).
3995 The built library should be placed in a directory different from
3996 the object directory.
3999 Copy the @code{ALI} files of the interface to the library directory,
4000 add in this copy an indication that it is an interface to a SAL
4001 (i.e., add a word @code{SL} on the line in the @code{ALI} file that starts
4002 with letter 'P') and make the modified copy of the @code{ALI} file
4006 Using SALs is not different from using other libraries
4007 (see @ref{8a,,Using a library}).
4009 @node Creating a Stand-alone Library to be used in a non-Ada context,Restrictions in Stand-alone Libraries,Building a Stand-alone Library,Stand-alone Ada Libraries
4010 @anchor{gnat_ugn/the_gnat_compilation_model creating-a-stand-alone-library-to-be-used-in-a-non-ada-context}@anchor{92}@anchor{gnat_ugn/the_gnat_compilation_model id44}@anchor{93}
4011 @subsubsection Creating a Stand-alone Library to be used in a non-Ada context
4014 It is easy to adapt the SAL build procedure discussed above for use of a SAL in
4017 The only extra step required is to ensure that library interface subprograms
4018 are compatible with the main program, by means of @code{pragma Export}
4019 or @code{pragma Convention}.
4021 Here is an example of simple library interface for use with C main program:
4024 package My_Package is
4026 procedure Do_Something;
4027 pragma Export (C, Do_Something, "do_something");
4029 procedure Do_Something_Else;
4030 pragma Export (C, Do_Something_Else, "do_something_else");
4035 On the foreign language side, you must provide a 'foreign' view of the
4036 library interface; remember that it should contain elaboration routines in
4037 addition to interface subprograms.
4039 The example below shows the content of @code{mylib_interface.h} (note
4040 that there is no rule for the naming of this file, any name can be used)
4043 /* the library elaboration procedure */
4044 extern void mylibinit (void);
4046 /* the library finalization procedure */
4047 extern void mylibfinal (void);
4049 /* the interface exported by the library */
4050 extern void do_something (void);
4051 extern void do_something_else (void);
4054 Libraries built as explained above can be used from any program, provided
4055 that the elaboration procedures (named @code{mylibinit} in the previous
4056 example) are called before the library services are used. Any number of
4057 libraries can be used simultaneously, as long as the elaboration
4058 procedure of each library is called.
4060 Below is an example of a C program that uses the @code{mylib} library.
4063 #include "mylib_interface.h"
4068 /* First, elaborate the library before using it */
4071 /* Main program, using the library exported entities */
4073 do_something_else ();
4075 /* Library finalization at the end of the program */
4081 Note that invoking any library finalization procedure generated by
4082 @code{gnatbind} shuts down the Ada run-time environment.
4084 finalization of all Ada libraries must be performed at the end of the program.
4085 No call to these libraries or to the Ada run-time library should be made
4086 after the finalization phase.
4088 Note also that special care must be taken with multi-tasks
4089 applications. The initialization and finalization routines are not
4090 protected against concurrent access. If such requirement is needed it
4091 must be ensured at the application level using a specific operating
4092 system services like a mutex or a critical-section.
4094 @node Restrictions in Stand-alone Libraries,,Creating a Stand-alone Library to be used in a non-Ada context,Stand-alone Ada Libraries
4095 @anchor{gnat_ugn/the_gnat_compilation_model id45}@anchor{94}@anchor{gnat_ugn/the_gnat_compilation_model restrictions-in-stand-alone-libraries}@anchor{95}
4096 @subsubsection Restrictions in Stand-alone Libraries
4099 The pragmas listed below should be used with caution inside libraries,
4100 as they can create incompatibilities with other Ada libraries:
4106 pragma @code{Locking_Policy}
4109 pragma @code{Partition_Elaboration_Policy}
4112 pragma @code{Queuing_Policy}
4115 pragma @code{Task_Dispatching_Policy}
4118 pragma @code{Unreserve_All_Interrupts}
4121 When using a library that contains such pragmas, the user must make sure
4122 that all libraries use the same pragmas with the same values. Otherwise,
4123 @code{Program_Error} will
4124 be raised during the elaboration of the conflicting
4125 libraries. The usage of these pragmas and its consequences for the user
4126 should therefore be well documented.
4128 Similarly, the traceback in the exception occurrence mechanism should be
4129 enabled or disabled in a consistent manner across all libraries.
4130 Otherwise, Program_Error will be raised during the elaboration of the
4131 conflicting libraries.
4133 If the @code{Version} or @code{Body_Version}
4134 attributes are used inside a library, then you need to
4135 perform a @code{gnatbind} step that specifies all @code{ALI} files in all
4136 libraries, so that version identifiers can be properly computed.
4137 In practice these attributes are rarely used, so this is unlikely
4138 to be a consideration.
4140 @node Rebuilding the GNAT Run-Time Library,,Stand-alone Ada Libraries,GNAT and Libraries
4141 @anchor{gnat_ugn/the_gnat_compilation_model id46}@anchor{96}@anchor{gnat_ugn/the_gnat_compilation_model rebuilding-the-gnat-run-time-library}@anchor{97}
4142 @subsection Rebuilding the GNAT Run-Time Library
4145 @geindex GNAT Run-Time Library
4148 @geindex Building the GNAT Run-Time Library
4150 @geindex Rebuilding the GNAT Run-Time Library
4152 @geindex Run-Time Library
4155 It may be useful to recompile the GNAT library in various contexts, the
4156 most important one being the use of partition-wide configuration pragmas
4157 such as @code{Normalize_Scalars}. A special Makefile called
4158 @code{Makefile.adalib} is provided to that effect and can be found in
4159 the directory containing the GNAT library. The location of this
4160 directory depends on the way the GNAT environment has been installed and can
4161 be determined by means of the command:
4167 The last entry in the object search path usually contains the
4168 gnat library. This Makefile contains its own documentation and in
4169 particular the set of instructions needed to rebuild a new library and
4172 @geindex Conditional compilation
4174 @node Conditional Compilation,Mixed Language Programming,GNAT and Libraries,The GNAT Compilation Model
4175 @anchor{gnat_ugn/the_gnat_compilation_model id47}@anchor{98}@anchor{gnat_ugn/the_gnat_compilation_model conditional-compilation}@anchor{16}
4176 @section Conditional Compilation
4179 This section presents some guidelines for modeling conditional compilation in Ada and describes the
4180 gnatprep preprocessor utility.
4182 @geindex Conditional compilation
4185 * Modeling Conditional Compilation in Ada::
4186 * Preprocessing with gnatprep::
4187 * Integrated Preprocessing::
4191 @node Modeling Conditional Compilation in Ada,Preprocessing with gnatprep,,Conditional Compilation
4192 @anchor{gnat_ugn/the_gnat_compilation_model modeling-conditional-compilation-in-ada}@anchor{99}@anchor{gnat_ugn/the_gnat_compilation_model id48}@anchor{9a}
4193 @subsection Modeling Conditional Compilation in Ada
4196 It is often necessary to arrange for a single source program
4197 to serve multiple purposes, where it is compiled in different
4198 ways to achieve these different goals. Some examples of the
4199 need for this feature are
4205 Adapting a program to a different hardware environment
4208 Adapting a program to a different target architecture
4211 Turning debugging features on and off
4214 Arranging for a program to compile with different compilers
4217 In C, or C++, the typical approach would be to use the preprocessor
4218 that is defined as part of the language. The Ada language does not
4219 contain such a feature. This is not an oversight, but rather a very
4220 deliberate design decision, based on the experience that overuse of
4221 the preprocessing features in C and C++ can result in programs that
4222 are extremely difficult to maintain. For example, if we have ten
4223 switches that can be on or off, this means that there are a thousand
4224 separate programs, any one of which might not even be syntactically
4225 correct, and even if syntactically correct, the resulting program
4226 might not work correctly. Testing all combinations can quickly become
4229 Nevertheless, the need to tailor programs certainly exists, and in
4230 this section we will discuss how this can
4231 be achieved using Ada in general, and GNAT in particular.
4234 * Use of Boolean Constants::
4235 * Debugging - A Special Case::
4236 * Conditionalizing Declarations::
4237 * Use of Alternative Implementations::
4242 @node Use of Boolean Constants,Debugging - A Special Case,,Modeling Conditional Compilation in Ada
4243 @anchor{gnat_ugn/the_gnat_compilation_model id49}@anchor{9b}@anchor{gnat_ugn/the_gnat_compilation_model use-of-boolean-constants}@anchor{9c}
4244 @subsubsection Use of Boolean Constants
4247 In the case where the difference is simply which code
4248 sequence is executed, the cleanest solution is to use Boolean
4249 constants to control which code is executed.
4252 FP_Initialize_Required : constant Boolean := True;
4254 if FP_Initialize_Required then
4259 Not only will the code inside the @code{if} statement not be executed if
4260 the constant Boolean is @code{False}, but it will also be completely
4261 deleted from the program.
4262 However, the code is only deleted after the @code{if} statement
4263 has been checked for syntactic and semantic correctness.
4264 (In contrast, with preprocessors the code is deleted before the
4265 compiler ever gets to see it, so it is not checked until the switch
4268 @geindex Preprocessors (contrasted with conditional compilation)
4270 Typically the Boolean constants will be in a separate package,
4275 FP_Initialize_Required : constant Boolean := True;
4276 Reset_Available : constant Boolean := False;
4281 The @code{Config} package exists in multiple forms for the various targets,
4282 with an appropriate script selecting the version of @code{Config} needed.
4283 Then any other unit requiring conditional compilation can do a @emph{with}
4284 of @code{Config} to make the constants visible.
4286 @node Debugging - A Special Case,Conditionalizing Declarations,Use of Boolean Constants,Modeling Conditional Compilation in Ada
4287 @anchor{gnat_ugn/the_gnat_compilation_model debugging-a-special-case}@anchor{9d}@anchor{gnat_ugn/the_gnat_compilation_model id50}@anchor{9e}
4288 @subsubsection Debugging - A Special Case
4291 A common use of conditional code is to execute statements (for example
4292 dynamic checks, or output of intermediate results) under control of a
4293 debug switch, so that the debugging behavior can be turned on and off.
4294 This can be done using a Boolean constant to control whether the code
4299 Put_Line ("got to the first stage!");
4306 if Debugging and then Temperature > 999.0 then
4307 raise Temperature_Crazy;
4311 @geindex pragma Assert
4313 Since this is a common case, there are special features to deal with
4314 this in a convenient manner. For the case of tests, Ada 2005 has added
4315 a pragma @code{Assert} that can be used for such tests. This pragma is modeled
4316 on the @code{Assert} pragma that has always been available in GNAT, so this
4317 feature may be used with GNAT even if you are not using Ada 2005 features.
4318 The use of pragma @code{Assert} is described in the
4319 @cite{GNAT_Reference_Manual}, but as an
4320 example, the last test could be written:
4323 pragma Assert (Temperature <= 999.0, "Temperature Crazy");
4329 pragma Assert (Temperature <= 999.0);
4332 In both cases, if assertions are active and the temperature is excessive,
4333 the exception @code{Assert_Failure} will be raised, with the given string in
4334 the first case or a string indicating the location of the pragma in the second
4335 case used as the exception message.
4337 @geindex pragma Assertion_Policy
4339 You can turn assertions on and off by using the @code{Assertion_Policy}
4342 @geindex -gnata switch
4344 This is an Ada 2005 pragma which is implemented in all modes by
4345 GNAT. Alternatively, you can use the @code{-gnata} switch
4346 to enable assertions from the command line, which applies to
4347 all versions of Ada.
4349 @geindex pragma Debug
4351 For the example above with the @code{Put_Line}, the GNAT-specific pragma
4352 @code{Debug} can be used:
4355 pragma Debug (Put_Line ("got to the first stage!"));
4358 If debug pragmas are enabled, the argument, which must be of the form of
4359 a procedure call, is executed (in this case, @code{Put_Line} will be called).
4360 Only one call can be present, but of course a special debugging procedure
4361 containing any code you like can be included in the program and then
4362 called in a pragma @code{Debug} argument as needed.
4364 One advantage of pragma @code{Debug} over the @code{if Debugging then}
4365 construct is that pragma @code{Debug} can appear in declarative contexts,
4366 such as at the very beginning of a procedure, before local declarations have
4369 @geindex pragma Debug_Policy
4371 Debug pragmas are enabled using either the @code{-gnata} switch that also
4372 controls assertions, or with a separate Debug_Policy pragma.
4374 The latter pragma is new in the Ada 2005 versions of GNAT (but it can be used
4375 in Ada 95 and Ada 83 programs as well), and is analogous to
4376 pragma @code{Assertion_Policy} to control assertions.
4378 @code{Assertion_Policy} and @code{Debug_Policy} are configuration pragmas,
4379 and thus they can appear in @code{gnat.adc} if you are not using a
4380 project file, or in the file designated to contain configuration pragmas
4382 They then apply to all subsequent compilations. In practice the use of
4383 the @code{-gnata} switch is often the most convenient method of controlling
4384 the status of these pragmas.
4386 Note that a pragma is not a statement, so in contexts where a statement
4387 sequence is required, you can't just write a pragma on its own. You have
4388 to add a @code{null} statement.
4392 ... -- some statements
4394 pragma Assert (Num_Cases < 10);
4399 @node Conditionalizing Declarations,Use of Alternative Implementations,Debugging - A Special Case,Modeling Conditional Compilation in Ada
4400 @anchor{gnat_ugn/the_gnat_compilation_model conditionalizing-declarations}@anchor{9f}@anchor{gnat_ugn/the_gnat_compilation_model id51}@anchor{a0}
4401 @subsubsection Conditionalizing Declarations
4404 In some cases it may be necessary to conditionalize declarations to meet
4405 different requirements. For example we might want a bit string whose length
4406 is set to meet some hardware message requirement.
4408 This may be possible using declare blocks controlled
4409 by conditional constants:
4412 if Small_Machine then
4414 X : Bit_String (1 .. 10);
4420 X : Large_Bit_String (1 .. 1000);
4427 Note that in this approach, both declarations are analyzed by the
4428 compiler so this can only be used where both declarations are legal,
4429 even though one of them will not be used.
4431 Another approach is to define integer constants, e.g., @code{Bits_Per_Word},
4432 or Boolean constants, e.g., @code{Little_Endian}, and then write declarations
4433 that are parameterized by these constants. For example
4437 Field1 at 0 range Boolean'Pos (Little_Endian) * 10 .. Bits_Per_Word;
4441 If @code{Bits_Per_Word} is set to 32, this generates either
4445 Field1 at 0 range 0 .. 32;
4449 for the big endian case, or
4453 Field1 at 0 range 10 .. 32;
4457 for the little endian case. Since a powerful subset of Ada expression
4458 notation is usable for creating static constants, clever use of this
4459 feature can often solve quite difficult problems in conditionalizing
4460 compilation (note incidentally that in Ada 95, the little endian
4461 constant was introduced as @code{System.Default_Bit_Order}, so you do not
4462 need to define this one yourself).
4464 @node Use of Alternative Implementations,Preprocessing,Conditionalizing Declarations,Modeling Conditional Compilation in Ada
4465 @anchor{gnat_ugn/the_gnat_compilation_model use-of-alternative-implementations}@anchor{a1}@anchor{gnat_ugn/the_gnat_compilation_model id52}@anchor{a2}
4466 @subsubsection Use of Alternative Implementations
4469 In some cases, none of the approaches described above are adequate. This
4470 can occur for example if the set of declarations required is radically
4471 different for two different configurations.
4473 In this situation, the official Ada way of dealing with conditionalizing
4474 such code is to write separate units for the different cases. As long as
4475 this does not result in excessive duplication of code, this can be done
4476 without creating maintenance problems. The approach is to share common
4477 code as far as possible, and then isolate the code and declarations
4478 that are different. Subunits are often a convenient method for breaking
4479 out a piece of a unit that is to be conditionalized, with separate files
4480 for different versions of the subunit for different targets, where the
4481 build script selects the right one to give to the compiler.
4483 @geindex Subunits (and conditional compilation)
4485 As an example, consider a situation where a new feature in Ada 2005
4486 allows something to be done in a really nice way. But your code must be able
4487 to compile with an Ada 95 compiler. Conceptually you want to say:
4491 ... neat Ada 2005 code
4493 ... not quite as neat Ada 95 code
4497 where @code{Ada_2005} is a Boolean constant.
4499 But this won't work when @code{Ada_2005} is set to @code{False},
4500 since the @code{then} clause will be illegal for an Ada 95 compiler.
4501 (Recall that although such unreachable code would eventually be deleted
4502 by the compiler, it still needs to be legal. If it uses features
4503 introduced in Ada 2005, it will be illegal in Ada 95.)
4508 procedure Insert is separate;
4511 Then we have two files for the subunit @code{Insert}, with the two sets of
4513 If the package containing this is called @code{File_Queries}, then we might
4520 @code{file_queries-insert-2005.adb}
4523 @code{file_queries-insert-95.adb}
4526 and the build script renames the appropriate file to @code{file_queries-insert.adb} and then carries out the compilation.
4528 This can also be done with project files' naming schemes. For example:
4531 for body ("File_Queries.Insert") use "file_queries-insert-2005.ada";
4534 Note also that with project files it is desirable to use a different extension
4535 than @code{ads} / @code{adb} for alternative versions. Otherwise a naming
4536 conflict may arise through another commonly used feature: to declare as part
4537 of the project a set of directories containing all the sources obeying the
4538 default naming scheme.
4540 The use of alternative units is certainly feasible in all situations,
4541 and for example the Ada part of the GNAT run-time is conditionalized
4542 based on the target architecture using this approach. As a specific example,
4543 consider the implementation of the AST feature in VMS. There is one
4544 spec: @code{s-asthan.ads} which is the same for all architectures, and three
4554 @item @code{s-asthan.adb}
4556 used for all non-VMS operating systems
4563 @item @code{s-asthan-vms-alpha.adb}
4565 used for VMS on the Alpha
4572 @item @code{s-asthan-vms-ia64.adb}
4574 used for VMS on the ia64
4578 The dummy version @code{s-asthan.adb} simply raises exceptions noting that
4579 this operating system feature is not available, and the two remaining
4580 versions interface with the corresponding versions of VMS to provide
4581 VMS-compatible AST handling. The GNAT build script knows the architecture
4582 and operating system, and automatically selects the right version,
4583 renaming it if necessary to @code{s-asthan.adb} before the run-time build.
4585 Another style for arranging alternative implementations is through Ada's
4586 access-to-subprogram facility.
4587 In case some functionality is to be conditionally included,
4588 you can declare an access-to-procedure variable @code{Ref} that is initialized
4589 to designate a 'do nothing' procedure, and then invoke @code{Ref.all}
4591 In some library package, set @code{Ref} to @code{Proc'Access} for some
4592 procedure @code{Proc} that performs the relevant processing.
4593 The initialization only occurs if the library package is included in the
4595 The same idea can also be implemented using tagged types and dispatching
4598 @node Preprocessing,,Use of Alternative Implementations,Modeling Conditional Compilation in Ada
4599 @anchor{gnat_ugn/the_gnat_compilation_model preprocessing}@anchor{a3}@anchor{gnat_ugn/the_gnat_compilation_model id53}@anchor{a4}
4600 @subsubsection Preprocessing
4603 @geindex Preprocessing
4605 Although it is quite possible to conditionalize code without the use of
4606 C-style preprocessing, as described earlier in this section, it is
4607 nevertheless convenient in some cases to use the C approach. Moreover,
4608 older Ada compilers have often provided some preprocessing capability,
4609 so legacy code may depend on this approach, even though it is not
4612 To accommodate such use, GNAT provides a preprocessor (modeled to a large
4613 extent on the various preprocessors that have been used
4614 with legacy code on other compilers, to enable easier transition).
4618 The preprocessor may be used in two separate modes. It can be used quite
4619 separately from the compiler, to generate a separate output source file
4620 that is then fed to the compiler as a separate step. This is the
4621 @code{gnatprep} utility, whose use is fully described in
4622 @ref{17,,Preprocessing with gnatprep}.
4624 The preprocessing language allows such constructs as
4627 #if DEBUG or else (PRIORITY > 4) then
4628 sequence of declarations
4630 completely different sequence of declarations
4634 The values of the symbols @code{DEBUG} and @code{PRIORITY} can be
4635 defined either on the command line or in a separate file.
4637 The other way of running the preprocessor is even closer to the C style and
4638 often more convenient. In this approach the preprocessing is integrated into
4639 the compilation process. The compiler is given the preprocessor input which
4640 includes @code{#if} lines etc, and then the compiler carries out the
4641 preprocessing internally and processes the resulting output.
4642 For more details on this approach, see @ref{18,,Integrated Preprocessing}.
4644 @node Preprocessing with gnatprep,Integrated Preprocessing,Modeling Conditional Compilation in Ada,Conditional Compilation
4645 @anchor{gnat_ugn/the_gnat_compilation_model id54}@anchor{a5}@anchor{gnat_ugn/the_gnat_compilation_model preprocessing-with-gnatprep}@anchor{17}
4646 @subsection Preprocessing with @code{gnatprep}
4651 @geindex Preprocessing (gnatprep)
4653 This section discusses how to use GNAT's @code{gnatprep} utility for simple
4655 Although designed for use with GNAT, @code{gnatprep} does not depend on any
4656 special GNAT features.
4657 For further discussion of conditional compilation in general, see
4658 @ref{16,,Conditional Compilation}.
4661 * Preprocessing Symbols::
4663 * Switches for gnatprep::
4664 * Form of Definitions File::
4665 * Form of Input Text for gnatprep::
4669 @node Preprocessing Symbols,Using gnatprep,,Preprocessing with gnatprep
4670 @anchor{gnat_ugn/the_gnat_compilation_model id55}@anchor{a6}@anchor{gnat_ugn/the_gnat_compilation_model preprocessing-symbols}@anchor{a7}
4671 @subsubsection Preprocessing Symbols
4674 Preprocessing symbols are defined in @emph{definition files} and referenced in the
4675 sources to be preprocessed. A preprocessing symbol is an identifier, following
4676 normal Ada (case-insensitive) rules for its syntax, with the restriction that
4677 all characters need to be in the ASCII set (no accented letters).
4679 @node Using gnatprep,Switches for gnatprep,Preprocessing Symbols,Preprocessing with gnatprep
4680 @anchor{gnat_ugn/the_gnat_compilation_model using-gnatprep}@anchor{a8}@anchor{gnat_ugn/the_gnat_compilation_model id56}@anchor{a9}
4681 @subsubsection Using @code{gnatprep}
4684 To call @code{gnatprep} use:
4687 $ gnatprep [ switches ] infile outfile [ deffile ]
4699 @item @emph{switches}
4701 is an optional sequence of switches as described in the next section.
4710 is the full name of the input file, which is an Ada source
4711 file containing preprocessor directives.
4718 @item @emph{outfile}
4720 is the full name of the output file, which is an Ada source
4721 in standard Ada form. When used with GNAT, this file name will
4722 normally have an @code{ads} or @code{adb} suffix.
4729 @item @code{deffile}
4731 is the full name of a text file containing definitions of
4732 preprocessing symbols to be referenced by the preprocessor. This argument is
4733 optional, and can be replaced by the use of the @code{-D} switch.
4737 @node Switches for gnatprep,Form of Definitions File,Using gnatprep,Preprocessing with gnatprep
4738 @anchor{gnat_ugn/the_gnat_compilation_model switches-for-gnatprep}@anchor{aa}@anchor{gnat_ugn/the_gnat_compilation_model id57}@anchor{ab}
4739 @subsubsection Switches for @code{gnatprep}
4742 @geindex --version (gnatprep)
4747 @item @code{--version}
4749 Display Copyright and version, then exit disregarding all other options.
4752 @geindex --help (gnatprep)
4759 If @code{--version} was not used, display usage and then exit disregarding
4763 @geindex -b (gnatprep)
4770 Causes both preprocessor lines and the lines deleted by
4771 preprocessing to be replaced by blank lines in the output source file,
4772 preserving line numbers in the output file.
4775 @geindex -c (gnatprep)
4782 Causes both preprocessor lines and the lines deleted
4783 by preprocessing to be retained in the output source as comments marked
4784 with the special string @code{"--! "}. This option will result in line numbers
4785 being preserved in the output file.
4788 @geindex -C (gnatprep)
4795 Causes comments to be scanned. Normally comments are ignored by gnatprep.
4796 If this option is specified, then comments are scanned and any $symbol
4797 substitutions performed as in program text. This is particularly useful
4798 when structured comments are used (e.g., for programs written in a
4799 pre-2014 version of the SPARK Ada subset). Note that this switch is not
4800 available when doing integrated preprocessing (it would be useless in
4801 this context since comments are ignored by the compiler in any case).
4804 @geindex -D (gnatprep)
4809 @item @code{-D@emph{symbol}[=@emph{value}]}
4811 Defines a new preprocessing symbol with the specified value. If no value is given
4812 on the command line, then symbol is considered to be @code{True}. This switch
4813 can be used in place of a definition file.
4816 @geindex -r (gnatprep)
4823 Causes a @code{Source_Reference} pragma to be generated that
4824 references the original input file, so that error messages will use
4825 the file name of this original file. The use of this switch implies
4826 that preprocessor lines are not to be removed from the file, so its
4827 use will force @code{-b} mode if @code{-c}
4828 has not been specified explicitly.
4830 Note that if the file to be preprocessed contains multiple units, then
4831 it will be necessary to @code{gnatchop} the output file from
4832 @code{gnatprep}. If a @code{Source_Reference} pragma is present
4833 in the preprocessed file, it will be respected by
4835 so that the final chopped files will correctly refer to the original
4836 input source file for @code{gnatprep}.
4839 @geindex -s (gnatprep)
4846 Causes a sorted list of symbol names and values to be
4847 listed on the standard output file.
4850 @geindex -T (gnatprep)
4857 Use LF as line terminators when writing files. By default the line terminator
4858 of the host (LF under unix, CR/LF under Windows) is used.
4861 @geindex -u (gnatprep)
4868 Causes undefined symbols to be treated as having the value FALSE in the context
4869 of a preprocessor test. In the absence of this option, an undefined symbol in
4870 a @code{#if} or @code{#elsif} test will be treated as an error.
4873 @geindex -v (gnatprep)
4880 Verbose mode: generates more output about work done.
4883 Note: if neither @code{-b} nor @code{-c} is present,
4884 then preprocessor lines and
4885 deleted lines are completely removed from the output, unless -r is
4886 specified, in which case -b is assumed.
4888 @node Form of Definitions File,Form of Input Text for gnatprep,Switches for gnatprep,Preprocessing with gnatprep
4889 @anchor{gnat_ugn/the_gnat_compilation_model form-of-definitions-file}@anchor{ac}@anchor{gnat_ugn/the_gnat_compilation_model id58}@anchor{ad}
4890 @subsubsection Form of Definitions File
4893 The definitions file contains lines of the form:
4899 where @code{symbol} is a preprocessing symbol, and @code{value} is one of the following:
4905 Empty, corresponding to a null substitution,
4908 A string literal using normal Ada syntax, or
4911 Any sequence of characters from the set @{letters, digits, period, underline@}.
4914 Comment lines may also appear in the definitions file, starting with
4915 the usual @code{--},
4916 and comments may be added to the definitions lines.
4918 @node Form of Input Text for gnatprep,,Form of Definitions File,Preprocessing with gnatprep
4919 @anchor{gnat_ugn/the_gnat_compilation_model id59}@anchor{ae}@anchor{gnat_ugn/the_gnat_compilation_model form-of-input-text-for-gnatprep}@anchor{af}
4920 @subsubsection Form of Input Text for @code{gnatprep}
4923 The input text may contain preprocessor conditional inclusion lines,
4924 as well as general symbol substitution sequences.
4926 The preprocessor conditional inclusion commands have the form:
4929 #if <expression> [then]
4931 #elsif <expression> [then]
4933 #elsif <expression> [then]
4941 In this example, <expression> is defined by the following grammar:
4944 <expression> ::= <symbol>
4945 <expression> ::= <symbol> = "<value>"
4946 <expression> ::= <symbol> = <symbol>
4947 <expression> ::= <symbol> = <integer>
4948 <expression> ::= <symbol> > <integer>
4949 <expression> ::= <symbol> >= <integer>
4950 <expression> ::= <symbol> < <integer>
4951 <expression> ::= <symbol> <= <integer>
4952 <expression> ::= <symbol> 'Defined
4953 <expression> ::= not <expression>
4954 <expression> ::= <expression> and <expression>
4955 <expression> ::= <expression> or <expression>
4956 <expression> ::= <expression> and then <expression>
4957 <expression> ::= <expression> or else <expression>
4958 <expression> ::= ( <expression> )
4961 Note the following restriction: it is not allowed to have "and" or "or"
4962 following "not" in the same expression without parentheses. For example, this
4969 This can be expressed instead as one of the following forms:
4976 For the first test (<expression> ::= <symbol>) the symbol must have
4977 either the value true or false, that is to say the right-hand of the
4978 symbol definition must be one of the (case-insensitive) literals
4979 @code{True} or @code{False}. If the value is true, then the
4980 corresponding lines are included, and if the value is false, they are
4983 When comparing a symbol to an integer, the integer is any non negative
4984 literal integer as defined in the Ada Reference Manual, such as 3, 16#FF# or
4985 2#11#. The symbol value must also be a non negative integer. Integer values
4986 in the range 0 .. 2**31-1 are supported.
4988 The test (<expression> ::= <symbol>'Defined) is true only if
4989 the symbol has been defined in the definition file or by a @code{-D}
4990 switch on the command line. Otherwise, the test is false.
4992 The equality tests are case insensitive, as are all the preprocessor lines.
4994 If the symbol referenced is not defined in the symbol definitions file,
4995 then the effect depends on whether or not switch @code{-u}
4996 is specified. If so, then the symbol is treated as if it had the value
4997 false and the test fails. If this switch is not specified, then
4998 it is an error to reference an undefined symbol. It is also an error to
4999 reference a symbol that is defined with a value other than @code{True}
5002 The use of the @code{not} operator inverts the sense of this logical test.
5003 The @code{not} operator cannot be combined with the @code{or} or @code{and}
5004 operators, without parentheses. For example, "if not X or Y then" is not
5005 allowed, but "if (not X) or Y then" and "if not (X or Y) then" are.
5007 The @code{then} keyword is optional as shown
5009 The @code{#} must be the first non-blank character on a line, but
5010 otherwise the format is free form. Spaces or tabs may appear between
5011 the @code{#} and the keyword. The keywords and the symbols are case
5012 insensitive as in normal Ada code. Comments may be used on a
5013 preprocessor line, but other than that, no other tokens may appear on a
5014 preprocessor line. Any number of @code{elsif} clauses can be present,
5015 including none at all. The @code{else} is optional, as in Ada.
5017 The @code{#} marking the start of a preprocessor line must be the first
5018 non-blank character on the line, i.e., it must be preceded only by
5019 spaces or horizontal tabs.
5021 Symbol substitution outside of preprocessor lines is obtained by using
5028 anywhere within a source line, except in a comment or within a
5029 string literal. The identifier
5030 following the @code{$} must match one of the symbols defined in the symbol
5031 definition file, and the result is to substitute the value of the
5032 symbol in place of @code{$symbol} in the output file.
5034 Note that although the substitution of strings within a string literal
5035 is not possible, it is possible to have a symbol whose defined value is
5036 a string literal. So instead of setting XYZ to @code{hello} and writing:
5039 Header : String := "$XYZ";
5042 you should set XYZ to @code{"hello"} and write:
5045 Header : String := $XYZ;
5048 and then the substitution will occur as desired.
5050 @node Integrated Preprocessing,,Preprocessing with gnatprep,Conditional Compilation
5051 @anchor{gnat_ugn/the_gnat_compilation_model id60}@anchor{b0}@anchor{gnat_ugn/the_gnat_compilation_model integrated-preprocessing}@anchor{18}
5052 @subsection Integrated Preprocessing
5055 As noted above, a file to be preprocessed consists of Ada source code
5056 in which preprocessing lines have been inserted. However,
5057 instead of using @code{gnatprep} to explicitly preprocess a file as a separate
5058 step before compilation, you can carry out the preprocessing implicitly
5059 as part of compilation. Such @emph{integrated preprocessing}, which is the common
5060 style with C, is performed when either or both of the following switches
5061 are passed to the compiler:
5069 @code{-gnatep}, which specifies the @emph{preprocessor data file}.
5070 This file dictates how the source files will be preprocessed (e.g., which
5071 symbol definition files apply to which sources).
5074 @code{-gnateD}, which defines values for preprocessing symbols.
5078 Integrated preprocessing applies only to Ada source files, it is
5079 not available for configuration pragma files.
5081 With integrated preprocessing, the output from the preprocessor is not,
5082 by default, written to any external file. Instead it is passed
5083 internally to the compiler. To preserve the result of
5084 preprocessing in a file, either run @code{gnatprep}
5085 in standalone mode or else supply the @code{-gnateG} switch
5086 (described below) to the compiler.
5088 When using project files:
5096 the builder switch @code{-x} should be used if any Ada source is
5097 compiled with @code{gnatep=}, so that the compiler finds the
5098 @emph{preprocessor data file}.
5101 the preprocessing data file and the symbol definition files should be
5102 located in the source directories of the project.
5106 Note that the @code{gnatmake} switch @code{-m} will almost
5107 always trigger recompilation for sources that are preprocessed,
5108 because @code{gnatmake} cannot compute the checksum of the source after
5111 The actual preprocessing function is described in detail in
5112 @ref{17,,Preprocessing with gnatprep}. This section explains the switches
5113 that relate to integrated preprocessing.
5115 @geindex -gnatep (gcc)
5120 @item @code{-gnatep=@emph{preprocessor_data_file}}
5122 This switch specifies the file name (without directory
5123 information) of the preprocessor data file. Either place this file
5124 in one of the source directories, or, when using project
5125 files, reference the project file's directory via the
5126 @code{project_name'Project_Dir} project attribute; e.g:
5133 for Switches ("Ada") use
5134 ("-gnatep=" & Prj'Project_Dir & "prep.def");
5140 A preprocessor data file is a text file that contains @emph{preprocessor
5141 control lines}. A preprocessor control line directs the preprocessing of
5142 either a particular source file, or, analogous to @code{others} in Ada,
5143 all sources not specified elsewhere in the preprocessor data file.
5144 A preprocessor control line
5145 can optionally identify a @emph{definition file} that assigns values to
5146 preprocessor symbols, as well as a list of switches that relate to
5148 Empty lines and comments (using Ada syntax) are also permitted, with no
5151 Here's an example of a preprocessor data file:
5156 "toto.adb" "prep.def" -u
5157 -- Preprocess toto.adb, using definition file prep.def
5158 -- Undefined symbols are treated as False
5161 -- Preprocess all other sources without using a definition file
5162 -- Suppressed lined are commented
5163 -- Symbol VERSION has the value V101
5165 "tata.adb" "prep2.def" -s
5166 -- Preprocess tata.adb, using definition file prep2.def
5167 -- List all symbols with their values
5171 A preprocessor control line has the following syntax:
5176 <preprocessor_control_line> ::=
5177 <preprocessor_input> [ <definition_file_name> ] @{ <switch> @}
5179 <preprocessor_input> ::= <source_file_name> | '*'
5181 <definition_file_name> ::= <string_literal>
5183 <source_file_name> := <string_literal>
5185 <switch> := (See below for list)
5189 Thus each preprocessor control line starts with either a literal string or
5196 A literal string is the file name (without directory information) of the source
5197 file that will be input to the preprocessor.
5200 The character '*' is a wild-card indicator; the additional parameters on the line
5201 indicate the preprocessing for all the sources
5202 that are not specified explicitly on other lines (the order of the lines is not
5206 It is an error to have two lines with the same file name or two
5207 lines starting with the character '*'.
5209 After the file name or '*', an optional literal string specifies the name of
5210 the definition file to be used for preprocessing
5211 (@ref{ac,,Form of Definitions File}). The definition files are found by the
5212 compiler in one of the source directories. In some cases, when compiling
5213 a source in a directory other than the current directory, if the definition
5214 file is in the current directory, it may be necessary to add the current
5215 directory as a source directory through the @code{-I} switch; otherwise
5216 the compiler would not find the definition file.
5218 Finally, switches similar to those of @code{gnatprep} may optionally appear:
5225 Causes both preprocessor lines and the lines deleted by
5226 preprocessing to be replaced by blank lines, preserving the line number.
5227 This switch is always implied; however, if specified after @code{-c}
5228 it cancels the effect of @code{-c}.
5232 Causes both preprocessor lines and the lines deleted
5233 by preprocessing to be retained as comments marked
5234 with the special string '@cite{--!}'.
5236 @item @code{-D@emph{symbol}=@emph{new_value}}
5238 Define or redefine @code{symbol} to have @code{new_value} as its value.
5239 The permitted form for @code{symbol} is either an Ada identifier, or any Ada reserved word
5240 aside from @code{if},
5241 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
5242 The permitted form for @code{new_value} is a literal string, an Ada identifier or any Ada reserved
5243 word. A symbol declared with this switch replaces a symbol with the
5244 same name defined in a definition file.
5248 Causes a sorted list of symbol names and values to be
5249 listed on the standard output file.
5253 Causes undefined symbols to be treated as having the value @code{FALSE}
5255 of a preprocessor test. In the absence of this option, an undefined symbol in
5256 a @code{#if} or @code{#elsif} test will be treated as an error.
5260 @geindex -gnateD (gcc)
5265 @item @code{-gnateD@emph{symbol}[=@emph{new_value}]}
5267 Define or redefine @code{symbol} to have @code{new_value} as its value. If no value
5268 is supplied, then the value of @code{symbol} is @code{True}.
5269 The form of @code{symbol} is an identifier, following normal Ada (case-insensitive)
5270 rules for its syntax, and @code{new_value} is either an arbitrary string between double
5271 quotes or any sequence (including an empty sequence) of characters from the
5272 set (letters, digits, period, underline).
5273 Ada reserved words may be used as symbols, with the exceptions of @code{if},
5274 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
5283 -gnateDFoo=\"Foo-Bar\"
5287 A symbol declared with this switch on the command line replaces a
5288 symbol with the same name either in a definition file or specified with a
5289 switch @code{-D} in the preprocessor data file.
5291 This switch is similar to switch @code{-D} of @code{gnatprep}.
5293 @item @code{-gnateG}
5295 When integrated preprocessing is performed on source file @code{filename.extension},
5296 create or overwrite @code{filename.extension.prep} to contain
5297 the result of the preprocessing.
5298 For example if the source file is @code{foo.adb} then
5299 the output file will be @code{foo.adb.prep}.
5302 @node Mixed Language Programming,GNAT and Other Compilation Models,Conditional Compilation,The GNAT Compilation Model
5303 @anchor{gnat_ugn/the_gnat_compilation_model mixed-language-programming}@anchor{44}@anchor{gnat_ugn/the_gnat_compilation_model id61}@anchor{b1}
5304 @section Mixed Language Programming
5307 @geindex Mixed Language Programming
5309 This section describes how to develop a mixed-language program,
5310 with a focus on combining Ada with C or C++.
5313 * Interfacing to C::
5314 * Calling Conventions::
5315 * Building Mixed Ada and C++ Programs::
5316 * Generating Ada Bindings for C and C++ headers::
5317 * Generating C Headers for Ada Specifications::
5321 @node Interfacing to C,Calling Conventions,,Mixed Language Programming
5322 @anchor{gnat_ugn/the_gnat_compilation_model interfacing-to-c}@anchor{b2}@anchor{gnat_ugn/the_gnat_compilation_model id62}@anchor{b3}
5323 @subsection Interfacing to C
5326 Interfacing Ada with a foreign language such as C involves using
5327 compiler directives to import and/or export entity definitions in each
5328 language -- using @code{extern} statements in C, for instance, and the
5329 @code{Import}, @code{Export}, and @code{Convention} pragmas in Ada.
5330 A full treatment of these topics is provided in Appendix B, section 1
5331 of the Ada Reference Manual.
5333 There are two ways to build a program using GNAT that contains some Ada
5334 sources and some foreign language sources, depending on whether or not
5335 the main subprogram is written in Ada. Here is a source example with
5336 the main subprogram in Ada:
5342 void print_num (int num)
5344 printf ("num is %d.\\n", num);
5352 /* num_from_Ada is declared in my_main.adb */
5353 extern int num_from_Ada;
5357 return num_from_Ada;
5363 procedure My_Main is
5365 -- Declare then export an Integer entity called num_from_Ada
5366 My_Num : Integer := 10;
5367 pragma Export (C, My_Num, "num_from_Ada");
5369 -- Declare an Ada function spec for Get_Num, then use
5370 -- C function get_num for the implementation.
5371 function Get_Num return Integer;
5372 pragma Import (C, Get_Num, "get_num");
5374 -- Declare an Ada procedure spec for Print_Num, then use
5375 -- C function print_num for the implementation.
5376 procedure Print_Num (Num : Integer);
5377 pragma Import (C, Print_Num, "print_num");
5380 Print_Num (Get_Num);
5384 To build this example:
5390 First compile the foreign language files to
5391 generate object files:
5399 Then, compile the Ada units to produce a set of object files and ALI
5403 $ gnatmake -c my_main.adb
5407 Run the Ada binder on the Ada main program:
5410 $ gnatbind my_main.ali
5414 Link the Ada main program, the Ada objects and the other language
5418 $ gnatlink my_main.ali file1.o file2.o
5422 The last three steps can be grouped in a single command:
5425 $ gnatmake my_main.adb -largs file1.o file2.o
5428 @geindex Binder output file
5430 If the main program is in a language other than Ada, then you may have
5431 more than one entry point into the Ada subsystem. You must use a special
5432 binder option to generate callable routines that initialize and
5433 finalize the Ada units (@ref{b4,,Binding with Non-Ada Main Programs}).
5434 Calls to the initialization and finalization routines must be inserted
5435 in the main program, or some other appropriate point in the code. The
5436 call to initialize the Ada units must occur before the first Ada
5437 subprogram is called, and the call to finalize the Ada units must occur
5438 after the last Ada subprogram returns. The binder will place the
5439 initialization and finalization subprograms into the
5440 @code{b~xxx.adb} file where they can be accessed by your C
5441 sources. To illustrate, we have the following example:
5445 extern void adainit (void);
5446 extern void adafinal (void);
5447 extern int add (int, int);
5448 extern int sub (int, int);
5450 int main (int argc, char *argv[])
5456 /* Should print "21 + 7 = 28" */
5457 printf ("%d + %d = %d\\n", a, b, add (a, b));
5459 /* Should print "21 - 7 = 14" */
5460 printf ("%d - %d = %d\\n", a, b, sub (a, b));
5469 function Add (A, B : Integer) return Integer;
5470 pragma Export (C, Add, "add");
5476 package body Unit1 is
5477 function Add (A, B : Integer) return Integer is
5487 function Sub (A, B : Integer) return Integer;
5488 pragma Export (C, Sub, "sub");
5494 package body Unit2 is
5495 function Sub (A, B : Integer) return Integer is
5502 The build procedure for this application is similar to the last
5509 First, compile the foreign language files to generate object files:
5516 Next, compile the Ada units to produce a set of object files and ALI
5520 $ gnatmake -c unit1.adb
5521 $ gnatmake -c unit2.adb
5525 Run the Ada binder on every generated ALI file. Make sure to use the
5526 @code{-n} option to specify a foreign main program:
5529 $ gnatbind -n unit1.ali unit2.ali
5533 Link the Ada main program, the Ada objects and the foreign language
5534 objects. You need only list the last ALI file here:
5537 $ gnatlink unit2.ali main.o -o exec_file
5540 This procedure yields a binary executable called @code{exec_file}.
5543 Depending on the circumstances (for example when your non-Ada main object
5544 does not provide symbol @code{main}), you may also need to instruct the
5545 GNAT linker not to include the standard startup objects by passing the
5546 @code{-nostartfiles} switch to @code{gnatlink}.
5548 @node Calling Conventions,Building Mixed Ada and C++ Programs,Interfacing to C,Mixed Language Programming
5549 @anchor{gnat_ugn/the_gnat_compilation_model calling-conventions}@anchor{b5}@anchor{gnat_ugn/the_gnat_compilation_model id63}@anchor{b6}
5550 @subsection Calling Conventions
5553 @geindex Foreign Languages
5555 @geindex Calling Conventions
5557 GNAT follows standard calling sequence conventions and will thus interface
5558 to any other language that also follows these conventions. The following
5559 Convention identifiers are recognized by GNAT:
5561 @geindex Interfacing to Ada
5563 @geindex Other Ada compilers
5565 @geindex Convention Ada
5572 This indicates that the standard Ada calling sequence will be
5573 used and all Ada data items may be passed without any limitations in the
5574 case where GNAT is used to generate both the caller and callee. It is also
5575 possible to mix GNAT generated code and code generated by another Ada
5576 compiler. In this case, the data types should be restricted to simple
5577 cases, including primitive types. Whether complex data types can be passed
5578 depends on the situation. Probably it is safe to pass simple arrays, such
5579 as arrays of integers or floats. Records may or may not work, depending
5580 on whether both compilers lay them out identically. Complex structures
5581 involving variant records, access parameters, tasks, or protected types,
5582 are unlikely to be able to be passed.
5584 Note that in the case of GNAT running
5585 on a platform that supports HP Ada 83, a higher degree of compatibility
5586 can be guaranteed, and in particular records are laid out in an identical
5587 manner in the two compilers. Note also that if output from two different
5588 compilers is mixed, the program is responsible for dealing with elaboration
5589 issues. Probably the safest approach is to write the main program in the
5590 version of Ada other than GNAT, so that it takes care of its own elaboration
5591 requirements, and then call the GNAT-generated adainit procedure to ensure
5592 elaboration of the GNAT components. Consult the documentation of the other
5593 Ada compiler for further details on elaboration.
5595 However, it is not possible to mix the tasking run time of GNAT and
5596 HP Ada 83, All the tasking operations must either be entirely within
5597 GNAT compiled sections of the program, or entirely within HP Ada 83
5598 compiled sections of the program.
5601 @geindex Interfacing to Assembly
5603 @geindex Convention Assembler
5608 @item @code{Assembler}
5610 Specifies assembler as the convention. In practice this has the
5611 same effect as convention Ada (but is not equivalent in the sense of being
5612 considered the same convention).
5615 @geindex Convention Asm
5624 Equivalent to Assembler.
5626 @geindex Interfacing to COBOL
5628 @geindex Convention COBOL
5638 Data will be passed according to the conventions described
5639 in section B.4 of the Ada Reference Manual.
5644 @geindex Interfacing to C
5646 @geindex Convention C
5653 Data will be passed according to the conventions described
5654 in section B.3 of the Ada Reference Manual.
5656 A note on interfacing to a C 'varargs' function:
5660 @geindex C varargs function
5662 @geindex Interfacing to C varargs function
5664 @geindex varargs function interfaces
5666 In C, @code{varargs} allows a function to take a variable number of
5667 arguments. There is no direct equivalent in this to Ada. One
5668 approach that can be used is to create a C wrapper for each
5669 different profile and then interface to this C wrapper. For
5670 example, to print an @code{int} value using @code{printf},
5671 create a C function @code{printfi} that takes two arguments, a
5672 pointer to a string and an int, and calls @code{printf}.
5673 Then in the Ada program, use pragma @code{Import} to
5674 interface to @code{printfi}.
5676 It may work on some platforms to directly interface to
5677 a @code{varargs} function by providing a specific Ada profile
5678 for a particular call. However, this does not work on
5679 all platforms, since there is no guarantee that the
5680 calling sequence for a two argument normal C function
5681 is the same as for calling a @code{varargs} C function with
5682 the same two arguments.
5686 @geindex Convention Default
5693 @item @code{Default}
5698 @geindex Convention External
5705 @item @code{External}
5712 @geindex Interfacing to C++
5714 @geindex Convention C++
5719 @item @code{C_Plus_Plus} (or @code{CPP})
5721 This stands for C++. For most purposes this is identical to C.
5722 See the separate description of the specialized GNAT pragmas relating to
5723 C++ interfacing for further details.
5728 @geindex Interfacing to Fortran
5730 @geindex Convention Fortran
5735 @item @code{Fortran}
5737 Data will be passed according to the conventions described
5738 in section B.5 of the Ada Reference Manual.
5740 @item @code{Intrinsic}
5742 This applies to an intrinsic operation, as defined in the Ada
5743 Reference Manual. If a pragma Import (Intrinsic) applies to a subprogram,
5744 this means that the body of the subprogram is provided by the compiler itself,
5745 usually by means of an efficient code sequence, and that the user does not
5746 supply an explicit body for it. In an application program, the pragma may
5747 be applied to the following sets of names:
5753 Rotate_Left, Rotate_Right, Shift_Left, Shift_Right, Shift_Right_Arithmetic.
5754 The corresponding subprogram declaration must have
5755 two formal parameters. The
5756 first one must be a signed integer type or a modular type with a binary
5757 modulus, and the second parameter must be of type Natural.
5758 The return type must be the same as the type of the first argument. The size
5759 of this type can only be 8, 16, 32, or 64.
5762 Binary arithmetic operators: '+', '-', '*', '/'.
5763 The corresponding operator declaration must have parameters and result type
5764 that have the same root numeric type (for example, all three are long_float
5765 types). This simplifies the definition of operations that use type checking
5766 to perform dimensional checks:
5770 type Distance is new Long_Float;
5771 type Time is new Long_Float;
5772 type Velocity is new Long_Float;
5773 function "/" (D : Distance; T : Time)
5775 pragma Import (Intrinsic, "/");
5777 This common idiom is often programmed with a generic definition and an
5778 explicit body. The pragma makes it simpler to introduce such declarations.
5779 It incurs no overhead in compilation time or code size, because it is
5780 implemented as a single machine instruction.
5787 General subprogram entities. This is used to bind an Ada subprogram
5789 a compiler builtin by name with back-ends where such interfaces are
5790 available. A typical example is the set of @code{__builtin} functions
5791 exposed by the GCC back-end, as in the following example:
5794 function builtin_sqrt (F : Float) return Float;
5795 pragma Import (Intrinsic, builtin_sqrt, "__builtin_sqrtf");
5798 Most of the GCC builtins are accessible this way, and as for other
5799 import conventions (e.g. C), it is the user's responsibility to ensure
5800 that the Ada subprogram profile matches the underlying builtin
5807 @geindex Convention Stdcall
5812 @item @code{Stdcall}
5814 This is relevant only to Windows implementations of GNAT,
5815 and specifies that the @code{Stdcall} calling sequence will be used,
5816 as defined by the NT API. Nevertheless, to ease building
5817 cross-platform bindings this convention will be handled as a @code{C} calling
5818 convention on non-Windows platforms.
5823 @geindex Convention DLL
5830 This is equivalent to @code{Stdcall}.
5835 @geindex Convention Win32
5842 This is equivalent to @code{Stdcall}.
5847 @geindex Convention Stubbed
5852 @item @code{Stubbed}
5854 This is a special convention that indicates that the compiler
5855 should provide a stub body that raises @code{Program_Error}.
5858 GNAT additionally provides a useful pragma @code{Convention_Identifier}
5859 that can be used to parameterize conventions and allow additional synonyms
5860 to be specified. For example if you have legacy code in which the convention
5861 identifier Fortran77 was used for Fortran, you can use the configuration
5865 pragma Convention_Identifier (Fortran77, Fortran);
5868 And from now on the identifier Fortran77 may be used as a convention
5869 identifier (for example in an @code{Import} pragma) with the same
5872 @node Building Mixed Ada and C++ Programs,Generating Ada Bindings for C and C++ headers,Calling Conventions,Mixed Language Programming
5873 @anchor{gnat_ugn/the_gnat_compilation_model id64}@anchor{b7}@anchor{gnat_ugn/the_gnat_compilation_model building-mixed-ada-and-c-programs}@anchor{b8}
5874 @subsection Building Mixed Ada and C++ Programs
5877 A programmer inexperienced with mixed-language development may find that
5878 building an application containing both Ada and C++ code can be a
5879 challenge. This section gives a few hints that should make this task easier.
5882 * Interfacing to C++::
5883 * Linking a Mixed C++ & Ada Program::
5884 * A Simple Example::
5885 * Interfacing with C++ constructors::
5886 * Interfacing with C++ at the Class Level::
5890 @node Interfacing to C++,Linking a Mixed C++ & Ada Program,,Building Mixed Ada and C++ Programs
5891 @anchor{gnat_ugn/the_gnat_compilation_model id65}@anchor{b9}@anchor{gnat_ugn/the_gnat_compilation_model id66}@anchor{ba}
5892 @subsubsection Interfacing to C++
5895 GNAT supports interfacing with the G++ compiler (or any C++ compiler
5896 generating code that is compatible with the G++ Application Binary
5897 Interface ---see @indicateurl{http://www.codesourcery.com/archives/cxx-abi}).
5899 Interfacing can be done at 3 levels: simple data, subprograms, and
5900 classes. In the first two cases, GNAT offers a specific @code{Convention C_Plus_Plus}
5901 (or @code{CPP}) that behaves exactly like @code{Convention C}.
5902 Usually, C++ mangles the names of subprograms. To generate proper mangled
5903 names automatically, see @ref{19,,Generating Ada Bindings for C and C++ headers}).
5904 This problem can also be addressed manually in two ways:
5910 by modifying the C++ code in order to force a C convention using
5911 the @code{extern "C"} syntax.
5914 by figuring out the mangled name (using e.g. @code{nm}) and using it as the
5915 Link_Name argument of the pragma import.
5918 Interfacing at the class level can be achieved by using the GNAT specific
5919 pragmas such as @code{CPP_Constructor}. See the @cite{GNAT_Reference_Manual} for additional information.
5921 @node Linking a Mixed C++ & Ada Program,A Simple Example,Interfacing to C++,Building Mixed Ada and C++ Programs
5922 @anchor{gnat_ugn/the_gnat_compilation_model linking-a-mixed-c-ada-program}@anchor{bb}@anchor{gnat_ugn/the_gnat_compilation_model linking-a-mixed-c-and-ada-program}@anchor{bc}
5923 @subsubsection Linking a Mixed C++ & Ada Program
5926 Usually the linker of the C++ development system must be used to link
5927 mixed applications because most C++ systems will resolve elaboration
5928 issues (such as calling constructors on global class instances)
5929 transparently during the link phase. GNAT has been adapted to ease the
5930 use of a foreign linker for the last phase. Three cases can be
5937 Using GNAT and G++ (GNU C++ compiler) from the same GCC installation:
5938 The C++ linker can simply be called by using the C++ specific driver
5941 Note that if the C++ code uses inline functions, you will need to
5942 compile your C++ code with the @code{-fkeep-inline-functions} switch in
5943 order to provide an existing function implementation that the Ada code can
5947 $ g++ -c -fkeep-inline-functions file1.C
5948 $ g++ -c -fkeep-inline-functions file2.C
5949 $ gnatmake ada_unit -largs file1.o file2.o --LINK=g++
5953 Using GNAT and G++ from two different GCC installations: If both
5954 compilers are on the :envvar`PATH`, the previous method may be used. It is
5955 important to note that environment variables such as
5956 @geindex C_INCLUDE_PATH
5957 @geindex environment variable; C_INCLUDE_PATH
5958 @code{C_INCLUDE_PATH},
5959 @geindex GCC_EXEC_PREFIX
5960 @geindex environment variable; GCC_EXEC_PREFIX
5961 @code{GCC_EXEC_PREFIX},
5962 @geindex BINUTILS_ROOT
5963 @geindex environment variable; BINUTILS_ROOT
5964 @code{BINUTILS_ROOT}, and
5966 @geindex environment variable; GCC_ROOT
5967 @code{GCC_ROOT} will affect both compilers
5968 at the same time and may make one of the two compilers operate
5969 improperly if set during invocation of the wrong compiler. It is also
5970 very important that the linker uses the proper @code{libgcc.a} GCC
5971 library -- that is, the one from the C++ compiler installation. The
5972 implicit link command as suggested in the @code{gnatmake} command
5973 from the former example can be replaced by an explicit link command with
5974 the full-verbosity option in order to verify which library is used:
5978 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=c++
5981 If there is a problem due to interfering environment variables, it can
5982 be worked around by using an intermediate script. The following example
5983 shows the proper script to use when GNAT has not been installed at its
5984 default location and g++ has been installed at its default location:
5992 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=./my_script
5996 Using a non-GNU C++ compiler: The commands previously described can be
5997 used to insure that the C++ linker is used. Nonetheless, you need to add
5998 a few more parameters to the link command line, depending on the exception
6001 If the @code{setjmp} / @code{longjmp} exception mechanism is used, only the paths
6002 to the @code{libgcc} libraries are required:
6007 CC $* gcc -print-file-name=libgcc.a gcc -print-file-name=libgcc_eh.a
6008 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
6011 where CC is the name of the non-GNU C++ compiler.
6013 If the "zero cost" exception mechanism is used, and the platform
6014 supports automatic registration of exception tables (e.g., Solaris),
6015 paths to more objects are required:
6020 CC gcc -print-file-name=crtbegin.o $* \\
6021 gcc -print-file-name=libgcc.a gcc -print-file-name=libgcc_eh.a \\
6022 gcc -print-file-name=crtend.o
6023 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
6026 If the "zero cost exception" mechanism is used, and the platform
6027 doesn't support automatic registration of exception tables (e.g., HP-UX
6028 or AIX), the simple approach described above will not work and
6029 a pre-linking phase using GNAT will be necessary.
6032 Another alternative is to use the @code{gprbuild} multi-language builder
6033 which has a large knowledge base and knows how to link Ada and C++ code
6034 together automatically in most cases.
6036 @node A Simple Example,Interfacing with C++ constructors,Linking a Mixed C++ & Ada Program,Building Mixed Ada and C++ Programs
6037 @anchor{gnat_ugn/the_gnat_compilation_model id67}@anchor{bd}@anchor{gnat_ugn/the_gnat_compilation_model a-simple-example}@anchor{be}
6038 @subsubsection A Simple Example
6041 The following example, provided as part of the GNAT examples, shows how
6042 to achieve procedural interfacing between Ada and C++ in both
6043 directions. The C++ class A has two methods. The first method is exported
6044 to Ada by the means of an extern C wrapper function. The second method
6045 calls an Ada subprogram. On the Ada side, The C++ calls are modelled by
6046 a limited record with a layout comparable to the C++ class. The Ada
6047 subprogram, in turn, calls the C++ method. So, starting from the C++
6048 main program, the process passes back and forth between the two
6051 Here are the compilation commands:
6054 $ gnatmake -c simple_cpp_interface
6057 $ gnatbind -n simple_cpp_interface
6058 $ gnatlink simple_cpp_interface -o cpp_main --LINK=g++ -lstdc++ ex7.o cpp_main.o
6061 Here are the corresponding sources:
6069 void adainit (void);
6070 void adafinal (void);
6071 void method1 (A *t);
6095 class A : public Origin @{
6097 void method1 (void);
6098 void method2 (int v);
6110 extern "C" @{ void ada_method2 (A *t, int v);@}
6112 void A::method1 (void)
6115 printf ("in A::method1, a_value = %d \\n",a_value);
6118 void A::method2 (int v)
6120 ada_method2 (this, v);
6121 printf ("in A::method2, a_value = %d \\n",a_value);
6127 printf ("in A::A, a_value = %d \\n",a_value);
6132 -- simple_cpp_interface.ads
6134 package Simple_Cpp_Interface is
6137 Vptr : System.Address;
6141 pragma Convention (C, A);
6143 procedure Method1 (This : in out A);
6144 pragma Import (C, Method1);
6146 procedure Ada_Method2 (This : in out A; V : Integer);
6147 pragma Export (C, Ada_Method2);
6149 end Simple_Cpp_Interface;
6153 -- simple_cpp_interface.adb
6154 package body Simple_Cpp_Interface is
6156 procedure Ada_Method2 (This : in out A; V : Integer) is
6162 end Simple_Cpp_Interface;
6165 @node Interfacing with C++ constructors,Interfacing with C++ at the Class Level,A Simple Example,Building Mixed Ada and C++ Programs
6166 @anchor{gnat_ugn/the_gnat_compilation_model id68}@anchor{bf}@anchor{gnat_ugn/the_gnat_compilation_model interfacing-with-c-constructors}@anchor{c0}
6167 @subsubsection Interfacing with C++ constructors
6170 In order to interface with C++ constructors GNAT provides the
6171 @code{pragma CPP_Constructor} (see the @cite{GNAT_Reference_Manual}
6172 for additional information).
6173 In this section we present some common uses of C++ constructors
6174 in mixed-languages programs in GNAT.
6176 Let us assume that we need to interface with the following
6184 virtual int Get_Value ();
6185 Root(); // Default constructor
6186 Root(int v); // 1st non-default constructor
6187 Root(int v, int w); // 2nd non-default constructor
6191 For this purpose we can write the following package spec (further
6192 information on how to build this spec is available in
6193 @ref{c1,,Interfacing with C++ at the Class Level} and
6194 @ref{19,,Generating Ada Bindings for C and C++ headers}).
6197 with Interfaces.C; use Interfaces.C;
6199 type Root is tagged limited record
6203 pragma Import (CPP, Root);
6205 function Get_Value (Obj : Root) return int;
6206 pragma Import (CPP, Get_Value);
6208 function Constructor return Root;
6209 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ev");
6211 function Constructor (v : Integer) return Root;
6212 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ei");
6214 function Constructor (v, w : Integer) return Root;
6215 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Eii");
6219 On the Ada side the constructor is represented by a function (whose
6220 name is arbitrary) that returns the classwide type corresponding to
6221 the imported C++ class. Although the constructor is described as a
6222 function, it is typically a procedure with an extra implicit argument
6223 (the object being initialized) at the implementation level. GNAT
6224 issues the appropriate call, whatever it is, to get the object
6225 properly initialized.
6227 Constructors can only appear in the following contexts:
6233 On the right side of an initialization of an object of type @code{T}.
6236 On the right side of an initialization of a record component of type @code{T}.
6239 In an Ada 2005 limited aggregate.
6242 In an Ada 2005 nested limited aggregate.
6245 In an Ada 2005 limited aggregate that initializes an object built in
6246 place by an extended return statement.
6249 In a declaration of an object whose type is a class imported from C++,
6250 either the default C++ constructor is implicitly called by GNAT, or
6251 else the required C++ constructor must be explicitly called in the
6252 expression that initializes the object. For example:
6256 Obj2 : Root := Constructor;
6257 Obj3 : Root := Constructor (v => 10);
6258 Obj4 : Root := Constructor (30, 40);
6261 The first two declarations are equivalent: in both cases the default C++
6262 constructor is invoked (in the former case the call to the constructor is
6263 implicit, and in the latter case the call is explicit in the object
6264 declaration). @code{Obj3} is initialized by the C++ non-default constructor
6265 that takes an integer argument, and @code{Obj4} is initialized by the
6266 non-default C++ constructor that takes two integers.
6268 Let us derive the imported C++ class in the Ada side. For example:
6271 type DT is new Root with record
6272 C_Value : Natural := 2009;
6276 In this case the components DT inherited from the C++ side must be
6277 initialized by a C++ constructor, and the additional Ada components
6278 of type DT are initialized by GNAT. The initialization of such an
6279 object is done either by default, or by means of a function returning
6280 an aggregate of type DT, or by means of an extension aggregate.
6284 Obj6 : DT := Function_Returning_DT (50);
6285 Obj7 : DT := (Constructor (30,40) with C_Value => 50);
6288 The declaration of @code{Obj5} invokes the default constructors: the
6289 C++ default constructor of the parent type takes care of the initialization
6290 of the components inherited from Root, and GNAT takes care of the default
6291 initialization of the additional Ada components of type DT (that is,
6292 @code{C_Value} is initialized to value 2009). The order of invocation of
6293 the constructors is consistent with the order of elaboration required by
6294 Ada and C++. That is, the constructor of the parent type is always called
6295 before the constructor of the derived type.
6297 Let us now consider a record that has components whose type is imported
6298 from C++. For example:
6301 type Rec1 is limited record
6302 Data1 : Root := Constructor (10);
6303 Value : Natural := 1000;
6306 type Rec2 (D : Integer := 20) is limited record
6308 Data2 : Root := Constructor (D, 30);
6312 The initialization of an object of type @code{Rec2} will call the
6313 non-default C++ constructors specified for the imported components.
6320 Using Ada 2005 we can use limited aggregates to initialize an object
6321 invoking C++ constructors that differ from those specified in the type
6322 declarations. For example:
6325 Obj9 : Rec2 := (Rec => (Data1 => Constructor (15, 16),
6330 The above declaration uses an Ada 2005 limited aggregate to
6331 initialize @code{Obj9}, and the C++ constructor that has two integer
6332 arguments is invoked to initialize the @code{Data1} component instead
6333 of the constructor specified in the declaration of type @code{Rec1}. In
6334 Ada 2005 the box in the aggregate indicates that unspecified components
6335 are initialized using the expression (if any) available in the component
6336 declaration. That is, in this case discriminant @code{D} is initialized
6337 to value @code{20}, @code{Value} is initialized to value 1000, and the
6338 non-default C++ constructor that handles two integers takes care of
6339 initializing component @code{Data2} with values @code{20,30}.
6341 In Ada 2005 we can use the extended return statement to build the Ada
6342 equivalent to C++ non-default constructors. For example:
6345 function Constructor (V : Integer) return Rec2 is
6347 return Obj : Rec2 := (Rec => (Data1 => Constructor (V, 20),
6350 -- Further actions required for construction of
6351 -- objects of type Rec2
6357 In this example the extended return statement construct is used to
6358 build in place the returned object whose components are initialized
6359 by means of a limited aggregate. Any further action associated with
6360 the constructor can be placed inside the construct.
6362 @node Interfacing with C++ at the Class Level,,Interfacing with C++ constructors,Building Mixed Ada and C++ Programs
6363 @anchor{gnat_ugn/the_gnat_compilation_model interfacing-with-c-at-the-class-level}@anchor{c1}@anchor{gnat_ugn/the_gnat_compilation_model id69}@anchor{c2}
6364 @subsubsection Interfacing with C++ at the Class Level
6367 In this section we demonstrate the GNAT features for interfacing with
6368 C++ by means of an example making use of Ada 2005 abstract interface
6369 types. This example consists of a classification of animals; classes
6370 have been used to model our main classification of animals, and
6371 interfaces provide support for the management of secondary
6372 classifications. We first demonstrate a case in which the types and
6373 constructors are defined on the C++ side and imported from the Ada
6374 side, and latter the reverse case.
6376 The root of our derivation will be the @code{Animal} class, with a
6377 single private attribute (the @code{Age} of the animal), a constructor,
6378 and two public primitives to set and get the value of this attribute.
6383 virtual void Set_Age (int New_Age);
6385 Animal() @{Age_Count = 0;@};
6391 Abstract interface types are defined in C++ by means of classes with pure
6392 virtual functions and no data members. In our example we will use two
6393 interfaces that provide support for the common management of @code{Carnivore}
6394 and @code{Domestic} animals:
6399 virtual int Number_Of_Teeth () = 0;
6404 virtual void Set_Owner (char* Name) = 0;
6408 Using these declarations, we can now say that a @code{Dog} is an animal that is
6409 both Carnivore and Domestic, that is:
6412 class Dog : Animal, Carnivore, Domestic @{
6414 virtual int Number_Of_Teeth ();
6415 virtual void Set_Owner (char* Name);
6417 Dog(); // Constructor
6424 In the following examples we will assume that the previous declarations are
6425 located in a file named @code{animals.h}. The following package demonstrates
6426 how to import these C++ declarations from the Ada side:
6429 with Interfaces.C.Strings; use Interfaces.C.Strings;
6431 type Carnivore is limited interface;
6432 pragma Convention (C_Plus_Plus, Carnivore);
6433 function Number_Of_Teeth (X : Carnivore)
6434 return Natural is abstract;
6436 type Domestic is limited interface;
6437 pragma Convention (C_Plus_Plus, Domestic);
6439 (X : in out Domestic;
6440 Name : Chars_Ptr) is abstract;
6442 type Animal is tagged limited record
6445 pragma Import (C_Plus_Plus, Animal);
6447 procedure Set_Age (X : in out Animal; Age : Integer);
6448 pragma Import (C_Plus_Plus, Set_Age);
6450 function Age (X : Animal) return Integer;
6451 pragma Import (C_Plus_Plus, Age);
6453 function New_Animal return Animal;
6454 pragma CPP_Constructor (New_Animal);
6455 pragma Import (CPP, New_Animal, "_ZN6AnimalC1Ev");
6457 type Dog is new Animal and Carnivore and Domestic with record
6458 Tooth_Count : Natural;
6461 pragma Import (C_Plus_Plus, Dog);
6463 function Number_Of_Teeth (A : Dog) return Natural;
6464 pragma Import (C_Plus_Plus, Number_Of_Teeth);
6466 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
6467 pragma Import (C_Plus_Plus, Set_Owner);
6469 function New_Dog return Dog;
6470 pragma CPP_Constructor (New_Dog);
6471 pragma Import (CPP, New_Dog, "_ZN3DogC2Ev");
6475 Thanks to the compatibility between GNAT run-time structures and the C++ ABI,
6476 interfacing with these C++ classes is easy. The only requirement is that all
6477 the primitives and components must be declared exactly in the same order in
6480 Regarding the abstract interfaces, we must indicate to the GNAT compiler by
6481 means of a @code{pragma Convention (C_Plus_Plus)}, the convention used to pass
6482 the arguments to the called primitives will be the same as for C++. For the
6483 imported classes we use @code{pragma Import} with convention @code{C_Plus_Plus}
6484 to indicate that they have been defined on the C++ side; this is required
6485 because the dispatch table associated with these tagged types will be built
6486 in the C++ side and therefore will not contain the predefined Ada primitives
6487 which Ada would otherwise expect.
6489 As the reader can see there is no need to indicate the C++ mangled names
6490 associated with each subprogram because it is assumed that all the calls to
6491 these primitives will be dispatching calls. The only exception is the
6492 constructor, which must be registered with the compiler by means of
6493 @code{pragma CPP_Constructor} and needs to provide its associated C++
6494 mangled name because the Ada compiler generates direct calls to it.
6496 With the above packages we can now declare objects of type Dog on the Ada side
6497 and dispatch calls to the corresponding subprograms on the C++ side. We can
6498 also extend the tagged type Dog with further fields and primitives, and
6499 override some of its C++ primitives on the Ada side. For example, here we have
6500 a type derivation defined on the Ada side that inherits all the dispatching
6501 primitives of the ancestor from the C++ side.
6504 with Animals; use Animals;
6505 package Vaccinated_Animals is
6506 type Vaccinated_Dog is new Dog with null record;
6507 function Vaccination_Expired (A : Vaccinated_Dog) return Boolean;
6508 end Vaccinated_Animals;
6511 It is important to note that, because of the ABI compatibility, the programmer
6512 does not need to add any further information to indicate either the object
6513 layout or the dispatch table entry associated with each dispatching operation.
6515 Now let us define all the types and constructors on the Ada side and export
6516 them to C++, using the same hierarchy of our previous example:
6519 with Interfaces.C.Strings;
6520 use Interfaces.C.Strings;
6522 type Carnivore is limited interface;
6523 pragma Convention (C_Plus_Plus, Carnivore);
6524 function Number_Of_Teeth (X : Carnivore)
6525 return Natural is abstract;
6527 type Domestic is limited interface;
6528 pragma Convention (C_Plus_Plus, Domestic);
6530 (X : in out Domestic;
6531 Name : Chars_Ptr) is abstract;
6533 type Animal is tagged record
6536 pragma Convention (C_Plus_Plus, Animal);
6538 procedure Set_Age (X : in out Animal; Age : Integer);
6539 pragma Export (C_Plus_Plus, Set_Age);
6541 function Age (X : Animal) return Integer;
6542 pragma Export (C_Plus_Plus, Age);
6544 function New_Animal return Animal'Class;
6545 pragma Export (C_Plus_Plus, New_Animal);
6547 type Dog is new Animal and Carnivore and Domestic with record
6548 Tooth_Count : Natural;
6549 Owner : String (1 .. 30);
6551 pragma Convention (C_Plus_Plus, Dog);
6553 function Number_Of_Teeth (A : Dog) return Natural;
6554 pragma Export (C_Plus_Plus, Number_Of_Teeth);
6556 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
6557 pragma Export (C_Plus_Plus, Set_Owner);
6559 function New_Dog return Dog'Class;
6560 pragma Export (C_Plus_Plus, New_Dog);
6564 Compared with our previous example the only differences are the use of
6565 @code{pragma Convention} (instead of @code{pragma Import}), and the use of
6566 @code{pragma Export} to indicate to the GNAT compiler that the primitives will
6567 be available to C++. Thanks to the ABI compatibility, on the C++ side there is
6568 nothing else to be done; as explained above, the only requirement is that all
6569 the primitives and components are declared in exactly the same order.
6571 For completeness, let us see a brief C++ main program that uses the
6572 declarations available in @code{animals.h} (presented in our first example) to
6573 import and use the declarations from the Ada side, properly initializing and
6574 finalizing the Ada run-time system along the way:
6577 #include "animals.h"
6579 using namespace std;
6581 void Check_Carnivore (Carnivore *obj) @{...@}
6582 void Check_Domestic (Domestic *obj) @{...@}
6583 void Check_Animal (Animal *obj) @{...@}
6584 void Check_Dog (Dog *obj) @{...@}
6587 void adainit (void);
6588 void adafinal (void);
6594 Dog *obj = new_dog(); // Ada constructor
6595 Check_Carnivore (obj); // Check secondary DT
6596 Check_Domestic (obj); // Check secondary DT
6597 Check_Animal (obj); // Check primary DT
6598 Check_Dog (obj); // Check primary DT
6603 adainit (); test(); adafinal ();
6608 @node Generating Ada Bindings for C and C++ headers,Generating C Headers for Ada Specifications,Building Mixed Ada and C++ Programs,Mixed Language Programming
6609 @anchor{gnat_ugn/the_gnat_compilation_model id70}@anchor{c3}@anchor{gnat_ugn/the_gnat_compilation_model generating-ada-bindings-for-c-and-c-headers}@anchor{19}
6610 @subsection Generating Ada Bindings for C and C++ headers
6613 @geindex Binding generation (for C and C++ headers)
6615 @geindex C headers (binding generation)
6617 @geindex C++ headers (binding generation)
6619 GNAT includes a binding generator for C and C++ headers which is
6620 intended to do 95% of the tedious work of generating Ada specs from C
6621 or C++ header files.
6623 Note that this capability is not intended to generate 100% correct Ada specs,
6624 and will is some cases require manual adjustments, although it can often
6625 be used out of the box in practice.
6627 Some of the known limitations include:
6633 only very simple character constant macros are translated into Ada
6634 constants. Function macros (macros with arguments) are partially translated
6635 as comments, to be completed manually if needed.
6638 some extensions (e.g. vector types) are not supported
6641 pointers to pointers or complex structures are mapped to System.Address
6644 identifiers with identical name (except casing) will generate compilation
6645 errors (e.g. @code{shm_get} vs @code{SHM_GET}).
6648 The code generated is using the Ada 2005 syntax, which makes it
6649 easier to interface with other languages than previous versions of Ada.
6652 * Running the Binding Generator::
6653 * Generating Bindings for C++ Headers::
6658 @node Running the Binding Generator,Generating Bindings for C++ Headers,,Generating Ada Bindings for C and C++ headers
6659 @anchor{gnat_ugn/the_gnat_compilation_model id71}@anchor{c4}@anchor{gnat_ugn/the_gnat_compilation_model running-the-binding-generator}@anchor{c5}
6660 @subsubsection Running the Binding Generator
6663 The binding generator is part of the @code{gcc} compiler and can be
6664 invoked via the @code{-fdump-ada-spec} switch, which will generate Ada
6665 spec files for the header files specified on the command line, and all
6666 header files needed by these files transitively. For example:
6669 $ g++ -c -fdump-ada-spec -C /usr/include/time.h
6670 $ gcc -c -gnat05 *.ads
6673 will generate, under GNU/Linux, the following files: @code{time_h.ads},
6674 @code{bits_time_h.ads}, @code{stddef_h.ads}, @code{bits_types_h.ads} which
6675 correspond to the files @code{/usr/include/time.h},
6676 @code{/usr/include/bits/time.h}, etc..., and will then compile these Ada specs
6679 The @code{-C} switch tells @code{gcc} to extract comments from headers,
6680 and will attempt to generate corresponding Ada comments.
6682 If you want to generate a single Ada file and not the transitive closure, you
6683 can use instead the @code{-fdump-ada-spec-slim} switch.
6685 You can optionally specify a parent unit, of which all generated units will
6686 be children, using @code{-fada-spec-parent=@emph{unit}}.
6688 Note that we recommend when possible to use the @emph{g++} driver to
6689 generate bindings, even for most C headers, since this will in general
6690 generate better Ada specs. For generating bindings for C++ headers, it is
6691 mandatory to use the @emph{g++} command, or @emph{gcc -x c++} which
6692 is equivalent in this case. If @emph{g++} cannot work on your C headers
6693 because of incompatibilities between C and C++, then you can fallback to
6696 For an example of better bindings generated from the C++ front-end,
6697 the name of the parameters (when available) are actually ignored by the C
6698 front-end. Consider the following C header:
6701 extern void foo (int variable);
6704 with the C front-end, @code{variable} is ignored, and the above is handled as:
6707 extern void foo (int);
6710 generating a generic:
6713 procedure foo (param1 : int);
6716 with the C++ front-end, the name is available, and we generate:
6719 procedure foo (variable : int);
6722 In some cases, the generated bindings will be more complete or more meaningful
6723 when defining some macros, which you can do via the @code{-D} switch. This
6724 is for example the case with @code{Xlib.h} under GNU/Linux:
6727 $ g++ -c -fdump-ada-spec -DXLIB_ILLEGAL_ACCESS -C /usr/include/X11/Xlib.h
6730 The above will generate more complete bindings than a straight call without
6731 the @code{-DXLIB_ILLEGAL_ACCESS} switch.
6733 In other cases, it is not possible to parse a header file in a stand-alone
6734 manner, because other include files need to be included first. In this
6735 case, the solution is to create a small header file including the needed
6736 @code{#include} and possible @code{#define} directives. For example, to
6737 generate Ada bindings for @code{readline/readline.h}, you need to first
6738 include @code{stdio.h}, so you can create a file with the following two
6739 lines in e.g. @code{readline1.h}:
6743 #include <readline/readline.h>
6746 and then generate Ada bindings from this file:
6749 $ g++ -c -fdump-ada-spec readline1.h
6752 @node Generating Bindings for C++ Headers,Switches,Running the Binding Generator,Generating Ada Bindings for C and C++ headers
6753 @anchor{gnat_ugn/the_gnat_compilation_model id72}@anchor{c6}@anchor{gnat_ugn/the_gnat_compilation_model generating-bindings-for-c-headers}@anchor{c7}
6754 @subsubsection Generating Bindings for C++ Headers
6757 Generating bindings for C++ headers is done using the same options, always
6758 with the @emph{g++} compiler. Note that generating Ada spec from C++ headers is a
6759 much more complex job and support for C++ headers is much more limited that
6760 support for C headers. As a result, you will need to modify the resulting
6761 bindings by hand more extensively when using C++ headers.
6763 In this mode, C++ classes will be mapped to Ada tagged types, constructors
6764 will be mapped using the @code{CPP_Constructor} pragma, and when possible,
6765 multiple inheritance of abstract classes will be mapped to Ada interfaces
6766 (see the @emph{Interfacing to C++} section in the @cite{GNAT Reference Manual}
6767 for additional information on interfacing to C++).
6769 For example, given the following C++ header file:
6774 virtual int Number_Of_Teeth () = 0;
6779 virtual void Set_Owner (char* Name) = 0;
6785 virtual void Set_Age (int New_Age);
6788 class Dog : Animal, Carnivore, Domestic @{
6793 virtual int Number_Of_Teeth ();
6794 virtual void Set_Owner (char* Name);
6800 The corresponding Ada code is generated:
6803 package Class_Carnivore is
6804 type Carnivore is limited interface;
6805 pragma Import (CPP, Carnivore);
6807 function Number_Of_Teeth (this : access Carnivore) return int is abstract;
6809 use Class_Carnivore;
6811 package Class_Domestic is
6812 type Domestic is limited interface;
6813 pragma Import (CPP, Domestic);
6816 (this : access Domestic;
6817 Name : Interfaces.C.Strings.chars_ptr) is abstract;
6821 package Class_Animal is
6822 type Animal is tagged limited record
6823 Age_Count : aliased int;
6825 pragma Import (CPP, Animal);
6827 procedure Set_Age (this : access Animal; New_Age : int);
6828 pragma Import (CPP, Set_Age, "_ZN6Animal7Set_AgeEi");
6832 package Class_Dog is
6833 type Dog is new Animal and Carnivore and Domestic with record
6834 Tooth_Count : aliased int;
6835 Owner : Interfaces.C.Strings.chars_ptr;
6837 pragma Import (CPP, Dog);
6839 function Number_Of_Teeth (this : access Dog) return int;
6840 pragma Import (CPP, Number_Of_Teeth, "_ZN3Dog15Number_Of_TeethEv");
6843 (this : access Dog; Name : Interfaces.C.Strings.chars_ptr);
6844 pragma Import (CPP, Set_Owner, "_ZN3Dog9Set_OwnerEPc");
6846 function New_Dog return Dog;
6847 pragma CPP_Constructor (New_Dog);
6848 pragma Import (CPP, New_Dog, "_ZN3DogC1Ev");
6853 @node Switches,,Generating Bindings for C++ Headers,Generating Ada Bindings for C and C++ headers
6854 @anchor{gnat_ugn/the_gnat_compilation_model switches}@anchor{c8}@anchor{gnat_ugn/the_gnat_compilation_model switches-for-ada-binding-generation}@anchor{c9}
6855 @subsubsection Switches
6858 @geindex -fdump-ada-spec (gcc)
6863 @item @code{-fdump-ada-spec}
6865 Generate Ada spec files for the given header files transitively (including
6866 all header files that these headers depend upon).
6869 @geindex -fdump-ada-spec-slim (gcc)
6874 @item @code{-fdump-ada-spec-slim}
6876 Generate Ada spec files for the header files specified on the command line
6880 @geindex -fada-spec-parent (gcc)
6885 @item @code{-fada-spec-parent=@emph{unit}}
6887 Specifies that all files generated by @code{-fdump-ada-spec} are
6888 to be child units of the specified parent unit.
6898 Extract comments from headers and generate Ada comments in the Ada spec files.
6901 @node Generating C Headers for Ada Specifications,,Generating Ada Bindings for C and C++ headers,Mixed Language Programming
6902 @anchor{gnat_ugn/the_gnat_compilation_model generating-c-headers-for-ada-specifications}@anchor{ca}@anchor{gnat_ugn/the_gnat_compilation_model id73}@anchor{cb}
6903 @subsection Generating C Headers for Ada Specifications
6906 @geindex Binding generation (for Ada specs)
6908 @geindex C headers (binding generation)
6910 GNAT includes a C header generator for Ada specifications which supports
6911 Ada types that have a direct mapping to C types. This includes in particular
6927 Composition of the above types
6930 Constant declarations
6936 Subprogram declarations
6940 * Running the C Header Generator::
6944 @node Running the C Header Generator,,,Generating C Headers for Ada Specifications
6945 @anchor{gnat_ugn/the_gnat_compilation_model running-the-c-header-generator}@anchor{cc}
6946 @subsubsection Running the C Header Generator
6949 The C header generator is part of the GNAT compiler and can be invoked via
6950 the @code{-gnatceg} combination of switches, which will generate a @code{.h}
6951 file corresponding to the given input file (Ada spec or body). Note that
6952 only spec files are processed in any case, so giving a spec or a body file
6953 as input is equivalent. For example:
6956 $ gcc -c -gnatceg pack1.ads
6959 will generate a self-contained file called @code{pack1.h} including
6960 common definitions from the Ada Standard package, followed by the
6961 definitions included in @code{pack1.ads}, as well as all the other units
6962 withed by this file.
6964 For instance, given the following Ada files:
6968 type Int is range 1 .. 10;
6977 Field1, Field2 : Pack2.Int;
6980 Global : Rec := (1, 2);
6982 procedure Proc1 (R : Rec);
6983 procedure Proc2 (R : in out Rec);
6987 The above @code{gcc} command will generate the following @code{pack1.h} file:
6990 /* Standard definitions skipped */
6993 typedef short_short_integer pack2__TintB;
6994 typedef pack2__TintB pack2__int;
6995 #endif /* PACK2_ADS */
6999 typedef struct _pack1__rec @{
7003 extern pack1__rec pack1__global;
7004 extern void pack1__proc1(const pack1__rec r);
7005 extern void pack1__proc2(pack1__rec *r);
7006 #endif /* PACK1_ADS */
7009 You can then @code{include} @code{pack1.h} from a C source file and use the types,
7010 call subprograms, reference objects, and constants.
7012 @node GNAT and Other Compilation Models,Using GNAT Files with External Tools,Mixed Language Programming,The GNAT Compilation Model
7013 @anchor{gnat_ugn/the_gnat_compilation_model id74}@anchor{cd}@anchor{gnat_ugn/the_gnat_compilation_model gnat-and-other-compilation-models}@anchor{45}
7014 @section GNAT and Other Compilation Models
7017 This section compares the GNAT model with the approaches taken in
7018 other environents, first the C/C++ model and then the mechanism that
7019 has been used in other Ada systems, in particular those traditionally
7023 * Comparison between GNAT and C/C++ Compilation Models::
7024 * Comparison between GNAT and Conventional Ada Library Models::
7028 @node Comparison between GNAT and C/C++ Compilation Models,Comparison between GNAT and Conventional Ada Library Models,,GNAT and Other Compilation Models
7029 @anchor{gnat_ugn/the_gnat_compilation_model comparison-between-gnat-and-c-c-compilation-models}@anchor{ce}@anchor{gnat_ugn/the_gnat_compilation_model id75}@anchor{cf}
7030 @subsection Comparison between GNAT and C/C++ Compilation Models
7033 The GNAT model of compilation is close to the C and C++ models. You can
7034 think of Ada specs as corresponding to header files in C. As in C, you
7035 don't need to compile specs; they are compiled when they are used. The
7036 Ada @emph{with} is similar in effect to the @code{#include} of a C
7039 One notable difference is that, in Ada, you may compile specs separately
7040 to check them for semantic and syntactic accuracy. This is not always
7041 possible with C headers because they are fragments of programs that have
7042 less specific syntactic or semantic rules.
7044 The other major difference is the requirement for running the binder,
7045 which performs two important functions. First, it checks for
7046 consistency. In C or C++, the only defense against assembling
7047 inconsistent programs lies outside the compiler, in a makefile, for
7048 example. The binder satisfies the Ada requirement that it be impossible
7049 to construct an inconsistent program when the compiler is used in normal
7052 @geindex Elaboration order control
7054 The other important function of the binder is to deal with elaboration
7055 issues. There are also elaboration issues in C++ that are handled
7056 automatically. This automatic handling has the advantage of being
7057 simpler to use, but the C++ programmer has no control over elaboration.
7058 Where @code{gnatbind} might complain there was no valid order of
7059 elaboration, a C++ compiler would simply construct a program that
7060 malfunctioned at run time.
7062 @node Comparison between GNAT and Conventional Ada Library Models,,Comparison between GNAT and C/C++ Compilation Models,GNAT and Other Compilation Models
7063 @anchor{gnat_ugn/the_gnat_compilation_model comparison-between-gnat-and-conventional-ada-library-models}@anchor{d0}@anchor{gnat_ugn/the_gnat_compilation_model id76}@anchor{d1}
7064 @subsection Comparison between GNAT and Conventional Ada Library Models
7067 This section is intended for Ada programmers who have
7068 used an Ada compiler implementing the traditional Ada library
7069 model, as described in the Ada Reference Manual.
7071 @geindex GNAT library
7073 In GNAT, there is no 'library' in the normal sense. Instead, the set of
7074 source files themselves acts as the library. Compiling Ada programs does
7075 not generate any centralized information, but rather an object file and
7076 a ALI file, which are of interest only to the binder and linker.
7077 In a traditional system, the compiler reads information not only from
7078 the source file being compiled, but also from the centralized library.
7079 This means that the effect of a compilation depends on what has been
7080 previously compiled. In particular:
7086 When a unit is @emph{with}ed, the unit seen by the compiler corresponds
7087 to the version of the unit most recently compiled into the library.
7090 Inlining is effective only if the necessary body has already been
7091 compiled into the library.
7094 Compiling a unit may obsolete other units in the library.
7097 In GNAT, compiling one unit never affects the compilation of any other
7098 units because the compiler reads only source files. Only changes to source
7099 files can affect the results of a compilation. In particular:
7105 When a unit is @emph{with}ed, the unit seen by the compiler corresponds
7106 to the source version of the unit that is currently accessible to the
7112 Inlining requires the appropriate source files for the package or
7113 subprogram bodies to be available to the compiler. Inlining is always
7114 effective, independent of the order in which units are compiled.
7117 Compiling a unit never affects any other compilations. The editing of
7118 sources may cause previous compilations to be out of date if they
7119 depended on the source file being modified.
7122 The most important result of these differences is that order of compilation
7123 is never significant in GNAT. There is no situation in which one is
7124 required to do one compilation before another. What shows up as order of
7125 compilation requirements in the traditional Ada library becomes, in
7126 GNAT, simple source dependencies; in other words, there is only a set
7127 of rules saying what source files must be present when a file is
7130 @node Using GNAT Files with External Tools,,GNAT and Other Compilation Models,The GNAT Compilation Model
7131 @anchor{gnat_ugn/the_gnat_compilation_model using-gnat-files-with-external-tools}@anchor{1a}@anchor{gnat_ugn/the_gnat_compilation_model id77}@anchor{d2}
7132 @section Using GNAT Files with External Tools
7135 This section explains how files that are produced by GNAT may be
7136 used with tools designed for other languages.
7139 * Using Other Utility Programs with GNAT::
7140 * The External Symbol Naming Scheme of GNAT::
7144 @node Using Other Utility Programs with GNAT,The External Symbol Naming Scheme of GNAT,,Using GNAT Files with External Tools
7145 @anchor{gnat_ugn/the_gnat_compilation_model using-other-utility-programs-with-gnat}@anchor{d3}@anchor{gnat_ugn/the_gnat_compilation_model id78}@anchor{d4}
7146 @subsection Using Other Utility Programs with GNAT
7149 The object files generated by GNAT are in standard system format and in
7150 particular the debugging information uses this format. This means
7151 programs generated by GNAT can be used with existing utilities that
7152 depend on these formats.
7154 In general, any utility program that works with C will also often work with
7155 Ada programs generated by GNAT. This includes software utilities such as
7156 gprof (a profiling program), gdb (the FSF debugger), and utilities such
7159 @node The External Symbol Naming Scheme of GNAT,,Using Other Utility Programs with GNAT,Using GNAT Files with External Tools
7160 @anchor{gnat_ugn/the_gnat_compilation_model the-external-symbol-naming-scheme-of-gnat}@anchor{d5}@anchor{gnat_ugn/the_gnat_compilation_model id79}@anchor{d6}
7161 @subsection The External Symbol Naming Scheme of GNAT
7164 In order to interpret the output from GNAT, when using tools that are
7165 originally intended for use with other languages, it is useful to
7166 understand the conventions used to generate link names from the Ada
7169 All link names are in all lowercase letters. With the exception of library
7170 procedure names, the mechanism used is simply to use the full expanded
7171 Ada name with dots replaced by double underscores. For example, suppose
7172 we have the following package spec:
7180 @geindex pragma Export
7182 The variable @code{MN} has a full expanded Ada name of @code{QRS.MN}, so
7183 the corresponding link name is @code{qrs__mn}.
7184 Of course if a @code{pragma Export} is used this may be overridden:
7189 pragma Export (Var1, C, External_Name => "var1_name");
7191 pragma Export (Var2, C, Link_Name => "var2_link_name");
7195 In this case, the link name for @code{Var1} is whatever link name the
7196 C compiler would assign for the C function @code{var1_name}. This typically
7197 would be either @code{var1_name} or @code{_var1_name}, depending on operating
7198 system conventions, but other possibilities exist. The link name for
7199 @code{Var2} is @code{var2_link_name}, and this is not operating system
7202 One exception occurs for library level procedures. A potential ambiguity
7203 arises between the required name @code{_main} for the C main program,
7204 and the name we would otherwise assign to an Ada library level procedure
7205 called @code{Main} (which might well not be the main program).
7207 To avoid this ambiguity, we attach the prefix @code{_ada_} to such
7208 names. So if we have a library level procedure such as:
7211 procedure Hello (S : String);
7214 the external name of this procedure will be @code{_ada_hello}.
7216 @c -- Example: A |withing| unit has a |with| clause, it |withs| a |withed| unit
7218 @node Building Executable Programs with GNAT,GNAT Utility Programs,The GNAT Compilation Model,Top
7219 @anchor{gnat_ugn/building_executable_programs_with_gnat building-executable-programs-with-gnat}@anchor{a}@anchor{gnat_ugn/building_executable_programs_with_gnat doc}@anchor{d7}@anchor{gnat_ugn/building_executable_programs_with_gnat id1}@anchor{d8}
7220 @chapter Building Executable Programs with GNAT
7223 This chapter describes first the gnatmake tool
7224 (@ref{1b,,Building with gnatmake}),
7225 which automatically determines the set of sources
7226 needed by an Ada compilation unit and executes the necessary
7227 (re)compilations, binding and linking.
7228 It also explains how to use each tool individually: the
7229 compiler (gcc, see @ref{1c,,Compiling with gcc}),
7230 binder (gnatbind, see @ref{1d,,Binding with gnatbind}),
7231 and linker (gnatlink, see @ref{1e,,Linking with gnatlink})
7232 to build executable programs.
7233 Finally, this chapter provides examples of
7234 how to make use of the general GNU make mechanism
7235 in a GNAT context (see @ref{1f,,Using the GNU make Utility}).
7239 * Building with gnatmake::
7240 * Compiling with gcc::
7241 * Compiler Switches::
7243 * Binding with gnatbind::
7244 * Linking with gnatlink::
7245 * Using the GNU make Utility::
7249 @node Building with gnatmake,Compiling with gcc,,Building Executable Programs with GNAT
7250 @anchor{gnat_ugn/building_executable_programs_with_gnat the-gnat-make-program-gnatmake}@anchor{1b}@anchor{gnat_ugn/building_executable_programs_with_gnat building-with-gnatmake}@anchor{d9}
7251 @section Building with @code{gnatmake}
7256 A typical development cycle when working on an Ada program consists of
7257 the following steps:
7263 Edit some sources to fix bugs;
7269 Compile all sources affected;
7272 Rebind and relink; and
7278 @geindex Dependency rules (compilation)
7280 The third step in particular can be tricky, because not only do the modified
7281 files have to be compiled, but any files depending on these files must also be
7282 recompiled. The dependency rules in Ada can be quite complex, especially
7283 in the presence of overloading, @code{use} clauses, generics and inlined
7286 @code{gnatmake} automatically takes care of the third and fourth steps
7287 of this process. It determines which sources need to be compiled,
7288 compiles them, and binds and links the resulting object files.
7290 Unlike some other Ada make programs, the dependencies are always
7291 accurately recomputed from the new sources. The source based approach of
7292 the GNAT compilation model makes this possible. This means that if
7293 changes to the source program cause corresponding changes in
7294 dependencies, they will always be tracked exactly correctly by
7297 Note that for advanced forms of project structure, we recommend creating
7298 a project file as explained in the @emph{GNAT_Project_Manager} chapter in the
7299 @emph{GPRbuild User's Guide}, and using the
7300 @code{gprbuild} tool which supports building with project files and works similarly
7304 * Running gnatmake::
7305 * Switches for gnatmake::
7306 * Mode Switches for gnatmake::
7307 * Notes on the Command Line::
7308 * How gnatmake Works::
7309 * Examples of gnatmake Usage::
7313 @node Running gnatmake,Switches for gnatmake,,Building with gnatmake
7314 @anchor{gnat_ugn/building_executable_programs_with_gnat running-gnatmake}@anchor{da}@anchor{gnat_ugn/building_executable_programs_with_gnat id2}@anchor{db}
7315 @subsection Running @code{gnatmake}
7318 The usual form of the @code{gnatmake} command is
7321 $ gnatmake [<switches>] <file_name> [<file_names>] [<mode_switches>]
7324 The only required argument is one @code{file_name}, which specifies
7325 a compilation unit that is a main program. Several @code{file_names} can be
7326 specified: this will result in several executables being built.
7327 If @code{switches} are present, they can be placed before the first
7328 @code{file_name}, between @code{file_names} or after the last @code{file_name}.
7329 If @code{mode_switches} are present, they must always be placed after
7330 the last @code{file_name} and all @code{switches}.
7332 If you are using standard file extensions (@code{.adb} and
7333 @code{.ads}), then the
7334 extension may be omitted from the @code{file_name} arguments. However, if
7335 you are using non-standard extensions, then it is required that the
7336 extension be given. A relative or absolute directory path can be
7337 specified in a @code{file_name}, in which case, the input source file will
7338 be searched for in the specified directory only. Otherwise, the input
7339 source file will first be searched in the directory where
7340 @code{gnatmake} was invoked and if it is not found, it will be search on
7341 the source path of the compiler as described in
7342 @ref{89,,Search Paths and the Run-Time Library (RTL)}.
7344 All @code{gnatmake} output (except when you specify @code{-M}) is sent to
7345 @code{stderr}. The output produced by the
7346 @code{-M} switch is sent to @code{stdout}.
7348 @node Switches for gnatmake,Mode Switches for gnatmake,Running gnatmake,Building with gnatmake
7349 @anchor{gnat_ugn/building_executable_programs_with_gnat switches-for-gnatmake}@anchor{dc}@anchor{gnat_ugn/building_executable_programs_with_gnat id3}@anchor{dd}
7350 @subsection Switches for @code{gnatmake}
7353 You may specify any of the following switches to @code{gnatmake}:
7355 @geindex --version (gnatmake)
7360 @item @code{--version}
7362 Display Copyright and version, then exit disregarding all other options.
7365 @geindex --help (gnatmake)
7372 If @code{--version} was not used, display usage, then exit disregarding
7376 @geindex --GCC=compiler_name (gnatmake)
7381 @item @code{--GCC=@emph{compiler_name}}
7383 Program used for compiling. The default is @code{gcc}. You need to use
7384 quotes around @code{compiler_name} if @code{compiler_name} contains
7385 spaces or other separator characters.
7386 As an example @code{--GCC="foo -x -y"}
7387 will instruct @code{gnatmake} to use @code{foo -x -y} as your
7388 compiler. A limitation of this syntax is that the name and path name of
7389 the executable itself must not include any embedded spaces. Note that
7390 switch @code{-c} is always inserted after your command name. Thus in the
7391 above example the compiler command that will be used by @code{gnatmake}
7392 will be @code{foo -c -x -y}. If several @code{--GCC=compiler_name} are
7393 used, only the last @code{compiler_name} is taken into account. However,
7394 all the additional switches are also taken into account. Thus,
7395 @code{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
7396 @code{--GCC="bar -x -y -z -t"}.
7399 @geindex --GNATBIND=binder_name (gnatmake)
7404 @item @code{--GNATBIND=@emph{binder_name}}
7406 Program used for binding. The default is @code{gnatbind}. You need to
7407 use quotes around @code{binder_name} if @code{binder_name} contains spaces
7408 or other separator characters.
7409 As an example @code{--GNATBIND="bar -x -y"}
7410 will instruct @code{gnatmake} to use @code{bar -x -y} as your
7411 binder. Binder switches that are normally appended by @code{gnatmake}
7412 to @code{gnatbind} are now appended to the end of @code{bar -x -y}.
7413 A limitation of this syntax is that the name and path name of the executable
7414 itself must not include any embedded spaces.
7417 @geindex --GNATLINK=linker_name (gnatmake)
7422 @item @code{--GNATLINK=@emph{linker_name}}
7424 Program used for linking. The default is @code{gnatlink}. You need to
7425 use quotes around @code{linker_name} if @code{linker_name} contains spaces
7426 or other separator characters.
7427 As an example @code{--GNATLINK="lan -x -y"}
7428 will instruct @code{gnatmake} to use @code{lan -x -y} as your
7429 linker. Linker switches that are normally appended by @code{gnatmake} to
7430 @code{gnatlink} are now appended to the end of @code{lan -x -y}.
7431 A limitation of this syntax is that the name and path name of the executable
7432 itself must not include any embedded spaces.
7434 @item @code{--create-map-file}
7436 When linking an executable, create a map file. The name of the map file
7437 has the same name as the executable with extension ".map".
7439 @item @code{--create-map-file=@emph{mapfile}}
7441 When linking an executable, create a map file with the specified name.
7444 @geindex --create-missing-dirs (gnatmake)
7449 @item @code{--create-missing-dirs}
7451 When using project files (@code{-P@emph{project}}), automatically create
7452 missing object directories, library directories and exec
7455 @item @code{--single-compile-per-obj-dir}
7457 Disallow simultaneous compilations in the same object directory when
7458 project files are used.
7460 @item @code{--subdirs=@emph{subdir}}
7462 Actual object directory of each project file is the subdirectory subdir of the
7463 object directory specified or defaulted in the project file.
7465 @item @code{--unchecked-shared-lib-imports}
7467 By default, shared library projects are not allowed to import static library
7468 projects. When this switch is used on the command line, this restriction is
7471 @item @code{--source-info=@emph{source info file}}
7473 Specify a source info file. This switch is active only when project files
7474 are used. If the source info file is specified as a relative path, then it is
7475 relative to the object directory of the main project. If the source info file
7476 does not exist, then after the Project Manager has successfully parsed and
7477 processed the project files and found the sources, it creates the source info
7478 file. If the source info file already exists and can be read successfully,
7479 then the Project Manager will get all the needed information about the sources
7480 from the source info file and will not look for them. This reduces the time
7481 to process the project files, especially when looking for sources that take a
7482 long time. If the source info file exists but cannot be parsed successfully,
7483 the Project Manager will attempt to recreate it. If the Project Manager fails
7484 to create the source info file, a message is issued, but gnatmake does not
7485 fail. @code{gnatmake} "trusts" the source info file. This means that
7486 if the source files have changed (addition, deletion, moving to a different
7487 source directory), then the source info file need to be deleted and recreated.
7490 @geindex -a (gnatmake)
7497 Consider all files in the make process, even the GNAT internal system
7498 files (for example, the predefined Ada library files), as well as any
7499 locked files. Locked files are files whose ALI file is write-protected.
7501 @code{gnatmake} does not check these files,
7502 because the assumption is that the GNAT internal files are properly up
7503 to date, and also that any write protected ALI files have been properly
7504 installed. Note that if there is an installation problem, such that one
7505 of these files is not up to date, it will be properly caught by the
7507 You may have to specify this switch if you are working on GNAT
7508 itself. The switch @code{-a} is also useful
7509 in conjunction with @code{-f}
7510 if you need to recompile an entire application,
7511 including run-time files, using special configuration pragmas,
7512 such as a @code{Normalize_Scalars} pragma.
7515 @code{gnatmake -a} compiles all GNAT
7517 @code{gcc -c -gnatpg} rather than @code{gcc -c}.
7520 @geindex -b (gnatmake)
7527 Bind only. Can be combined with @code{-c} to do
7528 compilation and binding, but no link.
7529 Can be combined with @code{-l}
7530 to do binding and linking. When not combined with
7532 all the units in the closure of the main program must have been previously
7533 compiled and must be up to date. The root unit specified by @code{file_name}
7534 may be given without extension, with the source extension or, if no GNAT
7535 Project File is specified, with the ALI file extension.
7538 @geindex -c (gnatmake)
7545 Compile only. Do not perform binding, except when @code{-b}
7546 is also specified. Do not perform linking, except if both
7548 @code{-l} are also specified.
7549 If the root unit specified by @code{file_name} is not a main unit, this is the
7550 default. Otherwise @code{gnatmake} will attempt binding and linking
7551 unless all objects are up to date and the executable is more recent than
7555 @geindex -C (gnatmake)
7562 Use a temporary mapping file. A mapping file is a way to communicate
7563 to the compiler two mappings: from unit names to file names (without
7564 any directory information) and from file names to path names (with
7565 full directory information). A mapping file can make the compiler's
7566 file searches faster, especially if there are many source directories,
7567 or the sources are read over a slow network connection. If
7568 @code{-P} is used, a mapping file is always used, so
7569 @code{-C} is unnecessary; in this case the mapping file
7570 is initially populated based on the project file. If
7571 @code{-C} is used without
7573 the mapping file is initially empty. Each invocation of the compiler
7574 will add any newly accessed sources to the mapping file.
7577 @geindex -C= (gnatmake)
7582 @item @code{-C=@emph{file}}
7584 Use a specific mapping file. The file, specified as a path name (absolute or
7585 relative) by this switch, should already exist, otherwise the switch is
7586 ineffective. The specified mapping file will be communicated to the compiler.
7587 This switch is not compatible with a project file
7588 (-P`file`) or with multiple compiling processes
7589 (-jnnn, when nnn is greater than 1).
7592 @geindex -d (gnatmake)
7599 Display progress for each source, up to date or not, as a single line:
7602 completed x out of y (zz%)
7605 If the file needs to be compiled this is displayed after the invocation of
7606 the compiler. These lines are displayed even in quiet output mode.
7609 @geindex -D (gnatmake)
7614 @item @code{-D @emph{dir}}
7616 Put all object files and ALI file in directory @code{dir}.
7617 If the @code{-D} switch is not used, all object files
7618 and ALI files go in the current working directory.
7620 This switch cannot be used when using a project file.
7623 @geindex -eI (gnatmake)
7628 @item @code{-eI@emph{nnn}}
7630 Indicates that the main source is a multi-unit source and the rank of the unit
7631 in the source file is nnn. nnn needs to be a positive number and a valid
7632 index in the source. This switch cannot be used when @code{gnatmake} is
7633 invoked for several mains.
7636 @geindex -eL (gnatmake)
7638 @geindex symbolic links
7645 Follow all symbolic links when processing project files.
7646 This should be used if your project uses symbolic links for files or
7647 directories, but is not needed in other cases.
7649 @geindex naming scheme
7651 This also assumes that no directory matches the naming scheme for files (for
7652 instance that you do not have a directory called "sources.ads" when using the
7653 default GNAT naming scheme).
7655 When you do not have to use this switch (i.e., by default), gnatmake is able to
7656 save a lot of system calls (several per source file and object file), which
7657 can result in a significant speed up to load and manipulate a project file,
7658 especially when using source files from a remote system.
7661 @geindex -eS (gnatmake)
7668 Output the commands for the compiler, the binder and the linker
7670 instead of standard error.
7673 @geindex -f (gnatmake)
7680 Force recompilations. Recompile all sources, even though some object
7681 files may be up to date, but don't recompile predefined or GNAT internal
7682 files or locked files (files with a write-protected ALI file),
7683 unless the @code{-a} switch is also specified.
7686 @geindex -F (gnatmake)
7693 When using project files, if some errors or warnings are detected during
7694 parsing and verbose mode is not in effect (no use of switch
7695 -v), then error lines start with the full path name of the project
7696 file, rather than its simple file name.
7699 @geindex -g (gnatmake)
7706 Enable debugging. This switch is simply passed to the compiler and to the
7710 @geindex -i (gnatmake)
7717 In normal mode, @code{gnatmake} compiles all object files and ALI files
7718 into the current directory. If the @code{-i} switch is used,
7719 then instead object files and ALI files that already exist are overwritten
7720 in place. This means that once a large project is organized into separate
7721 directories in the desired manner, then @code{gnatmake} will automatically
7722 maintain and update this organization. If no ALI files are found on the
7723 Ada object path (see @ref{89,,Search Paths and the Run-Time Library (RTL)}),
7724 the new object and ALI files are created in the
7725 directory containing the source being compiled. If another organization
7726 is desired, where objects and sources are kept in different directories,
7727 a useful technique is to create dummy ALI files in the desired directories.
7728 When detecting such a dummy file, @code{gnatmake} will be forced to
7729 recompile the corresponding source file, and it will be put the resulting
7730 object and ALI files in the directory where it found the dummy file.
7733 @geindex -j (gnatmake)
7735 @geindex Parallel make
7740 @item @code{-j@emph{n}}
7742 Use @code{n} processes to carry out the (re)compilations. On a multiprocessor
7743 machine compilations will occur in parallel. If @code{n} is 0, then the
7744 maximum number of parallel compilations is the number of core processors
7745 on the platform. In the event of compilation errors, messages from various
7746 compilations might get interspersed (but @code{gnatmake} will give you the
7747 full ordered list of failing compiles at the end). If this is problematic,
7748 rerun the make process with n set to 1 to get a clean list of messages.
7751 @geindex -k (gnatmake)
7758 Keep going. Continue as much as possible after a compilation error. To
7759 ease the programmer's task in case of compilation errors, the list of
7760 sources for which the compile fails is given when @code{gnatmake}
7763 If @code{gnatmake} is invoked with several @code{file_names} and with this
7764 switch, if there are compilation errors when building an executable,
7765 @code{gnatmake} will not attempt to build the following executables.
7768 @geindex -l (gnatmake)
7775 Link only. Can be combined with @code{-b} to binding
7776 and linking. Linking will not be performed if combined with
7778 but not with @code{-b}.
7779 When not combined with @code{-b}
7780 all the units in the closure of the main program must have been previously
7781 compiled and must be up to date, and the main program needs to have been bound.
7782 The root unit specified by @code{file_name}
7783 may be given without extension, with the source extension or, if no GNAT
7784 Project File is specified, with the ALI file extension.
7787 @geindex -m (gnatmake)
7794 Specify that the minimum necessary amount of recompilations
7795 be performed. In this mode @code{gnatmake} ignores time
7796 stamp differences when the only
7797 modifications to a source file consist in adding/removing comments,
7798 empty lines, spaces or tabs. This means that if you have changed the
7799 comments in a source file or have simply reformatted it, using this
7800 switch will tell @code{gnatmake} not to recompile files that depend on it
7801 (provided other sources on which these files depend have undergone no
7802 semantic modifications). Note that the debugging information may be
7803 out of date with respect to the sources if the @code{-m} switch causes
7804 a compilation to be switched, so the use of this switch represents a
7805 trade-off between compilation time and accurate debugging information.
7808 @geindex Dependencies
7809 @geindex producing list
7811 @geindex -M (gnatmake)
7818 Check if all objects are up to date. If they are, output the object
7819 dependences to @code{stdout} in a form that can be directly exploited in
7820 a @code{Makefile}. By default, each source file is prefixed with its
7821 (relative or absolute) directory name. This name is whatever you
7822 specified in the various @code{-aI}
7823 and @code{-I} switches. If you use
7824 @code{gnatmake -M} @code{-q}
7825 (see below), only the source file names,
7826 without relative paths, are output. If you just specify the @code{-M}
7827 switch, dependencies of the GNAT internal system files are omitted. This
7828 is typically what you want. If you also specify
7829 the @code{-a} switch,
7830 dependencies of the GNAT internal files are also listed. Note that
7831 dependencies of the objects in external Ada libraries (see
7832 switch @code{-aL@emph{dir}} in the following list)
7836 @geindex -n (gnatmake)
7843 Don't compile, bind, or link. Checks if all objects are up to date.
7844 If they are not, the full name of the first file that needs to be
7845 recompiled is printed.
7846 Repeated use of this option, followed by compiling the indicated source
7847 file, will eventually result in recompiling all required units.
7850 @geindex -o (gnatmake)
7855 @item @code{-o @emph{exec_name}}
7857 Output executable name. The name of the final executable program will be
7858 @code{exec_name}. If the @code{-o} switch is omitted the default
7859 name for the executable will be the name of the input file in appropriate form
7860 for an executable file on the host system.
7862 This switch cannot be used when invoking @code{gnatmake} with several
7866 @geindex -p (gnatmake)
7873 Same as @code{--create-missing-dirs}
7876 @geindex -P (gnatmake)
7881 @item @code{-P@emph{project}}
7883 Use project file @code{project}. Only one such switch can be used.
7887 @c :ref:`gnatmake_and_Project_Files`.
7889 @geindex -q (gnatmake)
7896 Quiet. When this flag is not set, the commands carried out by
7897 @code{gnatmake} are displayed.
7900 @geindex -s (gnatmake)
7907 Recompile if compiler switches have changed since last compilation.
7908 All compiler switches but -I and -o are taken into account in the
7910 orders between different 'first letter' switches are ignored, but
7911 orders between same switches are taken into account. For example,
7912 @code{-O -O2} is different than @code{-O2 -O}, but @code{-g -O}
7913 is equivalent to @code{-O -g}.
7915 This switch is recommended when Integrated Preprocessing is used.
7918 @geindex -u (gnatmake)
7925 Unique. Recompile at most the main files. It implies -c. Combined with
7926 -f, it is equivalent to calling the compiler directly. Note that using
7927 -u with a project file and no main has a special meaning.
7931 @c (See :ref:`Project_Files_and_Main_Subprograms`.)
7933 @geindex -U (gnatmake)
7940 When used without a project file or with one or several mains on the command
7941 line, is equivalent to -u. When used with a project file and no main
7942 on the command line, all sources of all project files are checked and compiled
7943 if not up to date, and libraries are rebuilt, if necessary.
7946 @geindex -v (gnatmake)
7953 Verbose. Display the reason for all recompilations @code{gnatmake}
7954 decides are necessary, with the highest verbosity level.
7957 @geindex -vl (gnatmake)
7964 Verbosity level Low. Display fewer lines than in verbosity Medium.
7967 @geindex -vm (gnatmake)
7974 Verbosity level Medium. Potentially display fewer lines than in verbosity High.
7977 @geindex -vm (gnatmake)
7984 Verbosity level High. Equivalent to -v.
7986 @item @code{-vP@emph{x}}
7988 Indicate the verbosity of the parsing of GNAT project files.
7989 See @ref{de,,Switches Related to Project Files}.
7992 @geindex -x (gnatmake)
7999 Indicate that sources that are not part of any Project File may be compiled.
8000 Normally, when using Project Files, only sources that are part of a Project
8001 File may be compile. When this switch is used, a source outside of all Project
8002 Files may be compiled. The ALI file and the object file will be put in the
8003 object directory of the main Project. The compilation switches used will only
8004 be those specified on the command line. Even when
8005 @code{-x} is used, mains specified on the
8006 command line need to be sources of a project file.
8008 @item @code{-X@emph{name}=@emph{value}}
8010 Indicate that external variable @code{name} has the value @code{value}.
8011 The Project Manager will use this value for occurrences of
8012 @code{external(name)} when parsing the project file.
8013 @ref{de,,Switches Related to Project Files}.
8016 @geindex -z (gnatmake)
8023 No main subprogram. Bind and link the program even if the unit name
8024 given on the command line is a package name. The resulting executable
8025 will execute the elaboration routines of the package and its closure,
8026 then the finalization routines.
8029 @subsubheading GCC switches
8032 Any uppercase or multi-character switch that is not a @code{gnatmake} switch
8033 is passed to @code{gcc} (e.g., @code{-O}, @code{-gnato,} etc.)
8035 @subsubheading Source and library search path switches
8038 @geindex -aI (gnatmake)
8043 @item @code{-aI@emph{dir}}
8045 When looking for source files also look in directory @code{dir}.
8046 The order in which source files search is undertaken is
8047 described in @ref{89,,Search Paths and the Run-Time Library (RTL)}.
8050 @geindex -aL (gnatmake)
8055 @item @code{-aL@emph{dir}}
8057 Consider @code{dir} as being an externally provided Ada library.
8058 Instructs @code{gnatmake} to skip compilation units whose @code{.ALI}
8059 files have been located in directory @code{dir}. This allows you to have
8060 missing bodies for the units in @code{dir} and to ignore out of date bodies
8061 for the same units. You still need to specify
8062 the location of the specs for these units by using the switches
8063 @code{-aI@emph{dir}} or @code{-I@emph{dir}}.
8064 Note: this switch is provided for compatibility with previous versions
8065 of @code{gnatmake}. The easier method of causing standard libraries
8066 to be excluded from consideration is to write-protect the corresponding
8070 @geindex -aO (gnatmake)
8075 @item @code{-aO@emph{dir}}
8077 When searching for library and object files, look in directory
8078 @code{dir}. The order in which library files are searched is described in
8079 @ref{8c,,Search Paths for gnatbind}.
8082 @geindex Search paths
8083 @geindex for gnatmake
8085 @geindex -A (gnatmake)
8090 @item @code{-A@emph{dir}}
8092 Equivalent to @code{-aL@emph{dir}} @code{-aI@emph{dir}}.
8094 @geindex -I (gnatmake)
8096 @item @code{-I@emph{dir}}
8098 Equivalent to @code{-aO@emph{dir} -aI@emph{dir}}.
8101 @geindex -I- (gnatmake)
8103 @geindex Source files
8104 @geindex suppressing search
8111 Do not look for source files in the directory containing the source
8112 file named in the command line.
8113 Do not look for ALI or object files in the directory
8114 where @code{gnatmake} was invoked.
8117 @geindex -L (gnatmake)
8119 @geindex Linker libraries
8124 @item @code{-L@emph{dir}}
8126 Add directory @code{dir} to the list of directories in which the linker
8127 will search for libraries. This is equivalent to
8128 @code{-largs} @code{-L@emph{dir}}.
8129 Furthermore, under Windows, the sources pointed to by the libraries path
8130 set in the registry are not searched for.
8133 @geindex -nostdinc (gnatmake)
8138 @item @code{-nostdinc}
8140 Do not look for source files in the system default directory.
8143 @geindex -nostdlib (gnatmake)
8148 @item @code{-nostdlib}
8150 Do not look for library files in the system default directory.
8153 @geindex --RTS (gnatmake)
8158 @item @code{--RTS=@emph{rts-path}}
8160 Specifies the default location of the run-time library. GNAT looks for the
8162 in the following directories, and stops as soon as a valid run-time is found
8163 (@code{adainclude} or @code{ada_source_path}, and @code{adalib} or
8164 @code{ada_object_path} present):
8170 @emph{<current directory>/$rts_path}
8173 @emph{<default-search-dir>/$rts_path}
8176 @emph{<default-search-dir>/rts-$rts_path}
8179 The selected path is handled like a normal RTS path.
8183 @node Mode Switches for gnatmake,Notes on the Command Line,Switches for gnatmake,Building with gnatmake
8184 @anchor{gnat_ugn/building_executable_programs_with_gnat id4}@anchor{df}@anchor{gnat_ugn/building_executable_programs_with_gnat mode-switches-for-gnatmake}@anchor{e0}
8185 @subsection Mode Switches for @code{gnatmake}
8188 The mode switches (referred to as @code{mode_switches}) allow the
8189 inclusion of switches that are to be passed to the compiler itself, the
8190 binder or the linker. The effect of a mode switch is to cause all
8191 subsequent switches up to the end of the switch list, or up to the next
8192 mode switch, to be interpreted as switches to be passed on to the
8193 designated component of GNAT.
8195 @geindex -cargs (gnatmake)
8200 @item @code{-cargs @emph{switches}}
8202 Compiler switches. Here @code{switches} is a list of switches
8203 that are valid switches for @code{gcc}. They will be passed on to
8204 all compile steps performed by @code{gnatmake}.
8207 @geindex -bargs (gnatmake)
8212 @item @code{-bargs @emph{switches}}
8214 Binder switches. Here @code{switches} is a list of switches
8215 that are valid switches for @code{gnatbind}. They will be passed on to
8216 all bind steps performed by @code{gnatmake}.
8219 @geindex -largs (gnatmake)
8224 @item @code{-largs @emph{switches}}
8226 Linker switches. Here @code{switches} is a list of switches
8227 that are valid switches for @code{gnatlink}. They will be passed on to
8228 all link steps performed by @code{gnatmake}.
8231 @geindex -margs (gnatmake)
8236 @item @code{-margs @emph{switches}}
8238 Make switches. The switches are directly interpreted by @code{gnatmake},
8239 regardless of any previous occurrence of @code{-cargs}, @code{-bargs}
8243 @node Notes on the Command Line,How gnatmake Works,Mode Switches for gnatmake,Building with gnatmake
8244 @anchor{gnat_ugn/building_executable_programs_with_gnat id5}@anchor{e1}@anchor{gnat_ugn/building_executable_programs_with_gnat notes-on-the-command-line}@anchor{e2}
8245 @subsection Notes on the Command Line
8248 This section contains some additional useful notes on the operation
8249 of the @code{gnatmake} command.
8251 @geindex Recompilation (by gnatmake)
8257 If @code{gnatmake} finds no ALI files, it recompiles the main program
8258 and all other units required by the main program.
8259 This means that @code{gnatmake}
8260 can be used for the initial compile, as well as during subsequent steps of
8261 the development cycle.
8264 If you enter @code{gnatmake foo.adb}, where @code{foo}
8265 is a subunit or body of a generic unit, @code{gnatmake} recompiles
8266 @code{foo.adb} (because it finds no ALI) and stops, issuing a
8270 In @code{gnatmake} the switch @code{-I}
8271 is used to specify both source and
8272 library file paths. Use @code{-aI}
8273 instead if you just want to specify
8274 source paths only and @code{-aO}
8275 if you want to specify library paths
8279 @code{gnatmake} will ignore any files whose ALI file is write-protected.
8280 This may conveniently be used to exclude standard libraries from
8281 consideration and in particular it means that the use of the
8282 @code{-f} switch will not recompile these files
8283 unless @code{-a} is also specified.
8286 @code{gnatmake} has been designed to make the use of Ada libraries
8287 particularly convenient. Assume you have an Ada library organized
8288 as follows: @emph{obj-dir} contains the objects and ALI files for
8289 of your Ada compilation units,
8290 whereas @emph{include-dir} contains the
8291 specs of these units, but no bodies. Then to compile a unit
8292 stored in @code{main.adb}, which uses this Ada library you would just type:
8295 $ gnatmake -aI`include-dir` -aL`obj-dir` main
8299 Using @code{gnatmake} along with the @code{-m (minimal recompilation)}
8300 switch provides a mechanism for avoiding unnecessary recompilations. Using
8302 you can update the comments/format of your
8303 source files without having to recompile everything. Note, however, that
8304 adding or deleting lines in a source files may render its debugging
8305 info obsolete. If the file in question is a spec, the impact is rather
8306 limited, as that debugging info will only be useful during the
8307 elaboration phase of your program. For bodies the impact can be more
8308 significant. In all events, your debugger will warn you if a source file
8309 is more recent than the corresponding object, and alert you to the fact
8310 that the debugging information may be out of date.
8313 @node How gnatmake Works,Examples of gnatmake Usage,Notes on the Command Line,Building with gnatmake
8314 @anchor{gnat_ugn/building_executable_programs_with_gnat id6}@anchor{e3}@anchor{gnat_ugn/building_executable_programs_with_gnat how-gnatmake-works}@anchor{e4}
8315 @subsection How @code{gnatmake} Works
8318 Generally @code{gnatmake} automatically performs all necessary
8319 recompilations and you don't need to worry about how it works. However,
8320 it may be useful to have some basic understanding of the @code{gnatmake}
8321 approach and in particular to understand how it uses the results of
8322 previous compilations without incorrectly depending on them.
8324 First a definition: an object file is considered @emph{up to date} if the
8325 corresponding ALI file exists and if all the source files listed in the
8326 dependency section of this ALI file have time stamps matching those in
8327 the ALI file. This means that neither the source file itself nor any
8328 files that it depends on have been modified, and hence there is no need
8329 to recompile this file.
8331 @code{gnatmake} works by first checking if the specified main unit is up
8332 to date. If so, no compilations are required for the main unit. If not,
8333 @code{gnatmake} compiles the main program to build a new ALI file that
8334 reflects the latest sources. Then the ALI file of the main unit is
8335 examined to find all the source files on which the main program depends,
8336 and @code{gnatmake} recursively applies the above procedure on all these
8339 This process ensures that @code{gnatmake} only trusts the dependencies
8340 in an existing ALI file if they are known to be correct. Otherwise it
8341 always recompiles to determine a new, guaranteed accurate set of
8342 dependencies. As a result the program is compiled 'upside down' from what may
8343 be more familiar as the required order of compilation in some other Ada
8344 systems. In particular, clients are compiled before the units on which
8345 they depend. The ability of GNAT to compile in any order is critical in
8346 allowing an order of compilation to be chosen that guarantees that
8347 @code{gnatmake} will recompute a correct set of new dependencies if
8350 When invoking @code{gnatmake} with several @code{file_names}, if a unit is
8351 imported by several of the executables, it will be recompiled at most once.
8353 Note: when using non-standard naming conventions
8354 (@ref{35,,Using Other File Names}), changing through a configuration pragmas
8355 file the version of a source and invoking @code{gnatmake} to recompile may
8356 have no effect, if the previous version of the source is still accessible
8357 by @code{gnatmake}. It may be necessary to use the switch
8360 @node Examples of gnatmake Usage,,How gnatmake Works,Building with gnatmake
8361 @anchor{gnat_ugn/building_executable_programs_with_gnat examples-of-gnatmake-usage}@anchor{e5}@anchor{gnat_ugn/building_executable_programs_with_gnat id7}@anchor{e6}
8362 @subsection Examples of @code{gnatmake} Usage
8368 @item @emph{gnatmake hello.adb}
8370 Compile all files necessary to bind and link the main program
8371 @code{hello.adb} (containing unit @code{Hello}) and bind and link the
8372 resulting object files to generate an executable file @code{hello}.
8374 @item @emph{gnatmake main1 main2 main3}
8376 Compile all files necessary to bind and link the main programs
8377 @code{main1.adb} (containing unit @code{Main1}), @code{main2.adb}
8378 (containing unit @code{Main2}) and @code{main3.adb}
8379 (containing unit @code{Main3}) and bind and link the resulting object files
8380 to generate three executable files @code{main1},
8381 @code{main2} and @code{main3}.
8383 @item @emph{gnatmake -q Main_Unit -cargs -O2 -bargs -l}
8385 Compile all files necessary to bind and link the main program unit
8386 @code{Main_Unit} (from file @code{main_unit.adb}). All compilations will
8387 be done with optimization level 2 and the order of elaboration will be
8388 listed by the binder. @code{gnatmake} will operate in quiet mode, not
8389 displaying commands it is executing.
8392 @node Compiling with gcc,Compiler Switches,Building with gnatmake,Building Executable Programs with GNAT
8393 @anchor{gnat_ugn/building_executable_programs_with_gnat compiling-with-gcc}@anchor{1c}@anchor{gnat_ugn/building_executable_programs_with_gnat id8}@anchor{e7}
8394 @section Compiling with @code{gcc}
8397 This section discusses how to compile Ada programs using the @code{gcc}
8398 command. It also describes the set of switches
8399 that can be used to control the behavior of the compiler.
8402 * Compiling Programs::
8403 * Search Paths and the Run-Time Library (RTL): Search Paths and the Run-Time Library RTL.
8404 * Order of Compilation Issues::
8409 @node Compiling Programs,Search Paths and the Run-Time Library RTL,,Compiling with gcc
8410 @anchor{gnat_ugn/building_executable_programs_with_gnat compiling-programs}@anchor{e8}@anchor{gnat_ugn/building_executable_programs_with_gnat id9}@anchor{e9}
8411 @subsection Compiling Programs
8414 The first step in creating an executable program is to compile the units
8415 of the program using the @code{gcc} command. You must compile the
8422 the body file (@code{.adb}) for a library level subprogram or generic
8426 the spec file (@code{.ads}) for a library level package or generic
8427 package that has no body
8430 the body file (@code{.adb}) for a library level package
8431 or generic package that has a body
8434 You need @emph{not} compile the following files
8440 the spec of a library unit which has a body
8446 because they are compiled as part of compiling related units. GNAT
8448 when the corresponding body is compiled, and subunits when the parent is
8451 @geindex cannot generate code
8453 If you attempt to compile any of these files, you will get one of the
8454 following error messages (where @code{fff} is the name of the file you
8460 cannot generate code for file `@w{`}fff`@w{`} (package spec)
8461 to check package spec, use -gnatc
8463 cannot generate code for file `@w{`}fff`@w{`} (missing subunits)
8464 to check parent unit, use -gnatc
8466 cannot generate code for file `@w{`}fff`@w{`} (subprogram spec)
8467 to check subprogram spec, use -gnatc
8469 cannot generate code for file `@w{`}fff`@w{`} (subunit)
8470 to check subunit, use -gnatc
8474 As indicated by the above error messages, if you want to submit
8475 one of these files to the compiler to check for correct semantics
8476 without generating code, then use the @code{-gnatc} switch.
8478 The basic command for compiling a file containing an Ada unit is:
8481 $ gcc -c [switches] <file name>
8484 where @code{file name} is the name of the Ada file (usually
8485 having an extension @code{.ads} for a spec or @code{.adb} for a body).
8487 @code{-c} switch to tell @code{gcc} to compile, but not link, the file.
8488 The result of a successful compilation is an object file, which has the
8489 same name as the source file but an extension of @code{.o} and an Ada
8490 Library Information (ALI) file, which also has the same name as the
8491 source file, but with @code{.ali} as the extension. GNAT creates these
8492 two output files in the current directory, but you may specify a source
8493 file in any directory using an absolute or relative path specification
8494 containing the directory information.
8496 TESTING: the @code{--foobar@emph{NN}} switch
8500 @code{gcc} is actually a driver program that looks at the extensions of
8501 the file arguments and loads the appropriate compiler. For example, the
8502 GNU C compiler is @code{cc1}, and the Ada compiler is @code{gnat1}.
8503 These programs are in directories known to the driver program (in some
8504 configurations via environment variables you set), but need not be in
8505 your path. The @code{gcc} driver also calls the assembler and any other
8506 utilities needed to complete the generation of the required object
8509 It is possible to supply several file names on the same @code{gcc}
8510 command. This causes @code{gcc} to call the appropriate compiler for
8511 each file. For example, the following command lists two separate
8512 files to be compiled:
8515 $ gcc -c x.adb y.adb
8518 calls @code{gnat1} (the Ada compiler) twice to compile @code{x.adb} and
8520 The compiler generates two object files @code{x.o} and @code{y.o}
8521 and the two ALI files @code{x.ali} and @code{y.ali}.
8523 Any switches apply to all the files listed, see @ref{ea,,Compiler Switches} for a
8524 list of available @code{gcc} switches.
8526 @node Search Paths and the Run-Time Library RTL,Order of Compilation Issues,Compiling Programs,Compiling with gcc
8527 @anchor{gnat_ugn/building_executable_programs_with_gnat id10}@anchor{eb}@anchor{gnat_ugn/building_executable_programs_with_gnat search-paths-and-the-run-time-library-rtl}@anchor{89}
8528 @subsection Search Paths and the Run-Time Library (RTL)
8531 With the GNAT source-based library system, the compiler must be able to
8532 find source files for units that are needed by the unit being compiled.
8533 Search paths are used to guide this process.
8535 The compiler compiles one source file whose name must be given
8536 explicitly on the command line. In other words, no searching is done
8537 for this file. To find all other source files that are needed (the most
8538 common being the specs of units), the compiler examines the following
8539 directories, in the following order:
8545 The directory containing the source file of the main unit being compiled
8546 (the file name on the command line).
8549 Each directory named by an @code{-I} switch given on the @code{gcc}
8550 command line, in the order given.
8552 @geindex ADA_PRJ_INCLUDE_FILE
8555 Each of the directories listed in the text file whose name is given
8557 @geindex ADA_PRJ_INCLUDE_FILE
8558 @geindex environment variable; ADA_PRJ_INCLUDE_FILE
8559 @code{ADA_PRJ_INCLUDE_FILE} environment variable.
8560 @geindex ADA_PRJ_INCLUDE_FILE
8561 @geindex environment variable; ADA_PRJ_INCLUDE_FILE
8562 @code{ADA_PRJ_INCLUDE_FILE} is normally set by gnatmake or by the gnat
8563 driver when project files are used. It should not normally be set
8566 @geindex ADA_INCLUDE_PATH
8569 Each of the directories listed in the value of the
8570 @geindex ADA_INCLUDE_PATH
8571 @geindex environment variable; ADA_INCLUDE_PATH
8572 @code{ADA_INCLUDE_PATH} environment variable.
8573 Construct this value
8576 @geindex environment variable; PATH
8577 @code{PATH} environment variable: a list of directory
8578 names separated by colons (semicolons when working with the NT version).
8581 The content of the @code{ada_source_path} file which is part of the GNAT
8582 installation tree and is used to store standard libraries such as the
8583 GNAT Run Time Library (RTL) source files.
8584 @ref{87,,Installing a library}
8587 Specifying the switch @code{-I-}
8588 inhibits the use of the directory
8589 containing the source file named in the command line. You can still
8590 have this directory on your search path, but in this case it must be
8591 explicitly requested with a @code{-I} switch.
8593 Specifying the switch @code{-nostdinc}
8594 inhibits the search of the default location for the GNAT Run Time
8595 Library (RTL) source files.
8597 The compiler outputs its object files and ALI files in the current
8599 Caution: The object file can be redirected with the @code{-o} switch;
8600 however, @code{gcc} and @code{gnat1} have not been coordinated on this
8601 so the @code{ALI} file will not go to the right place. Therefore, you should
8602 avoid using the @code{-o} switch.
8606 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
8607 children make up the GNAT RTL, together with the simple @code{System.IO}
8608 package used in the @code{"Hello World"} example. The sources for these units
8609 are needed by the compiler and are kept together in one directory. Not
8610 all of the bodies are needed, but all of the sources are kept together
8611 anyway. In a normal installation, you need not specify these directory
8612 names when compiling or binding. Either the environment variables or
8613 the built-in defaults cause these files to be found.
8615 In addition to the language-defined hierarchies (@code{System}, @code{Ada} and
8616 @code{Interfaces}), the GNAT distribution provides a fourth hierarchy,
8617 consisting of child units of @code{GNAT}. This is a collection of generally
8618 useful types, subprograms, etc. See the @cite{GNAT_Reference_Manual}
8619 for further details.
8621 Besides simplifying access to the RTL, a major use of search paths is
8622 in compiling sources from multiple directories. This can make
8623 development environments much more flexible.
8625 @node Order of Compilation Issues,Examples,Search Paths and the Run-Time Library RTL,Compiling with gcc
8626 @anchor{gnat_ugn/building_executable_programs_with_gnat id11}@anchor{ec}@anchor{gnat_ugn/building_executable_programs_with_gnat order-of-compilation-issues}@anchor{ed}
8627 @subsection Order of Compilation Issues
8630 If, in our earlier example, there was a spec for the @code{hello}
8631 procedure, it would be contained in the file @code{hello.ads}; yet this
8632 file would not have to be explicitly compiled. This is the result of the
8633 model we chose to implement library management. Some of the consequences
8634 of this model are as follows:
8640 There is no point in compiling specs (except for package
8641 specs with no bodies) because these are compiled as needed by clients. If
8642 you attempt a useless compilation, you will receive an error message.
8643 It is also useless to compile subunits because they are compiled as needed
8647 There are no order of compilation requirements: performing a
8648 compilation never obsoletes anything. The only way you can obsolete
8649 something and require recompilations is to modify one of the
8650 source files on which it depends.
8653 There is no library as such, apart from the ALI files
8654 (@ref{42,,The Ada Library Information Files}, for information on the format
8655 of these files). For now we find it convenient to create separate ALI files,
8656 but eventually the information therein may be incorporated into the object
8660 When you compile a unit, the source files for the specs of all units
8661 that it @emph{with}s, all its subunits, and the bodies of any generics it
8662 instantiates must be available (reachable by the search-paths mechanism
8663 described above), or you will receive a fatal error message.
8666 @node Examples,,Order of Compilation Issues,Compiling with gcc
8667 @anchor{gnat_ugn/building_executable_programs_with_gnat id12}@anchor{ee}@anchor{gnat_ugn/building_executable_programs_with_gnat examples}@anchor{ef}
8668 @subsection Examples
8671 The following are some typical Ada compilation command line examples:
8677 Compile body in file @code{xyz.adb} with all default options.
8680 $ gcc -c -O2 -gnata xyz-def.adb
8683 Compile the child unit package in file @code{xyz-def.adb} with extensive
8684 optimizations, and pragma @code{Assert}/@cite{Debug} statements
8688 $ gcc -c -gnatc abc-def.adb
8691 Compile the subunit in file @code{abc-def.adb} in semantic-checking-only
8694 @node Compiler Switches,Linker Switches,Compiling with gcc,Building Executable Programs with GNAT
8695 @anchor{gnat_ugn/building_executable_programs_with_gnat compiler-switches}@anchor{f0}@anchor{gnat_ugn/building_executable_programs_with_gnat switches-for-gcc}@anchor{ea}
8696 @section Compiler Switches
8699 The @code{gcc} command accepts switches that control the
8700 compilation process. These switches are fully described in this section:
8701 first an alphabetical listing of all switches with a brief description,
8702 and then functionally grouped sets of switches with more detailed
8705 More switches exist for GCC than those documented here, especially
8706 for specific targets. However, their use is not recommended as
8707 they may change code generation in ways that are incompatible with
8708 the Ada run-time library, or can cause inconsistencies between
8712 * Alphabetical List of All Switches::
8713 * Output and Error Message Control::
8714 * Warning Message Control::
8715 * Debugging and Assertion Control::
8716 * Validity Checking::
8719 * Using gcc for Syntax Checking::
8720 * Using gcc for Semantic Checking::
8721 * Compiling Different Versions of Ada::
8722 * Character Set Control::
8723 * File Naming Control::
8724 * Subprogram Inlining Control::
8725 * Auxiliary Output Control::
8726 * Debugging Control::
8727 * Exception Handling Control::
8728 * Units to Sources Mapping Files::
8729 * Code Generation Control::
8733 @node Alphabetical List of All Switches,Output and Error Message Control,,Compiler Switches
8734 @anchor{gnat_ugn/building_executable_programs_with_gnat id13}@anchor{f1}@anchor{gnat_ugn/building_executable_programs_with_gnat alphabetical-list-of-all-switches}@anchor{f2}
8735 @subsection Alphabetical List of All Switches
8743 @item @code{-b @emph{target}}
8745 Compile your program to run on @code{target}, which is the name of a
8746 system configuration. You must have a GNAT cross-compiler built if
8747 @code{target} is not the same as your host system.
8755 @item @code{-B@emph{dir}}
8757 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
8758 from @code{dir} instead of the default location. Only use this switch
8759 when multiple versions of the GNAT compiler are available.
8760 See the "Options for Directory Search" section in the
8761 @cite{Using the GNU Compiler Collection (GCC)} manual for further details.
8762 You would normally use the @code{-b} or @code{-V} switch instead.
8772 Compile. Always use this switch when compiling Ada programs.
8774 Note: for some other languages when using @code{gcc}, notably in
8775 the case of C and C++, it is possible to use
8776 use @code{gcc} without a @code{-c} switch to
8777 compile and link in one step. In the case of GNAT, you
8778 cannot use this approach, because the binder must be run
8779 and @code{gcc} cannot be used to run the GNAT binder.
8782 @geindex -fcallgraph-info (gcc)
8787 @item @code{-fcallgraph-info[=su,da]}
8789 Makes the compiler output callgraph information for the program, on a
8790 per-file basis. The information is generated in the VCG format. It can
8791 be decorated with additional, per-node and/or per-edge information, if a
8792 list of comma-separated markers is additionally specified. When the
8793 @code{su} marker is specified, the callgraph is decorated with stack usage
8794 information; it is equivalent to @code{-fstack-usage}. When the @code{da}
8795 marker is specified, the callgraph is decorated with information about
8796 dynamically allocated objects.
8799 @geindex -fdump-scos (gcc)
8804 @item @code{-fdump-scos}
8806 Generates SCO (Source Coverage Obligation) information in the ALI file.
8807 This information is used by advanced coverage tools. See unit @code{SCOs}
8808 in the compiler sources for details in files @code{scos.ads} and
8812 @geindex -fgnat-encodings (gcc)
8817 @item @code{-fgnat-encodings=[all|gdb|minimal]}
8819 This switch controls the balance between GNAT encodings and standard DWARF
8820 emitted in the debug information.
8823 @geindex -flto (gcc)
8828 @item @code{-flto[=@emph{n}]}
8830 Enables Link Time Optimization. This switch must be used in conjunction
8831 with the @code{-Ox} switches (but not with the @code{-gnatn} switch
8832 since it is a full replacement for the latter) and instructs the compiler
8833 to defer most optimizations until the link stage. The advantage of this
8834 approach is that the compiler can do a whole-program analysis and choose
8835 the best interprocedural optimization strategy based on a complete view
8836 of the program, instead of a fragmentary view with the usual approach.
8837 This can also speed up the compilation of big programs and reduce the
8838 size of the executable, compared with a traditional per-unit compilation
8839 with inlining across units enabled by the @code{-gnatn} switch.
8840 The drawback of this approach is that it may require more memory and that
8841 the debugging information generated by -g with it might be hardly usable.
8842 The switch, as well as the accompanying @code{-Ox} switches, must be
8843 specified both for the compilation and the link phases.
8844 If the @code{n} parameter is specified, the optimization and final code
8845 generation at link time are executed using @code{n} parallel jobs by
8846 means of an installed @code{make} program.
8849 @geindex -fno-inline (gcc)
8854 @item @code{-fno-inline}
8856 Suppresses all inlining, unless requested with pragma @code{Inline_Always}. The
8857 effect is enforced regardless of other optimization or inlining switches.
8858 Note that inlining can also be suppressed on a finer-grained basis with
8859 pragma @code{No_Inline}.
8862 @geindex -fno-inline-functions (gcc)
8867 @item @code{-fno-inline-functions}
8869 Suppresses automatic inlining of subprograms, which is enabled
8870 if @code{-O3} is used.
8873 @geindex -fno-inline-small-functions (gcc)
8878 @item @code{-fno-inline-small-functions}
8880 Suppresses automatic inlining of small subprograms, which is enabled
8881 if @code{-O2} is used.
8884 @geindex -fno-inline-functions-called-once (gcc)
8889 @item @code{-fno-inline-functions-called-once}
8891 Suppresses inlining of subprograms local to the unit and called once
8892 from within it, which is enabled if @code{-O1} is used.
8895 @geindex -fno-ivopts (gcc)
8900 @item @code{-fno-ivopts}
8902 Suppresses high-level loop induction variable optimizations, which are
8903 enabled if @code{-O1} is used. These optimizations are generally
8904 profitable but, for some specific cases of loops with numerous uses
8905 of the iteration variable that follow a common pattern, they may end
8906 up destroying the regularity that could be exploited at a lower level
8907 and thus producing inferior code.
8910 @geindex -fno-strict-aliasing (gcc)
8915 @item @code{-fno-strict-aliasing}
8917 Causes the compiler to avoid assumptions regarding non-aliasing
8918 of objects of different types. See
8919 @ref{f3,,Optimization and Strict Aliasing} for details.
8922 @geindex -fno-strict-overflow (gcc)
8927 @item @code{-fno-strict-overflow}
8929 Causes the compiler to avoid assumptions regarding the rules of signed
8930 integer overflow. These rules specify that signed integer overflow will
8931 result in a Constraint_Error exception at run time and are enforced in
8932 default mode by the compiler, so this switch should not be necessary in
8933 normal operating mode. It might be useful in conjunction with @code{-gnato0}
8934 for very peculiar cases of low-level programming.
8937 @geindex -fstack-check (gcc)
8942 @item @code{-fstack-check}
8944 Activates stack checking.
8945 See @ref{f4,,Stack Overflow Checking} for details.
8948 @geindex -fstack-usage (gcc)
8953 @item @code{-fstack-usage}
8955 Makes the compiler output stack usage information for the program, on a
8956 per-subprogram basis. See @ref{f5,,Static Stack Usage Analysis} for details.
8966 Generate debugging information. This information is stored in the object
8967 file and copied from there to the final executable file by the linker,
8968 where it can be read by the debugger. You must use the
8969 @code{-g} switch if you plan on using the debugger.
8972 @geindex -gnat05 (gcc)
8977 @item @code{-gnat05}
8979 Allow full Ada 2005 features.
8982 @geindex -gnat12 (gcc)
8987 @item @code{-gnat12}
8989 Allow full Ada 2012 features.
8992 @geindex -gnat83 (gcc)
8994 @geindex -gnat2005 (gcc)
8999 @item @code{-gnat2005}
9001 Allow full Ada 2005 features (same as @code{-gnat05})
9004 @geindex -gnat2012 (gcc)
9009 @item @code{-gnat2012}
9011 Allow full Ada 2012 features (same as @code{-gnat12})
9013 @item @code{-gnat83}
9015 Enforce Ada 83 restrictions.
9018 @geindex -gnat95 (gcc)
9023 @item @code{-gnat95}
9025 Enforce Ada 95 restrictions.
9027 Note: for compatibility with some Ada 95 compilers which support only
9028 the @code{overriding} keyword of Ada 2005, the @code{-gnatd.D} switch can
9029 be used along with @code{-gnat95} to achieve a similar effect with GNAT.
9031 @code{-gnatd.D} instructs GNAT to consider @code{overriding} as a keyword
9032 and handle its associated semantic checks, even in Ada 95 mode.
9035 @geindex -gnata (gcc)
9042 Assertions enabled. @code{Pragma Assert} and @code{pragma Debug} to be
9043 activated. Note that these pragmas can also be controlled using the
9044 configuration pragmas @code{Assertion_Policy} and @code{Debug_Policy}.
9045 It also activates pragmas @code{Check}, @code{Precondition}, and
9046 @code{Postcondition}. Note that these pragmas can also be controlled
9047 using the configuration pragma @code{Check_Policy}. In Ada 2012, it
9048 also activates all assertions defined in the RM as aspects: preconditions,
9049 postconditions, type invariants and (sub)type predicates. In all Ada modes,
9050 corresponding pragmas for type invariants and (sub)type predicates are
9051 also activated. The default is that all these assertions are disabled,
9052 and have no effect, other than being checked for syntactic validity, and
9053 in the case of subtype predicates, constructions such as membership tests
9054 still test predicates even if assertions are turned off.
9057 @geindex -gnatA (gcc)
9064 Avoid processing @code{gnat.adc}. If a @code{gnat.adc} file is present,
9068 @geindex -gnatb (gcc)
9075 Generate brief messages to @code{stderr} even if verbose mode set.
9078 @geindex -gnatB (gcc)
9085 Assume no invalid (bad) values except for 'Valid attribute use
9086 (@ref{f6,,Validity Checking}).
9089 @geindex -gnatc (gcc)
9096 Check syntax and semantics only (no code generation attempted). When the
9097 compiler is invoked by @code{gnatmake}, if the switch @code{-gnatc} is
9098 only given to the compiler (after @code{-cargs} or in package Compiler of
9099 the project file, @code{gnatmake} will fail because it will not find the
9100 object file after compilation. If @code{gnatmake} is called with
9101 @code{-gnatc} as a builder switch (before @code{-cargs} or in package
9102 Builder of the project file) then @code{gnatmake} will not fail because
9103 it will not look for the object files after compilation, and it will not try
9107 @geindex -gnatC (gcc)
9114 Generate CodePeer intermediate format (no code generation attempted).
9115 This switch will generate an intermediate representation suitable for
9116 use by CodePeer (@code{.scil} files). This switch is not compatible with
9117 code generation (it will, among other things, disable some switches such
9118 as -gnatn, and enable others such as -gnata).
9121 @geindex -gnatd (gcc)
9128 Specify debug options for the compiler. The string of characters after
9129 the @code{-gnatd} specify the specific debug options. The possible
9130 characters are 0-9, a-z, A-Z, optionally preceded by a dot. See
9131 compiler source file @code{debug.adb} for details of the implemented
9132 debug options. Certain debug options are relevant to applications
9133 programmers, and these are documented at appropriate points in this
9137 @geindex -gnatD[nn] (gcc)
9144 Create expanded source files for source level debugging. This switch
9145 also suppresses generation of cross-reference information
9146 (see @code{-gnatx}). Note that this switch is not allowed if a previous
9147 -gnatR switch has been given, since these two switches are not compatible.
9150 @geindex -gnateA (gcc)
9155 @item @code{-gnateA}
9157 Check that the actual parameters of a subprogram call are not aliases of one
9158 another. To qualify as aliasing, the actuals must denote objects of a composite
9159 type, their memory locations must be identical or overlapping, and at least one
9160 of the corresponding formal parameters must be of mode OUT or IN OUT.
9163 type Rec_Typ is record
9164 Data : Integer := 0;
9167 function Self (Val : Rec_Typ) return Rec_Typ is
9172 procedure Detect_Aliasing (Val_1 : in out Rec_Typ; Val_2 : Rec_Typ) is
9175 end Detect_Aliasing;
9179 Detect_Aliasing (Obj, Obj);
9180 Detect_Aliasing (Obj, Self (Obj));
9183 In the example above, the first call to @code{Detect_Aliasing} fails with a
9184 @code{Program_Error} at run time because the actuals for @code{Val_1} and
9185 @code{Val_2} denote the same object. The second call executes without raising
9186 an exception because @code{Self(Obj)} produces an anonymous object which does
9187 not share the memory location of @code{Obj}.
9190 @geindex -gnatec (gcc)
9195 @item @code{-gnatec=@emph{path}}
9197 Specify a configuration pragma file
9198 (the equal sign is optional)
9199 (@ref{79,,The Configuration Pragmas Files}).
9202 @geindex -gnateC (gcc)
9207 @item @code{-gnateC}
9209 Generate CodePeer messages in a compiler-like format. This switch is only
9210 effective if @code{-gnatcC} is also specified and requires an installation
9214 @geindex -gnated (gcc)
9219 @item @code{-gnated}
9221 Disable atomic synchronization
9224 @geindex -gnateD (gcc)
9229 @item @code{-gnateDsymbol[=@emph{value}]}
9231 Defines a symbol, associated with @code{value}, for preprocessing.
9232 (@ref{18,,Integrated Preprocessing}).
9235 @geindex -gnateE (gcc)
9240 @item @code{-gnateE}
9242 Generate extra information in exception messages. In particular, display
9243 extra column information and the value and range associated with index and
9244 range check failures, and extra column information for access checks.
9245 In cases where the compiler is able to determine at compile time that
9246 a check will fail, it gives a warning, and the extra information is not
9247 produced at run time.
9250 @geindex -gnatef (gcc)
9255 @item @code{-gnatef}
9257 Display full source path name in brief error messages.
9260 @geindex -gnateF (gcc)
9265 @item @code{-gnateF}
9267 Check for overflow on all floating-point operations, including those
9268 for unconstrained predefined types. See description of pragma
9269 @code{Check_Float_Overflow} in GNAT RM.
9272 @geindex -gnateg (gcc)
9279 The @code{-gnatc} switch must always be specified before this switch, e.g.
9280 @code{-gnatceg}. Generate a C header from the Ada input file. See
9281 @ref{ca,,Generating C Headers for Ada Specifications} for more
9285 @geindex -gnateG (gcc)
9290 @item @code{-gnateG}
9292 Save result of preprocessing in a text file.
9295 @geindex -gnatei (gcc)
9300 @item @code{-gnatei@emph{nnn}}
9302 Set maximum number of instantiations during compilation of a single unit to
9303 @code{nnn}. This may be useful in increasing the default maximum of 8000 for
9304 the rare case when a single unit legitimately exceeds this limit.
9307 @geindex -gnateI (gcc)
9312 @item @code{-gnateI@emph{nnn}}
9314 Indicates that the source is a multi-unit source and that the index of the
9315 unit to compile is @code{nnn}. @code{nnn} needs to be a positive number and need
9316 to be a valid index in the multi-unit source.
9319 @geindex -gnatel (gcc)
9324 @item @code{-gnatel}
9326 This switch can be used with the static elaboration model to issue info
9328 where implicit @code{pragma Elaborate} and @code{pragma Elaborate_All}
9329 are generated. This is useful in diagnosing elaboration circularities
9330 caused by these implicit pragmas when using the static elaboration
9331 model. See See the section in this guide on elaboration checking for
9332 further details. These messages are not generated by default, and are
9333 intended only for temporary use when debugging circularity problems.
9336 @geindex -gnatel (gcc)
9341 @item @code{-gnateL}
9343 This switch turns off the info messages about implicit elaboration pragmas.
9346 @geindex -gnatem (gcc)
9351 @item @code{-gnatem=@emph{path}}
9353 Specify a mapping file
9354 (the equal sign is optional)
9355 (@ref{f7,,Units to Sources Mapping Files}).
9358 @geindex -gnatep (gcc)
9363 @item @code{-gnatep=@emph{file}}
9365 Specify a preprocessing data file
9366 (the equal sign is optional)
9367 (@ref{18,,Integrated Preprocessing}).
9370 @geindex -gnateP (gcc)
9375 @item @code{-gnateP}
9377 Turn categorization dependency errors into warnings.
9378 Ada requires that units that WITH one another have compatible categories, for
9379 example a Pure unit cannot WITH a Preelaborate unit. If this switch is used,
9380 these errors become warnings (which can be ignored, or suppressed in the usual
9381 manner). This can be useful in some specialized circumstances such as the
9382 temporary use of special test software.
9385 @geindex -gnateS (gcc)
9390 @item @code{-gnateS}
9392 Synonym of @code{-fdump-scos}, kept for backwards compatibility.
9395 @geindex -gnatet=file (gcc)
9400 @item @code{-gnatet=@emph{path}}
9402 Generate target dependent information. The format of the output file is
9403 described in the section about switch @code{-gnateT}.
9406 @geindex -gnateT (gcc)
9411 @item @code{-gnateT=@emph{path}}
9413 Read target dependent information, such as endianness or sizes and alignments
9414 of base type. If this switch is passed, the default target dependent
9415 information of the compiler is replaced by the one read from the input file.
9416 This is used by tools other than the compiler, e.g. to do
9417 semantic analysis of programs that will run on some other target than
9418 the machine on which the tool is run.
9420 The following target dependent values should be defined,
9421 where @code{Nat} denotes a natural integer value, @code{Pos} denotes a
9422 positive integer value, and fields marked with a question mark are
9423 boolean fields, where a value of 0 is False, and a value of 1 is True:
9426 Bits_BE : Nat; -- Bits stored big-endian?
9427 Bits_Per_Unit : Pos; -- Bits in a storage unit
9428 Bits_Per_Word : Pos; -- Bits in a word
9429 Bytes_BE : Nat; -- Bytes stored big-endian?
9430 Char_Size : Pos; -- Standard.Character'Size
9431 Double_Float_Alignment : Nat; -- Alignment of double float
9432 Double_Scalar_Alignment : Nat; -- Alignment of double length scalar
9433 Double_Size : Pos; -- Standard.Long_Float'Size
9434 Float_Size : Pos; -- Standard.Float'Size
9435 Float_Words_BE : Nat; -- Float words stored big-endian?
9436 Int_Size : Pos; -- Standard.Integer'Size
9437 Long_Double_Size : Pos; -- Standard.Long_Long_Float'Size
9438 Long_Long_Size : Pos; -- Standard.Long_Long_Integer'Size
9439 Long_Size : Pos; -- Standard.Long_Integer'Size
9440 Maximum_Alignment : Pos; -- Maximum permitted alignment
9441 Max_Unaligned_Field : Pos; -- Maximum size for unaligned bit field
9442 Pointer_Size : Pos; -- System.Address'Size
9443 Short_Enums : Nat; -- Foreign enums use short size?
9444 Short_Size : Pos; -- Standard.Short_Integer'Size
9445 Strict_Alignment : Nat; -- Strict alignment?
9446 System_Allocator_Alignment : Nat; -- Alignment for malloc calls
9447 Wchar_T_Size : Pos; -- Interfaces.C.wchar_t'Size
9448 Words_BE : Nat; -- Words stored big-endian?
9451 @code{Bits_Per_Unit} is the number of bits in a storage unit, the equivalent of
9452 GCC macro @code{BITS_PER_UNIT} documented as follows: @cite{Define this macro to be the number of bits in an addressable storage unit (byte); normally 8.}
9454 @code{Bits_Per_Word} is the number of bits in a machine word, the equivalent of
9455 GCC macro @code{BITS_PER_WORD} documented as follows: @cite{Number of bits in a word; normally 32.}
9457 @code{Double_Scalar_Alignment} is the alignment for a scalar whose size is two
9458 machine words. It should be the same as the alignment for C @code{long_long} on
9461 @code{Maximum_Alignment} is the maximum alignment that the compiler might choose
9462 by default for a type or object, which is also the maximum alignment that can
9463 be specified in GNAT. It is computed for GCC backends as @code{BIGGEST_ALIGNMENT
9464 / BITS_PER_UNIT} where GCC macro @code{BIGGEST_ALIGNMENT} is documented as
9465 follows: @cite{Biggest alignment that any data type can require on this machine@comma{} in bits.}
9467 @code{Max_Unaligned_Field} is the maximum size for unaligned bit field, which is
9468 64 for the majority of GCC targets (but can be different on some targets like
9471 @code{Strict_Alignment} is the equivalent of GCC macro @code{STRICT_ALIGNMENT}
9472 documented as follows: @cite{Define this macro to be the value 1 if instructions will fail to work if given data not on the nominal alignment. If instructions will merely go slower in that case@comma{} define this macro as 0.}
9474 @code{System_Allocator_Alignment} is the guaranteed alignment of data returned
9475 by calls to @code{malloc}.
9477 The format of the input file is as follows. First come the values of
9478 the variables defined above, with one line per value:
9484 where @code{name} is the name of the parameter, spelled out in full,
9485 and cased as in the above list, and @code{value} is an unsigned decimal
9486 integer. Two or more blanks separates the name from the value.
9488 All the variables must be present, in alphabetical order (i.e. the
9489 same order as the list above).
9491 Then there is a blank line to separate the two parts of the file. Then
9492 come the lines showing the floating-point types to be registered, with
9493 one line per registered mode:
9496 name digs float_rep size alignment
9499 where @code{name} is the string name of the type (which can have
9500 single spaces embedded in the name (e.g. long double), @code{digs} is
9501 the number of digits for the floating-point type, @code{float_rep} is
9502 the float representation (I/V/A for IEEE-754-Binary, Vax_Native,
9503 AAMP), @code{size} is the size in bits, @code{alignment} is the
9504 alignment in bits. The name is followed by at least two blanks, fields
9505 are separated by at least one blank, and a LF character immediately
9506 follows the alignment field.
9508 Here is an example of a target parameterization file:
9516 Double_Float_Alignment 0
9517 Double_Scalar_Alignment 0
9522 Long_Double_Size 128
9525 Maximum_Alignment 16
9526 Max_Unaligned_Field 64
9530 System_Allocator_Alignment 16
9536 long double 18 I 80 128
9541 @geindex -gnateu (gcc)
9546 @item @code{-gnateu}
9548 Ignore unrecognized validity, warning, and style switches that
9549 appear after this switch is given. This may be useful when
9550 compiling sources developed on a later version of the compiler
9551 with an earlier version. Of course the earlier version must
9552 support this switch.
9555 @geindex -gnateV (gcc)
9560 @item @code{-gnateV}
9562 Check that all actual parameters of a subprogram call are valid according to
9563 the rules of validity checking (@ref{f6,,Validity Checking}).
9566 @geindex -gnateY (gcc)
9571 @item @code{-gnateY}
9573 Ignore all STYLE_CHECKS pragmas. Full legality checks
9574 are still carried out, but the pragmas have no effect
9575 on what style checks are active. This allows all style
9576 checking options to be controlled from the command line.
9579 @geindex -gnatE (gcc)
9586 Full dynamic elaboration checks.
9589 @geindex -gnatf (gcc)
9596 Full errors. Multiple errors per line, all undefined references, do not
9597 attempt to suppress cascaded errors.
9600 @geindex -gnatF (gcc)
9607 Externals names are folded to all uppercase.
9610 @geindex -gnatg (gcc)
9617 Internal GNAT implementation mode. This should not be used for applications
9618 programs, it is intended only for use by the compiler and its run-time
9619 library. For documentation, see the GNAT sources. Note that @code{-gnatg}
9620 implies @code{-gnatw.ge} and @code{-gnatyg} so that all standard
9621 warnings and all standard style options are turned on. All warnings and style
9622 messages are treated as errors.
9625 @geindex -gnatG[nn] (gcc)
9630 @item @code{-gnatG=nn}
9632 List generated expanded code in source form.
9635 @geindex -gnath (gcc)
9642 Output usage information. The output is written to @code{stdout}.
9645 @geindex -gnatH (gcc)
9652 Legacy elaboration-checking mode enabled. When this switch is in effect, the
9653 pre-18.x access-before-elaboration model becomes the de facto model.
9656 @geindex -gnati (gcc)
9661 @item @code{-gnati@emph{c}}
9663 Identifier character set (@code{c} = 1/2/3/4/8/9/p/f/n/w).
9664 For details of the possible selections for @code{c},
9665 see @ref{48,,Character Set Control}.
9668 @geindex -gnatI (gcc)
9675 Ignore representation clauses. When this switch is used,
9676 representation clauses are treated as comments. This is useful
9677 when initially porting code where you want to ignore rep clause
9678 problems, and also for compiling foreign code (particularly
9679 for use with ASIS). The representation clauses that are ignored
9680 are: enumeration_representation_clause, record_representation_clause,
9681 and attribute_definition_clause for the following attributes:
9682 Address, Alignment, Bit_Order, Component_Size, Machine_Radix,
9683 Object_Size, Scalar_Storage_Order, Size, Small, Stream_Size,
9684 and Value_Size. Pragma Default_Scalar_Storage_Order is also ignored.
9685 Note that this option should be used only for compiling -- the
9686 code is likely to malfunction at run time.
9688 Note that when @code{-gnatct} is used to generate trees for input
9689 into ASIS tools, these representation clauses are removed
9690 from the tree and ignored. This means that the tool will not see them.
9693 @geindex -gnatjnn (gcc)
9698 @item @code{-gnatj@emph{nn}}
9700 Reformat error messages to fit on @code{nn} character lines
9703 @geindex -gnatJ (gcc)
9710 Permissive elaboration-checking mode enabled. When this switch is in effect,
9711 the post-18.x access-before-elaboration model ignores potential issues with:
9720 Activations of tasks defined in instances
9726 Calls from within an instance to its enclosing context
9729 Calls through generic formal parameters
9732 Calls to subprograms defined in instances
9738 Indirect calls using 'Access
9747 Synchronous task suspension
9750 and does not emit compile-time diagnostics or run-time checks.
9753 @geindex -gnatk (gcc)
9758 @item @code{-gnatk=@emph{n}}
9760 Limit file names to @code{n} (1-999) characters (@code{k} = krunch).
9763 @geindex -gnatl (gcc)
9770 Output full source listing with embedded error messages.
9773 @geindex -gnatL (gcc)
9780 Used in conjunction with -gnatG or -gnatD to intersperse original
9781 source lines (as comment lines with line numbers) in the expanded
9785 @geindex -gnatm (gcc)
9790 @item @code{-gnatm=@emph{n}}
9792 Limit number of detected error or warning messages to @code{n}
9793 where @code{n} is in the range 1..999999. The default setting if
9794 no switch is given is 9999. If the number of warnings reaches this
9795 limit, then a message is output and further warnings are suppressed,
9796 but the compilation is continued. If the number of error messages
9797 reaches this limit, then a message is output and the compilation
9798 is abandoned. The equal sign here is optional. A value of zero
9799 means that no limit applies.
9802 @geindex -gnatn (gcc)
9807 @item @code{-gnatn[12]}
9809 Activate inlining across units for subprograms for which pragma @code{Inline}
9810 is specified. This inlining is performed by the GCC back-end. An optional
9811 digit sets the inlining level: 1 for moderate inlining across units
9812 or 2 for full inlining across units. If no inlining level is specified,
9813 the compiler will pick it based on the optimization level.
9816 @geindex -gnatN (gcc)
9823 Activate front end inlining for subprograms for which
9824 pragma @code{Inline} is specified. This inlining is performed
9825 by the front end and will be visible in the
9826 @code{-gnatG} output.
9828 When using a gcc-based back end (in practice this means using any version
9829 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
9830 @code{-gnatN} is deprecated, and the use of @code{-gnatn} is preferred.
9831 Historically front end inlining was more extensive than the gcc back end
9832 inlining, but that is no longer the case.
9835 @geindex -gnato0 (gcc)
9840 @item @code{-gnato0}
9842 Suppresses overflow checking. This causes the behavior of the compiler to
9843 match the default for older versions where overflow checking was suppressed
9844 by default. This is equivalent to having
9845 @code{pragma Suppress (Overflow_Check)} in a configuration pragma file.
9848 @geindex -gnato?? (gcc)
9853 @item @code{-gnato??}
9855 Set default mode for handling generation of code to avoid intermediate
9856 arithmetic overflow. Here @code{??} is two digits, a
9857 single digit, or nothing. Each digit is one of the digits @code{1}
9861 @multitable {xxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
9876 All intermediate overflows checked against base type (@code{STRICT})
9884 Minimize intermediate overflows (@code{MINIMIZED})
9892 Eliminate intermediate overflows (@code{ELIMINATED})
9897 If only one digit appears, then it applies to all
9898 cases; if two digits are given, then the first applies outside
9899 assertions, pre/postconditions, and type invariants, and the second
9900 applies within assertions, pre/postconditions, and type invariants.
9902 If no digits follow the @code{-gnato}, then it is equivalent to
9904 causing all intermediate overflows to be handled in strict
9907 This switch also causes arithmetic overflow checking to be performed
9908 (as though @code{pragma Unsuppress (Overflow_Check)} had been specified).
9910 The default if no option @code{-gnato} is given is that overflow handling
9911 is in @code{STRICT} mode (computations done using the base type), and that
9912 overflow checking is enabled.
9914 Note that division by zero is a separate check that is not
9915 controlled by this switch (divide-by-zero checking is on by default).
9917 See also @ref{f8,,Specifying the Desired Mode}.
9920 @geindex -gnatp (gcc)
9927 Suppress all checks. See @ref{f9,,Run-Time Checks} for details. This switch
9928 has no effect if cancelled by a subsequent @code{-gnat-p} switch.
9931 @geindex -gnat-p (gcc)
9936 @item @code{-gnat-p}
9938 Cancel effect of previous @code{-gnatp} switch.
9941 @geindex -gnatP (gcc)
9948 Enable polling. This is required on some systems (notably Windows NT) to
9949 obtain asynchronous abort and asynchronous transfer of control capability.
9950 See @code{Pragma_Polling} in the @cite{GNAT_Reference_Manual} for full
9954 @geindex -gnatq (gcc)
9961 Don't quit. Try semantics, even if parse errors.
9964 @geindex -gnatQ (gcc)
9971 Don't quit. Generate @code{ALI} and tree files even if illegalities.
9972 Note that code generation is still suppressed in the presence of any
9973 errors, so even with @code{-gnatQ} no object file is generated.
9976 @geindex -gnatr (gcc)
9983 Treat pragma Restrictions as Restriction_Warnings.
9986 @geindex -gnatR (gcc)
9991 @item @code{-gnatR[0|1|2|3|4][e][j][m][s]}
9993 Output representation information for declared types, objects and
9994 subprograms. Note that this switch is not allowed if a previous
9995 @code{-gnatD} switch has been given, since these two switches
9999 @geindex -gnats (gcc)
10004 @item @code{-gnats}
10009 @geindex -gnatS (gcc)
10014 @item @code{-gnatS}
10016 Print package Standard.
10019 @geindex -gnatt (gcc)
10024 @item @code{-gnatt}
10026 Generate tree output file.
10029 @geindex -gnatT (gcc)
10034 @item @code{-gnatT@emph{nnn}}
10036 All compiler tables start at @code{nnn} times usual starting size.
10039 @geindex -gnatu (gcc)
10044 @item @code{-gnatu}
10046 List units for this compilation.
10049 @geindex -gnatU (gcc)
10054 @item @code{-gnatU}
10056 Tag all error messages with the unique string 'error:'
10059 @geindex -gnatv (gcc)
10064 @item @code{-gnatv}
10066 Verbose mode. Full error output with source lines to @code{stdout}.
10069 @geindex -gnatV (gcc)
10074 @item @code{-gnatV}
10076 Control level of validity checking (@ref{f6,,Validity Checking}).
10079 @geindex -gnatw (gcc)
10084 @item @code{-gnatw@emph{xxx}}
10087 @code{xxx} is a string of option letters that denotes
10088 the exact warnings that
10089 are enabled or disabled (@ref{fa,,Warning Message Control}).
10092 @geindex -gnatW (gcc)
10097 @item @code{-gnatW@emph{e}}
10099 Wide character encoding method
10100 (@code{e}=n/h/u/s/e/8).
10103 @geindex -gnatx (gcc)
10108 @item @code{-gnatx}
10110 Suppress generation of cross-reference information.
10113 @geindex -gnatX (gcc)
10118 @item @code{-gnatX}
10120 Enable GNAT implementation extensions and latest Ada version.
10123 @geindex -gnaty (gcc)
10128 @item @code{-gnaty}
10130 Enable built-in style checks (@ref{fb,,Style Checking}).
10133 @geindex -gnatz (gcc)
10138 @item @code{-gnatz@emph{m}}
10140 Distribution stub generation and compilation
10141 (@code{m}=r/c for receiver/caller stubs).
10149 @item @code{-I@emph{dir}}
10153 Direct GNAT to search the @code{dir} directory for source files needed by
10154 the current compilation
10155 (see @ref{89,,Search Paths and the Run-Time Library (RTL)}).
10167 Except for the source file named in the command line, do not look for source
10168 files in the directory containing the source file named in the command line
10169 (see @ref{89,,Search Paths and the Run-Time Library (RTL)}).
10177 @item @code{-o @emph{file}}
10179 This switch is used in @code{gcc} to redirect the generated object file
10180 and its associated ALI file. Beware of this switch with GNAT, because it may
10181 cause the object file and ALI file to have different names which in turn
10182 may confuse the binder and the linker.
10185 @geindex -nostdinc (gcc)
10190 @item @code{-nostdinc}
10192 Inhibit the search of the default location for the GNAT Run Time
10193 Library (RTL) source files.
10196 @geindex -nostdlib (gcc)
10201 @item @code{-nostdlib}
10203 Inhibit the search of the default location for the GNAT Run Time
10204 Library (RTL) ALI files.
10212 @item @code{-O[@emph{n}]}
10214 @code{n} controls the optimization level:
10217 @multitable {xxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
10232 No optimization, the default setting if no @code{-O} appears
10240 Normal optimization, the default if you specify @code{-O} without an
10241 operand. A good compromise between code quality and compilation
10250 Extensive optimization, may improve execution time, possibly at
10251 the cost of substantially increased compilation time.
10259 Same as @code{-O2}, and also includes inline expansion for small
10260 subprograms in the same unit.
10268 Optimize space usage
10273 See also @ref{fc,,Optimization Levels}.
10276 @geindex -pass-exit-codes (gcc)
10281 @item @code{-pass-exit-codes}
10283 Catch exit codes from the compiler and use the most meaningful as
10287 @geindex --RTS (gcc)
10292 @item @code{--RTS=@emph{rts-path}}
10294 Specifies the default location of the run-time library. Same meaning as the
10295 equivalent @code{gnatmake} flag (@ref{dc,,Switches for gnatmake}).
10305 Used in place of @code{-c} to
10306 cause the assembler source file to be
10307 generated, using @code{.s} as the extension,
10308 instead of the object file.
10309 This may be useful if you need to examine the generated assembly code.
10312 @geindex -fverbose-asm (gcc)
10317 @item @code{-fverbose-asm}
10319 Used in conjunction with @code{-S}
10320 to cause the generated assembly code file to be annotated with variable
10321 names, making it significantly easier to follow.
10331 Show commands generated by the @code{gcc} driver. Normally used only for
10332 debugging purposes or if you need to be sure what version of the
10333 compiler you are executing.
10341 @item @code{-V @emph{ver}}
10343 Execute @code{ver} version of the compiler. This is the @code{gcc}
10344 version, not the GNAT version.
10354 Turn off warnings generated by the back end of the compiler. Use of
10355 this switch also causes the default for front end warnings to be set
10356 to suppress (as though @code{-gnatws} had appeared at the start of
10360 @geindex Combining GNAT switches
10362 You may combine a sequence of GNAT switches into a single switch. For
10363 example, the combined switch
10372 is equivalent to specifying the following sequence of switches:
10377 -gnato -gnatf -gnati3
10381 The following restrictions apply to the combination of switches
10388 The switch @code{-gnatc} if combined with other switches must come
10389 first in the string.
10392 The switch @code{-gnats} if combined with other switches must come
10393 first in the string.
10397 @code{-gnatzc} and @code{-gnatzr} may not be combined with any other
10398 switches, and only one of them may appear in the command line.
10401 The switch @code{-gnat-p} may not be combined with any other switch.
10404 Once a 'y' appears in the string (that is a use of the @code{-gnaty}
10405 switch), then all further characters in the switch are interpreted
10406 as style modifiers (see description of @code{-gnaty}).
10409 Once a 'd' appears in the string (that is a use of the @code{-gnatd}
10410 switch), then all further characters in the switch are interpreted
10411 as debug flags (see description of @code{-gnatd}).
10414 Once a 'w' appears in the string (that is a use of the @code{-gnatw}
10415 switch), then all further characters in the switch are interpreted
10416 as warning mode modifiers (see description of @code{-gnatw}).
10419 Once a 'V' appears in the string (that is a use of the @code{-gnatV}
10420 switch), then all further characters in the switch are interpreted
10421 as validity checking options (@ref{f6,,Validity Checking}).
10424 Option 'em', 'ec', 'ep', 'l=' and 'R' must be the last options in
10425 a combined list of options.
10428 @node Output and Error Message Control,Warning Message Control,Alphabetical List of All Switches,Compiler Switches
10429 @anchor{gnat_ugn/building_executable_programs_with_gnat id14}@anchor{fd}@anchor{gnat_ugn/building_executable_programs_with_gnat output-and-error-message-control}@anchor{fe}
10430 @subsection Output and Error Message Control
10435 The standard default format for error messages is called 'brief format'.
10436 Brief format messages are written to @code{stderr} (the standard error
10437 file) and have the following form:
10440 e.adb:3:04: Incorrect spelling of keyword "function"
10441 e.adb:4:20: ";" should be "is"
10444 The first integer after the file name is the line number in the file,
10445 and the second integer is the column number within the line.
10446 @code{GPS} can parse the error messages
10447 and point to the referenced character.
10448 The following switches provide control over the error message
10451 @geindex -gnatv (gcc)
10456 @item @code{-gnatv}
10458 The @code{v} stands for verbose.
10459 The effect of this setting is to write long-format error
10460 messages to @code{stdout} (the standard output file.
10461 The same program compiled with the
10462 @code{-gnatv} switch would generate:
10465 3. funcion X (Q : Integer)
10467 >>> Incorrect spelling of keyword "function"
10470 >>> ";" should be "is"
10473 The vertical bar indicates the location of the error, and the @code{>>>}
10474 prefix can be used to search for error messages. When this switch is
10475 used the only source lines output are those with errors.
10478 @geindex -gnatl (gcc)
10483 @item @code{-gnatl}
10485 The @code{l} stands for list.
10486 This switch causes a full listing of
10487 the file to be generated. In the case where a body is
10488 compiled, the corresponding spec is also listed, along
10489 with any subunits. Typical output from compiling a package
10490 body @code{p.adb} might look like:
10495 1. package body p is
10497 3. procedure a is separate;
10508 2. pragma Elaborate_Body
10529 When you specify the @code{-gnatv} or @code{-gnatl} switches and
10530 standard output is redirected, a brief summary is written to
10531 @code{stderr} (standard error) giving the number of error messages and
10532 warning messages generated.
10535 @geindex -gnatl=fname (gcc)
10540 @item @code{-gnatl=@emph{fname}}
10542 This has the same effect as @code{-gnatl} except that the output is
10543 written to a file instead of to standard output. If the given name
10544 @code{fname} does not start with a period, then it is the full name
10545 of the file to be written. If @code{fname} is an extension, it is
10546 appended to the name of the file being compiled. For example, if
10547 file @code{xyz.adb} is compiled with @code{-gnatl=.lst},
10548 then the output is written to file xyz.adb.lst.
10551 @geindex -gnatU (gcc)
10556 @item @code{-gnatU}
10558 This switch forces all error messages to be preceded by the unique
10559 string 'error:'. This means that error messages take a few more
10560 characters in space, but allows easy searching for and identification
10564 @geindex -gnatb (gcc)
10569 @item @code{-gnatb}
10571 The @code{b} stands for brief.
10572 This switch causes GNAT to generate the
10573 brief format error messages to @code{stderr} (the standard error
10574 file) as well as the verbose
10575 format message or full listing (which as usual is written to
10576 @code{stdout} (the standard output file).
10579 @geindex -gnatm (gcc)
10584 @item @code{-gnatm=@emph{n}}
10586 The @code{m} stands for maximum.
10587 @code{n} is a decimal integer in the
10588 range of 1 to 999999 and limits the number of error or warning
10589 messages to be generated. For example, using
10590 @code{-gnatm2} might yield
10593 e.adb:3:04: Incorrect spelling of keyword "function"
10594 e.adb:5:35: missing ".."
10595 fatal error: maximum number of errors detected
10596 compilation abandoned
10599 The default setting if
10600 no switch is given is 9999. If the number of warnings reaches this
10601 limit, then a message is output and further warnings are suppressed,
10602 but the compilation is continued. If the number of error messages
10603 reaches this limit, then a message is output and the compilation
10604 is abandoned. A value of zero means that no limit applies.
10606 Note that the equal sign is optional, so the switches
10607 @code{-gnatm2} and @code{-gnatm=2} are equivalent.
10610 @geindex -gnatf (gcc)
10615 @item @code{-gnatf}
10617 @geindex Error messages
10618 @geindex suppressing
10620 The @code{f} stands for full.
10621 Normally, the compiler suppresses error messages that are likely to be
10622 redundant. This switch causes all error
10623 messages to be generated. In particular, in the case of
10624 references to undefined variables. If a given variable is referenced
10625 several times, the normal format of messages is
10628 e.adb:7:07: "V" is undefined (more references follow)
10631 where the parenthetical comment warns that there are additional
10632 references to the variable @code{V}. Compiling the same program with the
10633 @code{-gnatf} switch yields
10636 e.adb:7:07: "V" is undefined
10637 e.adb:8:07: "V" is undefined
10638 e.adb:8:12: "V" is undefined
10639 e.adb:8:16: "V" is undefined
10640 e.adb:9:07: "V" is undefined
10641 e.adb:9:12: "V" is undefined
10644 The @code{-gnatf} switch also generates additional information for
10645 some error messages. Some examples are:
10651 Details on possibly non-portable unchecked conversion
10654 List possible interpretations for ambiguous calls
10657 Additional details on incorrect parameters
10661 @geindex -gnatjnn (gcc)
10666 @item @code{-gnatjnn}
10668 In normal operation mode (or if @code{-gnatj0} is used), then error messages
10669 with continuation lines are treated as though the continuation lines were
10670 separate messages (and so a warning with two continuation lines counts as
10671 three warnings, and is listed as three separate messages).
10673 If the @code{-gnatjnn} switch is used with a positive value for nn, then
10674 messages are output in a different manner. A message and all its continuation
10675 lines are treated as a unit, and count as only one warning or message in the
10676 statistics totals. Furthermore, the message is reformatted so that no line
10677 is longer than nn characters.
10680 @geindex -gnatq (gcc)
10685 @item @code{-gnatq}
10687 The @code{q} stands for quit (really 'don't quit').
10688 In normal operation mode, the compiler first parses the program and
10689 determines if there are any syntax errors. If there are, appropriate
10690 error messages are generated and compilation is immediately terminated.
10692 GNAT to continue with semantic analysis even if syntax errors have been
10693 found. This may enable the detection of more errors in a single run. On
10694 the other hand, the semantic analyzer is more likely to encounter some
10695 internal fatal error when given a syntactically invalid tree.
10698 @geindex -gnatQ (gcc)
10703 @item @code{-gnatQ}
10705 In normal operation mode, the @code{ALI} file is not generated if any
10706 illegalities are detected in the program. The use of @code{-gnatQ} forces
10707 generation of the @code{ALI} file. This file is marked as being in
10708 error, so it cannot be used for binding purposes, but it does contain
10709 reasonably complete cross-reference information, and thus may be useful
10710 for use by tools (e.g., semantic browsing tools or integrated development
10711 environments) that are driven from the @code{ALI} file. This switch
10712 implies @code{-gnatq}, since the semantic phase must be run to get a
10713 meaningful ALI file.
10715 In addition, if @code{-gnatt} is also specified, then the tree file is
10716 generated even if there are illegalities. It may be useful in this case
10717 to also specify @code{-gnatq} to ensure that full semantic processing
10718 occurs. The resulting tree file can be processed by ASIS, for the purpose
10719 of providing partial information about illegal units, but if the error
10720 causes the tree to be badly malformed, then ASIS may crash during the
10723 When @code{-gnatQ} is used and the generated @code{ALI} file is marked as
10724 being in error, @code{gnatmake} will attempt to recompile the source when it
10725 finds such an @code{ALI} file, including with switch @code{-gnatc}.
10727 Note that @code{-gnatQ} has no effect if @code{-gnats} is specified,
10728 since ALI files are never generated if @code{-gnats} is set.
10731 @node Warning Message Control,Debugging and Assertion Control,Output and Error Message Control,Compiler Switches
10732 @anchor{gnat_ugn/building_executable_programs_with_gnat warning-message-control}@anchor{fa}@anchor{gnat_ugn/building_executable_programs_with_gnat id15}@anchor{ff}
10733 @subsection Warning Message Control
10736 @geindex Warning messages
10738 In addition to error messages, which correspond to illegalities as defined
10739 in the Ada Reference Manual, the compiler detects two kinds of warning
10742 First, the compiler considers some constructs suspicious and generates a
10743 warning message to alert you to a possible error. Second, if the
10744 compiler detects a situation that is sure to raise an exception at
10745 run time, it generates a warning message. The following shows an example
10746 of warning messages:
10749 e.adb:4:24: warning: creation of object may raise Storage_Error
10750 e.adb:10:17: warning: static value out of range
10751 e.adb:10:17: warning: "Constraint_Error" will be raised at run time
10754 GNAT considers a large number of situations as appropriate
10755 for the generation of warning messages. As always, warnings are not
10756 definite indications of errors. For example, if you do an out-of-range
10757 assignment with the deliberate intention of raising a
10758 @code{Constraint_Error} exception, then the warning that may be
10759 issued does not indicate an error. Some of the situations for which GNAT
10760 issues warnings (at least some of the time) are given in the following
10761 list. This list is not complete, and new warnings are often added to
10762 subsequent versions of GNAT. The list is intended to give a general idea
10763 of the kinds of warnings that are generated.
10769 Possible infinitely recursive calls
10772 Out-of-range values being assigned
10775 Possible order of elaboration problems
10778 Size not a multiple of alignment for a record type
10781 Assertions (pragma Assert) that are sure to fail
10787 Address clauses with possibly unaligned values, or where an attempt is
10788 made to overlay a smaller variable with a larger one.
10791 Fixed-point type declarations with a null range
10794 Direct_IO or Sequential_IO instantiated with a type that has access values
10797 Variables that are never assigned a value
10800 Variables that are referenced before being initialized
10803 Task entries with no corresponding @code{accept} statement
10806 Duplicate accepts for the same task entry in a @code{select}
10809 Objects that take too much storage
10812 Unchecked conversion between types of differing sizes
10815 Missing @code{return} statement along some execution path in a function
10818 Incorrect (unrecognized) pragmas
10821 Incorrect external names
10824 Allocation from empty storage pool
10827 Potentially blocking operation in protected type
10830 Suspicious parenthesization of expressions
10833 Mismatching bounds in an aggregate
10836 Attempt to return local value by reference
10839 Premature instantiation of a generic body
10842 Attempt to pack aliased components
10845 Out of bounds array subscripts
10848 Wrong length on string assignment
10851 Violations of style rules if style checking is enabled
10854 Unused @emph{with} clauses
10857 @code{Bit_Order} usage that does not have any effect
10860 @code{Standard.Duration} used to resolve universal fixed expression
10863 Dereference of possibly null value
10866 Declaration that is likely to cause storage error
10869 Internal GNAT unit @emph{with}ed by application unit
10872 Values known to be out of range at compile time
10875 Unreferenced or unmodified variables. Note that a special
10876 exemption applies to variables which contain any of the substrings
10877 @code{DISCARD, DUMMY, IGNORE, JUNK, UNUSED}, in any casing. Such variables
10878 are considered likely to be intentionally used in a situation where
10879 otherwise a warning would be given, so warnings of this kind are
10880 always suppressed for such variables.
10883 Address overlays that could clobber memory
10886 Unexpected initialization when address clause present
10889 Bad alignment for address clause
10892 Useless type conversions
10895 Redundant assignment statements and other redundant constructs
10898 Useless exception handlers
10901 Accidental hiding of name by child unit
10904 Access before elaboration detected at compile time
10907 A range in a @code{for} loop that is known to be null or might be null
10910 The following section lists compiler switches that are available
10911 to control the handling of warning messages. It is also possible
10912 to exercise much finer control over what warnings are issued and
10913 suppressed using the GNAT pragma Warnings (see the description
10914 of the pragma in the @cite{GNAT_Reference_manual}).
10916 @geindex -gnatwa (gcc)
10921 @item @code{-gnatwa}
10923 @emph{Activate most optional warnings.}
10925 This switch activates most optional warning messages. See the remaining list
10926 in this section for details on optional warning messages that can be
10927 individually controlled. The warnings that are not turned on by this
10934 @code{-gnatwd} (implicit dereferencing)
10937 @code{-gnatw.d} (tag warnings with -gnatw switch)
10940 @code{-gnatwh} (hiding)
10943 @code{-gnatw.h} (holes in record layouts)
10946 @code{-gnatw.j} (late primitives of tagged types)
10949 @code{-gnatw.k} (redefinition of names in standard)
10952 @code{-gnatwl} (elaboration warnings)
10955 @code{-gnatw.l} (inherited aspects)
10958 @code{-gnatw.n} (atomic synchronization)
10961 @code{-gnatwo} (address clause overlay)
10964 @code{-gnatw.o} (values set by out parameters ignored)
10967 @code{-gnatw.q} (questionable layout of record types)
10970 @code{-gnatw.s} (overridden size clause)
10973 @code{-gnatwt} (tracking of deleted conditional code)
10976 @code{-gnatw.u} (unordered enumeration)
10979 @code{-gnatw.w} (use of Warnings Off)
10982 @code{-gnatw.y} (reasons for package needing body)
10985 All other optional warnings are turned on.
10988 @geindex -gnatwA (gcc)
10993 @item @code{-gnatwA}
10995 @emph{Suppress all optional errors.}
10997 This switch suppresses all optional warning messages, see remaining list
10998 in this section for details on optional warning messages that can be
10999 individually controlled. Note that unlike switch @code{-gnatws}, the
11000 use of switch @code{-gnatwA} does not suppress warnings that are
11001 normally given unconditionally and cannot be individually controlled
11002 (for example, the warning about a missing exit path in a function).
11003 Also, again unlike switch @code{-gnatws}, warnings suppressed by
11004 the use of switch @code{-gnatwA} can be individually turned back
11005 on. For example the use of switch @code{-gnatwA} followed by
11006 switch @code{-gnatwd} will suppress all optional warnings except
11007 the warnings for implicit dereferencing.
11010 @geindex -gnatw.a (gcc)
11015 @item @code{-gnatw.a}
11017 @emph{Activate warnings on failing assertions.}
11019 @geindex Assert failures
11021 This switch activates warnings for assertions where the compiler can tell at
11022 compile time that the assertion will fail. Note that this warning is given
11023 even if assertions are disabled. The default is that such warnings are
11027 @geindex -gnatw.A (gcc)
11032 @item @code{-gnatw.A}
11034 @emph{Suppress warnings on failing assertions.}
11036 @geindex Assert failures
11038 This switch suppresses warnings for assertions where the compiler can tell at
11039 compile time that the assertion will fail.
11042 @geindex -gnatwb (gcc)
11047 @item @code{-gnatwb}
11049 @emph{Activate warnings on bad fixed values.}
11051 @geindex Bad fixed values
11053 @geindex Fixed-point Small value
11055 @geindex Small value
11057 This switch activates warnings for static fixed-point expressions whose
11058 value is not an exact multiple of Small. Such values are implementation
11059 dependent, since an implementation is free to choose either of the multiples
11060 that surround the value. GNAT always chooses the closer one, but this is not
11061 required behavior, and it is better to specify a value that is an exact
11062 multiple, ensuring predictable execution. The default is that such warnings
11066 @geindex -gnatwB (gcc)
11071 @item @code{-gnatwB}
11073 @emph{Suppress warnings on bad fixed values.}
11075 This switch suppresses warnings for static fixed-point expressions whose
11076 value is not an exact multiple of Small.
11079 @geindex -gnatw.b (gcc)
11084 @item @code{-gnatw.b}
11086 @emph{Activate warnings on biased representation.}
11088 @geindex Biased representation
11090 This switch activates warnings when a size clause, value size clause, component
11091 clause, or component size clause forces the use of biased representation for an
11092 integer type (e.g. representing a range of 10..11 in a single bit by using 0/1
11093 to represent 10/11). The default is that such warnings are generated.
11096 @geindex -gnatwB (gcc)
11101 @item @code{-gnatw.B}
11103 @emph{Suppress warnings on biased representation.}
11105 This switch suppresses warnings for representation clauses that force the use
11106 of biased representation.
11109 @geindex -gnatwc (gcc)
11114 @item @code{-gnatwc}
11116 @emph{Activate warnings on conditionals.}
11118 @geindex Conditionals
11121 This switch activates warnings for conditional expressions used in
11122 tests that are known to be True or False at compile time. The default
11123 is that such warnings are not generated.
11124 Note that this warning does
11125 not get issued for the use of boolean variables or constants whose
11126 values are known at compile time, since this is a standard technique
11127 for conditional compilation in Ada, and this would generate too many
11128 false positive warnings.
11130 This warning option also activates a special test for comparisons using
11131 the operators '>=' and' <='.
11132 If the compiler can tell that only the equality condition is possible,
11133 then it will warn that the '>' or '<' part of the test
11134 is useless and that the operator could be replaced by '='.
11135 An example would be comparing a @code{Natural} variable <= 0.
11137 This warning option also generates warnings if
11138 one or both tests is optimized away in a membership test for integer
11139 values if the result can be determined at compile time. Range tests on
11140 enumeration types are not included, since it is common for such tests
11141 to include an end point.
11143 This warning can also be turned on using @code{-gnatwa}.
11146 @geindex -gnatwC (gcc)
11151 @item @code{-gnatwC}
11153 @emph{Suppress warnings on conditionals.}
11155 This switch suppresses warnings for conditional expressions used in
11156 tests that are known to be True or False at compile time.
11159 @geindex -gnatw.c (gcc)
11164 @item @code{-gnatw.c}
11166 @emph{Activate warnings on missing component clauses.}
11168 @geindex Component clause
11171 This switch activates warnings for record components where a record
11172 representation clause is present and has component clauses for the
11173 majority, but not all, of the components. A warning is given for each
11174 component for which no component clause is present.
11177 @geindex -gnatwC (gcc)
11182 @item @code{-gnatw.C}
11184 @emph{Suppress warnings on missing component clauses.}
11186 This switch suppresses warnings for record components that are
11187 missing a component clause in the situation described above.
11190 @geindex -gnatwd (gcc)
11195 @item @code{-gnatwd}
11197 @emph{Activate warnings on implicit dereferencing.}
11199 If this switch is set, then the use of a prefix of an access type
11200 in an indexed component, slice, or selected component without an
11201 explicit @code{.all} will generate a warning. With this warning
11202 enabled, access checks occur only at points where an explicit
11203 @code{.all} appears in the source code (assuming no warnings are
11204 generated as a result of this switch). The default is that such
11205 warnings are not generated.
11208 @geindex -gnatwD (gcc)
11213 @item @code{-gnatwD}
11215 @emph{Suppress warnings on implicit dereferencing.}
11217 @geindex Implicit dereferencing
11219 @geindex Dereferencing
11222 This switch suppresses warnings for implicit dereferences in
11223 indexed components, slices, and selected components.
11226 @geindex -gnatw.d (gcc)
11231 @item @code{-gnatw.d}
11233 @emph{Activate tagging of warning and info messages.}
11235 If this switch is set, then warning messages are tagged, with one of the
11245 Used to tag warnings controlled by the switch @code{-gnatwx} where x
11250 Used to tag warnings controlled by the switch @code{-gnatw.x} where x
11255 Used to tag elaboration information (info) messages generated when the
11256 static model of elaboration is used and the @code{-gnatel} switch is set.
11259 @emph{[restriction warning]}
11260 Used to tag warning messages for restriction violations, activated by use
11261 of the pragma @code{Restriction_Warnings}.
11264 @emph{[warning-as-error]}
11265 Used to tag warning messages that have been converted to error messages by
11266 use of the pragma Warning_As_Error. Note that such warnings are prefixed by
11267 the string "error: " rather than "warning: ".
11270 @emph{[enabled by default]}
11271 Used to tag all other warnings that are always given by default, unless
11272 warnings are completely suppressed using pragma @emph{Warnings(Off)} or
11273 the switch @code{-gnatws}.
11278 @geindex -gnatw.d (gcc)
11283 @item @code{-gnatw.D}
11285 @emph{Deactivate tagging of warning and info messages messages.}
11287 If this switch is set, then warning messages return to the default
11288 mode in which warnings and info messages are not tagged as described above for
11292 @geindex -gnatwe (gcc)
11295 @geindex treat as error
11300 @item @code{-gnatwe}
11302 @emph{Treat warnings and style checks as errors.}
11304 This switch causes warning messages and style check messages to be
11306 The warning string still appears, but the warning messages are counted
11307 as errors, and prevent the generation of an object file. Note that this
11308 is the only -gnatw switch that affects the handling of style check messages.
11309 Note also that this switch has no effect on info (information) messages, which
11310 are not treated as errors if this switch is present.
11313 @geindex -gnatw.e (gcc)
11318 @item @code{-gnatw.e}
11320 @emph{Activate every optional warning.}
11323 @geindex activate every optional warning
11325 This switch activates all optional warnings, including those which
11326 are not activated by @code{-gnatwa}. The use of this switch is not
11327 recommended for normal use. If you turn this switch on, it is almost
11328 certain that you will get large numbers of useless warnings. The
11329 warnings that are excluded from @code{-gnatwa} are typically highly
11330 specialized warnings that are suitable for use only in code that has
11331 been specifically designed according to specialized coding rules.
11334 @geindex -gnatwE (gcc)
11337 @geindex treat as error
11342 @item @code{-gnatwE}
11344 @emph{Treat all run-time exception warnings as errors.}
11346 This switch causes warning messages regarding errors that will be raised
11347 during run-time execution to be treated as errors.
11350 @geindex -gnatwf (gcc)
11355 @item @code{-gnatwf}
11357 @emph{Activate warnings on unreferenced formals.}
11360 @geindex unreferenced
11362 This switch causes a warning to be generated if a formal parameter
11363 is not referenced in the body of the subprogram. This warning can
11364 also be turned on using @code{-gnatwu}. The
11365 default is that these warnings are not generated.
11368 @geindex -gnatwF (gcc)
11373 @item @code{-gnatwF}
11375 @emph{Suppress warnings on unreferenced formals.}
11377 This switch suppresses warnings for unreferenced formal
11378 parameters. Note that the
11379 combination @code{-gnatwu} followed by @code{-gnatwF} has the
11380 effect of warning on unreferenced entities other than subprogram
11384 @geindex -gnatwg (gcc)
11389 @item @code{-gnatwg}
11391 @emph{Activate warnings on unrecognized pragmas.}
11394 @geindex unrecognized
11396 This switch causes a warning to be generated if an unrecognized
11397 pragma is encountered. Apart from issuing this warning, the
11398 pragma is ignored and has no effect. The default
11399 is that such warnings are issued (satisfying the Ada Reference
11400 Manual requirement that such warnings appear).
11403 @geindex -gnatwG (gcc)
11408 @item @code{-gnatwG}
11410 @emph{Suppress warnings on unrecognized pragmas.}
11412 This switch suppresses warnings for unrecognized pragmas.
11415 @geindex -gnatw.g (gcc)
11420 @item @code{-gnatw.g}
11422 @emph{Warnings used for GNAT sources.}
11424 This switch sets the warning categories that are used by the standard
11425 GNAT style. Currently this is equivalent to
11426 @code{-gnatwAao.q.s.CI.V.X.Z}
11427 but more warnings may be added in the future without advanced notice.
11430 @geindex -gnatwh (gcc)
11435 @item @code{-gnatwh}
11437 @emph{Activate warnings on hiding.}
11439 @geindex Hiding of Declarations
11441 This switch activates warnings on hiding declarations that are considered
11442 potentially confusing. Not all cases of hiding cause warnings; for example an
11443 overriding declaration hides an implicit declaration, which is just normal
11444 code. The default is that warnings on hiding are not generated.
11447 @geindex -gnatwH (gcc)
11452 @item @code{-gnatwH}
11454 @emph{Suppress warnings on hiding.}
11456 This switch suppresses warnings on hiding declarations.
11459 @geindex -gnatw.h (gcc)
11464 @item @code{-gnatw.h}
11466 @emph{Activate warnings on holes/gaps in records.}
11468 @geindex Record Representation (gaps)
11470 This switch activates warnings on component clauses in record
11471 representation clauses that leave holes (gaps) in the record layout.
11472 If this warning option is active, then record representation clauses
11473 should specify a contiguous layout, adding unused fill fields if needed.
11476 @geindex -gnatw.H (gcc)
11481 @item @code{-gnatw.H}
11483 @emph{Suppress warnings on holes/gaps in records.}
11485 This switch suppresses warnings on component clauses in record
11486 representation clauses that leave holes (haps) in the record layout.
11489 @geindex -gnatwi (gcc)
11494 @item @code{-gnatwi}
11496 @emph{Activate warnings on implementation units.}
11498 This switch activates warnings for a @emph{with} of an internal GNAT
11499 implementation unit, defined as any unit from the @code{Ada},
11500 @code{Interfaces}, @code{GNAT},
11502 hierarchies that is not
11503 documented in either the Ada Reference Manual or the GNAT
11504 Programmer's Reference Manual. Such units are intended only
11505 for internal implementation purposes and should not be @emph{with}ed
11506 by user programs. The default is that such warnings are generated
11509 @geindex -gnatwI (gcc)
11514 @item @code{-gnatwI}
11516 @emph{Disable warnings on implementation units.}
11518 This switch disables warnings for a @emph{with} of an internal GNAT
11519 implementation unit.
11522 @geindex -gnatw.i (gcc)
11527 @item @code{-gnatw.i}
11529 @emph{Activate warnings on overlapping actuals.}
11531 This switch enables a warning on statically detectable overlapping actuals in
11532 a subprogram call, when one of the actuals is an in-out parameter, and the
11533 types of the actuals are not by-copy types. This warning is off by default.
11536 @geindex -gnatw.I (gcc)
11541 @item @code{-gnatw.I}
11543 @emph{Disable warnings on overlapping actuals.}
11545 This switch disables warnings on overlapping actuals in a call..
11548 @geindex -gnatwj (gcc)
11553 @item @code{-gnatwj}
11555 @emph{Activate warnings on obsolescent features (Annex J).}
11558 @geindex obsolescent
11560 @geindex Obsolescent features
11562 If this warning option is activated, then warnings are generated for
11563 calls to subprograms marked with @code{pragma Obsolescent} and
11564 for use of features in Annex J of the Ada Reference Manual. In the
11565 case of Annex J, not all features are flagged. In particular use
11566 of the renamed packages (like @code{Text_IO}) and use of package
11567 @code{ASCII} are not flagged, since these are very common and
11568 would generate many annoying positive warnings. The default is that
11569 such warnings are not generated.
11571 In addition to the above cases, warnings are also generated for
11572 GNAT features that have been provided in past versions but which
11573 have been superseded (typically by features in the new Ada standard).
11574 For example, @code{pragma Ravenscar} will be flagged since its
11575 function is replaced by @code{pragma Profile(Ravenscar)}, and
11576 @code{pragma Interface_Name} will be flagged since its function
11577 is replaced by @code{pragma Import}.
11579 Note that this warning option functions differently from the
11580 restriction @code{No_Obsolescent_Features} in two respects.
11581 First, the restriction applies only to annex J features.
11582 Second, the restriction does flag uses of package @code{ASCII}.
11585 @geindex -gnatwJ (gcc)
11590 @item @code{-gnatwJ}
11592 @emph{Suppress warnings on obsolescent features (Annex J).}
11594 This switch disables warnings on use of obsolescent features.
11597 @geindex -gnatw.j (gcc)
11602 @item @code{-gnatw.j}
11604 @emph{Activate warnings on late declarations of tagged type primitives.}
11606 This switch activates warnings on visible primitives added to a
11607 tagged type after deriving a private extension from it.
11610 @geindex -gnatw.J (gcc)
11615 @item @code{-gnatw.J}
11617 @emph{Suppress warnings on late declarations of tagged type primitives.}
11619 This switch suppresses warnings on visible primitives added to a
11620 tagged type after deriving a private extension from it.
11623 @geindex -gnatwk (gcc)
11628 @item @code{-gnatwk}
11630 @emph{Activate warnings on variables that could be constants.}
11632 This switch activates warnings for variables that are initialized but
11633 never modified, and then could be declared constants. The default is that
11634 such warnings are not given.
11637 @geindex -gnatwK (gcc)
11642 @item @code{-gnatwK}
11644 @emph{Suppress warnings on variables that could be constants.}
11646 This switch disables warnings on variables that could be declared constants.
11649 @geindex -gnatw.k (gcc)
11654 @item @code{-gnatw.k}
11656 @emph{Activate warnings on redefinition of names in standard.}
11658 This switch activates warnings for declarations that declare a name that
11659 is defined in package Standard. Such declarations can be confusing,
11660 especially since the names in package Standard continue to be directly
11661 visible, meaning that use visibiliy on such redeclared names does not
11662 work as expected. Names of discriminants and components in records are
11663 not included in this check.
11666 @geindex -gnatwK (gcc)
11671 @item @code{-gnatw.K}
11673 @emph{Suppress warnings on redefinition of names in standard.}
11675 This switch activates warnings for declarations that declare a name that
11676 is defined in package Standard.
11679 @geindex -gnatwl (gcc)
11684 @item @code{-gnatwl}
11686 @emph{Activate warnings for elaboration pragmas.}
11688 @geindex Elaboration
11691 This switch activates warnings for possible elaboration problems,
11692 including suspicious use
11693 of @code{Elaborate} pragmas, when using the static elaboration model, and
11694 possible situations that may raise @code{Program_Error} when using the
11695 dynamic elaboration model.
11696 See the section in this guide on elaboration checking for further details.
11697 The default is that such warnings
11701 @geindex -gnatwL (gcc)
11706 @item @code{-gnatwL}
11708 @emph{Suppress warnings for elaboration pragmas.}
11710 This switch suppresses warnings for possible elaboration problems.
11713 @geindex -gnatw.l (gcc)
11718 @item @code{-gnatw.l}
11720 @emph{List inherited aspects.}
11722 This switch causes the compiler to list inherited invariants,
11723 preconditions, and postconditions from Type_Invariant'Class, Invariant'Class,
11724 Pre'Class, and Post'Class aspects. Also list inherited subtype predicates.
11727 @geindex -gnatw.L (gcc)
11732 @item @code{-gnatw.L}
11734 @emph{Suppress listing of inherited aspects.}
11736 This switch suppresses listing of inherited aspects.
11739 @geindex -gnatwm (gcc)
11744 @item @code{-gnatwm}
11746 @emph{Activate warnings on modified but unreferenced variables.}
11748 This switch activates warnings for variables that are assigned (using
11749 an initialization value or with one or more assignment statements) but
11750 whose value is never read. The warning is suppressed for volatile
11751 variables and also for variables that are renamings of other variables
11752 or for which an address clause is given.
11753 The default is that these warnings are not given.
11756 @geindex -gnatwM (gcc)
11761 @item @code{-gnatwM}
11763 @emph{Disable warnings on modified but unreferenced variables.}
11765 This switch disables warnings for variables that are assigned or
11766 initialized, but never read.
11769 @geindex -gnatw.m (gcc)
11774 @item @code{-gnatw.m}
11776 @emph{Activate warnings on suspicious modulus values.}
11778 This switch activates warnings for modulus values that seem suspicious.
11779 The cases caught are where the size is the same as the modulus (e.g.
11780 a modulus of 7 with a size of 7 bits), and modulus values of 32 or 64
11781 with no size clause. The guess in both cases is that 2**x was intended
11782 rather than x. In addition expressions of the form 2*x for small x
11783 generate a warning (the almost certainly accurate guess being that
11784 2**x was intended). The default is that these warnings are given.
11787 @geindex -gnatw.M (gcc)
11792 @item @code{-gnatw.M}
11794 @emph{Disable warnings on suspicious modulus values.}
11796 This switch disables warnings for suspicious modulus values.
11799 @geindex -gnatwn (gcc)
11804 @item @code{-gnatwn}
11806 @emph{Set normal warnings mode.}
11808 This switch sets normal warning mode, in which enabled warnings are
11809 issued and treated as warnings rather than errors. This is the default
11810 mode. the switch @code{-gnatwn} can be used to cancel the effect of
11811 an explicit @code{-gnatws} or
11812 @code{-gnatwe}. It also cancels the effect of the
11813 implicit @code{-gnatwe} that is activated by the
11814 use of @code{-gnatg}.
11817 @geindex -gnatw.n (gcc)
11819 @geindex Atomic Synchronization
11825 @item @code{-gnatw.n}
11827 @emph{Activate warnings on atomic synchronization.}
11829 This switch actives warnings when an access to an atomic variable
11830 requires the generation of atomic synchronization code. These
11831 warnings are off by default.
11834 @geindex -gnatw.N (gcc)
11839 @item @code{-gnatw.N}
11841 @emph{Suppress warnings on atomic synchronization.}
11843 @geindex Atomic Synchronization
11846 This switch suppresses warnings when an access to an atomic variable
11847 requires the generation of atomic synchronization code.
11850 @geindex -gnatwo (gcc)
11852 @geindex Address Clauses
11858 @item @code{-gnatwo}
11860 @emph{Activate warnings on address clause overlays.}
11862 This switch activates warnings for possibly unintended initialization
11863 effects of defining address clauses that cause one variable to overlap
11864 another. The default is that such warnings are generated.
11867 @geindex -gnatwO (gcc)
11872 @item @code{-gnatwO}
11874 @emph{Suppress warnings on address clause overlays.}
11876 This switch suppresses warnings on possibly unintended initialization
11877 effects of defining address clauses that cause one variable to overlap
11881 @geindex -gnatw.o (gcc)
11886 @item @code{-gnatw.o}
11888 @emph{Activate warnings on modified but unreferenced out parameters.}
11890 This switch activates warnings for variables that are modified by using
11891 them as actuals for a call to a procedure with an out mode formal, where
11892 the resulting assigned value is never read. It is applicable in the case
11893 where there is more than one out mode formal. If there is only one out
11894 mode formal, the warning is issued by default (controlled by -gnatwu).
11895 The warning is suppressed for volatile
11896 variables and also for variables that are renamings of other variables
11897 or for which an address clause is given.
11898 The default is that these warnings are not given.
11901 @geindex -gnatw.O (gcc)
11906 @item @code{-gnatw.O}
11908 @emph{Disable warnings on modified but unreferenced out parameters.}
11910 This switch suppresses warnings for variables that are modified by using
11911 them as actuals for a call to a procedure with an out mode formal, where
11912 the resulting assigned value is never read.
11915 @geindex -gnatwp (gcc)
11923 @item @code{-gnatwp}
11925 @emph{Activate warnings on ineffective pragma Inlines.}
11927 This switch activates warnings for failure of front end inlining
11928 (activated by @code{-gnatN}) to inline a particular call. There are
11929 many reasons for not being able to inline a call, including most
11930 commonly that the call is too complex to inline. The default is
11931 that such warnings are not given.
11932 Warnings on ineffective inlining by the gcc back-end can be activated
11933 separately, using the gcc switch -Winline.
11936 @geindex -gnatwP (gcc)
11941 @item @code{-gnatwP}
11943 @emph{Suppress warnings on ineffective pragma Inlines.}
11945 This switch suppresses warnings on ineffective pragma Inlines. If the
11946 inlining mechanism cannot inline a call, it will simply ignore the
11950 @geindex -gnatw.p (gcc)
11952 @geindex Parameter order
11958 @item @code{-gnatw.p}
11960 @emph{Activate warnings on parameter ordering.}
11962 This switch activates warnings for cases of suspicious parameter
11963 ordering when the list of arguments are all simple identifiers that
11964 match the names of the formals, but are in a different order. The
11965 warning is suppressed if any use of named parameter notation is used,
11966 so this is the appropriate way to suppress a false positive (and
11967 serves to emphasize that the "misordering" is deliberate). The
11968 default is that such warnings are not given.
11971 @geindex -gnatw.P (gcc)
11976 @item @code{-gnatw.P}
11978 @emph{Suppress warnings on parameter ordering.}
11980 This switch suppresses warnings on cases of suspicious parameter
11984 @geindex -gnatwq (gcc)
11986 @geindex Parentheses
11992 @item @code{-gnatwq}
11994 @emph{Activate warnings on questionable missing parentheses.}
11996 This switch activates warnings for cases where parentheses are not used and
11997 the result is potential ambiguity from a readers point of view. For example
11998 (not a > b) when a and b are modular means ((not a) > b) and very likely the
11999 programmer intended (not (a > b)). Similarly (-x mod 5) means (-(x mod 5)) and
12000 quite likely ((-x) mod 5) was intended. In such situations it seems best to
12001 follow the rule of always parenthesizing to make the association clear, and
12002 this warning switch warns if such parentheses are not present. The default
12003 is that these warnings are given.
12006 @geindex -gnatwQ (gcc)
12011 @item @code{-gnatwQ}
12013 @emph{Suppress warnings on questionable missing parentheses.}
12015 This switch suppresses warnings for cases where the association is not
12016 clear and the use of parentheses is preferred.
12019 @geindex -gnatw.q (gcc)
12027 @item @code{-gnatw.q}
12029 @emph{Activate warnings on questionable layout of record types.}
12031 This switch activates warnings for cases where the default layout of
12032 a record type, that is to say the layout of its components in textual
12033 order of the source code, would very likely cause inefficiencies in
12034 the code generated by the compiler, both in terms of space and speed
12035 during execution. One warning is issued for each problematic component
12036 without representation clause in the nonvariant part and then in each
12037 variant recursively, if any.
12039 The purpose of these warnings is neither to prescribe an optimal layout
12040 nor to force the use of representation clauses, but rather to get rid of
12041 the most blatant inefficiencies in the layout. Therefore, the default
12042 layout is matched against the following synthetic ordered layout and
12043 the deviations are flagged on a component-by-component basis:
12049 first all components or groups of components whose length is fixed
12050 and a multiple of the storage unit,
12053 then the remaining components whose length is fixed and not a multiple
12054 of the storage unit,
12057 then the remaining components whose length doesn't depend on discriminants
12058 (that is to say, with variable but uniform length for all objects),
12061 then all components whose length depends on discriminants,
12064 finally the variant part (if any),
12067 for the nonvariant part and for each variant recursively, if any.
12069 The exact wording of the warning depends on whether the compiler is allowed
12070 to reorder the components in the record type or precluded from doing it by
12071 means of pragma @code{No_Component_Reordering}.
12073 The default is that these warnings are not given.
12076 @geindex -gnatw.Q (gcc)
12081 @item @code{-gnatw.Q}
12083 @emph{Suppress warnings on questionable layout of record types.}
12085 This switch suppresses warnings for cases where the default layout of
12086 a record type would very likely cause inefficiencies.
12089 @geindex -gnatwr (gcc)
12094 @item @code{-gnatwr}
12096 @emph{Activate warnings on redundant constructs.}
12098 This switch activates warnings for redundant constructs. The following
12099 is the current list of constructs regarded as redundant:
12105 Assignment of an item to itself.
12108 Type conversion that converts an expression to its own type.
12111 Use of the attribute @code{Base} where @code{typ'Base} is the same
12115 Use of pragma @code{Pack} when all components are placed by a record
12116 representation clause.
12119 Exception handler containing only a reraise statement (raise with no
12120 operand) which has no effect.
12123 Use of the operator abs on an operand that is known at compile time
12127 Comparison of an object or (unary or binary) operation of boolean type to
12128 an explicit True value.
12131 The default is that warnings for redundant constructs are not given.
12134 @geindex -gnatwR (gcc)
12139 @item @code{-gnatwR}
12141 @emph{Suppress warnings on redundant constructs.}
12143 This switch suppresses warnings for redundant constructs.
12146 @geindex -gnatw.r (gcc)
12151 @item @code{-gnatw.r}
12153 @emph{Activate warnings for object renaming function.}
12155 This switch activates warnings for an object renaming that renames a
12156 function call, which is equivalent to a constant declaration (as
12157 opposed to renaming the function itself). The default is that these
12158 warnings are given.
12161 @geindex -gnatwT (gcc)
12166 @item @code{-gnatw.R}
12168 @emph{Suppress warnings for object renaming function.}
12170 This switch suppresses warnings for object renaming function.
12173 @geindex -gnatws (gcc)
12178 @item @code{-gnatws}
12180 @emph{Suppress all warnings.}
12182 This switch completely suppresses the
12183 output of all warning messages from the GNAT front end, including
12184 both warnings that can be controlled by switches described in this
12185 section, and those that are normally given unconditionally. The
12186 effect of this suppress action can only be cancelled by a subsequent
12187 use of the switch @code{-gnatwn}.
12189 Note that switch @code{-gnatws} does not suppress
12190 warnings from the @code{gcc} back end.
12191 To suppress these back end warnings as well, use the switch @code{-w}
12192 in addition to @code{-gnatws}. Also this switch has no effect on the
12193 handling of style check messages.
12196 @geindex -gnatw.s (gcc)
12198 @geindex Record Representation (component sizes)
12203 @item @code{-gnatw.s}
12205 @emph{Activate warnings on overridden size clauses.}
12207 This switch activates warnings on component clauses in record
12208 representation clauses where the length given overrides that
12209 specified by an explicit size clause for the component type. A
12210 warning is similarly given in the array case if a specified
12211 component size overrides an explicit size clause for the array
12215 @geindex -gnatw.S (gcc)
12220 @item @code{-gnatw.S}
12222 @emph{Suppress warnings on overridden size clauses.}
12224 This switch suppresses warnings on component clauses in record
12225 representation clauses that override size clauses, and similar
12226 warnings when an array component size overrides a size clause.
12229 @geindex -gnatwt (gcc)
12231 @geindex Deactivated code
12234 @geindex Deleted code
12240 @item @code{-gnatwt}
12242 @emph{Activate warnings for tracking of deleted conditional code.}
12244 This switch activates warnings for tracking of code in conditionals (IF and
12245 CASE statements) that is detected to be dead code which cannot be executed, and
12246 which is removed by the front end. This warning is off by default. This may be
12247 useful for detecting deactivated code in certified applications.
12250 @geindex -gnatwT (gcc)
12255 @item @code{-gnatwT}
12257 @emph{Suppress warnings for tracking of deleted conditional code.}
12259 This switch suppresses warnings for tracking of deleted conditional code.
12262 @geindex -gnatw.t (gcc)
12267 @item @code{-gnatw.t}
12269 @emph{Activate warnings on suspicious contracts.}
12271 This switch activates warnings on suspicious contracts. This includes
12272 warnings on suspicious postconditions (whether a pragma @code{Postcondition} or a
12273 @code{Post} aspect in Ada 2012) and suspicious contract cases (pragma or aspect
12274 @code{Contract_Cases}). A function postcondition or contract case is suspicious
12275 when no postcondition or contract case for this function mentions the result
12276 of the function. A procedure postcondition or contract case is suspicious
12277 when it only refers to the pre-state of the procedure, because in that case
12278 it should rather be expressed as a precondition. This switch also controls
12279 warnings on suspicious cases of expressions typically found in contracts like
12280 quantified expressions and uses of Update attribute. The default is that such
12281 warnings are generated.
12284 @geindex -gnatw.T (gcc)
12289 @item @code{-gnatw.T}
12291 @emph{Suppress warnings on suspicious contracts.}
12293 This switch suppresses warnings on suspicious contracts.
12296 @geindex -gnatwu (gcc)
12301 @item @code{-gnatwu}
12303 @emph{Activate warnings on unused entities.}
12305 This switch activates warnings to be generated for entities that
12306 are declared but not referenced, and for units that are @emph{with}ed
12308 referenced. In the case of packages, a warning is also generated if
12309 no entities in the package are referenced. This means that if a with'ed
12310 package is referenced but the only references are in @code{use}
12311 clauses or @code{renames}
12312 declarations, a warning is still generated. A warning is also generated
12313 for a generic package that is @emph{with}ed but never instantiated.
12314 In the case where a package or subprogram body is compiled, and there
12315 is a @emph{with} on the corresponding spec
12316 that is only referenced in the body,
12317 a warning is also generated, noting that the
12318 @emph{with} can be moved to the body. The default is that
12319 such warnings are not generated.
12320 This switch also activates warnings on unreferenced formals
12321 (it includes the effect of @code{-gnatwf}).
12324 @geindex -gnatwU (gcc)
12329 @item @code{-gnatwU}
12331 @emph{Suppress warnings on unused entities.}
12333 This switch suppresses warnings for unused entities and packages.
12334 It also turns off warnings on unreferenced formals (and thus includes
12335 the effect of @code{-gnatwF}).
12338 @geindex -gnatw.u (gcc)
12343 @item @code{-gnatw.u}
12345 @emph{Activate warnings on unordered enumeration types.}
12347 This switch causes enumeration types to be considered as conceptually
12348 unordered, unless an explicit pragma @code{Ordered} is given for the type.
12349 The effect is to generate warnings in clients that use explicit comparisons
12350 or subranges, since these constructs both treat objects of the type as
12351 ordered. (A @emph{client} is defined as a unit that is other than the unit in
12352 which the type is declared, or its body or subunits.) Please refer to
12353 the description of pragma @code{Ordered} in the
12354 @cite{GNAT Reference Manual} for further details.
12355 The default is that such warnings are not generated.
12358 @geindex -gnatw.U (gcc)
12363 @item @code{-gnatw.U}
12365 @emph{Deactivate warnings on unordered enumeration types.}
12367 This switch causes all enumeration types to be considered as ordered, so
12368 that no warnings are given for comparisons or subranges for any type.
12371 @geindex -gnatwv (gcc)
12373 @geindex Unassigned variable warnings
12378 @item @code{-gnatwv}
12380 @emph{Activate warnings on unassigned variables.}
12382 This switch activates warnings for access to variables which
12383 may not be properly initialized. The default is that
12384 such warnings are generated.
12387 @geindex -gnatwV (gcc)
12392 @item @code{-gnatwV}
12394 @emph{Suppress warnings on unassigned variables.}
12396 This switch suppresses warnings for access to variables which
12397 may not be properly initialized.
12398 For variables of a composite type, the warning can also be suppressed in
12399 Ada 2005 by using a default initialization with a box. For example, if
12400 Table is an array of records whose components are only partially uninitialized,
12401 then the following code:
12404 Tab : Table := (others => <>);
12407 will suppress warnings on subsequent statements that access components
12411 @geindex -gnatw.v (gcc)
12413 @geindex bit order warnings
12418 @item @code{-gnatw.v}
12420 @emph{Activate info messages for non-default bit order.}
12422 This switch activates messages (labeled "info", they are not warnings,
12423 just informational messages) about the effects of non-default bit-order
12424 on records to which a component clause is applied. The effect of specifying
12425 non-default bit ordering is a bit subtle (and changed with Ada 2005), so
12426 these messages, which are given by default, are useful in understanding the
12427 exact consequences of using this feature.
12430 @geindex -gnatw.V (gcc)
12435 @item @code{-gnatw.V}
12437 @emph{Suppress info messages for non-default bit order.}
12439 This switch suppresses information messages for the effects of specifying
12440 non-default bit order on record components with component clauses.
12443 @geindex -gnatww (gcc)
12445 @geindex String indexing warnings
12450 @item @code{-gnatww}
12452 @emph{Activate warnings on wrong low bound assumption.}
12454 This switch activates warnings for indexing an unconstrained string parameter
12455 with a literal or S'Length. This is a case where the code is assuming that the
12456 low bound is one, which is in general not true (for example when a slice is
12457 passed). The default is that such warnings are generated.
12460 @geindex -gnatwW (gcc)
12465 @item @code{-gnatwW}
12467 @emph{Suppress warnings on wrong low bound assumption.}
12469 This switch suppresses warnings for indexing an unconstrained string parameter
12470 with a literal or S'Length. Note that this warning can also be suppressed
12471 in a particular case by adding an assertion that the lower bound is 1,
12472 as shown in the following example:
12475 procedure K (S : String) is
12476 pragma Assert (S'First = 1);
12481 @geindex -gnatw.w (gcc)
12483 @geindex Warnings Off control
12488 @item @code{-gnatw.w}
12490 @emph{Activate warnings on Warnings Off pragmas.}
12492 This switch activates warnings for use of @code{pragma Warnings (Off, entity)}
12493 where either the pragma is entirely useless (because it suppresses no
12494 warnings), or it could be replaced by @code{pragma Unreferenced} or
12495 @code{pragma Unmodified}.
12496 Also activates warnings for the case of
12497 Warnings (Off, String), where either there is no matching
12498 Warnings (On, String), or the Warnings (Off) did not suppress any warning.
12499 The default is that these warnings are not given.
12502 @geindex -gnatw.W (gcc)
12507 @item @code{-gnatw.W}
12509 @emph{Suppress warnings on unnecessary Warnings Off pragmas.}
12511 This switch suppresses warnings for use of @code{pragma Warnings (Off, ...)}.
12514 @geindex -gnatwx (gcc)
12516 @geindex Export/Import pragma warnings
12521 @item @code{-gnatwx}
12523 @emph{Activate warnings on Export/Import pragmas.}
12525 This switch activates warnings on Export/Import pragmas when
12526 the compiler detects a possible conflict between the Ada and
12527 foreign language calling sequences. For example, the use of
12528 default parameters in a convention C procedure is dubious
12529 because the C compiler cannot supply the proper default, so
12530 a warning is issued. The default is that such warnings are
12534 @geindex -gnatwX (gcc)
12539 @item @code{-gnatwX}
12541 @emph{Suppress warnings on Export/Import pragmas.}
12543 This switch suppresses warnings on Export/Import pragmas.
12544 The sense of this is that you are telling the compiler that
12545 you know what you are doing in writing the pragma, and it
12546 should not complain at you.
12549 @geindex -gnatwm (gcc)
12554 @item @code{-gnatw.x}
12556 @emph{Activate warnings for No_Exception_Propagation mode.}
12558 This switch activates warnings for exception usage when pragma Restrictions
12559 (No_Exception_Propagation) is in effect. Warnings are given for implicit or
12560 explicit exception raises which are not covered by a local handler, and for
12561 exception handlers which do not cover a local raise. The default is that
12562 these warnings are given for units that contain exception handlers.
12564 @item @code{-gnatw.X}
12566 @emph{Disable warnings for No_Exception_Propagation mode.}
12568 This switch disables warnings for exception usage when pragma Restrictions
12569 (No_Exception_Propagation) is in effect.
12572 @geindex -gnatwy (gcc)
12574 @geindex Ada compatibility issues warnings
12579 @item @code{-gnatwy}
12581 @emph{Activate warnings for Ada compatibility issues.}
12583 For the most part, newer versions of Ada are upwards compatible
12584 with older versions. For example, Ada 2005 programs will almost
12585 always work when compiled as Ada 2012.
12586 However there are some exceptions (for example the fact that
12587 @code{some} is now a reserved word in Ada 2012). This
12588 switch activates several warnings to help in identifying
12589 and correcting such incompatibilities. The default is that
12590 these warnings are generated. Note that at one point Ada 2005
12591 was called Ada 0Y, hence the choice of character.
12594 @geindex -gnatwY (gcc)
12596 @geindex Ada compatibility issues warnings
12601 @item @code{-gnatwY}
12603 @emph{Disable warnings for Ada compatibility issues.}
12605 This switch suppresses the warnings intended to help in identifying
12606 incompatibilities between Ada language versions.
12609 @geindex -gnatw.y (gcc)
12611 @geindex Package spec needing body
12616 @item @code{-gnatw.y}
12618 @emph{Activate information messages for why package spec needs body.}
12620 There are a number of cases in which a package spec needs a body.
12621 For example, the use of pragma Elaborate_Body, or the declaration
12622 of a procedure specification requiring a completion. This switch
12623 causes information messages to be output showing why a package
12624 specification requires a body. This can be useful in the case of
12625 a large package specification which is unexpectedly requiring a
12626 body. The default is that such information messages are not output.
12629 @geindex -gnatw.Y (gcc)
12631 @geindex No information messages for why package spec needs body
12636 @item @code{-gnatw.Y}
12638 @emph{Disable information messages for why package spec needs body.}
12640 This switch suppresses the output of information messages showing why
12641 a package specification needs a body.
12644 @geindex -gnatwz (gcc)
12646 @geindex Unchecked_Conversion warnings
12651 @item @code{-gnatwz}
12653 @emph{Activate warnings on unchecked conversions.}
12655 This switch activates warnings for unchecked conversions
12656 where the types are known at compile time to have different
12657 sizes. The default is that such warnings are generated. Warnings are also
12658 generated for subprogram pointers with different conventions.
12661 @geindex -gnatwZ (gcc)
12666 @item @code{-gnatwZ}
12668 @emph{Suppress warnings on unchecked conversions.}
12670 This switch suppresses warnings for unchecked conversions
12671 where the types are known at compile time to have different
12672 sizes or conventions.
12675 @geindex -gnatw.z (gcc)
12677 @geindex Size/Alignment warnings
12682 @item @code{-gnatw.z}
12684 @emph{Activate warnings for size not a multiple of alignment.}
12686 This switch activates warnings for cases of record types with
12687 specified @code{Size} and @code{Alignment} attributes where the
12688 size is not a multiple of the alignment, resulting in an object
12689 size that is greater than the specified size. The default
12690 is that such warnings are generated.
12693 @geindex -gnatw.Z (gcc)
12695 @geindex Size/Alignment warnings
12700 @item @code{-gnatw.Z}
12702 @emph{Suppress warnings for size not a multiple of alignment.}
12704 This switch suppresses warnings for cases of record types with
12705 specified @code{Size} and @code{Alignment} attributes where the
12706 size is not a multiple of the alignment, resulting in an object
12707 size that is greater than the specified size.
12708 The warning can also be
12709 suppressed by giving an explicit @code{Object_Size} value.
12712 @geindex -Wunused (gcc)
12717 @item @code{-Wunused}
12719 The warnings controlled by the @code{-gnatw} switch are generated by
12720 the front end of the compiler. The GCC back end can provide
12721 additional warnings and they are controlled by the @code{-W} switch.
12722 For example, @code{-Wunused} activates back end
12723 warnings for entities that are declared but not referenced.
12726 @geindex -Wuninitialized (gcc)
12731 @item @code{-Wuninitialized}
12733 Similarly, @code{-Wuninitialized} activates
12734 the back end warning for uninitialized variables. This switch must be
12735 used in conjunction with an optimization level greater than zero.
12738 @geindex -Wstack-usage (gcc)
12743 @item @code{-Wstack-usage=@emph{len}}
12745 Warn if the stack usage of a subprogram might be larger than @code{len} bytes.
12746 See @ref{f5,,Static Stack Usage Analysis} for details.
12749 @geindex -Wall (gcc)
12756 This switch enables most warnings from the GCC back end.
12757 The code generator detects a number of warning situations that are missed
12758 by the GNAT front end, and this switch can be used to activate them.
12759 The use of this switch also sets the default front end warning mode to
12760 @code{-gnatwa}, that is, most front end warnings activated as well.
12770 Conversely, this switch suppresses warnings from the GCC back end.
12771 The use of this switch also sets the default front end warning mode to
12772 @code{-gnatws}, that is, front end warnings suppressed as well.
12775 @geindex -Werror (gcc)
12780 @item @code{-Werror}
12782 This switch causes warnings from the GCC back end to be treated as
12783 errors. The warning string still appears, but the warning messages are
12784 counted as errors, and prevent the generation of an object file.
12787 A string of warning parameters can be used in the same parameter. For example:
12793 will turn on all optional warnings except for unrecognized pragma warnings,
12794 and also specify that warnings should be treated as errors.
12796 When no switch @code{-gnatw} is used, this is equivalent to:
12943 @node Debugging and Assertion Control,Validity Checking,Warning Message Control,Compiler Switches
12944 @anchor{gnat_ugn/building_executable_programs_with_gnat debugging-and-assertion-control}@anchor{100}@anchor{gnat_ugn/building_executable_programs_with_gnat id16}@anchor{101}
12945 @subsection Debugging and Assertion Control
12948 @geindex -gnata (gcc)
12953 @item @code{-gnata}
12959 @geindex Assertions
12961 @geindex Precondition
12963 @geindex Postcondition
12965 @geindex Type invariants
12967 @geindex Subtype predicates
12969 The @code{-gnata} option is equivalent to the following @code{Assertion_Policy} pragma:
12972 pragma Assertion_Policy (Check);
12975 Which is a shorthand for:
12978 pragma Assertion_Policy
12980 Static_Predicate => Check,
12981 Dynamic_Predicate => Check,
12983 Pre'Class => Check,
12985 Post'Class => Check,
12986 Type_Invariant => Check,
12987 Type_Invariant'Class => Check);
12990 The pragmas @code{Assert} and @code{Debug} normally have no effect and
12991 are ignored. This switch, where @code{a} stands for 'assert', causes
12992 pragmas @code{Assert} and @code{Debug} to be activated. This switch also
12993 causes preconditions, postconditions, subtype predicates, and
12994 type invariants to be activated.
12996 The pragmas have the form:
12999 pragma Assert (<Boolean-expression> [, <static-string-expression>])
13000 pragma Debug (<procedure call>)
13001 pragma Type_Invariant (<type-local-name>, <Boolean-expression>)
13002 pragma Predicate (<type-local-name>, <Boolean-expression>)
13003 pragma Precondition (<Boolean-expression>, <string-expression>)
13004 pragma Postcondition (<Boolean-expression>, <string-expression>)
13007 The aspects have the form:
13010 with [Pre|Post|Type_Invariant|Dynamic_Predicate|Static_Predicate]
13011 => <Boolean-expression>;
13014 The @code{Assert} pragma causes @code{Boolean-expression} to be tested.
13015 If the result is @code{True}, the pragma has no effect (other than
13016 possible side effects from evaluating the expression). If the result is
13017 @code{False}, the exception @code{Assert_Failure} declared in the package
13018 @code{System.Assertions} is raised (passing @code{static-string-expression}, if
13019 present, as the message associated with the exception). If no string
13020 expression is given, the default is a string containing the file name and
13021 line number of the pragma.
13023 The @code{Debug} pragma causes @code{procedure} to be called. Note that
13024 @code{pragma Debug} may appear within a declaration sequence, allowing
13025 debugging procedures to be called between declarations.
13027 For the aspect specification, the @code{Boolean-expression} is evaluated.
13028 If the result is @code{True}, the aspect has no effect. If the result
13029 is @code{False}, the exception @code{Assert_Failure} is raised.
13032 @node Validity Checking,Style Checking,Debugging and Assertion Control,Compiler Switches
13033 @anchor{gnat_ugn/building_executable_programs_with_gnat validity-checking}@anchor{f6}@anchor{gnat_ugn/building_executable_programs_with_gnat id17}@anchor{102}
13034 @subsection Validity Checking
13037 @geindex Validity Checking
13039 The Ada Reference Manual defines the concept of invalid values (see
13040 RM 13.9.1). The primary source of invalid values is uninitialized
13041 variables. A scalar variable that is left uninitialized may contain
13042 an invalid value; the concept of invalid does not apply to access or
13045 It is an error to read an invalid value, but the RM does not require
13046 run-time checks to detect such errors, except for some minimal
13047 checking to prevent erroneous execution (i.e. unpredictable
13048 behavior). This corresponds to the @code{-gnatVd} switch below,
13049 which is the default. For example, by default, if the expression of a
13050 case statement is invalid, it will raise Constraint_Error rather than
13051 causing a wild jump, and if an array index on the left-hand side of an
13052 assignment is invalid, it will raise Constraint_Error rather than
13053 overwriting an arbitrary memory location.
13055 The @code{-gnatVa} may be used to enable additional validity checks,
13056 which are not required by the RM. These checks are often very
13057 expensive (which is why the RM does not require them). These checks
13058 are useful in tracking down uninitialized variables, but they are
13059 not usually recommended for production builds, and in particular
13060 we do not recommend using these extra validity checking options in
13061 combination with optimization, since this can confuse the optimizer.
13062 If performance is a consideration, leading to the need to optimize,
13063 then the validity checking options should not be used.
13065 The other @code{-gnatV@emph{x}} switches below allow finer-grained
13066 control; you can enable whichever validity checks you desire. However,
13067 for most debugging purposes, @code{-gnatVa} is sufficient, and the
13068 default @code{-gnatVd} (i.e. standard Ada behavior) is usually
13069 sufficient for non-debugging use.
13071 The @code{-gnatB} switch tells the compiler to assume that all
13072 values are valid (that is, within their declared subtype range)
13073 except in the context of a use of the Valid attribute. This means
13074 the compiler can generate more efficient code, since the range
13075 of values is better known at compile time. However, an uninitialized
13076 variable can cause wild jumps and memory corruption in this mode.
13078 The @code{-gnatV@emph{x}} switch allows control over the validity
13079 checking mode as described below.
13080 The @code{x} argument is a string of letters that
13081 indicate validity checks that are performed or not performed in addition
13082 to the default checks required by Ada as described above.
13084 @geindex -gnatVa (gcc)
13089 @item @code{-gnatVa}
13091 @emph{All validity checks.}
13093 All validity checks are turned on.
13094 That is, @code{-gnatVa} is
13095 equivalent to @code{gnatVcdfimorst}.
13098 @geindex -gnatVc (gcc)
13103 @item @code{-gnatVc}
13105 @emph{Validity checks for copies.}
13107 The right hand side of assignments, and the initializing values of
13108 object declarations are validity checked.
13111 @geindex -gnatVd (gcc)
13116 @item @code{-gnatVd}
13118 @emph{Default (RM) validity checks.}
13120 Some validity checks are done by default following normal Ada semantics
13121 (RM 13.9.1 (9-11)).
13122 A check is done in case statements that the expression is within the range
13123 of the subtype. If it is not, Constraint_Error is raised.
13124 For assignments to array components, a check is done that the expression used
13125 as index is within the range. If it is not, Constraint_Error is raised.
13126 Both these validity checks may be turned off using switch @code{-gnatVD}.
13127 They are turned on by default. If @code{-gnatVD} is specified, a subsequent
13128 switch @code{-gnatVd} will leave the checks turned on.
13129 Switch @code{-gnatVD} should be used only if you are sure that all such
13130 expressions have valid values. If you use this switch and invalid values
13131 are present, then the program is erroneous, and wild jumps or memory
13132 overwriting may occur.
13135 @geindex -gnatVe (gcc)
13140 @item @code{-gnatVe}
13142 @emph{Validity checks for elementary components.}
13144 In the absence of this switch, assignments to record or array components are
13145 not validity checked, even if validity checks for assignments generally
13146 (@code{-gnatVc}) are turned on. In Ada, assignment of composite values do not
13147 require valid data, but assignment of individual components does. So for
13148 example, there is a difference between copying the elements of an array with a
13149 slice assignment, compared to assigning element by element in a loop. This
13150 switch allows you to turn off validity checking for components, even when they
13151 are assigned component by component.
13154 @geindex -gnatVf (gcc)
13159 @item @code{-gnatVf}
13161 @emph{Validity checks for floating-point values.}
13163 In the absence of this switch, validity checking occurs only for discrete
13164 values. If @code{-gnatVf} is specified, then validity checking also applies
13165 for floating-point values, and NaNs and infinities are considered invalid,
13166 as well as out of range values for constrained types. Note that this means
13167 that standard IEEE infinity mode is not allowed. The exact contexts
13168 in which floating-point values are checked depends on the setting of other
13169 options. For example, @code{-gnatVif} or @code{-gnatVfi}
13170 (the order does not matter) specifies that floating-point parameters of mode
13171 @code{in} should be validity checked.
13174 @geindex -gnatVi (gcc)
13179 @item @code{-gnatVi}
13181 @emph{Validity checks for `@w{`}in`@w{`} mode parameters.}
13183 Arguments for parameters of mode @code{in} are validity checked in function
13184 and procedure calls at the point of call.
13187 @geindex -gnatVm (gcc)
13192 @item @code{-gnatVm}
13194 @emph{Validity checks for `@w{`}in out`@w{`} mode parameters.}
13196 Arguments for parameters of mode @code{in out} are validity checked in
13197 procedure calls at the point of call. The @code{'m'} here stands for
13198 modify, since this concerns parameters that can be modified by the call.
13199 Note that there is no specific option to test @code{out} parameters,
13200 but any reference within the subprogram will be tested in the usual
13201 manner, and if an invalid value is copied back, any reference to it
13202 will be subject to validity checking.
13205 @geindex -gnatVn (gcc)
13210 @item @code{-gnatVn}
13212 @emph{No validity checks.}
13214 This switch turns off all validity checking, including the default checking
13215 for case statements and left hand side subscripts. Note that the use of
13216 the switch @code{-gnatp} suppresses all run-time checks, including
13217 validity checks, and thus implies @code{-gnatVn}. When this switch
13218 is used, it cancels any other @code{-gnatV} previously issued.
13221 @geindex -gnatVo (gcc)
13226 @item @code{-gnatVo}
13228 @emph{Validity checks for operator and attribute operands.}
13230 Arguments for predefined operators and attributes are validity checked.
13231 This includes all operators in package @code{Standard},
13232 the shift operators defined as intrinsic in package @code{Interfaces}
13233 and operands for attributes such as @code{Pos}. Checks are also made
13234 on individual component values for composite comparisons, and on the
13235 expressions in type conversions and qualified expressions. Checks are
13236 also made on explicit ranges using @code{..} (e.g., slices, loops etc).
13239 @geindex -gnatVp (gcc)
13244 @item @code{-gnatVp}
13246 @emph{Validity checks for parameters.}
13248 This controls the treatment of parameters within a subprogram (as opposed
13249 to @code{-gnatVi} and @code{-gnatVm} which control validity testing
13250 of parameters on a call. If either of these call options is used, then
13251 normally an assumption is made within a subprogram that the input arguments
13252 have been validity checking at the point of call, and do not need checking
13253 again within a subprogram). If @code{-gnatVp} is set, then this assumption
13254 is not made, and parameters are not assumed to be valid, so their validity
13255 will be checked (or rechecked) within the subprogram.
13258 @geindex -gnatVr (gcc)
13263 @item @code{-gnatVr}
13265 @emph{Validity checks for function returns.}
13267 The expression in @code{return} statements in functions is validity
13271 @geindex -gnatVs (gcc)
13276 @item @code{-gnatVs}
13278 @emph{Validity checks for subscripts.}
13280 All subscripts expressions are checked for validity, whether they appear
13281 on the right side or left side (in default mode only left side subscripts
13282 are validity checked).
13285 @geindex -gnatVt (gcc)
13290 @item @code{-gnatVt}
13292 @emph{Validity checks for tests.}
13294 Expressions used as conditions in @code{if}, @code{while} or @code{exit}
13295 statements are checked, as well as guard expressions in entry calls.
13298 The @code{-gnatV} switch may be followed by a string of letters
13299 to turn on a series of validity checking options.
13300 For example, @code{-gnatVcr}
13301 specifies that in addition to the default validity checking, copies and
13302 function return expressions are to be validity checked.
13303 In order to make it easier to specify the desired combination of effects,
13304 the upper case letters @code{CDFIMORST} may
13305 be used to turn off the corresponding lower case option.
13306 Thus @code{-gnatVaM} turns on all validity checking options except for
13307 checking of @code{in out} parameters.
13309 The specification of additional validity checking generates extra code (and
13310 in the case of @code{-gnatVa} the code expansion can be substantial).
13311 However, these additional checks can be very useful in detecting
13312 uninitialized variables, incorrect use of unchecked conversion, and other
13313 errors leading to invalid values. The use of pragma @code{Initialize_Scalars}
13314 is useful in conjunction with the extra validity checking, since this
13315 ensures that wherever possible uninitialized variables have invalid values.
13317 See also the pragma @code{Validity_Checks} which allows modification of
13318 the validity checking mode at the program source level, and also allows for
13319 temporary disabling of validity checks.
13321 @node Style Checking,Run-Time Checks,Validity Checking,Compiler Switches
13322 @anchor{gnat_ugn/building_executable_programs_with_gnat id18}@anchor{103}@anchor{gnat_ugn/building_executable_programs_with_gnat style-checking}@anchor{fb}
13323 @subsection Style Checking
13326 @geindex Style checking
13328 @geindex -gnaty (gcc)
13330 The @code{-gnatyx} switch causes the compiler to
13331 enforce specified style rules. A limited set of style rules has been used
13332 in writing the GNAT sources themselves. This switch allows user programs
13333 to activate all or some of these checks. If the source program fails a
13334 specified style check, an appropriate message is given, preceded by
13335 the character sequence '(style)'. This message does not prevent
13336 successful compilation (unless the @code{-gnatwe} switch is used).
13338 Note that this is by no means intended to be a general facility for
13339 checking arbitrary coding standards. It is simply an embedding of the
13340 style rules we have chosen for the GNAT sources. If you are starting
13341 a project which does not have established style standards, you may
13342 find it useful to adopt the entire set of GNAT coding standards, or
13343 some subset of them.
13346 The string @code{x} is a sequence of letters or digits
13347 indicating the particular style
13348 checks to be performed. The following checks are defined:
13350 @geindex -gnaty[0-9] (gcc)
13355 @item @code{-gnaty0}
13357 @emph{Specify indentation level.}
13359 If a digit from 1-9 appears
13360 in the string after @code{-gnaty}
13361 then proper indentation is checked, with the digit indicating the
13362 indentation level required. A value of zero turns off this style check.
13363 The general style of required indentation is as specified by
13364 the examples in the Ada Reference Manual. Full line comments must be
13365 aligned with the @code{--} starting on a column that is a multiple of
13366 the alignment level, or they may be aligned the same way as the following
13367 non-blank line (this is useful when full line comments appear in the middle
13368 of a statement, or they may be aligned with the source line on the previous
13372 @geindex -gnatya (gcc)
13377 @item @code{-gnatya}
13379 @emph{Check attribute casing.}
13381 Attribute names, including the case of keywords such as @code{digits}
13382 used as attributes names, must be written in mixed case, that is, the
13383 initial letter and any letter following an underscore must be uppercase.
13384 All other letters must be lowercase.
13387 @geindex -gnatyA (gcc)
13392 @item @code{-gnatyA}
13394 @emph{Use of array index numbers in array attributes.}
13396 When using the array attributes First, Last, Range,
13397 or Length, the index number must be omitted for one-dimensional arrays
13398 and is required for multi-dimensional arrays.
13401 @geindex -gnatyb (gcc)
13406 @item @code{-gnatyb}
13408 @emph{Blanks not allowed at statement end.}
13410 Trailing blanks are not allowed at the end of statements. The purpose of this
13411 rule, together with h (no horizontal tabs), is to enforce a canonical format
13412 for the use of blanks to separate source tokens.
13415 @geindex -gnatyB (gcc)
13420 @item @code{-gnatyB}
13422 @emph{Check Boolean operators.}
13424 The use of AND/OR operators is not permitted except in the cases of modular
13425 operands, array operands, and simple stand-alone boolean variables or
13426 boolean constants. In all other cases @code{and then}/@cite{or else} are
13430 @geindex -gnatyc (gcc)
13435 @item @code{-gnatyc}
13437 @emph{Check comments, double space.}
13439 Comments must meet the following set of rules:
13445 The @code{--} that starts the column must either start in column one,
13446 or else at least one blank must precede this sequence.
13449 Comments that follow other tokens on a line must have at least one blank
13450 following the @code{--} at the start of the comment.
13453 Full line comments must have at least two blanks following the
13454 @code{--} that starts the comment, with the following exceptions.
13457 A line consisting only of the @code{--} characters, possibly preceded
13458 by blanks is permitted.
13461 A comment starting with @code{--x} where @code{x} is a special character
13463 This allows proper processing of the output from specialized tools
13464 such as @code{gnatprep} (where @code{--!} is used) and in earlier versions of the SPARK
13466 language (where @code{--#} is used). For the purposes of this rule, a
13467 special character is defined as being in one of the ASCII ranges
13468 @code{16#21#...16#2F#} or @code{16#3A#...16#3F#}.
13469 Note that this usage is not permitted
13470 in GNAT implementation units (i.e., when @code{-gnatg} is used).
13473 A line consisting entirely of minus signs, possibly preceded by blanks, is
13474 permitted. This allows the construction of box comments where lines of minus
13475 signs are used to form the top and bottom of the box.
13478 A comment that starts and ends with @code{--} is permitted as long as at
13479 least one blank follows the initial @code{--}. Together with the preceding
13480 rule, this allows the construction of box comments, as shown in the following
13484 ---------------------------
13485 -- This is a box comment --
13486 -- with two text lines. --
13487 ---------------------------
13492 @geindex -gnatyC (gcc)
13497 @item @code{-gnatyC}
13499 @emph{Check comments, single space.}
13501 This is identical to @code{c} except that only one space
13502 is required following the @code{--} of a comment instead of two.
13505 @geindex -gnatyd (gcc)
13510 @item @code{-gnatyd}
13512 @emph{Check no DOS line terminators present.}
13514 All lines must be terminated by a single ASCII.LF
13515 character (in particular the DOS line terminator sequence CR/LF is not
13519 @geindex -gnatye (gcc)
13524 @item @code{-gnatye}
13526 @emph{Check end/exit labels.}
13528 Optional labels on @code{end} statements ending subprograms and on
13529 @code{exit} statements exiting named loops, are required to be present.
13532 @geindex -gnatyf (gcc)
13537 @item @code{-gnatyf}
13539 @emph{No form feeds or vertical tabs.}
13541 Neither form feeds nor vertical tab characters are permitted
13542 in the source text.
13545 @geindex -gnatyg (gcc)
13550 @item @code{-gnatyg}
13552 @emph{GNAT style mode.}
13554 The set of style check switches is set to match that used by the GNAT sources.
13555 This may be useful when developing code that is eventually intended to be
13556 incorporated into GNAT. Currently this is equivalent to @code{-gnatwydISux})
13557 but additional style switches may be added to this set in the future without
13561 @geindex -gnatyh (gcc)
13566 @item @code{-gnatyh}
13568 @emph{No horizontal tabs.}
13570 Horizontal tab characters are not permitted in the source text.
13571 Together with the b (no blanks at end of line) check, this
13572 enforces a canonical form for the use of blanks to separate
13576 @geindex -gnatyi (gcc)
13581 @item @code{-gnatyi}
13583 @emph{Check if-then layout.}
13585 The keyword @code{then} must appear either on the same
13586 line as corresponding @code{if}, or on a line on its own, lined
13587 up under the @code{if}.
13590 @geindex -gnatyI (gcc)
13595 @item @code{-gnatyI}
13597 @emph{check mode IN keywords.}
13599 Mode @code{in} (the default mode) is not
13600 allowed to be given explicitly. @code{in out} is fine,
13601 but not @code{in} on its own.
13604 @geindex -gnatyk (gcc)
13609 @item @code{-gnatyk}
13611 @emph{Check keyword casing.}
13613 All keywords must be in lower case (with the exception of keywords
13614 such as @code{digits} used as attribute names to which this check
13618 @geindex -gnatyl (gcc)
13623 @item @code{-gnatyl}
13625 @emph{Check layout.}
13627 Layout of statement and declaration constructs must follow the
13628 recommendations in the Ada Reference Manual, as indicated by the
13629 form of the syntax rules. For example an @code{else} keyword must
13630 be lined up with the corresponding @code{if} keyword.
13632 There are two respects in which the style rule enforced by this check
13633 option are more liberal than those in the Ada Reference Manual. First
13634 in the case of record declarations, it is permissible to put the
13635 @code{record} keyword on the same line as the @code{type} keyword, and
13636 then the @code{end} in @code{end record} must line up under @code{type}.
13637 This is also permitted when the type declaration is split on two lines.
13638 For example, any of the following three layouts is acceptable:
13659 Second, in the case of a block statement, a permitted alternative
13660 is to put the block label on the same line as the @code{declare} or
13661 @code{begin} keyword, and then line the @code{end} keyword up under
13662 the block label. For example both the following are permitted:
13679 The same alternative format is allowed for loops. For example, both of
13680 the following are permitted:
13683 Clear : while J < 10 loop
13694 @geindex -gnatyLnnn (gcc)
13699 @item @code{-gnatyL}
13701 @emph{Set maximum nesting level.}
13703 The maximum level of nesting of constructs (including subprograms, loops,
13704 blocks, packages, and conditionals) may not exceed the given value
13705 @emph{nnn}. A value of zero disconnects this style check.
13708 @geindex -gnatym (gcc)
13713 @item @code{-gnatym}
13715 @emph{Check maximum line length.}
13717 The length of source lines must not exceed 79 characters, including
13718 any trailing blanks. The value of 79 allows convenient display on an
13719 80 character wide device or window, allowing for possible special
13720 treatment of 80 character lines. Note that this count is of
13721 characters in the source text. This means that a tab character counts
13722 as one character in this count and a wide character sequence counts as
13723 a single character (however many bytes are needed in the encoding).
13726 @geindex -gnatyMnnn (gcc)
13731 @item @code{-gnatyM}
13733 @emph{Set maximum line length.}
13735 The length of lines must not exceed the
13736 given value @emph{nnn}. The maximum value that can be specified is 32767.
13737 If neither style option for setting the line length is used, then the
13738 default is 255. This also controls the maximum length of lexical elements,
13739 where the only restriction is that they must fit on a single line.
13742 @geindex -gnatyn (gcc)
13747 @item @code{-gnatyn}
13749 @emph{Check casing of entities in Standard.}
13751 Any identifier from Standard must be cased
13752 to match the presentation in the Ada Reference Manual (for example,
13753 @code{Integer} and @code{ASCII.NUL}).
13756 @geindex -gnatyN (gcc)
13761 @item @code{-gnatyN}
13763 @emph{Turn off all style checks.}
13765 All style check options are turned off.
13768 @geindex -gnatyo (gcc)
13773 @item @code{-gnatyo}
13775 @emph{Check order of subprogram bodies.}
13777 All subprogram bodies in a given scope
13778 (e.g., a package body) must be in alphabetical order. The ordering
13779 rule uses normal Ada rules for comparing strings, ignoring casing
13780 of letters, except that if there is a trailing numeric suffix, then
13781 the value of this suffix is used in the ordering (e.g., Junk2 comes
13785 @geindex -gnatyO (gcc)
13790 @item @code{-gnatyO}
13792 @emph{Check that overriding subprograms are explicitly marked as such.}
13794 This applies to all subprograms of a derived type that override a primitive
13795 operation of the type, for both tagged and untagged types. In particular,
13796 the declaration of a primitive operation of a type extension that overrides
13797 an inherited operation must carry an overriding indicator. Another case is
13798 the declaration of a function that overrides a predefined operator (such
13799 as an equality operator).
13802 @geindex -gnatyp (gcc)
13807 @item @code{-gnatyp}
13809 @emph{Check pragma casing.}
13811 Pragma names must be written in mixed case, that is, the
13812 initial letter and any letter following an underscore must be uppercase.
13813 All other letters must be lowercase. An exception is that SPARK_Mode is
13814 allowed as an alternative for Spark_Mode.
13817 @geindex -gnatyr (gcc)
13822 @item @code{-gnatyr}
13824 @emph{Check references.}
13826 All identifier references must be cased in the same way as the
13827 corresponding declaration. No specific casing style is imposed on
13828 identifiers. The only requirement is for consistency of references
13832 @geindex -gnatys (gcc)
13837 @item @code{-gnatys}
13839 @emph{Check separate specs.}
13841 Separate declarations ('specs') are required for subprograms (a
13842 body is not allowed to serve as its own declaration). The only
13843 exception is that parameterless library level procedures are
13844 not required to have a separate declaration. This exception covers
13845 the most frequent form of main program procedures.
13848 @geindex -gnatyS (gcc)
13853 @item @code{-gnatyS}
13855 @emph{Check no statements after then/else.}
13857 No statements are allowed
13858 on the same line as a @code{then} or @code{else} keyword following the
13859 keyword in an @code{if} statement. @code{or else} and @code{and then} are not
13860 affected, and a special exception allows a pragma to appear after @code{else}.
13863 @geindex -gnatyt (gcc)
13868 @item @code{-gnatyt}
13870 @emph{Check token spacing.}
13872 The following token spacing rules are enforced:
13878 The keywords @code{abs} and @code{not} must be followed by a space.
13881 The token @code{=>} must be surrounded by spaces.
13884 The token @code{<>} must be preceded by a space or a left parenthesis.
13887 Binary operators other than @code{**} must be surrounded by spaces.
13888 There is no restriction on the layout of the @code{**} binary operator.
13891 Colon must be surrounded by spaces.
13894 Colon-equal (assignment, initialization) must be surrounded by spaces.
13897 Comma must be the first non-blank character on the line, or be
13898 immediately preceded by a non-blank character, and must be followed
13902 If the token preceding a left parenthesis ends with a letter or digit, then
13903 a space must separate the two tokens.
13906 If the token following a right parenthesis starts with a letter or digit, then
13907 a space must separate the two tokens.
13910 A right parenthesis must either be the first non-blank character on
13911 a line, or it must be preceded by a non-blank character.
13914 A semicolon must not be preceded by a space, and must not be followed by
13915 a non-blank character.
13918 A unary plus or minus may not be followed by a space.
13921 A vertical bar must be surrounded by spaces.
13924 Exactly one blank (and no other white space) must appear between
13925 a @code{not} token and a following @code{in} token.
13928 @geindex -gnatyu (gcc)
13933 @item @code{-gnatyu}
13935 @emph{Check unnecessary blank lines.}
13937 Unnecessary blank lines are not allowed. A blank line is considered
13938 unnecessary if it appears at the end of the file, or if more than
13939 one blank line occurs in sequence.
13942 @geindex -gnatyx (gcc)
13947 @item @code{-gnatyx}
13949 @emph{Check extra parentheses.}
13951 Unnecessary extra level of parentheses (C-style) are not allowed
13952 around conditions in @code{if} statements, @code{while} statements and
13953 @code{exit} statements.
13956 @geindex -gnatyy (gcc)
13961 @item @code{-gnatyy}
13963 @emph{Set all standard style check options.}
13965 This is equivalent to @code{gnaty3aAbcefhiklmnprst}, that is all checking
13966 options enabled with the exception of @code{-gnatyB}, @code{-gnatyd},
13967 @code{-gnatyI}, @code{-gnatyLnnn}, @code{-gnatyo}, @code{-gnatyO},
13968 @code{-gnatyS}, @code{-gnatyu}, and @code{-gnatyx}.
13971 @geindex -gnaty- (gcc)
13976 @item @code{-gnaty-}
13978 @emph{Remove style check options.}
13980 This causes any subsequent options in the string to act as canceling the
13981 corresponding style check option. To cancel maximum nesting level control,
13982 use the @code{L} parameter without any integer value after that, because any
13983 digit following @emph{-} in the parameter string of the @code{-gnaty}
13984 option will be treated as canceling the indentation check. The same is true
13985 for the @code{M} parameter. @code{y} and @code{N} parameters are not
13986 allowed after @emph{-}.
13989 @geindex -gnaty+ (gcc)
13994 @item @code{-gnaty+}
13996 @emph{Enable style check options.}
13998 This causes any subsequent options in the string to enable the corresponding
13999 style check option. That is, it cancels the effect of a previous -,
14003 @c end of switch description (leave this comment to ease automatic parsing for
14007 In the above rules, appearing in column one is always permitted, that is,
14008 counts as meeting either a requirement for a required preceding space,
14009 or as meeting a requirement for no preceding space.
14011 Appearing at the end of a line is also always permitted, that is, counts
14012 as meeting either a requirement for a following space, or as meeting
14013 a requirement for no following space.
14015 If any of these style rules is violated, a message is generated giving
14016 details on the violation. The initial characters of such messages are
14017 always '@cite{(style)}'. Note that these messages are treated as warning
14018 messages, so they normally do not prevent the generation of an object
14019 file. The @code{-gnatwe} switch can be used to treat warning messages,
14020 including style messages, as fatal errors.
14022 The switch @code{-gnaty} on its own (that is not
14023 followed by any letters or digits) is equivalent
14024 to the use of @code{-gnatyy} as described above, that is all
14025 built-in standard style check options are enabled.
14027 The switch @code{-gnatyN} clears any previously set style checks.
14029 @node Run-Time Checks,Using gcc for Syntax Checking,Style Checking,Compiler Switches
14030 @anchor{gnat_ugn/building_executable_programs_with_gnat run-time-checks}@anchor{f9}@anchor{gnat_ugn/building_executable_programs_with_gnat id19}@anchor{104}
14031 @subsection Run-Time Checks
14034 @geindex Division by zero
14036 @geindex Access before elaboration
14039 @geindex division by zero
14042 @geindex access before elaboration
14045 @geindex stack overflow checking
14047 By default, the following checks are suppressed: stack overflow
14048 checks, and checks for access before elaboration on subprogram
14049 calls. All other checks, including overflow checks, range checks and
14050 array bounds checks, are turned on by default. The following @code{gcc}
14051 switches refine this default behavior.
14053 @geindex -gnatp (gcc)
14058 @item @code{-gnatp}
14060 @geindex Suppressing checks
14063 @geindex suppressing
14065 This switch causes the unit to be compiled
14066 as though @code{pragma Suppress (All_checks)}
14067 had been present in the source. Validity checks are also eliminated (in
14068 other words @code{-gnatp} also implies @code{-gnatVn}.
14069 Use this switch to improve the performance
14070 of the code at the expense of safety in the presence of invalid data or
14073 Note that when checks are suppressed, the compiler is allowed, but not
14074 required, to omit the checking code. If the run-time cost of the
14075 checking code is zero or near-zero, the compiler will generate it even
14076 if checks are suppressed. In particular, if the compiler can prove
14077 that a certain check will necessarily fail, it will generate code to
14078 do an unconditional 'raise', even if checks are suppressed. The
14079 compiler warns in this case. Another case in which checks may not be
14080 eliminated is when they are embedded in certain run-time routines such
14081 as math library routines.
14083 Of course, run-time checks are omitted whenever the compiler can prove
14084 that they will not fail, whether or not checks are suppressed.
14086 Note that if you suppress a check that would have failed, program
14087 execution is erroneous, which means the behavior is totally
14088 unpredictable. The program might crash, or print wrong answers, or
14089 do anything else. It might even do exactly what you wanted it to do
14090 (and then it might start failing mysteriously next week or next
14091 year). The compiler will generate code based on the assumption that
14092 the condition being checked is true, which can result in erroneous
14093 execution if that assumption is wrong.
14095 The checks subject to suppression include all the checks defined by the Ada
14096 standard, the additional implementation defined checks @code{Alignment_Check},
14097 @code{Duplicated_Tag_Check}, @code{Predicate_Check}, @code{Container_Checks}, @code{Tampering_Check},
14098 and @code{Validity_Check}, as well as any checks introduced using @code{pragma Check_Name}.
14099 Note that @code{Atomic_Synchronization} is not automatically suppressed by use of this option.
14101 If the code depends on certain checks being active, you can use
14102 pragma @code{Unsuppress} either as a configuration pragma or as
14103 a local pragma to make sure that a specified check is performed
14104 even if @code{gnatp} is specified.
14106 The @code{-gnatp} switch has no effect if a subsequent
14107 @code{-gnat-p} switch appears.
14110 @geindex -gnat-p (gcc)
14112 @geindex Suppressing checks
14115 @geindex suppressing
14122 @item @code{-gnat-p}
14124 This switch cancels the effect of a previous @code{gnatp} switch.
14127 @geindex -gnato?? (gcc)
14129 @geindex Overflow checks
14131 @geindex Overflow mode
14139 @item @code{-gnato??}
14141 This switch controls the mode used for computing intermediate
14142 arithmetic integer operations, and also enables overflow checking.
14143 For a full description of overflow mode and checking control, see
14144 the 'Overflow Check Handling in GNAT' appendix in this
14147 Overflow checks are always enabled by this switch. The argument
14148 controls the mode, using the codes
14153 @item @emph{1 = STRICT}
14155 In STRICT mode, intermediate operations are always done using the
14156 base type, and overflow checking ensures that the result is within
14157 the base type range.
14159 @item @emph{2 = MINIMIZED}
14161 In MINIMIZED mode, overflows in intermediate operations are avoided
14162 where possible by using a larger integer type for the computation
14163 (typically @code{Long_Long_Integer}). Overflow checking ensures that
14164 the result fits in this larger integer type.
14166 @item @emph{3 = ELIMINATED}
14168 In ELIMINATED mode, overflows in intermediate operations are avoided
14169 by using multi-precision arithmetic. In this case, overflow checking
14170 has no effect on intermediate operations (since overflow is impossible).
14173 If two digits are present after @code{-gnato} then the first digit
14174 sets the mode for expressions outside assertions, and the second digit
14175 sets the mode for expressions within assertions. Here assertions is used
14176 in the technical sense (which includes for example precondition and
14177 postcondition expressions).
14179 If one digit is present, the corresponding mode is applicable to both
14180 expressions within and outside assertion expressions.
14182 If no digits are present, the default is to enable overflow checks
14183 and set STRICT mode for both kinds of expressions. This is compatible
14184 with the use of @code{-gnato} in previous versions of GNAT.
14186 @geindex Machine_Overflows
14188 Note that the @code{-gnato??} switch does not affect the code generated
14189 for any floating-point operations; it applies only to integer semantics.
14190 For floating-point, GNAT has the @code{Machine_Overflows}
14191 attribute set to @code{False} and the normal mode of operation is to
14192 generate IEEE NaN and infinite values on overflow or invalid operations
14193 (such as dividing 0.0 by 0.0).
14195 The reason that we distinguish overflow checking from other kinds of
14196 range constraint checking is that a failure of an overflow check, unlike
14197 for example the failure of a range check, can result in an incorrect
14198 value, but cannot cause random memory destruction (like an out of range
14199 subscript), or a wild jump (from an out of range case value). Overflow
14200 checking is also quite expensive in time and space, since in general it
14201 requires the use of double length arithmetic.
14203 Note again that the default is @code{-gnato11} (equivalent to @code{-gnato1}),
14204 so overflow checking is performed in STRICT mode by default.
14207 @geindex -gnatE (gcc)
14209 @geindex Elaboration checks
14212 @geindex elaboration
14217 @item @code{-gnatE}
14219 Enables dynamic checks for access-before-elaboration
14220 on subprogram calls and generic instantiations.
14221 Note that @code{-gnatE} is not necessary for safety, because in the
14222 default mode, GNAT ensures statically that the checks would not fail.
14223 For full details of the effect and use of this switch,
14224 @ref{1c,,Compiling with gcc}.
14227 @geindex -fstack-check (gcc)
14229 @geindex Stack Overflow Checking
14232 @geindex stack overflow checking
14237 @item @code{-fstack-check}
14239 Activates stack overflow checking. For full details of the effect and use of
14240 this switch see @ref{f4,,Stack Overflow Checking}.
14243 @geindex Unsuppress
14245 The setting of these switches only controls the default setting of the
14246 checks. You may modify them using either @code{Suppress} (to remove
14247 checks) or @code{Unsuppress} (to add back suppressed checks) pragmas in
14248 the program source.
14250 @node Using gcc for Syntax Checking,Using gcc for Semantic Checking,Run-Time Checks,Compiler Switches
14251 @anchor{gnat_ugn/building_executable_programs_with_gnat id20}@anchor{105}@anchor{gnat_ugn/building_executable_programs_with_gnat using-gcc-for-syntax-checking}@anchor{106}
14252 @subsection Using @code{gcc} for Syntax Checking
14255 @geindex -gnats (gcc)
14260 @item @code{-gnats}
14262 The @code{s} stands for 'syntax'.
14264 Run GNAT in syntax checking only mode. For
14265 example, the command
14268 $ gcc -c -gnats x.adb
14271 compiles file @code{x.adb} in syntax-check-only mode. You can check a
14272 series of files in a single command
14273 , and can use wild cards to specify such a group of files.
14274 Note that you must specify the @code{-c} (compile
14275 only) flag in addition to the @code{-gnats} flag.
14277 You may use other switches in conjunction with @code{-gnats}. In
14278 particular, @code{-gnatl} and @code{-gnatv} are useful to control the
14279 format of any generated error messages.
14281 When the source file is empty or contains only empty lines and/or comments,
14282 the output is a warning:
14285 $ gcc -c -gnats -x ada toto.txt
14286 toto.txt:1:01: warning: empty file, contains no compilation units
14290 Otherwise, the output is simply the error messages, if any. No object file or
14291 ALI file is generated by a syntax-only compilation. Also, no units other
14292 than the one specified are accessed. For example, if a unit @code{X}
14293 @emph{with}s a unit @code{Y}, compiling unit @code{X} in syntax
14294 check only mode does not access the source file containing unit
14297 @geindex Multiple units
14298 @geindex syntax checking
14300 Normally, GNAT allows only a single unit in a source file. However, this
14301 restriction does not apply in syntax-check-only mode, and it is possible
14302 to check a file containing multiple compilation units concatenated
14303 together. This is primarily used by the @code{gnatchop} utility
14304 (@ref{36,,Renaming Files with gnatchop}).
14307 @node Using gcc for Semantic Checking,Compiling Different Versions of Ada,Using gcc for Syntax Checking,Compiler Switches
14308 @anchor{gnat_ugn/building_executable_programs_with_gnat id21}@anchor{107}@anchor{gnat_ugn/building_executable_programs_with_gnat using-gcc-for-semantic-checking}@anchor{108}
14309 @subsection Using @code{gcc} for Semantic Checking
14312 @geindex -gnatc (gcc)
14317 @item @code{-gnatc}
14319 The @code{c} stands for 'check'.
14320 Causes the compiler to operate in semantic check mode,
14321 with full checking for all illegalities specified in the
14322 Ada Reference Manual, but without generation of any object code
14323 (no object file is generated).
14325 Because dependent files must be accessed, you must follow the GNAT
14326 semantic restrictions on file structuring to operate in this mode:
14332 The needed source files must be accessible
14333 (see @ref{89,,Search Paths and the Run-Time Library (RTL)}).
14336 Each file must contain only one compilation unit.
14339 The file name and unit name must match (@ref{52,,File Naming Rules}).
14342 The output consists of error messages as appropriate. No object file is
14343 generated. An @code{ALI} file is generated for use in the context of
14344 cross-reference tools, but this file is marked as not being suitable
14345 for binding (since no object file is generated).
14346 The checking corresponds exactly to the notion of
14347 legality in the Ada Reference Manual.
14349 Any unit can be compiled in semantics-checking-only mode, including
14350 units that would not normally be compiled (subunits,
14351 and specifications where a separate body is present).
14354 @node Compiling Different Versions of Ada,Character Set Control,Using gcc for Semantic Checking,Compiler Switches
14355 @anchor{gnat_ugn/building_executable_programs_with_gnat compiling-different-versions-of-ada}@anchor{6}@anchor{gnat_ugn/building_executable_programs_with_gnat id22}@anchor{109}
14356 @subsection Compiling Different Versions of Ada
14359 The switches described in this section allow you to explicitly specify
14360 the version of the Ada language that your programs are written in.
14361 The default mode is Ada 2012,
14362 but you can also specify Ada 95, Ada 2005 mode, or
14363 indicate Ada 83 compatibility mode.
14365 @geindex Compatibility with Ada 83
14367 @geindex -gnat83 (gcc)
14370 @geindex Ada 83 tests
14372 @geindex Ada 83 mode
14377 @item @code{-gnat83} (Ada 83 Compatibility Mode)
14379 Although GNAT is primarily an Ada 95 / Ada 2005 compiler, this switch
14380 specifies that the program is to be compiled in Ada 83 mode. With
14381 @code{-gnat83}, GNAT rejects most post-Ada 83 extensions and applies Ada 83
14382 semantics where this can be done easily.
14383 It is not possible to guarantee this switch does a perfect
14384 job; some subtle tests, such as are
14385 found in earlier ACVC tests (and that have been removed from the ACATS suite
14386 for Ada 95), might not compile correctly.
14387 Nevertheless, this switch may be useful in some circumstances, for example
14388 where, due to contractual reasons, existing code needs to be maintained
14389 using only Ada 83 features.
14391 With few exceptions (most notably the need to use @code{<>} on
14393 @geindex Generic formal parameters
14394 generic formal parameters,
14395 the use of the new Ada 95 / Ada 2005
14396 reserved words, and the use of packages
14397 with optional bodies), it is not necessary to specify the
14398 @code{-gnat83} switch when compiling Ada 83 programs, because, with rare
14399 exceptions, Ada 95 and Ada 2005 are upwardly compatible with Ada 83. Thus
14400 a correct Ada 83 program is usually also a correct program
14401 in these later versions of the language standard. For further information
14402 please refer to the @emph{Compatibility and Porting Guide} chapter in the
14403 @cite{GNAT Reference Manual}.
14406 @geindex -gnat95 (gcc)
14408 @geindex Ada 95 mode
14413 @item @code{-gnat95} (Ada 95 mode)
14415 This switch directs the compiler to implement the Ada 95 version of the
14417 Since Ada 95 is almost completely upwards
14418 compatible with Ada 83, Ada 83 programs may generally be compiled using
14419 this switch (see the description of the @code{-gnat83} switch for further
14420 information about Ada 83 mode).
14421 If an Ada 2005 program is compiled in Ada 95 mode,
14422 uses of the new Ada 2005 features will cause error
14423 messages or warnings.
14425 This switch also can be used to cancel the effect of a previous
14426 @code{-gnat83}, @code{-gnat05/2005}, or @code{-gnat12/2012}
14427 switch earlier in the command line.
14430 @geindex -gnat05 (gcc)
14432 @geindex -gnat2005 (gcc)
14434 @geindex Ada 2005 mode
14439 @item @code{-gnat05} or @code{-gnat2005} (Ada 2005 mode)
14441 This switch directs the compiler to implement the Ada 2005 version of the
14442 language, as documented in the official Ada standards document.
14443 Since Ada 2005 is almost completely upwards
14444 compatible with Ada 95 (and thus also with Ada 83), Ada 83 and Ada 95 programs
14445 may generally be compiled using this switch (see the description of the
14446 @code{-gnat83} and @code{-gnat95} switches for further
14450 @geindex -gnat12 (gcc)
14452 @geindex -gnat2012 (gcc)
14454 @geindex Ada 2012 mode
14459 @item @code{-gnat12} or @code{-gnat2012} (Ada 2012 mode)
14461 This switch directs the compiler to implement the Ada 2012 version of the
14462 language (also the default).
14463 Since Ada 2012 is almost completely upwards
14464 compatible with Ada 2005 (and thus also with Ada 83, and Ada 95),
14465 Ada 83 and Ada 95 programs
14466 may generally be compiled using this switch (see the description of the
14467 @code{-gnat83}, @code{-gnat95}, and @code{-gnat05/2005} switches
14468 for further information).
14471 @geindex -gnatX (gcc)
14473 @geindex Ada language extensions
14475 @geindex GNAT extensions
14480 @item @code{-gnatX} (Enable GNAT Extensions)
14482 This switch directs the compiler to implement the latest version of the
14483 language (currently Ada 2012) and also to enable certain GNAT implementation
14484 extensions that are not part of any Ada standard. For a full list of these
14485 extensions, see the GNAT reference manual.
14488 @node Character Set Control,File Naming Control,Compiling Different Versions of Ada,Compiler Switches
14489 @anchor{gnat_ugn/building_executable_programs_with_gnat id23}@anchor{10a}@anchor{gnat_ugn/building_executable_programs_with_gnat character-set-control}@anchor{48}
14490 @subsection Character Set Control
14493 @geindex -gnati (gcc)
14498 @item @code{-gnati@emph{c}}
14500 Normally GNAT recognizes the Latin-1 character set in source program
14501 identifiers, as described in the Ada Reference Manual.
14503 GNAT to recognize alternate character sets in identifiers. @code{c} is a
14504 single character indicating the character set, as follows:
14507 @multitable {xxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
14514 ISO 8859-1 (Latin-1) identifiers
14522 ISO 8859-2 (Latin-2) letters allowed in identifiers
14530 ISO 8859-3 (Latin-3) letters allowed in identifiers
14538 ISO 8859-4 (Latin-4) letters allowed in identifiers
14546 ISO 8859-5 (Cyrillic) letters allowed in identifiers
14554 ISO 8859-15 (Latin-9) letters allowed in identifiers
14562 IBM PC letters (code page 437) allowed in identifiers
14570 IBM PC letters (code page 850) allowed in identifiers
14578 Full upper-half codes allowed in identifiers
14586 No upper-half codes allowed in identifiers
14594 Wide-character codes (that is, codes greater than 255)
14595 allowed in identifiers
14600 See @ref{3e,,Foreign Language Representation} for full details on the
14601 implementation of these character sets.
14604 @geindex -gnatW (gcc)
14609 @item @code{-gnatW@emph{e}}
14611 Specify the method of encoding for wide characters.
14612 @code{e} is one of the following:
14615 @multitable {xxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
14622 Hex encoding (brackets coding also recognized)
14630 Upper half encoding (brackets encoding also recognized)
14638 Shift/JIS encoding (brackets encoding also recognized)
14646 EUC encoding (brackets encoding also recognized)
14654 UTF-8 encoding (brackets encoding also recognized)
14662 Brackets encoding only (default value)
14667 For full details on these encoding
14668 methods see @ref{4e,,Wide_Character Encodings}.
14669 Note that brackets coding is always accepted, even if one of the other
14670 options is specified, so for example @code{-gnatW8} specifies that both
14671 brackets and UTF-8 encodings will be recognized. The units that are
14672 with'ed directly or indirectly will be scanned using the specified
14673 representation scheme, and so if one of the non-brackets scheme is
14674 used, it must be used consistently throughout the program. However,
14675 since brackets encoding is always recognized, it may be conveniently
14676 used in standard libraries, allowing these libraries to be used with
14677 any of the available coding schemes.
14679 Note that brackets encoding only applies to program text. Within comments,
14680 brackets are considered to be normal graphic characters, and bracket sequences
14681 are never recognized as wide characters.
14683 If no @code{-gnatW?} parameter is present, then the default
14684 representation is normally Brackets encoding only. However, if the
14685 first three characters of the file are 16#EF# 16#BB# 16#BF# (the standard
14686 byte order mark or BOM for UTF-8), then these three characters are
14687 skipped and the default representation for the file is set to UTF-8.
14689 Note that the wide character representation that is specified (explicitly
14690 or by default) for the main program also acts as the default encoding used
14691 for Wide_Text_IO files if not specifically overridden by a WCEM form
14695 When no @code{-gnatW?} is specified, then characters (other than wide
14696 characters represented using brackets notation) are treated as 8-bit
14697 Latin-1 codes. The codes recognized are the Latin-1 graphic characters,
14698 and ASCII format effectors (CR, LF, HT, VT). Other lower half control
14699 characters in the range 16#00#..16#1F# are not accepted in program text
14700 or in comments. Upper half control characters (16#80#..16#9F#) are rejected
14701 in program text, but allowed and ignored in comments. Note in particular
14702 that the Next Line (NEL) character whose encoding is 16#85# is not recognized
14703 as an end of line in this default mode. If your source program contains
14704 instances of the NEL character used as a line terminator,
14705 you must use UTF-8 encoding for the whole
14706 source program. In default mode, all lines must be ended by a standard
14707 end of line sequence (CR, CR/LF, or LF).
14709 Note that the convention of simply accepting all upper half characters in
14710 comments means that programs that use standard ASCII for program text, but
14711 UTF-8 encoding for comments are accepted in default mode, providing that the
14712 comments are ended by an appropriate (CR, or CR/LF, or LF) line terminator.
14713 This is a common mode for many programs with foreign language comments.
14715 @node File Naming Control,Subprogram Inlining Control,Character Set Control,Compiler Switches
14716 @anchor{gnat_ugn/building_executable_programs_with_gnat file-naming-control}@anchor{10b}@anchor{gnat_ugn/building_executable_programs_with_gnat id24}@anchor{10c}
14717 @subsection File Naming Control
14720 @geindex -gnatk (gcc)
14725 @item @code{-gnatk@emph{n}}
14727 Activates file name 'krunching'. @code{n}, a decimal integer in the range
14728 1-999, indicates the maximum allowable length of a file name (not
14729 including the @code{.ads} or @code{.adb} extension). The default is not
14730 to enable file name krunching.
14732 For the source file naming rules, @ref{52,,File Naming Rules}.
14735 @node Subprogram Inlining Control,Auxiliary Output Control,File Naming Control,Compiler Switches
14736 @anchor{gnat_ugn/building_executable_programs_with_gnat subprogram-inlining-control}@anchor{10d}@anchor{gnat_ugn/building_executable_programs_with_gnat id25}@anchor{10e}
14737 @subsection Subprogram Inlining Control
14740 @geindex -gnatn (gcc)
14745 @item @code{-gnatn[12]}
14747 The @code{n} here is intended to suggest the first syllable of the word 'inline'.
14748 GNAT recognizes and processes @code{Inline} pragmas. However, for inlining to
14749 actually occur, optimization must be enabled and, by default, inlining of
14750 subprograms across units is not performed. If you want to additionally
14751 enable inlining of subprograms specified by pragma @code{Inline} across units,
14752 you must also specify this switch.
14754 In the absence of this switch, GNAT does not attempt inlining across units
14755 and does not access the bodies of subprograms for which @code{pragma Inline} is
14756 specified if they are not in the current unit.
14758 You can optionally specify the inlining level: 1 for moderate inlining across
14759 units, which is a good compromise between compilation times and performances
14760 at run time, or 2 for full inlining across units, which may bring about
14761 longer compilation times. If no inlining level is specified, the compiler will
14762 pick it based on the optimization level: 1 for @code{-O1}, @code{-O2} or
14763 @code{-Os} and 2 for @code{-O3}.
14765 If you specify this switch the compiler will access these bodies,
14766 creating an extra source dependency for the resulting object file, and
14767 where possible, the call will be inlined.
14768 For further details on when inlining is possible
14769 see @ref{10f,,Inlining of Subprograms}.
14772 @geindex -gnatN (gcc)
14777 @item @code{-gnatN}
14779 This switch activates front-end inlining which also
14780 generates additional dependencies.
14782 When using a gcc-based back end (in practice this means using any version
14783 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
14784 @code{-gnatN} is deprecated, and the use of @code{-gnatn} is preferred.
14785 Historically front end inlining was more extensive than the gcc back end
14786 inlining, but that is no longer the case.
14789 @node Auxiliary Output Control,Debugging Control,Subprogram Inlining Control,Compiler Switches
14790 @anchor{gnat_ugn/building_executable_programs_with_gnat auxiliary-output-control}@anchor{110}@anchor{gnat_ugn/building_executable_programs_with_gnat id26}@anchor{111}
14791 @subsection Auxiliary Output Control
14794 @geindex -gnatt (gcc)
14796 @geindex Writing internal trees
14798 @geindex Internal trees
14799 @geindex writing to file
14804 @item @code{-gnatt}
14806 Causes GNAT to write the internal tree for a unit to a file (with the
14807 extension @code{.adt}.
14808 This not normally required, but is used by separate analysis tools.
14810 these tools do the necessary compilations automatically, so you should
14811 not have to specify this switch in normal operation.
14812 Note that the combination of switches @code{-gnatct}
14813 generates a tree in the form required by ASIS applications.
14816 @geindex -gnatu (gcc)
14821 @item @code{-gnatu}
14823 Print a list of units required by this compilation on @code{stdout}.
14824 The listing includes all units on which the unit being compiled depends
14825 either directly or indirectly.
14828 @geindex -pass-exit-codes (gcc)
14833 @item @code{-pass-exit-codes}
14835 If this switch is not used, the exit code returned by @code{gcc} when
14836 compiling multiple files indicates whether all source files have
14837 been successfully used to generate object files or not.
14839 When @code{-pass-exit-codes} is used, @code{gcc} exits with an extended
14840 exit status and allows an integrated development environment to better
14841 react to a compilation failure. Those exit status are:
14844 @multitable {xxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
14851 There was an error in at least one source file.
14859 At least one source file did not generate an object file.
14867 The compiler died unexpectedly (internal error for example).
14875 An object file has been generated for every source file.
14881 @node Debugging Control,Exception Handling Control,Auxiliary Output Control,Compiler Switches
14882 @anchor{gnat_ugn/building_executable_programs_with_gnat debugging-control}@anchor{112}@anchor{gnat_ugn/building_executable_programs_with_gnat id27}@anchor{113}
14883 @subsection Debugging Control
14888 @geindex Debugging options
14891 @geindex -gnatd (gcc)
14896 @item @code{-gnatd@emph{x}}
14898 Activate internal debugging switches. @code{x} is a letter or digit, or
14899 string of letters or digits, which specifies the type of debugging
14900 outputs desired. Normally these are used only for internal development
14901 or system debugging purposes. You can find full documentation for these
14902 switches in the body of the @code{Debug} unit in the compiler source
14903 file @code{debug.adb}.
14906 @geindex -gnatG (gcc)
14911 @item @code{-gnatG[=@emph{nn}]}
14913 This switch causes the compiler to generate auxiliary output containing
14914 a pseudo-source listing of the generated expanded code. Like most Ada
14915 compilers, GNAT works by first transforming the high level Ada code into
14916 lower level constructs. For example, tasking operations are transformed
14917 into calls to the tasking run-time routines. A unique capability of GNAT
14918 is to list this expanded code in a form very close to normal Ada source.
14919 This is very useful in understanding the implications of various Ada
14920 usage on the efficiency of the generated code. There are many cases in
14921 Ada (e.g., the use of controlled types), where simple Ada statements can
14922 generate a lot of run-time code. By using @code{-gnatG} you can identify
14923 these cases, and consider whether it may be desirable to modify the coding
14924 approach to improve efficiency.
14926 The optional parameter @code{nn} if present after -gnatG specifies an
14927 alternative maximum line length that overrides the normal default of 72.
14928 This value is in the range 40-999999, values less than 40 being silently
14929 reset to 40. The equal sign is optional.
14931 The format of the output is very similar to standard Ada source, and is
14932 easily understood by an Ada programmer. The following special syntactic
14933 additions correspond to low level features used in the generated code that
14934 do not have any exact analogies in pure Ada source form. The following
14935 is a partial list of these special constructions. See the spec
14936 of package @code{Sprint} in file @code{sprint.ads} for a full list.
14938 @geindex -gnatL (gcc)
14940 If the switch @code{-gnatL} is used in conjunction with
14941 @code{-gnatG}, then the original source lines are interspersed
14942 in the expanded source (as comment lines with the original line number).
14947 @item @code{new @emph{xxx} [storage_pool = @emph{yyy}]}
14949 Shows the storage pool being used for an allocator.
14951 @item @code{at end @emph{procedure-name};}
14953 Shows the finalization (cleanup) procedure for a scope.
14955 @item @code{(if @emph{expr} then @emph{expr} else @emph{expr})}
14957 Conditional expression equivalent to the @code{x?y:z} construction in C.
14959 @item @code{@emph{target}^(@emph{source})}
14961 A conversion with floating-point truncation instead of rounding.
14963 @item @code{@emph{target}?(@emph{source})}
14965 A conversion that bypasses normal Ada semantic checking. In particular
14966 enumeration types and fixed-point types are treated simply as integers.
14968 @item @code{@emph{target}?^(@emph{source})}
14970 Combines the above two cases.
14973 @code{@emph{x} #/ @emph{y}}
14975 @code{@emph{x} #mod @emph{y}}
14977 @code{@emph{x} # @emph{y}}
14982 @item @code{@emph{x} #rem @emph{y}}
14984 A division or multiplication of fixed-point values which are treated as
14985 integers without any kind of scaling.
14987 @item @code{free @emph{expr} [storage_pool = @emph{xxx}]}
14989 Shows the storage pool associated with a @code{free} statement.
14991 @item @code{[subtype or type declaration]}
14993 Used to list an equivalent declaration for an internally generated
14994 type that is referenced elsewhere in the listing.
14996 @item @code{freeze @emph{type-name} [@emph{actions}]}
14998 Shows the point at which @code{type-name} is frozen, with possible
14999 associated actions to be performed at the freeze point.
15001 @item @code{reference @emph{itype}}
15003 Reference (and hence definition) to internal type @code{itype}.
15005 @item @code{@emph{function-name}! (@emph{arg}, @emph{arg}, @emph{arg})}
15007 Intrinsic function call.
15009 @item @code{@emph{label-name} : label}
15011 Declaration of label @code{labelname}.
15013 @item @code{#$ @emph{subprogram-name}}
15015 An implicit call to a run-time support routine
15016 (to meet the requirement of H.3.1(9) in a
15017 convenient manner).
15019 @item @code{@emph{expr} && @emph{expr} && @emph{expr} ... && @emph{expr}}
15021 A multiple concatenation (same effect as @code{expr} & @code{expr} &
15022 @code{expr}, but handled more efficiently).
15024 @item @code{[constraint_error]}
15026 Raise the @code{Constraint_Error} exception.
15028 @item @code{@emph{expression}'reference}
15030 A pointer to the result of evaluating @{expression@}.
15032 @item @code{@emph{target-type}!(@emph{source-expression})}
15034 An unchecked conversion of @code{source-expression} to @code{target-type}.
15036 @item @code{[@emph{numerator}/@emph{denominator}]}
15038 Used to represent internal real literals (that) have no exact
15039 representation in base 2-16 (for example, the result of compile time
15040 evaluation of the expression 1.0/27.0).
15044 @geindex -gnatD (gcc)
15049 @item @code{-gnatD[=nn]}
15051 When used in conjunction with @code{-gnatG}, this switch causes
15052 the expanded source, as described above for
15053 @code{-gnatG} to be written to files with names
15054 @code{xxx.dg}, where @code{xxx} is the normal file name,
15055 instead of to the standard output file. For
15056 example, if the source file name is @code{hello.adb}, then a file
15057 @code{hello.adb.dg} will be written. The debugging
15058 information generated by the @code{gcc} @code{-g} switch
15059 will refer to the generated @code{xxx.dg} file. This allows
15060 you to do source level debugging using the generated code which is
15061 sometimes useful for complex code, for example to find out exactly
15062 which part of a complex construction raised an exception. This switch
15063 also suppresses generation of cross-reference information (see
15064 @code{-gnatx}) since otherwise the cross-reference information
15065 would refer to the @code{.dg} file, which would cause
15066 confusion since this is not the original source file.
15068 Note that @code{-gnatD} actually implies @code{-gnatG}
15069 automatically, so it is not necessary to give both options.
15070 In other words @code{-gnatD} is equivalent to @code{-gnatDG}).
15072 @geindex -gnatL (gcc)
15074 If the switch @code{-gnatL} is used in conjunction with
15075 @code{-gnatDG}, then the original source lines are interspersed
15076 in the expanded source (as comment lines with the original line number).
15078 The optional parameter @code{nn} if present after -gnatD specifies an
15079 alternative maximum line length that overrides the normal default of 72.
15080 This value is in the range 40-999999, values less than 40 being silently
15081 reset to 40. The equal sign is optional.
15084 @geindex -gnatr (gcc)
15086 @geindex pragma Restrictions
15091 @item @code{-gnatr}
15093 This switch causes pragma Restrictions to be treated as Restriction_Warnings
15094 so that violation of restrictions causes warnings rather than illegalities.
15095 This is useful during the development process when new restrictions are added
15096 or investigated. The switch also causes pragma Profile to be treated as
15097 Profile_Warnings, and pragma Restricted_Run_Time and pragma Ravenscar set
15098 restriction warnings rather than restrictions.
15101 @geindex -gnatR (gcc)
15106 @item @code{-gnatR[0|1|2|3|4][e][j][m][s]}
15108 This switch controls output from the compiler of a listing showing
15109 representation information for declared types, objects and subprograms.
15110 For @code{-gnatR0}, no information is output (equivalent to omitting
15111 the @code{-gnatR} switch). For @code{-gnatR1} (which is the default,
15112 so @code{-gnatR} with no parameter has the same effect), size and
15113 alignment information is listed for declared array and record types.
15115 For @code{-gnatR2}, size and alignment information is listed for all
15116 declared types and objects. The @code{Linker_Section} is also listed for any
15117 entity for which the @code{Linker_Section} is set explicitly or implicitly (the
15118 latter case occurs for objects of a type for which a @code{Linker_Section}
15121 For @code{-gnatR3}, symbolic expressions for values that are computed
15122 at run time for records are included. These symbolic expressions have
15123 a mostly obvious format with #n being used to represent the value of the
15124 n'th discriminant. See source files @code{repinfo.ads/adb} in the
15125 GNAT sources for full details on the format of @code{-gnatR3} output.
15127 For @code{-gnatR4}, information for relevant compiler-generated types
15128 is also listed, i.e. when they are structurally part of other declared
15131 If the switch is followed by an @code{e} (e.g. @code{-gnatR2e}), then
15132 extended representation information for record sub-components of records
15135 If the switch is followed by an @code{m} (e.g. @code{-gnatRm}), then
15136 subprogram conventions and parameter passing mechanisms for all the
15137 subprograms are included.
15139 If the switch is followed by a @code{j} (e.g., @code{-gnatRj}), then
15140 the output is in the JSON data interchange format specified by the
15141 ECMA-404 standard. The semantic description of this JSON output is
15142 available in the specification of the Repinfo unit present in the
15145 If the switch is followed by an @code{s} (e.g., @code{-gnatR3s}), then
15146 the output is to a file with the name @code{file.rep} where @code{file} is
15147 the name of the corresponding source file, except if @code{j} is also
15148 specified, in which case the file name is @code{file.json}.
15150 Note that it is possible for record components to have zero size. In
15151 this case, the component clause uses an obvious extension of permitted
15152 Ada syntax, for example @code{at 0 range 0 .. -1}.
15155 @geindex -gnatS (gcc)
15160 @item @code{-gnatS}
15162 The use of the switch @code{-gnatS} for an
15163 Ada compilation will cause the compiler to output a
15164 representation of package Standard in a form very
15165 close to standard Ada. It is not quite possible to
15166 do this entirely in standard Ada (since new
15167 numeric base types cannot be created in standard
15168 Ada), but the output is easily
15169 readable to any Ada programmer, and is useful to
15170 determine the characteristics of target dependent
15171 types in package Standard.
15174 @geindex -gnatx (gcc)
15179 @item @code{-gnatx}
15181 Normally the compiler generates full cross-referencing information in
15182 the @code{ALI} file. This information is used by a number of tools,
15183 including @code{gnatfind} and @code{gnatxref}. The @code{-gnatx} switch
15184 suppresses this information. This saves some space and may slightly
15185 speed up compilation, but means that these tools cannot be used.
15188 @geindex -fgnat-encodings (gcc)
15193 @item @code{-fgnat-encodings=[all|gdb|minimal]}
15195 This switch controls the balance between GNAT encodings and standard DWARF
15196 emitted in the debug information.
15198 Historically, old debug formats like stabs were not powerful enough to
15199 express some Ada types (for instance, variant records or fixed-point types).
15200 To work around this, GNAT introduced proprietary encodings that embed the
15201 missing information ("GNAT encodings").
15203 Recent versions of the DWARF debug information format are now able to
15204 correctly describe most of these Ada constructs ("standard DWARF"). As
15205 third-party tools started to use this format, GNAT has been enhanced to
15206 generate it. However, most tools (including GDB) are still relying on GNAT
15209 To support all tools, GNAT needs to be versatile about the balance between
15210 generation of GNAT encodings and standard DWARF. This is what
15211 @code{-fgnat-encodings} is about.
15217 @code{=all}: Emit all GNAT encodings, and then emit as much standard DWARF as
15218 possible so it does not conflict with GNAT encodings.
15221 @code{=gdb}: Emit as much standard DWARF as possible as long as the current
15222 GDB handles it. Emit GNAT encodings for the rest.
15225 @code{=minimal}: Emit as much standard DWARF as possible and emit GNAT
15226 encodings for the rest.
15230 @node Exception Handling Control,Units to Sources Mapping Files,Debugging Control,Compiler Switches
15231 @anchor{gnat_ugn/building_executable_programs_with_gnat id28}@anchor{114}@anchor{gnat_ugn/building_executable_programs_with_gnat exception-handling-control}@anchor{115}
15232 @subsection Exception Handling Control
15235 GNAT uses two methods for handling exceptions at run time. The
15236 @code{setjmp/longjmp} method saves the context when entering
15237 a frame with an exception handler. Then when an exception is
15238 raised, the context can be restored immediately, without the
15239 need for tracing stack frames. This method provides very fast
15240 exception propagation, but introduces significant overhead for
15241 the use of exception handlers, even if no exception is raised.
15243 The other approach is called 'zero cost' exception handling.
15244 With this method, the compiler builds static tables to describe
15245 the exception ranges. No dynamic code is required when entering
15246 a frame containing an exception handler. When an exception is
15247 raised, the tables are used to control a back trace of the
15248 subprogram invocation stack to locate the required exception
15249 handler. This method has considerably poorer performance for
15250 the propagation of exceptions, but there is no overhead for
15251 exception handlers if no exception is raised. Note that in this
15252 mode and in the context of mixed Ada and C/C++ programming,
15253 to propagate an exception through a C/C++ code, the C/C++ code
15254 must be compiled with the @code{-funwind-tables} GCC's
15257 The following switches may be used to control which of the
15258 two exception handling methods is used.
15260 @geindex --RTS=sjlj (gnatmake)
15265 @item @code{--RTS=sjlj}
15267 This switch causes the setjmp/longjmp run-time (when available) to be used
15268 for exception handling. If the default
15269 mechanism for the target is zero cost exceptions, then
15270 this switch can be used to modify this default, and must be
15271 used for all units in the partition.
15272 This option is rarely used. One case in which it may be
15273 advantageous is if you have an application where exception
15274 raising is common and the overall performance of the
15275 application is improved by favoring exception propagation.
15278 @geindex --RTS=zcx (gnatmake)
15280 @geindex Zero Cost Exceptions
15285 @item @code{--RTS=zcx}
15287 This switch causes the zero cost approach to be used
15288 for exception handling. If this is the default mechanism for the
15289 target (see below), then this switch is unneeded. If the default
15290 mechanism for the target is setjmp/longjmp exceptions, then
15291 this switch can be used to modify this default, and must be
15292 used for all units in the partition.
15293 This option can only be used if the zero cost approach
15294 is available for the target in use, otherwise it will generate an error.
15297 The same option @code{--RTS} must be used both for @code{gcc}
15298 and @code{gnatbind}. Passing this option to @code{gnatmake}
15299 (@ref{dc,,Switches for gnatmake}) will ensure the required consistency
15300 through the compilation and binding steps.
15302 @node Units to Sources Mapping Files,Code Generation Control,Exception Handling Control,Compiler Switches
15303 @anchor{gnat_ugn/building_executable_programs_with_gnat id29}@anchor{116}@anchor{gnat_ugn/building_executable_programs_with_gnat units-to-sources-mapping-files}@anchor{f7}
15304 @subsection Units to Sources Mapping Files
15307 @geindex -gnatem (gcc)
15312 @item @code{-gnatem=@emph{path}}
15314 A mapping file is a way to communicate to the compiler two mappings:
15315 from unit names to file names (without any directory information) and from
15316 file names to path names (with full directory information). These mappings
15317 are used by the compiler to short-circuit the path search.
15319 The use of mapping files is not required for correct operation of the
15320 compiler, but mapping files can improve efficiency, particularly when
15321 sources are read over a slow network connection. In normal operation,
15322 you need not be concerned with the format or use of mapping files,
15323 and the @code{-gnatem} switch is not a switch that you would use
15324 explicitly. It is intended primarily for use by automatic tools such as
15325 @code{gnatmake} running under the project file facility. The
15326 description here of the format of mapping files is provided
15327 for completeness and for possible use by other tools.
15329 A mapping file is a sequence of sets of three lines. In each set, the
15330 first line is the unit name, in lower case, with @code{%s} appended
15331 for specs and @code{%b} appended for bodies; the second line is the
15332 file name; and the third line is the path name.
15339 /gnat/project1/sources/main.2.ada
15342 When the switch @code{-gnatem} is specified, the compiler will
15343 create in memory the two mappings from the specified file. If there is
15344 any problem (nonexistent file, truncated file or duplicate entries),
15345 no mapping will be created.
15347 Several @code{-gnatem} switches may be specified; however, only the
15348 last one on the command line will be taken into account.
15350 When using a project file, @code{gnatmake} creates a temporary
15351 mapping file and communicates it to the compiler using this switch.
15354 @node Code Generation Control,,Units to Sources Mapping Files,Compiler Switches
15355 @anchor{gnat_ugn/building_executable_programs_with_gnat code-generation-control}@anchor{117}@anchor{gnat_ugn/building_executable_programs_with_gnat id30}@anchor{118}
15356 @subsection Code Generation Control
15359 The GCC technology provides a wide range of target dependent
15360 @code{-m} switches for controlling
15361 details of code generation with respect to different versions of
15362 architectures. This includes variations in instruction sets (e.g.,
15363 different members of the power pc family), and different requirements
15364 for optimal arrangement of instructions (e.g., different members of
15365 the x86 family). The list of available @code{-m} switches may be
15366 found in the GCC documentation.
15368 Use of these @code{-m} switches may in some cases result in improved
15371 The GNAT technology is tested and qualified without any
15372 @code{-m} switches,
15373 so generally the most reliable approach is to avoid the use of these
15374 switches. However, we generally expect most of these switches to work
15375 successfully with GNAT, and many customers have reported successful
15376 use of these options.
15378 Our general advice is to avoid the use of @code{-m} switches unless
15379 special needs lead to requirements in this area. In particular,
15380 there is no point in using @code{-m} switches to improve performance
15381 unless you actually see a performance improvement.
15383 @node Linker Switches,Binding with gnatbind,Compiler Switches,Building Executable Programs with GNAT
15384 @anchor{gnat_ugn/building_executable_programs_with_gnat linker-switches}@anchor{119}@anchor{gnat_ugn/building_executable_programs_with_gnat id31}@anchor{11a}
15385 @section Linker Switches
15388 Linker switches can be specified after @code{-largs} builder switch.
15390 @geindex -fuse-ld=name
15395 @item @code{-fuse-ld=@emph{name}}
15397 Linker to be used. The default is @code{bfd} for @code{ld.bfd},
15398 the alternative being @code{gold} for @code{ld.gold}. The later is
15399 a more recent and faster linker, but only available on GNU/Linux
15403 @node Binding with gnatbind,Linking with gnatlink,Linker Switches,Building Executable Programs with GNAT
15404 @anchor{gnat_ugn/building_executable_programs_with_gnat binding-with-gnatbind}@anchor{1d}@anchor{gnat_ugn/building_executable_programs_with_gnat id32}@anchor{11b}
15405 @section Binding with @code{gnatbind}
15410 This chapter describes the GNAT binder, @code{gnatbind}, which is used
15411 to bind compiled GNAT objects.
15413 The @code{gnatbind} program performs four separate functions:
15419 Checks that a program is consistent, in accordance with the rules in
15420 Chapter 10 of the Ada Reference Manual. In particular, error
15421 messages are generated if a program uses inconsistent versions of a
15425 Checks that an acceptable order of elaboration exists for the program
15426 and issues an error message if it cannot find an order of elaboration
15427 that satisfies the rules in Chapter 10 of the Ada Language Manual.
15430 Generates a main program incorporating the given elaboration order.
15431 This program is a small Ada package (body and spec) that
15432 must be subsequently compiled
15433 using the GNAT compiler. The necessary compilation step is usually
15434 performed automatically by @code{gnatlink}. The two most important
15435 functions of this program
15436 are to call the elaboration routines of units in an appropriate order
15437 and to call the main program.
15440 Determines the set of object files required by the given main program.
15441 This information is output in the forms of comments in the generated program,
15442 to be read by the @code{gnatlink} utility used to link the Ada application.
15446 * Running gnatbind::
15447 * Switches for gnatbind::
15448 * Command-Line Access::
15449 * Search Paths for gnatbind::
15450 * Examples of gnatbind Usage::
15454 @node Running gnatbind,Switches for gnatbind,,Binding with gnatbind
15455 @anchor{gnat_ugn/building_executable_programs_with_gnat running-gnatbind}@anchor{11c}@anchor{gnat_ugn/building_executable_programs_with_gnat id33}@anchor{11d}
15456 @subsection Running @code{gnatbind}
15459 The form of the @code{gnatbind} command is
15462 $ gnatbind [ switches ] mainprog[.ali] [ switches ]
15465 where @code{mainprog.adb} is the Ada file containing the main program
15466 unit body. @code{gnatbind} constructs an Ada
15467 package in two files whose names are
15468 @code{b~mainprog.ads}, and @code{b~mainprog.adb}.
15469 For example, if given the
15470 parameter @code{hello.ali}, for a main program contained in file
15471 @code{hello.adb}, the binder output files would be @code{b~hello.ads}
15472 and @code{b~hello.adb}.
15474 When doing consistency checking, the binder takes into consideration
15475 any source files it can locate. For example, if the binder determines
15476 that the given main program requires the package @code{Pack}, whose
15478 file is @code{pack.ali} and whose corresponding source spec file is
15479 @code{pack.ads}, it attempts to locate the source file @code{pack.ads}
15480 (using the same search path conventions as previously described for the
15481 @code{gcc} command). If it can locate this source file, it checks that
15483 or source checksums of the source and its references to in @code{ALI} files
15484 match. In other words, any @code{ALI} files that mentions this spec must have
15485 resulted from compiling this version of the source file (or in the case
15486 where the source checksums match, a version close enough that the
15487 difference does not matter).
15489 @geindex Source files
15490 @geindex use by binder
15492 The effect of this consistency checking, which includes source files, is
15493 that the binder ensures that the program is consistent with the latest
15494 version of the source files that can be located at bind time. Editing a
15495 source file without compiling files that depend on the source file cause
15496 error messages to be generated by the binder.
15498 For example, suppose you have a main program @code{hello.adb} and a
15499 package @code{P}, from file @code{p.ads} and you perform the following
15506 Enter @code{gcc -c hello.adb} to compile the main program.
15509 Enter @code{gcc -c p.ads} to compile package @code{P}.
15512 Edit file @code{p.ads}.
15515 Enter @code{gnatbind hello}.
15518 At this point, the file @code{p.ali} contains an out-of-date time stamp
15519 because the file @code{p.ads} has been edited. The attempt at binding
15520 fails, and the binder generates the following error messages:
15523 error: "hello.adb" must be recompiled ("p.ads" has been modified)
15524 error: "p.ads" has been modified and must be recompiled
15527 Now both files must be recompiled as indicated, and then the bind can
15528 succeed, generating a main program. You need not normally be concerned
15529 with the contents of this file, but for reference purposes a sample
15530 binder output file is given in @ref{e,,Example of Binder Output File}.
15532 In most normal usage, the default mode of @code{gnatbind} which is to
15533 generate the main package in Ada, as described in the previous section.
15534 In particular, this means that any Ada programmer can read and understand
15535 the generated main program. It can also be debugged just like any other
15536 Ada code provided the @code{-g} switch is used for
15537 @code{gnatbind} and @code{gnatlink}.
15539 @node Switches for gnatbind,Command-Line Access,Running gnatbind,Binding with gnatbind
15540 @anchor{gnat_ugn/building_executable_programs_with_gnat id34}@anchor{11e}@anchor{gnat_ugn/building_executable_programs_with_gnat switches-for-gnatbind}@anchor{11f}
15541 @subsection Switches for @code{gnatbind}
15544 The following switches are available with @code{gnatbind}; details will
15545 be presented in subsequent sections.
15547 @geindex --version (gnatbind)
15552 @item @code{--version}
15554 Display Copyright and version, then exit disregarding all other options.
15557 @geindex --help (gnatbind)
15562 @item @code{--help}
15564 If @code{--version} was not used, display usage, then exit disregarding
15568 @geindex -a (gnatbind)
15575 Indicates that, if supported by the platform, the adainit procedure should
15576 be treated as an initialisation routine by the linker (a constructor). This
15577 is intended to be used by the Project Manager to automatically initialize
15578 shared Stand-Alone Libraries.
15581 @geindex -aO (gnatbind)
15588 Specify directory to be searched for ALI files.
15591 @geindex -aI (gnatbind)
15598 Specify directory to be searched for source file.
15601 @geindex -A (gnatbind)
15606 @item @code{-A[=@emph{filename}]}
15608 Output ALI list (to standard output or to the named file).
15611 @geindex -b (gnatbind)
15618 Generate brief messages to @code{stderr} even if verbose mode set.
15621 @geindex -c (gnatbind)
15628 Check only, no generation of binder output file.
15631 @geindex -dnn[k|m] (gnatbind)
15636 @item @code{-d@emph{nn}[k|m]}
15638 This switch can be used to change the default task stack size value
15639 to a specified size @code{nn}, which is expressed in bytes by default, or
15640 in kilobytes when suffixed with @code{k} or in megabytes when suffixed
15642 In the absence of a @code{[k|m]} suffix, this switch is equivalent,
15643 in effect, to completing all task specs with
15646 pragma Storage_Size (nn);
15649 When they do not already have such a pragma.
15652 @geindex -D (gnatbind)
15657 @item @code{-D@emph{nn}[k|m]}
15659 Set the default secondary stack size to @code{nn}. The suffix indicates whether
15660 the size is in bytes (no suffix), kilobytes (@code{k} suffix) or megabytes
15663 The secondary stack holds objects of unconstrained types that are returned by
15664 functions, for example unconstrained Strings. The size of the secondary stack
15665 can be dynamic or fixed depending on the target.
15667 For most targets, the secondary stack grows on demand and is implemented as
15668 a chain of blocks in the heap. In this case, the default secondary stack size
15669 determines the initial size of the secondary stack for each task and the
15670 smallest amount the secondary stack can grow by.
15672 For Ravenscar, ZFP, and Cert run-times the size of the secondary stack is
15673 fixed. This switch can be used to change the default size of these stacks.
15674 The default secondary stack size can be overridden on a per-task basis if
15675 individual tasks have different secondary stack requirements. This is
15676 achieved through the Secondary_Stack_Size aspect that takes the size of the
15677 secondary stack in bytes.
15680 @geindex -e (gnatbind)
15687 Output complete list of elaboration-order dependencies.
15690 @geindex -Ea (gnatbind)
15697 Store tracebacks in exception occurrences when the target supports it.
15698 The "a" is for "address"; tracebacks will contain hexadecimal addresses,
15699 unless symbolic tracebacks are enabled.
15701 See also the packages @code{GNAT.Traceback} and
15702 @code{GNAT.Traceback.Symbolic} for more information.
15703 Note that on x86 ports, you must not use @code{-fomit-frame-pointer}
15707 @geindex -Es (gnatbind)
15714 Store tracebacks in exception occurrences when the target supports it.
15715 The "s" is for "symbolic"; symbolic tracebacks are enabled.
15718 @geindex -E (gnatbind)
15725 Currently the same as @code{-Ea}.
15728 @geindex -f (gnatbind)
15733 @item @code{-f@emph{elab-order}}
15735 Force elaboration order.
15738 @geindex -F (gnatbind)
15745 Force the checks of elaboration flags. @code{gnatbind} does not normally
15746 generate checks of elaboration flags for the main executable, except when
15747 a Stand-Alone Library is used. However, there are cases when this cannot be
15748 detected by gnatbind. An example is importing an interface of a Stand-Alone
15749 Library through a pragma Import and only specifying through a linker switch
15750 this Stand-Alone Library. This switch is used to guarantee that elaboration
15751 flag checks are generated.
15754 @geindex -h (gnatbind)
15761 Output usage (help) information.
15763 @geindex -H32 (gnatbind)
15767 Use 32-bit allocations for @code{__gnat_malloc} (and thus for access types).
15768 For further details see @ref{120,,Dynamic Allocation Control}.
15770 @geindex -H64 (gnatbind)
15772 @geindex __gnat_malloc
15776 Use 64-bit allocations for @code{__gnat_malloc} (and thus for access types).
15777 For further details see @ref{120,,Dynamic Allocation Control}.
15779 @geindex -I (gnatbind)
15783 Specify directory to be searched for source and ALI files.
15785 @geindex -I- (gnatbind)
15789 Do not look for sources in the current directory where @code{gnatbind} was
15790 invoked, and do not look for ALI files in the directory containing the
15791 ALI file named in the @code{gnatbind} command line.
15793 @geindex -l (gnatbind)
15797 Output chosen elaboration order.
15799 @geindex -L (gnatbind)
15801 @item @code{-L@emph{xxx}}
15803 Bind the units for library building. In this case the @code{adainit} and
15804 @code{adafinal} procedures (@ref{b4,,Binding with Non-Ada Main Programs})
15805 are renamed to @code{@emph{xxx}init} and
15806 @code{@emph{xxx}final}.
15808 (@ref{15,,GNAT and Libraries}, for more details.)
15810 @geindex -M (gnatbind)
15812 @item @code{-M@emph{xyz}}
15814 Rename generated main program from main to xyz. This option is
15815 supported on cross environments only.
15817 @geindex -m (gnatbind)
15819 @item @code{-m@emph{n}}
15821 Limit number of detected errors or warnings to @code{n}, where @code{n} is
15822 in the range 1..999999. The default value if no switch is
15823 given is 9999. If the number of warnings reaches this limit, then a
15824 message is output and further warnings are suppressed, the bind
15825 continues in this case. If the number of errors reaches this
15826 limit, then a message is output and the bind is abandoned.
15827 A value of zero means that no limit is enforced. The equal
15830 @geindex -n (gnatbind)
15836 @geindex -nostdinc (gnatbind)
15838 @item @code{-nostdinc}
15840 Do not look for sources in the system default directory.
15842 @geindex -nostdlib (gnatbind)
15844 @item @code{-nostdlib}
15846 Do not look for library files in the system default directory.
15848 @geindex --RTS (gnatbind)
15850 @item @code{--RTS=@emph{rts-path}}
15852 Specifies the default location of the run-time library. Same meaning as the
15853 equivalent @code{gnatmake} flag (@ref{dc,,Switches for gnatmake}).
15855 @geindex -o (gnatbind)
15857 @item @code{-o @emph{file}}
15859 Name the output file @code{file} (default is @code{b~`xxx}.adb`).
15860 Note that if this option is used, then linking must be done manually,
15861 gnatlink cannot be used.
15863 @geindex -O (gnatbind)
15865 @item @code{-O[=@emph{filename}]}
15867 Output object list (to standard output or to the named file).
15869 @geindex -p (gnatbind)
15873 Pessimistic (worst-case) elaboration order.
15875 @geindex -P (gnatbind)
15879 Generate binder file suitable for CodePeer.
15881 @geindex -R (gnatbind)
15885 Output closure source list, which includes all non-run-time units that are
15886 included in the bind.
15888 @geindex -Ra (gnatbind)
15892 Like @code{-R} but the list includes run-time units.
15894 @geindex -s (gnatbind)
15898 Require all source files to be present.
15900 @geindex -S (gnatbind)
15902 @item @code{-S@emph{xxx}}
15904 Specifies the value to be used when detecting uninitialized scalar
15905 objects with pragma Initialize_Scalars.
15906 The @code{xxx} string specified with the switch is one of:
15912 @code{in} for an invalid value.
15914 If zero is invalid for the discrete type in question,
15915 then the scalar value is set to all zero bits.
15916 For signed discrete types, the largest possible negative value of
15917 the underlying scalar is set (i.e. a one bit followed by all zero bits).
15918 For unsigned discrete types, the underlying scalar value is set to all
15919 one bits. For floating-point types, a NaN value is set
15920 (see body of package System.Scalar_Values for exact values).
15923 @code{lo} for low value.
15925 If zero is invalid for the discrete type in question,
15926 then the scalar value is set to all zero bits.
15927 For signed discrete types, the largest possible negative value of
15928 the underlying scalar is set (i.e. a one bit followed by all zero bits).
15929 For unsigned discrete types, the underlying scalar value is set to all
15930 zero bits. For floating-point, a small value is set
15931 (see body of package System.Scalar_Values for exact values).
15934 @code{hi} for high value.
15936 If zero is invalid for the discrete type in question,
15937 then the scalar value is set to all one bits.
15938 For signed discrete types, the largest possible positive value of
15939 the underlying scalar is set (i.e. a zero bit followed by all one bits).
15940 For unsigned discrete types, the underlying scalar value is set to all
15941 one bits. For floating-point, a large value is set
15942 (see body of package System.Scalar_Values for exact values).
15945 @code{xx} for hex value (two hex digits).
15947 The underlying scalar is set to a value consisting of repeated bytes, whose
15948 value corresponds to the given value. For example if @code{BF} is given,
15949 then a 32-bit scalar value will be set to the bit patterm @code{16#BFBFBFBF#}.
15952 @geindex GNAT_INIT_SCALARS
15954 In addition, you can specify @code{-Sev} to indicate that the value is
15955 to be set at run time. In this case, the program will look for an environment
15956 variable of the form @code{GNAT_INIT_SCALARS=@emph{yy}}, where @code{yy} is one
15957 of @code{in/lo/hi/@emph{xx}} with the same meanings as above.
15958 If no environment variable is found, or if it does not have a valid value,
15959 then the default is @code{in} (invalid values).
15962 @geindex -static (gnatbind)
15967 @item @code{-static}
15969 Link against a static GNAT run-time.
15971 @geindex -shared (gnatbind)
15973 @item @code{-shared}
15975 Link against a shared GNAT run-time when available.
15977 @geindex -t (gnatbind)
15981 Tolerate time stamp and other consistency errors.
15983 @geindex -T (gnatbind)
15985 @item @code{-T@emph{n}}
15987 Set the time slice value to @code{n} milliseconds. If the system supports
15988 the specification of a specific time slice value, then the indicated value
15989 is used. If the system does not support specific time slice values, but
15990 does support some general notion of round-robin scheduling, then any
15991 nonzero value will activate round-robin scheduling.
15993 A value of zero is treated specially. It turns off time
15994 slicing, and in addition, indicates to the tasking run-time that the
15995 semantics should match as closely as possible the Annex D
15996 requirements of the Ada RM, and in particular sets the default
15997 scheduling policy to @code{FIFO_Within_Priorities}.
15999 @geindex -u (gnatbind)
16001 @item @code{-u@emph{n}}
16003 Enable dynamic stack usage, with @code{n} results stored and displayed
16004 at program termination. A result is generated when a task
16005 terminates. Results that can't be stored are displayed on the fly, at
16006 task termination. This option is currently not supported on Itanium
16007 platforms. (See @ref{121,,Dynamic Stack Usage Analysis} for details.)
16009 @geindex -v (gnatbind)
16013 Verbose mode. Write error messages, header, summary output to
16016 @geindex -V (gnatbind)
16018 @item @code{-V@emph{key}=@emph{value}}
16020 Store the given association of @code{key} to @code{value} in the bind environment.
16021 Values stored this way can be retrieved at run time using
16022 @code{GNAT.Bind_Environment}.
16024 @geindex -w (gnatbind)
16026 @item @code{-w@emph{x}}
16028 Warning mode; @code{x} = s/e for suppress/treat as error.
16030 @geindex -Wx (gnatbind)
16032 @item @code{-Wx@emph{e}}
16034 Override default wide character encoding for standard Text_IO files.
16036 @geindex -x (gnatbind)
16040 Exclude source files (check object consistency only).
16042 @geindex -Xnnn (gnatbind)
16044 @item @code{-X@emph{nnn}}
16046 Set default exit status value, normally 0 for POSIX compliance.
16048 @geindex -y (gnatbind)
16052 Enable leap seconds support in @code{Ada.Calendar} and its children.
16054 @geindex -z (gnatbind)
16058 No main subprogram.
16061 You may obtain this listing of switches by running @code{gnatbind} with
16065 * Consistency-Checking Modes::
16066 * Binder Error Message Control::
16067 * Elaboration Control::
16069 * Dynamic Allocation Control::
16070 * Binding with Non-Ada Main Programs::
16071 * Binding Programs with No Main Subprogram::
16075 @node Consistency-Checking Modes,Binder Error Message Control,,Switches for gnatbind
16076 @anchor{gnat_ugn/building_executable_programs_with_gnat consistency-checking-modes}@anchor{122}@anchor{gnat_ugn/building_executable_programs_with_gnat id35}@anchor{123}
16077 @subsubsection Consistency-Checking Modes
16080 As described earlier, by default @code{gnatbind} checks
16081 that object files are consistent with one another and are consistent
16082 with any source files it can locate. The following switches control binder
16087 @geindex -s (gnatbind)
16095 Require source files to be present. In this mode, the binder must be
16096 able to locate all source files that are referenced, in order to check
16097 their consistency. In normal mode, if a source file cannot be located it
16098 is simply ignored. If you specify this switch, a missing source
16101 @geindex -Wx (gnatbind)
16103 @item @code{-Wx@emph{e}}
16105 Override default wide character encoding for standard Text_IO files.
16106 Normally the default wide character encoding method used for standard
16107 [Wide_[Wide_]]Text_IO files is taken from the encoding specified for
16108 the main source input (see description of switch
16109 @code{-gnatWx} for the compiler). The
16110 use of this switch for the binder (which has the same set of
16111 possible arguments) overrides this default as specified.
16113 @geindex -x (gnatbind)
16117 Exclude source files. In this mode, the binder only checks that ALI
16118 files are consistent with one another. Source files are not accessed.
16119 The binder runs faster in this mode, and there is still a guarantee that
16120 the resulting program is self-consistent.
16121 If a source file has been edited since it was last compiled, and you
16122 specify this switch, the binder will not detect that the object
16123 file is out of date with respect to the source file. Note that this is the
16124 mode that is automatically used by @code{gnatmake} because in this
16125 case the checking against sources has already been performed by
16126 @code{gnatmake} in the course of compilation (i.e., before binding).
16129 @node Binder Error Message Control,Elaboration Control,Consistency-Checking Modes,Switches for gnatbind
16130 @anchor{gnat_ugn/building_executable_programs_with_gnat id36}@anchor{124}@anchor{gnat_ugn/building_executable_programs_with_gnat binder-error-message-control}@anchor{125}
16131 @subsubsection Binder Error Message Control
16134 The following switches provide control over the generation of error
16135 messages from the binder:
16139 @geindex -v (gnatbind)
16147 Verbose mode. In the normal mode, brief error messages are generated to
16148 @code{stderr}. If this switch is present, a header is written
16149 to @code{stdout} and any error messages are directed to @code{stdout}.
16150 All that is written to @code{stderr} is a brief summary message.
16152 @geindex -b (gnatbind)
16156 Generate brief error messages to @code{stderr} even if verbose mode is
16157 specified. This is relevant only when used with the
16160 @geindex -m (gnatbind)
16162 @item @code{-m@emph{n}}
16164 Limits the number of error messages to @code{n}, a decimal integer in the
16165 range 1-999. The binder terminates immediately if this limit is reached.
16167 @geindex -M (gnatbind)
16169 @item @code{-M@emph{xxx}}
16171 Renames the generated main program from @code{main} to @code{xxx}.
16172 This is useful in the case of some cross-building environments, where
16173 the actual main program is separate from the one generated
16174 by @code{gnatbind}.
16176 @geindex -ws (gnatbind)
16182 Suppress all warning messages.
16184 @geindex -we (gnatbind)
16188 Treat any warning messages as fatal errors.
16190 @geindex -t (gnatbind)
16192 @geindex Time stamp checks
16195 @geindex Binder consistency checks
16197 @geindex Consistency checks
16202 The binder performs a number of consistency checks including:
16208 Check that time stamps of a given source unit are consistent
16211 Check that checksums of a given source unit are consistent
16214 Check that consistent versions of @code{GNAT} were used for compilation
16217 Check consistency of configuration pragmas as required
16220 Normally failure of such checks, in accordance with the consistency
16221 requirements of the Ada Reference Manual, causes error messages to be
16222 generated which abort the binder and prevent the output of a binder
16223 file and subsequent link to obtain an executable.
16225 The @code{-t} switch converts these error messages
16226 into warnings, so that
16227 binding and linking can continue to completion even in the presence of such
16228 errors. The result may be a failed link (due to missing symbols), or a
16229 non-functional executable which has undefined semantics.
16233 This means that @code{-t} should be used only in unusual situations,
16239 @node Elaboration Control,Output Control,Binder Error Message Control,Switches for gnatbind
16240 @anchor{gnat_ugn/building_executable_programs_with_gnat id37}@anchor{126}@anchor{gnat_ugn/building_executable_programs_with_gnat elaboration-control}@anchor{127}
16241 @subsubsection Elaboration Control
16244 The following switches provide additional control over the elaboration
16245 order. For full details see @ref{f,,Elaboration Order Handling in GNAT}.
16247 @geindex -f (gnatbind)
16252 @item @code{-f@emph{elab-order}}
16254 Force elaboration order.
16256 @code{elab-order} should be the name of a "forced elaboration order file", that
16257 is, a text file containing library item names, one per line. A name of the
16258 form "some.unit%s" or "some.unit (spec)" denotes the spec of Some.Unit. A
16259 name of the form "some.unit%b" or "some.unit (body)" denotes the body of
16260 Some.Unit. Each pair of lines is taken to mean that there is an elaboration
16261 dependence of the second line on the first. For example, if the file
16271 then the spec of This will be elaborated before the body of This, and the
16272 body of This will be elaborated before the spec of That, and the spec of That
16273 will be elaborated before the body of That. The first and last of these three
16274 dependences are already required by Ada rules, so this file is really just
16275 forcing the body of This to be elaborated before the spec of That.
16277 The given order must be consistent with Ada rules, or else @code{gnatbind} will
16278 give elaboration cycle errors. For example, if you say x (body) should be
16279 elaborated before x (spec), there will be a cycle, because Ada rules require
16280 x (spec) to be elaborated before x (body); you can't have the spec and body
16281 both elaborated before each other.
16283 If you later add "with That;" to the body of This, there will be a cycle, in
16284 which case you should erase either "this (body)" or "that (spec)" from the
16285 above forced elaboration order file.
16287 Blank lines and Ada-style comments are ignored. Unit names that do not exist
16288 in the program are ignored. Units in the GNAT predefined library are also
16291 @geindex -p (gnatbind)
16295 Normally the binder attempts to choose an elaboration order that is
16296 likely to minimize the likelihood of an elaboration order error resulting
16297 in raising a @code{Program_Error} exception. This switch reverses the
16298 action of the binder, and requests that it deliberately choose an order
16299 that is likely to maximize the likelihood of an elaboration error.
16300 This is useful in ensuring portability and avoiding dependence on
16301 accidental fortuitous elaboration ordering.
16303 Normally it only makes sense to use the @code{-p}
16305 elaboration checking is used (@code{-gnatE} switch used for compilation).
16306 This is because in the default static elaboration mode, all necessary
16307 @code{Elaborate} and @code{Elaborate_All} pragmas are implicitly inserted.
16308 These implicit pragmas are still respected by the binder in
16309 @code{-p} mode, so a
16310 safe elaboration order is assured.
16312 Note that @code{-p} is not intended for
16313 production use; it is more for debugging/experimental use.
16316 @node Output Control,Dynamic Allocation Control,Elaboration Control,Switches for gnatbind
16317 @anchor{gnat_ugn/building_executable_programs_with_gnat output-control}@anchor{128}@anchor{gnat_ugn/building_executable_programs_with_gnat id38}@anchor{129}
16318 @subsubsection Output Control
16321 The following switches allow additional control over the output
16322 generated by the binder.
16326 @geindex -c (gnatbind)
16334 Check only. Do not generate the binder output file. In this mode the
16335 binder performs all error checks but does not generate an output file.
16337 @geindex -e (gnatbind)
16341 Output complete list of elaboration-order dependencies, showing the
16342 reason for each dependency. This output can be rather extensive but may
16343 be useful in diagnosing problems with elaboration order. The output is
16344 written to @code{stdout}.
16346 @geindex -h (gnatbind)
16350 Output usage information. The output is written to @code{stdout}.
16352 @geindex -K (gnatbind)
16356 Output linker options to @code{stdout}. Includes library search paths,
16357 contents of pragmas Ident and Linker_Options, and libraries added
16358 by @code{gnatbind}.
16360 @geindex -l (gnatbind)
16364 Output chosen elaboration order. The output is written to @code{stdout}.
16366 @geindex -O (gnatbind)
16370 Output full names of all the object files that must be linked to provide
16371 the Ada component of the program. The output is written to @code{stdout}.
16372 This list includes the files explicitly supplied and referenced by the user
16373 as well as implicitly referenced run-time unit files. The latter are
16374 omitted if the corresponding units reside in shared libraries. The
16375 directory names for the run-time units depend on the system configuration.
16377 @geindex -o (gnatbind)
16379 @item @code{-o @emph{file}}
16381 Set name of output file to @code{file} instead of the normal
16382 @code{b~`mainprog}.adb` default. Note that @code{file} denote the Ada
16383 binder generated body filename.
16384 Note that if this option is used, then linking must be done manually.
16385 It is not possible to use gnatlink in this case, since it cannot locate
16388 @geindex -r (gnatbind)
16392 Generate list of @code{pragma Restrictions} that could be applied to
16393 the current unit. This is useful for code audit purposes, and also may
16394 be used to improve code generation in some cases.
16397 @node Dynamic Allocation Control,Binding with Non-Ada Main Programs,Output Control,Switches for gnatbind
16398 @anchor{gnat_ugn/building_executable_programs_with_gnat dynamic-allocation-control}@anchor{120}@anchor{gnat_ugn/building_executable_programs_with_gnat id39}@anchor{12a}
16399 @subsubsection Dynamic Allocation Control
16402 The heap control switches -- @code{-H32} and @code{-H64} --
16403 determine whether dynamic allocation uses 32-bit or 64-bit memory.
16404 They only affect compiler-generated allocations via @code{__gnat_malloc};
16405 explicit calls to @code{malloc} and related functions from the C
16406 run-time library are unaffected.
16413 Allocate memory on 32-bit heap
16417 Allocate memory on 64-bit heap. This is the default
16418 unless explicitly overridden by a @code{'Size} clause on the access type.
16421 These switches are only effective on VMS platforms.
16423 @node Binding with Non-Ada Main Programs,Binding Programs with No Main Subprogram,Dynamic Allocation Control,Switches for gnatbind
16424 @anchor{gnat_ugn/building_executable_programs_with_gnat binding-with-non-ada-main-programs}@anchor{b4}@anchor{gnat_ugn/building_executable_programs_with_gnat id40}@anchor{12b}
16425 @subsubsection Binding with Non-Ada Main Programs
16428 The description so far has assumed that the main
16429 program is in Ada, and that the task of the binder is to generate a
16430 corresponding function @code{main} that invokes this Ada main
16431 program. GNAT also supports the building of executable programs where
16432 the main program is not in Ada, but some of the called routines are
16433 written in Ada and compiled using GNAT (@ref{44,,Mixed Language Programming}).
16434 The following switch is used in this situation:
16438 @geindex -n (gnatbind)
16446 No main program. The main program is not in Ada.
16449 In this case, most of the functions of the binder are still required,
16450 but instead of generating a main program, the binder generates a file
16451 containing the following callable routines:
16460 @item @code{adainit}
16462 You must call this routine to initialize the Ada part of the program by
16463 calling the necessary elaboration routines. A call to @code{adainit} is
16464 required before the first call to an Ada subprogram.
16466 Note that it is assumed that the basic execution environment must be setup
16467 to be appropriate for Ada execution at the point where the first Ada
16468 subprogram is called. In particular, if the Ada code will do any
16469 floating-point operations, then the FPU must be setup in an appropriate
16470 manner. For the case of the x86, for example, full precision mode is
16471 required. The procedure GNAT.Float_Control.Reset may be used to ensure
16472 that the FPU is in the right state.
16480 @item @code{adafinal}
16482 You must call this routine to perform any library-level finalization
16483 required by the Ada subprograms. A call to @code{adafinal} is required
16484 after the last call to an Ada subprogram, and before the program
16489 @geindex -n (gnatbind)
16492 @geindex multiple input files
16494 If the @code{-n} switch
16495 is given, more than one ALI file may appear on
16496 the command line for @code{gnatbind}. The normal @code{closure}
16497 calculation is performed for each of the specified units. Calculating
16498 the closure means finding out the set of units involved by tracing
16499 @emph{with} references. The reason it is necessary to be able to
16500 specify more than one ALI file is that a given program may invoke two or
16501 more quite separate groups of Ada units.
16503 The binder takes the name of its output file from the last specified ALI
16504 file, unless overridden by the use of the @code{-o file}.
16506 @geindex -o (gnatbind)
16508 The output is an Ada unit in source form that can be compiled with GNAT.
16509 This compilation occurs automatically as part of the @code{gnatlink}
16512 Currently the GNAT run-time requires a FPU using 80 bits mode
16513 precision. Under targets where this is not the default it is required to
16514 call GNAT.Float_Control.Reset before using floating point numbers (this
16515 include float computation, float input and output) in the Ada code. A
16516 side effect is that this could be the wrong mode for the foreign code
16517 where floating point computation could be broken after this call.
16519 @node Binding Programs with No Main Subprogram,,Binding with Non-Ada Main Programs,Switches for gnatbind
16520 @anchor{gnat_ugn/building_executable_programs_with_gnat binding-programs-with-no-main-subprogram}@anchor{12c}@anchor{gnat_ugn/building_executable_programs_with_gnat id41}@anchor{12d}
16521 @subsubsection Binding Programs with No Main Subprogram
16524 It is possible to have an Ada program which does not have a main
16525 subprogram. This program will call the elaboration routines of all the
16526 packages, then the finalization routines.
16528 The following switch is used to bind programs organized in this manner:
16532 @geindex -z (gnatbind)
16540 Normally the binder checks that the unit name given on the command line
16541 corresponds to a suitable main subprogram. When this switch is used,
16542 a list of ALI files can be given, and the execution of the program
16543 consists of elaboration of these units in an appropriate order. Note
16544 that the default wide character encoding method for standard Text_IO
16545 files is always set to Brackets if this switch is set (you can use
16547 @code{-Wx} to override this default).
16550 @node Command-Line Access,Search Paths for gnatbind,Switches for gnatbind,Binding with gnatbind
16551 @anchor{gnat_ugn/building_executable_programs_with_gnat id42}@anchor{12e}@anchor{gnat_ugn/building_executable_programs_with_gnat command-line-access}@anchor{12f}
16552 @subsection Command-Line Access
16555 The package @code{Ada.Command_Line} provides access to the command-line
16556 arguments and program name. In order for this interface to operate
16557 correctly, the two variables
16568 are declared in one of the GNAT library routines. These variables must
16569 be set from the actual @code{argc} and @code{argv} values passed to the
16570 main program. With no @emph{n} present, @code{gnatbind}
16571 generates the C main program to automatically set these variables.
16572 If the @emph{n} switch is used, there is no automatic way to
16573 set these variables. If they are not set, the procedures in
16574 @code{Ada.Command_Line} will not be available, and any attempt to use
16575 them will raise @code{Constraint_Error}. If command line access is
16576 required, your main program must set @code{gnat_argc} and
16577 @code{gnat_argv} from the @code{argc} and @code{argv} values passed to
16580 @node Search Paths for gnatbind,Examples of gnatbind Usage,Command-Line Access,Binding with gnatbind
16581 @anchor{gnat_ugn/building_executable_programs_with_gnat search-paths-for-gnatbind}@anchor{8c}@anchor{gnat_ugn/building_executable_programs_with_gnat id43}@anchor{130}
16582 @subsection Search Paths for @code{gnatbind}
16585 The binder takes the name of an ALI file as its argument and needs to
16586 locate source files as well as other ALI files to verify object consistency.
16588 For source files, it follows exactly the same search rules as @code{gcc}
16589 (see @ref{89,,Search Paths and the Run-Time Library (RTL)}). For ALI files the
16590 directories searched are:
16596 The directory containing the ALI file named in the command line, unless
16597 the switch @code{-I-} is specified.
16600 All directories specified by @code{-I}
16601 switches on the @code{gnatbind}
16602 command line, in the order given.
16604 @geindex ADA_PRJ_OBJECTS_FILE
16607 Each of the directories listed in the text file whose name is given
16609 @geindex ADA_PRJ_OBJECTS_FILE
16610 @geindex environment variable; ADA_PRJ_OBJECTS_FILE
16611 @code{ADA_PRJ_OBJECTS_FILE} environment variable.
16613 @geindex ADA_PRJ_OBJECTS_FILE
16614 @geindex environment variable; ADA_PRJ_OBJECTS_FILE
16615 @code{ADA_PRJ_OBJECTS_FILE} is normally set by gnatmake or by the gnat
16616 driver when project files are used. It should not normally be set
16619 @geindex ADA_OBJECTS_PATH
16622 Each of the directories listed in the value of the
16623 @geindex ADA_OBJECTS_PATH
16624 @geindex environment variable; ADA_OBJECTS_PATH
16625 @code{ADA_OBJECTS_PATH} environment variable.
16626 Construct this value
16629 @geindex environment variable; PATH
16630 @code{PATH} environment variable: a list of directory
16631 names separated by colons (semicolons when working with the NT version
16635 The content of the @code{ada_object_path} file which is part of the GNAT
16636 installation tree and is used to store standard libraries such as the
16637 GNAT Run-Time Library (RTL) unless the switch @code{-nostdlib} is
16638 specified. See @ref{87,,Installing a library}
16641 @geindex -I (gnatbind)
16643 @geindex -aI (gnatbind)
16645 @geindex -aO (gnatbind)
16647 In the binder the switch @code{-I}
16648 is used to specify both source and
16649 library file paths. Use @code{-aI}
16650 instead if you want to specify
16651 source paths only, and @code{-aO}
16652 if you want to specify library paths
16653 only. This means that for the binder
16654 @code{-I@emph{dir}} is equivalent to
16655 @code{-aI@emph{dir}}
16656 @code{-aO`@emph{dir}}.
16657 The binder generates the bind file (a C language source file) in the
16658 current working directory.
16664 @geindex Interfaces
16668 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
16669 children make up the GNAT Run-Time Library, together with the package
16670 GNAT and its children, which contain a set of useful additional
16671 library functions provided by GNAT. The sources for these units are
16672 needed by the compiler and are kept together in one directory. The ALI
16673 files and object files generated by compiling the RTL are needed by the
16674 binder and the linker and are kept together in one directory, typically
16675 different from the directory containing the sources. In a normal
16676 installation, you need not specify these directory names when compiling
16677 or binding. Either the environment variables or the built-in defaults
16678 cause these files to be found.
16680 Besides simplifying access to the RTL, a major use of search paths is
16681 in compiling sources from multiple directories. This can make
16682 development environments much more flexible.
16684 @node Examples of gnatbind Usage,,Search Paths for gnatbind,Binding with gnatbind
16685 @anchor{gnat_ugn/building_executable_programs_with_gnat id44}@anchor{131}@anchor{gnat_ugn/building_executable_programs_with_gnat examples-of-gnatbind-usage}@anchor{132}
16686 @subsection Examples of @code{gnatbind} Usage
16689 Here are some examples of @code{gnatbind} invovations:
16697 The main program @code{Hello} (source program in @code{hello.adb}) is
16698 bound using the standard switch settings. The generated main program is
16699 @code{b~hello.adb}. This is the normal, default use of the binder.
16702 gnatbind hello -o mainprog.adb
16705 The main program @code{Hello} (source program in @code{hello.adb}) is
16706 bound using the standard switch settings. The generated main program is
16707 @code{mainprog.adb} with the associated spec in
16708 @code{mainprog.ads}. Note that you must specify the body here not the
16709 spec. Note that if this option is used, then linking must be done manually,
16710 since gnatlink will not be able to find the generated file.
16713 @node Linking with gnatlink,Using the GNU make Utility,Binding with gnatbind,Building Executable Programs with GNAT
16714 @anchor{gnat_ugn/building_executable_programs_with_gnat id45}@anchor{133}@anchor{gnat_ugn/building_executable_programs_with_gnat linking-with-gnatlink}@anchor{1e}
16715 @section Linking with @code{gnatlink}
16720 This chapter discusses @code{gnatlink}, a tool that links
16721 an Ada program and builds an executable file. This utility
16722 invokes the system linker (via the @code{gcc} command)
16723 with a correct list of object files and library references.
16724 @code{gnatlink} automatically determines the list of files and
16725 references for the Ada part of a program. It uses the binder file
16726 generated by the @code{gnatbind} to determine this list.
16729 * Running gnatlink::
16730 * Switches for gnatlink::
16734 @node Running gnatlink,Switches for gnatlink,,Linking with gnatlink
16735 @anchor{gnat_ugn/building_executable_programs_with_gnat id46}@anchor{134}@anchor{gnat_ugn/building_executable_programs_with_gnat running-gnatlink}@anchor{135}
16736 @subsection Running @code{gnatlink}
16739 The form of the @code{gnatlink} command is
16742 $ gnatlink [ switches ] mainprog [.ali]
16743 [ non-Ada objects ] [ linker options ]
16746 The arguments of @code{gnatlink} (switches, main @code{ALI} file,
16748 or linker options) may be in any order, provided that no non-Ada object may
16749 be mistaken for a main @code{ALI} file.
16750 Any file name @code{F} without the @code{.ali}
16751 extension will be taken as the main @code{ALI} file if a file exists
16752 whose name is the concatenation of @code{F} and @code{.ali}.
16754 @code{mainprog.ali} references the ALI file of the main program.
16755 The @code{.ali} extension of this file can be omitted. From this
16756 reference, @code{gnatlink} locates the corresponding binder file
16757 @code{b~mainprog.adb} and, using the information in this file along
16758 with the list of non-Ada objects and linker options, constructs a
16759 linker command file to create the executable.
16761 The arguments other than the @code{gnatlink} switches and the main
16762 @code{ALI} file are passed to the linker uninterpreted.
16763 They typically include the names of
16764 object files for units written in other languages than Ada and any library
16765 references required to resolve references in any of these foreign language
16766 units, or in @code{Import} pragmas in any Ada units.
16768 @code{linker options} is an optional list of linker specific
16770 The default linker called by gnatlink is @code{gcc} which in
16771 turn calls the appropriate system linker.
16773 One useful option for the linker is @code{-s}: it reduces the size of the
16774 executable by removing all symbol table and relocation information from the
16777 Standard options for the linker such as @code{-lmy_lib} or
16778 @code{-Ldir} can be added as is.
16779 For options that are not recognized by
16780 @code{gcc} as linker options, use the @code{gcc} switches
16781 @code{-Xlinker} or @code{-Wl,}.
16783 Refer to the GCC documentation for
16786 Here is an example showing how to generate a linker map:
16789 $ gnatlink my_prog -Wl,-Map,MAPFILE
16792 Using @code{linker options} it is possible to set the program stack and
16794 See @ref{136,,Setting Stack Size from gnatlink} and
16795 @ref{137,,Setting Heap Size from gnatlink}.
16797 @code{gnatlink} determines the list of objects required by the Ada
16798 program and prepends them to the list of objects passed to the linker.
16799 @code{gnatlink} also gathers any arguments set by the use of
16800 @code{pragma Linker_Options} and adds them to the list of arguments
16801 presented to the linker.
16803 @node Switches for gnatlink,,Running gnatlink,Linking with gnatlink
16804 @anchor{gnat_ugn/building_executable_programs_with_gnat id47}@anchor{138}@anchor{gnat_ugn/building_executable_programs_with_gnat switches-for-gnatlink}@anchor{139}
16805 @subsection Switches for @code{gnatlink}
16808 The following switches are available with the @code{gnatlink} utility:
16810 @geindex --version (gnatlink)
16815 @item @code{--version}
16817 Display Copyright and version, then exit disregarding all other options.
16820 @geindex --help (gnatlink)
16825 @item @code{--help}
16827 If @code{--version} was not used, display usage, then exit disregarding
16831 @geindex Command line length
16833 @geindex -f (gnatlink)
16840 On some targets, the command line length is limited, and @code{gnatlink}
16841 will generate a separate file for the linker if the list of object files
16843 The @code{-f} switch forces this file
16844 to be generated even if
16845 the limit is not exceeded. This is useful in some cases to deal with
16846 special situations where the command line length is exceeded.
16849 @geindex Debugging information
16852 @geindex -g (gnatlink)
16859 The option to include debugging information causes the Ada bind file (in
16860 other words, @code{b~mainprog.adb}) to be compiled with @code{-g}.
16861 In addition, the binder does not delete the @code{b~mainprog.adb},
16862 @code{b~mainprog.o} and @code{b~mainprog.ali} files.
16863 Without @code{-g}, the binder removes these files by default.
16866 @geindex -n (gnatlink)
16873 Do not compile the file generated by the binder. This may be used when
16874 a link is rerun with different options, but there is no need to recompile
16878 @geindex -v (gnatlink)
16885 Verbose mode. Causes additional information to be output, including a full
16886 list of the included object files.
16887 This switch option is most useful when you want
16888 to see what set of object files are being used in the link step.
16891 @geindex -v -v (gnatlink)
16898 Very verbose mode. Requests that the compiler operate in verbose mode when
16899 it compiles the binder file, and that the system linker run in verbose mode.
16902 @geindex -o (gnatlink)
16907 @item @code{-o @emph{exec-name}}
16909 @code{exec-name} specifies an alternate name for the generated
16910 executable program. If this switch is omitted, the executable has the same
16911 name as the main unit. For example, @code{gnatlink try.ali} creates
16912 an executable called @code{try}.
16915 @geindex -B (gnatlink)
16920 @item @code{-B@emph{dir}}
16922 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
16923 from @code{dir} instead of the default location. Only use this switch
16924 when multiple versions of the GNAT compiler are available.
16925 See the @code{Directory Options} section in @cite{The_GNU_Compiler_Collection}
16926 for further details. You would normally use the @code{-b} or
16927 @code{-V} switch instead.
16930 @geindex -M (gnatlink)
16937 When linking an executable, create a map file. The name of the map file
16938 has the same name as the executable with extension ".map".
16941 @geindex -M= (gnatlink)
16946 @item @code{-M=@emph{mapfile}}
16948 When linking an executable, create a map file. The name of the map file is
16952 @geindex --GCC=compiler_name (gnatlink)
16957 @item @code{--GCC=@emph{compiler_name}}
16959 Program used for compiling the binder file. The default is
16960 @code{gcc}. You need to use quotes around @code{compiler_name} if
16961 @code{compiler_name} contains spaces or other separator characters.
16962 As an example @code{--GCC="foo -x -y"} will instruct @code{gnatlink} to
16963 use @code{foo -x -y} as your compiler. Note that switch @code{-c} is always
16964 inserted after your command name. Thus in the above example the compiler
16965 command that will be used by @code{gnatlink} will be @code{foo -c -x -y}.
16966 A limitation of this syntax is that the name and path name of the executable
16967 itself must not include any embedded spaces. If the compiler executable is
16968 different from the default one (gcc or <prefix>-gcc), then the back-end
16969 switches in the ALI file are not used to compile the binder generated source.
16970 For example, this is the case with @code{--GCC="foo -x -y"}. But the back end
16971 switches will be used for @code{--GCC="gcc -gnatv"}. If several
16972 @code{--GCC=compiler_name} are used, only the last @code{compiler_name}
16973 is taken into account. However, all the additional switches are also taken
16974 into account. Thus,
16975 @code{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
16976 @code{--GCC="bar -x -y -z -t"}.
16979 @geindex --LINK= (gnatlink)
16984 @item @code{--LINK=@emph{name}}
16986 @code{name} is the name of the linker to be invoked. This is especially
16987 useful in mixed language programs since languages such as C++ require
16988 their own linker to be used. When this switch is omitted, the default
16989 name for the linker is @code{gcc}. When this switch is used, the
16990 specified linker is called instead of @code{gcc} with exactly the same
16991 parameters that would have been passed to @code{gcc} so if the desired
16992 linker requires different parameters it is necessary to use a wrapper
16993 script that massages the parameters before invoking the real linker. It
16994 may be useful to control the exact invocation by using the verbose
16998 @node Using the GNU make Utility,,Linking with gnatlink,Building Executable Programs with GNAT
16999 @anchor{gnat_ugn/building_executable_programs_with_gnat using-the-gnu-make-utility}@anchor{1f}@anchor{gnat_ugn/building_executable_programs_with_gnat id48}@anchor{13a}
17000 @section Using the GNU @code{make} Utility
17003 @geindex make (GNU)
17006 This chapter offers some examples of makefiles that solve specific
17007 problems. It does not explain how to write a makefile, nor does it try to replace the
17008 @code{gnatmake} utility (@ref{1b,,Building with gnatmake}).
17010 All the examples in this section are specific to the GNU version of
17011 make. Although @code{make} is a standard utility, and the basic language
17012 is the same, these examples use some advanced features found only in
17016 * Using gnatmake in a Makefile::
17017 * Automatically Creating a List of Directories::
17018 * Generating the Command Line Switches::
17019 * Overcoming Command Line Length Limits::
17023 @node Using gnatmake in a Makefile,Automatically Creating a List of Directories,,Using the GNU make Utility
17024 @anchor{gnat_ugn/building_executable_programs_with_gnat using-gnatmake-in-a-makefile}@anchor{13b}@anchor{gnat_ugn/building_executable_programs_with_gnat id49}@anchor{13c}
17025 @subsection Using gnatmake in a Makefile
17028 @c index makefile (GNU make)
17030 Complex project organizations can be handled in a very powerful way by
17031 using GNU make combined with gnatmake. For instance, here is a Makefile
17032 which allows you to build each subsystem of a big project into a separate
17033 shared library. Such a makefile allows you to significantly reduce the link
17034 time of very big applications while maintaining full coherence at
17035 each step of the build process.
17037 The list of dependencies are handled automatically by
17038 @code{gnatmake}. The Makefile is simply used to call gnatmake in each of
17039 the appropriate directories.
17041 Note that you should also read the example on how to automatically
17042 create the list of directories
17043 (@ref{13d,,Automatically Creating a List of Directories})
17044 which might help you in case your project has a lot of subdirectories.
17047 ## This Makefile is intended to be used with the following directory
17049 ## - The sources are split into a series of csc (computer software components)
17050 ## Each of these csc is put in its own directory.
17051 ## Their name are referenced by the directory names.
17052 ## They will be compiled into shared library (although this would also work
17053 ## with static libraries
17054 ## - The main program (and possibly other packages that do not belong to any
17055 ## csc is put in the top level directory (where the Makefile is).
17056 ## toplevel_dir __ first_csc (sources) __ lib (will contain the library)
17057 ## \\_ second_csc (sources) __ lib (will contain the library)
17059 ## Although this Makefile is build for shared library, it is easy to modify
17060 ## to build partial link objects instead (modify the lines with -shared and
17063 ## With this makefile, you can change any file in the system or add any new
17064 ## file, and everything will be recompiled correctly (only the relevant shared
17065 ## objects will be recompiled, and the main program will be re-linked).
17067 # The list of computer software component for your project. This might be
17068 # generated automatically.
17071 # Name of the main program (no extension)
17074 # If we need to build objects with -fPIC, uncomment the following line
17077 # The following variable should give the directory containing libgnat.so
17078 # You can get this directory through 'gnatls -v'. This is usually the last
17079 # directory in the Object_Path.
17082 # The directories for the libraries
17083 # (This macro expands the list of CSC to the list of shared libraries, you
17084 # could simply use the expanded form:
17085 # LIB_DIR=aa/lib/libaa.so bb/lib/libbb.so cc/lib/libcc.so
17086 LIB_DIR=$@{foreach dir,$@{CSC_LIST@},$@{dir@}/lib/lib$@{dir@}.so@}
17088 $@{MAIN@}: objects $@{LIB_DIR@}
17089 gnatbind $@{MAIN@} $@{CSC_LIST:%=-aO%/lib@} -shared
17090 gnatlink $@{MAIN@} $@{CSC_LIST:%=-l%@}
17093 # recompile the sources
17094 gnatmake -c -i $@{MAIN@}.adb $@{NEED_FPIC@} $@{CSC_LIST:%=-I%@}
17096 # Note: In a future version of GNAT, the following commands will be simplified
17097 # by a new tool, gnatmlib
17099 mkdir -p $@{dir $@@ @}
17100 cd $@{dir $@@ @} && gcc -shared -o $@{notdir $@@ @} ../*.o -L$@{GLIB@} -lgnat
17101 cd $@{dir $@@ @} && cp -f ../*.ali .
17103 # The dependencies for the modules
17104 # Note that we have to force the expansion of *.o, since in some cases
17105 # make won't be able to do it itself.
17106 aa/lib/libaa.so: $@{wildcard aa/*.o@}
17107 bb/lib/libbb.so: $@{wildcard bb/*.o@}
17108 cc/lib/libcc.so: $@{wildcard cc/*.o@}
17110 # Make sure all of the shared libraries are in the path before starting the
17113 LD_LIBRARY_PATH=`pwd`/aa/lib:`pwd`/bb/lib:`pwd`/cc/lib ./$@{MAIN@}
17116 $@{RM@} -rf $@{CSC_LIST:%=%/lib@}
17117 $@{RM@} $@{CSC_LIST:%=%/*.ali@}
17118 $@{RM@} $@{CSC_LIST:%=%/*.o@}
17119 $@{RM@} *.o *.ali $@{MAIN@}
17122 @node Automatically Creating a List of Directories,Generating the Command Line Switches,Using gnatmake in a Makefile,Using the GNU make Utility
17123 @anchor{gnat_ugn/building_executable_programs_with_gnat id50}@anchor{13e}@anchor{gnat_ugn/building_executable_programs_with_gnat automatically-creating-a-list-of-directories}@anchor{13d}
17124 @subsection Automatically Creating a List of Directories
17127 In most makefiles, you will have to specify a list of directories, and
17128 store it in a variable. For small projects, it is often easier to
17129 specify each of them by hand, since you then have full control over what
17130 is the proper order for these directories, which ones should be
17133 However, in larger projects, which might involve hundreds of
17134 subdirectories, it might be more convenient to generate this list
17137 The example below presents two methods. The first one, although less
17138 general, gives you more control over the list. It involves wildcard
17139 characters, that are automatically expanded by @code{make}. Its
17140 shortcoming is that you need to explicitly specify some of the
17141 organization of your project, such as for instance the directory tree
17142 depth, whether some directories are found in a separate tree, etc.
17144 The second method is the most general one. It requires an external
17145 program, called @code{find}, which is standard on all Unix systems. All
17146 the directories found under a given root directory will be added to the
17150 # The examples below are based on the following directory hierarchy:
17151 # All the directories can contain any number of files
17152 # ROOT_DIRECTORY -> a -> aa -> aaa
17155 # -> b -> ba -> baa
17158 # This Makefile creates a variable called DIRS, that can be reused any time
17159 # you need this list (see the other examples in this section)
17161 # The root of your project's directory hierarchy
17165 # First method: specify explicitly the list of directories
17166 # This allows you to specify any subset of all the directories you need.
17169 DIRS := a/aa/ a/ab/ b/ba/
17172 # Second method: use wildcards
17173 # Note that the argument(s) to wildcard below should end with a '/'.
17174 # Since wildcards also return file names, we have to filter them out
17175 # to avoid duplicate directory names.
17176 # We thus use make's `@w{`}dir`@w{`} and `@w{`}sort`@w{`} functions.
17177 # It sets DIRs to the following value (note that the directories aaa and baa
17178 # are not given, unless you change the arguments to wildcard).
17179 # DIRS= ./a/a/ ./b/ ./a/aa/ ./a/ab/ ./a/ac/ ./b/ba/ ./b/bb/ ./b/bc/
17182 DIRS := $@{sort $@{dir $@{wildcard $@{ROOT_DIRECTORY@}/*/
17183 $@{ROOT_DIRECTORY@}/*/*/@}@}@}
17186 # Third method: use an external program
17187 # This command is much faster if run on local disks, avoiding NFS slowdowns.
17188 # This is the most complete command: it sets DIRs to the following value:
17189 # DIRS= ./a ./a/aa ./a/aa/aaa ./a/ab ./a/ac ./b ./b/ba ./b/ba/baa ./b/bb ./b/bc
17192 DIRS := $@{shell find $@{ROOT_DIRECTORY@} -type d -print@}
17195 @node Generating the Command Line Switches,Overcoming Command Line Length Limits,Automatically Creating a List of Directories,Using the GNU make Utility
17196 @anchor{gnat_ugn/building_executable_programs_with_gnat id51}@anchor{13f}@anchor{gnat_ugn/building_executable_programs_with_gnat generating-the-command-line-switches}@anchor{140}
17197 @subsection Generating the Command Line Switches
17200 Once you have created the list of directories as explained in the
17201 previous section (@ref{13d,,Automatically Creating a List of Directories}),
17202 you can easily generate the command line arguments to pass to gnatmake.
17204 For the sake of completeness, this example assumes that the source path
17205 is not the same as the object path, and that you have two separate lists
17209 # see "Automatically creating a list of directories" to create
17214 GNATMAKE_SWITCHES := $@{patsubst %,-aI%,$@{SOURCE_DIRS@}@}
17215 GNATMAKE_SWITCHES += $@{patsubst %,-aO%,$@{OBJECT_DIRS@}@}
17218 gnatmake $@{GNATMAKE_SWITCHES@} main_unit
17221 @node Overcoming Command Line Length Limits,,Generating the Command Line Switches,Using the GNU make Utility
17222 @anchor{gnat_ugn/building_executable_programs_with_gnat overcoming-command-line-length-limits}@anchor{141}@anchor{gnat_ugn/building_executable_programs_with_gnat id52}@anchor{142}
17223 @subsection Overcoming Command Line Length Limits
17226 One problem that might be encountered on big projects is that many
17227 operating systems limit the length of the command line. It is thus hard to give
17228 gnatmake the list of source and object directories.
17230 This example shows how you can set up environment variables, which will
17231 make @code{gnatmake} behave exactly as if the directories had been
17232 specified on the command line, but have a much higher length limit (or
17233 even none on most systems).
17235 It assumes that you have created a list of directories in your Makefile,
17236 using one of the methods presented in
17237 @ref{13d,,Automatically Creating a List of Directories}.
17238 For the sake of completeness, we assume that the object
17239 path (where the ALI files are found) is different from the sources patch.
17241 Note a small trick in the Makefile below: for efficiency reasons, we
17242 create two temporary variables (SOURCE_LIST and OBJECT_LIST), that are
17243 expanded immediately by @code{make}. This way we overcome the standard
17244 make behavior which is to expand the variables only when they are
17247 On Windows, if you are using the standard Windows command shell, you must
17248 replace colons with semicolons in the assignments to these variables.
17251 # In this example, we create both ADA_INCLUDE_PATH and ADA_OBJECTS_PATH.
17252 # This is the same thing as putting the -I arguments on the command line.
17253 # (the equivalent of using -aI on the command line would be to define
17254 # only ADA_INCLUDE_PATH, the equivalent of -aO is ADA_OBJECTS_PATH).
17255 # You can of course have different values for these variables.
17257 # Note also that we need to keep the previous values of these variables, since
17258 # they might have been set before running 'make' to specify where the GNAT
17259 # library is installed.
17261 # see "Automatically creating a list of directories" to create these
17267 space:=$@{empty@} $@{empty@}
17268 SOURCE_LIST := $@{subst $@{space@},:,$@{SOURCE_DIRS@}@}
17269 OBJECT_LIST := $@{subst $@{space@},:,$@{OBJECT_DIRS@}@}
17270 ADA_INCLUDE_PATH += $@{SOURCE_LIST@}
17271 ADA_OBJECTS_PATH += $@{OBJECT_LIST@}
17272 export ADA_INCLUDE_PATH
17273 export ADA_OBJECTS_PATH
17279 @node GNAT Utility Programs,GNAT and Program Execution,Building Executable Programs with GNAT,Top
17280 @anchor{gnat_ugn/gnat_utility_programs doc}@anchor{143}@anchor{gnat_ugn/gnat_utility_programs gnat-utility-programs}@anchor{b}@anchor{gnat_ugn/gnat_utility_programs id1}@anchor{144}
17281 @chapter GNAT Utility Programs
17284 This chapter describes a number of utility programs:
17291 @ref{20,,The File Cleanup Utility gnatclean}
17294 @ref{21,,The GNAT Library Browser gnatls}
17297 @ref{22,,The Cross-Referencing Tools gnatxref and gnatfind}
17300 @ref{23,,The Ada to HTML Converter gnathtml}
17303 Other GNAT utilities are described elsewhere in this manual:
17309 @ref{59,,Handling Arbitrary File Naming Conventions with gnatname}
17312 @ref{63,,File Name Krunching with gnatkr}
17315 @ref{36,,Renaming Files with gnatchop}
17318 @ref{17,,Preprocessing with gnatprep}
17322 * The File Cleanup Utility gnatclean::
17323 * The GNAT Library Browser gnatls::
17324 * The Cross-Referencing Tools gnatxref and gnatfind::
17325 * The Ada to HTML Converter gnathtml::
17329 @node The File Cleanup Utility gnatclean,The GNAT Library Browser gnatls,,GNAT Utility Programs
17330 @anchor{gnat_ugn/gnat_utility_programs id2}@anchor{145}@anchor{gnat_ugn/gnat_utility_programs the-file-cleanup-utility-gnatclean}@anchor{20}
17331 @section The File Cleanup Utility @code{gnatclean}
17334 @geindex File cleanup tool
17338 @code{gnatclean} is a tool that allows the deletion of files produced by the
17339 compiler, binder and linker, including ALI files, object files, tree files,
17340 expanded source files, library files, interface copy source files, binder
17341 generated files and executable files.
17344 * Running gnatclean::
17345 * Switches for gnatclean::
17349 @node Running gnatclean,Switches for gnatclean,,The File Cleanup Utility gnatclean
17350 @anchor{gnat_ugn/gnat_utility_programs running-gnatclean}@anchor{146}@anchor{gnat_ugn/gnat_utility_programs id3}@anchor{147}
17351 @subsection Running @code{gnatclean}
17354 The @code{gnatclean} command has the form:
17359 $ gnatclean switches names
17363 where @code{names} is a list of source file names. Suffixes @code{.ads} and
17364 @code{adb} may be omitted. If a project file is specified using switch
17365 @code{-P}, then @code{names} may be completely omitted.
17367 In normal mode, @code{gnatclean} delete the files produced by the compiler and,
17368 if switch @code{-c} is not specified, by the binder and
17369 the linker. In informative-only mode, specified by switch
17370 @code{-n}, the list of files that would have been deleted in
17371 normal mode is listed, but no file is actually deleted.
17373 @node Switches for gnatclean,,Running gnatclean,The File Cleanup Utility gnatclean
17374 @anchor{gnat_ugn/gnat_utility_programs id4}@anchor{148}@anchor{gnat_ugn/gnat_utility_programs switches-for-gnatclean}@anchor{149}
17375 @subsection Switches for @code{gnatclean}
17378 @code{gnatclean} recognizes the following switches:
17380 @geindex --version (gnatclean)
17385 @item @code{--version}
17387 Display copyright and version, then exit disregarding all other options.
17390 @geindex --help (gnatclean)
17395 @item @code{--help}
17397 If @code{--version} was not used, display usage, then exit disregarding
17400 @item @code{--subdirs=@emph{subdir}}
17402 Actual object directory of each project file is the subdirectory subdir of the
17403 object directory specified or defaulted in the project file.
17405 @item @code{--unchecked-shared-lib-imports}
17407 By default, shared library projects are not allowed to import static library
17408 projects. When this switch is used on the command line, this restriction is
17412 @geindex -c (gnatclean)
17419 Only attempt to delete the files produced by the compiler, not those produced
17420 by the binder or the linker. The files that are not to be deleted are library
17421 files, interface copy files, binder generated files and executable files.
17424 @geindex -D (gnatclean)
17429 @item @code{-D @emph{dir}}
17431 Indicate that ALI and object files should normally be found in directory @code{dir}.
17434 @geindex -F (gnatclean)
17441 When using project files, if some errors or warnings are detected during
17442 parsing and verbose mode is not in effect (no use of switch
17443 -v), then error lines start with the full path name of the project
17444 file, rather than its simple file name.
17447 @geindex -h (gnatclean)
17454 Output a message explaining the usage of @code{gnatclean}.
17457 @geindex -n (gnatclean)
17464 Informative-only mode. Do not delete any files. Output the list of the files
17465 that would have been deleted if this switch was not specified.
17468 @geindex -P (gnatclean)
17473 @item @code{-P@emph{project}}
17475 Use project file @code{project}. Only one such switch can be used.
17476 When cleaning a project file, the files produced by the compilation of the
17477 immediate sources or inherited sources of the project files are to be
17478 deleted. This is not depending on the presence or not of executable names
17479 on the command line.
17482 @geindex -q (gnatclean)
17489 Quiet output. If there are no errors, do not output anything, except in
17490 verbose mode (switch -v) or in informative-only mode
17494 @geindex -r (gnatclean)
17501 When a project file is specified (using switch -P),
17502 clean all imported and extended project files, recursively. If this switch
17503 is not specified, only the files related to the main project file are to be
17504 deleted. This switch has no effect if no project file is specified.
17507 @geindex -v (gnatclean)
17517 @geindex -vP (gnatclean)
17522 @item @code{-vP@emph{x}}
17524 Indicates the verbosity of the parsing of GNAT project files.
17525 @ref{de,,Switches Related to Project Files}.
17528 @geindex -X (gnatclean)
17533 @item @code{-X@emph{name}=@emph{value}}
17535 Indicates that external variable @code{name} has the value @code{value}.
17536 The Project Manager will use this value for occurrences of
17537 @code{external(name)} when parsing the project file.
17538 See @ref{de,,Switches Related to Project Files}.
17541 @geindex -aO (gnatclean)
17546 @item @code{-aO@emph{dir}}
17548 When searching for ALI and object files, look in directory @code{dir}.
17551 @geindex -I (gnatclean)
17556 @item @code{-I@emph{dir}}
17558 Equivalent to @code{-aO@emph{dir}}.
17561 @geindex -I- (gnatclean)
17563 @geindex Source files
17564 @geindex suppressing search
17571 Do not look for ALI or object files in the directory
17572 where @code{gnatclean} was invoked.
17575 @node The GNAT Library Browser gnatls,The Cross-Referencing Tools gnatxref and gnatfind,The File Cleanup Utility gnatclean,GNAT Utility Programs
17576 @anchor{gnat_ugn/gnat_utility_programs the-gnat-library-browser-gnatls}@anchor{21}@anchor{gnat_ugn/gnat_utility_programs id5}@anchor{14a}
17577 @section The GNAT Library Browser @code{gnatls}
17580 @geindex Library browser
17584 @code{gnatls} is a tool that outputs information about compiled
17585 units. It gives the relationship between objects, unit names and source
17586 files. It can also be used to check the source dependencies of a unit
17587 as well as various characteristics.
17591 * Switches for gnatls::
17592 * Example of gnatls Usage::
17596 @node Running gnatls,Switches for gnatls,,The GNAT Library Browser gnatls
17597 @anchor{gnat_ugn/gnat_utility_programs id6}@anchor{14b}@anchor{gnat_ugn/gnat_utility_programs running-gnatls}@anchor{14c}
17598 @subsection Running @code{gnatls}
17601 The @code{gnatls} command has the form
17606 $ gnatls switches object_or_ali_file
17610 The main argument is the list of object or @code{ali} files
17611 (see @ref{42,,The Ada Library Information Files})
17612 for which information is requested.
17614 In normal mode, without additional option, @code{gnatls} produces a
17615 four-column listing. Each line represents information for a specific
17616 object. The first column gives the full path of the object, the second
17617 column gives the name of the principal unit in this object, the third
17618 column gives the status of the source and the fourth column gives the
17619 full path of the source representing this unit.
17620 Here is a simple example of use:
17626 ./demo1.o demo1 DIF demo1.adb
17627 ./demo2.o demo2 OK demo2.adb
17628 ./hello.o h1 OK hello.adb
17629 ./instr-child.o instr.child MOK instr-child.adb
17630 ./instr.o instr OK instr.adb
17631 ./tef.o tef DIF tef.adb
17632 ./text_io_example.o text_io_example OK text_io_example.adb
17633 ./tgef.o tgef DIF tgef.adb
17637 The first line can be interpreted as follows: the main unit which is
17639 object file @code{demo1.o} is demo1, whose main source is in
17640 @code{demo1.adb}. Furthermore, the version of the source used for the
17641 compilation of demo1 has been modified (DIF). Each source file has a status
17642 qualifier which can be:
17647 @item @emph{OK (unchanged)}
17649 The version of the source file used for the compilation of the
17650 specified unit corresponds exactly to the actual source file.
17652 @item @emph{MOK (slightly modified)}
17654 The version of the source file used for the compilation of the
17655 specified unit differs from the actual source file but not enough to
17656 require recompilation. If you use gnatmake with the option
17657 @code{-m} (minimal recompilation), a file marked
17658 MOK will not be recompiled.
17660 @item @emph{DIF (modified)}
17662 No version of the source found on the path corresponds to the source
17663 used to build this object.
17665 @item @emph{??? (file not found)}
17667 No source file was found for this unit.
17669 @item @emph{HID (hidden, unchanged version not first on PATH)}
17671 The version of the source that corresponds exactly to the source used
17672 for compilation has been found on the path but it is hidden by another
17673 version of the same source that has been modified.
17676 @node Switches for gnatls,Example of gnatls Usage,Running gnatls,The GNAT Library Browser gnatls
17677 @anchor{gnat_ugn/gnat_utility_programs id7}@anchor{14d}@anchor{gnat_ugn/gnat_utility_programs switches-for-gnatls}@anchor{14e}
17678 @subsection Switches for @code{gnatls}
17681 @code{gnatls} recognizes the following switches:
17683 @geindex --version (gnatls)
17688 @item @code{--version}
17690 Display copyright and version, then exit disregarding all other options.
17693 @geindex --help (gnatls)
17698 @item @code{--help}
17700 If @code{--version} was not used, display usage, then exit disregarding
17704 @geindex -a (gnatls)
17711 Consider all units, including those of the predefined Ada library.
17712 Especially useful with @code{-d}.
17715 @geindex -d (gnatls)
17722 List sources from which specified units depend on.
17725 @geindex -h (gnatls)
17732 Output the list of options.
17735 @geindex -o (gnatls)
17742 Only output information about object files.
17745 @geindex -s (gnatls)
17752 Only output information about source files.
17755 @geindex -u (gnatls)
17762 Only output information about compilation units.
17765 @geindex -files (gnatls)
17770 @item @code{-files=@emph{file}}
17772 Take as arguments the files listed in text file @code{file}.
17773 Text file @code{file} may contain empty lines that are ignored.
17774 Each nonempty line should contain the name of an existing file.
17775 Several such switches may be specified simultaneously.
17778 @geindex -aO (gnatls)
17780 @geindex -aI (gnatls)
17782 @geindex -I (gnatls)
17784 @geindex -I- (gnatls)
17789 @item @code{-aO@emph{dir}}, @code{-aI@emph{dir}}, @code{-I@emph{dir}}, @code{-I-}, @code{-nostdinc}
17791 Source path manipulation. Same meaning as the equivalent @code{gnatmake}
17792 flags (@ref{dc,,Switches for gnatmake}).
17795 @geindex -aP (gnatls)
17800 @item @code{-aP@emph{dir}}
17802 Add @code{dir} at the beginning of the project search dir.
17805 @geindex --RTS (gnatls)
17810 @item @code{--RTS=@emph{rts-path}}
17812 Specifies the default location of the runtime library. Same meaning as the
17813 equivalent @code{gnatmake} flag (@ref{dc,,Switches for gnatmake}).
17816 @geindex -v (gnatls)
17823 Verbose mode. Output the complete source, object and project paths. Do not use
17824 the default column layout but instead use long format giving as much as
17825 information possible on each requested units, including special
17826 characteristics such as:
17832 @emph{Preelaborable}: The unit is preelaborable in the Ada sense.
17835 @emph{No_Elab_Code}: No elaboration code has been produced by the compiler for this unit.
17838 @emph{Pure}: The unit is pure in the Ada sense.
17841 @emph{Elaborate_Body}: The unit contains a pragma Elaborate_Body.
17844 @emph{Remote_Types}: The unit contains a pragma Remote_Types.
17847 @emph{Shared_Passive}: The unit contains a pragma Shared_Passive.
17850 @emph{Predefined}: This unit is part of the predefined environment and cannot be modified
17854 @emph{Remote_Call_Interface}: The unit contains a pragma Remote_Call_Interface.
17858 @node Example of gnatls Usage,,Switches for gnatls,The GNAT Library Browser gnatls
17859 @anchor{gnat_ugn/gnat_utility_programs id8}@anchor{14f}@anchor{gnat_ugn/gnat_utility_programs example-of-gnatls-usage}@anchor{150}
17860 @subsection Example of @code{gnatls} Usage
17863 Example of using the verbose switch. Note how the source and
17864 object paths are affected by the -I switch.
17869 $ gnatls -v -I.. demo1.o
17871 GNATLS 5.03w (20041123-34)
17872 Copyright 1997-2004 Free Software Foundation, Inc.
17874 Source Search Path:
17875 <Current_Directory>
17877 /home/comar/local/adainclude/
17879 Object Search Path:
17880 <Current_Directory>
17882 /home/comar/local/lib/gcc-lib/x86-linux/3.4.3/adalib/
17884 Project Search Path:
17885 <Current_Directory>
17886 /home/comar/local/lib/gnat/
17891 Kind => subprogram body
17892 Flags => No_Elab_Code
17893 Source => demo1.adb modified
17897 The following is an example of use of the dependency list.
17898 Note the use of the -s switch
17899 which gives a straight list of source files. This can be useful for
17900 building specialized scripts.
17905 $ gnatls -d demo2.o
17906 ./demo2.o demo2 OK demo2.adb
17912 $ gnatls -d -s -a demo1.o
17914 /home/comar/local/adainclude/ada.ads
17915 /home/comar/local/adainclude/a-finali.ads
17916 /home/comar/local/adainclude/a-filico.ads
17917 /home/comar/local/adainclude/a-stream.ads
17918 /home/comar/local/adainclude/a-tags.ads
17921 /home/comar/local/adainclude/gnat.ads
17922 /home/comar/local/adainclude/g-io.ads
17924 /home/comar/local/adainclude/system.ads
17925 /home/comar/local/adainclude/s-exctab.ads
17926 /home/comar/local/adainclude/s-finimp.ads
17927 /home/comar/local/adainclude/s-finroo.ads
17928 /home/comar/local/adainclude/s-secsta.ads
17929 /home/comar/local/adainclude/s-stalib.ads
17930 /home/comar/local/adainclude/s-stoele.ads
17931 /home/comar/local/adainclude/s-stratt.ads
17932 /home/comar/local/adainclude/s-tasoli.ads
17933 /home/comar/local/adainclude/s-unstyp.ads
17934 /home/comar/local/adainclude/unchconv.ads
17938 @node The Cross-Referencing Tools gnatxref and gnatfind,The Ada to HTML Converter gnathtml,The GNAT Library Browser gnatls,GNAT Utility Programs
17939 @anchor{gnat_ugn/gnat_utility_programs the-cross-referencing-tools-gnatxref-and-gnatfind}@anchor{22}@anchor{gnat_ugn/gnat_utility_programs id9}@anchor{151}
17940 @section The Cross-Referencing Tools @code{gnatxref} and @code{gnatfind}
17947 The compiler generates cross-referencing information (unless
17948 you set the @code{-gnatx} switch), which are saved in the @code{.ali} files.
17949 This information indicates where in the source each entity is declared and
17950 referenced. Note that entities in package Standard are not included, but
17951 entities in all other predefined units are included in the output.
17953 Before using any of these two tools, you need to compile successfully your
17954 application, so that GNAT gets a chance to generate the cross-referencing
17957 The two tools @code{gnatxref} and @code{gnatfind} take advantage of this
17958 information to provide the user with the capability to easily locate the
17959 declaration and references to an entity. These tools are quite similar,
17960 the difference being that @code{gnatfind} is intended for locating
17961 definitions and/or references to a specified entity or entities, whereas
17962 @code{gnatxref} is oriented to generating a full report of all
17965 To use these tools, you must not compile your application using the
17966 @code{-gnatx} switch on the @code{gnatmake} command line
17967 (see @ref{1b,,Building with gnatmake}). Otherwise, cross-referencing
17968 information will not be generated.
17971 * gnatxref Switches::
17972 * gnatfind Switches::
17973 * Configuration Files for gnatxref and gnatfind::
17974 * Regular Expressions in gnatfind and gnatxref::
17975 * Examples of gnatxref Usage::
17976 * Examples of gnatfind Usage::
17980 @node gnatxref Switches,gnatfind Switches,,The Cross-Referencing Tools gnatxref and gnatfind
17981 @anchor{gnat_ugn/gnat_utility_programs id10}@anchor{152}@anchor{gnat_ugn/gnat_utility_programs gnatxref-switches}@anchor{153}
17982 @subsection @code{gnatxref} Switches
17985 The command invocation for @code{gnatxref} is:
17990 $ gnatxref [ switches ] sourcefile1 [ sourcefile2 ... ]
17999 @item @code{sourcefile1} [, @code{sourcefile2} ...]
18001 identify the source files for which a report is to be generated. The
18002 @code{with}ed units will be processed too. You must provide at least one file.
18004 These file names are considered to be regular expressions, so for instance
18005 specifying @code{source*.adb} is the same as giving every file in the current
18006 directory whose name starts with @code{source} and whose extension is
18009 You shouldn't specify any directory name, just base names. @code{gnatxref}
18010 and @code{gnatfind} will be able to locate these files by themselves using
18011 the source path. If you specify directories, no result is produced.
18014 The following switches are available for @code{gnatxref}:
18016 @geindex --version (gnatxref)
18021 @item @code{--version}
18023 Display copyright and version, then exit disregarding all other options.
18026 @geindex --help (gnatxref)
18031 @item @code{--help}
18033 If @code{--version} was not used, display usage, then exit disregarding
18037 @geindex -a (gnatxref)
18044 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
18045 the read-only files found in the library search path. Otherwise, these files
18046 will be ignored. This option can be used to protect Gnat sources or your own
18047 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
18048 much faster, and their output much smaller. Read-only here refers to access
18049 or permissions status in the file system for the current user.
18052 @geindex -aIDIR (gnatxref)
18057 @item @code{-aI@emph{DIR}}
18059 When looking for source files also look in directory DIR. The order in which
18060 source file search is undertaken is the same as for @code{gnatmake}.
18063 @geindex -aODIR (gnatxref)
18068 @item @code{aO@emph{DIR}}
18070 When -searching for library and object files, look in directory
18071 DIR. The order in which library files are searched is the same as for
18075 @geindex -nostdinc (gnatxref)
18080 @item @code{-nostdinc}
18082 Do not look for sources in the system default directory.
18085 @geindex -nostdlib (gnatxref)
18090 @item @code{-nostdlib}
18092 Do not look for library files in the system default directory.
18095 @geindex --ext (gnatxref)
18100 @item @code{--ext=@emph{extension}}
18102 Specify an alternate ali file extension. The default is @code{ali} and other
18103 extensions (e.g. @code{gli} for C/C++ sources) may be specified via this switch.
18104 Note that if this switch overrides the default, only the new extension will
18108 @geindex --RTS (gnatxref)
18113 @item @code{--RTS=@emph{rts-path}}
18115 Specifies the default location of the runtime library. Same meaning as the
18116 equivalent @code{gnatmake} flag (@ref{dc,,Switches for gnatmake}).
18119 @geindex -d (gnatxref)
18126 If this switch is set @code{gnatxref} will output the parent type
18127 reference for each matching derived types.
18130 @geindex -f (gnatxref)
18137 If this switch is set, the output file names will be preceded by their
18138 directory (if the file was found in the search path). If this switch is
18139 not set, the directory will not be printed.
18142 @geindex -g (gnatxref)
18149 If this switch is set, information is output only for library-level
18150 entities, ignoring local entities. The use of this switch may accelerate
18151 @code{gnatfind} and @code{gnatxref}.
18154 @geindex -IDIR (gnatxref)
18159 @item @code{-I@emph{DIR}}
18161 Equivalent to @code{-aODIR -aIDIR}.
18164 @geindex -pFILE (gnatxref)
18169 @item @code{-p@emph{FILE}}
18171 Specify a configuration file to use to list the source and object directories.
18173 If a file is specified, then the content of the source directory and object
18174 directory lines are added as if they had been specified respectively
18175 by @code{-aI} and @code{-aO}.
18177 See @ref{154,,Configuration Files for gnatxref and gnatfind} for the syntax
18178 of this configuration file.
18182 Output only unused symbols. This may be really useful if you give your
18183 main compilation unit on the command line, as @code{gnatxref} will then
18184 display every unused entity and 'with'ed package.
18188 Instead of producing the default output, @code{gnatxref} will generate a
18189 @code{tags} file that can be used by vi. For examples how to use this
18190 feature, see @ref{155,,Examples of gnatxref Usage}. The tags file is output
18191 to the standard output, thus you will have to redirect it to a file.
18194 All these switches may be in any order on the command line, and may even
18195 appear after the file names. They need not be separated by spaces, thus
18196 you can say @code{gnatxref -ag} instead of @code{gnatxref -a -g}.
18198 @node gnatfind Switches,Configuration Files for gnatxref and gnatfind,gnatxref Switches,The Cross-Referencing Tools gnatxref and gnatfind
18199 @anchor{gnat_ugn/gnat_utility_programs id11}@anchor{156}@anchor{gnat_ugn/gnat_utility_programs gnatfind-switches}@anchor{157}
18200 @subsection @code{gnatfind} Switches
18203 The command invocation for @code{gnatfind} is:
18208 $ gnatfind [ switches ] pattern[:sourcefile[:line[:column]]]
18213 with the following iterpretation of the command arguments:
18218 @item @emph{pattern}
18220 An entity will be output only if it matches the regular expression found
18221 in @emph{pattern}, see @ref{158,,Regular Expressions in gnatfind and gnatxref}.
18223 Omitting the pattern is equivalent to specifying @code{*}, which
18224 will match any entity. Note that if you do not provide a pattern, you
18225 have to provide both a sourcefile and a line.
18227 Entity names are given in Latin-1, with uppercase/lowercase equivalence
18228 for matching purposes. At the current time there is no support for
18229 8-bit codes other than Latin-1, or for wide characters in identifiers.
18231 @item @emph{sourcefile}
18233 @code{gnatfind} will look for references, bodies or declarations
18234 of symbols referenced in @code{sourcefile}, at line @code{line}
18235 and column @code{column}. See @ref{159,,Examples of gnatfind Usage}
18236 for syntax examples.
18240 A decimal integer identifying the line number containing
18241 the reference to the entity (or entities) to be located.
18243 @item @emph{column}
18245 A decimal integer identifying the exact location on the
18246 line of the first character of the identifier for the
18247 entity reference. Columns are numbered from 1.
18249 @item @emph{file1 file2 ...}
18251 The search will be restricted to these source files. If none are given, then
18252 the search will be conducted for every library file in the search path.
18253 These files must appear only after the pattern or sourcefile.
18255 These file names are considered to be regular expressions, so for instance
18256 specifying @code{source*.adb} is the same as giving every file in the current
18257 directory whose name starts with @code{source} and whose extension is
18260 The location of the spec of the entity will always be displayed, even if it
18261 isn't in one of @code{file1}, @code{file2}, ... The
18262 occurrences of the entity in the separate units of the ones given on the
18263 command line will also be displayed.
18265 Note that if you specify at least one file in this part, @code{gnatfind} may
18266 sometimes not be able to find the body of the subprograms.
18269 At least one of 'sourcefile' or 'pattern' has to be present on
18272 The following switches are available:
18274 @geindex --version (gnatfind)
18279 @item @code{--version}
18281 Display copyright and version, then exit disregarding all other options.
18284 @geindex --help (gnatfind)
18289 @item @code{--help}
18291 If @code{--version} was not used, display usage, then exit disregarding
18295 @geindex -a (gnatfind)
18302 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
18303 the read-only files found in the library search path. Otherwise, these files
18304 will be ignored. This option can be used to protect Gnat sources or your own
18305 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
18306 much faster, and their output much smaller. Read-only here refers to access
18307 or permission status in the file system for the current user.
18310 @geindex -aIDIR (gnatfind)
18315 @item @code{-aI@emph{DIR}}
18317 When looking for source files also look in directory DIR. The order in which
18318 source file search is undertaken is the same as for @code{gnatmake}.
18321 @geindex -aODIR (gnatfind)
18326 @item @code{-aO@emph{DIR}}
18328 When searching for library and object files, look in directory
18329 DIR. The order in which library files are searched is the same as for
18333 @geindex -nostdinc (gnatfind)
18338 @item @code{-nostdinc}
18340 Do not look for sources in the system default directory.
18343 @geindex -nostdlib (gnatfind)
18348 @item @code{-nostdlib}
18350 Do not look for library files in the system default directory.
18353 @geindex --ext (gnatfind)
18358 @item @code{--ext=@emph{extension}}
18360 Specify an alternate ali file extension. The default is @code{ali} and other
18361 extensions may be specified via this switch. Note that if this switch
18362 overrides the default, only the new extension will be considered.
18365 @geindex --RTS (gnatfind)
18370 @item @code{--RTS=@emph{rts-path}}
18372 Specifies the default location of the runtime library. Same meaning as the
18373 equivalent @code{gnatmake} flag (@ref{dc,,Switches for gnatmake}).
18376 @geindex -d (gnatfind)
18383 If this switch is set, then @code{gnatfind} will output the parent type
18384 reference for each matching derived types.
18387 @geindex -e (gnatfind)
18394 By default, @code{gnatfind} accept the simple regular expression set for
18395 @code{pattern}. If this switch is set, then the pattern will be
18396 considered as full Unix-style regular expression.
18399 @geindex -f (gnatfind)
18406 If this switch is set, the output file names will be preceded by their
18407 directory (if the file was found in the search path). If this switch is
18408 not set, the directory will not be printed.
18411 @geindex -g (gnatfind)
18418 If this switch is set, information is output only for library-level
18419 entities, ignoring local entities. The use of this switch may accelerate
18420 @code{gnatfind} and @code{gnatxref}.
18423 @geindex -IDIR (gnatfind)
18428 @item @code{-I@emph{DIR}}
18430 Equivalent to @code{-aODIR -aIDIR}.
18433 @geindex -pFILE (gnatfind)
18438 @item @code{-p@emph{FILE}}
18440 Specify a configuration file to use to list the source and object directories.
18442 If a file is specified, then the content of the source directory and object
18443 directory lines are added as if they had been specified respectively
18444 by @code{-aI} and @code{-aO}.
18446 See @ref{154,,Configuration Files for gnatxref and gnatfind} for the syntax
18447 of this configuration file.
18450 @geindex -r (gnatfind)
18457 By default, @code{gnatfind} will output only the information about the
18458 declaration, body or type completion of the entities. If this switch is
18459 set, the @code{gnatfind} will locate every reference to the entities in
18460 the files specified on the command line (or in every file in the search
18461 path if no file is given on the command line).
18464 @geindex -s (gnatfind)
18471 If this switch is set, then @code{gnatfind} will output the content
18472 of the Ada source file lines were the entity was found.
18475 @geindex -t (gnatfind)
18482 If this switch is set, then @code{gnatfind} will output the type hierarchy for
18483 the specified type. It act like -d option but recursively from parent
18484 type to parent type. When this switch is set it is not possible to
18485 specify more than one file.
18488 All these switches may be in any order on the command line, and may even
18489 appear after the file names. They need not be separated by spaces, thus
18490 you can say @code{gnatxref -ag} instead of
18491 @code{gnatxref -a -g}.
18493 As stated previously, @code{gnatfind} will search in every directory in the
18494 search path. You can force it to look only in the current directory if
18495 you specify @code{*} at the end of the command line.
18497 @node Configuration Files for gnatxref and gnatfind,Regular Expressions in gnatfind and gnatxref,gnatfind Switches,The Cross-Referencing Tools gnatxref and gnatfind
18498 @anchor{gnat_ugn/gnat_utility_programs configuration-files-for-gnatxref-and-gnatfind}@anchor{154}@anchor{gnat_ugn/gnat_utility_programs id12}@anchor{15a}
18499 @subsection Configuration Files for @code{gnatxref} and @code{gnatfind}
18502 Configuration files are used by @code{gnatxref} and @code{gnatfind} to specify
18503 the list of source and object directories to consider. They can be
18504 specified via the @code{-p} switch.
18506 The following lines can be included, in any order in the file:
18515 @item @emph{src_dir=DIR}
18517 [default: @code{"./"}].
18518 Specifies a directory where to look for source files. Multiple @code{src_dir}
18519 lines can be specified and they will be searched in the order they
18527 @item @emph{obj_dir=DIR}
18529 [default: @code{"./"}].
18530 Specifies a directory where to look for object and library files. Multiple
18531 @code{obj_dir} lines can be specified, and they will be searched in the order
18536 Any other line will be silently ignored.
18538 @node Regular Expressions in gnatfind and gnatxref,Examples of gnatxref Usage,Configuration Files for gnatxref and gnatfind,The Cross-Referencing Tools gnatxref and gnatfind
18539 @anchor{gnat_ugn/gnat_utility_programs id13}@anchor{15b}@anchor{gnat_ugn/gnat_utility_programs regular-expressions-in-gnatfind-and-gnatxref}@anchor{158}
18540 @subsection Regular Expressions in @code{gnatfind} and @code{gnatxref}
18543 As specified in the section about @code{gnatfind}, the pattern can be a
18544 regular expression. Two kinds of regular expressions
18554 @item @emph{Globbing pattern}
18556 These are the most common regular expression. They are the same as are
18557 generally used in a Unix shell command line, or in a DOS session.
18559 Here is a more formal grammar:
18563 term ::= elmt -- matches elmt
18564 term ::= elmt elmt -- concatenation (elmt then elmt)
18565 term ::= * -- any string of 0 or more characters
18566 term ::= ? -- matches any character
18567 term ::= [char @{char@}] -- matches any character listed
18568 term ::= [char - char] -- matches any character in range
18576 @item @emph{Full regular expression}
18578 The second set of regular expressions is much more powerful. This is the
18579 type of regular expressions recognized by utilities such as @code{grep}.
18581 The following is the form of a regular expression, expressed in same BNF
18582 style as is found in the Ada Reference Manual:
18585 regexp ::= term @{| term@} -- alternation (term or term ...)
18587 term ::= item @{item@} -- concatenation (item then item)
18589 item ::= elmt -- match elmt
18590 item ::= elmt * -- zero or more elmt's
18591 item ::= elmt + -- one or more elmt's
18592 item ::= elmt ? -- matches elmt or nothing
18594 elmt ::= nschar -- matches given character
18595 elmt ::= [nschar @{nschar@}] -- matches any character listed
18596 elmt ::= [^ nschar @{nschar@}] -- matches any character not listed
18597 elmt ::= [char - char] -- matches chars in given range
18598 elmt ::= \\ char -- matches given character
18599 elmt ::= . -- matches any single character
18600 elmt ::= ( regexp ) -- parens used for grouping
18602 char ::= any character, including special characters
18603 nschar ::= any character except ()[].*+?^
18606 Here are a few examples:
18613 @item @code{abcde|fghi}
18615 will match any of the two strings @code{abcde} and @code{fghi},
18619 will match any string like @code{abd}, @code{abcd}, @code{abccd},
18620 @code{abcccd}, and so on,
18622 @item @code{[a-z]+}
18624 will match any string which has only lowercase characters in it (and at
18625 least one character.
18631 @node Examples of gnatxref Usage,Examples of gnatfind Usage,Regular Expressions in gnatfind and gnatxref,The Cross-Referencing Tools gnatxref and gnatfind
18632 @anchor{gnat_ugn/gnat_utility_programs examples-of-gnatxref-usage}@anchor{155}@anchor{gnat_ugn/gnat_utility_programs id14}@anchor{15c}
18633 @subsection Examples of @code{gnatxref} Usage
18638 * Using gnatxref with vi::
18642 @node General Usage,Using gnatxref with vi,,Examples of gnatxref Usage
18643 @anchor{gnat_ugn/gnat_utility_programs general-usage}@anchor{15d}
18644 @subsubsection General Usage
18647 For the following examples, we will consider the following units:
18655 3: procedure Foo (B : in Integer);
18662 1: package body Main is
18663 2: procedure Foo (B : in Integer) is
18674 2: procedure Print (B : Integer);
18679 The first thing to do is to recompile your application (for instance, in
18680 that case just by doing a @code{gnatmake main}, so that GNAT generates
18681 the cross-referencing information.
18682 You can then issue any of the following commands:
18690 @code{gnatxref main.adb}
18691 @code{gnatxref} generates cross-reference information for main.adb
18692 and every unit 'with'ed by main.adb.
18694 The output would be:
18702 Decl: main.ads 3:20
18703 Body: main.adb 2:20
18704 Ref: main.adb 4:13 5:13 6:19
18707 Ref: main.adb 6:8 7:8
18717 Decl: main.ads 3:15
18718 Body: main.adb 2:15
18721 Body: main.adb 1:14
18724 Ref: main.adb 6:12 7:12
18728 This shows that the entity @code{Main} is declared in main.ads, line 2, column 9,
18729 its body is in main.adb, line 1, column 14 and is not referenced any where.
18731 The entity @code{Print} is declared in @code{bar.ads}, line 2, column 15 and it
18732 is referenced in @code{main.adb}, line 6 column 12 and line 7 column 12.
18735 @code{gnatxref package1.adb package2.ads}
18736 @code{gnatxref} will generates cross-reference information for
18737 @code{package1.adb}, @code{package2.ads} and any other package @code{with}ed by any
18742 @node Using gnatxref with vi,,General Usage,Examples of gnatxref Usage
18743 @anchor{gnat_ugn/gnat_utility_programs using-gnatxref-with-vi}@anchor{15e}
18744 @subsubsection Using @code{gnatxref} with @code{vi}
18747 @code{gnatxref} can generate a tags file output, which can be used
18748 directly from @code{vi}. Note that the standard version of @code{vi}
18749 will not work properly with overloaded symbols. Consider using another
18750 free implementation of @code{vi}, such as @code{vim}.
18755 $ gnatxref -v gnatfind.adb > tags
18759 The following command will generate the tags file for @code{gnatfind} itself
18760 (if the sources are in the search path!):
18765 $ gnatxref -v gnatfind.adb > tags
18769 From @code{vi}, you can then use the command @code{:tag @emph{entity}}
18770 (replacing @code{entity} by whatever you are looking for), and vi will
18771 display a new file with the corresponding declaration of entity.
18773 @node Examples of gnatfind Usage,,Examples of gnatxref Usage,The Cross-Referencing Tools gnatxref and gnatfind
18774 @anchor{gnat_ugn/gnat_utility_programs id15}@anchor{15f}@anchor{gnat_ugn/gnat_utility_programs examples-of-gnatfind-usage}@anchor{159}
18775 @subsection Examples of @code{gnatfind} Usage
18782 @code{gnatfind -f xyz:main.adb}
18783 Find declarations for all entities xyz referenced at least once in
18784 main.adb. The references are search in every library file in the search
18787 The directories will be printed as well (as the @code{-f}
18790 The output will look like:
18795 directory/main.ads:106:14: xyz <= declaration
18796 directory/main.adb:24:10: xyz <= body
18797 directory/foo.ads:45:23: xyz <= declaration
18801 I.e., one of the entities xyz found in main.adb is declared at
18802 line 12 of main.ads (and its body is in main.adb), and another one is
18803 declared at line 45 of foo.ads
18806 @code{gnatfind -fs xyz:main.adb}
18807 This is the same command as the previous one, but @code{gnatfind} will
18808 display the content of the Ada source file lines.
18810 The output will look like:
18813 directory/main.ads:106:14: xyz <= declaration
18815 directory/main.adb:24:10: xyz <= body
18817 directory/foo.ads:45:23: xyz <= declaration
18821 This can make it easier to find exactly the location your are looking
18825 @code{gnatfind -r "*x*":main.ads:123 foo.adb}
18826 Find references to all entities containing an x that are
18827 referenced on line 123 of main.ads.
18828 The references will be searched only in main.ads and foo.adb.
18831 @code{gnatfind main.ads:123}
18832 Find declarations and bodies for all entities that are referenced on
18833 line 123 of main.ads.
18835 This is the same as @code{gnatfind "*":main.adb:123`}
18838 @code{gnatfind mydir/main.adb:123:45}
18839 Find the declaration for the entity referenced at column 45 in
18840 line 123 of file main.adb in directory mydir. Note that it
18841 is usual to omit the identifier name when the column is given,
18842 since the column position identifies a unique reference.
18844 The column has to be the beginning of the identifier, and should not
18845 point to any character in the middle of the identifier.
18848 @node The Ada to HTML Converter gnathtml,,The Cross-Referencing Tools gnatxref and gnatfind,GNAT Utility Programs
18849 @anchor{gnat_ugn/gnat_utility_programs the-ada-to-html-converter-gnathtml}@anchor{23}@anchor{gnat_ugn/gnat_utility_programs id16}@anchor{160}
18850 @section The Ada to HTML Converter @code{gnathtml}
18855 @code{gnathtml} is a Perl script that allows Ada source files to be browsed using
18856 standard Web browsers. For installation information, see @ref{161,,Installing gnathtml}.
18858 Ada reserved keywords are highlighted in a bold font and Ada comments in
18859 a blue font. Unless your program was compiled with the gcc @code{-gnatx}
18860 switch to suppress the generation of cross-referencing information, user
18861 defined variables and types will appear in a different color; you will
18862 be able to click on any identifier and go to its declaration.
18865 * Invoking gnathtml::
18866 * Installing gnathtml::
18870 @node Invoking gnathtml,Installing gnathtml,,The Ada to HTML Converter gnathtml
18871 @anchor{gnat_ugn/gnat_utility_programs invoking-gnathtml}@anchor{162}@anchor{gnat_ugn/gnat_utility_programs id17}@anchor{163}
18872 @subsection Invoking @code{gnathtml}
18875 The command line is as follows:
18880 $ perl gnathtml.pl [ switches ] ada-files
18884 You can specify as many Ada files as you want. @code{gnathtml} will generate
18885 an html file for every ada file, and a global file called @code{index.htm}.
18886 This file is an index of every identifier defined in the files.
18888 The following switches are available:
18890 @geindex -83 (gnathtml)
18897 Only the Ada 83 subset of keywords will be highlighted.
18900 @geindex -cc (gnathtml)
18905 @item @code{cc @emph{color}}
18907 This option allows you to change the color used for comments. The default
18908 value is green. The color argument can be any name accepted by html.
18911 @geindex -d (gnathtml)
18918 If the Ada files depend on some other files (for instance through
18919 @code{with} clauses, the latter files will also be converted to html.
18920 Only the files in the user project will be converted to html, not the files
18921 in the run-time library itself.
18924 @geindex -D (gnathtml)
18931 This command is the same as @code{-d} above, but @code{gnathtml} will
18932 also look for files in the run-time library, and generate html files for them.
18935 @geindex -ext (gnathtml)
18940 @item @code{ext @emph{extension}}
18942 This option allows you to change the extension of the generated HTML files.
18943 If you do not specify an extension, it will default to @code{htm}.
18946 @geindex -f (gnathtml)
18953 By default, gnathtml will generate html links only for global entities
18954 ('with'ed units, global variables and types,...). If you specify
18955 @code{-f} on the command line, then links will be generated for local
18959 @geindex -l (gnathtml)
18964 @item @code{l @emph{number}}
18966 If this switch is provided and @code{number} is not 0, then
18967 @code{gnathtml} will number the html files every @code{number} line.
18970 @geindex -I (gnathtml)
18975 @item @code{I @emph{dir}}
18977 Specify a directory to search for library files (@code{.ALI} files) and
18978 source files. You can provide several -I switches on the command line,
18979 and the directories will be parsed in the order of the command line.
18982 @geindex -o (gnathtml)
18987 @item @code{o @emph{dir}}
18989 Specify the output directory for html files. By default, gnathtml will
18990 saved the generated html files in a subdirectory named @code{html/}.
18993 @geindex -p (gnathtml)
18998 @item @code{p @emph{file}}
19000 If you are using Emacs and the most recent Emacs Ada mode, which provides
19001 a full Integrated Development Environment for compiling, checking,
19002 running and debugging applications, you may use @code{.gpr} files
19003 to give the directories where Emacs can find sources and object files.
19005 Using this switch, you can tell gnathtml to use these files.
19006 This allows you to get an html version of your application, even if it
19007 is spread over multiple directories.
19010 @geindex -sc (gnathtml)
19015 @item @code{sc @emph{color}}
19017 This switch allows you to change the color used for symbol
19019 The default value is red. The color argument can be any name accepted by html.
19022 @geindex -t (gnathtml)
19027 @item @code{t @emph{file}}
19029 This switch provides the name of a file. This file contains a list of
19030 file names to be converted, and the effect is exactly as though they had
19031 appeared explicitly on the command line. This
19032 is the recommended way to work around the command line length limit on some
19036 @node Installing gnathtml,,Invoking gnathtml,The Ada to HTML Converter gnathtml
19037 @anchor{gnat_ugn/gnat_utility_programs installing-gnathtml}@anchor{161}@anchor{gnat_ugn/gnat_utility_programs id18}@anchor{164}
19038 @subsection Installing @code{gnathtml}
19041 @code{Perl} needs to be installed on your machine to run this script.
19042 @code{Perl} is freely available for almost every architecture and
19043 operating system via the Internet.
19045 On Unix systems, you may want to modify the first line of the script
19046 @code{gnathtml}, to explicitly specify where Perl
19047 is located. The syntax of this line is:
19052 #!full_path_name_to_perl
19056 Alternatively, you may run the script using the following command line:
19061 $ perl gnathtml.pl [ switches ] files
19065 @c -- +---------------------------------------------------------------------+
19067 @c -- | The following sections are present only in the PRO and GPL editions |
19069 @c -- +---------------------------------------------------------------------+
19079 @c -- Example: A |withing| unit has a |with| clause, it |withs| a |withed| unit
19081 @node GNAT and Program Execution,Platform-Specific Information,GNAT Utility Programs,Top
19082 @anchor{gnat_ugn/gnat_and_program_execution gnat-and-program-execution}@anchor{c}@anchor{gnat_ugn/gnat_and_program_execution doc}@anchor{165}@anchor{gnat_ugn/gnat_and_program_execution id1}@anchor{166}
19083 @chapter GNAT and Program Execution
19086 This chapter covers several topics:
19092 @ref{167,,Running and Debugging Ada Programs}
19095 @ref{25,,Profiling}
19098 @ref{168,,Improving Performance}
19101 @ref{169,,Overflow Check Handling in GNAT}
19104 @ref{16a,,Performing Dimensionality Analysis in GNAT}
19107 @ref{16b,,Stack Related Facilities}
19110 @ref{16c,,Memory Management Issues}
19114 * Running and Debugging Ada Programs::
19116 * Improving Performance::
19117 * Overflow Check Handling in GNAT::
19118 * Performing Dimensionality Analysis in GNAT::
19119 * Stack Related Facilities::
19120 * Memory Management Issues::
19124 @node Running and Debugging Ada Programs,Profiling,,GNAT and Program Execution
19125 @anchor{gnat_ugn/gnat_and_program_execution id2}@anchor{167}@anchor{gnat_ugn/gnat_and_program_execution running-and-debugging-ada-programs}@anchor{24}
19126 @section Running and Debugging Ada Programs
19131 This section discusses how to debug Ada programs.
19133 An incorrect Ada program may be handled in three ways by the GNAT compiler:
19139 The illegality may be a violation of the static semantics of Ada. In
19140 that case GNAT diagnoses the constructs in the program that are illegal.
19141 It is then a straightforward matter for the user to modify those parts of
19145 The illegality may be a violation of the dynamic semantics of Ada. In
19146 that case the program compiles and executes, but may generate incorrect
19147 results, or may terminate abnormally with some exception.
19150 When presented with a program that contains convoluted errors, GNAT
19151 itself may terminate abnormally without providing full diagnostics on
19152 the incorrect user program.
19160 * The GNAT Debugger GDB::
19162 * Introduction to GDB Commands::
19163 * Using Ada Expressions::
19164 * Calling User-Defined Subprograms::
19165 * Using the next Command in a Function::
19166 * Stopping When Ada Exceptions Are Raised::
19168 * Debugging Generic Units::
19169 * Remote Debugging with gdbserver::
19170 * GNAT Abnormal Termination or Failure to Terminate::
19171 * Naming Conventions for GNAT Source Files::
19172 * Getting Internal Debugging Information::
19173 * Stack Traceback::
19174 * Pretty-Printers for the GNAT runtime::
19178 @node The GNAT Debugger GDB,Running GDB,,Running and Debugging Ada Programs
19179 @anchor{gnat_ugn/gnat_and_program_execution the-gnat-debugger-gdb}@anchor{16d}@anchor{gnat_ugn/gnat_and_program_execution id3}@anchor{16e}
19180 @subsection The GNAT Debugger GDB
19183 @code{GDB} is a general purpose, platform-independent debugger that
19184 can be used to debug mixed-language programs compiled with @code{gcc},
19185 and in particular is capable of debugging Ada programs compiled with
19186 GNAT. The latest versions of @code{GDB} are Ada-aware and can handle
19187 complex Ada data structures.
19189 See @cite{Debugging with GDB},
19190 for full details on the usage of @code{GDB}, including a section on
19191 its usage on programs. This manual should be consulted for full
19192 details. The section that follows is a brief introduction to the
19193 philosophy and use of @code{GDB}.
19195 When GNAT programs are compiled, the compiler optionally writes debugging
19196 information into the generated object file, including information on
19197 line numbers, and on declared types and variables. This information is
19198 separate from the generated code. It makes the object files considerably
19199 larger, but it does not add to the size of the actual executable that
19200 will be loaded into memory, and has no impact on run-time performance. The
19201 generation of debug information is triggered by the use of the
19202 @code{-g} switch in the @code{gcc} or @code{gnatmake} command
19203 used to carry out the compilations. It is important to emphasize that
19204 the use of these options does not change the generated code.
19206 The debugging information is written in standard system formats that
19207 are used by many tools, including debuggers and profilers. The format
19208 of the information is typically designed to describe C types and
19209 semantics, but GNAT implements a translation scheme which allows full
19210 details about Ada types and variables to be encoded into these
19211 standard C formats. Details of this encoding scheme may be found in
19212 the file exp_dbug.ads in the GNAT source distribution. However, the
19213 details of this encoding are, in general, of no interest to a user,
19214 since @code{GDB} automatically performs the necessary decoding.
19216 When a program is bound and linked, the debugging information is
19217 collected from the object files, and stored in the executable image of
19218 the program. Again, this process significantly increases the size of
19219 the generated executable file, but it does not increase the size of
19220 the executable program itself. Furthermore, if this program is run in
19221 the normal manner, it runs exactly as if the debug information were
19222 not present, and takes no more actual memory.
19224 However, if the program is run under control of @code{GDB}, the
19225 debugger is activated. The image of the program is loaded, at which
19226 point it is ready to run. If a run command is given, then the program
19227 will run exactly as it would have if @code{GDB} were not present. This
19228 is a crucial part of the @code{GDB} design philosophy. @code{GDB} is
19229 entirely non-intrusive until a breakpoint is encountered. If no
19230 breakpoint is ever hit, the program will run exactly as it would if no
19231 debugger were present. When a breakpoint is hit, @code{GDB} accesses
19232 the debugging information and can respond to user commands to inspect
19233 variables, and more generally to report on the state of execution.
19235 @node Running GDB,Introduction to GDB Commands,The GNAT Debugger GDB,Running and Debugging Ada Programs
19236 @anchor{gnat_ugn/gnat_and_program_execution id4}@anchor{16f}@anchor{gnat_ugn/gnat_and_program_execution running-gdb}@anchor{170}
19237 @subsection Running GDB
19240 This section describes how to initiate the debugger.
19242 The debugger can be launched from a @code{GPS} menu or
19243 directly from the command line. The description below covers the latter use.
19244 All the commands shown can be used in the @code{GPS} debug console window,
19245 but there are usually more GUI-based ways to achieve the same effect.
19247 The command to run @code{GDB} is
19256 where @code{program} is the name of the executable file. This
19257 activates the debugger and results in a prompt for debugger commands.
19258 The simplest command is simply @code{run}, which causes the program to run
19259 exactly as if the debugger were not present. The following section
19260 describes some of the additional commands that can be given to @code{GDB}.
19262 @node Introduction to GDB Commands,Using Ada Expressions,Running GDB,Running and Debugging Ada Programs
19263 @anchor{gnat_ugn/gnat_and_program_execution introduction-to-gdb-commands}@anchor{171}@anchor{gnat_ugn/gnat_and_program_execution id5}@anchor{172}
19264 @subsection Introduction to GDB Commands
19267 @code{GDB} contains a large repertoire of commands.
19268 See @cite{Debugging with GDB} for extensive documentation on the use
19269 of these commands, together with examples of their use. Furthermore,
19270 the command @emph{help} invoked from within GDB activates a simple help
19271 facility which summarizes the available commands and their options.
19272 In this section we summarize a few of the most commonly
19273 used commands to give an idea of what @code{GDB} is about. You should create
19274 a simple program with debugging information and experiment with the use of
19275 these @code{GDB} commands on the program as you read through the
19285 @item @code{set args @emph{arguments}}
19287 The @emph{arguments} list above is a list of arguments to be passed to
19288 the program on a subsequent run command, just as though the arguments
19289 had been entered on a normal invocation of the program. The @code{set args}
19290 command is not needed if the program does not require arguments.
19299 The @code{run} command causes execution of the program to start from
19300 the beginning. If the program is already running, that is to say if
19301 you are currently positioned at a breakpoint, then a prompt will ask
19302 for confirmation that you want to abandon the current execution and
19310 @item @code{breakpoint @emph{location}}
19312 The breakpoint command sets a breakpoint, that is to say a point at which
19313 execution will halt and @code{GDB} will await further
19314 commands. @emph{location} is
19315 either a line number within a file, given in the format @code{file:linenumber},
19316 or it is the name of a subprogram. If you request that a breakpoint be set on
19317 a subprogram that is overloaded, a prompt will ask you to specify on which of
19318 those subprograms you want to breakpoint. You can also
19319 specify that all of them should be breakpointed. If the program is run
19320 and execution encounters the breakpoint, then the program
19321 stops and @code{GDB} signals that the breakpoint was encountered by
19322 printing the line of code before which the program is halted.
19329 @item @code{catch exception @emph{name}}
19331 This command causes the program execution to stop whenever exception
19332 @code{name} is raised. If @code{name} is omitted, then the execution is
19333 suspended when any exception is raised.
19340 @item @code{print @emph{expression}}
19342 This will print the value of the given expression. Most simple
19343 Ada expression formats are properly handled by @code{GDB}, so the expression
19344 can contain function calls, variables, operators, and attribute references.
19351 @item @code{continue}
19353 Continues execution following a breakpoint, until the next breakpoint or the
19354 termination of the program.
19363 Executes a single line after a breakpoint. If the next statement
19364 is a subprogram call, execution continues into (the first statement of)
19365 the called subprogram.
19374 Executes a single line. If this line is a subprogram call, executes and
19375 returns from the call.
19384 Lists a few lines around the current source location. In practice, it
19385 is usually more convenient to have a separate edit window open with the
19386 relevant source file displayed. Successive applications of this command
19387 print subsequent lines. The command can be given an argument which is a
19388 line number, in which case it displays a few lines around the specified one.
19395 @item @code{backtrace}
19397 Displays a backtrace of the call chain. This command is typically
19398 used after a breakpoint has occurred, to examine the sequence of calls that
19399 leads to the current breakpoint. The display includes one line for each
19400 activation record (frame) corresponding to an active subprogram.
19409 At a breakpoint, @code{GDB} can display the values of variables local
19410 to the current frame. The command @code{up} can be used to
19411 examine the contents of other active frames, by moving the focus up
19412 the stack, that is to say from callee to caller, one frame at a time.
19421 Moves the focus of @code{GDB} down from the frame currently being
19422 examined to the frame of its callee (the reverse of the previous command),
19429 @item @code{frame @emph{n}}
19431 Inspect the frame with the given number. The value 0 denotes the frame
19432 of the current breakpoint, that is to say the top of the call stack.
19441 Kills the child process in which the program is running under GDB.
19442 This may be useful for several purposes:
19448 It allows you to recompile and relink your program, since on many systems
19449 you cannot regenerate an executable file while it is running in a process.
19452 You can run your program outside the debugger, on systems that do not
19453 permit executing a program outside GDB while breakpoints are set
19457 It allows you to debug a core dump rather than a running process.
19462 The above list is a very short introduction to the commands that
19463 @code{GDB} provides. Important additional capabilities, including conditional
19464 breakpoints, the ability to execute command sequences on a breakpoint,
19465 the ability to debug at the machine instruction level and many other
19466 features are described in detail in @cite{Debugging with GDB}.
19467 Note that most commands can be abbreviated
19468 (for example, c for continue, bt for backtrace).
19470 @node Using Ada Expressions,Calling User-Defined Subprograms,Introduction to GDB Commands,Running and Debugging Ada Programs
19471 @anchor{gnat_ugn/gnat_and_program_execution id6}@anchor{173}@anchor{gnat_ugn/gnat_and_program_execution using-ada-expressions}@anchor{174}
19472 @subsection Using Ada Expressions
19475 @geindex Ada expressions (in gdb)
19477 @code{GDB} supports a fairly large subset of Ada expression syntax, with some
19478 extensions. The philosophy behind the design of this subset is
19486 That @code{GDB} should provide basic literals and access to operations for
19487 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
19488 leaving more sophisticated computations to subprograms written into the
19489 program (which therefore may be called from @code{GDB}).
19492 That type safety and strict adherence to Ada language restrictions
19493 are not particularly relevant in a debugging context.
19496 That brevity is important to the @code{GDB} user.
19500 Thus, for brevity, the debugger acts as if there were
19501 implicit @code{with} and @code{use} clauses in effect for all user-written
19502 packages, thus making it unnecessary to fully qualify most names with
19503 their packages, regardless of context. Where this causes ambiguity,
19504 @code{GDB} asks the user's intent.
19506 For details on the supported Ada syntax, see @cite{Debugging with GDB}.
19508 @node Calling User-Defined Subprograms,Using the next Command in a Function,Using Ada Expressions,Running and Debugging Ada Programs
19509 @anchor{gnat_ugn/gnat_and_program_execution id7}@anchor{175}@anchor{gnat_ugn/gnat_and_program_execution calling-user-defined-subprograms}@anchor{176}
19510 @subsection Calling User-Defined Subprograms
19513 An important capability of @code{GDB} is the ability to call user-defined
19514 subprograms while debugging. This is achieved simply by entering
19515 a subprogram call statement in the form:
19520 call subprogram-name (parameters)
19524 The keyword @code{call} can be omitted in the normal case where the
19525 @code{subprogram-name} does not coincide with any of the predefined
19526 @code{GDB} commands.
19528 The effect is to invoke the given subprogram, passing it the
19529 list of parameters that is supplied. The parameters can be expressions and
19530 can include variables from the program being debugged. The
19531 subprogram must be defined
19532 at the library level within your program, and @code{GDB} will call the
19533 subprogram within the environment of your program execution (which
19534 means that the subprogram is free to access or even modify variables
19535 within your program).
19537 The most important use of this facility is in allowing the inclusion of
19538 debugging routines that are tailored to particular data structures
19539 in your program. Such debugging routines can be written to provide a suitably
19540 high-level description of an abstract type, rather than a low-level dump
19541 of its physical layout. After all, the standard
19542 @code{GDB print} command only knows the physical layout of your
19543 types, not their abstract meaning. Debugging routines can provide information
19544 at the desired semantic level and are thus enormously useful.
19546 For example, when debugging GNAT itself, it is crucial to have access to
19547 the contents of the tree nodes used to represent the program internally.
19548 But tree nodes are represented simply by an integer value (which in turn
19549 is an index into a table of nodes).
19550 Using the @code{print} command on a tree node would simply print this integer
19551 value, which is not very useful. But the PN routine (defined in file
19552 treepr.adb in the GNAT sources) takes a tree node as input, and displays
19553 a useful high level representation of the tree node, which includes the
19554 syntactic category of the node, its position in the source, the integers
19555 that denote descendant nodes and parent node, as well as varied
19556 semantic information. To study this example in more detail, you might want to
19557 look at the body of the PN procedure in the stated file.
19559 Another useful application of this capability is to deal with situations of
19560 complex data which are not handled suitably by GDB. For example, if you specify
19561 Convention Fortran for a multi-dimensional array, GDB does not know that
19562 the ordering of array elements has been switched and will not properly
19563 address the array elements. In such a case, instead of trying to print the
19564 elements directly from GDB, you can write a callable procedure that prints
19565 the elements in the desired format.
19567 @node Using the next Command in a Function,Stopping When Ada Exceptions Are Raised,Calling User-Defined Subprograms,Running and Debugging Ada Programs
19568 @anchor{gnat_ugn/gnat_and_program_execution using-the-next-command-in-a-function}@anchor{177}@anchor{gnat_ugn/gnat_and_program_execution id8}@anchor{178}
19569 @subsection Using the @emph{next} Command in a Function
19572 When you use the @code{next} command in a function, the current source
19573 location will advance to the next statement as usual. A special case
19574 arises in the case of a @code{return} statement.
19576 Part of the code for a return statement is the 'epilogue' of the function.
19577 This is the code that returns to the caller. There is only one copy of
19578 this epilogue code, and it is typically associated with the last return
19579 statement in the function if there is more than one return. In some
19580 implementations, this epilogue is associated with the first statement
19583 The result is that if you use the @code{next} command from a return
19584 statement that is not the last return statement of the function you
19585 may see a strange apparent jump to the last return statement or to
19586 the start of the function. You should simply ignore this odd jump.
19587 The value returned is always that from the first return statement
19588 that was stepped through.
19590 @node Stopping When Ada Exceptions Are Raised,Ada Tasks,Using the next Command in a Function,Running and Debugging Ada Programs
19591 @anchor{gnat_ugn/gnat_and_program_execution stopping-when-ada-exceptions-are-raised}@anchor{179}@anchor{gnat_ugn/gnat_and_program_execution id9}@anchor{17a}
19592 @subsection Stopping When Ada Exceptions Are Raised
19595 @geindex Exceptions (in gdb)
19597 You can set catchpoints that stop the program execution when your program
19598 raises selected exceptions.
19607 @item @code{catch exception}
19609 Set a catchpoint that stops execution whenever (any task in the) program
19610 raises any exception.
19617 @item @code{catch exception @emph{name}}
19619 Set a catchpoint that stops execution whenever (any task in the) program
19620 raises the exception @emph{name}.
19627 @item @code{catch exception unhandled}
19629 Set a catchpoint that stops executing whenever (any task in the) program
19630 raises an exception for which there is no handler.
19637 @item @code{info exceptions}, @code{info exceptions @emph{regexp}}
19639 The @code{info exceptions} command permits the user to examine all defined
19640 exceptions within Ada programs. With a regular expression, @emph{regexp}, as
19641 argument, prints out only those exceptions whose name matches @emph{regexp}.
19645 @geindex Tasks (in gdb)
19647 @node Ada Tasks,Debugging Generic Units,Stopping When Ada Exceptions Are Raised,Running and Debugging Ada Programs
19648 @anchor{gnat_ugn/gnat_and_program_execution ada-tasks}@anchor{17b}@anchor{gnat_ugn/gnat_and_program_execution id10}@anchor{17c}
19649 @subsection Ada Tasks
19652 @code{GDB} allows the following task-related commands:
19661 @item @code{info tasks}
19663 This command shows a list of current Ada tasks, as in the following example:
19667 ID TID P-ID Thread Pri State Name
19668 1 8088000 0 807e000 15 Child Activation Wait main_task
19669 2 80a4000 1 80ae000 15 Accept/Select Wait b
19670 3 809a800 1 80a4800 15 Child Activation Wait a
19671 * 4 80ae800 3 80b8000 15 Running c
19674 In this listing, the asterisk before the first task indicates it to be the
19675 currently running task. The first column lists the task ID that is used
19676 to refer to tasks in the following commands.
19680 @geindex Breakpoints and tasks
19686 @code{break`@w{`}*linespec* `@w{`}task} @emph{taskid}, @code{break} @emph{linespec} @code{task} @emph{taskid} @code{if} ...
19690 These commands are like the @code{break ... thread ...}.
19691 @emph{linespec} specifies source lines.
19693 Use the qualifier @code{task @emph{taskid}} with a breakpoint command
19694 to specify that you only want @code{GDB} to stop the program when a
19695 particular Ada task reaches this breakpoint. @emph{taskid} is one of the
19696 numeric task identifiers assigned by @code{GDB}, shown in the first
19697 column of the @code{info tasks} display.
19699 If you do not specify @code{task @emph{taskid}} when you set a
19700 breakpoint, the breakpoint applies to @emph{all} tasks of your
19703 You can use the @code{task} qualifier on conditional breakpoints as
19704 well; in this case, place @code{task @emph{taskid}} before the
19705 breakpoint condition (before the @code{if}).
19709 @geindex Task switching (in gdb)
19715 @code{task @emph{taskno}}
19719 This command allows switching to the task referred by @emph{taskno}. In
19720 particular, this allows browsing of the backtrace of the specified
19721 task. It is advisable to switch back to the original task before
19722 continuing execution otherwise the scheduling of the program may be
19727 For more detailed information on the tasking support,
19728 see @cite{Debugging with GDB}.
19730 @geindex Debugging Generic Units
19734 @node Debugging Generic Units,Remote Debugging with gdbserver,Ada Tasks,Running and Debugging Ada Programs
19735 @anchor{gnat_ugn/gnat_and_program_execution debugging-generic-units}@anchor{17d}@anchor{gnat_ugn/gnat_and_program_execution id11}@anchor{17e}
19736 @subsection Debugging Generic Units
19739 GNAT always uses code expansion for generic instantiation. This means that
19740 each time an instantiation occurs, a complete copy of the original code is
19741 made, with appropriate substitutions of formals by actuals.
19743 It is not possible to refer to the original generic entities in
19744 @code{GDB}, but it is always possible to debug a particular instance of
19745 a generic, by using the appropriate expanded names. For example, if we have
19752 generic package k is
19753 procedure kp (v1 : in out integer);
19757 procedure kp (v1 : in out integer) is
19763 package k1 is new k;
19764 package k2 is new k;
19766 var : integer := 1;
19777 Then to break on a call to procedure kp in the k2 instance, simply
19783 (gdb) break g.k2.kp
19787 When the breakpoint occurs, you can step through the code of the
19788 instance in the normal manner and examine the values of local variables, as for
19791 @geindex Remote Debugging with gdbserver
19793 @node Remote Debugging with gdbserver,GNAT Abnormal Termination or Failure to Terminate,Debugging Generic Units,Running and Debugging Ada Programs
19794 @anchor{gnat_ugn/gnat_and_program_execution remote-debugging-with-gdbserver}@anchor{17f}@anchor{gnat_ugn/gnat_and_program_execution id12}@anchor{180}
19795 @subsection Remote Debugging with gdbserver
19798 On platforms where gdbserver is supported, it is possible to use this tool
19799 to debug your application remotely. This can be useful in situations
19800 where the program needs to be run on a target host that is different
19801 from the host used for development, particularly when the target has
19802 a limited amount of resources (either CPU and/or memory).
19804 To do so, start your program using gdbserver on the target machine.
19805 gdbserver then automatically suspends the execution of your program
19806 at its entry point, waiting for a debugger to connect to it. The
19807 following commands starts an application and tells gdbserver to
19808 wait for a connection with the debugger on localhost port 4444.
19813 $ gdbserver localhost:4444 program
19814 Process program created; pid = 5685
19815 Listening on port 4444
19819 Once gdbserver has started listening, we can tell the debugger to establish
19820 a connection with this gdbserver, and then start the same debugging session
19821 as if the program was being debugged on the same host, directly under
19822 the control of GDB.
19828 (gdb) target remote targethost:4444
19829 Remote debugging using targethost:4444
19830 0x00007f29936d0af0 in ?? () from /lib64/ld-linux-x86-64.so.
19832 Breakpoint 1 at 0x401f0c: file foo.adb, line 3.
19836 Breakpoint 1, foo () at foo.adb:4
19841 It is also possible to use gdbserver to attach to an already running
19842 program, in which case the execution of that program is simply suspended
19843 until the connection between the debugger and gdbserver is established.
19845 For more information on how to use gdbserver, see the @emph{Using the gdbserver Program}
19846 section in @cite{Debugging with GDB}.
19847 GNAT provides support for gdbserver on x86-linux, x86-windows and x86_64-linux.
19849 @geindex Abnormal Termination or Failure to Terminate
19851 @node GNAT Abnormal Termination or Failure to Terminate,Naming Conventions for GNAT Source Files,Remote Debugging with gdbserver,Running and Debugging Ada Programs
19852 @anchor{gnat_ugn/gnat_and_program_execution gnat-abnormal-termination-or-failure-to-terminate}@anchor{181}@anchor{gnat_ugn/gnat_and_program_execution id13}@anchor{182}
19853 @subsection GNAT Abnormal Termination or Failure to Terminate
19856 When presented with programs that contain serious errors in syntax
19858 GNAT may on rare occasions experience problems in operation, such
19860 segmentation fault or illegal memory access, raising an internal
19861 exception, terminating abnormally, or failing to terminate at all.
19862 In such cases, you can activate
19863 various features of GNAT that can help you pinpoint the construct in your
19864 program that is the likely source of the problem.
19866 The following strategies are presented in increasing order of
19867 difficulty, corresponding to your experience in using GNAT and your
19868 familiarity with compiler internals.
19874 Run @code{gcc} with the @code{-gnatf}. This first
19875 switch causes all errors on a given line to be reported. In its absence,
19876 only the first error on a line is displayed.
19878 The @code{-gnatdO} switch causes errors to be displayed as soon as they
19879 are encountered, rather than after compilation is terminated. If GNAT
19880 terminates prematurely or goes into an infinite loop, the last error
19881 message displayed may help to pinpoint the culprit.
19884 Run @code{gcc} with the @code{-v} (verbose) switch. In this
19885 mode, @code{gcc} produces ongoing information about the progress of the
19886 compilation and provides the name of each procedure as code is
19887 generated. This switch allows you to find which Ada procedure was being
19888 compiled when it encountered a code generation problem.
19891 @geindex -gnatdc switch
19897 Run @code{gcc} with the @code{-gnatdc} switch. This is a GNAT specific
19898 switch that does for the front-end what @code{-v} does
19899 for the back end. The system prints the name of each unit,
19900 either a compilation unit or nested unit, as it is being analyzed.
19903 Finally, you can start
19904 @code{gdb} directly on the @code{gnat1} executable. @code{gnat1} is the
19905 front-end of GNAT, and can be run independently (normally it is just
19906 called from @code{gcc}). You can use @code{gdb} on @code{gnat1} as you
19907 would on a C program (but @ref{16d,,The GNAT Debugger GDB} for caveats). The
19908 @code{where} command is the first line of attack; the variable
19909 @code{lineno} (seen by @code{print lineno}), used by the second phase of
19910 @code{gnat1} and by the @code{gcc} backend, indicates the source line at
19911 which the execution stopped, and @code{input_file name} indicates the name of
19915 @node Naming Conventions for GNAT Source Files,Getting Internal Debugging Information,GNAT Abnormal Termination or Failure to Terminate,Running and Debugging Ada Programs
19916 @anchor{gnat_ugn/gnat_and_program_execution naming-conventions-for-gnat-source-files}@anchor{183}@anchor{gnat_ugn/gnat_and_program_execution id14}@anchor{184}
19917 @subsection Naming Conventions for GNAT Source Files
19920 In order to examine the workings of the GNAT system, the following
19921 brief description of its organization may be helpful:
19927 Files with prefix @code{sc} contain the lexical scanner.
19930 All files prefixed with @code{par} are components of the parser. The
19931 numbers correspond to chapters of the Ada Reference Manual. For example,
19932 parsing of select statements can be found in @code{par-ch9.adb}.
19935 All files prefixed with @code{sem} perform semantic analysis. The
19936 numbers correspond to chapters of the Ada standard. For example, all
19937 issues involving context clauses can be found in @code{sem_ch10.adb}. In
19938 addition, some features of the language require sufficient special processing
19939 to justify their own semantic files: sem_aggr for aggregates, sem_disp for
19940 dynamic dispatching, etc.
19943 All files prefixed with @code{exp} perform normalization and
19944 expansion of the intermediate representation (abstract syntax tree, or AST).
19945 these files use the same numbering scheme as the parser and semantics files.
19946 For example, the construction of record initialization procedures is done in
19947 @code{exp_ch3.adb}.
19950 The files prefixed with @code{bind} implement the binder, which
19951 verifies the consistency of the compilation, determines an order of
19952 elaboration, and generates the bind file.
19955 The files @code{atree.ads} and @code{atree.adb} detail the low-level
19956 data structures used by the front-end.
19959 The files @code{sinfo.ads} and @code{sinfo.adb} detail the structure of
19960 the abstract syntax tree as produced by the parser.
19963 The files @code{einfo.ads} and @code{einfo.adb} detail the attributes of
19964 all entities, computed during semantic analysis.
19967 Library management issues are dealt with in files with prefix
19970 @geindex Annex A (in Ada Reference Manual)
19973 Ada files with the prefix @code{a-} are children of @code{Ada}, as
19974 defined in Annex A.
19976 @geindex Annex B (in Ada reference Manual)
19979 Files with prefix @code{i-} are children of @code{Interfaces}, as
19980 defined in Annex B.
19982 @geindex System (package in Ada Reference Manual)
19985 Files with prefix @code{s-} are children of @code{System}. This includes
19986 both language-defined children and GNAT run-time routines.
19988 @geindex GNAT (package)
19991 Files with prefix @code{g-} are children of @code{GNAT}. These are useful
19992 general-purpose packages, fully documented in their specs. All
19993 the other @code{.c} files are modifications of common @code{gcc} files.
19996 @node Getting Internal Debugging Information,Stack Traceback,Naming Conventions for GNAT Source Files,Running and Debugging Ada Programs
19997 @anchor{gnat_ugn/gnat_and_program_execution id15}@anchor{185}@anchor{gnat_ugn/gnat_and_program_execution getting-internal-debugging-information}@anchor{186}
19998 @subsection Getting Internal Debugging Information
20001 Most compilers have internal debugging switches and modes. GNAT
20002 does also, except GNAT internal debugging switches and modes are not
20003 secret. A summary and full description of all the compiler and binder
20004 debug flags are in the file @code{debug.adb}. You must obtain the
20005 sources of the compiler to see the full detailed effects of these flags.
20007 The switches that print the source of the program (reconstructed from
20008 the internal tree) are of general interest for user programs, as are the
20010 the full internal tree, and the entity table (the symbol table
20011 information). The reconstructed source provides a readable version of the
20012 program after the front-end has completed analysis and expansion,
20013 and is useful when studying the performance of specific constructs.
20014 For example, constraint checks are indicated, complex aggregates
20015 are replaced with loops and assignments, and tasking primitives
20016 are replaced with run-time calls.
20020 @geindex stack traceback
20022 @geindex stack unwinding
20024 @node Stack Traceback,Pretty-Printers for the GNAT runtime,Getting Internal Debugging Information,Running and Debugging Ada Programs
20025 @anchor{gnat_ugn/gnat_and_program_execution stack-traceback}@anchor{187}@anchor{gnat_ugn/gnat_and_program_execution id16}@anchor{188}
20026 @subsection Stack Traceback
20029 Traceback is a mechanism to display the sequence of subprogram calls that
20030 leads to a specified execution point in a program. Often (but not always)
20031 the execution point is an instruction at which an exception has been raised.
20032 This mechanism is also known as @emph{stack unwinding} because it obtains
20033 its information by scanning the run-time stack and recovering the activation
20034 records of all active subprograms. Stack unwinding is one of the most
20035 important tools for program debugging.
20037 The first entry stored in traceback corresponds to the deepest calling level,
20038 that is to say the subprogram currently executing the instruction
20039 from which we want to obtain the traceback.
20041 Note that there is no runtime performance penalty when stack traceback
20042 is enabled, and no exception is raised during program execution.
20045 @geindex non-symbolic
20048 * Non-Symbolic Traceback::
20049 * Symbolic Traceback::
20053 @node Non-Symbolic Traceback,Symbolic Traceback,,Stack Traceback
20054 @anchor{gnat_ugn/gnat_and_program_execution non-symbolic-traceback}@anchor{189}@anchor{gnat_ugn/gnat_and_program_execution id17}@anchor{18a}
20055 @subsubsection Non-Symbolic Traceback
20058 Note: this feature is not supported on all platforms. See
20059 @code{GNAT.Traceback} spec in @code{g-traceb.ads}
20060 for a complete list of supported platforms.
20062 @subsubheading Tracebacks From an Unhandled Exception
20065 A runtime non-symbolic traceback is a list of addresses of call instructions.
20066 To enable this feature you must use the @code{-E}
20067 @code{gnatbind} option. With this option a stack traceback is stored as part
20068 of exception information. You can retrieve this information using the
20069 @code{addr2line} tool.
20071 Here is a simple example:
20080 raise Constraint_Error;
20094 $ gnatmake stb -bargs -E
20097 Execution terminated by unhandled exception
20098 Exception name: CONSTRAINT_ERROR
20100 Call stack traceback locations:
20101 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
20105 As we see the traceback lists a sequence of addresses for the unhandled
20106 exception @code{CONSTRAINT_ERROR} raised in procedure P1. It is easy to
20107 guess that this exception come from procedure P1. To translate these
20108 addresses into the source lines where the calls appear, the
20109 @code{addr2line} tool, described below, is invaluable. The use of this tool
20110 requires the program to be compiled with debug information.
20115 $ gnatmake -g stb -bargs -E
20118 Execution terminated by unhandled exception
20119 Exception name: CONSTRAINT_ERROR
20121 Call stack traceback locations:
20122 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
20124 $ addr2line --exe=stb 0x401373 0x40138b 0x40139c 0x401335 0x4011c4
20125 0x4011f1 0x77e892a4
20127 00401373 at d:/stb/stb.adb:5
20128 0040138B at d:/stb/stb.adb:10
20129 0040139C at d:/stb/stb.adb:14
20130 00401335 at d:/stb/b~stb.adb:104
20131 004011C4 at /build/.../crt1.c:200
20132 004011F1 at /build/.../crt1.c:222
20133 77E892A4 in ?? at ??:0
20137 The @code{addr2line} tool has several other useful options:
20142 @multitable {xxxxxxxxxxxxxxxxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
20149 to get the function name corresponding to any location
20153 @code{--demangle=gnat}
20157 to use the gnat decoding mode for the function names.
20158 Note that for binutils version 2.9.x the option is
20159 simply @code{--demangle}.
20165 $ addr2line --exe=stb --functions --demangle=gnat 0x401373 0x40138b
20166 0x40139c 0x401335 0x4011c4 0x4011f1
20168 00401373 in stb.p1 at d:/stb/stb.adb:5
20169 0040138B in stb.p2 at d:/stb/stb.adb:10
20170 0040139C in stb at d:/stb/stb.adb:14
20171 00401335 in main at d:/stb/b~stb.adb:104
20172 004011C4 in <__mingw_CRTStartup> at /build/.../crt1.c:200
20173 004011F1 in <mainCRTStartup> at /build/.../crt1.c:222
20177 From this traceback we can see that the exception was raised in
20178 @code{stb.adb} at line 5, which was reached from a procedure call in
20179 @code{stb.adb} at line 10, and so on. The @code{b~std.adb} is the binder file,
20180 which contains the call to the main program.
20181 @ref{11c,,Running gnatbind}. The remaining entries are assorted runtime routines,
20182 and the output will vary from platform to platform.
20184 It is also possible to use @code{GDB} with these traceback addresses to debug
20185 the program. For example, we can break at a given code location, as reported
20186 in the stack traceback:
20195 Furthermore, this feature is not implemented inside Windows DLL. Only
20196 the non-symbolic traceback is reported in this case.
20201 (gdb) break *0x401373
20202 Breakpoint 1 at 0x401373: file stb.adb, line 5.
20206 It is important to note that the stack traceback addresses
20207 do not change when debug information is included. This is particularly useful
20208 because it makes it possible to release software without debug information (to
20209 minimize object size), get a field report that includes a stack traceback
20210 whenever an internal bug occurs, and then be able to retrieve the sequence
20211 of calls with the same program compiled with debug information.
20213 @subsubheading Tracebacks From Exception Occurrences
20216 Non-symbolic tracebacks are obtained by using the @code{-E} binder argument.
20217 The stack traceback is attached to the exception information string, and can
20218 be retrieved in an exception handler within the Ada program, by means of the
20219 Ada facilities defined in @code{Ada.Exceptions}. Here is a simple example:
20225 with Ada.Exceptions;
20230 use Ada.Exceptions;
20238 Text_IO.Put_Line (Exception_Information (E));
20252 This program will output:
20259 Exception name: CONSTRAINT_ERROR
20260 Message: stb.adb:12
20261 Call stack traceback locations:
20262 0x4015e4 0x401633 0x401644 0x401461 0x4011c4 0x4011f1 0x77e892a4
20266 @subsubheading Tracebacks From Anywhere in a Program
20269 It is also possible to retrieve a stack traceback from anywhere in a
20270 program. For this you need to
20271 use the @code{GNAT.Traceback} API. This package includes a procedure called
20272 @code{Call_Chain} that computes a complete stack traceback, as well as useful
20273 display procedures described below. It is not necessary to use the
20274 @code{-E} @code{gnatbind} option in this case, because the stack traceback mechanism
20275 is invoked explicitly.
20277 In the following example we compute a traceback at a specific location in
20278 the program, and we display it using @code{GNAT.Debug_Utilities.Image} to
20279 convert addresses to strings:
20285 with GNAT.Traceback;
20286 with GNAT.Debug_Utilities;
20292 use GNAT.Traceback;
20295 TB : Tracebacks_Array (1 .. 10);
20296 -- We are asking for a maximum of 10 stack frames.
20298 -- Len will receive the actual number of stack frames returned.
20300 Call_Chain (TB, Len);
20302 Text_IO.Put ("In STB.P1 : ");
20304 for K in 1 .. Len loop
20305 Text_IO.Put (Debug_Utilities.Image (TB (K)));
20326 In STB.P1 : 16#0040_F1E4# 16#0040_14F2# 16#0040_170B# 16#0040_171C#
20327 16#0040_1461# 16#0040_11C4# 16#0040_11F1# 16#77E8_92A4#
20331 You can then get further information by invoking the @code{addr2line}
20332 tool as described earlier (note that the hexadecimal addresses
20333 need to be specified in C format, with a leading '0x').
20338 @node Symbolic Traceback,,Non-Symbolic Traceback,Stack Traceback
20339 @anchor{gnat_ugn/gnat_and_program_execution id18}@anchor{18b}@anchor{gnat_ugn/gnat_and_program_execution symbolic-traceback}@anchor{18c}
20340 @subsubsection Symbolic Traceback
20343 A symbolic traceback is a stack traceback in which procedure names are
20344 associated with each code location.
20346 Note that this feature is not supported on all platforms. See
20347 @code{GNAT.Traceback.Symbolic} spec in @code{g-trasym.ads} for a complete
20348 list of currently supported platforms.
20350 Note that the symbolic traceback requires that the program be compiled
20351 with debug information. If it is not compiled with debug information
20352 only the non-symbolic information will be valid.
20354 @subsubheading Tracebacks From Exception Occurrences
20357 Here is an example:
20363 with GNAT.Traceback.Symbolic;
20369 raise Constraint_Error;
20386 Ada.Text_IO.Put_Line (GNAT.Traceback.Symbolic.Symbolic_Traceback (E));
20391 $ gnatmake -g .\stb -bargs -E
20394 0040149F in stb.p1 at stb.adb:8
20395 004014B7 in stb.p2 at stb.adb:13
20396 004014CF in stb.p3 at stb.adb:18
20397 004015DD in ada.stb at stb.adb:22
20398 00401461 in main at b~stb.adb:168
20399 004011C4 in __mingw_CRTStartup at crt1.c:200
20400 004011F1 in mainCRTStartup at crt1.c:222
20401 77E892A4 in ?? at ??:0
20405 In the above example the @code{.\} syntax in the @code{gnatmake} command
20406 is currently required by @code{addr2line} for files that are in
20407 the current working directory.
20408 Moreover, the exact sequence of linker options may vary from platform
20410 The above @code{-largs} section is for Windows platforms. By contrast,
20411 under Unix there is no need for the @code{-largs} section.
20412 Differences across platforms are due to details of linker implementation.
20414 @subsubheading Tracebacks From Anywhere in a Program
20417 It is possible to get a symbolic stack traceback
20418 from anywhere in a program, just as for non-symbolic tracebacks.
20419 The first step is to obtain a non-symbolic
20420 traceback, and then call @code{Symbolic_Traceback} to compute the symbolic
20421 information. Here is an example:
20427 with GNAT.Traceback;
20428 with GNAT.Traceback.Symbolic;
20433 use GNAT.Traceback;
20434 use GNAT.Traceback.Symbolic;
20437 TB : Tracebacks_Array (1 .. 10);
20438 -- We are asking for a maximum of 10 stack frames.
20440 -- Len will receive the actual number of stack frames returned.
20442 Call_Chain (TB, Len);
20443 Text_IO.Put_Line (Symbolic_Traceback (TB (1 .. Len)));
20457 @subsubheading Automatic Symbolic Tracebacks
20460 Symbolic tracebacks may also be enabled by using the -Es switch to gnatbind (as
20461 in @code{gprbuild -g ... -bargs -Es}).
20462 This will cause the Exception_Information to contain a symbolic traceback,
20463 which will also be printed if an unhandled exception terminates the
20466 @node Pretty-Printers for the GNAT runtime,,Stack Traceback,Running and Debugging Ada Programs
20467 @anchor{gnat_ugn/gnat_and_program_execution id19}@anchor{18d}@anchor{gnat_ugn/gnat_and_program_execution pretty-printers-for-the-gnat-runtime}@anchor{18e}
20468 @subsection Pretty-Printers for the GNAT runtime
20471 As discussed in @cite{Calling User-Defined Subprograms}, GDB's
20472 @code{print} command only knows about the physical layout of program data
20473 structures and therefore normally displays only low-level dumps, which
20474 are often hard to understand.
20476 An example of this is when trying to display the contents of an Ada
20477 standard container, such as @code{Ada.Containers.Ordered_Maps.Map}:
20482 with Ada.Containers.Ordered_Maps;
20485 package Int_To_Nat is
20486 new Ada.Containers.Ordered_Maps (Integer, Natural);
20488 Map : Int_To_Nat.Map;
20490 Map.Insert (1, 10);
20491 Map.Insert (2, 20);
20492 Map.Insert (3, 30);
20494 Map.Clear; -- BREAK HERE
20499 When this program is built with debugging information and run under
20500 GDB up to the @code{Map.Clear} statement, trying to print @code{Map} will
20501 yield information that is only relevant to the developers of our standard
20523 Fortunately, GDB has a feature called pretty-printers@footnote{http://docs.adacore.com/gdb-docs/html/gdb.html#Pretty_002dPrinter-Introduction},
20524 which allows customizing how GDB displays data structures. The GDB
20525 shipped with GNAT embeds such pretty-printers for the most common
20526 containers in the standard library. To enable them, either run the
20527 following command manually under GDB or add it to your @code{.gdbinit} file:
20532 python import gnatdbg; gnatdbg.setup()
20536 Once this is done, GDB's @code{print} command will automatically use
20537 these pretty-printers when appropriate. Using the previous example:
20543 $1 = pp.int_to_nat.map of length 3 = @{
20551 Pretty-printers are invoked each time GDB tries to display a value,
20552 including when displaying the arguments of a called subprogram (in
20553 GDB's @code{backtrace} command) or when printing the value returned by a
20554 function (in GDB's @code{finish} command).
20556 To display a value without involving pretty-printers, @code{print} can be
20557 invoked with its @code{/r} option:
20568 Finer control of pretty-printers is also possible: see GDB's online documentation@footnote{http://docs.adacore.com/gdb-docs/html/gdb.html#Pretty_002dPrinter-Commands}
20569 for more information.
20573 @node Profiling,Improving Performance,Running and Debugging Ada Programs,GNAT and Program Execution
20574 @anchor{gnat_ugn/gnat_and_program_execution profiling}@anchor{25}@anchor{gnat_ugn/gnat_and_program_execution id20}@anchor{18f}
20578 This section describes how to use the the @code{gprof} profiler tool on Ada
20586 * Profiling an Ada Program with gprof::
20590 @node Profiling an Ada Program with gprof,,,Profiling
20591 @anchor{gnat_ugn/gnat_and_program_execution id21}@anchor{190}@anchor{gnat_ugn/gnat_and_program_execution profiling-an-ada-program-with-gprof}@anchor{191}
20592 @subsection Profiling an Ada Program with gprof
20595 This section is not meant to be an exhaustive documentation of @code{gprof}.
20596 Full documentation for it can be found in the @cite{GNU Profiler User's Guide}
20597 documentation that is part of this GNAT distribution.
20599 Profiling a program helps determine the parts of a program that are executed
20600 most often, and are therefore the most time-consuming.
20602 @code{gprof} is the standard GNU profiling tool; it has been enhanced to
20603 better handle Ada programs and multitasking.
20604 It is currently supported on the following platforms
20616 In order to profile a program using @code{gprof}, several steps are needed:
20622 Instrument the code, which requires a full recompilation of the project with the
20626 Execute the program under the analysis conditions, i.e. with the desired
20630 Analyze the results using the @code{gprof} tool.
20633 The following sections detail the different steps, and indicate how
20634 to interpret the results.
20637 * Compilation for profiling::
20638 * Program execution::
20640 * Interpretation of profiling results::
20644 @node Compilation for profiling,Program execution,,Profiling an Ada Program with gprof
20645 @anchor{gnat_ugn/gnat_and_program_execution id22}@anchor{192}@anchor{gnat_ugn/gnat_and_program_execution compilation-for-profiling}@anchor{193}
20646 @subsubsection Compilation for profiling
20650 @geindex for profiling
20652 @geindex -pg (gnatlink)
20653 @geindex for profiling
20655 In order to profile a program the first step is to tell the compiler
20656 to generate the necessary profiling information. The compiler switch to be used
20657 is @code{-pg}, which must be added to other compilation switches. This
20658 switch needs to be specified both during compilation and link stages, and can
20659 be specified once when using gnatmake:
20664 $ gnatmake -f -pg -P my_project
20668 Note that only the objects that were compiled with the @code{-pg} switch will
20669 be profiled; if you need to profile your whole project, use the @code{-f}
20670 gnatmake switch to force full recompilation.
20672 @node Program execution,Running gprof,Compilation for profiling,Profiling an Ada Program with gprof
20673 @anchor{gnat_ugn/gnat_and_program_execution program-execution}@anchor{194}@anchor{gnat_ugn/gnat_and_program_execution id23}@anchor{195}
20674 @subsubsection Program execution
20677 Once the program has been compiled for profiling, you can run it as usual.
20679 The only constraint imposed by profiling is that the program must terminate
20680 normally. An interrupted program (via a Ctrl-C, kill, etc.) will not be
20683 Once the program completes execution, a data file called @code{gmon.out} is
20684 generated in the directory where the program was launched from. If this file
20685 already exists, it will be overwritten.
20687 @node Running gprof,Interpretation of profiling results,Program execution,Profiling an Ada Program with gprof
20688 @anchor{gnat_ugn/gnat_and_program_execution running-gprof}@anchor{196}@anchor{gnat_ugn/gnat_and_program_execution id24}@anchor{197}
20689 @subsubsection Running gprof
20692 The @code{gprof} tool is called as follow:
20697 $ gprof my_prog gmon.out
20710 The complete form of the gprof command line is the following:
20715 $ gprof [switches] [executable [data-file]]
20719 @code{gprof} supports numerous switches. The order of these
20720 switch does not matter. The full list of options can be found in
20721 the GNU Profiler User's Guide documentation that comes with this documentation.
20723 The following is the subset of those switches that is most relevant:
20725 @geindex --demangle (gprof)
20730 @item @code{--demangle[=@emph{style}]}, @code{--no-demangle}
20732 These options control whether symbol names should be demangled when
20733 printing output. The default is to demangle C++ symbols. The
20734 @code{--no-demangle} option may be used to turn off demangling. Different
20735 compilers have different mangling styles. The optional demangling style
20736 argument can be used to choose an appropriate demangling style for your
20737 compiler, in particular Ada symbols generated by GNAT can be demangled using
20738 @code{--demangle=gnat}.
20741 @geindex -e (gprof)
20746 @item @code{-e @emph{function_name}}
20748 The @code{-e @emph{function}} option tells @code{gprof} not to print
20749 information about the function @code{function_name} (and its
20750 children...) in the call graph. The function will still be listed
20751 as a child of any functions that call it, but its index number will be
20752 shown as @code{[not printed]}. More than one @code{-e} option may be
20753 given; only one @code{function_name} may be indicated with each @code{-e}
20757 @geindex -E (gprof)
20762 @item @code{-E @emph{function_name}}
20764 The @code{-E @emph{function}} option works like the @code{-e} option, but
20765 execution time spent in the function (and children who were not called from
20766 anywhere else), will not be used to compute the percentages-of-time for
20767 the call graph. More than one @code{-E} option may be given; only one
20768 @code{function_name} may be indicated with each @code{-E`} option.
20771 @geindex -f (gprof)
20776 @item @code{-f @emph{function_name}}
20778 The @code{-f @emph{function}} option causes @code{gprof} to limit the
20779 call graph to the function @code{function_name} and its children (and
20780 their children...). More than one @code{-f} option may be given;
20781 only one @code{function_name} may be indicated with each @code{-f}
20785 @geindex -F (gprof)
20790 @item @code{-F @emph{function_name}}
20792 The @code{-F @emph{function}} option works like the @code{-f} option, but
20793 only time spent in the function and its children (and their
20794 children...) will be used to determine total-time and
20795 percentages-of-time for the call graph. More than one @code{-F} option
20796 may be given; only one @code{function_name} may be indicated with each
20797 @code{-F} option. The @code{-F} option overrides the @code{-E} option.
20800 @node Interpretation of profiling results,,Running gprof,Profiling an Ada Program with gprof
20801 @anchor{gnat_ugn/gnat_and_program_execution id25}@anchor{198}@anchor{gnat_ugn/gnat_and_program_execution interpretation-of-profiling-results}@anchor{199}
20802 @subsubsection Interpretation of profiling results
20805 The results of the profiling analysis are represented by two arrays: the
20806 'flat profile' and the 'call graph'. Full documentation of those outputs
20807 can be found in the GNU Profiler User's Guide.
20809 The flat profile shows the time spent in each function of the program, and how
20810 many time it has been called. This allows you to locate easily the most
20811 time-consuming functions.
20813 The call graph shows, for each subprogram, the subprograms that call it,
20814 and the subprograms that it calls. It also provides an estimate of the time
20815 spent in each of those callers/called subprograms.
20817 @node Improving Performance,Overflow Check Handling in GNAT,Profiling,GNAT and Program Execution
20818 @anchor{gnat_ugn/gnat_and_program_execution improving-performance}@anchor{26}@anchor{gnat_ugn/gnat_and_program_execution id26}@anchor{168}
20819 @section Improving Performance
20822 @geindex Improving performance
20824 This section presents several topics related to program performance.
20825 It first describes some of the tradeoffs that need to be considered
20826 and some of the techniques for making your program run faster.
20829 It then documents the unused subprogram/data elimination feature,
20830 which can reduce the size of program executables.
20833 * Performance Considerations::
20834 * Text_IO Suggestions::
20835 * Reducing Size of Executables with Unused Subprogram/Data Elimination::
20839 @node Performance Considerations,Text_IO Suggestions,,Improving Performance
20840 @anchor{gnat_ugn/gnat_and_program_execution performance-considerations}@anchor{19a}@anchor{gnat_ugn/gnat_and_program_execution id27}@anchor{19b}
20841 @subsection Performance Considerations
20844 The GNAT system provides a number of options that allow a trade-off
20851 performance of the generated code
20854 speed of compilation
20857 minimization of dependences and recompilation
20860 the degree of run-time checking.
20863 The defaults (if no options are selected) aim at improving the speed
20864 of compilation and minimizing dependences, at the expense of performance
20865 of the generated code:
20874 no inlining of subprogram calls
20877 all run-time checks enabled except overflow and elaboration checks
20880 These options are suitable for most program development purposes. This
20881 section describes how you can modify these choices, and also provides
20882 some guidelines on debugging optimized code.
20885 * Controlling Run-Time Checks::
20886 * Use of Restrictions::
20887 * Optimization Levels::
20888 * Debugging Optimized Code::
20889 * Inlining of Subprograms::
20890 * Floating_Point_Operations::
20891 * Vectorization of loops::
20892 * Other Optimization Switches::
20893 * Optimization and Strict Aliasing::
20894 * Aliased Variables and Optimization::
20895 * Atomic Variables and Optimization::
20896 * Passive Task Optimization::
20900 @node Controlling Run-Time Checks,Use of Restrictions,,Performance Considerations
20901 @anchor{gnat_ugn/gnat_and_program_execution id28}@anchor{19c}@anchor{gnat_ugn/gnat_and_program_execution controlling-run-time-checks}@anchor{19d}
20902 @subsubsection Controlling Run-Time Checks
20905 By default, GNAT generates all run-time checks, except stack overflow
20906 checks, and checks for access before elaboration on subprogram
20907 calls. The latter are not required in default mode, because all
20908 necessary checking is done at compile time.
20910 @geindex -gnatp (gcc)
20912 @geindex -gnato (gcc)
20914 The gnat switch, @code{-gnatp} allows this default to be modified. See
20915 @ref{f9,,Run-Time Checks}.
20917 Our experience is that the default is suitable for most development
20920 Elaboration checks are off by default, and also not needed by default, since
20921 GNAT uses a static elaboration analysis approach that avoids the need for
20922 run-time checking. This manual contains a full chapter discussing the issue
20923 of elaboration checks, and if the default is not satisfactory for your use,
20924 you should read this chapter.
20926 For validity checks, the minimal checks required by the Ada Reference
20927 Manual (for case statements and assignments to array elements) are on
20928 by default. These can be suppressed by use of the @code{-gnatVn} switch.
20929 Note that in Ada 83, there were no validity checks, so if the Ada 83 mode
20930 is acceptable (or when comparing GNAT performance with an Ada 83 compiler),
20931 it may be reasonable to routinely use @code{-gnatVn}. Validity checks
20932 are also suppressed entirely if @code{-gnatp} is used.
20934 @geindex Overflow checks
20941 @geindex Unsuppress
20943 @geindex pragma Suppress
20945 @geindex pragma Unsuppress
20947 Note that the setting of the switches controls the default setting of
20948 the checks. They may be modified using either @code{pragma Suppress} (to
20949 remove checks) or @code{pragma Unsuppress} (to add back suppressed
20950 checks) in the program source.
20952 @node Use of Restrictions,Optimization Levels,Controlling Run-Time Checks,Performance Considerations
20953 @anchor{gnat_ugn/gnat_and_program_execution id29}@anchor{19e}@anchor{gnat_ugn/gnat_and_program_execution use-of-restrictions}@anchor{19f}
20954 @subsubsection Use of Restrictions
20957 The use of pragma Restrictions allows you to control which features are
20958 permitted in your program. Apart from the obvious point that if you avoid
20959 relatively expensive features like finalization (enforceable by the use
20960 of pragma Restrictions (No_Finalization), the use of this pragma does not
20961 affect the generated code in most cases.
20963 One notable exception to this rule is that the possibility of task abort
20964 results in some distributed overhead, particularly if finalization or
20965 exception handlers are used. The reason is that certain sections of code
20966 have to be marked as non-abortable.
20968 If you use neither the @code{abort} statement, nor asynchronous transfer
20969 of control (@code{select ... then abort}), then this distributed overhead
20970 is removed, which may have a general positive effect in improving
20971 overall performance. Especially code involving frequent use of tasking
20972 constructs and controlled types will show much improved performance.
20973 The relevant restrictions pragmas are
20978 pragma Restrictions (No_Abort_Statements);
20979 pragma Restrictions (Max_Asynchronous_Select_Nesting => 0);
20983 It is recommended that these restriction pragmas be used if possible. Note
20984 that this also means that you can write code without worrying about the
20985 possibility of an immediate abort at any point.
20987 @node Optimization Levels,Debugging Optimized Code,Use of Restrictions,Performance Considerations
20988 @anchor{gnat_ugn/gnat_and_program_execution id30}@anchor{1a0}@anchor{gnat_ugn/gnat_and_program_execution optimization-levels}@anchor{fc}
20989 @subsubsection Optimization Levels
20994 Without any optimization option,
20995 the compiler's goal is to reduce the cost of
20996 compilation and to make debugging produce the expected results.
20997 Statements are independent: if you stop the program with a breakpoint between
20998 statements, you can then assign a new value to any variable or change
20999 the program counter to any other statement in the subprogram and get exactly
21000 the results you would expect from the source code.
21002 Turning on optimization makes the compiler attempt to improve the
21003 performance and/or code size at the expense of compilation time and
21004 possibly the ability to debug the program.
21006 If you use multiple
21007 -O options, with or without level numbers,
21008 the last such option is the one that is effective.
21010 The default is optimization off. This results in the fastest compile
21011 times, but GNAT makes absolutely no attempt to optimize, and the
21012 generated programs are considerably larger and slower than when
21013 optimization is enabled. You can use the
21014 @code{-O} switch (the permitted forms are @code{-O0}, @code{-O1}
21015 @code{-O2}, @code{-O3}, and @code{-Os})
21016 to @code{gcc} to control the optimization level:
21027 No optimization (the default);
21028 generates unoptimized code but has
21029 the fastest compilation time.
21031 Note that many other compilers do substantial optimization even
21032 if 'no optimization' is specified. With gcc, it is very unusual
21033 to use @code{-O0} for production if execution time is of any concern,
21034 since @code{-O0} means (almost) no optimization. This difference
21035 between gcc and other compilers should be kept in mind when
21036 doing performance comparisons.
21045 Moderate optimization;
21046 optimizes reasonably well but does not
21047 degrade compilation time significantly.
21057 generates highly optimized code and has
21058 the slowest compilation time.
21067 Full optimization as in @code{-O2};
21068 also uses more aggressive automatic inlining of subprograms within a unit
21069 (@ref{10f,,Inlining of Subprograms}) and attempts to vectorize loops.
21078 Optimize space usage (code and data) of resulting program.
21082 Higher optimization levels perform more global transformations on the
21083 program and apply more expensive analysis algorithms in order to generate
21084 faster and more compact code. The price in compilation time, and the
21085 resulting improvement in execution time,
21086 both depend on the particular application and the hardware environment.
21087 You should experiment to find the best level for your application.
21089 Since the precise set of optimizations done at each level will vary from
21090 release to release (and sometime from target to target), it is best to think
21091 of the optimization settings in general terms.
21092 See the @emph{Options That Control Optimization} section in
21093 @cite{Using the GNU Compiler Collection (GCC)}
21095 the @code{-O} settings and a number of @code{-f} options that
21096 individually enable or disable specific optimizations.
21098 Unlike some other compilation systems, @code{gcc} has
21099 been tested extensively at all optimization levels. There are some bugs
21100 which appear only with optimization turned on, but there have also been
21101 bugs which show up only in @emph{unoptimized} code. Selecting a lower
21102 level of optimization does not improve the reliability of the code
21103 generator, which in practice is highly reliable at all optimization
21106 Note regarding the use of @code{-O3}: The use of this optimization level
21107 ought not to be automatically preferred over that of level @code{-O2},
21108 since it often results in larger executables which may run more slowly.
21109 See further discussion of this point in @ref{10f,,Inlining of Subprograms}.
21111 @node Debugging Optimized Code,Inlining of Subprograms,Optimization Levels,Performance Considerations
21112 @anchor{gnat_ugn/gnat_and_program_execution debugging-optimized-code}@anchor{1a1}@anchor{gnat_ugn/gnat_and_program_execution id31}@anchor{1a2}
21113 @subsubsection Debugging Optimized Code
21116 @geindex Debugging optimized code
21118 @geindex Optimization and debugging
21120 Although it is possible to do a reasonable amount of debugging at
21121 nonzero optimization levels,
21122 the higher the level the more likely that
21123 source-level constructs will have been eliminated by optimization.
21124 For example, if a loop is strength-reduced, the loop
21125 control variable may be completely eliminated and thus cannot be
21126 displayed in the debugger.
21127 This can only happen at @code{-O2} or @code{-O3}.
21128 Explicit temporary variables that you code might be eliminated at
21129 level @code{-O1} or higher.
21133 The use of the @code{-g} switch,
21134 which is needed for source-level debugging,
21135 affects the size of the program executable on disk,
21136 and indeed the debugging information can be quite large.
21137 However, it has no effect on the generated code (and thus does not
21138 degrade performance)
21140 Since the compiler generates debugging tables for a compilation unit before
21141 it performs optimizations, the optimizing transformations may invalidate some
21142 of the debugging data. You therefore need to anticipate certain
21143 anomalous situations that may arise while debugging optimized code.
21144 These are the most common cases:
21150 @emph{The 'hopping Program Counter':} Repeated @code{step} or @code{next}
21152 the PC bouncing back and forth in the code. This may result from any of
21153 the following optimizations:
21159 @emph{Common subexpression elimination:} using a single instance of code for a
21160 quantity that the source computes several times. As a result you
21161 may not be able to stop on what looks like a statement.
21164 @emph{Invariant code motion:} moving an expression that does not change within a
21165 loop, to the beginning of the loop.
21168 @emph{Instruction scheduling:} moving instructions so as to
21169 overlap loads and stores (typically) with other code, or in
21170 general to move computations of values closer to their uses. Often
21171 this causes you to pass an assignment statement without the assignment
21172 happening and then later bounce back to the statement when the
21173 value is actually needed. Placing a breakpoint on a line of code
21174 and then stepping over it may, therefore, not always cause all the
21175 expected side-effects.
21179 @emph{The 'big leap':} More commonly known as @emph{cross-jumping}, in which
21180 two identical pieces of code are merged and the program counter suddenly
21181 jumps to a statement that is not supposed to be executed, simply because
21182 it (and the code following) translates to the same thing as the code
21183 that @emph{was} supposed to be executed. This effect is typically seen in
21184 sequences that end in a jump, such as a @code{goto}, a @code{return}, or
21185 a @code{break} in a C @code{switch} statement.
21188 @emph{The 'roving variable':} The symptom is an unexpected value in a variable.
21189 There are various reasons for this effect:
21195 In a subprogram prologue, a parameter may not yet have been moved to its
21199 A variable may be dead, and its register re-used. This is
21200 probably the most common cause.
21203 As mentioned above, the assignment of a value to a variable may
21207 A variable may be eliminated entirely by value propagation or
21208 other means. In this case, GCC may incorrectly generate debugging
21209 information for the variable
21212 In general, when an unexpected value appears for a local variable or parameter
21213 you should first ascertain if that value was actually computed by
21214 your program, as opposed to being incorrectly reported by the debugger.
21216 array elements in an object designated by an access value
21217 are generally less of a problem, once you have ascertained that the access
21219 Typically, this means checking variables in the preceding code and in the
21220 calling subprogram to verify that the value observed is explainable from other
21221 values (one must apply the procedure recursively to those
21222 other values); or re-running the code and stopping a little earlier
21223 (perhaps before the call) and stepping to better see how the variable obtained
21224 the value in question; or continuing to step @emph{from} the point of the
21225 strange value to see if code motion had simply moved the variable's
21229 In light of such anomalies, a recommended technique is to use @code{-O0}
21230 early in the software development cycle, when extensive debugging capabilities
21231 are most needed, and then move to @code{-O1} and later @code{-O2} as
21232 the debugger becomes less critical.
21233 Whether to use the @code{-g} switch in the release version is
21234 a release management issue.
21235 Note that if you use @code{-g} you can then use the @code{strip} program
21236 on the resulting executable,
21237 which removes both debugging information and global symbols.
21239 @node Inlining of Subprograms,Floating_Point_Operations,Debugging Optimized Code,Performance Considerations
21240 @anchor{gnat_ugn/gnat_and_program_execution id32}@anchor{1a3}@anchor{gnat_ugn/gnat_and_program_execution inlining-of-subprograms}@anchor{10f}
21241 @subsubsection Inlining of Subprograms
21244 A call to a subprogram in the current unit is inlined if all the
21245 following conditions are met:
21251 The optimization level is at least @code{-O1}.
21254 The called subprogram is suitable for inlining: It must be small enough
21255 and not contain something that @code{gcc} cannot support in inlined
21258 @geindex pragma Inline
21263 Any one of the following applies: @code{pragma Inline} is applied to the
21264 subprogram; the subprogram is local to the unit and called once from
21265 within it; the subprogram is small and optimization level @code{-O2} is
21266 specified; optimization level @code{-O3} is specified.
21269 Calls to subprograms in @emph{with}ed units are normally not inlined.
21270 To achieve actual inlining (that is, replacement of the call by the code
21271 in the body of the subprogram), the following conditions must all be true:
21277 The optimization level is at least @code{-O1}.
21280 The called subprogram is suitable for inlining: It must be small enough
21281 and not contain something that @code{gcc} cannot support in inlined
21285 There is a @code{pragma Inline} for the subprogram.
21288 The @code{-gnatn} switch is used on the command line.
21291 Even if all these conditions are met, it may not be possible for
21292 the compiler to inline the call, due to the length of the body,
21293 or features in the body that make it impossible for the compiler
21294 to do the inlining.
21296 Note that specifying the @code{-gnatn} switch causes additional
21297 compilation dependencies. Consider the following:
21319 With the default behavior (no @code{-gnatn} switch specified), the
21320 compilation of the @code{Main} procedure depends only on its own source,
21321 @code{main.adb}, and the spec of the package in file @code{r.ads}. This
21322 means that editing the body of @code{R} does not require recompiling
21325 On the other hand, the call @code{R.Q} is not inlined under these
21326 circumstances. If the @code{-gnatn} switch is present when @code{Main}
21327 is compiled, the call will be inlined if the body of @code{Q} is small
21328 enough, but now @code{Main} depends on the body of @code{R} in
21329 @code{r.adb} as well as on the spec. This means that if this body is edited,
21330 the main program must be recompiled. Note that this extra dependency
21331 occurs whether or not the call is in fact inlined by @code{gcc}.
21333 The use of front end inlining with @code{-gnatN} generates similar
21334 additional dependencies.
21336 @geindex -fno-inline (gcc)
21338 Note: The @code{-fno-inline} switch overrides all other conditions and ensures that
21339 no inlining occurs, unless requested with pragma Inline_Always for @code{gcc}
21340 back-ends. The extra dependences resulting from @code{-gnatn} will still be active,
21341 even if this switch is used to suppress the resulting inlining actions.
21343 @geindex -fno-inline-functions (gcc)
21345 Note: The @code{-fno-inline-functions} switch can be used to prevent
21346 automatic inlining of subprograms if @code{-O3} is used.
21348 @geindex -fno-inline-small-functions (gcc)
21350 Note: The @code{-fno-inline-small-functions} switch can be used to prevent
21351 automatic inlining of small subprograms if @code{-O2} is used.
21353 @geindex -fno-inline-functions-called-once (gcc)
21355 Note: The @code{-fno-inline-functions-called-once} switch
21356 can be used to prevent inlining of subprograms local to the unit
21357 and called once from within it if @code{-O1} is used.
21359 Note regarding the use of @code{-O3}: @code{-gnatn} is made up of two
21360 sub-switches @code{-gnatn1} and @code{-gnatn2} that can be directly
21361 specified in lieu of it, @code{-gnatn} being translated into one of them
21362 based on the optimization level. With @code{-O2} or below, @code{-gnatn}
21363 is equivalent to @code{-gnatn1} which activates pragma @code{Inline} with
21364 moderate inlining across modules. With @code{-O3}, @code{-gnatn} is
21365 equivalent to @code{-gnatn2} which activates pragma @code{Inline} with
21366 full inlining across modules. If you have used pragma @code{Inline} in
21367 appropriate cases, then it is usually much better to use @code{-O2}
21368 and @code{-gnatn} and avoid the use of @code{-O3} which has the additional
21369 effect of inlining subprograms you did not think should be inlined. We have
21370 found that the use of @code{-O3} may slow down the compilation and increase
21371 the code size by performing excessive inlining, leading to increased
21372 instruction cache pressure from the increased code size and thus minor
21373 performance improvements. So the bottom line here is that you should not
21374 automatically assume that @code{-O3} is better than @code{-O2}, and
21375 indeed you should use @code{-O3} only if tests show that it actually
21376 improves performance for your program.
21378 @node Floating_Point_Operations,Vectorization of loops,Inlining of Subprograms,Performance Considerations
21379 @anchor{gnat_ugn/gnat_and_program_execution floating-point-operations}@anchor{1a4}@anchor{gnat_ugn/gnat_and_program_execution id33}@anchor{1a5}
21380 @subsubsection Floating_Point_Operations
21383 @geindex Floating-Point Operations
21385 On almost all targets, GNAT maps Float and Long_Float to the 32-bit and
21386 64-bit standard IEEE floating-point representations, and operations will
21387 use standard IEEE arithmetic as provided by the processor. On most, but
21388 not all, architectures, the attribute Machine_Overflows is False for these
21389 types, meaning that the semantics of overflow is implementation-defined.
21390 In the case of GNAT, these semantics correspond to the normal IEEE
21391 treatment of infinities and NaN (not a number) values. For example,
21392 1.0 / 0.0 yields plus infinitiy and 0.0 / 0.0 yields a NaN. By
21393 avoiding explicit overflow checks, the performance is greatly improved
21394 on many targets. However, if required, floating-point overflow can be
21395 enabled by the use of the pragma Check_Float_Overflow.
21397 Another consideration that applies specifically to x86 32-bit
21398 architectures is which form of floating-point arithmetic is used.
21399 By default the operations use the old style x86 floating-point,
21400 which implements an 80-bit extended precision form (on these
21401 architectures the type Long_Long_Float corresponds to that form).
21402 In addition, generation of efficient code in this mode means that
21403 the extended precision form will be used for intermediate results.
21404 This may be helpful in improving the final precision of a complex
21405 expression. However it means that the results obtained on the x86
21406 will be different from those on other architectures, and for some
21407 algorithms, the extra intermediate precision can be detrimental.
21409 In addition to this old-style floating-point, all modern x86 chips
21410 implement an alternative floating-point operation model referred
21411 to as SSE2. In this model there is no extended form, and furthermore
21412 execution performance is significantly enhanced. To force GNAT to use
21413 this more modern form, use both of the switches:
21417 -msse2 -mfpmath=sse
21420 A unit compiled with these switches will automatically use the more
21421 efficient SSE2 instruction set for Float and Long_Float operations.
21422 Note that the ABI has the same form for both floating-point models,
21423 so it is permissible to mix units compiled with and without these
21426 @node Vectorization of loops,Other Optimization Switches,Floating_Point_Operations,Performance Considerations
21427 @anchor{gnat_ugn/gnat_and_program_execution id34}@anchor{1a6}@anchor{gnat_ugn/gnat_and_program_execution vectorization-of-loops}@anchor{1a7}
21428 @subsubsection Vectorization of loops
21431 @geindex Optimization Switches
21433 You can take advantage of the auto-vectorizer present in the @code{gcc}
21434 back end to vectorize loops with GNAT. The corresponding command line switch
21435 is @code{-ftree-vectorize} but, as it is enabled by default at @code{-O3}
21436 and other aggressive optimizations helpful for vectorization also are enabled
21437 by default at this level, using @code{-O3} directly is recommended.
21439 You also need to make sure that the target architecture features a supported
21440 SIMD instruction set. For example, for the x86 architecture, you should at
21441 least specify @code{-msse2} to get significant vectorization (but you don't
21442 need to specify it for x86-64 as it is part of the base 64-bit architecture).
21443 Similarly, for the PowerPC architecture, you should specify @code{-maltivec}.
21445 The preferred loop form for vectorization is the @code{for} iteration scheme.
21446 Loops with a @code{while} iteration scheme can also be vectorized if they are
21447 very simple, but the vectorizer will quickly give up otherwise. With either
21448 iteration scheme, the flow of control must be straight, in particular no
21449 @code{exit} statement may appear in the loop body. The loop may however
21450 contain a single nested loop, if it can be vectorized when considered alone:
21455 A : array (1..4, 1..4) of Long_Float;
21456 S : array (1..4) of Long_Float;
21460 for I in A'Range(1) loop
21461 for J in A'Range(2) loop
21462 S (I) := S (I) + A (I, J);
21469 The vectorizable operations depend on the targeted SIMD instruction set, but
21470 the adding and some of the multiplying operators are generally supported, as
21471 well as the logical operators for modular types. Note that compiling
21472 with @code{-gnatp} might well reveal cases where some checks do thwart
21475 Type conversions may also prevent vectorization if they involve semantics that
21476 are not directly supported by the code generator or the SIMD instruction set.
21477 A typical example is direct conversion from floating-point to integer types.
21478 The solution in this case is to use the following idiom:
21483 Integer (S'Truncation (F))
21487 if @code{S} is the subtype of floating-point object @code{F}.
21489 In most cases, the vectorizable loops are loops that iterate over arrays.
21490 All kinds of array types are supported, i.e. constrained array types with
21496 type Array_Type is array (1 .. 4) of Long_Float;
21500 constrained array types with dynamic bounds:
21505 type Array_Type is array (1 .. Q.N) of Long_Float;
21507 type Array_Type is array (Q.K .. 4) of Long_Float;
21509 type Array_Type is array (Q.K .. Q.N) of Long_Float;
21513 or unconstrained array types:
21518 type Array_Type is array (Positive range <>) of Long_Float;
21522 The quality of the generated code decreases when the dynamic aspect of the
21523 array type increases, the worst code being generated for unconstrained array
21524 types. This is so because, the less information the compiler has about the
21525 bounds of the array, the more fallback code it needs to generate in order to
21526 fix things up at run time.
21528 It is possible to specify that a given loop should be subject to vectorization
21529 preferably to other optimizations by means of pragma @code{Loop_Optimize}:
21534 pragma Loop_Optimize (Vector);
21538 placed immediately within the loop will convey the appropriate hint to the
21539 compiler for this loop.
21541 It is also possible to help the compiler generate better vectorized code
21542 for a given loop by asserting that there are no loop-carried dependencies
21543 in the loop. Consider for example the procedure:
21548 type Arr is array (1 .. 4) of Long_Float;
21550 procedure Add (X, Y : not null access Arr; R : not null access Arr) is
21552 for I in Arr'Range loop
21553 R(I) := X(I) + Y(I);
21559 By default, the compiler cannot unconditionally vectorize the loop because
21560 assigning to a component of the array designated by R in one iteration could
21561 change the value read from the components of the array designated by X or Y
21562 in a later iteration. As a result, the compiler will generate two versions
21563 of the loop in the object code, one vectorized and the other not vectorized,
21564 as well as a test to select the appropriate version at run time. This can
21565 be overcome by another hint:
21570 pragma Loop_Optimize (Ivdep);
21574 placed immediately within the loop will tell the compiler that it can safely
21575 omit the non-vectorized version of the loop as well as the run-time test.
21577 @node Other Optimization Switches,Optimization and Strict Aliasing,Vectorization of loops,Performance Considerations
21578 @anchor{gnat_ugn/gnat_and_program_execution other-optimization-switches}@anchor{1a8}@anchor{gnat_ugn/gnat_and_program_execution id35}@anchor{1a9}
21579 @subsubsection Other Optimization Switches
21582 @geindex Optimization Switches
21584 Since GNAT uses the @code{gcc} back end, all the specialized
21585 @code{gcc} optimization switches are potentially usable. These switches
21586 have not been extensively tested with GNAT but can generally be expected
21587 to work. Examples of switches in this category are @code{-funroll-loops}
21588 and the various target-specific @code{-m} options (in particular, it has
21589 been observed that @code{-march=xxx} can significantly improve performance
21590 on appropriate machines). For full details of these switches, see
21591 the @emph{Submodel Options} section in the @emph{Hardware Models and Configurations}
21592 chapter of @cite{Using the GNU Compiler Collection (GCC)}.
21594 @node Optimization and Strict Aliasing,Aliased Variables and Optimization,Other Optimization Switches,Performance Considerations
21595 @anchor{gnat_ugn/gnat_and_program_execution optimization-and-strict-aliasing}@anchor{f3}@anchor{gnat_ugn/gnat_and_program_execution id36}@anchor{1aa}
21596 @subsubsection Optimization and Strict Aliasing
21601 @geindex Strict Aliasing
21603 @geindex No_Strict_Aliasing
21605 The strong typing capabilities of Ada allow an optimizer to generate
21606 efficient code in situations where other languages would be forced to
21607 make worst case assumptions preventing such optimizations. Consider
21608 the following example:
21614 type Int1 is new Integer;
21615 type Int2 is new Integer;
21616 type Int1A is access Int1;
21617 type Int2A is access Int2;
21624 for J in Data'Range loop
21625 if Data (J) = Int1V.all then
21626 Int2V.all := Int2V.all + 1;
21634 In this example, since the variable @code{Int1V} can only access objects
21635 of type @code{Int1}, and @code{Int2V} can only access objects of type
21636 @code{Int2}, there is no possibility that the assignment to
21637 @code{Int2V.all} affects the value of @code{Int1V.all}. This means that
21638 the compiler optimizer can "know" that the value @code{Int1V.all} is constant
21639 for all iterations of the loop and avoid the extra memory reference
21640 required to dereference it each time through the loop.
21642 This kind of optimization, called strict aliasing analysis, is
21643 triggered by specifying an optimization level of @code{-O2} or
21644 higher or @code{-Os} and allows GNAT to generate more efficient code
21645 when access values are involved.
21647 However, although this optimization is always correct in terms of
21648 the formal semantics of the Ada Reference Manual, difficulties can
21649 arise if features like @code{Unchecked_Conversion} are used to break
21650 the typing system. Consider the following complete program example:
21656 type int1 is new integer;
21657 type int2 is new integer;
21658 type a1 is access int1;
21659 type a2 is access int2;
21664 function to_a2 (Input : a1) return a2;
21667 with Unchecked_Conversion;
21669 function to_a2 (Input : a1) return a2 is
21671 new Unchecked_Conversion (a1, a2);
21673 return to_a2u (Input);
21679 with Text_IO; use Text_IO;
21681 v1 : a1 := new int1;
21682 v2 : a2 := to_a2 (v1);
21686 put_line (int1'image (v1.all));
21691 This program prints out 0 in @code{-O0} or @code{-O1}
21692 mode, but it prints out 1 in @code{-O2} mode. That's
21693 because in strict aliasing mode, the compiler can and
21694 does assume that the assignment to @code{v2.all} could not
21695 affect the value of @code{v1.all}, since different types
21698 This behavior is not a case of non-conformance with the standard, since
21699 the Ada RM specifies that an unchecked conversion where the resulting
21700 bit pattern is not a correct value of the target type can result in an
21701 abnormal value and attempting to reference an abnormal value makes the
21702 execution of a program erroneous. That's the case here since the result
21703 does not point to an object of type @code{int2}. This means that the
21704 effect is entirely unpredictable.
21706 However, although that explanation may satisfy a language
21707 lawyer, in practice an applications programmer expects an
21708 unchecked conversion involving pointers to create true
21709 aliases and the behavior of printing 1 seems plain wrong.
21710 In this case, the strict aliasing optimization is unwelcome.
21712 Indeed the compiler recognizes this possibility, and the
21713 unchecked conversion generates a warning:
21718 p2.adb:5:07: warning: possible aliasing problem with type "a2"
21719 p2.adb:5:07: warning: use -fno-strict-aliasing switch for references
21720 p2.adb:5:07: warning: or use "pragma No_Strict_Aliasing (a2);"
21724 Unfortunately the problem is recognized when compiling the body of
21725 package @code{p2}, but the actual "bad" code is generated while
21726 compiling the body of @code{m} and this latter compilation does not see
21727 the suspicious @code{Unchecked_Conversion}.
21729 As implied by the warning message, there are approaches you can use to
21730 avoid the unwanted strict aliasing optimization in a case like this.
21732 One possibility is to simply avoid the use of @code{-O2}, but
21733 that is a bit drastic, since it throws away a number of useful
21734 optimizations that do not involve strict aliasing assumptions.
21736 A less drastic approach is to compile the program using the
21737 option @code{-fno-strict-aliasing}. Actually it is only the
21738 unit containing the dereferencing of the suspicious pointer
21739 that needs to be compiled. So in this case, if we compile
21740 unit @code{m} with this switch, then we get the expected
21741 value of zero printed. Analyzing which units might need
21742 the switch can be painful, so a more reasonable approach
21743 is to compile the entire program with options @code{-O2}
21744 and @code{-fno-strict-aliasing}. If the performance is
21745 satisfactory with this combination of options, then the
21746 advantage is that the entire issue of possible "wrong"
21747 optimization due to strict aliasing is avoided.
21749 To avoid the use of compiler switches, the configuration
21750 pragma @code{No_Strict_Aliasing} with no parameters may be
21751 used to specify that for all access types, the strict
21752 aliasing optimization should be suppressed.
21754 However, these approaches are still overkill, in that they causes
21755 all manipulations of all access values to be deoptimized. A more
21756 refined approach is to concentrate attention on the specific
21757 access type identified as problematic.
21759 First, if a careful analysis of uses of the pointer shows
21760 that there are no possible problematic references, then
21761 the warning can be suppressed by bracketing the
21762 instantiation of @code{Unchecked_Conversion} to turn
21768 pragma Warnings (Off);
21770 new Unchecked_Conversion (a1, a2);
21771 pragma Warnings (On);
21775 Of course that approach is not appropriate for this particular
21776 example, since indeed there is a problematic reference. In this
21777 case we can take one of two other approaches.
21779 The first possibility is to move the instantiation of unchecked
21780 conversion to the unit in which the type is declared. In
21781 this example, we would move the instantiation of
21782 @code{Unchecked_Conversion} from the body of package
21783 @code{p2} to the spec of package @code{p1}. Now the
21784 warning disappears. That's because any use of the
21785 access type knows there is a suspicious unchecked
21786 conversion, and the strict aliasing optimization
21787 is automatically suppressed for the type.
21789 If it is not practical to move the unchecked conversion to the same unit
21790 in which the destination access type is declared (perhaps because the
21791 source type is not visible in that unit), you may use pragma
21792 @code{No_Strict_Aliasing} for the type. This pragma must occur in the
21793 same declarative sequence as the declaration of the access type:
21798 type a2 is access int2;
21799 pragma No_Strict_Aliasing (a2);
21803 Here again, the compiler now knows that the strict aliasing optimization
21804 should be suppressed for any reference to type @code{a2} and the
21805 expected behavior is obtained.
21807 Finally, note that although the compiler can generate warnings for
21808 simple cases of unchecked conversions, there are tricker and more
21809 indirect ways of creating type incorrect aliases which the compiler
21810 cannot detect. Examples are the use of address overlays and unchecked
21811 conversions involving composite types containing access types as
21812 components. In such cases, no warnings are generated, but there can
21813 still be aliasing problems. One safe coding practice is to forbid the
21814 use of address clauses for type overlaying, and to allow unchecked
21815 conversion only for primitive types. This is not really a significant
21816 restriction since any possible desired effect can be achieved by
21817 unchecked conversion of access values.
21819 The aliasing analysis done in strict aliasing mode can certainly
21820 have significant benefits. We have seen cases of large scale
21821 application code where the time is increased by up to 5% by turning
21822 this optimization off. If you have code that includes significant
21823 usage of unchecked conversion, you might want to just stick with
21824 @code{-O1} and avoid the entire issue. If you get adequate
21825 performance at this level of optimization level, that's probably
21826 the safest approach. If tests show that you really need higher
21827 levels of optimization, then you can experiment with @code{-O2}
21828 and @code{-O2 -fno-strict-aliasing} to see how much effect this
21829 has on size and speed of the code. If you really need to use
21830 @code{-O2} with strict aliasing in effect, then you should
21831 review any uses of unchecked conversion of access types,
21832 particularly if you are getting the warnings described above.
21834 @node Aliased Variables and Optimization,Atomic Variables and Optimization,Optimization and Strict Aliasing,Performance Considerations
21835 @anchor{gnat_ugn/gnat_and_program_execution id37}@anchor{1ab}@anchor{gnat_ugn/gnat_and_program_execution aliased-variables-and-optimization}@anchor{1ac}
21836 @subsubsection Aliased Variables and Optimization
21841 There are scenarios in which programs may
21842 use low level techniques to modify variables
21843 that otherwise might be considered to be unassigned. For example,
21844 a variable can be passed to a procedure by reference, which takes
21845 the address of the parameter and uses the address to modify the
21846 variable's value, even though it is passed as an IN parameter.
21847 Consider the following example:
21853 Max_Length : constant Natural := 16;
21854 type Char_Ptr is access all Character;
21856 procedure Get_String(Buffer: Char_Ptr; Size : Integer);
21857 pragma Import (C, Get_String, "get_string");
21859 Name : aliased String (1 .. Max_Length) := (others => ' ');
21862 function Addr (S : String) return Char_Ptr is
21863 function To_Char_Ptr is
21864 new Ada.Unchecked_Conversion (System.Address, Char_Ptr);
21866 return To_Char_Ptr (S (S'First)'Address);
21870 Temp := Addr (Name);
21871 Get_String (Temp, Max_Length);
21876 where Get_String is a C function that uses the address in Temp to
21877 modify the variable @code{Name}. This code is dubious, and arguably
21878 erroneous, and the compiler would be entitled to assume that
21879 @code{Name} is never modified, and generate code accordingly.
21881 However, in practice, this would cause some existing code that
21882 seems to work with no optimization to start failing at high
21883 levels of optimzization.
21885 What the compiler does for such cases is to assume that marking
21886 a variable as aliased indicates that some "funny business" may
21887 be going on. The optimizer recognizes the aliased keyword and
21888 inhibits optimizations that assume the value cannot be assigned.
21889 This means that the above example will in fact "work" reliably,
21890 that is, it will produce the expected results.
21892 @node Atomic Variables and Optimization,Passive Task Optimization,Aliased Variables and Optimization,Performance Considerations
21893 @anchor{gnat_ugn/gnat_and_program_execution atomic-variables-and-optimization}@anchor{1ad}@anchor{gnat_ugn/gnat_and_program_execution id38}@anchor{1ae}
21894 @subsubsection Atomic Variables and Optimization
21899 There are two considerations with regard to performance when
21900 atomic variables are used.
21902 First, the RM only guarantees that access to atomic variables
21903 be atomic, it has nothing to say about how this is achieved,
21904 though there is a strong implication that this should not be
21905 achieved by explicit locking code. Indeed GNAT will never
21906 generate any locking code for atomic variable access (it will
21907 simply reject any attempt to make a variable or type atomic
21908 if the atomic access cannot be achieved without such locking code).
21910 That being said, it is important to understand that you cannot
21911 assume that the entire variable will always be accessed. Consider
21918 A,B,C,D : Character;
21921 for R'Alignment use 4;
21924 pragma Atomic (RV);
21931 You cannot assume that the reference to @code{RV.B}
21932 will read the entire 32-bit
21933 variable with a single load instruction. It is perfectly legitimate if
21934 the hardware allows it to do a byte read of just the B field. This read
21935 is still atomic, which is all the RM requires. GNAT can and does take
21936 advantage of this, depending on the architecture and optimization level.
21937 Any assumption to the contrary is non-portable and risky. Even if you
21938 examine the assembly language and see a full 32-bit load, this might
21939 change in a future version of the compiler.
21941 If your application requires that all accesses to @code{RV} in this
21942 example be full 32-bit loads, you need to make a copy for the access
21949 RV_Copy : constant R := RV;
21956 Now the reference to RV must read the whole variable.
21957 Actually one can imagine some compiler which figures
21958 out that the whole copy is not required (because only
21959 the B field is actually accessed), but GNAT
21960 certainly won't do that, and we don't know of any
21961 compiler that would not handle this right, and the
21962 above code will in practice work portably across
21963 all architectures (that permit the Atomic declaration).
21965 The second issue with atomic variables has to do with
21966 the possible requirement of generating synchronization
21967 code. For more details on this, consult the sections on
21968 the pragmas Enable/Disable_Atomic_Synchronization in the
21969 GNAT Reference Manual. If performance is critical, and
21970 such synchronization code is not required, it may be
21971 useful to disable it.
21973 @node Passive Task Optimization,,Atomic Variables and Optimization,Performance Considerations
21974 @anchor{gnat_ugn/gnat_and_program_execution passive-task-optimization}@anchor{1af}@anchor{gnat_ugn/gnat_and_program_execution id39}@anchor{1b0}
21975 @subsubsection Passive Task Optimization
21978 @geindex Passive Task
21980 A passive task is one which is sufficiently simple that
21981 in theory a compiler could recognize it an implement it
21982 efficiently without creating a new thread. The original design
21983 of Ada 83 had in mind this kind of passive task optimization, but
21984 only a few Ada 83 compilers attempted it. The problem was that
21985 it was difficult to determine the exact conditions under which
21986 the optimization was possible. The result is a very fragile
21987 optimization where a very minor change in the program can
21988 suddenly silently make a task non-optimizable.
21990 With the revisiting of this issue in Ada 95, there was general
21991 agreement that this approach was fundamentally flawed, and the
21992 notion of protected types was introduced. When using protected
21993 types, the restrictions are well defined, and you KNOW that the
21994 operations will be optimized, and furthermore this optimized
21995 performance is fully portable.
21997 Although it would theoretically be possible for GNAT to attempt to
21998 do this optimization, but it really doesn't make sense in the
21999 context of Ada 95, and none of the Ada 95 compilers implement
22000 this optimization as far as we know. In particular GNAT never
22001 attempts to perform this optimization.
22003 In any new Ada 95 code that is written, you should always
22004 use protected types in place of tasks that might be able to
22005 be optimized in this manner.
22006 Of course this does not help if you have legacy Ada 83 code
22007 that depends on this optimization, but it is unusual to encounter
22008 a case where the performance gains from this optimization
22011 Your program should work correctly without this optimization. If
22012 you have performance problems, then the most practical
22013 approach is to figure out exactly where these performance problems
22014 arise, and update those particular tasks to be protected types. Note
22015 that typically clients of the tasks who call entries, will not have
22016 to be modified, only the task definition itself.
22018 @node Text_IO Suggestions,Reducing Size of Executables with Unused Subprogram/Data Elimination,Performance Considerations,Improving Performance
22019 @anchor{gnat_ugn/gnat_and_program_execution text-io-suggestions}@anchor{1b1}@anchor{gnat_ugn/gnat_and_program_execution id40}@anchor{1b2}
22020 @subsection @code{Text_IO} Suggestions
22023 @geindex Text_IO and performance
22025 The @code{Ada.Text_IO} package has fairly high overheads due in part to
22026 the requirement of maintaining page and line counts. If performance
22027 is critical, a recommendation is to use @code{Stream_IO} instead of
22028 @code{Text_IO} for volume output, since this package has less overhead.
22030 If @code{Text_IO} must be used, note that by default output to the standard
22031 output and standard error files is unbuffered (this provides better
22032 behavior when output statements are used for debugging, or if the
22033 progress of a program is observed by tracking the output, e.g. by
22034 using the Unix @emph{tail -f} command to watch redirected output.
22036 If you are generating large volumes of output with @code{Text_IO} and
22037 performance is an important factor, use a designated file instead
22038 of the standard output file, or change the standard output file to
22039 be buffered using @code{Interfaces.C_Streams.setvbuf}.
22041 @node Reducing Size of Executables with Unused Subprogram/Data Elimination,,Text_IO Suggestions,Improving Performance
22042 @anchor{gnat_ugn/gnat_and_program_execution id41}@anchor{1b3}@anchor{gnat_ugn/gnat_and_program_execution reducing-size-of-executables-with-unused-subprogram-data-elimination}@anchor{1b4}
22043 @subsection Reducing Size of Executables with Unused Subprogram/Data Elimination
22046 @geindex Uunused subprogram/data elimination
22048 This section describes how you can eliminate unused subprograms and data from
22049 your executable just by setting options at compilation time.
22052 * About unused subprogram/data elimination::
22053 * Compilation options::
22054 * Example of unused subprogram/data elimination::
22058 @node About unused subprogram/data elimination,Compilation options,,Reducing Size of Executables with Unused Subprogram/Data Elimination
22059 @anchor{gnat_ugn/gnat_and_program_execution id42}@anchor{1b5}@anchor{gnat_ugn/gnat_and_program_execution about-unused-subprogram-data-elimination}@anchor{1b6}
22060 @subsubsection About unused subprogram/data elimination
22063 By default, an executable contains all code and data of its composing objects
22064 (directly linked or coming from statically linked libraries), even data or code
22065 never used by this executable.
22067 This feature will allow you to eliminate such unused code from your
22068 executable, making it smaller (in disk and in memory).
22070 This functionality is available on all Linux platforms except for the IA-64
22071 architecture and on all cross platforms using the ELF binary file format.
22072 In both cases GNU binutils version 2.16 or later are required to enable it.
22074 @node Compilation options,Example of unused subprogram/data elimination,About unused subprogram/data elimination,Reducing Size of Executables with Unused Subprogram/Data Elimination
22075 @anchor{gnat_ugn/gnat_and_program_execution id43}@anchor{1b7}@anchor{gnat_ugn/gnat_and_program_execution compilation-options}@anchor{1b8}
22076 @subsubsection Compilation options
22079 The operation of eliminating the unused code and data from the final executable
22080 is directly performed by the linker.
22082 @geindex -ffunction-sections (gcc)
22084 @geindex -fdata-sections (gcc)
22086 In order to do this, it has to work with objects compiled with the
22088 @code{-ffunction-sections} @code{-fdata-sections}.
22090 These options are usable with C and Ada files.
22091 They will place respectively each
22092 function or data in a separate section in the resulting object file.
22094 Once the objects and static libraries are created with these options, the
22095 linker can perform the dead code elimination. You can do this by setting
22096 the @code{-Wl,--gc-sections} option to gcc command or in the
22097 @code{-largs} section of @code{gnatmake}. This will perform a
22098 garbage collection of code and data never referenced.
22100 If the linker performs a partial link (@code{-r} linker option), then you
22101 will need to provide the entry point using the @code{-e} / @code{--entry}
22104 Note that objects compiled without the @code{-ffunction-sections} and
22105 @code{-fdata-sections} options can still be linked with the executable.
22106 However, no dead code elimination will be performed on those objects (they will
22109 The GNAT static library is now compiled with -ffunction-sections and
22110 -fdata-sections on some platforms. This allows you to eliminate the unused code
22111 and data of the GNAT library from your executable.
22113 @node Example of unused subprogram/data elimination,,Compilation options,Reducing Size of Executables with Unused Subprogram/Data Elimination
22114 @anchor{gnat_ugn/gnat_and_program_execution example-of-unused-subprogram-data-elimination}@anchor{1b9}@anchor{gnat_ugn/gnat_and_program_execution id44}@anchor{1ba}
22115 @subsubsection Example of unused subprogram/data elimination
22118 Here is a simple example:
22131 Used_Data : Integer;
22132 Unused_Data : Integer;
22134 procedure Used (Data : Integer);
22135 procedure Unused (Data : Integer);
22138 package body Aux is
22139 procedure Used (Data : Integer) is
22144 procedure Unused (Data : Integer) is
22146 Unused_Data := Data;
22152 @code{Unused} and @code{Unused_Data} are never referenced in this code
22153 excerpt, and hence they may be safely removed from the final executable.
22160 $ nm test | grep used
22161 020015f0 T aux__unused
22162 02005d88 B aux__unused_data
22163 020015cc T aux__used
22164 02005d84 B aux__used_data
22166 $ gnatmake test -cargs -fdata-sections -ffunction-sections \\
22167 -largs -Wl,--gc-sections
22169 $ nm test | grep used
22170 02005350 T aux__used
22171 0201ffe0 B aux__used_data
22175 It can be observed that the procedure @code{Unused} and the object
22176 @code{Unused_Data} are removed by the linker when using the
22177 appropriate options.
22179 @geindex Overflow checks
22181 @geindex Checks (overflow)
22184 @node Overflow Check Handling in GNAT,Performing Dimensionality Analysis in GNAT,Improving Performance,GNAT and Program Execution
22185 @anchor{gnat_ugn/gnat_and_program_execution id50}@anchor{169}@anchor{gnat_ugn/gnat_and_program_execution overflow-check-handling-in-gnat}@anchor{27}
22186 @section Overflow Check Handling in GNAT
22189 This section explains how to control the handling of overflow checks.
22193 * Management of Overflows in GNAT::
22194 * Specifying the Desired Mode::
22195 * Default Settings::
22196 * Implementation Notes::
22200 @node Background,Management of Overflows in GNAT,,Overflow Check Handling in GNAT
22201 @anchor{gnat_ugn/gnat_and_program_execution id51}@anchor{1bb}@anchor{gnat_ugn/gnat_and_program_execution background}@anchor{1bc}
22202 @subsection Background
22205 Overflow checks are checks that the compiler may make to ensure
22206 that intermediate results are not out of range. For example:
22217 If @code{A} has the value @code{Integer'Last}, then the addition may cause
22218 overflow since the result is out of range of the type @code{Integer}.
22219 In this case @code{Constraint_Error} will be raised if checks are
22222 A trickier situation arises in examples like the following:
22233 where @code{A} is @code{Integer'Last} and @code{C} is @code{-1}.
22234 Now the final result of the expression on the right hand side is
22235 @code{Integer'Last} which is in range, but the question arises whether the
22236 intermediate addition of @code{(A + 1)} raises an overflow error.
22238 The (perhaps surprising) answer is that the Ada language
22239 definition does not answer this question. Instead it leaves
22240 it up to the implementation to do one of two things if overflow
22241 checks are enabled.
22247 raise an exception (@code{Constraint_Error}), or
22250 yield the correct mathematical result which is then used in
22251 subsequent operations.
22254 If the compiler chooses the first approach, then the assignment of this
22255 example will indeed raise @code{Constraint_Error} if overflow checking is
22256 enabled, or result in erroneous execution if overflow checks are suppressed.
22258 But if the compiler
22259 chooses the second approach, then it can perform both additions yielding
22260 the correct mathematical result, which is in range, so no exception
22261 will be raised, and the right result is obtained, regardless of whether
22262 overflow checks are suppressed.
22264 Note that in the first example an
22265 exception will be raised in either case, since if the compiler
22266 gives the correct mathematical result for the addition, it will
22267 be out of range of the target type of the assignment, and thus
22268 fails the range check.
22270 This lack of specified behavior in the handling of overflow for
22271 intermediate results is a source of non-portability, and can thus
22272 be problematic when programs are ported. Most typically this arises
22273 in a situation where the original compiler did not raise an exception,
22274 and then the application is moved to a compiler where the check is
22275 performed on the intermediate result and an unexpected exception is
22278 Furthermore, when using Ada 2012's preconditions and other
22279 assertion forms, another issue arises. Consider:
22284 procedure P (A, B : Integer) with
22285 Pre => A + B <= Integer'Last;
22289 One often wants to regard arithmetic in a context like this from
22290 a mathematical point of view. So for example, if the two actual parameters
22291 for a call to @code{P} are both @code{Integer'Last}, then
22292 the precondition should be regarded as False. If we are executing
22293 in a mode with run-time checks enabled for preconditions, then we would
22294 like this precondition to fail, rather than raising an exception
22295 because of the intermediate overflow.
22297 However, the language definition leaves the specification of
22298 whether the above condition fails (raising @code{Assert_Error}) or
22299 causes an intermediate overflow (raising @code{Constraint_Error})
22300 up to the implementation.
22302 The situation is worse in a case such as the following:
22307 procedure Q (A, B, C : Integer) with
22308 Pre => A + B + C <= Integer'Last;
22317 Q (A => Integer'Last, B => 1, C => -1);
22321 From a mathematical point of view the precondition
22322 is True, but at run time we may (but are not guaranteed to) get an
22323 exception raised because of the intermediate overflow (and we really
22324 would prefer this precondition to be considered True at run time).
22326 @node Management of Overflows in GNAT,Specifying the Desired Mode,Background,Overflow Check Handling in GNAT
22327 @anchor{gnat_ugn/gnat_and_program_execution management-of-overflows-in-gnat}@anchor{1bd}@anchor{gnat_ugn/gnat_and_program_execution id52}@anchor{1be}
22328 @subsection Management of Overflows in GNAT
22331 To deal with the portability issue, and with the problem of
22332 mathematical versus run-time interpretation of the expressions in
22333 assertions, GNAT provides comprehensive control over the handling
22334 of intermediate overflow. GNAT can operate in three modes, and
22335 furthemore, permits separate selection of operating modes for
22336 the expressions within assertions (here the term 'assertions'
22337 is used in the technical sense, which includes preconditions and so forth)
22338 and for expressions appearing outside assertions.
22340 The three modes are:
22346 @emph{Use base type for intermediate operations} (@code{STRICT})
22348 In this mode, all intermediate results for predefined arithmetic
22349 operators are computed using the base type, and the result must
22350 be in range of the base type. If this is not the
22351 case then either an exception is raised (if overflow checks are
22352 enabled) or the execution is erroneous (if overflow checks are suppressed).
22353 This is the normal default mode.
22356 @emph{Most intermediate overflows avoided} (@code{MINIMIZED})
22358 In this mode, the compiler attempts to avoid intermediate overflows by
22359 using a larger integer type, typically @code{Long_Long_Integer},
22360 as the type in which arithmetic is
22361 performed for predefined arithmetic operators. This may be slightly more
22363 run time (compared to suppressing intermediate overflow checks), though
22364 the cost is negligible on modern 64-bit machines. For the examples given
22365 earlier, no intermediate overflows would have resulted in exceptions,
22366 since the intermediate results are all in the range of
22367 @code{Long_Long_Integer} (typically 64-bits on nearly all implementations
22368 of GNAT). In addition, if checks are enabled, this reduces the number of
22369 checks that must be made, so this choice may actually result in an
22370 improvement in space and time behavior.
22372 However, there are cases where @code{Long_Long_Integer} is not large
22373 enough, consider the following example:
22378 procedure R (A, B, C, D : Integer) with
22379 Pre => (A**2 * B**2) / (C**2 * D**2) <= 10;
22383 where @code{A} = @code{B} = @code{C} = @code{D} = @code{Integer'Last}.
22384 Now the intermediate results are
22385 out of the range of @code{Long_Long_Integer} even though the final result
22386 is in range and the precondition is True (from a mathematical point
22387 of view). In such a case, operating in this mode, an overflow occurs
22388 for the intermediate computation (which is why this mode
22389 says @emph{most} intermediate overflows are avoided). In this case,
22390 an exception is raised if overflow checks are enabled, and the
22391 execution is erroneous if overflow checks are suppressed.
22394 @emph{All intermediate overflows avoided} (@code{ELIMINATED})
22396 In this mode, the compiler avoids all intermediate overflows
22397 by using arbitrary precision arithmetic as required. In this
22398 mode, the above example with @code{A**2 * B**2} would
22399 not cause intermediate overflow, because the intermediate result
22400 would be evaluated using sufficient precision, and the result
22401 of evaluating the precondition would be True.
22403 This mode has the advantage of avoiding any intermediate
22404 overflows, but at the expense of significant run-time overhead,
22405 including the use of a library (included automatically in this
22406 mode) for multiple-precision arithmetic.
22408 This mode provides cleaner semantics for assertions, since now
22409 the run-time behavior emulates true arithmetic behavior for the
22410 predefined arithmetic operators, meaning that there is never a
22411 conflict between the mathematical view of the assertion, and its
22414 Note that in this mode, the behavior is unaffected by whether or
22415 not overflow checks are suppressed, since overflow does not occur.
22416 It is possible for gigantic intermediate expressions to raise
22417 @code{Storage_Error} as a result of attempting to compute the
22418 results of such expressions (e.g. @code{Integer'Last ** Integer'Last})
22419 but overflow is impossible.
22422 Note that these modes apply only to the evaluation of predefined
22423 arithmetic, membership, and comparison operators for signed integer
22426 For fixed-point arithmetic, checks can be suppressed. But if checks
22428 then fixed-point values are always checked for overflow against the
22429 base type for intermediate expressions (that is such checks always
22430 operate in the equivalent of @code{STRICT} mode).
22432 For floating-point, on nearly all architectures, @code{Machine_Overflows}
22433 is False, and IEEE infinities are generated, so overflow exceptions
22434 are never raised. If you want to avoid infinities, and check that
22435 final results of expressions are in range, then you can declare a
22436 constrained floating-point type, and range checks will be carried
22437 out in the normal manner (with infinite values always failing all
22440 @node Specifying the Desired Mode,Default Settings,Management of Overflows in GNAT,Overflow Check Handling in GNAT
22441 @anchor{gnat_ugn/gnat_and_program_execution specifying-the-desired-mode}@anchor{f8}@anchor{gnat_ugn/gnat_and_program_execution id53}@anchor{1bf}
22442 @subsection Specifying the Desired Mode
22445 @geindex pragma Overflow_Mode
22447 The desired mode of for handling intermediate overflow can be specified using
22448 either the @code{Overflow_Mode} pragma or an equivalent compiler switch.
22449 The pragma has the form
22454 pragma Overflow_Mode ([General =>] MODE [, [Assertions =>] MODE]);
22458 where @code{MODE} is one of
22464 @code{STRICT}: intermediate overflows checked (using base type)
22467 @code{MINIMIZED}: minimize intermediate overflows
22470 @code{ELIMINATED}: eliminate intermediate overflows
22473 The case is ignored, so @code{MINIMIZED}, @code{Minimized} and
22474 @code{minimized} all have the same effect.
22476 If only the @code{General} parameter is present, then the given @code{MODE} applies
22477 to expressions both within and outside assertions. If both arguments
22478 are present, then @code{General} applies to expressions outside assertions,
22479 and @code{Assertions} applies to expressions within assertions. For example:
22484 pragma Overflow_Mode
22485 (General => Minimized, Assertions => Eliminated);
22489 specifies that general expressions outside assertions be evaluated
22490 in 'minimize intermediate overflows' mode, and expressions within
22491 assertions be evaluated in 'eliminate intermediate overflows' mode.
22492 This is often a reasonable choice, avoiding excessive overhead
22493 outside assertions, but assuring a high degree of portability
22494 when importing code from another compiler, while incurring
22495 the extra overhead for assertion expressions to ensure that
22496 the behavior at run time matches the expected mathematical
22499 The @code{Overflow_Mode} pragma has the same scoping and placement
22500 rules as pragma @code{Suppress}, so it can occur either as a
22501 configuration pragma, specifying a default for the whole
22502 program, or in a declarative scope, where it applies to the
22503 remaining declarations and statements in that scope.
22505 Note that pragma @code{Overflow_Mode} does not affect whether
22506 overflow checks are enabled or suppressed. It only controls the
22507 method used to compute intermediate values. To control whether
22508 overflow checking is enabled or suppressed, use pragma @code{Suppress}
22509 or @code{Unsuppress} in the usual manner.
22511 @geindex -gnato? (gcc)
22513 @geindex -gnato?? (gcc)
22515 Additionally, a compiler switch @code{-gnato?} or @code{-gnato??}
22516 can be used to control the checking mode default (which can be subsequently
22517 overridden using pragmas).
22519 Here @code{?} is one of the digits @code{1} through @code{3}:
22524 @multitable {xxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
22531 use base type for intermediate operations (@code{STRICT})
22539 minimize intermediate overflows (@code{MINIMIZED})
22547 eliminate intermediate overflows (@code{ELIMINATED})
22553 As with the pragma, if only one digit appears then it applies to all
22554 cases; if two digits are given, then the first applies outside
22555 assertions, and the second within assertions. Thus the equivalent
22556 of the example pragma above would be
22559 If no digits follow the @code{-gnato}, then it is equivalent to
22561 causing all intermediate operations to be computed using the base
22562 type (@code{STRICT} mode).
22564 @node Default Settings,Implementation Notes,Specifying the Desired Mode,Overflow Check Handling in GNAT
22565 @anchor{gnat_ugn/gnat_and_program_execution id54}@anchor{1c0}@anchor{gnat_ugn/gnat_and_program_execution default-settings}@anchor{1c1}
22566 @subsection Default Settings
22569 The default mode for overflow checks is
22578 which causes all computations both inside and outside assertions to use
22581 This retains compatibility with previous versions of
22582 GNAT which suppressed overflow checks by default and always
22583 used the base type for computation of intermediate results.
22585 @c Sphinx allows no emphasis within :index: role. As a workaround we
22586 @c point the index to "switch" and use emphasis for "-gnato".
22589 @geindex -gnato (gcc)
22590 switch @code{-gnato} (with no digits following)
22600 which causes overflow checking of all intermediate overflows
22601 both inside and outside assertions against the base type.
22603 The pragma @code{Suppress (Overflow_Check)} disables overflow
22604 checking, but it has no effect on the method used for computing
22605 intermediate results.
22607 The pragma @code{Unsuppress (Overflow_Check)} enables overflow
22608 checking, but it has no effect on the method used for computing
22609 intermediate results.
22611 @node Implementation Notes,,Default Settings,Overflow Check Handling in GNAT
22612 @anchor{gnat_ugn/gnat_and_program_execution id55}@anchor{1c2}@anchor{gnat_ugn/gnat_and_program_execution implementation-notes}@anchor{1c3}
22613 @subsection Implementation Notes
22616 In practice on typical 64-bit machines, the @code{MINIMIZED} mode is
22617 reasonably efficient, and can be generally used. It also helps
22618 to ensure compatibility with code imported from some other
22621 Setting all intermediate overflows checking (@code{CHECKED} mode)
22622 makes sense if you want to
22623 make sure that your code is compatible with any other possible
22624 Ada implementation. This may be useful in ensuring portability
22625 for code that is to be exported to some other compiler than GNAT.
22627 The Ada standard allows the reassociation of expressions at
22628 the same precedence level if no parentheses are present. For
22629 example, @code{A+B+C} parses as though it were @code{(A+B)+C}, but
22630 the compiler can reintepret this as @code{A+(B+C)}, possibly
22631 introducing or eliminating an overflow exception. The GNAT
22632 compiler never takes advantage of this freedom, and the
22633 expression @code{A+B+C} will be evaluated as @code{(A+B)+C}.
22634 If you need the other order, you can write the parentheses
22635 explicitly @code{A+(B+C)} and GNAT will respect this order.
22637 The use of @code{ELIMINATED} mode will cause the compiler to
22638 automatically include an appropriate arbitrary precision
22639 integer arithmetic package. The compiler will make calls
22640 to this package, though only in cases where it cannot be
22641 sure that @code{Long_Long_Integer} is sufficient to guard against
22642 intermediate overflows. This package does not use dynamic
22643 alllocation, but it does use the secondary stack, so an
22644 appropriate secondary stack package must be present (this
22645 is always true for standard full Ada, but may require
22646 specific steps for restricted run times such as ZFP).
22648 Although @code{ELIMINATED} mode causes expressions to use arbitrary
22649 precision arithmetic, avoiding overflow, the final result
22650 must be in an appropriate range. This is true even if the
22651 final result is of type @code{[Long_[Long_]]Integer'Base}, which
22652 still has the same bounds as its associated constrained
22655 Currently, the @code{ELIMINATED} mode is only available on target
22656 platforms for which @code{Long_Long_Integer} is 64-bits (nearly all GNAT
22659 @node Performing Dimensionality Analysis in GNAT,Stack Related Facilities,Overflow Check Handling in GNAT,GNAT and Program Execution
22660 @anchor{gnat_ugn/gnat_and_program_execution id56}@anchor{16a}@anchor{gnat_ugn/gnat_and_program_execution performing-dimensionality-analysis-in-gnat}@anchor{28}
22661 @section Performing Dimensionality Analysis in GNAT
22664 @geindex Dimensionality analysis
22666 The GNAT compiler supports dimensionality checking. The user can
22667 specify physical units for objects, and the compiler will verify that uses
22668 of these objects are compatible with their dimensions, in a fashion that is
22669 familiar to engineering practice. The dimensions of algebraic expressions
22670 (including powers with static exponents) are computed from their constituents.
22672 @geindex Dimension_System aspect
22674 @geindex Dimension aspect
22676 This feature depends on Ada 2012 aspect specifications, and is available from
22677 version 7.0.1 of GNAT onwards.
22678 The GNAT-specific aspect @code{Dimension_System}
22679 allows you to define a system of units; the aspect @code{Dimension}
22680 then allows the user to declare dimensioned quantities within a given system.
22681 (These aspects are described in the @emph{Implementation Defined Aspects}
22682 chapter of the @emph{GNAT Reference Manual}).
22684 The major advantage of this model is that it does not require the declaration of
22685 multiple operators for all possible combinations of types: it is only necessary
22686 to use the proper subtypes in object declarations.
22688 @geindex System.Dim.Mks package (GNAT library)
22690 @geindex MKS_Type type
22692 The simplest way to impose dimensionality checking on a computation is to make
22693 use of one of the instantiations of the package @code{System.Dim.Generic_Mks}, which
22694 are part of the GNAT library. This generic package defines a floating-point
22695 type @code{MKS_Type}, for which a sequence of dimension names are specified,
22696 together with their conventional abbreviations. The following should be read
22697 together with the full specification of the package, in file
22698 @code{s-digemk.ads}.
22702 @geindex s-digemk.ads file
22705 type Mks_Type is new Float_Type
22707 Dimension_System => (
22708 (Unit_Name => Meter, Unit_Symbol => 'm', Dim_Symbol => 'L'),
22709 (Unit_Name => Kilogram, Unit_Symbol => "kg", Dim_Symbol => 'M'),
22710 (Unit_Name => Second, Unit_Symbol => 's', Dim_Symbol => 'T'),
22711 (Unit_Name => Ampere, Unit_Symbol => 'A', Dim_Symbol => 'I'),
22712 (Unit_Name => Kelvin, Unit_Symbol => 'K', Dim_Symbol => "Theta"),
22713 (Unit_Name => Mole, Unit_Symbol => "mol", Dim_Symbol => 'N'),
22714 (Unit_Name => Candela, Unit_Symbol => "cd", Dim_Symbol => 'J'));
22718 The package then defines a series of subtypes that correspond to these
22719 conventional units. For example:
22724 subtype Length is Mks_Type
22726 Dimension => (Symbol => 'm', Meter => 1, others => 0);
22730 and similarly for @code{Mass}, @code{Time}, @code{Electric_Current},
22731 @code{Thermodynamic_Temperature}, @code{Amount_Of_Substance}, and
22732 @code{Luminous_Intensity} (the standard set of units of the SI system).
22734 The package also defines conventional names for values of each unit, for
22740 m : constant Length := 1.0;
22741 kg : constant Mass := 1.0;
22742 s : constant Time := 1.0;
22743 A : constant Electric_Current := 1.0;
22747 as well as useful multiples of these units:
22752 cm : constant Length := 1.0E-02;
22753 g : constant Mass := 1.0E-03;
22754 min : constant Time := 60.0;
22755 day : constant Time := 60.0 * 24.0 * min;
22760 There are three instantiations of @code{System.Dim.Generic_Mks} defined in the
22767 @code{System.Dim.Float_Mks} based on @code{Float} defined in @code{s-diflmk.ads}.
22770 @code{System.Dim.Long_Mks} based on @code{Long_Float} defined in @code{s-dilomk.ads}.
22773 @code{System.Dim.Mks} based on @code{Long_Long_Float} defined in @code{s-dimmks.ads}.
22776 Using one of these packages, you can then define a derived unit by providing
22777 the aspect that specifies its dimensions within the MKS system, as well as the
22778 string to be used for output of a value of that unit:
22783 subtype Acceleration is Mks_Type
22784 with Dimension => ("m/sec^2",
22791 Here is a complete example of use:
22796 with System.Dim.MKS; use System.Dim.Mks;
22797 with System.Dim.Mks_IO; use System.Dim.Mks_IO;
22798 with Text_IO; use Text_IO;
22799 procedure Free_Fall is
22800 subtype Acceleration is Mks_Type
22801 with Dimension => ("m/sec^2", 1, 0, -2, others => 0);
22802 G : constant acceleration := 9.81 * m / (s ** 2);
22803 T : Time := 10.0*s;
22807 Put ("Gravitational constant: ");
22808 Put (G, Aft => 2, Exp => 0); Put_Line ("");
22809 Distance := 0.5 * G * T ** 2;
22810 Put ("distance travelled in 10 seconds of free fall ");
22811 Put (Distance, Aft => 2, Exp => 0);
22817 Execution of this program yields:
22822 Gravitational constant: 9.81 m/sec^2
22823 distance travelled in 10 seconds of free fall 490.50 m
22827 However, incorrect assignments such as:
22833 Distance := 5.0 * kg;
22837 are rejected with the following diagnoses:
22843 >>> dimensions mismatch in assignment
22844 >>> left-hand side has dimension [L]
22845 >>> right-hand side is dimensionless
22847 Distance := 5.0 * kg:
22848 >>> dimensions mismatch in assignment
22849 >>> left-hand side has dimension [L]
22850 >>> right-hand side has dimension [M]
22854 The dimensions of an expression are properly displayed, even if there is
22855 no explicit subtype for it. If we add to the program:
22860 Put ("Final velocity: ");
22861 Put (G * T, Aft =>2, Exp =>0);
22866 then the output includes:
22871 Final velocity: 98.10 m.s**(-1)
22874 @geindex Dimensionable type
22876 @geindex Dimensioned subtype
22879 The type @code{Mks_Type} is said to be a @emph{dimensionable type} since it has a
22880 @code{Dimension_System} aspect, and the subtypes @code{Length}, @code{Mass}, etc.,
22881 are said to be @emph{dimensioned subtypes} since each one has a @code{Dimension}
22886 @geindex Dimension Vector (for a dimensioned subtype)
22888 @geindex Dimension aspect
22890 @geindex Dimension_System aspect
22893 The @code{Dimension} aspect of a dimensioned subtype @code{S} defines a mapping
22894 from the base type's Unit_Names to integer (or, more generally, rational)
22895 values. This mapping is the @emph{dimension vector} (also referred to as the
22896 @emph{dimensionality}) for that subtype, denoted by @code{DV(S)}, and thus for each
22897 object of that subtype. Intuitively, the value specified for each
22898 @code{Unit_Name} is the exponent associated with that unit; a zero value
22899 means that the unit is not used. For example:
22905 Acc : Acceleration;
22913 Here @code{DV(Acc)} = @code{DV(Acceleration)} =
22914 @code{(Meter=>1, Kilogram=>0, Second=>-2, Ampere=>0, Kelvin=>0, Mole=>0, Candela=>0)}.
22915 Symbolically, we can express this as @code{Meter / Second**2}.
22917 The dimension vector of an arithmetic expression is synthesized from the
22918 dimension vectors of its components, with compile-time dimensionality checks
22919 that help prevent mismatches such as using an @code{Acceleration} where a
22920 @code{Length} is required.
22922 The dimension vector of the result of an arithmetic expression @emph{expr}, or
22923 @code{DV(@emph{expr})}, is defined as follows, assuming conventional
22924 mathematical definitions for the vector operations that are used:
22930 If @emph{expr} is of the type @emph{universal_real}, or is not of a dimensioned subtype,
22931 then @emph{expr} is dimensionless; @code{DV(@emph{expr})} is the empty vector.
22934 @code{DV(@emph{op expr})}, where @emph{op} is a unary operator, is @code{DV(@emph{expr})}
22937 @code{DV(@emph{expr1 op expr2})} where @emph{op} is "+" or "-" is @code{DV(@emph{expr1})}
22938 provided that @code{DV(@emph{expr1})} = @code{DV(@emph{expr2})}.
22939 If this condition is not met then the construct is illegal.
22942 @code{DV(@emph{expr1} * @emph{expr2})} is @code{DV(@emph{expr1})} + @code{DV(@emph{expr2})},
22943 and @code{DV(@emph{expr1} / @emph{expr2})} = @code{DV(@emph{expr1})} - @code{DV(@emph{expr2})}.
22944 In this context if one of the @emph{expr}s is dimensionless then its empty
22945 dimension vector is treated as @code{(others => 0)}.
22948 @code{DV(@emph{expr} ** @emph{power})} is @emph{power} * @code{DV(@emph{expr})},
22949 provided that @emph{power} is a static rational value. If this condition is not
22950 met then the construct is illegal.
22953 Note that, by the above rules, it is illegal to use binary "+" or "-" to
22954 combine a dimensioned and dimensionless value. Thus an expression such as
22955 @code{acc-10.0} is illegal, where @code{acc} is an object of subtype
22956 @code{Acceleration}.
22958 The dimensionality checks for relationals use the same rules as
22959 for "+" and "-", except when comparing to a literal; thus
22977 and is thus illegal, but
22986 is accepted with a warning. Analogously a conditional expression requires the
22987 same dimension vector for each branch (with no exception for literals).
22989 The dimension vector of a type conversion @code{T(@emph{expr})} is defined
22990 as follows, based on the nature of @code{T}:
22996 If @code{T} is a dimensioned subtype then @code{DV(T(@emph{expr}))} is @code{DV(T)}
22997 provided that either @emph{expr} is dimensionless or
22998 @code{DV(T)} = @code{DV(@emph{expr})}. The conversion is illegal
22999 if @emph{expr} is dimensioned and @code{DV(@emph{expr})} /= @code{DV(T)}.
23000 Note that vector equality does not require that the corresponding
23001 Unit_Names be the same.
23003 As a consequence of the above rule, it is possible to convert between
23004 different dimension systems that follow the same international system
23005 of units, with the seven physical components given in the standard order
23006 (length, mass, time, etc.). Thus a length in meters can be converted to
23007 a length in inches (with a suitable conversion factor) but cannot be
23008 converted, for example, to a mass in pounds.
23011 If @code{T} is the base type for @emph{expr} (and the dimensionless root type of
23012 the dimension system), then @code{DV(T(@emph{expr}))} is @code{DV(expr)}.
23013 Thus, if @emph{expr} is of a dimensioned subtype of @code{T}, the conversion may
23014 be regarded as a "view conversion" that preserves dimensionality.
23016 This rule makes it possible to write generic code that can be instantiated
23017 with compatible dimensioned subtypes. The generic unit will contain
23018 conversions that will consequently be present in instantiations, but
23019 conversions to the base type will preserve dimensionality and make it
23020 possible to write generic code that is correct with respect to
23024 Otherwise (i.e., @code{T} is neither a dimensioned subtype nor a dimensionable
23025 base type), @code{DV(T(@emph{expr}))} is the empty vector. Thus a dimensioned
23026 value can be explicitly converted to a non-dimensioned subtype, which
23027 of course then escapes dimensionality analysis.
23030 The dimension vector for a type qualification @code{T'(@emph{expr})} is the same
23031 as for the type conversion @code{T(@emph{expr})}.
23033 An assignment statement
23042 requires @code{DV(Source)} = @code{DV(Target)}, and analogously for parameter
23043 passing (the dimension vector for the actual parameter must be equal to the
23044 dimension vector for the formal parameter).
23046 @node Stack Related Facilities,Memory Management Issues,Performing Dimensionality Analysis in GNAT,GNAT and Program Execution
23047 @anchor{gnat_ugn/gnat_and_program_execution id57}@anchor{16b}@anchor{gnat_ugn/gnat_and_program_execution stack-related-facilities}@anchor{29}
23048 @section Stack Related Facilities
23051 This section describes some useful tools associated with stack
23052 checking and analysis. In
23053 particular, it deals with dynamic and static stack usage measurements.
23056 * Stack Overflow Checking::
23057 * Static Stack Usage Analysis::
23058 * Dynamic Stack Usage Analysis::
23062 @node Stack Overflow Checking,Static Stack Usage Analysis,,Stack Related Facilities
23063 @anchor{gnat_ugn/gnat_and_program_execution id58}@anchor{1c4}@anchor{gnat_ugn/gnat_and_program_execution stack-overflow-checking}@anchor{f4}
23064 @subsection Stack Overflow Checking
23067 @geindex Stack Overflow Checking
23069 @geindex -fstack-check (gcc)
23071 For most operating systems, @code{gcc} does not perform stack overflow
23072 checking by default. This means that if the main environment task or
23073 some other task exceeds the available stack space, then unpredictable
23074 behavior will occur. Most native systems offer some level of protection by
23075 adding a guard page at the end of each task stack. This mechanism is usually
23076 not enough for dealing properly with stack overflow situations because
23077 a large local variable could "jump" above the guard page.
23078 Furthermore, when the
23079 guard page is hit, there may not be any space left on the stack for executing
23080 the exception propagation code. Enabling stack checking avoids
23083 To activate stack checking, compile all units with the @code{gcc} option
23084 @code{-fstack-check}. For example:
23089 $ gcc -c -fstack-check package1.adb
23093 Units compiled with this option will generate extra instructions to check
23094 that any use of the stack (for procedure calls or for declaring local
23095 variables in declare blocks) does not exceed the available stack space.
23096 If the space is exceeded, then a @code{Storage_Error} exception is raised.
23098 For declared tasks, the default stack size is defined by the GNAT runtime,
23099 whose size may be modified at bind time through the @code{-d} bind switch
23100 (@ref{11f,,Switches for gnatbind}). Task specific stack sizes may be set using the
23101 @code{Storage_Size} pragma.
23103 For the environment task, the stack size is determined by the operating system.
23104 Consequently, to modify the size of the environment task please refer to your
23105 operating system documentation.
23107 @node Static Stack Usage Analysis,Dynamic Stack Usage Analysis,Stack Overflow Checking,Stack Related Facilities
23108 @anchor{gnat_ugn/gnat_and_program_execution static-stack-usage-analysis}@anchor{f5}@anchor{gnat_ugn/gnat_and_program_execution id59}@anchor{1c5}
23109 @subsection Static Stack Usage Analysis
23112 @geindex Static Stack Usage Analysis
23114 @geindex -fstack-usage
23116 A unit compiled with @code{-fstack-usage} will generate an extra file
23118 the maximum amount of stack used, on a per-function basis.
23119 The file has the same
23120 basename as the target object file with a @code{.su} extension.
23121 Each line of this file is made up of three fields:
23127 The name of the function.
23133 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
23136 The second field corresponds to the size of the known part of the function
23139 The qualifier @code{static} means that the function frame size
23141 It usually means that all local variables have a static size.
23142 In this case, the second field is a reliable measure of the function stack
23145 The qualifier @code{dynamic} means that the function frame size is not static.
23146 It happens mainly when some local variables have a dynamic size. When this
23147 qualifier appears alone, the second field is not a reliable measure
23148 of the function stack analysis. When it is qualified with @code{bounded}, it
23149 means that the second field is a reliable maximum of the function stack
23152 A unit compiled with @code{-Wstack-usage} will issue a warning for each
23153 subprogram whose stack usage might be larger than the specified amount of
23154 bytes. The wording is in keeping with the qualifier documented above.
23156 @node Dynamic Stack Usage Analysis,,Static Stack Usage Analysis,Stack Related Facilities
23157 @anchor{gnat_ugn/gnat_and_program_execution dynamic-stack-usage-analysis}@anchor{121}@anchor{gnat_ugn/gnat_and_program_execution id60}@anchor{1c6}
23158 @subsection Dynamic Stack Usage Analysis
23161 It is possible to measure the maximum amount of stack used by a task, by
23162 adding a switch to @code{gnatbind}, as:
23167 $ gnatbind -u0 file
23171 With this option, at each task termination, its stack usage is output on
23173 It is not always convenient to output the stack usage when the program
23174 is still running. Hence, it is possible to delay this output until program
23175 termination. for a given number of tasks specified as the argument of the
23176 @code{-u} option. For instance:
23181 $ gnatbind -u100 file
23185 will buffer the stack usage information of the first 100 tasks to terminate and
23186 output this info at program termination. Results are displayed in four
23192 Index | Task Name | Stack Size | Stack Usage
23202 @emph{Index} is a number associated with each task.
23205 @emph{Task Name} is the name of the task analyzed.
23208 @emph{Stack Size} is the maximum size for the stack.
23211 @emph{Stack Usage} is the measure done by the stack analyzer.
23212 In order to prevent overflow, the stack
23213 is not entirely analyzed, and it's not possible to know exactly how
23214 much has actually been used.
23217 By default the environment task stack, the stack that contains the main unit,
23218 is not processed. To enable processing of the environment task stack, the
23219 environment variable GNAT_STACK_LIMIT needs to be set to the maximum size of
23220 the environment task stack. This amount is given in kilobytes. For example:
23225 $ set GNAT_STACK_LIMIT 1600
23229 would specify to the analyzer that the environment task stack has a limit
23230 of 1.6 megabytes. Any stack usage beyond this will be ignored by the analysis.
23232 The package @code{GNAT.Task_Stack_Usage} provides facilities to get
23233 stack-usage reports at run time. See its body for the details.
23235 @node Memory Management Issues,,Stack Related Facilities,GNAT and Program Execution
23236 @anchor{gnat_ugn/gnat_and_program_execution id61}@anchor{16c}@anchor{gnat_ugn/gnat_and_program_execution memory-management-issues}@anchor{2a}
23237 @section Memory Management Issues
23240 This section describes some useful memory pools provided in the GNAT library
23241 and in particular the GNAT Debug Pool facility, which can be used to detect
23242 incorrect uses of access values (including 'dangling references').
23246 * Some Useful Memory Pools::
23247 * The GNAT Debug Pool Facility::
23251 @node Some Useful Memory Pools,The GNAT Debug Pool Facility,,Memory Management Issues
23252 @anchor{gnat_ugn/gnat_and_program_execution id62}@anchor{1c7}@anchor{gnat_ugn/gnat_and_program_execution some-useful-memory-pools}@anchor{1c8}
23253 @subsection Some Useful Memory Pools
23256 @geindex Memory Pool
23261 The @code{System.Pool_Global} package offers the Unbounded_No_Reclaim_Pool
23262 storage pool. Allocations use the standard system call @code{malloc} while
23263 deallocations use the standard system call @code{free}. No reclamation is
23264 performed when the pool goes out of scope. For performance reasons, the
23265 standard default Ada allocators/deallocators do not use any explicit storage
23266 pools but if they did, they could use this storage pool without any change in
23267 behavior. That is why this storage pool is used when the user
23268 manages to make the default implicit allocator explicit as in this example:
23273 type T1 is access Something;
23274 -- no Storage pool is defined for T2
23276 type T2 is access Something_Else;
23277 for T2'Storage_Pool use T1'Storage_Pool;
23278 -- the above is equivalent to
23279 for T2'Storage_Pool use System.Pool_Global.Global_Pool_Object;
23283 The @code{System.Pool_Local} package offers the @code{Unbounded_Reclaim_Pool} storage
23284 pool. The allocation strategy is similar to @code{Pool_Local}
23285 except that the all
23286 storage allocated with this pool is reclaimed when the pool object goes out of
23287 scope. This pool provides a explicit mechanism similar to the implicit one
23288 provided by several Ada 83 compilers for allocations performed through a local
23289 access type and whose purpose was to reclaim memory when exiting the
23290 scope of a given local access. As an example, the following program does not
23291 leak memory even though it does not perform explicit deallocation:
23296 with System.Pool_Local;
23297 procedure Pooloc1 is
23298 procedure Internal is
23299 type A is access Integer;
23300 X : System.Pool_Local.Unbounded_Reclaim_Pool;
23301 for A'Storage_Pool use X;
23304 for I in 1 .. 50 loop
23309 for I in 1 .. 100 loop
23316 The @code{System.Pool_Size} package implements the @code{Stack_Bounded_Pool} used when
23317 @code{Storage_Size} is specified for an access type.
23318 The whole storage for the pool is
23319 allocated at once, usually on the stack at the point where the access type is
23320 elaborated. It is automatically reclaimed when exiting the scope where the
23321 access type is defined. This package is not intended to be used directly by the
23322 user and it is implicitly used for each such declaration:
23327 type T1 is access Something;
23328 for T1'Storage_Size use 10_000;
23332 @node The GNAT Debug Pool Facility,,Some Useful Memory Pools,Memory Management Issues
23333 @anchor{gnat_ugn/gnat_and_program_execution id63}@anchor{1c9}@anchor{gnat_ugn/gnat_and_program_execution the-gnat-debug-pool-facility}@anchor{1ca}
23334 @subsection The GNAT Debug Pool Facility
23337 @geindex Debug Pool
23341 @geindex memory corruption
23343 The use of unchecked deallocation and unchecked conversion can easily
23344 lead to incorrect memory references. The problems generated by such
23345 references are usually difficult to tackle because the symptoms can be
23346 very remote from the origin of the problem. In such cases, it is
23347 very helpful to detect the problem as early as possible. This is the
23348 purpose of the Storage Pool provided by @code{GNAT.Debug_Pools}.
23350 In order to use the GNAT specific debugging pool, the user must
23351 associate a debug pool object with each of the access types that may be
23352 related to suspected memory problems. See Ada Reference Manual 13.11.
23357 type Ptr is access Some_Type;
23358 Pool : GNAT.Debug_Pools.Debug_Pool;
23359 for Ptr'Storage_Pool use Pool;
23363 @code{GNAT.Debug_Pools} is derived from a GNAT-specific kind of
23364 pool: the @code{Checked_Pool}. Such pools, like standard Ada storage pools,
23365 allow the user to redefine allocation and deallocation strategies. They
23366 also provide a checkpoint for each dereference, through the use of
23367 the primitive operation @code{Dereference} which is implicitly called at
23368 each dereference of an access value.
23370 Once an access type has been associated with a debug pool, operations on
23371 values of the type may raise four distinct exceptions,
23372 which correspond to four potential kinds of memory corruption:
23378 @code{GNAT.Debug_Pools.Accessing_Not_Allocated_Storage}
23381 @code{GNAT.Debug_Pools.Accessing_Deallocated_Storage}
23384 @code{GNAT.Debug_Pools.Freeing_Not_Allocated_Storage}
23387 @code{GNAT.Debug_Pools.Freeing_Deallocated_Storage}
23390 For types associated with a Debug_Pool, dynamic allocation is performed using
23391 the standard GNAT allocation routine. References to all allocated chunks of
23392 memory are kept in an internal dictionary. Several deallocation strategies are
23393 provided, whereupon the user can choose to release the memory to the system,
23394 keep it allocated for further invalid access checks, or fill it with an easily
23395 recognizable pattern for debug sessions. The memory pattern is the old IBM
23396 hexadecimal convention: @code{16#DEADBEEF#}.
23398 See the documentation in the file g-debpoo.ads for more information on the
23399 various strategies.
23401 Upon each dereference, a check is made that the access value denotes a
23402 properly allocated memory location. Here is a complete example of use of
23403 @code{Debug_Pools}, that includes typical instances of memory corruption:
23408 with Gnat.Io; use Gnat.Io;
23409 with Unchecked_Deallocation;
23410 with Unchecked_Conversion;
23411 with GNAT.Debug_Pools;
23412 with System.Storage_Elements;
23413 with Ada.Exceptions; use Ada.Exceptions;
23414 procedure Debug_Pool_Test is
23416 type T is access Integer;
23417 type U is access all T;
23419 P : GNAT.Debug_Pools.Debug_Pool;
23420 for T'Storage_Pool use P;
23422 procedure Free is new Unchecked_Deallocation (Integer, T);
23423 function UC is new Unchecked_Conversion (U, T);
23426 procedure Info is new GNAT.Debug_Pools.Print_Info(Put_Line);
23436 Put_Line (Integer'Image(B.all));
23438 when E : others => Put_Line ("raised: " & Exception_Name (E));
23443 when E : others => Put_Line ("raised: " & Exception_Name (E));
23447 Put_Line (Integer'Image(B.all));
23449 when E : others => Put_Line ("raised: " & Exception_Name (E));
23454 when E : others => Put_Line ("raised: " & Exception_Name (E));
23457 end Debug_Pool_Test;
23461 The debug pool mechanism provides the following precise diagnostics on the
23462 execution of this erroneous program:
23468 Total allocated bytes : 0
23469 Total deallocated bytes : 0
23470 Current Water Mark: 0
23474 Total allocated bytes : 8
23475 Total deallocated bytes : 0
23476 Current Water Mark: 8
23479 raised: GNAT.DEBUG_POOLS.ACCESSING_DEALLOCATED_STORAGE
23480 raised: GNAT.DEBUG_POOLS.FREEING_DEALLOCATED_STORAGE
23481 raised: GNAT.DEBUG_POOLS.ACCESSING_NOT_ALLOCATED_STORAGE
23482 raised: GNAT.DEBUG_POOLS.FREEING_NOT_ALLOCATED_STORAGE
23484 Total allocated bytes : 8
23485 Total deallocated bytes : 4
23486 Current Water Mark: 4
23492 @c -- Non-breaking space in running text
23493 @c -- E.g. Ada |nbsp| 95
23495 @node Platform-Specific Information,Example of Binder Output File,GNAT and Program Execution,Top
23496 @anchor{gnat_ugn/platform_specific_information platform-specific-information}@anchor{d}@anchor{gnat_ugn/platform_specific_information doc}@anchor{1cb}@anchor{gnat_ugn/platform_specific_information id1}@anchor{1cc}
23497 @chapter Platform-Specific Information
23500 This appendix contains information relating to the implementation
23501 of run-time libraries on various platforms and also covers
23502 topics related to the GNAT implementation on Windows and Mac OS.
23505 * Run-Time Libraries::
23506 * Specifying a Run-Time Library::
23507 * GNU/Linux Topics::
23508 * Microsoft Windows Topics::
23513 @node Run-Time Libraries,Specifying a Run-Time Library,,Platform-Specific Information
23514 @anchor{gnat_ugn/platform_specific_information id2}@anchor{1cd}@anchor{gnat_ugn/platform_specific_information run-time-libraries}@anchor{2b}
23515 @section Run-Time Libraries
23518 @geindex Tasking and threads libraries
23520 @geindex Threads libraries and tasking
23522 @geindex Run-time libraries (platform-specific information)
23524 The GNAT run-time implementation may vary with respect to both the
23525 underlying threads library and the exception-handling scheme.
23526 For threads support, the default run-time will bind to the thread
23527 package of the underlying operating system.
23529 For exception handling, either or both of two models are supplied:
23533 @geindex Zero-Cost Exceptions
23535 @geindex ZCX (Zero-Cost Exceptions)
23542 @strong{Zero-Cost Exceptions} ("ZCX"),
23543 which uses binder-generated tables that
23544 are interrogated at run time to locate a handler.
23546 @geindex setjmp/longjmp Exception Model
23548 @geindex SJLJ (setjmp/longjmp Exception Model)
23551 @strong{setjmp / longjmp} ('SJLJ'),
23552 which uses dynamically-set data to establish
23553 the set of handlers
23556 Most programs should experience a substantial speed improvement by
23557 being compiled with a ZCX run-time.
23558 This is especially true for
23559 tasking applications or applications with many exception handlers.
23560 Note however that the ZCX run-time does not support asynchronous abort
23561 of tasks (@code{abort} and @code{select-then-abort} constructs) and will instead
23562 implement abort by polling points in the runtime. You can also add additional
23563 polling points explicitly if needed in your application via @code{pragma
23566 This section summarizes which combinations of threads and exception support
23567 are supplied on various GNAT platforms.
23570 * Summary of Run-Time Configurations::
23574 @node Summary of Run-Time Configurations,,,Run-Time Libraries
23575 @anchor{gnat_ugn/platform_specific_information summary-of-run-time-configurations}@anchor{1ce}@anchor{gnat_ugn/platform_specific_information id3}@anchor{1cf}
23576 @subsection Summary of Run-Time Configurations
23580 @multitable {xxxxxxxxxxxxxxxxxxx} {xxxxxxxxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxx} {xxxxxxxxxxxxxx}
23637 native Win32 threads
23649 native Win32 threads
23674 @node Specifying a Run-Time Library,GNU/Linux Topics,Run-Time Libraries,Platform-Specific Information
23675 @anchor{gnat_ugn/platform_specific_information specifying-a-run-time-library}@anchor{1d0}@anchor{gnat_ugn/platform_specific_information id4}@anchor{1d1}
23676 @section Specifying a Run-Time Library
23679 The @code{adainclude} subdirectory containing the sources of the GNAT
23680 run-time library, and the @code{adalib} subdirectory containing the
23681 @code{ALI} files and the static and/or shared GNAT library, are located
23682 in the gcc target-dependent area:
23687 target=$prefix/lib/gcc/gcc-*dumpmachine*/gcc-*dumpversion*/
23691 As indicated above, on some platforms several run-time libraries are supplied.
23692 These libraries are installed in the target dependent area and
23693 contain a complete source and binary subdirectory. The detailed description
23694 below explains the differences between the different libraries in terms of
23695 their thread support.
23697 The default run-time library (when GNAT is installed) is @emph{rts-native}.
23698 This default run-time is selected by the means of soft links.
23699 For example on x86-linux:
23702 @c -- $(target-dir)
23704 @c -- +--- adainclude----------+
23706 @c -- +--- adalib-----------+ |
23708 @c -- +--- rts-native | |
23710 @c -- | +--- adainclude <---+
23712 @c -- | +--- adalib <----+
23714 @c -- +--- rts-sjlj
23716 @c -- +--- adainclude
23724 _______/ / \ \_________________
23727 ADAINCLUDE ADALIB rts-native rts-sjlj
23732 +-------------> adainclude adalib adainclude adalib
23735 +---------------------+
23737 Run-Time Library Directory Structure
23738 (Upper-case names and dotted/dashed arrows represent soft links)
23741 If the @emph{rts-sjlj} library is to be selected on a permanent basis,
23742 these soft links can be modified with the following commands:
23748 $ rm -f adainclude adalib
23749 $ ln -s rts-sjlj/adainclude adainclude
23750 $ ln -s rts-sjlj/adalib adalib
23754 Alternatively, you can specify @code{rts-sjlj/adainclude} in the file
23755 @code{$target/ada_source_path} and @code{rts-sjlj/adalib} in
23756 @code{$target/ada_object_path}.
23758 @geindex --RTS option
23760 Selecting another run-time library temporarily can be
23761 achieved by using the @code{--RTS} switch, e.g., @code{--RTS=sjlj}
23762 @anchor{gnat_ugn/platform_specific_information choosing-the-scheduling-policy}@anchor{1d2}
23763 @geindex SCHED_FIFO scheduling policy
23765 @geindex SCHED_RR scheduling policy
23767 @geindex SCHED_OTHER scheduling policy
23770 * Choosing the Scheduling Policy::
23774 @node Choosing the Scheduling Policy,,,Specifying a Run-Time Library
23775 @anchor{gnat_ugn/platform_specific_information id5}@anchor{1d3}
23776 @subsection Choosing the Scheduling Policy
23779 When using a POSIX threads implementation, you have a choice of several
23780 scheduling policies: @code{SCHED_FIFO}, @code{SCHED_RR} and @code{SCHED_OTHER}.
23782 Typically, the default is @code{SCHED_OTHER}, while using @code{SCHED_FIFO}
23783 or @code{SCHED_RR} requires special (e.g., root) privileges.
23785 @geindex pragma Time_Slice
23787 @geindex -T0 option
23789 @geindex pragma Task_Dispatching_Policy
23791 By default, GNAT uses the @code{SCHED_OTHER} policy. To specify
23793 you can use one of the following:
23799 @code{pragma Time_Slice (0.0)}
23802 the corresponding binder option @code{-T0}
23805 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
23808 To specify @code{SCHED_RR},
23809 you should use @code{pragma Time_Slice} with a
23810 value greater than 0.0, or else use the corresponding @code{-T}
23813 To make sure a program is running as root, you can put something like
23814 this in a library package body in your application:
23819 function geteuid return Integer;
23820 pragma Import (C, geteuid, "geteuid");
23821 Ignore : constant Boolean :=
23822 (if geteuid = 0 then True else raise Program_Error with "must be root");
23826 It gets the effective user id, and if it's not 0 (i.e. root), it raises
23833 @node GNU/Linux Topics,Microsoft Windows Topics,Specifying a Run-Time Library,Platform-Specific Information
23834 @anchor{gnat_ugn/platform_specific_information id6}@anchor{1d4}@anchor{gnat_ugn/platform_specific_information gnu-linux-topics}@anchor{1d5}
23835 @section GNU/Linux Topics
23838 This section describes topics that are specific to GNU/Linux platforms.
23841 * Required Packages on GNU/Linux::
23845 @node Required Packages on GNU/Linux,,,GNU/Linux Topics
23846 @anchor{gnat_ugn/platform_specific_information id7}@anchor{1d6}@anchor{gnat_ugn/platform_specific_information required-packages-on-gnu-linux}@anchor{1d7}
23847 @subsection Required Packages on GNU/Linux
23850 GNAT requires the C library developer's package to be installed.
23851 The name of of that package depends on your GNU/Linux distribution:
23857 RedHat, SUSE: @code{glibc-devel};
23860 Debian, Ubuntu: @code{libc6-dev} (normally installed by default).
23863 If using the 32-bit version of GNAT on a 64-bit version of GNU/Linux,
23864 you'll need the 32-bit version of the following packages:
23870 RedHat, SUSE: @code{glibc.i686}, @code{glibc-devel.i686}, @code{ncurses-libs.i686}
23873 Debian, Ubuntu: @code{libc6:i386}, @code{libc6-dev:i386}, @code{lib32ncursesw5}
23876 Other GNU/Linux distributions might be choosing a different name
23877 for those packages.
23881 @node Microsoft Windows Topics,Mac OS Topics,GNU/Linux Topics,Platform-Specific Information
23882 @anchor{gnat_ugn/platform_specific_information microsoft-windows-topics}@anchor{2c}@anchor{gnat_ugn/platform_specific_information id8}@anchor{1d8}
23883 @section Microsoft Windows Topics
23886 This section describes topics that are specific to the Microsoft Windows
23894 * Using GNAT on Windows::
23895 * Using a network installation of GNAT::
23896 * CONSOLE and WINDOWS subsystems::
23897 * Temporary Files::
23898 * Disabling Command Line Argument Expansion::
23899 * Mixed-Language Programming on Windows::
23900 * Windows Specific Add-Ons::
23904 @node Using GNAT on Windows,Using a network installation of GNAT,,Microsoft Windows Topics
23905 @anchor{gnat_ugn/platform_specific_information using-gnat-on-windows}@anchor{1d9}@anchor{gnat_ugn/platform_specific_information id9}@anchor{1da}
23906 @subsection Using GNAT on Windows
23909 One of the strengths of the GNAT technology is that its tool set
23910 (@code{gcc}, @code{gnatbind}, @code{gnatlink}, @code{gnatmake}, the
23911 @code{gdb} debugger, etc.) is used in the same way regardless of the
23914 On Windows this tool set is complemented by a number of Microsoft-specific
23915 tools that have been provided to facilitate interoperability with Windows
23916 when this is required. With these tools:
23922 You can build applications using the @code{CONSOLE} or @code{WINDOWS}
23926 You can use any Dynamically Linked Library (DLL) in your Ada code (both
23927 relocatable and non-relocatable DLLs are supported).
23930 You can build Ada DLLs for use in other applications. These applications
23931 can be written in a language other than Ada (e.g., C, C++, etc). Again both
23932 relocatable and non-relocatable Ada DLLs are supported.
23935 You can include Windows resources in your Ada application.
23938 You can use or create COM/DCOM objects.
23941 Immediately below are listed all known general GNAT-for-Windows restrictions.
23942 Other restrictions about specific features like Windows Resources and DLLs
23943 are listed in separate sections below.
23949 It is not possible to use @code{GetLastError} and @code{SetLastError}
23950 when tasking, protected records, or exceptions are used. In these
23951 cases, in order to implement Ada semantics, the GNAT run-time system
23952 calls certain Win32 routines that set the last error variable to 0 upon
23953 success. It should be possible to use @code{GetLastError} and
23954 @code{SetLastError} when tasking, protected record, and exception
23955 features are not used, but it is not guaranteed to work.
23958 It is not possible to link against Microsoft C++ libraries except for
23959 import libraries. Interfacing must be done by the mean of DLLs.
23962 It is possible to link against Microsoft C libraries. Yet the preferred
23963 solution is to use C/C++ compiler that comes with GNAT, since it
23964 doesn't require having two different development environments and makes the
23965 inter-language debugging experience smoother.
23968 When the compilation environment is located on FAT32 drives, users may
23969 experience recompilations of the source files that have not changed if
23970 Daylight Saving Time (DST) state has changed since the last time files
23971 were compiled. NTFS drives do not have this problem.
23974 No components of the GNAT toolset use any entries in the Windows
23975 registry. The only entries that can be created are file associations and
23976 PATH settings, provided the user has chosen to create them at installation
23977 time, as well as some minimal book-keeping information needed to correctly
23978 uninstall or integrate different GNAT products.
23981 @node Using a network installation of GNAT,CONSOLE and WINDOWS subsystems,Using GNAT on Windows,Microsoft Windows Topics
23982 @anchor{gnat_ugn/platform_specific_information id10}@anchor{1db}@anchor{gnat_ugn/platform_specific_information using-a-network-installation-of-gnat}@anchor{1dc}
23983 @subsection Using a network installation of GNAT
23986 Make sure the system on which GNAT is installed is accessible from the
23987 current machine, i.e., the install location is shared over the network.
23988 Shared resources are accessed on Windows by means of UNC paths, which
23989 have the format @code{\\\\server\\sharename\\path}
23991 In order to use such a network installation, simply add the UNC path of the
23992 @code{bin} directory of your GNAT installation in front of your PATH. For
23993 example, if GNAT is installed in @code{\GNAT} directory of a share location
23994 called @code{c-drive} on a machine @code{LOKI}, the following command will
24000 $ path \\loki\c-drive\gnat\bin;%path%`
24004 Be aware that every compilation using the network installation results in the
24005 transfer of large amounts of data across the network and will likely cause
24006 serious performance penalty.
24008 @node CONSOLE and WINDOWS subsystems,Temporary Files,Using a network installation of GNAT,Microsoft Windows Topics
24009 @anchor{gnat_ugn/platform_specific_information id11}@anchor{1dd}@anchor{gnat_ugn/platform_specific_information console-and-windows-subsystems}@anchor{1de}
24010 @subsection CONSOLE and WINDOWS subsystems
24013 @geindex CONSOLE Subsystem
24015 @geindex WINDOWS Subsystem
24019 There are two main subsystems under Windows. The @code{CONSOLE} subsystem
24020 (which is the default subsystem) will always create a console when
24021 launching the application. This is not something desirable when the
24022 application has a Windows GUI. To get rid of this console the
24023 application must be using the @code{WINDOWS} subsystem. To do so
24024 the @code{-mwindows} linker option must be specified.
24029 $ gnatmake winprog -largs -mwindows
24033 @node Temporary Files,Disabling Command Line Argument Expansion,CONSOLE and WINDOWS subsystems,Microsoft Windows Topics
24034 @anchor{gnat_ugn/platform_specific_information id12}@anchor{1df}@anchor{gnat_ugn/platform_specific_information temporary-files}@anchor{1e0}
24035 @subsection Temporary Files
24038 @geindex Temporary files
24040 It is possible to control where temporary files gets created by setting
24043 @geindex environment variable; TMP
24044 @code{TMP} environment variable. The file will be created:
24050 Under the directory pointed to by the
24052 @geindex environment variable; TMP
24053 @code{TMP} environment variable if
24054 this directory exists.
24057 Under @code{c:\temp}, if the
24059 @geindex environment variable; TMP
24060 @code{TMP} environment variable is not
24061 set (or not pointing to a directory) and if this directory exists.
24064 Under the current working directory otherwise.
24067 This allows you to determine exactly where the temporary
24068 file will be created. This is particularly useful in networked
24069 environments where you may not have write access to some
24072 @node Disabling Command Line Argument Expansion,Mixed-Language Programming on Windows,Temporary Files,Microsoft Windows Topics
24073 @anchor{gnat_ugn/platform_specific_information disabling-command-line-argument-expansion}@anchor{1e1}
24074 @subsection Disabling Command Line Argument Expansion
24077 @geindex Command Line Argument Expansion
24079 By default, an executable compiled for the Windows platform will do
24080 the following postprocessing on the arguments passed on the command
24087 If the argument contains the characters @code{*} and/or @code{?}, then
24088 file expansion will be attempted. For example, if the current directory
24089 contains @code{a.txt} and @code{b.txt}, then when calling:
24092 $ my_ada_program *.txt
24095 The following arguments will effectively be passed to the main program
24096 (for example when using @code{Ada.Command_Line.Argument}):
24099 Ada.Command_Line.Argument (1) -> "a.txt"
24100 Ada.Command_Line.Argument (2) -> "b.txt"
24104 Filename expansion can be disabled for a given argument by using single
24105 quotes. Thus, calling:
24108 $ my_ada_program '*.txt'
24114 Ada.Command_Line.Argument (1) -> "*.txt"
24118 Note that if the program is launched from a shell such as Cygwin Bash
24119 then quote removal might be performed by the shell.
24121 In some contexts it might be useful to disable this feature (for example if
24122 the program performs its own argument expansion). In order to do this, a C
24123 symbol needs to be defined and set to @code{0}. You can do this by
24124 adding the following code fragment in one of your Ada units:
24127 Do_Argv_Expansion : Integer := 0;
24128 pragma Export (C, Do_Argv_Expansion, "__gnat_do_argv_expansion");
24131 The results of previous examples will be respectively:
24134 Ada.Command_Line.Argument (1) -> "*.txt"
24140 Ada.Command_Line.Argument (1) -> "'*.txt'"
24143 @node Mixed-Language Programming on Windows,Windows Specific Add-Ons,Disabling Command Line Argument Expansion,Microsoft Windows Topics
24144 @anchor{gnat_ugn/platform_specific_information id13}@anchor{1e2}@anchor{gnat_ugn/platform_specific_information mixed-language-programming-on-windows}@anchor{1e3}
24145 @subsection Mixed-Language Programming on Windows
24148 Developing pure Ada applications on Windows is no different than on
24149 other GNAT-supported platforms. However, when developing or porting an
24150 application that contains a mix of Ada and C/C++, the choice of your
24151 Windows C/C++ development environment conditions your overall
24152 interoperability strategy.
24154 If you use @code{gcc} or Microsoft C to compile the non-Ada part of
24155 your application, there are no Windows-specific restrictions that
24156 affect the overall interoperability with your Ada code. If you do want
24157 to use the Microsoft tools for your C++ code, you have two choices:
24163 Encapsulate your C++ code in a DLL to be linked with your Ada
24164 application. In this case, use the Microsoft or whatever environment to
24165 build the DLL and use GNAT to build your executable
24166 (@ref{1e4,,Using DLLs with GNAT}).
24169 Or you can encapsulate your Ada code in a DLL to be linked with the
24170 other part of your application. In this case, use GNAT to build the DLL
24171 (@ref{1e5,,Building DLLs with GNAT Project files}) and use the Microsoft
24172 or whatever environment to build your executable.
24175 In addition to the description about C main in
24176 @ref{44,,Mixed Language Programming} section, if the C main uses a
24177 stand-alone library it is required on x86-windows to
24178 setup the SEH context. For this the C main must looks like this:
24184 extern void adainit (void);
24185 extern void adafinal (void);
24186 extern void __gnat_initialize(void*);
24187 extern void call_to_ada (void);
24189 int main (int argc, char *argv[])
24193 /* Initialize the SEH context */
24194 __gnat_initialize (&SEH);
24198 /* Then call Ada services in the stand-alone library */
24207 Note that this is not needed on x86_64-windows where the Windows
24208 native SEH support is used.
24211 * Windows Calling Conventions::
24212 * Introduction to Dynamic Link Libraries (DLLs): Introduction to Dynamic Link Libraries DLLs.
24213 * Using DLLs with GNAT::
24214 * Building DLLs with GNAT Project files::
24215 * Building DLLs with GNAT::
24216 * Building DLLs with gnatdll::
24217 * Ada DLLs and Finalization::
24218 * Creating a Spec for Ada DLLs::
24219 * GNAT and Windows Resources::
24220 * Using GNAT DLLs from Microsoft Visual Studio Applications::
24221 * Debugging a DLL::
24222 * Setting Stack Size from gnatlink::
24223 * Setting Heap Size from gnatlink::
24227 @node Windows Calling Conventions,Introduction to Dynamic Link Libraries DLLs,,Mixed-Language Programming on Windows
24228 @anchor{gnat_ugn/platform_specific_information windows-calling-conventions}@anchor{1e6}@anchor{gnat_ugn/platform_specific_information id14}@anchor{1e7}
24229 @subsubsection Windows Calling Conventions
24236 This section pertain only to Win32. On Win64 there is a single native
24237 calling convention. All convention specifiers are ignored on this
24240 When a subprogram @code{F} (caller) calls a subprogram @code{G}
24241 (callee), there are several ways to push @code{G}'s parameters on the
24242 stack and there are several possible scenarios to clean up the stack
24243 upon @code{G}'s return. A calling convention is an agreed upon software
24244 protocol whereby the responsibilities between the caller (@code{F}) and
24245 the callee (@code{G}) are clearly defined. Several calling conventions
24246 are available for Windows:
24252 @code{C} (Microsoft defined)
24255 @code{Stdcall} (Microsoft defined)
24258 @code{Win32} (GNAT specific)
24261 @code{DLL} (GNAT specific)
24265 * C Calling Convention::
24266 * Stdcall Calling Convention::
24267 * Win32 Calling Convention::
24268 * DLL Calling Convention::
24272 @node C Calling Convention,Stdcall Calling Convention,,Windows Calling Conventions
24273 @anchor{gnat_ugn/platform_specific_information c-calling-convention}@anchor{1e8}@anchor{gnat_ugn/platform_specific_information id15}@anchor{1e9}
24274 @subsubsection @code{C} Calling Convention
24277 This is the default calling convention used when interfacing to C/C++
24278 routines compiled with either @code{gcc} or Microsoft Visual C++.
24280 In the @code{C} calling convention subprogram parameters are pushed on the
24281 stack by the caller from right to left. The caller itself is in charge of
24282 cleaning up the stack after the call. In addition, the name of a routine
24283 with @code{C} calling convention is mangled by adding a leading underscore.
24285 The name to use on the Ada side when importing (or exporting) a routine
24286 with @code{C} calling convention is the name of the routine. For
24287 instance the C function:
24292 int get_val (long);
24296 should be imported from Ada as follows:
24301 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
24302 pragma Import (C, Get_Val, External_Name => "get_val");
24306 Note that in this particular case the @code{External_Name} parameter could
24307 have been omitted since, when missing, this parameter is taken to be the
24308 name of the Ada entity in lower case. When the @code{Link_Name} parameter
24309 is missing, as in the above example, this parameter is set to be the
24310 @code{External_Name} with a leading underscore.
24312 When importing a variable defined in C, you should always use the @code{C}
24313 calling convention unless the object containing the variable is part of a
24314 DLL (in which case you should use the @code{Stdcall} calling
24315 convention, @ref{1ea,,Stdcall Calling Convention}).
24317 @node Stdcall Calling Convention,Win32 Calling Convention,C Calling Convention,Windows Calling Conventions
24318 @anchor{gnat_ugn/platform_specific_information stdcall-calling-convention}@anchor{1ea}@anchor{gnat_ugn/platform_specific_information id16}@anchor{1eb}
24319 @subsubsection @code{Stdcall} Calling Convention
24322 This convention, which was the calling convention used for Pascal
24323 programs, is used by Microsoft for all the routines in the Win32 API for
24324 efficiency reasons. It must be used to import any routine for which this
24325 convention was specified.
24327 In the @code{Stdcall} calling convention subprogram parameters are pushed
24328 on the stack by the caller from right to left. The callee (and not the
24329 caller) is in charge of cleaning the stack on routine exit. In addition,
24330 the name of a routine with @code{Stdcall} calling convention is mangled by
24331 adding a leading underscore (as for the @code{C} calling convention) and a
24332 trailing @code{@@@emph{nn}}, where @code{nn} is the overall size (in
24333 bytes) of the parameters passed to the routine.
24335 The name to use on the Ada side when importing a C routine with a
24336 @code{Stdcall} calling convention is the name of the C routine. The leading
24337 underscore and trailing @code{@@@emph{nn}} are added automatically by
24338 the compiler. For instance the Win32 function:
24343 APIENTRY int get_val (long);
24347 should be imported from Ada as follows:
24352 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
24353 pragma Import (Stdcall, Get_Val);
24354 -- On the x86 a long is 4 bytes, so the Link_Name is "_get_val@@4"
24358 As for the @code{C} calling convention, when the @code{External_Name}
24359 parameter is missing, it is taken to be the name of the Ada entity in lower
24360 case. If instead of writing the above import pragma you write:
24365 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
24366 pragma Import (Stdcall, Get_Val, External_Name => "retrieve_val");
24370 then the imported routine is @code{_retrieve_val@@4}. However, if instead
24371 of specifying the @code{External_Name} parameter you specify the
24372 @code{Link_Name} as in the following example:
24377 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
24378 pragma Import (Stdcall, Get_Val, Link_Name => "retrieve_val");
24382 then the imported routine is @code{retrieve_val}, that is, there is no
24383 decoration at all. No leading underscore and no Stdcall suffix
24384 @code{@@@emph{nn}}.
24386 This is especially important as in some special cases a DLL's entry
24387 point name lacks a trailing @code{@@@emph{nn}} while the exported
24388 name generated for a call has it.
24390 It is also possible to import variables defined in a DLL by using an
24391 import pragma for a variable. As an example, if a DLL contains a
24392 variable defined as:
24401 then, to access this variable from Ada you should write:
24406 My_Var : Interfaces.C.int;
24407 pragma Import (Stdcall, My_Var);
24411 Note that to ease building cross-platform bindings this convention
24412 will be handled as a @code{C} calling convention on non-Windows platforms.
24414 @node Win32 Calling Convention,DLL Calling Convention,Stdcall Calling Convention,Windows Calling Conventions
24415 @anchor{gnat_ugn/platform_specific_information win32-calling-convention}@anchor{1ec}@anchor{gnat_ugn/platform_specific_information id17}@anchor{1ed}
24416 @subsubsection @code{Win32} Calling Convention
24419 This convention, which is GNAT-specific is fully equivalent to the
24420 @code{Stdcall} calling convention described above.
24422 @node DLL Calling Convention,,Win32 Calling Convention,Windows Calling Conventions
24423 @anchor{gnat_ugn/platform_specific_information id18}@anchor{1ee}@anchor{gnat_ugn/platform_specific_information dll-calling-convention}@anchor{1ef}
24424 @subsubsection @code{DLL} Calling Convention
24427 This convention, which is GNAT-specific is fully equivalent to the
24428 @code{Stdcall} calling convention described above.
24430 @node Introduction to Dynamic Link Libraries DLLs,Using DLLs with GNAT,Windows Calling Conventions,Mixed-Language Programming on Windows
24431 @anchor{gnat_ugn/platform_specific_information id19}@anchor{1f0}@anchor{gnat_ugn/platform_specific_information introduction-to-dynamic-link-libraries-dlls}@anchor{1f1}
24432 @subsubsection Introduction to Dynamic Link Libraries (DLLs)
24437 A Dynamically Linked Library (DLL) is a library that can be shared by
24438 several applications running under Windows. A DLL can contain any number of
24439 routines and variables.
24441 One advantage of DLLs is that you can change and enhance them without
24442 forcing all the applications that depend on them to be relinked or
24443 recompiled. However, you should be aware than all calls to DLL routines are
24444 slower since, as you will understand below, such calls are indirect.
24446 To illustrate the remainder of this section, suppose that an application
24447 wants to use the services of a DLL @code{API.dll}. To use the services
24448 provided by @code{API.dll} you must statically link against the DLL or
24449 an import library which contains a jump table with an entry for each
24450 routine and variable exported by the DLL. In the Microsoft world this
24451 import library is called @code{API.lib}. When using GNAT this import
24452 library is called either @code{libAPI.dll.a}, @code{libapi.dll.a},
24453 @code{libAPI.a} or @code{libapi.a} (names are case insensitive).
24455 After you have linked your application with the DLL or the import library
24456 and you run your application, here is what happens:
24462 Your application is loaded into memory.
24465 The DLL @code{API.dll} is mapped into the address space of your
24466 application. This means that:
24472 The DLL will use the stack of the calling thread.
24475 The DLL will use the virtual address space of the calling process.
24478 The DLL will allocate memory from the virtual address space of the calling
24482 Handles (pointers) can be safely exchanged between routines in the DLL
24483 routines and routines in the application using the DLL.
24487 The entries in the jump table (from the import library @code{libAPI.dll.a}
24488 or @code{API.lib} or automatically created when linking against a DLL)
24489 which is part of your application are initialized with the addresses
24490 of the routines and variables in @code{API.dll}.
24493 If present in @code{API.dll}, routines @code{DllMain} or
24494 @code{DllMainCRTStartup} are invoked. These routines typically contain
24495 the initialization code needed for the well-being of the routines and
24496 variables exported by the DLL.
24499 There is an additional point which is worth mentioning. In the Windows
24500 world there are two kind of DLLs: relocatable and non-relocatable
24501 DLLs. Non-relocatable DLLs can only be loaded at a very specific address
24502 in the target application address space. If the addresses of two
24503 non-relocatable DLLs overlap and these happen to be used by the same
24504 application, a conflict will occur and the application will run
24505 incorrectly. Hence, when possible, it is always preferable to use and
24506 build relocatable DLLs. Both relocatable and non-relocatable DLLs are
24507 supported by GNAT. Note that the @code{-s} linker option (see GNU Linker
24508 User's Guide) removes the debugging symbols from the DLL but the DLL can
24509 still be relocated.
24511 As a side note, an interesting difference between Microsoft DLLs and
24512 Unix shared libraries, is the fact that on most Unix systems all public
24513 routines are exported by default in a Unix shared library, while under
24514 Windows it is possible (but not required) to list exported routines in
24515 a definition file (see @ref{1f2,,The Definition File}).
24517 @node Using DLLs with GNAT,Building DLLs with GNAT Project files,Introduction to Dynamic Link Libraries DLLs,Mixed-Language Programming on Windows
24518 @anchor{gnat_ugn/platform_specific_information id20}@anchor{1f3}@anchor{gnat_ugn/platform_specific_information using-dlls-with-gnat}@anchor{1e4}
24519 @subsubsection Using DLLs with GNAT
24522 To use the services of a DLL, say @code{API.dll}, in your Ada application
24529 The Ada spec for the routines and/or variables you want to access in
24530 @code{API.dll}. If not available this Ada spec must be built from the C/C++
24531 header files provided with the DLL.
24534 The import library (@code{libAPI.dll.a} or @code{API.lib}). As previously
24535 mentioned an import library is a statically linked library containing the
24536 import table which will be filled at load time to point to the actual
24537 @code{API.dll} routines. Sometimes you don't have an import library for the
24538 DLL you want to use. The following sections will explain how to build
24539 one. Note that this is optional.
24542 The actual DLL, @code{API.dll}.
24545 Once you have all the above, to compile an Ada application that uses the
24546 services of @code{API.dll} and whose main subprogram is @code{My_Ada_App},
24547 you simply issue the command
24552 $ gnatmake my_ada_app -largs -lAPI
24556 The argument @code{-largs -lAPI} at the end of the @code{gnatmake} command
24557 tells the GNAT linker to look for an import library. The linker will
24558 look for a library name in this specific order:
24564 @code{libAPI.dll.a}
24582 The first three are the GNU style import libraries. The third is the
24583 Microsoft style import libraries. The last two are the actual DLL names.
24585 Note that if the Ada package spec for @code{API.dll} contains the
24591 pragma Linker_Options ("-lAPI");
24595 you do not have to add @code{-largs -lAPI} at the end of the
24596 @code{gnatmake} command.
24598 If any one of the items above is missing you will have to create it
24599 yourself. The following sections explain how to do so using as an
24600 example a fictitious DLL called @code{API.dll}.
24603 * Creating an Ada Spec for the DLL Services::
24604 * Creating an Import Library::
24608 @node Creating an Ada Spec for the DLL Services,Creating an Import Library,,Using DLLs with GNAT
24609 @anchor{gnat_ugn/platform_specific_information id21}@anchor{1f4}@anchor{gnat_ugn/platform_specific_information creating-an-ada-spec-for-the-dll-services}@anchor{1f5}
24610 @subsubsection Creating an Ada Spec for the DLL Services
24613 A DLL typically comes with a C/C++ header file which provides the
24614 definitions of the routines and variables exported by the DLL. The Ada
24615 equivalent of this header file is a package spec that contains definitions
24616 for the imported entities. If the DLL you intend to use does not come with
24617 an Ada spec you have to generate one such spec yourself. For example if
24618 the header file of @code{API.dll} is a file @code{api.h} containing the
24619 following two definitions:
24629 then the equivalent Ada spec could be:
24634 with Interfaces.C.Strings;
24639 function Get (Str : C.Strings.Chars_Ptr) return C.int;
24642 pragma Import (C, Get);
24643 pragma Import (DLL, Some_Var);
24648 @node Creating an Import Library,,Creating an Ada Spec for the DLL Services,Using DLLs with GNAT
24649 @anchor{gnat_ugn/platform_specific_information id22}@anchor{1f6}@anchor{gnat_ugn/platform_specific_information creating-an-import-library}@anchor{1f7}
24650 @subsubsection Creating an Import Library
24653 @geindex Import library
24655 If a Microsoft-style import library @code{API.lib} or a GNAT-style
24656 import library @code{libAPI.dll.a} or @code{libAPI.a} is available
24657 with @code{API.dll} you can skip this section. You can also skip this
24658 section if @code{API.dll} or @code{libAPI.dll} is built with GNU tools
24659 as in this case it is possible to link directly against the
24660 DLL. Otherwise read on.
24662 @geindex Definition file
24663 @anchor{gnat_ugn/platform_specific_information the-definition-file}@anchor{1f2}
24664 @subsubheading The Definition File
24667 As previously mentioned, and unlike Unix systems, the list of symbols
24668 that are exported from a DLL must be provided explicitly in Windows.
24669 The main goal of a definition file is precisely that: list the symbols
24670 exported by a DLL. A definition file (usually a file with a @code{.def}
24671 suffix) has the following structure:
24676 [LIBRARY `@w{`}name`@w{`}]
24677 [DESCRIPTION `@w{`}string`@w{`}]
24679 `@w{`}symbol1`@w{`}
24680 `@w{`}symbol2`@w{`}
24688 @item @emph{LIBRARY name}
24690 This section, which is optional, gives the name of the DLL.
24692 @item @emph{DESCRIPTION string}
24694 This section, which is optional, gives a description string that will be
24695 embedded in the import library.
24697 @item @emph{EXPORTS}
24699 This section gives the list of exported symbols (procedures, functions or
24700 variables). For instance in the case of @code{API.dll} the @code{EXPORTS}
24701 section of @code{API.def} looks like:
24710 Note that you must specify the correct suffix (@code{@@@emph{nn}})
24711 (see @ref{1e6,,Windows Calling Conventions}) for a Stdcall
24712 calling convention function in the exported symbols list.
24714 There can actually be other sections in a definition file, but these
24715 sections are not relevant to the discussion at hand.
24716 @anchor{gnat_ugn/platform_specific_information create-def-file-automatically}@anchor{1f8}
24717 @subsubheading Creating a Definition File Automatically
24720 You can automatically create the definition file @code{API.def}
24721 (see @ref{1f2,,The Definition File}) from a DLL.
24722 For that use the @code{dlltool} program as follows:
24727 $ dlltool API.dll -z API.def --export-all-symbols
24730 Note that if some routines in the DLL have the @code{Stdcall} convention
24731 (@ref{1e6,,Windows Calling Conventions}) with stripped @code{@@@emph{nn}}
24732 suffix then you'll have to edit @code{api.def} to add it, and specify
24733 @code{-k} to @code{gnatdll} when creating the import library.
24735 Here are some hints to find the right @code{@@@emph{nn}} suffix.
24741 If you have the Microsoft import library (.lib), it is possible to get
24742 the right symbols by using Microsoft @code{dumpbin} tool (see the
24743 corresponding Microsoft documentation for further details).
24746 $ dumpbin /exports api.lib
24750 If you have a message about a missing symbol at link time the compiler
24751 tells you what symbol is expected. You just have to go back to the
24752 definition file and add the right suffix.
24755 @anchor{gnat_ugn/platform_specific_information gnat-style-import-library}@anchor{1f9}
24756 @subsubheading GNAT-Style Import Library
24759 To create a static import library from @code{API.dll} with the GNAT tools
24760 you should create the .def file, then use @code{gnatdll} tool
24761 (see @ref{1fa,,Using gnatdll}) as follows:
24766 $ gnatdll -e API.def -d API.dll
24769 @code{gnatdll} takes as input a definition file @code{API.def} and the
24770 name of the DLL containing the services listed in the definition file
24771 @code{API.dll}. The name of the static import library generated is
24772 computed from the name of the definition file as follows: if the
24773 definition file name is @code{xyz.def}, the import library name will
24774 be @code{libxyz.a}. Note that in the previous example option
24775 @code{-e} could have been removed because the name of the definition
24776 file (before the @code{.def} suffix) is the same as the name of the
24777 DLL (@ref{1fa,,Using gnatdll} for more information about @code{gnatdll}).
24779 @anchor{gnat_ugn/platform_specific_information msvs-style-import-library}@anchor{1fb}
24780 @subsubheading Microsoft-Style Import Library
24783 A Microsoft import library is needed only if you plan to make an
24784 Ada DLL available to applications developed with Microsoft
24785 tools (@ref{1e3,,Mixed-Language Programming on Windows}).
24787 To create a Microsoft-style import library for @code{API.dll} you
24788 should create the .def file, then build the actual import library using
24789 Microsoft's @code{lib} utility:
24794 $ lib -machine:IX86 -def:API.def -out:API.lib
24797 If you use the above command the definition file @code{API.def} must
24798 contain a line giving the name of the DLL:
24804 See the Microsoft documentation for further details about the usage of
24808 @node Building DLLs with GNAT Project files,Building DLLs with GNAT,Using DLLs with GNAT,Mixed-Language Programming on Windows
24809 @anchor{gnat_ugn/platform_specific_information id23}@anchor{1fc}@anchor{gnat_ugn/platform_specific_information building-dlls-with-gnat-project-files}@anchor{1e5}
24810 @subsubsection Building DLLs with GNAT Project files
24816 There is nothing specific to Windows in the build process.
24817 See the @emph{Library Projects} section in the @emph{GNAT Project Manager}
24818 chapter of the @emph{GPRbuild User's Guide}.
24820 Due to a system limitation, it is not possible under Windows to create threads
24821 when inside the @code{DllMain} routine which is used for auto-initialization
24822 of shared libraries, so it is not possible to have library level tasks in SALs.
24824 @node Building DLLs with GNAT,Building DLLs with gnatdll,Building DLLs with GNAT Project files,Mixed-Language Programming on Windows
24825 @anchor{gnat_ugn/platform_specific_information building-dlls-with-gnat}@anchor{1fd}@anchor{gnat_ugn/platform_specific_information id24}@anchor{1fe}
24826 @subsubsection Building DLLs with GNAT
24832 This section explain how to build DLLs using the GNAT built-in DLL
24833 support. With the following procedure it is straight forward to build
24834 and use DLLs with GNAT.
24840 Building object files.
24841 The first step is to build all objects files that are to be included
24842 into the DLL. This is done by using the standard @code{gnatmake} tool.
24846 To build the DLL you must use the @code{gcc} @code{-shared} and
24847 @code{-shared-libgcc} options. It is quite simple to use this method:
24850 $ gcc -shared -shared-libgcc -o api.dll obj1.o obj2.o ...
24853 It is important to note that in this case all symbols found in the
24854 object files are automatically exported. It is possible to restrict
24855 the set of symbols to export by passing to @code{gcc} a definition
24856 file (see @ref{1f2,,The Definition File}).
24860 $ gcc -shared -shared-libgcc -o api.dll api.def obj1.o obj2.o ...
24863 If you use a definition file you must export the elaboration procedures
24864 for every package that required one. Elaboration procedures are named
24865 using the package name followed by "_E".
24868 Preparing DLL to be used.
24869 For the DLL to be used by client programs the bodies must be hidden
24870 from it and the .ali set with read-only attribute. This is very important
24871 otherwise GNAT will recompile all packages and will not actually use
24872 the code in the DLL. For example:
24876 $ copy *.ads *.ali api.dll apilib
24877 $ attrib +R apilib\\*.ali
24881 At this point it is possible to use the DLL by directly linking
24882 against it. Note that you must use the GNAT shared runtime when using
24883 GNAT shared libraries. This is achieved by using the @code{-shared} binder
24889 $ gnatmake main -Iapilib -bargs -shared -largs -Lapilib -lAPI
24893 @node Building DLLs with gnatdll,Ada DLLs and Finalization,Building DLLs with GNAT,Mixed-Language Programming on Windows
24894 @anchor{gnat_ugn/platform_specific_information building-dlls-with-gnatdll}@anchor{1ff}@anchor{gnat_ugn/platform_specific_information id25}@anchor{200}
24895 @subsubsection Building DLLs with gnatdll
24901 Note that it is preferred to use GNAT Project files
24902 (@ref{1e5,,Building DLLs with GNAT Project files}) or the built-in GNAT
24903 DLL support (@ref{1fd,,Building DLLs with GNAT}) or to build DLLs.
24905 This section explains how to build DLLs containing Ada code using
24906 @code{gnatdll}. These DLLs will be referred to as Ada DLLs in the
24907 remainder of this section.
24909 The steps required to build an Ada DLL that is to be used by Ada as well as
24910 non-Ada applications are as follows:
24916 You need to mark each Ada entity exported by the DLL with a @code{C} or
24917 @code{Stdcall} calling convention to avoid any Ada name mangling for the
24918 entities exported by the DLL
24919 (see @ref{201,,Exporting Ada Entities}). You can
24920 skip this step if you plan to use the Ada DLL only from Ada applications.
24923 Your Ada code must export an initialization routine which calls the routine
24924 @code{adainit} generated by @code{gnatbind} to perform the elaboration of
24925 the Ada code in the DLL (@ref{202,,Ada DLLs and Elaboration}). The initialization
24926 routine exported by the Ada DLL must be invoked by the clients of the DLL
24927 to initialize the DLL.
24930 When useful, the DLL should also export a finalization routine which calls
24931 routine @code{adafinal} generated by @code{gnatbind} to perform the
24932 finalization of the Ada code in the DLL (@ref{203,,Ada DLLs and Finalization}).
24933 The finalization routine exported by the Ada DLL must be invoked by the
24934 clients of the DLL when the DLL services are no further needed.
24937 You must provide a spec for the services exported by the Ada DLL in each
24938 of the programming languages to which you plan to make the DLL available.
24941 You must provide a definition file listing the exported entities
24942 (@ref{1f2,,The Definition File}).
24945 Finally you must use @code{gnatdll} to produce the DLL and the import
24946 library (@ref{1fa,,Using gnatdll}).
24949 Note that a relocatable DLL stripped using the @code{strip}
24950 binutils tool will not be relocatable anymore. To build a DLL without
24951 debug information pass @code{-largs -s} to @code{gnatdll}. This
24952 restriction does not apply to a DLL built using a Library Project.
24953 See the @emph{Library Projects} section in the @emph{GNAT Project Manager}
24954 chapter of the @emph{GPRbuild User's Guide}.
24956 @c Limitations_When_Using_Ada_DLLs_from Ada:
24959 * Limitations When Using Ada DLLs from Ada::
24960 * Exporting Ada Entities::
24961 * Ada DLLs and Elaboration::
24965 @node Limitations When Using Ada DLLs from Ada,Exporting Ada Entities,,Building DLLs with gnatdll
24966 @anchor{gnat_ugn/platform_specific_information limitations-when-using-ada-dlls-from-ada}@anchor{204}
24967 @subsubsection Limitations When Using Ada DLLs from Ada
24970 When using Ada DLLs from Ada applications there is a limitation users
24971 should be aware of. Because on Windows the GNAT run-time is not in a DLL of
24972 its own, each Ada DLL includes a part of the GNAT run-time. Specifically,
24973 each Ada DLL includes the services of the GNAT run-time that are necessary
24974 to the Ada code inside the DLL. As a result, when an Ada program uses an
24975 Ada DLL there are two independent GNAT run-times: one in the Ada DLL and
24976 one in the main program.
24978 It is therefore not possible to exchange GNAT run-time objects between the
24979 Ada DLL and the main Ada program. Example of GNAT run-time objects are file
24980 handles (e.g., @code{Text_IO.File_Type}), tasks types, protected objects
24983 It is completely safe to exchange plain elementary, array or record types,
24984 Windows object handles, etc.
24986 @node Exporting Ada Entities,Ada DLLs and Elaboration,Limitations When Using Ada DLLs from Ada,Building DLLs with gnatdll
24987 @anchor{gnat_ugn/platform_specific_information exporting-ada-entities}@anchor{201}@anchor{gnat_ugn/platform_specific_information id26}@anchor{205}
24988 @subsubsection Exporting Ada Entities
24991 @geindex Export table
24993 Building a DLL is a way to encapsulate a set of services usable from any
24994 application. As a result, the Ada entities exported by a DLL should be
24995 exported with the @code{C} or @code{Stdcall} calling conventions to avoid
24996 any Ada name mangling. As an example here is an Ada package
24997 @code{API}, spec and body, exporting two procedures, a function, and a
25003 with Interfaces.C; use Interfaces;
25005 Count : C.int := 0;
25006 function Factorial (Val : C.int) return C.int;
25008 procedure Initialize_API;
25009 procedure Finalize_API;
25010 -- Initialization & Finalization routines. More in the next section.
25012 pragma Export (C, Initialize_API);
25013 pragma Export (C, Finalize_API);
25014 pragma Export (C, Count);
25015 pragma Export (C, Factorial);
25020 package body API is
25021 function Factorial (Val : C.int) return C.int is
25024 Count := Count + 1;
25025 for K in 1 .. Val loop
25031 procedure Initialize_API is
25033 pragma Import (C, Adainit);
25036 end Initialize_API;
25038 procedure Finalize_API is
25039 procedure Adafinal;
25040 pragma Import (C, Adafinal);
25048 If the Ada DLL you are building will only be used by Ada applications
25049 you do not have to export Ada entities with a @code{C} or @code{Stdcall}
25050 convention. As an example, the previous package could be written as
25057 Count : Integer := 0;
25058 function Factorial (Val : Integer) return Integer;
25060 procedure Initialize_API;
25061 procedure Finalize_API;
25062 -- Initialization and Finalization routines.
25067 package body API is
25068 function Factorial (Val : Integer) return Integer is
25069 Fact : Integer := 1;
25071 Count := Count + 1;
25072 for K in 1 .. Val loop
25079 -- The remainder of this package body is unchanged.
25084 Note that if you do not export the Ada entities with a @code{C} or
25085 @code{Stdcall} convention you will have to provide the mangled Ada names
25086 in the definition file of the Ada DLL
25087 (@ref{206,,Creating the Definition File}).
25089 @node Ada DLLs and Elaboration,,Exporting Ada Entities,Building DLLs with gnatdll
25090 @anchor{gnat_ugn/platform_specific_information ada-dlls-and-elaboration}@anchor{202}@anchor{gnat_ugn/platform_specific_information id27}@anchor{207}
25091 @subsubsection Ada DLLs and Elaboration
25094 @geindex DLLs and elaboration
25096 The DLL that you are building contains your Ada code as well as all the
25097 routines in the Ada library that are needed by it. The first thing a
25098 user of your DLL must do is elaborate the Ada code
25099 (@ref{f,,Elaboration Order Handling in GNAT}).
25101 To achieve this you must export an initialization routine
25102 (@code{Initialize_API} in the previous example), which must be invoked
25103 before using any of the DLL services. This elaboration routine must call
25104 the Ada elaboration routine @code{adainit} generated by the GNAT binder
25105 (@ref{b4,,Binding with Non-Ada Main Programs}). See the body of
25106 @code{Initialize_Api} for an example. Note that the GNAT binder is
25107 automatically invoked during the DLL build process by the @code{gnatdll}
25108 tool (@ref{1fa,,Using gnatdll}).
25110 When a DLL is loaded, Windows systematically invokes a routine called
25111 @code{DllMain}. It would therefore be possible to call @code{adainit}
25112 directly from @code{DllMain} without having to provide an explicit
25113 initialization routine. Unfortunately, it is not possible to call
25114 @code{adainit} from the @code{DllMain} if your program has library level
25115 tasks because access to the @code{DllMain} entry point is serialized by
25116 the system (that is, only a single thread can execute 'through' it at a
25117 time), which means that the GNAT run-time will deadlock waiting for the
25118 newly created task to complete its initialization.
25120 @node Ada DLLs and Finalization,Creating a Spec for Ada DLLs,Building DLLs with gnatdll,Mixed-Language Programming on Windows
25121 @anchor{gnat_ugn/platform_specific_information id28}@anchor{208}@anchor{gnat_ugn/platform_specific_information ada-dlls-and-finalization}@anchor{203}
25122 @subsubsection Ada DLLs and Finalization
25125 @geindex DLLs and finalization
25127 When the services of an Ada DLL are no longer needed, the client code should
25128 invoke the DLL finalization routine, if available. The DLL finalization
25129 routine is in charge of releasing all resources acquired by the DLL. In the
25130 case of the Ada code contained in the DLL, this is achieved by calling
25131 routine @code{adafinal} generated by the GNAT binder
25132 (@ref{b4,,Binding with Non-Ada Main Programs}).
25133 See the body of @code{Finalize_Api} for an
25134 example. As already pointed out the GNAT binder is automatically invoked
25135 during the DLL build process by the @code{gnatdll} tool
25136 (@ref{1fa,,Using gnatdll}).
25138 @node Creating a Spec for Ada DLLs,GNAT and Windows Resources,Ada DLLs and Finalization,Mixed-Language Programming on Windows
25139 @anchor{gnat_ugn/platform_specific_information id29}@anchor{209}@anchor{gnat_ugn/platform_specific_information creating-a-spec-for-ada-dlls}@anchor{20a}
25140 @subsubsection Creating a Spec for Ada DLLs
25143 To use the services exported by the Ada DLL from another programming
25144 language (e.g., C), you have to translate the specs of the exported Ada
25145 entities in that language. For instance in the case of @code{API.dll},
25146 the corresponding C header file could look like:
25151 extern int *_imp__count;
25152 #define count (*_imp__count)
25153 int factorial (int);
25157 It is important to understand that when building an Ada DLL to be used by
25158 other Ada applications, you need two different specs for the packages
25159 contained in the DLL: one for building the DLL and the other for using
25160 the DLL. This is because the @code{DLL} calling convention is needed to
25161 use a variable defined in a DLL, but when building the DLL, the variable
25162 must have either the @code{Ada} or @code{C} calling convention. As an
25163 example consider a DLL comprising the following package @code{API}:
25169 Count : Integer := 0;
25171 -- Remainder of the package omitted.
25176 After producing a DLL containing package @code{API}, the spec that
25177 must be used to import @code{API.Count} from Ada code outside of the
25185 pragma Import (DLL, Count);
25191 * Creating the Definition File::
25196 @node Creating the Definition File,Using gnatdll,,Creating a Spec for Ada DLLs
25197 @anchor{gnat_ugn/platform_specific_information creating-the-definition-file}@anchor{206}@anchor{gnat_ugn/platform_specific_information id30}@anchor{20b}
25198 @subsubsection Creating the Definition File
25201 The definition file is the last file needed to build the DLL. It lists
25202 the exported symbols. As an example, the definition file for a DLL
25203 containing only package @code{API} (where all the entities are exported
25204 with a @code{C} calling convention) is:
25217 If the @code{C} calling convention is missing from package @code{API},
25218 then the definition file contains the mangled Ada names of the above
25219 entities, which in this case are:
25228 api__initialize_api
25232 @node Using gnatdll,,Creating the Definition File,Creating a Spec for Ada DLLs
25233 @anchor{gnat_ugn/platform_specific_information using-gnatdll}@anchor{1fa}@anchor{gnat_ugn/platform_specific_information id31}@anchor{20c}
25234 @subsubsection Using @code{gnatdll}
25239 @code{gnatdll} is a tool to automate the DLL build process once all the Ada
25240 and non-Ada sources that make up your DLL have been compiled.
25241 @code{gnatdll} is actually in charge of two distinct tasks: build the
25242 static import library for the DLL and the actual DLL. The form of the
25243 @code{gnatdll} command is
25248 $ gnatdll [ switches ] list-of-files [ -largs opts ]
25252 where @code{list-of-files} is a list of ALI and object files. The object
25253 file list must be the exact list of objects corresponding to the non-Ada
25254 sources whose services are to be included in the DLL. The ALI file list
25255 must be the exact list of ALI files for the corresponding Ada sources
25256 whose services are to be included in the DLL. If @code{list-of-files} is
25257 missing, only the static import library is generated.
25259 You may specify any of the following switches to @code{gnatdll}:
25263 @geindex -a (gnatdll)
25269 @item @code{-a[@emph{address}]}
25271 Build a non-relocatable DLL at @code{address}. If @code{address} is not
25272 specified the default address @code{0x11000000} will be used. By default,
25273 when this switch is missing, @code{gnatdll} builds relocatable DLL. We
25274 advise the reader to build relocatable DLL.
25276 @geindex -b (gnatdll)
25278 @item @code{-b @emph{address}}
25280 Set the relocatable DLL base address. By default the address is
25283 @geindex -bargs (gnatdll)
25285 @item @code{-bargs @emph{opts}}
25287 Binder options. Pass @code{opts} to the binder.
25289 @geindex -d (gnatdll)
25291 @item @code{-d @emph{dllfile}}
25293 @code{dllfile} is the name of the DLL. This switch must be present for
25294 @code{gnatdll} to do anything. The name of the generated import library is
25295 obtained algorithmically from @code{dllfile} as shown in the following
25296 example: if @code{dllfile} is @code{xyz.dll}, the import library name is
25297 @code{libxyz.dll.a}. The name of the definition file to use (if not specified
25298 by option @code{-e}) is obtained algorithmically from @code{dllfile}
25299 as shown in the following example:
25300 if @code{dllfile} is @code{xyz.dll}, the definition
25301 file used is @code{xyz.def}.
25303 @geindex -e (gnatdll)
25305 @item @code{-e @emph{deffile}}
25307 @code{deffile} is the name of the definition file.
25309 @geindex -g (gnatdll)
25313 Generate debugging information. This information is stored in the object
25314 file and copied from there to the final DLL file by the linker,
25315 where it can be read by the debugger. You must use the
25316 @code{-g} switch if you plan on using the debugger or the symbolic
25319 @geindex -h (gnatdll)
25323 Help mode. Displays @code{gnatdll} switch usage information.
25325 @geindex -I (gnatdll)
25327 @item @code{-I@emph{dir}}
25329 Direct @code{gnatdll} to search the @code{dir} directory for source and
25330 object files needed to build the DLL.
25331 (@ref{89,,Search Paths and the Run-Time Library (RTL)}).
25333 @geindex -k (gnatdll)
25337 Removes the @code{@@@emph{nn}} suffix from the import library's exported
25338 names, but keeps them for the link names. You must specify this
25339 option if you want to use a @code{Stdcall} function in a DLL for which
25340 the @code{@@@emph{nn}} suffix has been removed. This is the case for most
25341 of the Windows NT DLL for example. This option has no effect when
25342 @code{-n} option is specified.
25344 @geindex -l (gnatdll)
25346 @item @code{-l @emph{file}}
25348 The list of ALI and object files used to build the DLL are listed in
25349 @code{file}, instead of being given in the command line. Each line in
25350 @code{file} contains the name of an ALI or object file.
25352 @geindex -n (gnatdll)
25356 No Import. Do not create the import library.
25358 @geindex -q (gnatdll)
25362 Quiet mode. Do not display unnecessary messages.
25364 @geindex -v (gnatdll)
25368 Verbose mode. Display extra information.
25370 @geindex -largs (gnatdll)
25372 @item @code{-largs @emph{opts}}
25374 Linker options. Pass @code{opts} to the linker.
25377 @subsubheading @code{gnatdll} Example
25380 As an example the command to build a relocatable DLL from @code{api.adb}
25381 once @code{api.adb} has been compiled and @code{api.def} created is
25386 $ gnatdll -d api.dll api.ali
25390 The above command creates two files: @code{libapi.dll.a} (the import
25391 library) and @code{api.dll} (the actual DLL). If you want to create
25392 only the DLL, just type:
25397 $ gnatdll -d api.dll -n api.ali
25401 Alternatively if you want to create just the import library, type:
25406 $ gnatdll -d api.dll
25410 @subsubheading @code{gnatdll} behind the Scenes
25413 This section details the steps involved in creating a DLL. @code{gnatdll}
25414 does these steps for you. Unless you are interested in understanding what
25415 goes on behind the scenes, you should skip this section.
25417 We use the previous example of a DLL containing the Ada package @code{API},
25418 to illustrate the steps necessary to build a DLL. The starting point is a
25419 set of objects that will make up the DLL and the corresponding ALI
25420 files. In the case of this example this means that @code{api.o} and
25421 @code{api.ali} are available. To build a relocatable DLL, @code{gnatdll} does
25428 @code{gnatdll} builds the base file (@code{api.base}). A base file gives
25429 the information necessary to generate relocation information for the
25434 $ gnatlink api -o api.jnk -mdll -Wl,--base-file,api.base
25437 In addition to the base file, the @code{gnatlink} command generates an
25438 output file @code{api.jnk} which can be discarded. The @code{-mdll} switch
25439 asks @code{gnatlink} to generate the routines @code{DllMain} and
25440 @code{DllMainCRTStartup} that are called by the Windows loader when the DLL
25441 is loaded into memory.
25444 @code{gnatdll} uses @code{dlltool} (see @ref{20d,,Using dlltool}) to build the
25445 export table (@code{api.exp}). The export table contains the relocation
25446 information in a form which can be used during the final link to ensure
25447 that the Windows loader is able to place the DLL anywhere in memory.
25450 $ dlltool --dllname api.dll --def api.def --base-file api.base \\
25451 --output-exp api.exp
25455 @code{gnatdll} builds the base file using the new export table. Note that
25456 @code{gnatbind} must be called once again since the binder generated file
25457 has been deleted during the previous call to @code{gnatlink}.
25461 $ gnatlink api -o api.jnk api.exp -mdll
25462 -Wl,--base-file,api.base
25466 @code{gnatdll} builds the new export table using the new base file and
25467 generates the DLL import library @code{libAPI.dll.a}.
25470 $ dlltool --dllname api.dll --def api.def --base-file api.base \\
25471 --output-exp api.exp --output-lib libAPI.a
25475 Finally @code{gnatdll} builds the relocatable DLL using the final export
25480 $ gnatlink api api.exp -o api.dll -mdll
25483 @anchor{gnat_ugn/platform_specific_information using-dlltool}@anchor{20d}
25484 @subsubheading Using @code{dlltool}
25487 @code{dlltool} is the low-level tool used by @code{gnatdll} to build
25488 DLLs and static import libraries. This section summarizes the most
25489 common @code{dlltool} switches. The form of the @code{dlltool} command
25495 $ dlltool [`switches`]
25499 @code{dlltool} switches include:
25501 @geindex --base-file (dlltool)
25506 @item @code{--base-file @emph{basefile}}
25508 Read the base file @code{basefile} generated by the linker. This switch
25509 is used to create a relocatable DLL.
25512 @geindex --def (dlltool)
25517 @item @code{--def @emph{deffile}}
25519 Read the definition file.
25522 @geindex --dllname (dlltool)
25527 @item @code{--dllname @emph{name}}
25529 Gives the name of the DLL. This switch is used to embed the name of the
25530 DLL in the static import library generated by @code{dlltool} with switch
25531 @code{--output-lib}.
25534 @geindex -k (dlltool)
25541 Kill @code{@@@emph{nn}} from exported names
25542 (@ref{1e6,,Windows Calling Conventions}
25543 for a discussion about @code{Stdcall}-style symbols.
25546 @geindex --help (dlltool)
25551 @item @code{--help}
25553 Prints the @code{dlltool} switches with a concise description.
25556 @geindex --output-exp (dlltool)
25561 @item @code{--output-exp @emph{exportfile}}
25563 Generate an export file @code{exportfile}. The export file contains the
25564 export table (list of symbols in the DLL) and is used to create the DLL.
25567 @geindex --output-lib (dlltool)
25572 @item @code{--output-lib @emph{libfile}}
25574 Generate a static import library @code{libfile}.
25577 @geindex -v (dlltool)
25587 @geindex --as (dlltool)
25592 @item @code{--as @emph{assembler-name}}
25594 Use @code{assembler-name} as the assembler. The default is @code{as}.
25597 @node GNAT and Windows Resources,Using GNAT DLLs from Microsoft Visual Studio Applications,Creating a Spec for Ada DLLs,Mixed-Language Programming on Windows
25598 @anchor{gnat_ugn/platform_specific_information gnat-and-windows-resources}@anchor{20e}@anchor{gnat_ugn/platform_specific_information id32}@anchor{20f}
25599 @subsubsection GNAT and Windows Resources
25605 Resources are an easy way to add Windows specific objects to your
25606 application. The objects that can be added as resources include:
25636 version information
25639 For example, a version information resource can be defined as follow and
25640 embedded into an executable or DLL:
25642 A version information resource can be used to embed information into an
25643 executable or a DLL. These information can be viewed using the file properties
25644 from the Windows Explorer. Here is an example of a version information
25651 FILEVERSION 1,0,0,0
25652 PRODUCTVERSION 1,0,0,0
25654 BLOCK "StringFileInfo"
25658 VALUE "CompanyName", "My Company Name"
25659 VALUE "FileDescription", "My application"
25660 VALUE "FileVersion", "1.0"
25661 VALUE "InternalName", "my_app"
25662 VALUE "LegalCopyright", "My Name"
25663 VALUE "OriginalFilename", "my_app.exe"
25664 VALUE "ProductName", "My App"
25665 VALUE "ProductVersion", "1.0"
25669 BLOCK "VarFileInfo"
25671 VALUE "Translation", 0x809, 1252
25677 The value @code{0809} (langID) is for the U.K English language and
25678 @code{04E4} (charsetID), which is equal to @code{1252} decimal, for
25681 This section explains how to build, compile and use resources. Note that this
25682 section does not cover all resource objects, for a complete description see
25683 the corresponding Microsoft documentation.
25686 * Building Resources::
25687 * Compiling Resources::
25688 * Using Resources::
25692 @node Building Resources,Compiling Resources,,GNAT and Windows Resources
25693 @anchor{gnat_ugn/platform_specific_information building-resources}@anchor{210}@anchor{gnat_ugn/platform_specific_information id33}@anchor{211}
25694 @subsubsection Building Resources
25700 A resource file is an ASCII file. By convention resource files have an
25701 @code{.rc} extension.
25702 The easiest way to build a resource file is to use Microsoft tools
25703 such as @code{imagedit.exe} to build bitmaps, icons and cursors and
25704 @code{dlgedit.exe} to build dialogs.
25705 It is always possible to build an @code{.rc} file yourself by writing a
25708 It is not our objective to explain how to write a resource file. A
25709 complete description of the resource script language can be found in the
25710 Microsoft documentation.
25712 @node Compiling Resources,Using Resources,Building Resources,GNAT and Windows Resources
25713 @anchor{gnat_ugn/platform_specific_information compiling-resources}@anchor{212}@anchor{gnat_ugn/platform_specific_information id34}@anchor{213}
25714 @subsubsection Compiling Resources
25724 This section describes how to build a GNAT-compatible (COFF) object file
25725 containing the resources. This is done using the Resource Compiler
25726 @code{windres} as follows:
25731 $ windres -i myres.rc -o myres.o
25735 By default @code{windres} will run @code{gcc} to preprocess the @code{.rc}
25736 file. You can specify an alternate preprocessor (usually named
25737 @code{cpp.exe}) using the @code{windres} @code{--preprocessor}
25738 parameter. A list of all possible options may be obtained by entering
25739 the command @code{windres} @code{--help}.
25741 It is also possible to use the Microsoft resource compiler @code{rc.exe}
25742 to produce a @code{.res} file (binary resource file). See the
25743 corresponding Microsoft documentation for further details. In this case
25744 you need to use @code{windres} to translate the @code{.res} file to a
25745 GNAT-compatible object file as follows:
25750 $ windres -i myres.res -o myres.o
25754 @node Using Resources,,Compiling Resources,GNAT and Windows Resources
25755 @anchor{gnat_ugn/platform_specific_information using-resources}@anchor{214}@anchor{gnat_ugn/platform_specific_information id35}@anchor{215}
25756 @subsubsection Using Resources
25762 To include the resource file in your program just add the
25763 GNAT-compatible object file for the resource(s) to the linker
25764 arguments. With @code{gnatmake} this is done by using the @code{-largs}
25770 $ gnatmake myprog -largs myres.o
25774 @node Using GNAT DLLs from Microsoft Visual Studio Applications,Debugging a DLL,GNAT and Windows Resources,Mixed-Language Programming on Windows
25775 @anchor{gnat_ugn/platform_specific_information using-gnat-dll-from-msvs}@anchor{216}@anchor{gnat_ugn/platform_specific_information using-gnat-dlls-from-microsoft-visual-studio-applications}@anchor{217}
25776 @subsubsection Using GNAT DLLs from Microsoft Visual Studio Applications
25779 @geindex Microsoft Visual Studio
25780 @geindex use with GNAT DLLs
25782 This section describes a common case of mixed GNAT/Microsoft Visual Studio
25783 application development, where the main program is developed using MSVS, and
25784 is linked with a DLL developed using GNAT. Such a mixed application should
25785 be developed following the general guidelines outlined above; below is the
25786 cookbook-style sequence of steps to follow:
25792 First develop and build the GNAT shared library using a library project
25793 (let's assume the project is @code{mylib.gpr}, producing the library @code{libmylib.dll}):
25799 $ gprbuild -p mylib.gpr
25807 Produce a .def file for the symbols you need to interface with, either by
25808 hand or automatically with possibly some manual adjustments
25809 (see @ref{1f8,,Creating Definition File Automatically}):
25815 $ dlltool libmylib.dll -z libmylib.def --export-all-symbols
25823 Make sure that MSVS command-line tools are accessible on the path.
25826 Create the Microsoft-style import library (see @ref{1fb,,MSVS-Style Import Library}):
25832 $ lib -machine:IX86 -def:libmylib.def -out:libmylib.lib
25836 If you are using a 64-bit toolchain, the above becomes...
25841 $ lib -machine:X64 -def:libmylib.def -out:libmylib.lib
25855 $ cl /O2 /MD main.c libmylib.lib
25863 Before running the executable, make sure you have set the PATH to the DLL,
25864 or copy the DLL into into the directory containing the .exe.
25867 @node Debugging a DLL,Setting Stack Size from gnatlink,Using GNAT DLLs from Microsoft Visual Studio Applications,Mixed-Language Programming on Windows
25868 @anchor{gnat_ugn/platform_specific_information id36}@anchor{218}@anchor{gnat_ugn/platform_specific_information debugging-a-dll}@anchor{219}
25869 @subsubsection Debugging a DLL
25872 @geindex DLL debugging
25874 Debugging a DLL is similar to debugging a standard program. But
25875 we have to deal with two different executable parts: the DLL and the
25876 program that uses it. We have the following four possibilities:
25882 The program and the DLL are built with GCC/GNAT.
25885 The program is built with foreign tools and the DLL is built with
25889 The program is built with GCC/GNAT and the DLL is built with
25893 In this section we address only cases one and two above.
25894 There is no point in trying to debug
25895 a DLL with GNU/GDB, if there is no GDB-compatible debugging
25896 information in it. To do so you must use a debugger compatible with the
25897 tools suite used to build the DLL.
25900 * Program and DLL Both Built with GCC/GNAT::
25901 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
25905 @node Program and DLL Both Built with GCC/GNAT,Program Built with Foreign Tools and DLL Built with GCC/GNAT,,Debugging a DLL
25906 @anchor{gnat_ugn/platform_specific_information id37}@anchor{21a}@anchor{gnat_ugn/platform_specific_information program-and-dll-both-built-with-gcc-gnat}@anchor{21b}
25907 @subsubsection Program and DLL Both Built with GCC/GNAT
25910 This is the simplest case. Both the DLL and the program have @code{GDB}
25911 compatible debugging information. It is then possible to break anywhere in
25912 the process. Let's suppose here that the main procedure is named
25913 @code{ada_main} and that in the DLL there is an entry point named
25916 The DLL (@ref{1f1,,Introduction to Dynamic Link Libraries (DLLs)}) and
25917 program must have been built with the debugging information (see GNAT -g
25918 switch). Here are the step-by-step instructions for debugging it:
25924 Launch @code{GDB} on the main program.
25931 Start the program and stop at the beginning of the main procedure
25937 This step is required to be able to set a breakpoint inside the DLL. As long
25938 as the program is not run, the DLL is not loaded. This has the
25939 consequence that the DLL debugging information is also not loaded, so it is not
25940 possible to set a breakpoint in the DLL.
25943 Set a breakpoint inside the DLL
25946 (gdb) break ada_dll
25951 At this stage a breakpoint is set inside the DLL. From there on
25952 you can use the standard approach to debug the whole program
25953 (@ref{24,,Running and Debugging Ada Programs}).
25955 @node Program Built with Foreign Tools and DLL Built with GCC/GNAT,,Program and DLL Both Built with GCC/GNAT,Debugging a DLL
25956 @anchor{gnat_ugn/platform_specific_information program-built-with-foreign-tools-and-dll-built-with-gcc-gnat}@anchor{21c}@anchor{gnat_ugn/platform_specific_information id38}@anchor{21d}
25957 @subsubsection Program Built with Foreign Tools and DLL Built with GCC/GNAT
25960 In this case things are slightly more complex because it is not possible to
25961 start the main program and then break at the beginning to load the DLL and the
25962 associated DLL debugging information. It is not possible to break at the
25963 beginning of the program because there is no @code{GDB} debugging information,
25964 and therefore there is no direct way of getting initial control. This
25965 section addresses this issue by describing some methods that can be used
25966 to break somewhere in the DLL to debug it.
25968 First suppose that the main procedure is named @code{main} (this is for
25969 example some C code built with Microsoft Visual C) and that there is a
25970 DLL named @code{test.dll} containing an Ada entry point named
25973 The DLL (see @ref{1f1,,Introduction to Dynamic Link Libraries (DLLs)}) must have
25974 been built with debugging information (see the GNAT @code{-g} option).
25976 @subsubheading Debugging the DLL Directly
25983 Find out the executable starting address
25986 $ objdump --file-header main.exe
25989 The starting address is reported on the last line. For example:
25992 main.exe: file format pei-i386
25993 architecture: i386, flags 0x0000010a:
25994 EXEC_P, HAS_DEBUG, D_PAGED
25995 start address 0x00401010
25999 Launch the debugger on the executable.
26006 Set a breakpoint at the starting address, and launch the program.
26009 $ (gdb) break *0x00401010
26013 The program will stop at the given address.
26016 Set a breakpoint on a DLL subroutine.
26019 (gdb) break ada_dll.adb:45
26022 Or if you want to break using a symbol on the DLL, you need first to
26023 select the Ada language (language used by the DLL).
26026 (gdb) set language ada
26027 (gdb) break ada_dll
26031 Continue the program.
26037 This will run the program until it reaches the breakpoint that has been
26038 set. From that point you can use the standard way to debug a program
26039 as described in (@ref{24,,Running and Debugging Ada Programs}).
26042 It is also possible to debug the DLL by attaching to a running process.
26044 @subsubheading Attaching to a Running Process
26047 @geindex DLL debugging
26048 @geindex attach to process
26050 With @code{GDB} it is always possible to debug a running process by
26051 attaching to it. It is possible to debug a DLL this way. The limitation
26052 of this approach is that the DLL must run long enough to perform the
26053 attach operation. It may be useful for instance to insert a time wasting
26054 loop in the code of the DLL to meet this criterion.
26060 Launch the main program @code{main.exe}.
26067 Use the Windows @emph{Task Manager} to find the process ID. Let's say
26068 that the process PID for @code{main.exe} is 208.
26078 Attach to the running process to be debugged.
26085 Load the process debugging information.
26088 (gdb) symbol-file main.exe
26092 Break somewhere in the DLL.
26095 (gdb) break ada_dll
26099 Continue process execution.
26106 This last step will resume the process execution, and stop at
26107 the breakpoint we have set. From there you can use the standard
26108 approach to debug a program as described in
26109 @ref{24,,Running and Debugging Ada Programs}.
26111 @node Setting Stack Size from gnatlink,Setting Heap Size from gnatlink,Debugging a DLL,Mixed-Language Programming on Windows
26112 @anchor{gnat_ugn/platform_specific_information setting-stack-size-from-gnatlink}@anchor{136}@anchor{gnat_ugn/platform_specific_information id39}@anchor{21e}
26113 @subsubsection Setting Stack Size from @code{gnatlink}
26116 It is possible to specify the program stack size at link time. On modern
26117 versions of Windows, starting with XP, this is mostly useful to set the size of
26118 the main stack (environment task). The other task stacks are set with pragma
26119 Storage_Size or with the @emph{gnatbind -d} command.
26121 Since older versions of Windows (2000, NT4, etc.) do not allow setting the
26122 reserve size of individual tasks, the link-time stack size applies to all
26123 tasks, and pragma Storage_Size has no effect.
26124 In particular, Stack Overflow checks are made against this
26125 link-time specified size.
26127 This setting can be done with @code{gnatlink} using either of the following:
26133 @code{-Xlinker} linker option
26136 $ gnatlink hello -Xlinker --stack=0x10000,0x1000
26139 This sets the stack reserve size to 0x10000 bytes and the stack commit
26140 size to 0x1000 bytes.
26143 @code{-Wl} linker option
26146 $ gnatlink hello -Wl,--stack=0x1000000
26149 This sets the stack reserve size to 0x1000000 bytes. Note that with
26150 @code{-Wl} option it is not possible to set the stack commit size
26151 because the comma is a separator for this option.
26154 @node Setting Heap Size from gnatlink,,Setting Stack Size from gnatlink,Mixed-Language Programming on Windows
26155 @anchor{gnat_ugn/platform_specific_information setting-heap-size-from-gnatlink}@anchor{137}@anchor{gnat_ugn/platform_specific_information id40}@anchor{21f}
26156 @subsubsection Setting Heap Size from @code{gnatlink}
26159 Under Windows systems, it is possible to specify the program heap size from
26160 @code{gnatlink} using either of the following:
26166 @code{-Xlinker} linker option
26169 $ gnatlink hello -Xlinker --heap=0x10000,0x1000
26172 This sets the heap reserve size to 0x10000 bytes and the heap commit
26173 size to 0x1000 bytes.
26176 @code{-Wl} linker option
26179 $ gnatlink hello -Wl,--heap=0x1000000
26182 This sets the heap reserve size to 0x1000000 bytes. Note that with
26183 @code{-Wl} option it is not possible to set the heap commit size
26184 because the comma is a separator for this option.
26187 @node Windows Specific Add-Ons,,Mixed-Language Programming on Windows,Microsoft Windows Topics
26188 @anchor{gnat_ugn/platform_specific_information windows-specific-add-ons}@anchor{220}@anchor{gnat_ugn/platform_specific_information win32-specific-addons}@anchor{221}
26189 @subsection Windows Specific Add-Ons
26192 This section describes the Windows specific add-ons.
26200 @node Win32Ada,wPOSIX,,Windows Specific Add-Ons
26201 @anchor{gnat_ugn/platform_specific_information win32ada}@anchor{222}@anchor{gnat_ugn/platform_specific_information id41}@anchor{223}
26202 @subsubsection Win32Ada
26205 Win32Ada is a binding for the Microsoft Win32 API. This binding can be
26206 easily installed from the provided installer. To use the Win32Ada
26207 binding you need to use a project file, and adding a single with_clause
26208 will give you full access to the Win32Ada binding sources and ensure
26209 that the proper libraries are passed to the linker.
26216 for Sources use ...;
26221 To build the application you just need to call gprbuild for the
26222 application's project, here p.gpr:
26231 @node wPOSIX,,Win32Ada,Windows Specific Add-Ons
26232 @anchor{gnat_ugn/platform_specific_information id42}@anchor{224}@anchor{gnat_ugn/platform_specific_information wposix}@anchor{225}
26233 @subsubsection wPOSIX
26236 wPOSIX is a minimal POSIX binding whose goal is to help with building
26237 cross-platforms applications. This binding is not complete though, as
26238 the Win32 API does not provide the necessary support for all POSIX APIs.
26240 To use the wPOSIX binding you need to use a project file, and adding
26241 a single with_clause will give you full access to the wPOSIX binding
26242 sources and ensure that the proper libraries are passed to the linker.
26249 for Sources use ...;
26254 To build the application you just need to call gprbuild for the
26255 application's project, here p.gpr:
26264 @node Mac OS Topics,,Microsoft Windows Topics,Platform-Specific Information
26265 @anchor{gnat_ugn/platform_specific_information mac-os-topics}@anchor{2d}@anchor{gnat_ugn/platform_specific_information id43}@anchor{226}
26266 @section Mac OS Topics
26271 This section describes topics that are specific to Apple's OS X
26275 * Codesigning the Debugger::
26279 @node Codesigning the Debugger,,,Mac OS Topics
26280 @anchor{gnat_ugn/platform_specific_information codesigning-the-debugger}@anchor{227}
26281 @subsection Codesigning the Debugger
26284 The Darwin Kernel requires the debugger to have special permissions
26285 before it is allowed to control other processes. These permissions
26286 are granted by codesigning the GDB executable. Without these
26287 permissions, the debugger will report error messages such as:
26290 Starting program: /x/y/foo
26291 Unable to find Mach task port for process-id 28885: (os/kern) failure (0x5).
26292 (please check gdb is codesigned - see taskgated(8))
26295 Codesigning requires a certificate. The following procedure explains
26302 Start the Keychain Access application (in
26303 /Applications/Utilities/Keychain Access.app)
26306 Select the Keychain Access -> Certificate Assistant ->
26307 Create a Certificate... menu
26316 Choose a name for the new certificate (this procedure will use
26317 "gdb-cert" as an example)
26320 Set "Identity Type" to "Self Signed Root"
26323 Set "Certificate Type" to "Code Signing"
26326 Activate the "Let me override defaults" option
26330 Click several times on "Continue" until the "Specify a Location
26331 For The Certificate" screen appears, then set "Keychain" to "System"
26334 Click on "Continue" until the certificate is created
26337 Finally, in the view, double-click on the new certificate,
26338 and set "When using this certificate" to "Always Trust"
26341 Exit the Keychain Access application and restart the computer
26342 (this is unfortunately required)
26345 Once a certificate has been created, the debugger can be codesigned
26346 as follow. In a Terminal, run the following command:
26351 $ codesign -f -s "gdb-cert" <gnat_install_prefix>/bin/gdb
26355 where "gdb-cert" should be replaced by the actual certificate
26356 name chosen above, and <gnat_install_prefix> should be replaced by
26357 the location where you installed GNAT. Also, be sure that users are
26358 in the Unix group @code{_developer}.
26360 @node Example of Binder Output File,Elaboration Order Handling in GNAT,Platform-Specific Information,Top
26361 @anchor{gnat_ugn/example_of_binder_output example-of-binder-output-file}@anchor{e}@anchor{gnat_ugn/example_of_binder_output doc}@anchor{228}@anchor{gnat_ugn/example_of_binder_output id1}@anchor{229}
26362 @chapter Example of Binder Output File
26365 @geindex Binder output (example)
26367 This Appendix displays the source code for the output file
26368 generated by @emph{gnatbind} for a simple 'Hello World' program.
26369 Comments have been added for clarification purposes.
26372 -- The package is called Ada_Main unless this name is actually used
26373 -- as a unit name in the partition, in which case some other unique
26378 package ada_main is
26379 pragma Warnings (Off);
26381 -- The main program saves the parameters (argument count,
26382 -- argument values, environment pointer) in global variables
26383 -- for later access by other units including
26384 -- Ada.Command_Line.
26386 gnat_argc : Integer;
26387 gnat_argv : System.Address;
26388 gnat_envp : System.Address;
26390 -- The actual variables are stored in a library routine. This
26391 -- is useful for some shared library situations, where there
26392 -- are problems if variables are not in the library.
26394 pragma Import (C, gnat_argc);
26395 pragma Import (C, gnat_argv);
26396 pragma Import (C, gnat_envp);
26398 -- The exit status is similarly an external location
26400 gnat_exit_status : Integer;
26401 pragma Import (C, gnat_exit_status);
26403 GNAT_Version : constant String :=
26404 "GNAT Version: Pro 7.4.0w (20141119-49)" & ASCII.NUL;
26405 pragma Export (C, GNAT_Version, "__gnat_version");
26407 Ada_Main_Program_Name : constant String := "_ada_hello" & ASCII.NUL;
26408 pragma Export (C, Ada_Main_Program_Name, "__gnat_ada_main_program_name");
26410 -- This is the generated adainit routine that performs
26411 -- initialization at the start of execution. In the case
26412 -- where Ada is the main program, this main program makes
26413 -- a call to adainit at program startup.
26416 pragma Export (C, adainit, "adainit");
26418 -- This is the generated adafinal routine that performs
26419 -- finalization at the end of execution. In the case where
26420 -- Ada is the main program, this main program makes a call
26421 -- to adafinal at program termination.
26423 procedure adafinal;
26424 pragma Export (C, adafinal, "adafinal");
26426 -- This routine is called at the start of execution. It is
26427 -- a dummy routine that is used by the debugger to breakpoint
26428 -- at the start of execution.
26430 -- This is the actual generated main program (it would be
26431 -- suppressed if the no main program switch were used). As
26432 -- required by standard system conventions, this program has
26433 -- the external name main.
26437 argv : System.Address;
26438 envp : System.Address)
26440 pragma Export (C, main, "main");
26442 -- The following set of constants give the version
26443 -- identification values for every unit in the bound
26444 -- partition. This identification is computed from all
26445 -- dependent semantic units, and corresponds to the
26446 -- string that would be returned by use of the
26447 -- Body_Version or Version attributes.
26449 -- The following Export pragmas export the version numbers
26450 -- with symbolic names ending in B (for body) or S
26451 -- (for spec) so that they can be located in a link. The
26452 -- information provided here is sufficient to track down
26453 -- the exact versions of units used in a given build.
26455 type Version_32 is mod 2 ** 32;
26456 u00001 : constant Version_32 := 16#8ad6e54a#;
26457 pragma Export (C, u00001, "helloB");
26458 u00002 : constant Version_32 := 16#fbff4c67#;
26459 pragma Export (C, u00002, "system__standard_libraryB");
26460 u00003 : constant Version_32 := 16#1ec6fd90#;
26461 pragma Export (C, u00003, "system__standard_libraryS");
26462 u00004 : constant Version_32 := 16#3ffc8e18#;
26463 pragma Export (C, u00004, "adaS");
26464 u00005 : constant Version_32 := 16#28f088c2#;
26465 pragma Export (C, u00005, "ada__text_ioB");
26466 u00006 : constant Version_32 := 16#f372c8ac#;
26467 pragma Export (C, u00006, "ada__text_ioS");
26468 u00007 : constant Version_32 := 16#2c143749#;
26469 pragma Export (C, u00007, "ada__exceptionsB");
26470 u00008 : constant Version_32 := 16#f4f0cce8#;
26471 pragma Export (C, u00008, "ada__exceptionsS");
26472 u00009 : constant Version_32 := 16#a46739c0#;
26473 pragma Export (C, u00009, "ada__exceptions__last_chance_handlerB");
26474 u00010 : constant Version_32 := 16#3aac8c92#;
26475 pragma Export (C, u00010, "ada__exceptions__last_chance_handlerS");
26476 u00011 : constant Version_32 := 16#1d274481#;
26477 pragma Export (C, u00011, "systemS");
26478 u00012 : constant Version_32 := 16#a207fefe#;
26479 pragma Export (C, u00012, "system__soft_linksB");
26480 u00013 : constant Version_32 := 16#467d9556#;
26481 pragma Export (C, u00013, "system__soft_linksS");
26482 u00014 : constant Version_32 := 16#b01dad17#;
26483 pragma Export (C, u00014, "system__parametersB");
26484 u00015 : constant Version_32 := 16#630d49fe#;
26485 pragma Export (C, u00015, "system__parametersS");
26486 u00016 : constant Version_32 := 16#b19b6653#;
26487 pragma Export (C, u00016, "system__secondary_stackB");
26488 u00017 : constant Version_32 := 16#b6468be8#;
26489 pragma Export (C, u00017, "system__secondary_stackS");
26490 u00018 : constant Version_32 := 16#39a03df9#;
26491 pragma Export (C, u00018, "system__storage_elementsB");
26492 u00019 : constant Version_32 := 16#30e40e85#;
26493 pragma Export (C, u00019, "system__storage_elementsS");
26494 u00020 : constant Version_32 := 16#41837d1e#;
26495 pragma Export (C, u00020, "system__stack_checkingB");
26496 u00021 : constant Version_32 := 16#93982f69#;
26497 pragma Export (C, u00021, "system__stack_checkingS");
26498 u00022 : constant Version_32 := 16#393398c1#;
26499 pragma Export (C, u00022, "system__exception_tableB");
26500 u00023 : constant Version_32 := 16#b33e2294#;
26501 pragma Export (C, u00023, "system__exception_tableS");
26502 u00024 : constant Version_32 := 16#ce4af020#;
26503 pragma Export (C, u00024, "system__exceptionsB");
26504 u00025 : constant Version_32 := 16#75442977#;
26505 pragma Export (C, u00025, "system__exceptionsS");
26506 u00026 : constant Version_32 := 16#37d758f1#;
26507 pragma Export (C, u00026, "system__exceptions__machineS");
26508 u00027 : constant Version_32 := 16#b895431d#;
26509 pragma Export (C, u00027, "system__exceptions_debugB");
26510 u00028 : constant Version_32 := 16#aec55d3f#;
26511 pragma Export (C, u00028, "system__exceptions_debugS");
26512 u00029 : constant Version_32 := 16#570325c8#;
26513 pragma Export (C, u00029, "system__img_intB");
26514 u00030 : constant Version_32 := 16#1ffca443#;
26515 pragma Export (C, u00030, "system__img_intS");
26516 u00031 : constant Version_32 := 16#b98c3e16#;
26517 pragma Export (C, u00031, "system__tracebackB");
26518 u00032 : constant Version_32 := 16#831a9d5a#;
26519 pragma Export (C, u00032, "system__tracebackS");
26520 u00033 : constant Version_32 := 16#9ed49525#;
26521 pragma Export (C, u00033, "system__traceback_entriesB");
26522 u00034 : constant Version_32 := 16#1d7cb2f1#;
26523 pragma Export (C, u00034, "system__traceback_entriesS");
26524 u00035 : constant Version_32 := 16#8c33a517#;
26525 pragma Export (C, u00035, "system__wch_conB");
26526 u00036 : constant Version_32 := 16#065a6653#;
26527 pragma Export (C, u00036, "system__wch_conS");
26528 u00037 : constant Version_32 := 16#9721e840#;
26529 pragma Export (C, u00037, "system__wch_stwB");
26530 u00038 : constant Version_32 := 16#2b4b4a52#;
26531 pragma Export (C, u00038, "system__wch_stwS");
26532 u00039 : constant Version_32 := 16#92b797cb#;
26533 pragma Export (C, u00039, "system__wch_cnvB");
26534 u00040 : constant Version_32 := 16#09eddca0#;
26535 pragma Export (C, u00040, "system__wch_cnvS");
26536 u00041 : constant Version_32 := 16#6033a23f#;
26537 pragma Export (C, u00041, "interfacesS");
26538 u00042 : constant Version_32 := 16#ece6fdb6#;
26539 pragma Export (C, u00042, "system__wch_jisB");
26540 u00043 : constant Version_32 := 16#899dc581#;
26541 pragma Export (C, u00043, "system__wch_jisS");
26542 u00044 : constant Version_32 := 16#10558b11#;
26543 pragma Export (C, u00044, "ada__streamsB");
26544 u00045 : constant Version_32 := 16#2e6701ab#;
26545 pragma Export (C, u00045, "ada__streamsS");
26546 u00046 : constant Version_32 := 16#db5c917c#;
26547 pragma Export (C, u00046, "ada__io_exceptionsS");
26548 u00047 : constant Version_32 := 16#12c8cd7d#;
26549 pragma Export (C, u00047, "ada__tagsB");
26550 u00048 : constant Version_32 := 16#ce72c228#;
26551 pragma Export (C, u00048, "ada__tagsS");
26552 u00049 : constant Version_32 := 16#c3335bfd#;
26553 pragma Export (C, u00049, "system__htableB");
26554 u00050 : constant Version_32 := 16#99e5f76b#;
26555 pragma Export (C, u00050, "system__htableS");
26556 u00051 : constant Version_32 := 16#089f5cd0#;
26557 pragma Export (C, u00051, "system__string_hashB");
26558 u00052 : constant Version_32 := 16#3bbb9c15#;
26559 pragma Export (C, u00052, "system__string_hashS");
26560 u00053 : constant Version_32 := 16#807fe041#;
26561 pragma Export (C, u00053, "system__unsigned_typesS");
26562 u00054 : constant Version_32 := 16#d27be59e#;
26563 pragma Export (C, u00054, "system__val_lluB");
26564 u00055 : constant Version_32 := 16#fa8db733#;
26565 pragma Export (C, u00055, "system__val_lluS");
26566 u00056 : constant Version_32 := 16#27b600b2#;
26567 pragma Export (C, u00056, "system__val_utilB");
26568 u00057 : constant Version_32 := 16#b187f27f#;
26569 pragma Export (C, u00057, "system__val_utilS");
26570 u00058 : constant Version_32 := 16#d1060688#;
26571 pragma Export (C, u00058, "system__case_utilB");
26572 u00059 : constant Version_32 := 16#392e2d56#;
26573 pragma Export (C, u00059, "system__case_utilS");
26574 u00060 : constant Version_32 := 16#84a27f0d#;
26575 pragma Export (C, u00060, "interfaces__c_streamsB");
26576 u00061 : constant Version_32 := 16#8bb5f2c0#;
26577 pragma Export (C, u00061, "interfaces__c_streamsS");
26578 u00062 : constant Version_32 := 16#6db6928f#;
26579 pragma Export (C, u00062, "system__crtlS");
26580 u00063 : constant Version_32 := 16#4e6a342b#;
26581 pragma Export (C, u00063, "system__file_ioB");
26582 u00064 : constant Version_32 := 16#ba56a5e4#;
26583 pragma Export (C, u00064, "system__file_ioS");
26584 u00065 : constant Version_32 := 16#b7ab275c#;
26585 pragma Export (C, u00065, "ada__finalizationB");
26586 u00066 : constant Version_32 := 16#19f764ca#;
26587 pragma Export (C, u00066, "ada__finalizationS");
26588 u00067 : constant Version_32 := 16#95817ed8#;
26589 pragma Export (C, u00067, "system__finalization_rootB");
26590 u00068 : constant Version_32 := 16#52d53711#;
26591 pragma Export (C, u00068, "system__finalization_rootS");
26592 u00069 : constant Version_32 := 16#769e25e6#;
26593 pragma Export (C, u00069, "interfaces__cB");
26594 u00070 : constant Version_32 := 16#4a38bedb#;
26595 pragma Export (C, u00070, "interfaces__cS");
26596 u00071 : constant Version_32 := 16#07e6ee66#;
26597 pragma Export (C, u00071, "system__os_libB");
26598 u00072 : constant Version_32 := 16#d7b69782#;
26599 pragma Export (C, u00072, "system__os_libS");
26600 u00073 : constant Version_32 := 16#1a817b8e#;
26601 pragma Export (C, u00073, "system__stringsB");
26602 u00074 : constant Version_32 := 16#639855e7#;
26603 pragma Export (C, u00074, "system__stringsS");
26604 u00075 : constant Version_32 := 16#e0b8de29#;
26605 pragma Export (C, u00075, "system__file_control_blockS");
26606 u00076 : constant Version_32 := 16#b5b2aca1#;
26607 pragma Export (C, u00076, "system__finalization_mastersB");
26608 u00077 : constant Version_32 := 16#69316dc1#;
26609 pragma Export (C, u00077, "system__finalization_mastersS");
26610 u00078 : constant Version_32 := 16#57a37a42#;
26611 pragma Export (C, u00078, "system__address_imageB");
26612 u00079 : constant Version_32 := 16#bccbd9bb#;
26613 pragma Export (C, u00079, "system__address_imageS");
26614 u00080 : constant Version_32 := 16#7268f812#;
26615 pragma Export (C, u00080, "system__img_boolB");
26616 u00081 : constant Version_32 := 16#e8fe356a#;
26617 pragma Export (C, u00081, "system__img_boolS");
26618 u00082 : constant Version_32 := 16#d7aac20c#;
26619 pragma Export (C, u00082, "system__ioB");
26620 u00083 : constant Version_32 := 16#8365b3ce#;
26621 pragma Export (C, u00083, "system__ioS");
26622 u00084 : constant Version_32 := 16#6d4d969a#;
26623 pragma Export (C, u00084, "system__storage_poolsB");
26624 u00085 : constant Version_32 := 16#e87cc305#;
26625 pragma Export (C, u00085, "system__storage_poolsS");
26626 u00086 : constant Version_32 := 16#e34550ca#;
26627 pragma Export (C, u00086, "system__pool_globalB");
26628 u00087 : constant Version_32 := 16#c88d2d16#;
26629 pragma Export (C, u00087, "system__pool_globalS");
26630 u00088 : constant Version_32 := 16#9d39c675#;
26631 pragma Export (C, u00088, "system__memoryB");
26632 u00089 : constant Version_32 := 16#445a22b5#;
26633 pragma Export (C, u00089, "system__memoryS");
26634 u00090 : constant Version_32 := 16#6a859064#;
26635 pragma Export (C, u00090, "system__storage_pools__subpoolsB");
26636 u00091 : constant Version_32 := 16#e3b008dc#;
26637 pragma Export (C, u00091, "system__storage_pools__subpoolsS");
26638 u00092 : constant Version_32 := 16#63f11652#;
26639 pragma Export (C, u00092, "system__storage_pools__subpools__finalizationB");
26640 u00093 : constant Version_32 := 16#fe2f4b3a#;
26641 pragma Export (C, u00093, "system__storage_pools__subpools__finalizationS");
26643 -- BEGIN ELABORATION ORDER
26647 -- system.case_util%s
26648 -- system.case_util%b
26650 -- system.img_bool%s
26651 -- system.img_bool%b
26652 -- system.img_int%s
26653 -- system.img_int%b
26656 -- system.parameters%s
26657 -- system.parameters%b
26659 -- interfaces.c_streams%s
26660 -- interfaces.c_streams%b
26661 -- system.standard_library%s
26662 -- system.exceptions_debug%s
26663 -- system.exceptions_debug%b
26664 -- system.storage_elements%s
26665 -- system.storage_elements%b
26666 -- system.stack_checking%s
26667 -- system.stack_checking%b
26668 -- system.string_hash%s
26669 -- system.string_hash%b
26671 -- system.strings%s
26672 -- system.strings%b
26674 -- system.traceback_entries%s
26675 -- system.traceback_entries%b
26676 -- ada.exceptions%s
26677 -- system.soft_links%s
26678 -- system.unsigned_types%s
26679 -- system.val_llu%s
26680 -- system.val_util%s
26681 -- system.val_util%b
26682 -- system.val_llu%b
26683 -- system.wch_con%s
26684 -- system.wch_con%b
26685 -- system.wch_cnv%s
26686 -- system.wch_jis%s
26687 -- system.wch_jis%b
26688 -- system.wch_cnv%b
26689 -- system.wch_stw%s
26690 -- system.wch_stw%b
26691 -- ada.exceptions.last_chance_handler%s
26692 -- ada.exceptions.last_chance_handler%b
26693 -- system.address_image%s
26694 -- system.exception_table%s
26695 -- system.exception_table%b
26696 -- ada.io_exceptions%s
26701 -- system.exceptions%s
26702 -- system.exceptions%b
26703 -- system.exceptions.machine%s
26704 -- system.finalization_root%s
26705 -- system.finalization_root%b
26706 -- ada.finalization%s
26707 -- ada.finalization%b
26708 -- system.storage_pools%s
26709 -- system.storage_pools%b
26710 -- system.finalization_masters%s
26711 -- system.storage_pools.subpools%s
26712 -- system.storage_pools.subpools.finalization%s
26713 -- system.storage_pools.subpools.finalization%b
26716 -- system.standard_library%b
26717 -- system.pool_global%s
26718 -- system.pool_global%b
26719 -- system.file_control_block%s
26720 -- system.file_io%s
26721 -- system.secondary_stack%s
26722 -- system.file_io%b
26723 -- system.storage_pools.subpools%b
26724 -- system.finalization_masters%b
26727 -- system.soft_links%b
26729 -- system.secondary_stack%b
26730 -- system.address_image%b
26731 -- system.traceback%s
26732 -- ada.exceptions%b
26733 -- system.traceback%b
26737 -- END ELABORATION ORDER
26744 -- The following source file name pragmas allow the generated file
26745 -- names to be unique for different main programs. They are needed
26746 -- since the package name will always be Ada_Main.
26748 pragma Source_File_Name (ada_main, Spec_File_Name => "b~hello.ads");
26749 pragma Source_File_Name (ada_main, Body_File_Name => "b~hello.adb");
26751 pragma Suppress (Overflow_Check);
26752 with Ada.Exceptions;
26754 -- Generated package body for Ada_Main starts here
26756 package body ada_main is
26757 pragma Warnings (Off);
26759 -- These values are reference counter associated to units which have
26760 -- been elaborated. It is also used to avoid elaborating the
26761 -- same unit twice.
26763 E72 : Short_Integer; pragma Import (Ada, E72, "system__os_lib_E");
26764 E13 : Short_Integer; pragma Import (Ada, E13, "system__soft_links_E");
26765 E23 : Short_Integer; pragma Import (Ada, E23, "system__exception_table_E");
26766 E46 : Short_Integer; pragma Import (Ada, E46, "ada__io_exceptions_E");
26767 E48 : Short_Integer; pragma Import (Ada, E48, "ada__tags_E");
26768 E45 : Short_Integer; pragma Import (Ada, E45, "ada__streams_E");
26769 E70 : Short_Integer; pragma Import (Ada, E70, "interfaces__c_E");
26770 E25 : Short_Integer; pragma Import (Ada, E25, "system__exceptions_E");
26771 E68 : Short_Integer; pragma Import (Ada, E68, "system__finalization_root_E");
26772 E66 : Short_Integer; pragma Import (Ada, E66, "ada__finalization_E");
26773 E85 : Short_Integer; pragma Import (Ada, E85, "system__storage_pools_E");
26774 E77 : Short_Integer; pragma Import (Ada, E77, "system__finalization_masters_E");
26775 E91 : Short_Integer; pragma Import (Ada, E91, "system__storage_pools__subpools_E");
26776 E87 : Short_Integer; pragma Import (Ada, E87, "system__pool_global_E");
26777 E75 : Short_Integer; pragma Import (Ada, E75, "system__file_control_block_E");
26778 E64 : Short_Integer; pragma Import (Ada, E64, "system__file_io_E");
26779 E17 : Short_Integer; pragma Import (Ada, E17, "system__secondary_stack_E");
26780 E06 : Short_Integer; pragma Import (Ada, E06, "ada__text_io_E");
26782 Local_Priority_Specific_Dispatching : constant String := "";
26783 Local_Interrupt_States : constant String := "";
26785 Is_Elaborated : Boolean := False;
26787 procedure finalize_library is
26792 pragma Import (Ada, F1, "ada__text_io__finalize_spec");
26800 pragma Import (Ada, F2, "system__file_io__finalize_body");
26807 pragma Import (Ada, F3, "system__file_control_block__finalize_spec");
26815 pragma Import (Ada, F4, "system__pool_global__finalize_spec");
26821 pragma Import (Ada, F5, "system__storage_pools__subpools__finalize_spec");
26827 pragma Import (Ada, F6, "system__finalization_masters__finalize_spec");
26832 procedure Reraise_Library_Exception_If_Any;
26833 pragma Import (Ada, Reraise_Library_Exception_If_Any, "__gnat_reraise_library_exception_if_any");
26835 Reraise_Library_Exception_If_Any;
26837 end finalize_library;
26843 procedure adainit is
26845 Main_Priority : Integer;
26846 pragma Import (C, Main_Priority, "__gl_main_priority");
26847 Time_Slice_Value : Integer;
26848 pragma Import (C, Time_Slice_Value, "__gl_time_slice_val");
26849 WC_Encoding : Character;
26850 pragma Import (C, WC_Encoding, "__gl_wc_encoding");
26851 Locking_Policy : Character;
26852 pragma Import (C, Locking_Policy, "__gl_locking_policy");
26853 Queuing_Policy : Character;
26854 pragma Import (C, Queuing_Policy, "__gl_queuing_policy");
26855 Task_Dispatching_Policy : Character;
26856 pragma Import (C, Task_Dispatching_Policy, "__gl_task_dispatching_policy");
26857 Priority_Specific_Dispatching : System.Address;
26858 pragma Import (C, Priority_Specific_Dispatching, "__gl_priority_specific_dispatching");
26859 Num_Specific_Dispatching : Integer;
26860 pragma Import (C, Num_Specific_Dispatching, "__gl_num_specific_dispatching");
26861 Main_CPU : Integer;
26862 pragma Import (C, Main_CPU, "__gl_main_cpu");
26863 Interrupt_States : System.Address;
26864 pragma Import (C, Interrupt_States, "__gl_interrupt_states");
26865 Num_Interrupt_States : Integer;
26866 pragma Import (C, Num_Interrupt_States, "__gl_num_interrupt_states");
26867 Unreserve_All_Interrupts : Integer;
26868 pragma Import (C, Unreserve_All_Interrupts, "__gl_unreserve_all_interrupts");
26869 Detect_Blocking : Integer;
26870 pragma Import (C, Detect_Blocking, "__gl_detect_blocking");
26871 Default_Stack_Size : Integer;
26872 pragma Import (C, Default_Stack_Size, "__gl_default_stack_size");
26873 Leap_Seconds_Support : Integer;
26874 pragma Import (C, Leap_Seconds_Support, "__gl_leap_seconds_support");
26876 procedure Runtime_Initialize;
26877 pragma Import (C, Runtime_Initialize, "__gnat_runtime_initialize");
26879 Finalize_Library_Objects : No_Param_Proc;
26880 pragma Import (C, Finalize_Library_Objects, "__gnat_finalize_library_objects");
26882 -- Start of processing for adainit
26886 -- Record various information for this partition. The values
26887 -- are derived by the binder from information stored in the ali
26888 -- files by the compiler.
26890 if Is_Elaborated then
26893 Is_Elaborated := True;
26894 Main_Priority := -1;
26895 Time_Slice_Value := -1;
26896 WC_Encoding := 'b';
26897 Locking_Policy := ' ';
26898 Queuing_Policy := ' ';
26899 Task_Dispatching_Policy := ' ';
26900 Priority_Specific_Dispatching :=
26901 Local_Priority_Specific_Dispatching'Address;
26902 Num_Specific_Dispatching := 0;
26904 Interrupt_States := Local_Interrupt_States'Address;
26905 Num_Interrupt_States := 0;
26906 Unreserve_All_Interrupts := 0;
26907 Detect_Blocking := 0;
26908 Default_Stack_Size := -1;
26909 Leap_Seconds_Support := 0;
26911 Runtime_Initialize;
26913 Finalize_Library_Objects := finalize_library'access;
26915 -- Now we have the elaboration calls for all units in the partition.
26916 -- The Elab_Spec and Elab_Body attributes generate references to the
26917 -- implicit elaboration procedures generated by the compiler for
26918 -- each unit that requires elaboration. Increment a counter of
26919 -- reference for each unit.
26921 System.Soft_Links'Elab_Spec;
26922 System.Exception_Table'Elab_Body;
26924 Ada.Io_Exceptions'Elab_Spec;
26926 Ada.Tags'Elab_Spec;
26927 Ada.Streams'Elab_Spec;
26929 Interfaces.C'Elab_Spec;
26930 System.Exceptions'Elab_Spec;
26932 System.Finalization_Root'Elab_Spec;
26934 Ada.Finalization'Elab_Spec;
26936 System.Storage_Pools'Elab_Spec;
26938 System.Finalization_Masters'Elab_Spec;
26939 System.Storage_Pools.Subpools'Elab_Spec;
26940 System.Pool_Global'Elab_Spec;
26942 System.File_Control_Block'Elab_Spec;
26944 System.File_Io'Elab_Body;
26947 System.Finalization_Masters'Elab_Body;
26950 Ada.Tags'Elab_Body;
26952 System.Soft_Links'Elab_Body;
26954 System.Os_Lib'Elab_Body;
26956 System.Secondary_Stack'Elab_Body;
26958 Ada.Text_Io'Elab_Spec;
26959 Ada.Text_Io'Elab_Body;
26967 procedure adafinal is
26968 procedure s_stalib_adafinal;
26969 pragma Import (C, s_stalib_adafinal, "system__standard_library__adafinal");
26971 procedure Runtime_Finalize;
26972 pragma Import (C, Runtime_Finalize, "__gnat_runtime_finalize");
26975 if not Is_Elaborated then
26978 Is_Elaborated := False;
26983 -- We get to the main program of the partition by using
26984 -- pragma Import because if we try to with the unit and
26985 -- call it Ada style, then not only do we waste time
26986 -- recompiling it, but also, we don't really know the right
26987 -- switches (e.g.@@: identifier character set) to be used
26990 procedure Ada_Main_Program;
26991 pragma Import (Ada, Ada_Main_Program, "_ada_hello");
26997 -- main is actually a function, as in the ANSI C standard,
26998 -- defined to return the exit status. The three parameters
26999 -- are the argument count, argument values and environment
27004 argv : System.Address;
27005 envp : System.Address)
27008 -- The initialize routine performs low level system
27009 -- initialization using a standard library routine which
27010 -- sets up signal handling and performs any other
27011 -- required setup. The routine can be found in file
27014 procedure initialize;
27015 pragma Import (C, initialize, "__gnat_initialize");
27017 -- The finalize routine performs low level system
27018 -- finalization using a standard library routine. The
27019 -- routine is found in file a-final.c and in the standard
27020 -- distribution is a dummy routine that does nothing, so
27021 -- really this is a hook for special user finalization.
27023 procedure finalize;
27024 pragma Import (C, finalize, "__gnat_finalize");
27026 -- The following is to initialize the SEH exceptions
27028 SEH : aliased array (1 .. 2) of Integer;
27030 Ensure_Reference : aliased System.Address := Ada_Main_Program_Name'Address;
27031 pragma Volatile (Ensure_Reference);
27033 -- Start of processing for main
27036 -- Save global variables
27042 -- Call low level system initialization
27044 Initialize (SEH'Address);
27046 -- Call our generated Ada initialization routine
27050 -- Now we call the main program of the partition
27054 -- Perform Ada finalization
27058 -- Perform low level system finalization
27062 -- Return the proper exit status
27063 return (gnat_exit_status);
27066 -- This section is entirely comments, so it has no effect on the
27067 -- compilation of the Ada_Main package. It provides the list of
27068 -- object files and linker options, as well as some standard
27069 -- libraries needed for the link. The gnatlink utility parses
27070 -- this b~hello.adb file to read these comment lines to generate
27071 -- the appropriate command line arguments for the call to the
27072 -- system linker. The BEGIN/END lines are used for sentinels for
27073 -- this parsing operation.
27075 -- The exact file names will of course depend on the environment,
27076 -- host/target and location of files on the host system.
27078 -- BEGIN Object file/option list
27081 -- -L/usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/
27082 -- /usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/libgnat.a
27083 -- END Object file/option list
27088 The Ada code in the above example is exactly what is generated by the
27089 binder. We have added comments to more clearly indicate the function
27090 of each part of the generated @code{Ada_Main} package.
27092 The code is standard Ada in all respects, and can be processed by any
27093 tools that handle Ada. In particular, it is possible to use the debugger
27094 in Ada mode to debug the generated @code{Ada_Main} package. For example,
27095 suppose that for reasons that you do not understand, your program is crashing
27096 during elaboration of the body of @code{Ada.Text_IO}. To locate this bug,
27097 you can place a breakpoint on the call:
27102 Ada.Text_Io'Elab_Body;
27106 and trace the elaboration routine for this package to find out where
27107 the problem might be (more usually of course you would be debugging
27108 elaboration code in your own application).
27110 @c -- Example: A |withing| unit has a |with| clause, it |withs| a |withed| unit
27112 @node Elaboration Order Handling in GNAT,Inline Assembler,Example of Binder Output File,Top
27113 @anchor{gnat_ugn/elaboration_order_handling_in_gnat elaboration-order-handling-in-gnat}@anchor{f}@anchor{gnat_ugn/elaboration_order_handling_in_gnat doc}@anchor{22a}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id1}@anchor{22b}
27114 @chapter Elaboration Order Handling in GNAT
27117 @geindex Order of elaboration
27119 @geindex Elaboration control
27121 This appendix describes the handling of elaboration code in Ada and GNAT, and
27122 discusses how the order of elaboration of program units can be controlled in
27123 GNAT, either automatically or with explicit programming features.
27126 * Elaboration Code::
27127 * Elaboration Order::
27128 * Checking the Elaboration Order::
27129 * Controlling the Elaboration Order in Ada::
27130 * Controlling the Elaboration Order in GNAT::
27131 * Common Elaboration-model Traits::
27132 * Dynamic Elaboration Model in GNAT::
27133 * Static Elaboration Model in GNAT::
27134 * SPARK Elaboration Model in GNAT::
27135 * Legacy Elaboration Model in GNAT::
27136 * Mixing Elaboration Models::
27137 * Elaboration Circularities::
27138 * Resolving Elaboration Circularities::
27139 * Resolving Task Issues::
27140 * Elaboration-related Compiler Switches::
27141 * Summary of Procedures for Elaboration Control::
27142 * Inspecting the Chosen Elaboration Order::
27146 @node Elaboration Code,Elaboration Order,,Elaboration Order Handling in GNAT
27147 @anchor{gnat_ugn/elaboration_order_handling_in_gnat elaboration-code}@anchor{22c}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id2}@anchor{22d}
27148 @section Elaboration Code
27151 Ada defines the term @emph{execution} as the process by which a construct achieves
27152 its run-time effect. This process is also referred to as @strong{elaboration} for
27153 declarations and @emph{evaluation} for expressions.
27155 The execution model in Ada allows for certain sections of an Ada program to be
27156 executed prior to execution of the program itself, primarily with the intent of
27157 initializing data. These sections are referred to as @strong{elaboration code}.
27158 Elaboration code is executed as follows:
27164 All partitions of an Ada program are executed in parallel with one another,
27165 possibly in a separate address space, and possibly on a separate computer.
27168 The execution of a partition involves running the environment task for that
27172 The environment task executes all elaboration code (if available) for all
27173 units within that partition. This code is said to be executed at
27174 @strong{elaboration time}.
27177 The environment task executes the Ada program (if available) for that
27181 In addition to the Ada terminology, this appendix defines the following terms:
27189 A construct that is elaborated or executed by elaboration code is referred to
27190 as an @emph{elaboration scenario} or simply a @strong{scenario}. GNAT recognizes the
27191 following scenarios:
27197 @code{'Access} of entries, operators, and subprograms
27200 Activation of tasks
27203 Calls to entries, operators, and subprograms
27206 Instantiations of generic templates
27212 A construct elaborated by a scenario is referred to as @emph{elaboration target}
27213 or simply @strong{target}. GNAT recognizes the following targets:
27219 For @code{'Access} of entries, operators, and subprograms, the target is the
27220 entry, operator, or subprogram being aliased.
27223 For activation of tasks, the target is the task body
27226 For calls to entries, operators, and subprograms, the target is the entry,
27227 operator, or subprogram being invoked.
27230 For instantiations of generic templates, the target is the generic template
27231 being instantiated.
27235 Elaboration code may appear in two distinct contexts:
27241 @emph{Library level}
27243 A scenario appears at the library level when it is encapsulated by a package
27244 [body] compilation unit, ignoring any other package [body] declarations in
27253 Val : ... := Server.Func;
27258 In the example above, the call to @code{Server.Func} is an elaboration scenario
27259 because it appears at the library level of package @code{Client}. Note that the
27260 declaration of package @code{Nested} is ignored according to the definition
27261 given above. As a result, the call to @code{Server.Func} will be executed when
27262 the spec of unit @code{Client} is elaborated.
27265 @emph{Package body statements}
27267 A scenario appears within the statement sequence of a package body when it is
27268 bounded by the region starting from the @code{begin} keyword of the package body
27269 and ending at the @code{end} keyword of the package body.
27272 package body Client is
27282 In the example above, the call to @code{Proc} is an elaboration scenario because
27283 it appears within the statement sequence of package body @code{Client}. As a
27284 result, the call to @code{Proc} will be executed when the body of @code{Client} is
27288 @node Elaboration Order,Checking the Elaboration Order,Elaboration Code,Elaboration Order Handling in GNAT
27289 @anchor{gnat_ugn/elaboration_order_handling_in_gnat elaboration-order}@anchor{22e}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id3}@anchor{22f}
27290 @section Elaboration Order
27293 The sequence by which the elaboration code of all units within a partition is
27294 executed is referred to as @strong{elaboration order}.
27296 Within a single unit, elaboration code is executed in sequential order.
27299 package body Client is
27300 Result : ... := Server.Func;
27303 package Inst is new Server.Gen;
27305 Inst.Eval (Result);
27312 In the example above, the elaboration order within package body @code{Client} is
27319 The object declaration of @code{Result} is elaborated.
27325 Function @code{Server.Func} is invoked.
27329 The subprogram body of @code{Proc} is elaborated.
27332 Procedure @code{Proc} is invoked.
27338 Generic unit @code{Server.Gen} is instantiated as @code{Inst}.
27341 Instance @code{Inst} is elaborated.
27344 Procedure @code{Inst.Eval} is invoked.
27348 The elaboration order of all units within a partition depends on the following
27355 @emph{with}ed units
27361 preelaborability of units
27364 presence of elaboration control pragmas
27367 A program may have several elaboration orders depending on its structure.
27371 function Func (Index : Integer) return Integer;
27376 package body Server is
27377 Results : array (1 .. 5) of Integer := (1, 2, 3, 4, 5);
27379 function Func (Index : Integer) return Integer is
27381 return Results (Index);
27389 Val : constant Integer := Server.Func (3);
27395 procedure Main is begin null; end Main;
27398 The following elaboration order exhibits a fundamental problem referred to as
27399 @emph{access-before-elaboration} or simply @strong{ABE}.
27408 The elaboration of @code{Server}'s spec materializes function @code{Func}, making it
27409 callable. The elaboration of @code{Client}'s spec elaborates the declaration of
27410 @code{Val}. This invokes function @code{Server.Func}, however the body of
27411 @code{Server.Func} has not been elaborated yet because @code{Server}'s body comes
27412 after @code{Client}'s spec in the elaboration order. As a result, the value of
27413 constant @code{Val} is now undefined.
27415 Without any guarantees from the language, an undetected ABE problem may hinder
27416 proper initialization of data, which in turn may lead to undefined behavior at
27417 run time. To prevent such ABE problems, Ada employs dynamic checks in the same
27418 vein as index or null exclusion checks. A failed ABE check raises exception
27419 @code{Program_Error}.
27421 The following elaboration order avoids the ABE problem and the program can be
27422 successfully elaborated.
27431 Ada states that a total elaboration order must exist, but it does not define
27432 what this order is. A compiler is thus tasked with choosing a suitable
27433 elaboration order which satisfies the dependencies imposed by @emph{with} clauses,
27434 unit categorization, and elaboration control pragmas. Ideally an order which
27435 avoids ABE problems should be chosen, however a compiler may not always find
27436 such an order due to complications with respect to control and data flow.
27438 @node Checking the Elaboration Order,Controlling the Elaboration Order in Ada,Elaboration Order,Elaboration Order Handling in GNAT
27439 @anchor{gnat_ugn/elaboration_order_handling_in_gnat id4}@anchor{230}@anchor{gnat_ugn/elaboration_order_handling_in_gnat checking-the-elaboration-order}@anchor{231}
27440 @section Checking the Elaboration Order
27443 To avoid placing the entire elaboration order burden on the programmer, Ada
27444 provides three lines of defense:
27450 @emph{Static semantics}
27452 Static semantic rules restrict the possible choice of elaboration order. For
27453 instance, if unit Client @emph{with}s unit Server, then the spec of Server is
27454 always elaborated prior to Client. The same principle applies to child units
27455 - the spec of a parent unit is always elaborated prior to the child unit.
27458 @emph{Dynamic semantics}
27460 Dynamic checks are performed at run time, to ensure that a target is
27461 elaborated prior to a scenario that executes it, thus avoiding ABE problems.
27462 A failed run-time check raises exception @code{Program_Error}. The following
27463 restrictions apply:
27469 @emph{Restrictions on calls}
27471 An entry, operator, or subprogram can be called from elaboration code only
27472 when the corresponding body has been elaborated.
27475 @emph{Restrictions on instantiations}
27477 A generic unit can be instantiated by elaboration code only when the
27478 corresponding body has been elaborated.
27481 @emph{Restrictions on task activation}
27483 A task can be activated by elaboration code only when the body of the
27484 associated task type has been elaborated.
27487 The restrictions above can be summarized by the following rule:
27489 @emph{If a target has a body, then this body must be elaborated prior to the
27490 execution of the scenario that invokes, instantiates, or activates the
27494 @emph{Elaboration control}
27496 Pragmas are provided for the programmer to specify the desired elaboration
27500 @node Controlling the Elaboration Order in Ada,Controlling the Elaboration Order in GNAT,Checking the Elaboration Order,Elaboration Order Handling in GNAT
27501 @anchor{gnat_ugn/elaboration_order_handling_in_gnat controlling-the-elaboration-order-in-ada}@anchor{232}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id5}@anchor{233}
27502 @section Controlling the Elaboration Order in Ada
27505 Ada provides several idioms and pragmas to aid the programmer with specifying
27506 the desired elaboration order and avoiding ABE problems altogether.
27512 @emph{Packages without a body}
27514 A library package which does not require a completing body does not suffer
27520 type Element is private;
27521 package Containers is
27522 type Element_Array is array (1 .. 10) of Element;
27527 In the example above, package @code{Pack} does not require a body because it
27528 does not contain any constructs which require completion in a body. As a
27529 result, generic @code{Pack.Containers} can be instantiated without encountering
27533 @geindex pragma Pure
27541 Pragma @code{Pure} places sufficient restrictions on a unit to guarantee that no
27542 scenario within the unit can result in an ABE problem.
27545 @geindex pragma Preelaborate
27551 @emph{pragma Preelaborate}
27553 Pragma @code{Preelaborate} is slightly less restrictive than pragma @code{Pure},
27554 but still strong enough to prevent ABE problems within a unit.
27557 @geindex pragma Elaborate_Body
27563 @emph{pragma Elaborate_Body}
27565 Pragma @code{Elaborate_Body} requires that the body of a unit is elaborated
27566 immediately after its spec. This restriction guarantees that no client
27567 scenario can execute a server target before the target body has been
27568 elaborated because the spec and body are effectively "glued" together.
27572 pragma Elaborate_Body;
27574 function Func return Integer;
27579 package body Server is
27580 function Func return Integer is
27590 Val : constant Integer := Server.Func;
27594 In the example above, pragma @code{Elaborate_Body} guarantees the following
27603 because the spec of @code{Server} must be elaborated prior to @code{Client} by
27604 virtue of the @emph{with} clause, and in addition the body of @code{Server} must be
27605 elaborated immediately after the spec of @code{Server}.
27607 Removing pragma @code{Elaborate_Body} could result in the following incorrect
27616 where @code{Client} invokes @code{Server.Func}, but the body of @code{Server.Func} has
27617 not been elaborated yet.
27620 The pragmas outlined above allow a server unit to guarantee safe elaboration
27621 use by client units. Thus it is a good rule to mark units as @code{Pure} or
27622 @code{Preelaborate}, and if this is not possible, mark them as @code{Elaborate_Body}.
27624 There are however situations where @code{Pure}, @code{Preelaborate}, and
27625 @code{Elaborate_Body} are not applicable. Ada provides another set of pragmas for
27626 use by client units to help ensure the elaboration safety of server units they
27629 @geindex pragma Elaborate (Unit)
27635 @emph{pragma Elaborate (Unit)}
27637 Pragma @code{Elaborate} can be placed in the context clauses of a unit, after a
27638 @emph{with} clause. It guarantees that both the spec and body of its argument will
27639 be elaborated prior to the unit with the pragma. Note that other unrelated
27640 units may be elaborated in between the spec and the body.
27644 function Func return Integer;
27649 package body Server is
27650 function Func return Integer is
27659 pragma Elaborate (Server);
27661 Val : constant Integer := Server.Func;
27665 In the example above, pragma @code{Elaborate} guarantees the following
27674 Removing pragma @code{Elaborate} could result in the following incorrect
27683 where @code{Client} invokes @code{Server.Func}, but the body of @code{Server.Func}
27684 has not been elaborated yet.
27687 @geindex pragma Elaborate_All (Unit)
27693 @emph{pragma Elaborate_All (Unit)}
27695 Pragma @code{Elaborate_All} is placed in the context clauses of a unit, after
27696 a @emph{with} clause. It guarantees that both the spec and body of its argument
27697 will be elaborated prior to the unit with the pragma, as well as all units
27698 @emph{with}ed by the spec and body of the argument, recursively. Note that other
27699 unrelated units may be elaborated in between the spec and the body.
27703 function Factorial (Val : Natural) return Natural;
27708 package body Math is
27709 function Factorial (Val : Natural) return Natural is
27717 package Computer is
27718 type Operation_Kind is (None, Op_Factorial);
27722 Op : Operation_Kind) return Natural;
27728 package body Computer is
27731 Op : Operation_Kind) return Natural
27733 if Op = Op_Factorial then
27734 return Math.Factorial (Val);
27744 pragma Elaborate_All (Computer);
27746 Val : constant Natural :=
27747 Computer.Compute (123, Computer.Op_Factorial);
27751 In the example above, pragma @code{Elaborate_All} can result in the following
27762 Note that there are several allowable suborders for the specs and bodies of
27763 @code{Math} and @code{Computer}, but the point is that these specs and bodies will
27764 be elaborated prior to @code{Client}.
27766 Removing pragma @code{Elaborate_All} could result in the following incorrect
27777 where @code{Client} invokes @code{Computer.Compute}, which in turn invokes
27778 @code{Math.Factorial}, but the body of @code{Math.Factorial} has not been
27782 All pragmas shown above can be summarized by the following rule:
27784 @emph{If a client unit elaborates a server target directly or indirectly, then if
27785 the server unit requires a body and does not have pragma Pure, Preelaborate,
27786 or Elaborate_Body, then the client unit should have pragma Elaborate or
27787 Elaborate_All for the server unit.}
27789 If the rule outlined above is not followed, then a program may fall in one of
27790 the following states:
27796 @emph{No elaboration order exists}
27798 In this case a compiler must diagnose the situation, and refuse to build an
27799 executable program.
27802 @emph{One or more incorrect elaboration orders exist}
27804 In this case a compiler can build an executable program, but
27805 @code{Program_Error} will be raised when the program is run.
27808 @emph{Several elaboration orders exist, some correct, some incorrect}
27810 In this case the programmer has not controlled the elaboration order. As a
27811 result, a compiler may or may not pick one of the correct orders, and the
27812 program may or may not raise @code{Program_Error} when it is run. This is the
27813 worst possible state because the program may fail on another compiler, or
27814 even another version of the same compiler.
27817 @emph{One or more correct orders exist}
27819 In this case a compiler can build an executable program, and the program is
27820 run successfully. This state may be guaranteed by following the outlined
27821 rules, or may be the result of good program architecture.
27824 Note that one additional advantage of using @code{Elaborate} and @code{Elaborate_All}
27825 is that the program continues to stay in the last state (one or more correct
27826 orders exist) even if maintenance changes the bodies of targets.
27828 @node Controlling the Elaboration Order in GNAT,Common Elaboration-model Traits,Controlling the Elaboration Order in Ada,Elaboration Order Handling in GNAT
27829 @anchor{gnat_ugn/elaboration_order_handling_in_gnat id6}@anchor{234}@anchor{gnat_ugn/elaboration_order_handling_in_gnat controlling-the-elaboration-order-in-gnat}@anchor{235}
27830 @section Controlling the Elaboration Order in GNAT
27833 In addition to Ada semantics and rules synthesized from them, GNAT offers
27834 three elaboration models to aid the programmer with specifying the correct
27835 elaboration order and to diagnose elaboration problems.
27837 @geindex Dynamic elaboration model
27843 @emph{Dynamic elaboration model}
27845 This is the most permissive of the three elaboration models. When the
27846 dynamic model is in effect, GNAT assumes that all code within all units in
27847 a partition is elaboration code. GNAT performs very few diagnostics and
27848 generates run-time checks to verify the elaboration order of a program. This
27849 behavior is identical to that specified by the Ada Reference Manual. The
27850 dynamic model is enabled with compiler switch @code{-gnatE}.
27853 @geindex Static elaboration model
27859 @emph{Static elaboration model}
27861 This is the middle ground of the three models. When the static model is in
27862 effect, GNAT performs extensive diagnostics on a unit-by-unit basis for all
27863 scenarios that elaborate or execute internal targets. GNAT also generates
27864 run-time checks for all external targets and for all scenarios that may
27865 exhibit ABE problems. Finally, GNAT installs implicit @code{Elaborate} and
27866 @code{Elaborate_All} pragmas for server units based on the dependencies of
27867 client units. The static model is the default model in GNAT.
27870 @geindex SPARK elaboration model
27876 @emph{SPARK elaboration model}
27878 This is the most conservative of the three models and enforces the SPARK
27879 rules of elaboration as defined in the SPARK Reference Manual, section 7.7.
27880 The SPARK model is in effect only when a scenario and a target reside in a
27881 region subject to SPARK_Mode On, otherwise the dynamic or static model is in
27885 @geindex Legacy elaboration model
27891 @emph{Legacy elaboration model}
27893 In addition to the three elaboration models outlined above, GNAT provides the
27894 elaboration model of pre-18.x versions referred to as @cite{legacy elaboration model}. The legacy elaboration model is enabled with compiler switch
27898 @geindex Relaxed elaboration mode
27900 The dynamic, legacy, and static models can be relaxed using compiler switch
27901 @code{-gnatJ}, making them more permissive. Note that in this mode, GNAT
27902 may not diagnose certain elaboration issues or install run-time checks.
27904 @node Common Elaboration-model Traits,Dynamic Elaboration Model in GNAT,Controlling the Elaboration Order in GNAT,Elaboration Order Handling in GNAT
27905 @anchor{gnat_ugn/elaboration_order_handling_in_gnat common-elaboration-model-traits}@anchor{236}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id7}@anchor{237}
27906 @section Common Elaboration-model Traits
27909 All three GNAT models are able to detect elaboration problems related to
27910 dispatching calls and a particular kind of ABE referred to as @emph{guaranteed ABE}.
27916 @emph{Dispatching calls}
27918 GNAT installs run-time checks for each primitive subprogram of each tagged
27919 type defined in a partition on the assumption that a dispatching call
27920 invoked at elaboration time will execute one of these primitives. As a
27921 result, a dispatching call that executes a primitive whose body has not
27922 been elaborated yet will raise exception @code{Program_Error} at run time. The
27923 checks can be suppressed using pragma @code{Suppress (Elaboration_Check)}.
27926 @emph{Guaranteed ABE}
27928 A guaranteed ABE arises when the body of a target is not elaborated early
27929 enough, and causes all scenarios that directly execute the target to fail.
27932 package body Guaranteed_ABE is
27933 function ABE return Integer;
27935 Val : constant Integer := ABE;
27937 function ABE return Integer is
27941 end Guaranteed_ABE;
27944 In the example above, the elaboration of @code{Guaranteed_ABE}'s body elaborates
27945 the declaration of @code{Val}. This invokes function @code{ABE}, however the body
27946 of @code{ABE} has not been elaborated yet. GNAT emits similar diagnostics in all
27950 1. package body Guaranteed_ABE is
27951 2. function ABE return Integer;
27953 4. Val : constant Integer := ABE;
27955 >>> warning: cannot call "ABE" before body seen
27956 >>> warning: Program_Error will be raised at run time
27959 6. function ABE return Integer is
27963 10. end Guaranteed_ABE;
27967 Note that GNAT emits warnings rather than hard errors whenever it encounters an
27968 elaboration problem. This is because the elaboration model in effect may be too
27969 conservative, or a particular scenario may not be elaborated or executed due to
27970 data and control flow. The warnings can be suppressed selectively with @code{pragma
27971 Warnigns (Off)} or globally with compiler switch @code{-gnatwL}.
27973 @node Dynamic Elaboration Model in GNAT,Static Elaboration Model in GNAT,Common Elaboration-model Traits,Elaboration Order Handling in GNAT
27974 @anchor{gnat_ugn/elaboration_order_handling_in_gnat dynamic-elaboration-model-in-gnat}@anchor{238}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id8}@anchor{239}
27975 @section Dynamic Elaboration Model in GNAT
27978 The dynamic model assumes that all code within all units in a partition is
27979 elaboration code. As a result, run-time checks are installed for each scenario
27980 regardless of whether the target is internal or external. The checks can be
27981 suppressed using pragma @code{Suppress (Elaboration_Check)}. This behavior is
27982 identical to that specified by the Ada Reference Manual. The following example
27983 showcases run-time checks installed by GNAT to verify the elaboration state of
27984 package @code{Dynamic_Model}.
27988 package body Dynamic_Model is
27994 <check that the body of Server.Gen is elaborated>
27995 package Inst is new Server.Gen;
27997 T : Server.Task_Type;
28000 <check that the body of Server.Task_Type is elaborated>
28002 <check that the body of Server.Proc is elaborated>
28007 The checks verify that the body of a target has been successfully elaborated
28008 before a scenario activates, calls, or instantiates a target.
28010 Note that no scenario within package @code{Dynamic_Model} calls procedure @code{API}.
28011 In fact, procedure @code{API} may not be invoked by elaboration code within the
28012 partition, however the dynamic model assumes that this can happen.
28014 The dynamic model emits very few diagnostics, but can make suggestions on
28015 missing @code{Elaborate} and @code{Elaborate_All} pragmas for library-level
28016 scenarios. This information is available when compiler switch @code{-gnatel}
28021 2. package body Dynamic_Model is
28022 3. Val : constant Integer := Server.Func;
28024 >>> info: call to "Func" during elaboration
28025 >>> info: missing pragma "Elaborate_All" for unit "Server"
28027 4. end Dynamic_Model;
28030 @node Static Elaboration Model in GNAT,SPARK Elaboration Model in GNAT,Dynamic Elaboration Model in GNAT,Elaboration Order Handling in GNAT
28031 @anchor{gnat_ugn/elaboration_order_handling_in_gnat static-elaboration-model-in-gnat}@anchor{23a}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id9}@anchor{23b}
28032 @section Static Elaboration Model in GNAT
28035 In contrast to the dynamic model, the static model is more precise in its
28036 analysis of elaboration code. The model makes a clear distinction between
28037 internal and external targets, and resorts to different diagnostics and
28038 run-time checks based on the nature of the target.
28044 @emph{Internal targets}
28046 The static model performs extensive diagnostics on scenarios which elaborate
28047 or execute internal targets. The warnings resulting from these diagnostics
28048 are enabled by default, but can be suppressed selectively with @code{pragma
28049 Warnings (Off)} or globally with compiler switch @code{-gnatwL}.
28052 1. package body Static_Model is
28054 3. with function Func return Integer;
28056 5. Val : constant Integer := Func;
28059 8. function ABE return Integer;
28061 10. function Cause_ABE return Boolean is
28062 11. package Inst is new Gen (ABE);
28064 >>> warning: in instantiation at line 5
28065 >>> warning: cannot call "ABE" before body seen
28066 >>> warning: Program_Error may be raised at run time
28067 >>> warning: body of unit "Static_Model" elaborated
28068 >>> warning: function "Cause_ABE" called at line 16
28069 >>> warning: function "ABE" called at line 5, instance at line 11
28075 16. Val : constant Boolean := Cause_ABE;
28077 18. function ABE return Integer is
28081 22. end Static_Model;
28084 The example above illustrates an ABE problem within package @code{Static_Model},
28085 which is hidden by several layers of indirection. The elaboration of package
28086 body @code{Static_Model} elaborates the declaration of @code{Val}. This invokes
28087 function @code{Cause_ABE}, which instantiates generic unit @code{Gen} as @code{Inst}.
28088 The elaboration of @code{Inst} invokes function @code{ABE}, however the body of
28089 @code{ABE} has not been elaborated yet.
28092 @emph{External targets}
28094 The static model installs run-time checks to verify the elaboration status
28095 of server targets only when the scenario that elaborates or executes that
28096 target is part of the elaboration code of the client unit. The checks can be
28097 suppressed using pragma @code{Suppress (Elaboration_Check)}.
28101 package body Static_Model is
28103 with function Func return Integer;
28105 Val : constant Integer := Func;
28108 function Call_Func return Boolean is
28109 <check that the body of Server.Func is elaborated>
28110 package Inst is new Gen (Server.Func);
28115 Val : constant Boolean := Call_Func;
28119 In the example above, the elaboration of package body @code{Static_Model}
28120 elaborates the declaration of @code{Val}. This invokes function @code{Call_Func},
28121 which instantiates generic unit @code{Gen} as @code{Inst}. The elaboration of
28122 @code{Inst} invokes function @code{Server.Func}. Since @code{Server.Func} is an
28123 external target, GNAT installs a run-time check to verify that its body has
28126 In addition to checks, the static model installs implicit @code{Elaborate} and
28127 @code{Elaborate_All} pragmas to guarantee safe elaboration use of server units.
28128 This information is available when compiler switch @code{-gnatel} is in
28133 2. package body Static_Model is
28135 4. with function Func return Integer;
28137 6. Val : constant Integer := Func;
28140 9. function Call_Func return Boolean is
28141 10. package Inst is new Gen (Server.Func);
28143 >>> info: instantiation of "Gen" during elaboration
28144 >>> info: in instantiation at line 6
28145 >>> info: call to "Func" during elaboration
28146 >>> info: in instantiation at line 6
28147 >>> info: implicit pragma "Elaborate_All" generated for unit "Server"
28148 >>> info: body of unit "Static_Model" elaborated
28149 >>> info: function "Call_Func" called at line 15
28150 >>> info: function "Func" called at line 6, instance at line 10
28156 15. Val : constant Boolean := Call_Func;
28158 >>> info: call to "Call_Func" during elaboration
28160 16. end Static_Model;
28163 In the example above, the elaboration of package body @code{Static_Model}
28164 elaborates the declaration of @code{Val}. This invokes function @code{Call_Func},
28165 which instantiates generic unit @code{Gen} as @code{Inst}. The elaboration of
28166 @code{Inst} invokes function @code{Server.Func}. Since @code{Server.Func} is an
28167 external target, GNAT installs an implicit @code{Elaborate_All} pragma for unit
28168 @code{Server}. The pragma guarantees that both the spec and body of @code{Server},
28169 along with any additional dependencies that @code{Server} may require, are
28170 elaborated prior to the body of @code{Static_Model}.
28173 @node SPARK Elaboration Model in GNAT,Legacy Elaboration Model in GNAT,Static Elaboration Model in GNAT,Elaboration Order Handling in GNAT
28174 @anchor{gnat_ugn/elaboration_order_handling_in_gnat id10}@anchor{23c}@anchor{gnat_ugn/elaboration_order_handling_in_gnat spark-elaboration-model-in-gnat}@anchor{23d}
28175 @section SPARK Elaboration Model in GNAT
28178 The SPARK model is identical to the static model in its handling of internal
28179 targets. The SPARK model, however, requires explicit @code{Elaborate} or
28180 @code{Elaborate_All} pragmas to be present in the program when a target is
28181 external, and compiler switch @code{-gnatd.v} is in effect.
28185 2. package body SPARK_Model with SPARK_Mode is
28186 3. Val : constant Integer := Server.Func;
28188 >>> call to "Func" during elaboration in SPARK
28189 >>> unit "SPARK_Model" requires pragma "Elaborate_All" for "Server"
28190 >>> body of unit "SPARK_Model" elaborated
28191 >>> function "Func" called at line 3
28193 4. end SPARK_Model;
28196 @node Legacy Elaboration Model in GNAT,Mixing Elaboration Models,SPARK Elaboration Model in GNAT,Elaboration Order Handling in GNAT
28197 @anchor{gnat_ugn/elaboration_order_handling_in_gnat legacy-elaboration-model-in-gnat}@anchor{23e}
28198 @section Legacy Elaboration Model in GNAT
28201 The legacy elaboration model is provided for compatibility with code bases
28202 developed with pre-18.x versions of GNAT. It is similar in functionality to
28203 the dynamic and static models of post-18.x version of GNAT, but may differ
28204 in terms of diagnostics and run-time checks. The legacy elaboration model is
28205 enabled with compiler switch @code{-gnatH}.
28207 @node Mixing Elaboration Models,Elaboration Circularities,Legacy Elaboration Model in GNAT,Elaboration Order Handling in GNAT
28208 @anchor{gnat_ugn/elaboration_order_handling_in_gnat mixing-elaboration-models}@anchor{23f}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id11}@anchor{240}
28209 @section Mixing Elaboration Models
28212 It is possible to mix units compiled with a different elaboration model,
28213 however the following rules must be observed:
28219 A client unit compiled with the dynamic model can only @emph{with} a server unit
28220 that meets at least one of the following criteria:
28226 The server unit is compiled with the dynamic model.
28229 The server unit is a GNAT implementation unit from the Ada, GNAT,
28230 Interfaces, or System hierarchies.
28233 The server unit has pragma @code{Pure} or @code{Preelaborate}.
28236 The client unit has an explicit @code{Elaborate_All} pragma for the server
28241 These rules ensure that elaboration checks are not omitted. If the rules are
28242 violated, the binder emits a warning:
28245 warning: "x.ads" has dynamic elaboration checks and with's
28246 warning: "y.ads" which has static elaboration checks
28249 The warnings can be suppressed by binder switch @code{-ws}.
28251 @node Elaboration Circularities,Resolving Elaboration Circularities,Mixing Elaboration Models,Elaboration Order Handling in GNAT
28252 @anchor{gnat_ugn/elaboration_order_handling_in_gnat id12}@anchor{241}@anchor{gnat_ugn/elaboration_order_handling_in_gnat elaboration-circularities}@anchor{242}
28253 @section Elaboration Circularities
28256 If the binder cannot find an acceptable elaboration order, it outputs detailed
28257 diagnostics describing an @strong{elaboration circularity}.
28261 function Func return Integer;
28267 package body Server is
28268 function Func return Integer is
28278 Val : constant Integer := Server.Func;
28284 procedure Main is begin null; end Main;
28288 error: elaboration circularity detected
28289 info: "server (body)" must be elaborated before "client (spec)"
28290 info: reason: implicit Elaborate_All in unit "client (spec)"
28291 info: recompile "client (spec)" with -gnatel for full details
28292 info: "server (body)"
28293 info: must be elaborated along with its spec:
28294 info: "server (spec)"
28295 info: which is withed by:
28296 info: "client (spec)"
28297 info: "client (spec)" must be elaborated before "server (body)"
28298 info: reason: with clause
28301 In the example above, @code{Client} must be elaborated prior to @code{Main} by virtue
28302 of a @emph{with} clause. The elaboration of @code{Client} invokes @code{Server.Func}, and
28303 static model generates an implicit @code{Elaborate_All} pragma for @code{Server}. The
28304 pragma implies that both the spec and body of @code{Server}, along with any units
28305 they @emph{with}, must be elaborated prior to @code{Client}. However, @code{Server}'s body
28306 @emph{with}s @code{Client}, implying that @code{Client} must be elaborated prior to
28307 @code{Server}. The end result is that @code{Client} must be elaborated prior to
28308 @code{Client}, and this leads to a circularity.
28310 @node Resolving Elaboration Circularities,Resolving Task Issues,Elaboration Circularities,Elaboration Order Handling in GNAT
28311 @anchor{gnat_ugn/elaboration_order_handling_in_gnat id13}@anchor{243}@anchor{gnat_ugn/elaboration_order_handling_in_gnat resolving-elaboration-circularities}@anchor{244}
28312 @section Resolving Elaboration Circularities
28315 When faced with an elaboration circularity, a programmer has several options
28322 @emph{Fix the program}
28324 The most desirable option from the point of view of long-term maintenance
28325 is to rearrange the program so that the elaboration problems are avoided.
28326 One useful technique is to place the elaboration code into separate child
28327 packages. Another is to move some of the initialization code to explicitly
28328 invoked subprograms, where the program controls the order of initialization
28329 explicitly. Although this is the most desirable option, it may be impractical
28330 and involve too much modification, especially in the case of complex legacy
28334 @emph{Switch to more permissive elaboration model}
28336 If the compilation was performed using the static model, enable the dynamic
28337 model with compiler switch @code{-gnatE}. GNAT will no longer generate
28338 implicit @code{Elaborate} and @code{Elaborate_All} pragmas, resulting in a behavior
28339 identical to that specified by the Ada Reference Manual. The binder will
28340 generate an executable program that may or may not raise @code{Program_Error},
28341 and it is the programmer's responsibility to ensure that it does not raise
28342 @code{Program_Error}.
28344 If the compilation was performed using a post-18.x version of GNAT, consider
28345 using the legacy elaboration model, in the following order:
28351 Use the relaxed static elaboration model, with compiler switch
28355 Use the relaxed dynamic elaboration model, with compiler switches
28356 @code{-gnatE} @code{-gnatJ}.
28359 Use the legacy static elaboration model, with compiler switch
28363 Use the legacy dynamic elaboration model, with compiler switches
28364 @code{-gnatE} @code{-gnatH}.
28368 @emph{Suppress all elaboration checks}
28370 The drawback of run-time checks is that they generate overhead at run time,
28371 both in space and time. If the programmer is absolutely sure that a program
28372 will not raise an elaboration-related @code{Program_Error}, then using the
28373 pragma @code{Suppress (Elaboration_Check)} globally (as a configuration pragma)
28374 will eliminate all run-time checks.
28377 @emph{Suppress elaboration checks selectively}
28379 If a scenario cannot possibly lead to an elaboration @code{Program_Error},
28380 and the binder nevertheless complains about implicit @code{Elaborate} and
28381 @code{Elaborate_All} pragmas that lead to elaboration circularities, it
28382 is possible to suppress the generation of implicit @code{Elaborate} and
28383 @code{Elaborate_All} pragmas, as well as run-time checks. Clearly this can
28384 be unsafe, and it is the responsibility of the programmer to make sure
28385 that the resulting program has no elaboration anomalies. Pragma
28386 @code{Suppress (Elaboration_Check)} can be used with different levels of
28387 granularity to achieve these effects.
28393 @emph{Target suppression}
28395 When the pragma is placed in a declarative part, without a second argument
28396 naming an entity, it will suppress implicit @code{Elaborate} and
28397 @code{Elaborate_All} pragma generation, as well as run-time checks, on all
28398 targets within the region.
28401 package Range_Suppress is
28402 pragma Suppress (Elaboration_Check);
28404 function Func return Integer;
28409 pragma Unsuppress (Elaboration_Check);
28412 end Range_Suppress;
28415 In the example above, a pair of Suppress/Unsuppress pragmas define a region
28416 of suppression within package @code{Range_Suppress}. As a result, no implicit
28417 @code{Elaborate} and @code{Elaborate_All} pragmas, nor any run-time checks, will
28418 be generated by callers of @code{Func} and instantiators of @code{Gen}. Note that
28419 task type @code{Tsk} is not within this region.
28421 An alternative to the region-based suppression is to use multiple
28422 @code{Suppress} pragmas with arguments naming specific entities for which
28423 elaboration checks should be suppressed:
28426 package Range_Suppress is
28427 function Func return Integer;
28428 pragma Suppress (Elaboration_Check, Func);
28432 pragma Suppress (Elaboration_Check, Gen);
28435 end Range_Suppress;
28439 @emph{Scenario suppression}
28441 When the pragma @code{Suppress} is placed in a declarative or statement
28442 part, without an entity argument, it will suppress implicit @code{Elaborate}
28443 and @code{Elaborate_All} pragma generation, as well as run-time checks, on
28444 all scenarios within the region.
28448 package body Range_Suppress is
28449 pragma Suppress (Elaboration_Check);
28451 function Func return Integer is
28453 return Server.Func;
28461 pragma Unsuppress (Elaboration_Check);
28467 end Range_Suppress;
28470 In the example above, a pair of Suppress/Unsuppress pragmas define a region
28471 of suppression within package body @code{Range_Suppress}. As a result, the
28472 calls to @code{Server.Func} in @code{Func} and @code{Server.Proc} in @code{Gen} will
28473 not generate any implicit @code{Elaborate} and @code{Elaborate_All} pragmas or
28478 @node Resolving Task Issues,Elaboration-related Compiler Switches,Resolving Elaboration Circularities,Elaboration Order Handling in GNAT
28479 @anchor{gnat_ugn/elaboration_order_handling_in_gnat id14}@anchor{245}@anchor{gnat_ugn/elaboration_order_handling_in_gnat resolving-task-issues}@anchor{246}
28480 @section Resolving Task Issues
28483 The model of execution in Ada dictates that elaboration must first take place,
28484 and only then can the main program be started. Tasks which are activated during
28485 elaboration violate this model and may lead to serious concurrent problems at
28488 A task can be activated in two different ways:
28494 The task is created by an allocator in which case it is activated immediately
28495 after the allocator is evaluated.
28498 The task is declared at the library level or within some nested master in
28499 which case it is activated before starting execution of the statement
28500 sequence of the master defining the task.
28503 Since the elaboration of a partition is performed by the environment task
28504 servicing that partition, any tasks activated during elaboration may be in
28505 a race with the environment task, and lead to unpredictable state and behavior.
28506 The static model seeks to avoid such interactions by assuming that all code in
28507 the task body is executed at elaboration time, if the task was activated by
28516 type My_Int is new Integer;
28518 function Ident (M : My_Int) return My_Int;
28524 package body Decls is
28525 task body Lib_Task is
28531 function Ident (M : My_Int) return My_Int is
28541 procedure Put_Val (Arg : Decls.My_Int);
28546 with Ada.Text_IO; use Ada.Text_IO;
28547 package body Utils is
28548 procedure Put_Val (Arg : Decls.My_Int) is
28550 Put_Line (Arg'Img);
28559 Decls.Lib_Task.Start;
28563 When the above example is compiled with the static model, an elaboration
28564 circularity arises:
28567 error: elaboration circularity detected
28568 info: "decls (body)" must be elaborated before "decls (body)"
28569 info: reason: implicit Elaborate_All in unit "decls (body)"
28570 info: recompile "decls (body)" with -gnatel for full details
28571 info: "decls (body)"
28572 info: must be elaborated along with its spec:
28573 info: "decls (spec)"
28574 info: which is withed by:
28575 info: "utils (spec)"
28576 info: which is withed by:
28577 info: "decls (body)"
28580 In the above example, @code{Decls} must be elaborated prior to @code{Main} by virtue
28581 of a with clause. The elaboration of @code{Decls} activates task @code{Lib_Task}. The
28582 static model conservatibely assumes that all code within the body of
28583 @code{Lib_Task} is executed, and generates an implicit @code{Elaborate_All} pragma
28584 for @code{Units} due to the call to @code{Utils.Put_Val}. The pragma implies that
28585 both the spec and body of @code{Utils}, along with any units they @emph{with},
28586 must be elaborated prior to @code{Decls}. However, @code{Utils}'s spec @emph{with}s
28587 @code{Decls}, implying that @code{Decls} must be elaborated before @code{Utils}. The end
28588 result is that @code{Utils} must be elaborated prior to @code{Utils}, and this
28589 leads to a circularity.
28591 In reality, the example above will not exhibit an ABE problem at run time.
28592 When the body of task @code{Lib_Task} is activated, execution will wait for entry
28593 @code{Start} to be accepted, and the call to @code{Utils.Put_Val} will not take place
28594 at elaboration time. Task @code{Lib_Task} will resume its execution after the main
28595 program is executed because @code{Main} performs a rendezvous with
28596 @code{Lib_Task.Start}, and at that point all units have already been elaborated.
28597 As a result, the static model may seem overly conservative, partly because it
28598 does not take control and data flow into account.
28600 When faced with a task elaboration circularity, a programmer has several
28607 @emph{Use the dynamic model}
28609 The dynamic model does not generate implicit @code{Elaborate} and
28610 @code{Elaborate_All} pragmas. Instead, it will install checks prior to every
28611 call in the example above, thus verifying the successful elaboration of
28612 @code{Utils.Put_Val} in case the call to it takes place at elaboration time.
28613 The dynamic model is enabled with compiler switch @code{-gnatE}.
28616 @emph{Isolate the tasks}
28618 Relocating tasks in their own separate package could decouple them from
28619 dependencies that would otherwise cause an elaboration circularity. The
28620 example above can be rewritten as follows:
28623 package Decls1 is -- new
28632 package body Decls1 is -- new
28633 task body Lib_Task is
28642 package Decls2 is -- new
28643 type My_Int is new Integer;
28644 function Ident (M : My_Int) return My_Int;
28650 package body Decls2 is -- new
28651 function Ident (M : My_Int) return My_Int is
28661 procedure Put_Val (Arg : Decls2.My_Int);
28666 with Ada.Text_IO; use Ada.Text_IO;
28667 package body Utils is
28668 procedure Put_Val (Arg : Decls2.My_Int) is
28670 Put_Line (Arg'Img);
28679 Decls1.Lib_Task.Start;
28684 @emph{Declare the tasks}
28686 The original example uses a single task declaration for @code{Lib_Task}. An
28687 explicit task type declaration and a properly placed task object could avoid
28688 the dependencies that would otherwise cause an elaboration circularity. The
28689 example can be rewritten as follows:
28693 task type Lib_Task is -- new
28697 type My_Int is new Integer;
28699 function Ident (M : My_Int) return My_Int;
28705 package body Decls is
28706 task body Lib_Task is
28712 function Ident (M : My_Int) return My_Int is
28722 procedure Put_Val (Arg : Decls.My_Int);
28727 with Ada.Text_IO; use Ada.Text_IO;
28728 package body Utils is
28729 procedure Put_Val (Arg : Decls.My_Int) is
28731 Put_Line (Arg'Img);
28738 package Obj_Decls is -- new
28739 Task_Obj : Decls.Lib_Task;
28747 Obj_Decls.Task_Obj.Start; -- new
28752 @emph{Use restriction No_Entry_Calls_In_Elaboration_Code}
28754 The issue exhibited in the original example under this section revolves
28755 around the body of @code{Lib_Task} blocking on an accept statement. There is
28756 no rule to prevent elaboration code from performing entry calls, however in
28757 practice this is highly unusual. In addition, the pattern of starting tasks
28758 at elaboration time and then immediately blocking on accept or select
28759 statements is quite common.
28761 If a programmer knows that elaboration code will not perform any entry
28762 calls, then the programmer can indicate that the static model should not
28763 process the remainder of a task body once an accept or select statement has
28764 been encountered. This behavior can be specified by a configuration pragma:
28767 pragma Restrictions (No_Entry_Calls_In_Elaboration_Code);
28770 In addition to the change in behavior with respect to task bodies, the
28771 static model will verify that no entry calls take place at elaboration time.
28774 @node Elaboration-related Compiler Switches,Summary of Procedures for Elaboration Control,Resolving Task Issues,Elaboration Order Handling in GNAT
28775 @anchor{gnat_ugn/elaboration_order_handling_in_gnat elaboration-related-compiler-switches}@anchor{247}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id15}@anchor{248}
28776 @section Elaboration-related Compiler Switches
28779 GNAT has several switches that affect the elaboration model and consequently
28780 the elaboration order chosen by the binder.
28782 @geindex -gnatE (gnat)
28787 @item @code{-gnatE}
28789 Dynamic elaboration checking mode enabled
28791 When this switch is in effect, GNAT activates the dynamic elaboration model.
28794 @geindex -gnatel (gnat)
28799 @item @code{-gnatel}
28801 Turn on info messages on generated Elaborate[_All] pragmas
28803 When this switch is in effect, GNAT will emit the following supplementary
28804 information depending on the elaboration model in effect.
28810 @emph{Dynamic model}
28812 GNAT will indicate missing @code{Elaborate} and @code{Elaborate_All} pragmas for
28813 all library-level scenarios within the partition.
28816 @emph{Static model}
28818 GNAT will indicate all scenarios executed during elaboration. In addition,
28819 it will provide detailed traceback when an implicit @code{Elaborate} or
28820 @code{Elaborate_All} pragma is generated.
28825 GNAT will indicate how an elaboration requirement is met by the context of
28826 a unit. This diagnostic requires compiler switch @code{-gnatd.v}.
28829 1. with Server; pragma Elaborate_All (Server);
28830 2. package Client with SPARK_Mode is
28831 3. Val : constant Integer := Server.Func;
28833 >>> info: call to "Func" during elaboration in SPARK
28834 >>> info: "Elaborate_All" requirement for unit "Server" met by pragma at line 1
28841 @geindex -gnatH (gnat)
28846 @item @code{-gnatH}
28848 Legacy elaboration checking mode enabled
28850 When this switch is in effect, GNAT will utilize the pre-18.x elaboration
28854 @geindex -gnatJ (gnat)
28859 @item @code{-gnatJ}
28861 Relaxed elaboration checking mode enabled
28863 When this switch is in effect, GNAT will not process certain scenarios,
28864 resulting in a more permissive elaboration model. Note that this may
28865 eliminate some diagnostics and run-time checks.
28868 @geindex -gnatw.f (gnat)
28873 @item @code{-gnatw.f}
28875 Turn on warnings for suspicious Subp'Access
28877 When this switch is in effect, GNAT will treat @code{'Access} of an entry,
28878 operator, or subprogram as a potential call to the target and issue warnings:
28881 1. package body Attribute_Call is
28882 2. function Func return Integer;
28883 3. type Func_Ptr is access function return Integer;
28885 5. Ptr : constant Func_Ptr := Func'Access;
28887 >>> warning: "Access" attribute of "Func" before body seen
28888 >>> warning: possible Program_Error on later references
28889 >>> warning: body of unit "Attribute_Call" elaborated
28890 >>> warning: "Access" of "Func" taken at line 5
28893 7. function Func return Integer is
28897 11. end Attribute_Call;
28900 In the example above, the elaboration of declaration @code{Ptr} is assigned
28901 @code{Func'Access} before the body of @code{Func} has been elaborated.
28904 @geindex -gnatwl (gnat)
28909 @item @code{-gnatwl}
28911 Turn on warnings for elaboration problems
28913 When this switch is in effect, GNAT emits diagnostics in the form of warnings
28914 concerning various elaboration problems. The warnings are enabled by default.
28915 The switch is provided in case all warnings are suppressed, but elaboration
28916 warnings are still desired.
28918 @item @code{-gnatwL}
28920 Turn off warnings for elaboration problems
28922 When this switch is in effect, GNAT no longer emits any diagnostics in the
28923 form of warnings. Selective suppression of elaboration problems is possible
28924 using @code{pragma Warnings (Off)}.
28927 1. package body Selective_Suppression is
28928 2. function ABE return Integer;
28930 4. Val_1 : constant Integer := ABE;
28932 >>> warning: cannot call "ABE" before body seen
28933 >>> warning: Program_Error will be raised at run time
28936 6. pragma Warnings (Off);
28937 7. Val_2 : constant Integer := ABE;
28938 8. pragma Warnings (On);
28940 10. function ABE return Integer is
28944 14. end Selective_Suppression;
28947 Note that suppressing elaboration warnings does not eliminate run-time
28948 checks. The example above will still fail at run time with an ABE.
28951 @node Summary of Procedures for Elaboration Control,Inspecting the Chosen Elaboration Order,Elaboration-related Compiler Switches,Elaboration Order Handling in GNAT
28952 @anchor{gnat_ugn/elaboration_order_handling_in_gnat summary-of-procedures-for-elaboration-control}@anchor{249}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id16}@anchor{24a}
28953 @section Summary of Procedures for Elaboration Control
28956 A programmer should first compile the program with the default options, using
28957 none of the binder or compiler switches. If the binder succeeds in finding an
28958 elaboration order, then apart from possible cases involing dispatching calls
28959 and access-to-subprogram types, the program is free of elaboration errors.
28961 If it is important for the program to be portable to compilers other than GNAT,
28962 then the programmer should use compiler switch @code{-gnatel} and consider
28963 the messages about missing or implicitly created @code{Elaborate} and
28964 @code{Elaborate_All} pragmas.
28966 If the binder reports an elaboration circularity, the programmer has several
28973 Ensure that elaboration warnings are enabled. This will allow the static
28974 model to output trace information of elaboration issues. The trace
28975 information could shed light on previously unforeseen dependencies, as well
28976 as their origins. Elaboration warnings are enabled with compiler switch
28980 Use switch @code{-gnatel} to obtain messages on generated implicit
28981 @code{Elaborate} and @code{Elaborate_All} pragmas. The trace information could
28982 indicate why a server unit must be elaborated prior to a client unit.
28985 If the warnings produced by the static model indicate that a task is
28986 involved, consider the options in section @ref{245,,Resolving Task Issues}.
28989 If none of the steps outlined above resolve the circularity, use a more
28990 permissive elaboration model, in the following order:
28996 Use the dynamic elaboration model, with compiler switch @code{-gnatE}.
28999 Use the legacy static elaboration model, with compiler switch
29003 Use the legacy dynamic elaboration model, with compiler switches
29004 @code{-gnatH} @code{-gnatE}.
29007 Use the relaxed legacy static elaboration model, with compiler switches
29008 @code{-gnatH} @code{-gnatJ}.
29011 Use the relaxed legacy dynamic elaboration model, with compiler switches
29012 @code{-gnatH} @code{-gnatJ} @code{-gnatE}.
29016 @node Inspecting the Chosen Elaboration Order,,Summary of Procedures for Elaboration Control,Elaboration Order Handling in GNAT
29017 @anchor{gnat_ugn/elaboration_order_handling_in_gnat inspecting-the-chosen-elaboration-order}@anchor{24b}@anchor{gnat_ugn/elaboration_order_handling_in_gnat id17}@anchor{24c}
29018 @section Inspecting the Chosen Elaboration Order
29021 To see the elaboration order chosen by the binder, inspect the contents of file
29022 @cite{b~xxx.adb}. On certain targets, this file appears as @cite{b_xxx.adb}. The
29023 elaboration order appears as a sequence of calls to @code{Elab_Body} and
29024 @code{Elab_Spec}, interspersed with assignments to @cite{Exxx} which indicates that a
29025 particular unit is elaborated. For example:
29028 System.Soft_Links'Elab_Body;
29030 System.Secondary_Stack'Elab_Body;
29032 System.Exception_Table'Elab_Body;
29034 Ada.Io_Exceptions'Elab_Spec;
29036 Ada.Tags'Elab_Spec;
29037 Ada.Streams'Elab_Spec;
29039 Interfaces.C'Elab_Spec;
29041 System.Finalization_Root'Elab_Spec;
29043 System.Os_Lib'Elab_Body;
29045 System.Finalization_Implementation'Elab_Spec;
29046 System.Finalization_Implementation'Elab_Body;
29048 Ada.Finalization'Elab_Spec;
29050 Ada.Finalization.List_Controller'Elab_Spec;
29052 System.File_Control_Block'Elab_Spec;
29054 System.File_Io'Elab_Body;
29056 Ada.Tags'Elab_Body;
29058 Ada.Text_Io'Elab_Spec;
29059 Ada.Text_Io'Elab_Body;
29063 Note also binder switch @code{-l}, which outputs the chosen elaboration
29064 order and provides a more readable form of the above:
29070 system.case_util (spec)
29071 system.case_util (body)
29072 system.concat_2 (spec)
29073 system.concat_2 (body)
29074 system.concat_3 (spec)
29075 system.concat_3 (body)
29076 system.htable (spec)
29077 system.parameters (spec)
29078 system.parameters (body)
29080 interfaces.c_streams (spec)
29081 interfaces.c_streams (body)
29082 system.restrictions (spec)
29083 system.restrictions (body)
29084 system.standard_library (spec)
29085 system.exceptions (spec)
29086 system.exceptions (body)
29087 system.storage_elements (spec)
29088 system.storage_elements (body)
29089 system.secondary_stack (spec)
29090 system.stack_checking (spec)
29091 system.stack_checking (body)
29092 system.string_hash (spec)
29093 system.string_hash (body)
29094 system.htable (body)
29095 system.strings (spec)
29096 system.strings (body)
29097 system.traceback (spec)
29098 system.traceback (body)
29099 system.traceback_entries (spec)
29100 system.traceback_entries (body)
29101 ada.exceptions (spec)
29102 ada.exceptions.last_chance_handler (spec)
29103 system.soft_links (spec)
29104 system.soft_links (body)
29105 ada.exceptions.last_chance_handler (body)
29106 system.secondary_stack (body)
29107 system.exception_table (spec)
29108 system.exception_table (body)
29109 ada.io_exceptions (spec)
29112 interfaces.c (spec)
29113 interfaces.c (body)
29114 system.finalization_root (spec)
29115 system.finalization_root (body)
29116 system.memory (spec)
29117 system.memory (body)
29118 system.standard_library (body)
29119 system.os_lib (spec)
29120 system.os_lib (body)
29121 system.unsigned_types (spec)
29122 system.stream_attributes (spec)
29123 system.stream_attributes (body)
29124 system.finalization_implementation (spec)
29125 system.finalization_implementation (body)
29126 ada.finalization (spec)
29127 ada.finalization (body)
29128 ada.finalization.list_controller (spec)
29129 ada.finalization.list_controller (body)
29130 system.file_control_block (spec)
29131 system.file_io (spec)
29132 system.file_io (body)
29133 system.val_uns (spec)
29134 system.val_util (spec)
29135 system.val_util (body)
29136 system.val_uns (body)
29137 system.wch_con (spec)
29138 system.wch_con (body)
29139 system.wch_cnv (spec)
29140 system.wch_jis (spec)
29141 system.wch_jis (body)
29142 system.wch_cnv (body)
29143 system.wch_stw (spec)
29144 system.wch_stw (body)
29146 ada.exceptions (body)
29153 @node Inline Assembler,GNU Free Documentation License,Elaboration Order Handling in GNAT,Top
29154 @anchor{gnat_ugn/inline_assembler inline-assembler}@anchor{10}@anchor{gnat_ugn/inline_assembler doc}@anchor{24d}@anchor{gnat_ugn/inline_assembler id1}@anchor{24e}
29155 @chapter Inline Assembler
29158 @geindex Inline Assembler
29160 If you need to write low-level software that interacts directly
29161 with the hardware, Ada provides two ways to incorporate assembly
29162 language code into your program. First, you can import and invoke
29163 external routines written in assembly language, an Ada feature fully
29164 supported by GNAT. However, for small sections of code it may be simpler
29165 or more efficient to include assembly language statements directly
29166 in your Ada source program, using the facilities of the implementation-defined
29167 package @code{System.Machine_Code}, which incorporates the gcc
29168 Inline Assembler. The Inline Assembler approach offers a number of advantages,
29169 including the following:
29175 No need to use non-Ada tools
29178 Consistent interface over different targets
29181 Automatic usage of the proper calling conventions
29184 Access to Ada constants and variables
29187 Definition of intrinsic routines
29190 Possibility of inlining a subprogram comprising assembler code
29193 Code optimizer can take Inline Assembler code into account
29196 This appendix presents a series of examples to show you how to use
29197 the Inline Assembler. Although it focuses on the Intel x86,
29198 the general approach applies also to other processors.
29199 It is assumed that you are familiar with Ada
29200 and with assembly language programming.
29203 * Basic Assembler Syntax::
29204 * A Simple Example of Inline Assembler::
29205 * Output Variables in Inline Assembler::
29206 * Input Variables in Inline Assembler::
29207 * Inlining Inline Assembler Code::
29208 * Other Asm Functionality::
29212 @node Basic Assembler Syntax,A Simple Example of Inline Assembler,,Inline Assembler
29213 @anchor{gnat_ugn/inline_assembler id2}@anchor{24f}@anchor{gnat_ugn/inline_assembler basic-assembler-syntax}@anchor{250}
29214 @section Basic Assembler Syntax
29217 The assembler used by GNAT and gcc is based not on the Intel assembly
29218 language, but rather on a language that descends from the AT&T Unix
29219 assembler @code{as} (and which is often referred to as 'AT&T syntax').
29220 The following table summarizes the main features of @code{as} syntax
29221 and points out the differences from the Intel conventions.
29222 See the gcc @code{as} and @code{gas} (an @code{as} macro
29223 pre-processor) documentation for further information.
29227 @emph{Register names}@w{ }
29229 gcc / @code{as}: Prefix with '%'; for example @code{%eax}@w{ }
29230 Intel: No extra punctuation; for example @code{eax}@w{ }
29238 @emph{Immediate operand}@w{ }
29240 gcc / @code{as}: Prefix with '$'; for example @code{$4}@w{ }
29241 Intel: No extra punctuation; for example @code{4}@w{ }
29249 @emph{Address}@w{ }
29251 gcc / @code{as}: Prefix with '$'; for example @code{$loc}@w{ }
29252 Intel: No extra punctuation; for example @code{loc}@w{ }
29260 @emph{Memory contents}@w{ }
29262 gcc / @code{as}: No extra punctuation; for example @code{loc}@w{ }
29263 Intel: Square brackets; for example @code{[loc]}@w{ }
29271 @emph{Register contents}@w{ }
29273 gcc / @code{as}: Parentheses; for example @code{(%eax)}@w{ }
29274 Intel: Square brackets; for example @code{[eax]}@w{ }
29282 @emph{Hexadecimal numbers}@w{ }
29284 gcc / @code{as}: Leading '0x' (C language syntax); for example @code{0xA0}@w{ }
29285 Intel: Trailing 'h'; for example @code{A0h}@w{ }
29293 @emph{Operand size}@w{ }
29295 gcc / @code{as}: Explicit in op code; for example @code{movw} to move a 16-bit word@w{ }
29296 Intel: Implicit, deduced by assembler; for example @code{mov}@w{ }
29304 @emph{Instruction repetition}@w{ }
29306 gcc / @code{as}: Split into two lines; for example@w{ }
29311 Intel: Keep on one line; for example @code{rep stosl}@w{ }
29319 @emph{Order of operands}@w{ }
29321 gcc / @code{as}: Source first; for example @code{movw $4, %eax}@w{ }
29322 Intel: Destination first; for example @code{mov eax, 4}@w{ }
29328 @node A Simple Example of Inline Assembler,Output Variables in Inline Assembler,Basic Assembler Syntax,Inline Assembler
29329 @anchor{gnat_ugn/inline_assembler a-simple-example-of-inline-assembler}@anchor{251}@anchor{gnat_ugn/inline_assembler id3}@anchor{252}
29330 @section A Simple Example of Inline Assembler
29333 The following example will generate a single assembly language statement,
29334 @code{nop}, which does nothing. Despite its lack of run-time effect,
29335 the example will be useful in illustrating the basics of
29336 the Inline Assembler facility.
29341 with System.Machine_Code; use System.Machine_Code;
29342 procedure Nothing is
29349 @code{Asm} is a procedure declared in package @code{System.Machine_Code};
29350 here it takes one parameter, a @emph{template string} that must be a static
29351 expression and that will form the generated instruction.
29352 @code{Asm} may be regarded as a compile-time procedure that parses
29353 the template string and additional parameters (none here),
29354 from which it generates a sequence of assembly language instructions.
29356 The examples in this chapter will illustrate several of the forms
29357 for invoking @code{Asm}; a complete specification of the syntax
29358 is found in the @code{Machine_Code_Insertions} section of the
29359 @cite{GNAT Reference Manual}.
29361 Under the standard GNAT conventions, the @code{Nothing} procedure
29362 should be in a file named @code{nothing.adb}.
29363 You can build the executable in the usual way:
29372 However, the interesting aspect of this example is not its run-time behavior
29373 but rather the generated assembly code.
29374 To see this output, invoke the compiler as follows:
29379 $ gcc -c -S -fomit-frame-pointer -gnatp nothing.adb
29383 where the options are:
29394 compile only (no bind or link)
29403 generate assembler listing
29410 @item @code{-fomit-frame-pointer}
29412 do not set up separate stack frames
29419 @item @code{-gnatp}
29421 do not add runtime checks
29425 This gives a human-readable assembler version of the code. The resulting
29426 file will have the same name as the Ada source file, but with a @code{.s}
29427 extension. In our example, the file @code{nothing.s} has the following
29433 .file "nothing.adb"
29435 ___gnu_compiled_ada:
29438 .globl __ada_nothing
29450 The assembly code you included is clearly indicated by
29451 the compiler, between the @code{#APP} and @code{#NO_APP}
29452 delimiters. The character before the 'APP' and 'NOAPP'
29453 can differ on different targets. For example, GNU/Linux uses '#APP' while
29454 on NT you will see '/APP'.
29456 If you make a mistake in your assembler code (such as using the
29457 wrong size modifier, or using a wrong operand for the instruction) GNAT
29458 will report this error in a temporary file, which will be deleted when
29459 the compilation is finished. Generating an assembler file will help
29460 in such cases, since you can assemble this file separately using the
29461 @code{as} assembler that comes with gcc.
29463 Assembling the file using the command
29472 will give you error messages whose lines correspond to the assembler
29473 input file, so you can easily find and correct any mistakes you made.
29474 If there are no errors, @code{as} will generate an object file
29475 @code{nothing.out}.
29477 @node Output Variables in Inline Assembler,Input Variables in Inline Assembler,A Simple Example of Inline Assembler,Inline Assembler
29478 @anchor{gnat_ugn/inline_assembler id4}@anchor{253}@anchor{gnat_ugn/inline_assembler output-variables-in-inline-assembler}@anchor{254}
29479 @section Output Variables in Inline Assembler
29482 The examples in this section, showing how to access the processor flags,
29483 illustrate how to specify the destination operands for assembly language
29489 with Interfaces; use Interfaces;
29490 with Ada.Text_IO; use Ada.Text_IO;
29491 with System.Machine_Code; use System.Machine_Code;
29492 procedure Get_Flags is
29493 Flags : Unsigned_32;
29496 Asm ("pushfl" & LF & HT & -- push flags on stack
29497 "popl %%eax" & LF & HT & -- load eax with flags
29498 "movl %%eax, %0", -- store flags in variable
29499 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
29500 Put_Line ("Flags register:" & Flags'Img);
29505 In order to have a nicely aligned assembly listing, we have separated
29506 multiple assembler statements in the Asm template string with linefeed
29507 (ASCII.LF) and horizontal tab (ASCII.HT) characters.
29508 The resulting section of the assembly output file is:
29516 movl %eax, -40(%ebp)
29521 It would have been legal to write the Asm invocation as:
29526 Asm ("pushfl popl %%eax movl %%eax, %0")
29530 but in the generated assembler file, this would come out as:
29536 pushfl popl %eax movl %eax, -40(%ebp)
29541 which is not so convenient for the human reader.
29543 We use Ada comments
29544 at the end of each line to explain what the assembler instructions
29545 actually do. This is a useful convention.
29547 When writing Inline Assembler instructions, you need to precede each register
29548 and variable name with a percent sign. Since the assembler already requires
29549 a percent sign at the beginning of a register name, you need two consecutive
29550 percent signs for such names in the Asm template string, thus @code{%%eax}.
29551 In the generated assembly code, one of the percent signs will be stripped off.
29553 Names such as @code{%0}, @code{%1}, @code{%2}, etc., denote input or output
29554 variables: operands you later define using @code{Input} or @code{Output}
29555 parameters to @code{Asm}.
29556 An output variable is illustrated in
29557 the third statement in the Asm template string:
29566 The intent is to store the contents of the eax register in a variable that can
29567 be accessed in Ada. Simply writing @code{movl %%eax, Flags} would not
29568 necessarily work, since the compiler might optimize by using a register
29569 to hold Flags, and the expansion of the @code{movl} instruction would not be
29570 aware of this optimization. The solution is not to store the result directly
29571 but rather to advise the compiler to choose the correct operand form;
29572 that is the purpose of the @code{%0} output variable.
29574 Information about the output variable is supplied in the @code{Outputs}
29575 parameter to @code{Asm}:
29580 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
29584 The output is defined by the @code{Asm_Output} attribute of the target type;
29585 the general format is
29590 Type'Asm_Output (constraint_string, variable_name)
29594 The constraint string directs the compiler how
29595 to store/access the associated variable. In the example
29600 Unsigned_32'Asm_Output ("=m", Flags);
29604 the @code{"m"} (memory) constraint tells the compiler that the variable
29605 @code{Flags} should be stored in a memory variable, thus preventing
29606 the optimizer from keeping it in a register. In contrast,
29611 Unsigned_32'Asm_Output ("=r", Flags);
29615 uses the @code{"r"} (register) constraint, telling the compiler to
29616 store the variable in a register.
29618 If the constraint is preceded by the equal character '=', it tells
29619 the compiler that the variable will be used to store data into it.
29621 In the @code{Get_Flags} example, we used the @code{"g"} (global) constraint,
29622 allowing the optimizer to choose whatever it deems best.
29624 There are a fairly large number of constraints, but the ones that are
29625 most useful (for the Intel x86 processor) are the following:
29630 @multitable {xxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx}
29645 global (i.e., can be stored anywhere)
29717 use one of eax, ebx, ecx or edx
29725 use one of eax, ebx, ecx, edx, esi or edi
29731 The full set of constraints is described in the gcc and @code{as}
29732 documentation; note that it is possible to combine certain constraints
29733 in one constraint string.
29735 You specify the association of an output variable with an assembler operand
29736 through the @code{%@emph{n}} notation, where @emph{n} is a non-negative
29742 Asm ("pushfl" & LF & HT & -- push flags on stack
29743 "popl %%eax" & LF & HT & -- load eax with flags
29744 "movl %%eax, %0", -- store flags in variable
29745 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
29749 @code{%0} will be replaced in the expanded code by the appropriate operand,
29751 the compiler decided for the @code{Flags} variable.
29753 In general, you may have any number of output variables:
29759 Count the operands starting at 0; thus @code{%0}, @code{%1}, etc.
29762 Specify the @code{Outputs} parameter as a parenthesized comma-separated list
29763 of @code{Asm_Output} attributes
29771 Asm ("movl %%eax, %0" & LF & HT &
29772 "movl %%ebx, %1" & LF & HT &
29774 Outputs => (Unsigned_32'Asm_Output ("=g", Var_A), -- %0 = Var_A
29775 Unsigned_32'Asm_Output ("=g", Var_B), -- %1 = Var_B
29776 Unsigned_32'Asm_Output ("=g", Var_C))); -- %2 = Var_C
29780 where @code{Var_A}, @code{Var_B}, and @code{Var_C} are variables
29781 in the Ada program.
29783 As a variation on the @code{Get_Flags} example, we can use the constraints
29784 string to direct the compiler to store the eax register into the @code{Flags}
29785 variable, instead of including the store instruction explicitly in the
29786 @code{Asm} template string:
29791 with Interfaces; use Interfaces;
29792 with Ada.Text_IO; use Ada.Text_IO;
29793 with System.Machine_Code; use System.Machine_Code;
29794 procedure Get_Flags_2 is
29795 Flags : Unsigned_32;
29798 Asm ("pushfl" & LF & HT & -- push flags on stack
29799 "popl %%eax", -- save flags in eax
29800 Outputs => Unsigned_32'Asm_Output ("=a", Flags));
29801 Put_Line ("Flags register:" & Flags'Img);
29806 The @code{"a"} constraint tells the compiler that the @code{Flags}
29807 variable will come from the eax register. Here is the resulting code:
29816 movl %eax,-40(%ebp)
29820 The compiler generated the store of eax into Flags after
29821 expanding the assembler code.
29823 Actually, there was no need to pop the flags into the eax register;
29824 more simply, we could just pop the flags directly into the program variable:
29829 with Interfaces; use Interfaces;
29830 with Ada.Text_IO; use Ada.Text_IO;
29831 with System.Machine_Code; use System.Machine_Code;
29832 procedure Get_Flags_3 is
29833 Flags : Unsigned_32;
29836 Asm ("pushfl" & LF & HT & -- push flags on stack
29837 "pop %0", -- save flags in Flags
29838 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
29839 Put_Line ("Flags register:" & Flags'Img);
29844 @node Input Variables in Inline Assembler,Inlining Inline Assembler Code,Output Variables in Inline Assembler,Inline Assembler
29845 @anchor{gnat_ugn/inline_assembler id5}@anchor{255}@anchor{gnat_ugn/inline_assembler input-variables-in-inline-assembler}@anchor{256}
29846 @section Input Variables in Inline Assembler
29849 The example in this section illustrates how to specify the source operands
29850 for assembly language statements.
29851 The program simply increments its input value by 1:
29856 with Interfaces; use Interfaces;
29857 with Ada.Text_IO; use Ada.Text_IO;
29858 with System.Machine_Code; use System.Machine_Code;
29859 procedure Increment is
29861 function Incr (Value : Unsigned_32) return Unsigned_32 is
29862 Result : Unsigned_32;
29865 Outputs => Unsigned_32'Asm_Output ("=a", Result),
29866 Inputs => Unsigned_32'Asm_Input ("a", Value));
29870 Value : Unsigned_32;
29874 Put_Line ("Value before is" & Value'Img);
29875 Value := Incr (Value);
29876 Put_Line ("Value after is" & Value'Img);
29881 The @code{Outputs} parameter to @code{Asm} specifies
29882 that the result will be in the eax register and that it is to be stored
29883 in the @code{Result} variable.
29885 The @code{Inputs} parameter looks much like the @code{Outputs} parameter,
29886 but with an @code{Asm_Input} attribute.
29887 The @code{"="} constraint, indicating an output value, is not present.
29889 You can have multiple input variables, in the same way that you can have more
29890 than one output variable.
29892 The parameter count (%0, %1) etc, still starts at the first output statement,
29893 and continues with the input statements.
29895 Just as the @code{Outputs} parameter causes the register to be stored into the
29896 target variable after execution of the assembler statements, so does the
29897 @code{Inputs} parameter cause its variable to be loaded into the register
29898 before execution of the assembler statements.
29900 Thus the effect of the @code{Asm} invocation is:
29906 load the 32-bit value of @code{Value} into eax
29909 execute the @code{incl %eax} instruction
29912 store the contents of eax into the @code{Result} variable
29915 The resulting assembler file (with @code{-O2} optimization) contains:
29920 _increment__incr.1:
29933 @node Inlining Inline Assembler Code,Other Asm Functionality,Input Variables in Inline Assembler,Inline Assembler
29934 @anchor{gnat_ugn/inline_assembler id6}@anchor{257}@anchor{gnat_ugn/inline_assembler inlining-inline-assembler-code}@anchor{258}
29935 @section Inlining Inline Assembler Code
29938 For a short subprogram such as the @code{Incr} function in the previous
29939 section, the overhead of the call and return (creating / deleting the stack
29940 frame) can be significant, compared to the amount of code in the subprogram
29941 body. A solution is to apply Ada's @code{Inline} pragma to the subprogram,
29942 which directs the compiler to expand invocations of the subprogram at the
29943 point(s) of call, instead of setting up a stack frame for out-of-line calls.
29944 Here is the resulting program:
29949 with Interfaces; use Interfaces;
29950 with Ada.Text_IO; use Ada.Text_IO;
29951 with System.Machine_Code; use System.Machine_Code;
29952 procedure Increment_2 is
29954 function Incr (Value : Unsigned_32) return Unsigned_32 is
29955 Result : Unsigned_32;
29958 Outputs => Unsigned_32'Asm_Output ("=a", Result),
29959 Inputs => Unsigned_32'Asm_Input ("a", Value));
29962 pragma Inline (Increment);
29964 Value : Unsigned_32;
29968 Put_Line ("Value before is" & Value'Img);
29969 Value := Increment (Value);
29970 Put_Line ("Value after is" & Value'Img);
29975 Compile the program with both optimization (@code{-O2}) and inlining
29976 (@code{-gnatn}) enabled.
29978 The @code{Incr} function is still compiled as usual, but at the
29979 point in @code{Increment} where our function used to be called:
29985 call _increment__incr.1
29989 the code for the function body directly appears:
30002 thus saving the overhead of stack frame setup and an out-of-line call.
30004 @node Other Asm Functionality,,Inlining Inline Assembler Code,Inline Assembler
30005 @anchor{gnat_ugn/inline_assembler other-asm-functionality}@anchor{259}@anchor{gnat_ugn/inline_assembler id7}@anchor{25a}
30006 @section Other @code{Asm} Functionality
30009 This section describes two important parameters to the @code{Asm}
30010 procedure: @code{Clobber}, which identifies register usage;
30011 and @code{Volatile}, which inhibits unwanted optimizations.
30014 * The Clobber Parameter::
30015 * The Volatile Parameter::
30019 @node The Clobber Parameter,The Volatile Parameter,,Other Asm Functionality
30020 @anchor{gnat_ugn/inline_assembler the-clobber-parameter}@anchor{25b}@anchor{gnat_ugn/inline_assembler id8}@anchor{25c}
30021 @subsection The @code{Clobber} Parameter
30024 One of the dangers of intermixing assembly language and a compiled language
30025 such as Ada is that the compiler needs to be aware of which registers are
30026 being used by the assembly code. In some cases, such as the earlier examples,
30027 the constraint string is sufficient to indicate register usage (e.g.,
30029 the eax register). But more generally, the compiler needs an explicit
30030 identification of the registers that are used by the Inline Assembly
30033 Using a register that the compiler doesn't know about
30034 could be a side effect of an instruction (like @code{mull}
30035 storing its result in both eax and edx).
30036 It can also arise from explicit register usage in your
30037 assembly code; for example:
30042 Asm ("movl %0, %%ebx" & LF & HT &
30044 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
30045 Inputs => Unsigned_32'Asm_Input ("g", Var_In));
30049 where the compiler (since it does not analyze the @code{Asm} template string)
30050 does not know you are using the ebx register.
30052 In such cases you need to supply the @code{Clobber} parameter to @code{Asm},
30053 to identify the registers that will be used by your assembly code:
30058 Asm ("movl %0, %%ebx" & LF & HT &
30060 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
30061 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
30066 The Clobber parameter is a static string expression specifying the
30067 register(s) you are using. Note that register names are @emph{not} prefixed
30068 by a percent sign. Also, if more than one register is used then their names
30069 are separated by commas; e.g., @code{"eax, ebx"}
30071 The @code{Clobber} parameter has several additional uses:
30077 Use 'register' name @code{cc} to indicate that flags might have changed
30080 Use 'register' name @code{memory} if you changed a memory location
30083 @node The Volatile Parameter,,The Clobber Parameter,Other Asm Functionality
30084 @anchor{gnat_ugn/inline_assembler the-volatile-parameter}@anchor{25d}@anchor{gnat_ugn/inline_assembler id9}@anchor{25e}
30085 @subsection The @code{Volatile} Parameter
30088 @geindex Volatile parameter
30090 Compiler optimizations in the presence of Inline Assembler may sometimes have
30091 unwanted effects. For example, when an @code{Asm} invocation with an input
30092 variable is inside a loop, the compiler might move the loading of the input
30093 variable outside the loop, regarding it as a one-time initialization.
30095 If this effect is not desired, you can disable such optimizations by setting
30096 the @code{Volatile} parameter to @code{True}; for example:
30101 Asm ("movl %0, %%ebx" & LF & HT &
30103 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
30104 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
30110 By default, @code{Volatile} is set to @code{False} unless there is no
30111 @code{Outputs} parameter.
30113 Although setting @code{Volatile} to @code{True} prevents unwanted
30114 optimizations, it will also disable other optimizations that might be
30115 important for efficiency. In general, you should set @code{Volatile}
30116 to @code{True} only if the compiler's optimizations have created
30119 @node GNU Free Documentation License,Index,Inline Assembler,Top
30120 @anchor{share/gnu_free_documentation_license gnu-fdl}@anchor{1}@anchor{share/gnu_free_documentation_license doc}@anchor{25f}@anchor{share/gnu_free_documentation_license gnu-free-documentation-license}@anchor{260}
30121 @chapter GNU Free Documentation License
30124 Version 1.3, 3 November 2008
30126 Copyright 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc
30127 @indicateurl{http://fsf.org/}
30129 Everyone is permitted to copy and distribute verbatim copies of this
30130 license document, but changing it is not allowed.
30134 The purpose of this License is to make a manual, textbook, or other
30135 functional and useful document "free" in the sense of freedom: to
30136 assure everyone the effective freedom to copy and redistribute it,
30137 with or without modifying it, either commercially or noncommercially.
30138 Secondarily, this License preserves for the author and publisher a way
30139 to get credit for their work, while not being considered responsible
30140 for modifications made by others.
30142 This License is a kind of "copyleft", which means that derivative
30143 works of the document must themselves be free in the same sense. It
30144 complements the GNU General Public License, which is a copyleft
30145 license designed for free software.
30147 We have designed this License in order to use it for manuals for free
30148 software, because free software needs free documentation: a free
30149 program should come with manuals providing the same freedoms that the
30150 software does. But this License is not limited to software manuals;
30151 it can be used for any textual work, regardless of subject matter or
30152 whether it is published as a printed book. We recommend this License
30153 principally for works whose purpose is instruction or reference.
30155 @strong{1. APPLICABILITY AND DEFINITIONS}
30157 This License applies to any manual or other work, in any medium, that
30158 contains a notice placed by the copyright holder saying it can be
30159 distributed under the terms of this License. Such a notice grants a
30160 world-wide, royalty-free license, unlimited in duration, to use that
30161 work under the conditions stated herein. The @strong{Document}, below,
30162 refers to any such manual or work. Any member of the public is a
30163 licensee, and is addressed as "@strong{you}". You accept the license if you
30164 copy, modify or distribute the work in a way requiring permission
30165 under copyright law.
30167 A "@strong{Modified Version}" of the Document means any work containing the
30168 Document or a portion of it, either copied verbatim, or with
30169 modifications and/or translated into another language.
30171 A "@strong{Secondary Section}" is a named appendix or a front-matter section of
30172 the Document that deals exclusively with the relationship of the
30173 publishers or authors of the Document to the Document's overall subject
30174 (or to related matters) and contains nothing that could fall directly
30175 within that overall subject. (Thus, if the Document is in part a
30176 textbook of mathematics, a Secondary Section may not explain any
30177 mathematics.) The relationship could be a matter of historical
30178 connection with the subject or with related matters, or of legal,
30179 commercial, philosophical, ethical or political position regarding
30182 The "@strong{Invariant Sections}" are certain Secondary Sections whose titles
30183 are designated, as being those of Invariant Sections, in the notice
30184 that says that the Document is released under this License. If a
30185 section does not fit the above definition of Secondary then it is not
30186 allowed to be designated as Invariant. The Document may contain zero
30187 Invariant Sections. If the Document does not identify any Invariant
30188 Sections then there are none.
30190 The "@strong{Cover Texts}" are certain short passages of text that are listed,
30191 as Front-Cover Texts or Back-Cover Texts, in the notice that says that
30192 the Document is released under this License. A Front-Cover Text may
30193 be at most 5 words, and a Back-Cover Text may be at most 25 words.
30195 A "@strong{Transparent}" copy of the Document means a machine-readable copy,
30196 represented in a format whose specification is available to the
30197 general public, that is suitable for revising the document
30198 straightforwardly with generic text editors or (for images composed of
30199 pixels) generic paint programs or (for drawings) some widely available
30200 drawing editor, and that is suitable for input to text formatters or
30201 for automatic translation to a variety of formats suitable for input
30202 to text formatters. A copy made in an otherwise Transparent file
30203 format whose markup, or absence of markup, has been arranged to thwart
30204 or discourage subsequent modification by readers is not Transparent.
30205 An image format is not Transparent if used for any substantial amount
30206 of text. A copy that is not "Transparent" is called @strong{Opaque}.
30208 Examples of suitable formats for Transparent copies include plain
30209 ASCII without markup, Texinfo input format, LaTeX input format, SGML
30210 or XML using a publicly available DTD, and standard-conforming simple
30211 HTML, PostScript or PDF designed for human modification. Examples of
30212 transparent image formats include PNG, XCF and JPG. Opaque formats
30213 include proprietary formats that can be read and edited only by
30214 proprietary word processors, SGML or XML for which the DTD and/or
30215 processing tools are not generally available, and the
30216 machine-generated HTML, PostScript or PDF produced by some word
30217 processors for output purposes only.
30219 The "@strong{Title Page}" means, for a printed book, the title page itself,
30220 plus such following pages as are needed to hold, legibly, the material
30221 this License requires to appear in the title page. For works in
30222 formats which do not have any title page as such, "Title Page" means
30223 the text near the most prominent appearance of the work's title,
30224 preceding the beginning of the body of the text.
30226 The "@strong{publisher}" means any person or entity that distributes
30227 copies of the Document to the public.
30229 A section "@strong{Entitled XYZ}" means a named subunit of the Document whose
30230 title either is precisely XYZ or contains XYZ in parentheses following
30231 text that translates XYZ in another language. (Here XYZ stands for a
30232 specific section name mentioned below, such as "@strong{Acknowledgements}",
30233 "@strong{Dedications}", "@strong{Endorsements}", or "@strong{History}".)
30234 To "@strong{Preserve the Title}"
30235 of such a section when you modify the Document means that it remains a
30236 section "Entitled XYZ" according to this definition.
30238 The Document may include Warranty Disclaimers next to the notice which
30239 states that this License applies to the Document. These Warranty
30240 Disclaimers are considered to be included by reference in this
30241 License, but only as regards disclaiming warranties: any other
30242 implication that these Warranty Disclaimers may have is void and has
30243 no effect on the meaning of this License.
30245 @strong{2. VERBATIM COPYING}
30247 You may copy and distribute the Document in any medium, either
30248 commercially or noncommercially, provided that this License, the
30249 copyright notices, and the license notice saying this License applies
30250 to the Document are reproduced in all copies, and that you add no other
30251 conditions whatsoever to those of this License. You may not use
30252 technical measures to obstruct or control the reading or further
30253 copying of the copies you make or distribute. However, you may accept
30254 compensation in exchange for copies. If you distribute a large enough
30255 number of copies you must also follow the conditions in section 3.
30257 You may also lend copies, under the same conditions stated above, and
30258 you may publicly display copies.
30260 @strong{3. COPYING IN QUANTITY}
30262 If you publish printed copies (or copies in media that commonly have
30263 printed covers) of the Document, numbering more than 100, and the
30264 Document's license notice requires Cover Texts, you must enclose the
30265 copies in covers that carry, clearly and legibly, all these Cover
30266 Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
30267 the back cover. Both covers must also clearly and legibly identify
30268 you as the publisher of these copies. The front cover must present
30269 the full title with all words of the title equally prominent and
30270 visible. You may add other material on the covers in addition.
30271 Copying with changes limited to the covers, as long as they preserve
30272 the title of the Document and satisfy these conditions, can be treated
30273 as verbatim copying in other respects.
30275 If the required texts for either cover are too voluminous to fit
30276 legibly, you should put the first ones listed (as many as fit
30277 reasonably) on the actual cover, and continue the rest onto adjacent
30280 If you publish or distribute Opaque copies of the Document numbering
30281 more than 100, you must either include a machine-readable Transparent
30282 copy along with each Opaque copy, or state in or with each Opaque copy
30283 a computer-network location from which the general network-using
30284 public has access to download using public-standard network protocols
30285 a complete Transparent copy of the Document, free of added material.
30286 If you use the latter option, you must take reasonably prudent steps,
30287 when you begin distribution of Opaque copies in quantity, to ensure
30288 that this Transparent copy will remain thus accessible at the stated
30289 location until at least one year after the last time you distribute an
30290 Opaque copy (directly or through your agents or retailers) of that
30291 edition to the public.
30293 It is requested, but not required, that you contact the authors of the
30294 Document well before redistributing any large number of copies, to give
30295 them a chance to provide you with an updated version of the Document.
30297 @strong{4. MODIFICATIONS}
30299 You may copy and distribute a Modified Version of the Document under
30300 the conditions of sections 2 and 3 above, provided that you release
30301 the Modified Version under precisely this License, with the Modified
30302 Version filling the role of the Document, thus licensing distribution
30303 and modification of the Modified Version to whoever possesses a copy
30304 of it. In addition, you must do these things in the Modified Version:
30310 Use in the Title Page (and on the covers, if any) a title distinct
30311 from that of the Document, and from those of previous versions
30312 (which should, if there were any, be listed in the History section
30313 of the Document). You may use the same title as a previous version
30314 if the original publisher of that version gives permission.
30317 List on the Title Page, as authors, one or more persons or entities
30318 responsible for authorship of the modifications in the Modified
30319 Version, together with at least five of the principal authors of the
30320 Document (all of its principal authors, if it has fewer than five),
30321 unless they release you from this requirement.
30324 State on the Title page the name of the publisher of the
30325 Modified Version, as the publisher.
30328 Preserve all the copyright notices of the Document.
30331 Add an appropriate copyright notice for your modifications
30332 adjacent to the other copyright notices.
30335 Include, immediately after the copyright notices, a license notice
30336 giving the public permission to use the Modified Version under the
30337 terms of this License, in the form shown in the Addendum below.
30340 Preserve in that license notice the full lists of Invariant Sections
30341 and required Cover Texts given in the Document's license notice.
30344 Include an unaltered copy of this License.
30347 Preserve the section Entitled "History", Preserve its Title, and add
30348 to it an item stating at least the title, year, new authors, and
30349 publisher of the Modified Version as given on the Title Page. If
30350 there is no section Entitled "History" in the Document, create one
30351 stating the title, year, authors, and publisher of the Document as
30352 given on its Title Page, then add an item describing the Modified
30353 Version as stated in the previous sentence.
30356 Preserve the network location, if any, given in the Document for
30357 public access to a Transparent copy of the Document, and likewise
30358 the network locations given in the Document for previous versions
30359 it was based on. These may be placed in the "History" section.
30360 You may omit a network location for a work that was published at
30361 least four years before the Document itself, or if the original
30362 publisher of the version it refers to gives permission.
30365 For any section Entitled "Acknowledgements" or "Dedications",
30366 Preserve the Title of the section, and preserve in the section all
30367 the substance and tone of each of the contributor acknowledgements
30368 and/or dedications given therein.
30371 Preserve all the Invariant Sections of the Document,
30372 unaltered in their text and in their titles. Section numbers
30373 or the equivalent are not considered part of the section titles.
30376 Delete any section Entitled "Endorsements". Such a section
30377 may not be included in the Modified Version.
30380 Do not retitle any existing section to be Entitled "Endorsements"
30381 or to conflict in title with any Invariant Section.
30384 Preserve any Warranty Disclaimers.
30387 If the Modified Version includes new front-matter sections or
30388 appendices that qualify as Secondary Sections and contain no material
30389 copied from the Document, you may at your option designate some or all
30390 of these sections as invariant. To do this, add their titles to the
30391 list of Invariant Sections in the Modified Version's license notice.
30392 These titles must be distinct from any other section titles.
30394 You may add a section Entitled "Endorsements", provided it contains
30395 nothing but endorsements of your Modified Version by various
30396 parties---for example, statements of peer review or that the text has
30397 been approved by an organization as the authoritative definition of a
30400 You may add a passage of up to five words as a Front-Cover Text, and a
30401 passage of up to 25 words as a Back-Cover Text, to the end of the list
30402 of Cover Texts in the Modified Version. Only one passage of
30403 Front-Cover Text and one of Back-Cover Text may be added by (or
30404 through arrangements made by) any one entity. If the Document already
30405 includes a cover text for the same cover, previously added by you or
30406 by arrangement made by the same entity you are acting on behalf of,
30407 you may not add another; but you may replace the old one, on explicit
30408 permission from the previous publisher that added the old one.
30410 The author(s) and publisher(s) of the Document do not by this License
30411 give permission to use their names for publicity for or to assert or
30412 imply endorsement of any Modified Version.
30414 @strong{5. COMBINING DOCUMENTS}
30416 You may combine the Document with other documents released under this
30417 License, under the terms defined in section 4 above for modified
30418 versions, provided that you include in the combination all of the
30419 Invariant Sections of all of the original documents, unmodified, and
30420 list them all as Invariant Sections of your combined work in its
30421 license notice, and that you preserve all their Warranty Disclaimers.
30423 The combined work need only contain one copy of this License, and
30424 multiple identical Invariant Sections may be replaced with a single
30425 copy. If there are multiple Invariant Sections with the same name but
30426 different contents, make the title of each such section unique by
30427 adding at the end of it, in parentheses, the name of the original
30428 author or publisher of that section if known, or else a unique number.
30429 Make the same adjustment to the section titles in the list of
30430 Invariant Sections in the license notice of the combined work.
30432 In the combination, you must combine any sections Entitled "History"
30433 in the various original documents, forming one section Entitled
30434 "History"; likewise combine any sections Entitled "Acknowledgements",
30435 and any sections Entitled "Dedications". You must delete all sections
30436 Entitled "Endorsements".
30438 @strong{6. COLLECTIONS OF DOCUMENTS}
30440 You may make a collection consisting of the Document and other documents
30441 released under this License, and replace the individual copies of this
30442 License in the various documents with a single copy that is included in
30443 the collection, provided that you follow the rules of this License for
30444 verbatim copying of each of the documents in all other respects.
30446 You may extract a single document from such a collection, and distribute
30447 it individually under this License, provided you insert a copy of this
30448 License into the extracted document, and follow this License in all
30449 other respects regarding verbatim copying of that document.
30451 @strong{7. AGGREGATION WITH INDEPENDENT WORKS}
30453 A compilation of the Document or its derivatives with other separate
30454 and independent documents or works, in or on a volume of a storage or
30455 distribution medium, is called an "aggregate" if the copyright
30456 resulting from the compilation is not used to limit the legal rights
30457 of the compilation's users beyond what the individual works permit.
30458 When the Document is included in an aggregate, this License does not
30459 apply to the other works in the aggregate which are not themselves
30460 derivative works of the Document.
30462 If the Cover Text requirement of section 3 is applicable to these
30463 copies of the Document, then if the Document is less than one half of
30464 the entire aggregate, the Document's Cover Texts may be placed on
30465 covers that bracket the Document within the aggregate, or the
30466 electronic equivalent of covers if the Document is in electronic form.
30467 Otherwise they must appear on printed covers that bracket the whole
30470 @strong{8. TRANSLATION}
30472 Translation is considered a kind of modification, so you may
30473 distribute translations of the Document under the terms of section 4.
30474 Replacing Invariant Sections with translations requires special
30475 permission from their copyright holders, but you may include
30476 translations of some or all Invariant Sections in addition to the
30477 original versions of these Invariant Sections. You may include a
30478 translation of this License, and all the license notices in the
30479 Document, and any Warranty Disclaimers, provided that you also include
30480 the original English version of this License and the original versions
30481 of those notices and disclaimers. In case of a disagreement between
30482 the translation and the original version of this License or a notice
30483 or disclaimer, the original version will prevail.
30485 If a section in the Document is Entitled "Acknowledgements",
30486 "Dedications", or "History", the requirement (section 4) to Preserve
30487 its Title (section 1) will typically require changing the actual
30490 @strong{9. TERMINATION}
30492 You may not copy, modify, sublicense, or distribute the Document
30493 except as expressly provided under this License. Any attempt
30494 otherwise to copy, modify, sublicense, or distribute it is void, and
30495 will automatically terminate your rights under this License.
30497 However, if you cease all violation of this License, then your license
30498 from a particular copyright holder is reinstated (a) provisionally,
30499 unless and until the copyright holder explicitly and finally
30500 terminates your license, and (b) permanently, if the copyright holder
30501 fails to notify you of the violation by some reasonable means prior to
30502 60 days after the cessation.
30504 Moreover, your license from a particular copyright holder is
30505 reinstated permanently if the copyright holder notifies you of the
30506 violation by some reasonable means, this is the first time you have
30507 received notice of violation of this License (for any work) from that
30508 copyright holder, and you cure the violation prior to 30 days after
30509 your receipt of the notice.
30511 Termination of your rights under this section does not terminate the
30512 licenses of parties who have received copies or rights from you under
30513 this License. If your rights have been terminated and not permanently
30514 reinstated, receipt of a copy of some or all of the same material does
30515 not give you any rights to use it.
30517 @strong{10. FUTURE REVISIONS OF THIS LICENSE}
30519 The Free Software Foundation may publish new, revised versions
30520 of the GNU Free Documentation License from time to time. Such new
30521 versions will be similar in spirit to the present version, but may
30522 differ in detail to address new problems or concerns. See
30523 @indicateurl{http://www.gnu.org/copyleft/}.
30525 Each version of the License is given a distinguishing version number.
30526 If the Document specifies that a particular numbered version of this
30527 License "or any later version" applies to it, you have the option of
30528 following the terms and conditions either of that specified version or
30529 of any later version that has been published (not as a draft) by the
30530 Free Software Foundation. If the Document does not specify a version
30531 number of this License, you may choose any version ever published (not
30532 as a draft) by the Free Software Foundation. If the Document
30533 specifies that a proxy can decide which future versions of this
30534 License can be used, that proxy's public statement of acceptance of a
30535 version permanently authorizes you to choose that version for the
30538 @strong{11. RELICENSING}
30540 "Massive Multiauthor Collaboration Site" (or "MMC Site") means any
30541 World Wide Web server that publishes copyrightable works and also
30542 provides prominent facilities for anybody to edit those works. A
30543 public wiki that anybody can edit is an example of such a server. A
30544 "Massive Multiauthor Collaboration" (or "MMC") contained in the
30545 site means any set of copyrightable works thus published on the MMC
30548 "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0
30549 license published by Creative Commons Corporation, a not-for-profit
30550 corporation with a principal place of business in San Francisco,
30551 California, as well as future copyleft versions of that license
30552 published by that same organization.
30554 "Incorporate" means to publish or republish a Document, in whole or
30555 in part, as part of another Document.
30557 An MMC is "eligible for relicensing" if it is licensed under this
30558 License, and if all works that were first published under this License
30559 somewhere other than this MMC, and subsequently incorporated in whole
30560 or in part into the MMC, (1) had no cover texts or invariant sections,
30561 and (2) were thus incorporated prior to November 1, 2008.
30563 The operator of an MMC Site may republish an MMC contained in the site
30564 under CC-BY-SA on the same site at any time before August 1, 2009,
30565 provided the MMC is eligible for relicensing.
30567 @strong{ADDENDUM: How to use this License for your documents}
30569 To use this License in a document you have written, include a copy of
30570 the License in the document and put the following copyright and
30571 license notices just after the title page:
30575 Copyright © YEAR YOUR NAME.
30576 Permission is granted to copy, distribute and/or modify this document
30577 under the terms of the GNU Free Documentation License, Version 1.3
30578 or any later version published by the Free Software Foundation;
30579 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.
30580 A copy of the license is included in the section entitled "GNU
30581 Free Documentation License".
30584 If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts,
30585 replace the "with ... Texts." line with this:
30589 with the Invariant Sections being LIST THEIR TITLES, with the
30590 Front-Cover Texts being LIST, and with the Back-Cover Texts being LIST.
30593 If you have Invariant Sections without Cover Texts, or some other
30594 combination of the three, merge those two alternatives to suit the
30597 If your document contains nontrivial examples of program code, we
30598 recommend releasing these examples in parallel under your choice of
30599 free software license, such as the GNU General Public License,
30600 to permit their use in free software.
30602 @node Index,,GNU Free Documentation License,Top
30609 @anchor{gnat_ugn/gnat_utility_programs switches-related-to-project-files}@w{ }