gcc.git
16 years agore PR middle-end/37010 (-mno-accumulate-outgoing-args doesn't work with stack alignment)
H.J. Lu [Wed, 6 Aug 2008 15:43:46 +0000 (15:43 +0000)]
re PR middle-end/37010 (-mno-accumulate-outgoing-args doesn't work with stack alignment)

gcc/

2008-08-06  H.J. Lu  <hongjiu.lu@intel.com>

PR middle-end/37010
* calls.c (expand_call): Use the biggest preferred stack
boundary.

gcc/testsuite/

2008-08-06  H.J. Lu  <hongjiu.lu@intel.com>

PR middle-end/37010
* gcc.dg/torture/stackalign/push-1.c: New.

From-SVN: r138808

16 years agore PR target/36613 (likely codegen bug)
Michael Matz [Wed, 6 Aug 2008 15:34:45 +0000 (15:34 +0000)]
re PR target/36613 (likely codegen bug)

        PR target/36613

        * reload.c (push_reload): Merge in,out,in_reg,out_reg members
        for reused reload, instead of overwriting them.

        * gcc.target/i386/pr36613.c: New testcase.

From-SVN: r138807

16 years agore PR middle-end/37009 (No need to align stack when incoming stack is aligned)
H.J. Lu [Wed, 6 Aug 2008 15:29:37 +0000 (08:29 -0700)]
re PR middle-end/37009 (No need to align stack when incoming stack is aligned)

gcc/

2008-08-06  H.J. Lu  <hongjiu.lu@intel.com>

PR middle-end/37009
* cfgexpand.c (expand_stack_alignment): Check parm_stack_boundary
for incoming stack boundary.

* function.c (assign_parm_find_entry_rtl): Update
parm_stack_boundary.

* function.h (rtl_data): Add parm_stack_boundary.

* config/i386/i386.c (ix86_finalize_stack_realign_flags): Check
parm_stack_boundary for incoming stack boundary.

gcc/testsuite/

2008-08-06  H.J. Lu  <hongjiu.lu@intel.com>

PR middle-end/37009
* gcc.dg/torture/stackalign/alloca-2.c: New.
* gcc.dg/torture/stackalign/alloca-3.c: Likewise.
* gcc.dg/torture/stackalign/alloca-4.c: Likewise.
* gcc.dg/torture/stackalign/vararg-3.c: Likewise.
* gcc.target/i386/incoming-1.c: Likewise.
* gcc.target/i386/incoming-2.c: Likewise.
* gcc.target/i386/incoming-3.c: Likewise.
* gcc.target/i386/incoming-4.c: Likewise.
* gcc.target/i386/incoming-5.c: Likewise.

From-SVN: r138806

16 years agojump.c (rtx_renumbered_equal_p): Do not call subreg_regno_offset for unrepresentable...
Joseph Myers [Wed, 6 Aug 2008 15:20:14 +0000 (16:20 +0100)]
jump.c (rtx_renumbered_equal_p): Do not call subreg_regno_offset for unrepresentable subregs or treat them as...

* jump.c (rtx_renumbered_equal_p): Do not call subreg_regno_offset
for unrepresentable subregs or treat them as equal to other regs
or subregs with the same register number.

From-SVN: r138804

16 years agoMake-lang.in: Use GCC_FOR_TARGET when dealing with s-oscons-tmplt.i.
Samuel Tardieu [Wed, 6 Aug 2008 13:55:07 +0000 (13:55 +0000)]
Make-lang.in: Use GCC_FOR_TARGET when dealing with s-oscons-tmplt.i.

    gcc/ada/
* gcc-interface/Make-lang.in: Use GCC_FOR_TARGET when dealing
with s-oscons-tmplt.i.

    libada/
* Makefile.in: Pass FLAGS_TO_PASS to sub-make for target
oscons.

From-SVN: r138800

16 years agoMake-lang.in (OSCONS_CPPFLAGS): Remove.
Samuel Tardieu [Wed, 6 Aug 2008 13:54:31 +0000 (13:54 +0000)]
Make-lang.in (OSCONS_CPPFLAGS): Remove.

    gcc/ada/
* gcc-interface/Make-lang.in (OSCONS_CPPFLAGS): Remove.

From-SVN: r138799

16 years agoacinclude.m4 ([GLIBCXX_CHECK_INT64_T]): Also check whether int64_t is actually a...
Paolo Carlini [Wed, 6 Aug 2008 12:25:00 +0000 (12:25 +0000)]
acinclude.m4 ([GLIBCXX_CHECK_INT64_T]): Also check whether int64_t is actually a typedef to long or long long.

2008-08-06  Paolo Carlini  <paolo.carlini@oracle.com>

* acinclude.m4 ([GLIBCXX_CHECK_INT64_T]): Also check whether int64_t
is actually a typedef to long or long long.
* include/bits/postypes.h: If int64_t is actually a typedef for
long or long long don't include <stdint.h> unnecessarily.
* configure: Regenerate.
* config.h.in: Likewise.

From-SVN: r138797

16 years agore PR middle-end/35432 (ICE with zero-sized array)
Aldy Hernandez [Wed, 6 Aug 2008 10:25:22 +0000 (10:25 +0000)]
re PR middle-end/35432 (ICE with zero-sized array)

        PR middle-end/35432
        * gimplify.c (gimplify_modify_expr): Do not optimize zero-sized types
        if want_value.
 testsuite/
        * gcc.c-torture/compile/pr35432.c: New file.

From-SVN: r138793

16 years agoMakefile.in, [...]: Remove GNAT.Sockets.Constants.
Arnaud Charlet [Wed, 6 Aug 2008 10:08:09 +0000 (12:08 +0200)]
Makefile.in, [...]: Remove GNAT.Sockets.Constants.

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* gcc-interface/Makefile.in,
g-socthi-vxworks.adb, g-socthi-mingw.adb, g-sttsne-vxworks.adb,
g-socthi.adb, g-socket.adb, g-socket.ads, g-sothco.ads,
g-soccon-linux-x86.ads, g-soccon-vxworks.ads, g-soccon-mingw.ads,
g-soccon-hpux-ia64.ads, g-soccon-irix.ads, g-soccon-linux-64.ads,
g-soccon-aix.ads, g-soccon-solaris.ads, g-soccon-lynxos.ads,
g-soccon-vms.ads, g-soccon.ads, g-soccon-freebsd.ads,
g-soccon-linux-ppc.ads, g-soccon-tru64.ads, g-soccon-hpux.ads,
g-soccon-solaris-64.ads, gen-oscons.c, g-soccon-darwin.ads,
g-soccon-mingw-64.ads, g-soccon-linux-mips.ads, g-soccon-rtems.ads:
Remove GNAT.Sockets.Constants. This internal package is replaced by
System.OS_Constants.

From-SVN: r138792

16 years agognat_ugn.texi: Clarify -gnato documentation
Robert Dewar [Wed, 6 Aug 2008 10:03:52 +0000 (12:03 +0200)]
gnat_ugn.texi: Clarify -gnato documentation

2008-08-06  Robert Dewar  <dewar@adacore.com>

* gnat_ugn.texi: Clarify -gnato documentation

From-SVN: r138791

16 years agosem_ch3.adb (Analyze_Component_Declaration): Protect against misuse of incomplete...
Ed Schonberg [Wed, 6 Aug 2008 10:03:21 +0000 (12:03 +0200)]
sem_ch3.adb (Analyze_Component_Declaration): Protect against misuse of incomplete type.

2008-08-06  Ed Schonberg  <schonberg@adacore.com>

* sem_ch3.adb (Analyze_Component_Declaration): Protect against misuse
of incomplete type.

* sem_ch8.adb (Analyze_Object_Renaming): Diagnose properly a renaming
of a formal parameter of an incomplete type. Improve error message for
other improper uses of incomplete types.

From-SVN: r138790

16 years agoMakefile.in: Now generate s-oscons-$(THREAD_KIND).
Thomas Quinot [Wed, 6 Aug 2008 09:43:33 +0000 (09:43 +0000)]
Makefile.in: Now generate s-oscons-$(THREAD_KIND).

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* Makefile.in: Now generate s-oscons-$(THREAD_KIND).

From-SVN: r138789

16 years agoMakefile.in, [...]: Remove obsolete targets referencing gen-soccon When generating...
Thomas Quinot [Wed, 6 Aug 2008 09:43:10 +0000 (11:43 +0200)]
Makefile.in, [...]: Remove obsolete targets referencing gen-soccon When generating s-oscons.ads...

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* gcc-interface/Makefile.in, gcc-interface/Make-lang.in:
Remove obsolete targets referencing gen-soccon
When generating s-oscons.ads, use a file name that includes the
THREAD_KIND, to ensure that the (potentially different) version from a
previous build with a different threads flavour does not get reused.

From-SVN: r138788

16 years agoMakefile.in, [...]: Remove obsolete targets referencing gen-soccon When generating...
Thomas Quinot [Wed, 6 Aug 2008 09:42:31 +0000 (11:42 +0200)]
Makefile.in, [...]: Remove obsolete targets referencing gen-soccon When generating s-oscons.ads...

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* gcc-interface/Makefile.in, gcc-interface/Make-lang.in:
Remove obsolete targets referencing gen-soccon
When generating s-oscons.ads, use a file name that includes the
THREAD_KIND, to ensure that the (potentially different) version from a
previous build with a different threads flavour does not get reused.

From-SVN: r138787

16 years agos-fileio.adb (Open): When file open operation fails...
Arnaud Charlet [Wed, 6 Aug 2008 09:41:04 +0000 (11:41 +0200)]
s-fileio.adb (Open): When file open operation fails...

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* s-fileio.adb (Open): When file open operation fails, raise Name_Error
only when the operating system reports a non-existing file or directory
(ENOENT), otherwise raise Name_Error.

From-SVN: r138786

16 years agoexp_ch11.adb: Minor reformatting
Thomas Quinot [Wed, 6 Aug 2008 09:35:27 +0000 (11:35 +0200)]
exp_ch11.adb: Minor reformatting

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* exp_ch11.adb: Minor reformatting

From-SVN: r138785

16 years agosem_ch3.adb (Access_Subprogram_Declaration): If the return type is incomplete...
Ed Schonberg [Wed, 6 Aug 2008 09:35:17 +0000 (11:35 +0200)]
sem_ch3.adb (Access_Subprogram_Declaration): If the return type is incomplete...

2008-08-06  Ed Schonberg  <schonberg@adacore.com>

* sem_ch3.adb (Access_Subprogram_Declaration): If the return type is
incomplete, add the access_to_subprogram type to the list of private
dependents only if the incomplete type will be completed in the current
scope.
(Build_Discriminant_Constraints): If the type of the discriminant is
access_to_variable, reject a constraint that is access_to_constant.

From-SVN: r138784

16 years agosem_res.adb: Minor reformatting
Thomas Quinot [Wed, 6 Aug 2008 09:35:06 +0000 (11:35 +0200)]
sem_res.adb: Minor reformatting

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* sem_res.adb: Minor reformatting

From-SVN: r138783

16 years agoNew tests.
Arnaud Charlet [Wed, 6 Aug 2008 09:16:07 +0000 (09:16 +0000)]
New tests.

From-SVN: r138782

16 years agog-socket-dummy.adb, [...]: New files.
Arnaud Charlet [Wed, 6 Aug 2008 08:57:21 +0000 (10:57 +0200)]
g-socket-dummy.adb, [...]: New files.

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* g-socket-dummy.adb, g-socket-dummy.ads, g-sothco-dummy.adb,
g-sothco-dummy.ads, g-socthi-dummy.adb, g-socthi-dummy.ads,
g-sttsne-dummy.ads: New files.

* gcc-interface/Makefile.in, Makefile.rtl: Use placeholder sources
with pragma Unimplemented_Unit for sockets packages on Nucleus.

From-SVN: r138781

16 years agog-comlin.adb (Set_Command_Line): Now that aliases can contain parameters, always...
Jerome Lambourg [Wed, 6 Aug 2008 08:52:44 +0000 (10:52 +0200)]
g-comlin.adb (Set_Command_Line): Now that aliases can contain parameters, always specify the expected separator.

2008-08-06  Jerome Lambourg  <lambourg@adacore.com>

* g-comlin.adb (Set_Command_Line): Now that aliases can contain
parameters, always specify the expected separator.

From-SVN: r138780

16 years agosem_res.adb (Resolve_Call): Use base type to determine whether a dereference is neede...
Ed Schonberg [Wed, 6 Aug 2008 08:52:36 +0000 (10:52 +0200)]
sem_res.adb (Resolve_Call): Use base type to determine whether a dereference is needed because a subtype of...

2008-08-06  Ed Schonberg  <schonberg@adacore.com>

* sem_res.adb (Resolve_Call): Use base type to determine whether a
dereference is needed because a subtype of an access_to_subprogram is
simply an access-subtype

From-SVN: r138779

16 years agoexp_disp (Expand_Interface_Actuals): Adds missing support for expansion of calls...
Javier Miranda [Wed, 6 Aug 2008 08:52:28 +0000 (10:52 +0200)]
exp_disp (Expand_Interface_Actuals): Adds missing support for expansion of calls to subprograms using selected...

2008-08-06  Javier Miranda  <miranda@adacore.com>

* exp_disp (Expand_Interface_Actuals): Adds missing support for
expansion of calls to subprograms using selected components.

From-SVN: r138778

16 years agoadaint.c: Another fix for ACL support on Windows.
Pascal Obry [Wed, 6 Aug 2008 08:52:19 +0000 (10:52 +0200)]
adaint.c: Another fix for ACL support on Windows.

2008-08-06  Pascal Obry  <obry@adacore.com>

* adaint.c: Another fix for ACL support on Windows.

From-SVN: r138777

16 years agosem_ch3.adb: Minor reformatting
Thomas Quinot [Wed, 6 Aug 2008 08:52:10 +0000 (10:52 +0200)]
sem_ch3.adb: Minor reformatting

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* sem_ch3.adb: Minor reformatting

From-SVN: r138776

16 years agog-comlin.adb (Define_Switch, [...]): New.
Arnaud Charlet [Wed, 6 Aug 2008 08:33:21 +0000 (10:33 +0200)]
g-comlin.adb (Define_Switch, [...]): New.

2008-08-06  Jerome Lambourg  <lambourg@adacore.com>

* g-comlin.adb (Define_Switch, Get_Switches): New.
(Can_Have_Parameter, Require_Parameter, Actual_Switch): New, used when
ungrouping switches.
(For_Each_Simple_Switch): Allow more control over parameters handling.
This generic method now allows ungrouping of switches with parameters
and switches with more than one letter after the prefix.
(Set_Command_Line): Take care of switches that are prefixed with a
switch handling parameters without delimiter (-gnatya and -gnaty3 for
example).
(Add_Switch, Remove_Switch): Handle parameters possibly present inside
a group, as in gnaty3aM80 (3 and 80 are parameters). Report status of
the operation.
(Start, Alias_Switches, Group_Switches): Take care of parameters
possibly present inside a group.

* g-comlin.ads (Define_Switch): New method used to define a list of
expected switches, that are necessary for correctly ungrouping switches
with more that one character after the prefix.
(Get_Switches): Method that builds a getopt string from the list of
switches as set previously by Define_Switch.
(Add_Switch, Remove_Switch): New versions of the methods, reporting the
status of the operation. Also allow the removal of switches with
parameters only.
(Command_Line_Configuration_Record): Maintain a list of expected
switches.

From-SVN: r138775

16 years agosem.adb: Minor rewording (comment)
Thomas Quinot [Wed, 6 Aug 2008 08:32:52 +0000 (10:32 +0200)]
sem.adb: Minor rewording (comment)

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* sem.adb: Minor rewording (comment)

From-SVN: r138774

16 years agosem_ch3.adb (Find_Type_Name): protect against duplicate incomplete declaration for...
Ed Schonberg [Wed, 6 Aug 2008 08:32:42 +0000 (10:32 +0200)]
sem_ch3.adb (Find_Type_Name): protect against duplicate incomplete declaration for the same type.

2008-08-06  Ed Schonberg  <schonberg@adacore.com>

* sem_ch3.adb (Find_Type_Name): protect against duplicate incomplete
declaration for the same type.

From-SVN: r138773

16 years agog-awk.adb (Finalize): Do not use directly objects of the type in the finalization...
Ed Schonberg [Wed, 6 Aug 2008 08:32:32 +0000 (10:32 +0200)]
g-awk.adb (Finalize): Do not use directly objects of the type in the finalization routine to prevent...

2008-08-06  Ed Schonberg  <schonberg@adacore.com>

* g-awk.adb (Finalize): Do not use directly objects of the type in the
finalization routine to prevent elaboration order anomalies in new
finalization scheme.

From-SVN: r138772

16 years agos-stausa.adb (Fill_Stack): Fixed pragma assert and top pattern mark in the case of...
Quentin Ochem [Wed, 6 Aug 2008 08:32:09 +0000 (10:32 +0200)]
s-stausa.adb (Fill_Stack): Fixed pragma assert and top pattern mark in the case of an empty pattern size.

2008-08-06  Quentin Ochem  <ochem@adacore.com>

* s-stausa.adb (Fill_Stack): Fixed pragma assert and top pattern mark
in the case of an empty pattern size.
(Compute_Result): Do not do any computation in the case of an empty
pattern size.
(Report_Result): Fixed computation of the overflow guard.

From-SVN: r138771

16 years agognat_ugn.texi: Document compiler switch -gnateG
Vincent Celier [Wed, 6 Aug 2008 08:32:00 +0000 (10:32 +0200)]
gnat_ugn.texi: Document compiler switch -gnateG

2008-08-06  Vincent Celier  <celier@adacore.com>

* gnat_ugn.texi: Document compiler switch -gnateG

From-SVN: r138770

16 years agoxnmake.adb: Use new XUtil package for platform independent text output.
Thomas Quinot [Wed, 6 Aug 2008 08:31:51 +0000 (10:31 +0200)]
xnmake.adb: Use new XUtil package for platform independent text output.

2008-08-06  Thomas Quinot  <quinot@adacore.com>

* xnmake.adb: Use new XUtil package for platform independent text
output.

From-SVN: r138769

16 years agodecl.c (gnat_to_gnu_param): Force 32bit descriptor if TARGET_MALLOC64 clear.
Doug Rupp [Wed, 6 Aug 2008 08:07:32 +0000 (08:07 +0000)]
decl.c (gnat_to_gnu_param): Force 32bit descriptor if TARGET_MALLOC64 clear.

2008-08-06  Doug Rupp  <rupp@adacore.com>

* gcc-interface/decl.c (gnat_to_gnu_param): Force 32bit descriptor if
TARGET_MALLOC64 clear.

* gcc-interface/utils2.c (build_call_alloc_dealloc): Force 32bit malloc
if TARGET_MALLOC64 clear.

* gcc-interface/gigi.h (TARGET_ABI_OPEN_VMS): Move here from utils2.c
(TARGET_MALLC64): New macro. Default to clear.

From-SVN: r138768

16 years ago2008-08-06 Doug Rupp <rupp@adacore.com>
Doug Rupp [Wed, 6 Aug 2008 07:57:04 +0000 (07:57 +0000)]
2008-08-06  Doug Rupp  <rupp@adacore.com>

* gcc-interface/utils2.c (snames.h) Include
(TARGET_ABI_OPEN_VMS): Initialize.
(build_call_alloc_dealloc); [TARGET_ABI_OPEN_VMS] Allocate on 32bit heap
for Convention C.

From-SVN: r138767

16 years agoexp_disp (Expand_Interface_Conversion): Freeze the entity associated with the target...
Arnaud Charlet [Wed, 6 Aug 2008 07:56:23 +0000 (09:56 +0200)]
exp_disp (Expand_Interface_Conversion): Freeze the entity associated with the target interface before expanding the...

2008-08-06  Javier Miranda  <miranda@adacore.com>

* exp_disp (Expand_Interface_Conversion): Freeze the entity associated
with the target interface before expanding the code of the interface
conversion.

From-SVN: r138766

16 years agosem_ch3.adb (Process_Discriminants): diagnose redundant or improper null exclusion...
Ed Schonberg [Wed, 6 Aug 2008 07:56:04 +0000 (09:56 +0200)]
sem_ch3.adb (Process_Discriminants): diagnose redundant or improper null exclusion in a discriminant declaration

2008-08-06  Ed Schonberg  <schonberg@adacore.com>

* sem_ch3.adb (Process_Discriminants): diagnose redundant or improper
null exclusion in a discriminant declaration

* sem_ch8.adb (Analyze_Object_Renaming): diagnose null exclusion
indicators when type is not an access type.

* sem_ch12.adb (Formal_Object_Declaration): diagnose null exclusion
indicators when type is not an access type.

From-SVN: r138765

16 years agopredict.c (maybe_hot_frequency_p): When profile is absent, all frequencies might...
Jan Hubicka [Wed, 6 Aug 2008 07:50:52 +0000 (09:50 +0200)]
predict.c (maybe_hot_frequency_p): When profile is absent, all frequencies might be hot.

* predict.c (maybe_hot_frequency_p): When profile is absent, all
frequencies might be hot.

From-SVN: r138764

16 years agoreload.c (find_reloads): Force constants into literal pool also if they are wrapped...
Andreas Krebbel [Wed, 6 Aug 2008 06:51:11 +0000 (06:51 +0000)]
reload.c (find_reloads): Force constants into literal pool also if they are wrapped in a SUBREG.

2008-08-06  Andreas Krebbel  <krebbel1@de.ibm.com>

* reload.c (find_reloads): Force constants into literal pool
also if they are wrapped in a SUBREG.

2008-08-06  Andreas Krebbel  <krebbel1@de.ibm.com>

  * gcc.c-torture/compile/20080806-1.c: New testcase.

From-SVN: r138763

16 years agore PR target/35659 (Miscompiled code with -O2 (but not with -O2 -funroll-loops) on...
Maxim Kuvyrkov [Wed, 6 Aug 2008 06:23:47 +0000 (06:23 +0000)]
re PR target/35659 (Miscompiled code with -O2 (but not with -O2 -funroll-loops) on ia64)

PR target/35659
* haifa-sched.c (sched_insn_is_legitimate_for_speculation_p): Move ...
* sched-deps.c (sched_insn_is_legitimate_for_speculation_p): ... here.
Don't allow predicated instructions for data speculation.
* sched-int.h (sched_insn_is_legitimate_for_speculation_p): Move
declaration.

From-SVN: r138759

16 years agohaifa-sched.c (extend_global): Split to extend_global_data and extend_region_data.
Maxim Kuvyrkov [Wed, 6 Aug 2008 06:22:33 +0000 (06:22 +0000)]
haifa-sched.c (extend_global): Split to extend_global_data and extend_region_data.

* haifa-sched.c (extend_global): Split to extend_global_data and
extend_region_data.  Update all uses.
(extend_all): Rename to extend_block_data.

* gcc.target/ia64/20080802-1.c: New test.

From-SVN: r138758

16 years agosched-rgn.c (new_ready): Check if instruction can be speculatively scheduled before...
Maxim Kuvyrkov [Wed, 6 Aug 2008 06:20:02 +0000 (06:20 +0000)]
sched-rgn.c (new_ready): Check if instruction can be speculatively scheduled before attempting speculation.

* sched-rgn.c (new_ready): Check if instruction can be
speculatively scheduled before attempting speculation.
(debug_rgn_dependencies): Remove wrongful assert.

From-SVN: r138757

16 years agore PR c++/37016 (member function pointer failure with optimization)
Jason Merrill [Wed, 6 Aug 2008 02:25:20 +0000 (22:25 -0400)]
re PR c++/37016 (member function pointer failure with optimization)

        PR c++/37016
        * decl.c (build_ptrmemfunc_type): Don't require structural
        comparison of PMF types.
        * tree.c (cp_build_qualified_type_real): Don't clear
        a valid TYPE_PTRMEMFUNC_TYPE.
        * typeck.c (cp_build_unary_op): Still do build_ptrmemfunc in
        templates.

From-SVN: r138756

16 years agoDaily bump.
GCC Administrator [Wed, 6 Aug 2008 00:16:51 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r138751

16 years ago* config/xtensa/t-xtensa: Remove dependency for gt-xtensa.h.
Bob Wilson [Tue, 5 Aug 2008 23:16:51 +0000 (23:16 +0000)]
* config/xtensa/t-xtensa: Remove dependency for gt-xtensa.h.

From-SVN: r138748

16 years agoxtensa.c (xtensa_va_start): Unshare valist.
Bob Wilson [Tue, 5 Aug 2008 23:06:33 +0000 (23:06 +0000)]
xtensa.c (xtensa_va_start): Unshare valist.

* config/xtensa/xtensa.c (xtensa_va_start): Unshare valist.
(xtensa_gimplify_va_arg_expr): Unshare valist, orig_ndx, ndx, array,
va_size, and type_size.

From-SVN: r138747

16 years agore PR c++/37016 (member function pointer failure with optimization)
Jason Merrill [Tue, 5 Aug 2008 22:22:00 +0000 (18:22 -0400)]
re PR c++/37016 (member function pointer failure with optimization)

        PR c++/37016
        * tree-ssa.c (useless_type_conversion_p_1): Call langhook
        if TYPE_STRUCTURAL_EQUALITY_P is true for both types.

From-SVN: r138740

16 years agoconfigure.ac (HAVE_GAS_CFI_DIRECTIVE): Check .cfi_personality.
Richard Henderson [Tue, 5 Aug 2008 21:25:48 +0000 (14:25 -0700)]
configure.ac (HAVE_GAS_CFI_DIRECTIVE): Check .cfi_personality.

        * configure.ac (HAVE_GAS_CFI_DIRECTIVE): Check .cfi_personality.
        * configure: Rebuild.

From-SVN: r138738

16 years agotestsuite_api.h (diamond_derivation<>::test): Expect ambiguity together with the...
Paolo Carlini [Tue, 5 Aug 2008 20:39:48 +0000 (20:39 +0000)]
testsuite_api.h (diamond_derivation<>::test): Expect ambiguity together with the standard exception classes.

2008-08-05  Paolo Carlini  <paolo.carlini@oracle.com>

* testsuite/util/testsuite_api.h (diamond_derivation<>::test):
Expect ambiguity together with the standard exception classes.
* testsuite/27_io/ios_base/failure/cons_virtual_derivation.cc: Remove
xfail.
* testsuite/19_diagnostics/logic_error/cons_virtual_derivation.cc:
Likewise.
* testsuite/19_diagnostics/runtime_error/cons_virtual_derivation.cc:
Likewise.
* testsuite/19_diagnostics/system_error/cons_virtual_derivation.cc:
Likewise.
* testsuite/18_support/bad_alloc/cons_virtual_derivation.cc: Likewise.
* testsuite/18_support/bad_cast/cons_virtual_derivation.cc: Likewise.
* testsuite/18_support/bad_exception/cons_virtual_derivation.cc:
Likewise.
* testsuite/18_support/bad_typeid/cons_virtual_derivation.cc:
Likewise.
* testsuite/ext/concurrence_lock_error/cons_virtual_derivation.cc:
Likewise.
* testsuite/ext/forced_exception_error/cons_virtual_derivation.cc:
Likewise.
* testsuite/ext/concurrence_unlock_error/cons_virtual_derivation.cc:
Likewise.
* testsuite/20_util/bad_function_call/cons_virtual_derivation.cc:
Likewise.

From-SVN: r138737

16 years agore PR tree-optimization/37024 (bad tail call from tuples merge)
Andrew Pinski [Tue, 5 Aug 2008 18:23:07 +0000 (18:23 +0000)]
re PR tree-optimization/37024 (bad tail call from tuples merge)

2008-08-05  Andrew Pinski  <andrew_pinski@playstation.sony.com>

        PR tree-opt/37024
        * tree-tailcall.c (process_assignment): Use gimple_assign_cast_p
        instead of IS_CONVERT_EXPR_CODE_P for seeing if the assignment
        is a conversion.

2008-08-05  Andrew Pinski  <andrew_pinski@playstation.sony.com>

        PR tree-opt/37024
        * gcc.dg/tree-ssa/tailcall-4.c: New testcase.

From-SVN: r138735

16 years agoMakefile.in (c-cppbuiltin.o): Depend on debug.h.
Richard Henderson [Tue, 5 Aug 2008 17:24:37 +0000 (10:24 -0700)]
Makefile.in (c-cppbuiltin.o): Depend on debug.h.

        * Makefile.in (c-cppbuiltin.o): Depend on debug.h.
        * c-cppbuiltin.c (c_cpp_builtins): Define __GCC_HAVE_DWARF2_CFI_ASM.
        * doc/cpp.texi (__GCC_HAVE_DWARF2_CFI_ASM): Document it.
        * common.opt (fdwarf2-cfi-asm): New.
        * configure.ac (HAVE_GAS_CFI_DIRECTIVE): New.
        * config.in, configure: Rebuild.
        * dwarf2asm.c (dw2_asm_output_data_raw): New.
        (dw2_asm_output_data_uleb128_raw, dw2_asm_output_data_sleb128_raw): New.
        (dw2_force_const_mem): Externalize.
        * dwarf2asm.h: Update.
        * dwarf2out.c (dwarf2out_cfi_label): If flag_dwarf2_cfi_asm, don't
        generate a real label.
        (output_cfi_directive): New.
        (add_fde_cfi): If flag_dwarf2_cfi_asm, use it.
        (output_call_frame_info): Do nothing if flag_dwarf2_cfi_asm.
        (dwarf2out_begin_prologue): Emit .cfi_startproc, .cfi_personality,
        and .cfi_lsda.
        (dwarf2out_end_epilogue): Emit .cfi_endproc.
        (output_loc_operands_raw, output_loc_sequence_raw): New.
        (output_cfa_loc_raw): New.

From-SVN: r138733

16 years agoHACKING: Update instructions how to build gcj/javaprims.h.
Matthias Klose [Tue, 5 Aug 2008 15:46:22 +0000 (15:46 +0000)]
HACKING: Update instructions how to build gcj/javaprims.h.

2008-08-05  Matthias Klose  <doko@ubuntu.com>

        * HACKING: Update instructions how to build gcj/javaprims.h.

From-SVN: r138732

16 years agosem_ch3.adb (Process_Subtype): An allocator is a valid construct that can carry a...
Arnaud Charlet [Tue, 5 Aug 2008 15:14:29 +0000 (17:14 +0200)]
sem_ch3.adb (Process_Subtype): An allocator is a valid construct that can carry a null exclusion indicator...

2008-08-05  Ed Schonberg  <schonberg@adacore.com>

* sem_ch3.adb (Process_Subtype): An allocator is a valid construct that
can carry a null exclusion indicator, and on which an error may be
posted if the indicator is redundant.

* sem_ch8.adb (Analyze_Object_Renaming): Verify that a null exclusion
does not apply to a subtype mark that already excludes null.

* sem_ch12.adb (Formal_Object_Declaration): Verify that a null
exclusion does not apply to a subtype mark that already excludes null.

From-SVN: r138728

16 years agovms_data.ads: Translation for /POINTER_SIZE qualifier.
Doug Rupp [Tue, 5 Aug 2008 15:13:25 +0000 (17:13 +0200)]
vms_data.ads: Translation for /POINTER_SIZE qualifier.

2008-08-05  Doug Rupp  <rupp@adacore.com>

* vms_data.ads: Translation for /POINTER_SIZE qualifier.

From-SVN: r138727

16 years agoMakefile.rtl: Compile s-oscons.ads as part of the runtime library.
Thomas Quinot [Tue, 5 Aug 2008 15:13:05 +0000 (17:13 +0200)]
Makefile.rtl: Compile s-oscons.ads as part of the runtime library.

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* Makefile.rtl: Compile s-oscons.ads as part of the runtime library.

From-SVN: r138726

16 years ago2008-08-05 Ed Schonberg <schonberg@adacore.com>
Ed Schonberg [Tue, 5 Aug 2008 15:11:19 +0000 (17:11 +0200)]
2008-08-05  Ed Schonberg  <schonberg@adacore.com>

* freeze.adb:
(Freeze_Entity): A deferred constant does not violate the restriction
No_Default_Initialization,

From-SVN: r138725

16 years agoRemove test, premature.
Arnaud Charlet [Tue, 5 Aug 2008 14:49:51 +0000 (16:49 +0200)]
Remove test, premature.

From-SVN: r138724

16 years agosem_ch5.adb (Analyze_Assignment_Statement): Apply conversion to right-hand side when...
Arnaud Charlet [Tue, 5 Aug 2008 14:37:44 +0000 (16:37 +0200)]
sem_ch5.adb (Analyze_Assignment_Statement): Apply conversion to right-hand side when it is an anonymous access_to_subprogram...

2008-08-05  Ed Schonberg  <schonberg@adacore.com>

* sem_ch5.adb (Analyze_Assignment_Statement): Apply conversion to
right-hand side when it is an anonymous access_to_subprogram, to force
static accessibility check when needed.

From-SVN: r138723

16 years agosem_attr.adb: (Analyze_Attribute...
Ed Schonberg [Tue, 5 Aug 2008 14:37:19 +0000 (16:37 +0200)]
sem_attr.adb: (Analyze_Attribute...

2008-08-05  Ed Schonberg  <schonberg@adacore.com>

* sem_attr.adb:
(Analyze_Attribute, case 'Result): handle properly the case where some
operand of the expression in a post-condition generates a transient
block.

From-SVN: r138722

16 years agoadaint.c: Remove support for readable attribute on vxworks and nucleus
Pascal Obry [Tue, 5 Aug 2008 14:37:10 +0000 (16:37 +0200)]
adaint.c: Remove support for readable attribute on vxworks and nucleus

2008-08-05  Pascal Obry  <obry@adacore.com>

* adaint.c: Remove support for readable attribute on vxworks and nucleus

From-SVN: r138721

16 years agogsocket.h: Make this file includable in a Nucleus environment, which does not support...
Thomas Quinot [Tue, 5 Aug 2008 14:37:01 +0000 (16:37 +0200)]
gsocket.h: Make this file includable in a Nucleus environment, which does not support sockets.

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* gsocket.h: Make this file includable in a Nucleus environment, which
does not support sockets.

* socket.c: Remove Nucleus-specific hack.

From-SVN: r138720

16 years ago* gnat.dg/post_block.adb: New test.
Arnaud Charlet [Tue, 5 Aug 2008 14:36:21 +0000 (14:36 +0000)]
* gnat.dg/post_block.adb: New test.

From-SVN: r138719

16 years agoinvoke.texi: Document new ARM -mfpu= and -mcpu= options.
Paul Brook [Tue, 5 Aug 2008 14:24:08 +0000 (14:24 +0000)]
invoke.texi: Document new ARM -mfpu= and -mcpu= options.

2008-08-05  Paul Brook  <paul@codesourcery.com>

gcc/
* doc/invoke.texi: Document new ARM -mfpu= and -mcpu= options.
* config/arm/arm.c (all_fpus): Add vfpv3 and vfpv3-d16.
(fp_model_for_fpu): Add entry for FPUTYPE_VFP3D16.
(arm_file_start): Add FPUTYPE_VFP3D16.  Rename vfp3 to vfpv3.
* config/arm/arm.h (TARGET_VFPD32): Define.
(TARGET_VFP3): Use TARGET_VFPD32.
(fputype): Add FPUTYPE_VFP3D16.
(LAST_VFP_REGNUM): Use TARGET_VFPD32.
* config/arm/constraints.md ("w"): Use TARGET_VFPD32.
* config/arm/arm-cores.def: Add cortex-r4f.
* config/arm/arm-tune.md: Regenerate.

From-SVN: r138718

16 years agoug_words: Add entries for -gnatw.b/-gnatw.B
Arnaud Charlet [Tue, 5 Aug 2008 14:10:54 +0000 (16:10 +0200)]
ug_words: Add entries for -gnatw.b/-gnatw.B

2008-08-05  Robert Dewar  <dewar@adacore.com>

* ug_words: Add entries for -gnatw.b/-gnatw.B

* vms_data.ads: Add entries for -gnatw.b/-gnatw.B

From-SVN: r138717

16 years agognat_ugn.texi: Changing the description of the gnatcheck metrics rule according to...
Sergey Rybin [Tue, 5 Aug 2008 14:10:38 +0000 (16:10 +0200)]
gnat_ugn.texi: Changing the description of the gnatcheck metrics rule according to the change...

2008-08-05  Sergey Rybin  <rybin@adacore.com>

* gnat_ugn.texi: Changing the description of the gnatcheck metrics
rule according to the change in the rule option.
Add documentation for -gnatw.b/-gnatw.B

From-SVN: r138716

16 years agospu_spu_mfcio.h: Wrap in extern "C" if __cplusplus.
Ulrich Weigand [Tue, 5 Aug 2008 14:04:56 +0000 (14:04 +0000)]
spu_spu_mfcio.h: Wrap in extern "C" if __cplusplus.

* config/spu_spu_mfcio.h: Wrap in extern "C" if __cplusplus.
Reword some comments throughout the file.

(MFC_MIN_DMA_LIST_ELEMENTS): New define.
(MFC_MAX_DMA_LIST_ELEMENTS): Likewise.
(MFC_MIN_DMA_LIST_SIZE): Redefine in terms of MFC_MIN_DMA_LIST_ELEMENTS.
(MFC_MAX_DMA_LIST_SIZE): Redefine in terms of MFC_MAX_DMA_LIST_ELEMENTS.

(MFC_START_ENABLE): Remove PPU-only define.
(MFC_PUTS_CMD, MFC_PUTFS_CMD, MFC_PUTBS_CMD): Likewise.
(MFC_GETS_CMD, MFC_GETFS_CMD, MFC_GETBS_CMD): Likewise.

(MFC_PUTB_CMD, MFC_PUTF_CMD): Reimplement using symbolic constants.
(MFC_PUTL_CMD, MFC_PUTLB_CMD, MFC_PUTLF_CMD): Likewise.
(MFC_PUTR_CMD, MFC_PUTRB_CMD, MFC_PUTRF_CMD): Likewise.
(MFC_PUTRL_CMD, MFC_PUTRLB_CMD, MFC_PUTRLF_CMD): Likewise.
(MFC_GETB_CMD, MFC_GETF_CMD): Likewise.
(MFC_GETL_CMD, MFC_GETLB_CMD, MFC_GETLF_CMD): Likewise.
(MFC_SNDSIGB_CMD, MFC_SNDSIGF_CMD): Likewise.

(MFC_SDCRT_CMD, MFC_SDCRTST_CMD): New defines.
(MFC_SDCRZ_CMD, MFC_SDCRST_CMD, MFC_SDCRF_CMD): Likewise.
(mfc_sdcrt, mfc_sdcrtst): Likewise.
(mfc_sdcrz, mfc_sdcrst, mfc_sdcrf): Likewise.

(spu_read_machine_status): Fix typo.

From-SVN: r138714

16 years agospu.h (CANNOT_CHANGE_MODE_CLASS): Allow (multi)word-sized SUBREG of multi-word hard...
Ulrich Weigand [Tue, 5 Aug 2008 13:53:20 +0000 (13:53 +0000)]
spu.h (CANNOT_CHANGE_MODE_CLASS): Allow (multi)word-sized SUBREG of multi-word hard register.

* config/spu/spu.h (CANNOT_CHANGE_MODE_CLASS): Allow (multi)word-sized
SUBREG of multi-word hard register.
* config/spu/spu.c (valid_subreg): Likewise.
(adjust_operand): Handle SUBREGs of multi-word hard registers.

From-SVN: r138711

16 years agoa-wtdeio.adb (Put (Current_Output)): Use Fore in the call to Put (File).
Vincent Celier [Tue, 5 Aug 2008 13:53:04 +0000 (13:53 +0000)]
a-wtdeio.adb (Put (Current_Output)): Use Fore in the call to Put (File).

2008-08-05  Vincent Celier  <celier@adacore.com>

* a-wtdeio.adb (Put (Current_Output)): Use Fore in the call to Put
(File).

* a-ztdeio.adb: Ditto.

From-SVN: r138710

16 years agoadaint.c, [...]: Add support for the readable attribute.
Arnaud Charlet [Tue, 5 Aug 2008 13:42:47 +0000 (15:42 +0200)]
adaint.c, [...]: Add support for the readable attribute.

2008-08-05  Pascal Obry  <obry@adacore.com>

* adaint.c, adaint.h, s-os_lib.adb, s-os_lib.ads: Add support for the
readable attribute.

From-SVN: r138709

16 years ago2008-08-05 Vincent Celier <celier@adacore.com>
Vincent Celier [Tue, 5 Aug 2008 13:42:08 +0000 (15:42 +0200)]
2008-08-05  Vincent Celier  <celier@adacore.com>

* s-wchwts.adb:
(Wide_String_To_String): Returns a String with the same 'First as its
parameter S.
(Wide_Wide_String_To_String): Ditto

* s-wchwts.ads:
(Wide_String_To_String): Document that the lowest index of the returned
String is equal to S'First.

From-SVN: r138708

16 years agoMakefile.in (gnatlib*): Now depend on oscons target.
Thomas Quinot [Tue, 5 Aug 2008 13:29:05 +0000 (13:29 +0000)]
Makefile.in (gnatlib*): Now depend on oscons target.

* Makefile.in (gnatlib*): Now depend on oscons target.
(oscons): New target.

From-SVN: r138706

16 years agoxoscons.adb, [...]: New files.
Thomas Quinot [Tue, 5 Aug 2008 13:26:24 +0000 (13:26 +0000)]
xoscons.adb, [...]: New files.

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* xoscons.adb, xutil.ads, xutil.adb, s-oscons-tmplt.c: New files.

* gcc-interface/Makefile.in, gcc-interface/Make-lang.in: Generate
s-oscons.ads

From-SVN: r138705

16 years agoopt.ads (Warn_On_Biased_Representation): New flag
Arnaud Charlet [Tue, 5 Aug 2008 13:24:27 +0000 (15:24 +0200)]
opt.ads (Warn_On_Biased_Representation): New flag

2008-08-05  Robert Dewar  <dewar@adacore.com>

* opt.ads (Warn_On_Biased_Representation): New flag

* sem_ch13.adb:
(Analyze_Attribute_Definition_Clause): Issue warning when biased
representation is required.
(Minimum_Size): Don't allow biasing if enum rep clause case

* sem_warn.adb:
(Set_Dot_Warning_Switch): Add handling of -gnatw.b/B switches
(Set_Warning_Switch): Include -gnatw.b in -gnatwa, -gnatw.B in gnatws

* usage.adb: Add lines for -gnatw.b/B switches

From-SVN: r138704

16 years agoadaint.c: In UNIX cases do not call __gnat_stat but stat directly.
Pascal Obry [Tue, 5 Aug 2008 13:24:02 +0000 (15:24 +0200)]
adaint.c: In UNIX cases do not call __gnat_stat but stat directly.

2008-08-05  Pascal Obry  <obry@adacore.com>

* adaint.c: In UNIX cases do not call __gnat_stat but stat directly.

From-SVN: r138703

16 years agosocket.c: Minor reformatting.
Thomas Quinot [Tue, 5 Aug 2008 13:23:53 +0000 (15:23 +0200)]
socket.c: Minor reformatting.

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* socket.c: Minor reformatting.

From-SVN: r138702

16 years agoa-coinve.adb: Reorder the code to avoid uninitialized warning.
Pascal Obry [Tue, 5 Aug 2008 13:23:44 +0000 (15:23 +0200)]
a-coinve.adb: Reorder the code to avoid uninitialized warning.

2008-08-05  Pascal Obry  <obry@adacore.com>

* a-coinve.adb: Reorder the code to avoid uninitialized warning.

From-SVN: r138701

16 years agore PR middle-end/37026 (GCC ICE in tree_to_gimple_tuple)
Richard Guenther [Tue, 5 Aug 2008 11:42:33 +0000 (11:42 +0000)]
re PR middle-end/37026 (GCC ICE in tree_to_gimple_tuple)

2008-08-05  Richard Guenther  <rguenther@suse.de>

PR middle-end/37026
* gcc.c-torture/compile/pr37026.c: New testcase.

From-SVN: r138700

16 years agosem_ch12.adb (Validate_Array_Type_Instance): Only apply complex visibility check...
Arnaud Charlet [Tue, 5 Aug 2008 09:40:19 +0000 (11:40 +0200)]
sem_ch12.adb (Validate_Array_Type_Instance): Only apply complex visibility check on the component type if the simple test fails.

2008-08-05  Ed Schonberg  <schonberg@adacore.com>

* sem_ch12.adb (Validate_Array_Type_Instance): Only apply complex
visibility check on the component type if the simple test fails.

From-SVN: r138697

16 years agoprj-nmsc.adb: Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:39:23 +0000 (11:39 +0200)]
prj-nmsc.adb: Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* prj-nmsc.adb: Minor reformatting

From-SVN: r138696

16 years agosem_ch3.adb: Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:39:05 +0000 (11:39 +0200)]
sem_ch3.adb: Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* sem_ch3.adb: Minor reformatting

From-SVN: r138695

16 years ago* gnat.dg/not_null.adb: New test.
Arnaud Charlet [Tue, 5 Aug 2008 09:33:21 +0000 (09:33 +0000)]
* gnat.dg/not_null.adb: New test.

From-SVN: r138694

16 years agosem_ch3.adb (Analyze_Object_Declaration): Avoid type Any_Access in unresolved initial...
Arnaud Charlet [Tue, 5 Aug 2008 09:30:53 +0000 (11:30 +0200)]
sem_ch3.adb (Analyze_Object_Declaration): Avoid type Any_Access in unresolved initial value of "null"...

2008-08-05  Bob Duff  <duff@adacore.com>

* sem_ch3.adb (Analyze_Object_Declaration): Avoid type Any_Access in
unresolved initial value of "null", because it causes implicitly
generated "=" operators to be ambiguous, and because this type should
not be passed to gigi.

From-SVN: r138693

16 years agoprj.ads: Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:29:28 +0000 (11:29 +0200)]
prj.ads: Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* prj.ads: Minor reformatting

From-SVN: r138692

16 years agos-fileio.adb: Minor code reorganization Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:29:20 +0000 (11:29 +0200)]
s-fileio.adb: Minor code reorganization Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* s-fileio.adb: Minor code reorganization
Minor reformatting

From-SVN: r138691

16 years agos-os_lib.adb: Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:29:12 +0000 (11:29 +0200)]
s-os_lib.adb: Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* s-os_lib.adb: Minor reformatting

From-SVN: r138690

16 years agoprj-attr.ads: Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:29:03 +0000 (11:29 +0200)]
prj-attr.ads: Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* prj-attr.ads: Minor reformatting

From-SVN: r138689

16 years agomlib.adb: Minor code reorganization Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:28:55 +0000 (11:28 +0200)]
mlib.adb: Minor code reorganization Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* mlib.adb: Minor code reorganization
Minor reformatting

From-SVN: r138688

16 years agog-sercom.adb, [...] (Data_Bits): Change literals B7 and B8 to CS7 and CS8.
Thomas Quinot [Tue, 5 Aug 2008 09:28:44 +0000 (11:28 +0200)]
g-sercom.adb, [...] (Data_Bits): Change literals B7 and B8 to CS7 and CS8.

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* g-sercom.adb, g-sercom.ads, g-sercom-mingw.adb,
g-sercom-linux.adb (Data_Bits): Change literals B7 and B8 to CS7 and
CS8.

From-SVN: r138687

16 years agomake.adb: Minor reformatting
Robert Dewar [Tue, 5 Aug 2008 09:28:12 +0000 (11:28 +0200)]
make.adb: Minor reformatting

2008-08-05  Robert Dewar  <dewar@adacore.com>

* make.adb: Minor reformatting

From-SVN: r138686

16 years agochecks.ads: Minor comment fix
Bob Duff [Tue, 5 Aug 2008 09:28:03 +0000 (11:28 +0200)]
checks.ads: Minor comment fix

2008-08-05  Bob Duff  <duff@adacore.com>

* checks.ads: Minor comment fix

From-SVN: r138685

16 years agoinit.c (__gnat_install_handler for linux): If we are building the Xenomai run time...
Jose Ruiz [Tue, 5 Aug 2008 09:27:49 +0000 (11:27 +0200)]
init.c (__gnat_install_handler for linux): If we are building the Xenomai run time then we need to do two additional things...

2008-08-05  Jose Ruiz  <ruiz@adacore.com>

* init.c (__gnat_install_handler for linux): If we are building the
Xenomai run time then we need to do two additional things: avoid
memory swapping and transform the Linux environment task into a native
Xenomai task.

* gcc-interface/Makefile.in (LIBGNAT_TARGET_PAIRS for xenomai run
time): Use interface to Xenomai native skin and avoid linux-specific
way of setting CPU affinity.
(EH_MECHANISM for the xenomai run time): Use sjlj exception mechanism.

From-SVN: r138684

16 years agomake.adb (Switches_Of): Check for Switches (others), before checking for Default_Swit...
Arnaud Charlet [Tue, 5 Aug 2008 09:14:48 +0000 (11:14 +0200)]
make.adb (Switches_Of): Check for Switches (others), before checking for Default_Switches ("Ada").

2008-08-05  Vincent Celier  <celier@adacore.com>

* make.adb (Switches_Of): Check for Switches (others), before checking
for Default_Switches ("Ada").
(Gnatmake): Use Builder'Switches (others) in preference to
Builder'Default_Switches ("Ada") if there are several mains.

* prj-attr-pm.adb:
(Add_Attribute): Add component Others_Allowed in Attribute_Record
aggregate.

* prj-attr.adb:
Add markers to indicates that attributes Switches allow others as index
(Others_Allowed_For): New Boolean function, returning True for
attributes with the mark.
(Initialize): Recognize optional letter 'O' as the marker for
associative array attributes where others is allowed as the index.

* prj-attr.ads:
(Others_Allowed_For): New Boolean function
(Attribute_Record): New Boolean component Others_Allowed

* prj-dect.adb:
(Parse_Attribute_Declaration): For associative array attribute where
others is allowed as the index, allow others as an index.

* prj-nmsc.adb:
(Process_Binder): Skip associative array attributes with index others
(Process_Compiler): Ditto

* prj-util.adb:
(Value_Of (Index, In_Array)): Make no attempt to put in lower case when
index is All_Other_Names.

* prj.ads:
(All_Other_Names): New constant

From-SVN: r138683

16 years agogen-oscons.c: Adapt for VMS where termios.h is not available.
Vasiliy Fofanov [Tue, 5 Aug 2008 09:14:13 +0000 (11:14 +0200)]
gen-oscons.c: Adapt for VMS where termios.h is not available.

2008-08-05  Vasiliy Fofanov  <fofanov@adacore.com>

* gen-oscons.c: Adapt for VMS where termios.h is not available.

From-SVN: r138682

16 years ago2008-08-05 Vincent Celier <celier@adacore.com>
Vincent Celier [Tue, 5 Aug 2008 09:14:03 +0000 (11:14 +0200)]
2008-08-05  Vincent Celier  <celier@adacore.com>

* prj-proc.adb:
(Process_Declarative_Items): Skip associative array attribute when index
is reserved word "others".

From-SVN: r138681

16 years agomlib.adb: Update comments.
Vincent Celier [Tue, 5 Aug 2008 09:13:53 +0000 (11:13 +0200)]
mlib.adb: Update comments.

2008-08-05  Vincent Celier  <celier@adacore.com>

* mlib.adb: Update comments.

From-SVN: r138680

16 years agogen-soccon.c: Rename to gen-oscons.c
Thomas Quinot [Tue, 5 Aug 2008 09:05:08 +0000 (09:05 +0000)]
gen-soccon.c: Rename to gen-oscons.c

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* gen-soccon.c: Rename to gen-oscons.c

* gen-oscons.c: New file. Now generate System.OS_Constants instead of
GNAT.Sockets.Constants.
Add new constants for GNAT.Serial_Communications and System.File_IO.

From-SVN: r138679

16 years agoa-rttiev.adb: Minor reformatting (comments)
Thomas Quinot [Tue, 5 Aug 2008 09:04:18 +0000 (11:04 +0200)]
a-rttiev.adb: Minor reformatting (comments)

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* a-rttiev.adb: Minor reformatting (comments)

From-SVN: r138678

16 years agoadaint.c, [...]: Fix the Set_Read_Only Win32 implementation.
Arnaud Charlet [Tue, 5 Aug 2008 08:41:30 +0000 (10:41 +0200)]
adaint.c, [...]: Fix the Set_Read_Only Win32 implementation.

2008-08-05  Pascal Obry  <obry@adacore.com>

* adaint.c, adaint.h, s-os_lib.adb, s-os_lib.ads: Fix the
Set_Read_Only Win32 implementation.

From-SVN: r138676

16 years agosem_util.adb (Collect_Interfaces_Info): Minor reformating.
Javier Miranda [Tue, 5 Aug 2008 08:41:07 +0000 (10:41 +0200)]
sem_util.adb (Collect_Interfaces_Info): Minor reformating.

2008-08-05  Javier Miranda  <miranda@adacore.com>

* sem_util.adb (Collect_Interfaces_Info): Minor reformating.
* exp_ch3.adb (Build_Offset_To_Top_Functions): Code cleanup: the
implementation of this routine has been simplified.

From-SVN: r138675

16 years agoexp_strm.adb: Minor reformatting (comments)
Thomas Quinot [Tue, 5 Aug 2008 08:34:58 +0000 (08:34 +0000)]
exp_strm.adb: Minor reformatting (comments)

2008-08-05  Thomas Quinot  <quinot@adacore.com>

* exp_strm.adb: Minor reformatting (comments)

* sem_ch12.adb: Minor reformatting.

From-SVN: r138674

16 years ago* gnat.dg/access_discr2.adb: New test.
Arnaud Charlet [Tue, 5 Aug 2008 08:21:05 +0000 (08:21 +0000)]
* gnat.dg/access_discr2.adb: New test.

From-SVN: r138673