re PR preprocessor/31869 (stringifying empty macros)

libcpp:
PR preprocessor/31869
* macro.c (stringify_arg): Handle NULL source token in padding
token where previous padding token did not have source token with
preceding whitespace.

gcc/testsuite:
* gcc.dg/cpp/strify5.c: New test.

From-SVN: r145989

fold-const.c (build_range_check): Properly deal with enumeral and boolean base types.

* fold-const.c (build_range_check): Properly deal with enumeral and
boolean base types.

From-SVN: r145988

invoke.texi (max_gcse_passes): Remove documentation.

* doc/invoke.texi (max_gcse_passes): Remove documentation.
* params.def (PARAM_MAX_GCSE_PASSES): Remove.
* params.h (MAX_GCSE_PASSES): Remove.
* gcse.c (gcse_main): Run CPROP1, PRE or HOIST, and CPROP2
in sequence.  Remove ability to run multiple passes.
(bypass_jumps): Report run as third CPROP pass.

From-SVN: r145987

re PR middle-end/39651 (New cleanup test failures)

PR middle-end/39651
* except.c (can_throw_external): Look at each insn in a SEQUENCE
when deciding whether the whole SEQUENCE can throw.

From-SVN: r145986

re PR target/39740 (unrecognizable insn on alpha using -O3 and -std=c99)

2009-04-12  Uros Bizjak  <ubizjak@gmail.com>

PR target/39740
* config/alpha/predicates.md (local_symbolic_operand): Return 1 for
offseted label references.

testsuite/ChangeLog:

PR target/39740
* gcc.target/alpha/pr39740.c: New test.

From-SVN: r145985

Daily bump.

From-SVN: r145972

* tree-ssa-pre.c (eliminate): Fix call of update_stmt.

From-SVN: r145961

re PR fortran/37746 (bounds check of string dummy arguments)

2009-04-11  Daniel Kraft  <d@domob.eu>

PR fortran/37746
* gfortran.h (struct gfc_charlen): New field `passed_length' to store
the actual passed string length for dummy arguments.
* trans-decl.c (gfc_create_string_length): Formatting fixes and added
assertion, moved a local variable into the innermost block it is needed.
(create_function_arglist): Removed TODO about the check being
implemented and initialize cl->passed_length here.
(add_argument_checking): New method.
(gfc_generate_function_code): Call the argument checking method.

2009-04-11  Daniel Kraft  <d@domob.eu>

PR fortran/37746
* gfortran.dg/bounds_check_strlen_1.f90: New test.
* gfortran.dg/bounds_check_strlen_2.f90: New test.
* gfortran.dg/bounds_check_strlen_3.f90: New test.
* gfortran.dg/bounds_check_strlen_4.f90: New test.
* gfortran.dg/bounds_check_strlen_5.f90: New test.
* gfortran.dg/bounds_check_strlen_6.f90: New test.
* gfortran.dg/bounds_check_strlen_7.f90: New test.
* gfortran.fortran-torture/execute/intrinsic_index.f90: Fix wrong
expected string length that failed with -fbounds-check now.
* gfortran.fortran-torture/execute/intrinsic_trim.f90: Ditto.

From-SVN: r145958

re PR fortran/39692 (f95: conflict between EXTERNAL and POINTER)

2009-04-11  Janus Weil  <janus@gcc.gnu.org>

PR fortran/39692
* symbol.c (check_conflict): Reject procedure pointers for -std=f95.

2009-04-11  Janus Weil  <janus@gcc.gnu.org>

PR fortran/39692
* gfortran.dg/proc_ptr_14.f90: New.

From-SVN: r145955

re PR middle-end/39732 (-fprofile-generate -O1: ICE: verify_stmts failed, ADDRESSABLE bit not set on pointers passed to std::copy)

2009-04-11  Richard Guenther  <rguenther@suse.de>

PR middle-end/39732
* tree-inline.c (declare_return_variable): Mark DECL_BY_REFERENCE
return variables as TREE_ADDRESSABLE.

* g++.dg/torture/pr39732.C: New testcase.

From-SVN: r145954

resolve.c (resolve_global_procedure): Enable whole-file checking for procedures that are declared later in the file.

gcc/fortran/:
2009-04-11  Daniel Franke  <frake.daniel@gmail.com>

        * resolve.c (resolve_global_procedure): Enable whole-file checking for
        procedures that are declared later in the file.

gcc/testsuite:
2009-04-11  Daniel Franke  <franke.daniel@gmail.com>

        * gfortran.dg/whole_file_5.f90: New.
        * gfortran.dg/whole_file_6.f90: New.

From-SVN: r145953

re PR tree-optimization/39713 (ICE in get_expr_value_id)

2009-04-11  Richard Guenther  <rguenther@suse.de>

PR tree-optimization/39713
* tree-ssa-sccvn.c (vn_get_expr_for): Make sure built
reference trees have SSA_NAME operands.

* g++.dg/torture/pr39713.C: New testcase.

From-SVN: r145951

re PR c/39712 (type mismatch in address expression)

2009-04-11  Richard Guenther  <rguenther@suse.de>

PR c/39712
* c-gimplify.c (c_gimplify_expr): Adjust check for mismatched
address expressions.

* gcc.dg/pr39712.c: New testcase.

From-SVN: r145950

cygwin-stdint.h (INT_LEAST32_TYPE): Update to match changes in Cygwin 1.7 (UINT_LEAST32_TYPE...

* config/i386/cygwin-stdint.h (INT_LEAST32_TYPE):  Update to
match changes in Cygwin 1.7
(UINT_LEAST32_TYPE, INT_FAST16_TYPE, INT_FAST32_TYPE,
UINT_FAST16_TYPE, UINT_FAST32_TYPE):  Likewise.

From-SVN: r145949

re PR middle-end/39701 (Revision 145846 caused many test failures)

2009-04-10  Paolo Bonzini  <bonzini@gnu.org>

PR tree-optimization/39701
* doc/invoke.texi (Optimization Options): Document change in
meaning and initialization of -fdelete-null-pointer-checks.

From-SVN: r145948

Daily bump.

From-SVN: r145947

re PR middle-end/39701 (Revision 145846 caused many test failures)

2009-04-10  H.J. Lu  <hongjiu.lu@intel.com>

PR middle-end/39701
* common.opt (-fdelete-null-pointer-checks): Initialize to 1.

* opts.c (decode_options): Don't set flag_delete_null_pointer_checks
here.

* doc/invoke.texi: Update -fdelete-null-pointer-checks.

From-SVN: r145937

re PR c++/28301 (ICE with broken specialization)

gcc/cp/

2009-04-10  Jason Merrill  <jason@redhat.com>

PR c++/28301
* parser.c (cp_parser_skip_to_end_of_block_or_statement): Return
if we see a close brace without an open brace.

gcc/testsuite/

2009-04-10  H.J. Lu  <hongjiu.lu@intel.com>

PR c++/28301
* g++.dg/cpp0x/enum2.C: Updated.
* g++.dg/debug/pr22514.C: Likewise.
* g++.dg/parse/enum2.C: Likewise.
* g++.dg/parse/enum3.C: Likewise.
* g++.dg/template/crash79.C: Likewise.
* g++.old-deja/g++.jason/cond.C: Likewise.

* g++.dg/template/pr28301.C: New.

From-SVN: r145936

interrupt_handler.c: Change from compile to assemble.

2009-04-10  Chao-ying Fu  <fu@mips.com>

* gcc.target/mips/interrupt_handler.c: Change from compile to
assemble.

From-SVN: r145935

tm.texi (Instruction Output): Document TARGET_ASM_FINAL_POSTSCAN_INSN.

2009-04-10  Chao-ying Fu  <fu@mips.com>

* doc/tm.texi (Instruction Output): Document
TARGET_ASM_FINAL_POSTSCAN_INSN.
* target.h (final_postscan_insn): New field in asm_out.
* target-def.h (TARGET_ASM_FINAL_POSTSCAN_INSN): New define.
(TARGET_ASM_OUT): Add TARGET_ASM_FINAL_POSTSCAN_INSN.
* final.c (final_scan_insn): Call
targetm.asm_out.final_postscan_insn after outputting
an asm macro and a normal instruction.

* config/mips/mips.h (FINAL_PRESCAN_INSN): New define.
* config/mips/mips-protos.h (mips_final_prescan_insn): Declare.
* config/mips/mips.c (mips_at_reg_p): New for_each_rtx callback.
(mips_final_prescan_insn, mips_final_postscan_insn): New functions.
(TARGET_ASM_FINAL_POSTSCAN_INSN): New define.

From-SVN: r145934

forward-1.x: Fix x86_64 entry typos.

* objc/execute/forward-1.x: Fix x86_64 entry typos.

From-SVN: r145932

re PR middle-end/39701 (Revision 145846 caused many test failures)

2009-04-10  Paolo Bonzini  <bonzini@gnu.org>

PR middle-end/39701
* trans.c (gfc_allocate_with_status): Fix type mismatches
on "pstat == 0".

From-SVN: r145930

re PR middle-end/39701 (Revision 145846 caused many test failures)

2009-04-10  Paolo Bonzini  <bonzini@gnu.org>

PR middle-end/39701
        * fold-const.c (tree_single_nonzero_warnv_p): Pass non-static
        variables as non-NULL even with -fdelete-null-pointer-checks.

From-SVN: r145927

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_warn.ads, sem_warn.adb (Check_Low_Bound_Tested): Catch more cases
for warning suppression.

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_ch8.adb (Use_One_Type): If the two use_type clauses are
identical, there is no redudancy to check.

2009-04-10  Gary Dismukes  <dismukes@adacore.com>

* exp_ch5.adb (Expand_N_Extended_Return_Statement): Delete redundant
calls initializing SS_Allocator (which is initialized in following
code).
(Expand_Simple_Function_Return): Add comment about False value for
Comes_From_Source on secondary-stack allocator.

* exp_ch9.adb (Build_Entry_Family_Name): Add comment.
(Build_Entry_Name): Add comment.

From-SVN: r145926

Revert unwanted change in previous commit.

From-SVN: r145925

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* einfo.ads, einfo.adb (Low_Bound_Tested): New name for Low_Bound_Known

* sem_prag.adb (Analyze_Pragma, case Check): Remove check for lower
bound tested, since this is now done more generally in Sem_Res.

* sem_res.adb (Resolve_Comparison_Op): Add call to
Check_Lower_Bound_Tested.
(Resolve_Equality_Op): Add call to Check_Lower_Bound_Tested

* sem_warn.ads, sem_warn.adb (Check_Low_Bound_Tested): New procedure
(Low_Bound_Tested): New name for Low_Bound_Known flag

* exp_ch5.adb: Minor reformatting

* exp_ch4.adb:
Add comments on copying the Comes_From_Source flag for allocators

* sinfo.ads:
Add comments on copying the Comes_From_Source flag for allocators

* exp_ch6.adb (Make_Build_In_Place_Call_In_Allocator): Copy
Comes_From_Source flag from old allocator to new one.

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_ch6.ads: Address missing documentation query

2009-04-10  Vincent Celier  <celier@adacore.com>

* prj-attr.adb:
Add new Linker attributes Max_Command_Line_Length, Response_File_Format
and Response_File_Switches.

* prj-nmsc.adb (Process_Linker): Process new attributes
Max_Command_Line_Length, Response_File_Format and
Response_File_Switches.

* prj.ads (Response_File_Format): New enumeration type
(Project_Configuration): New componants Max_Command_Line_Length,
Resp_File_Format and Resp_File_Options.

* snames.ads-tmpl: Add new standard names for linking response files
for gprbuild: GNU, None, Object_List, Option_List,
Max_Command_Line_Length, Response_File_Format and
Response_File_Switches.

2009-04-10  Geert Bosch  <bosch@adacore.com>

* system-aix.ads, system-darwin-ppc.ads, system-darwin-x86.ads,
system-freebsd-x86.ads, system-hpux.ads, system-hpux-ia64.ads,
system-irix-n32.ads, system-irix-o32.ads, system-linux-alpha.ads,
system-linux-hppa.ads, system-linux-ia64.ads, system-linux-ppc.ads,
system-linux-s390.ads, system-linux-s390x.ads, system-linux-sh4.ads,
system-linux-sparc.ads, system-linux-x86_64.ads, system-linux-x86.ads,
system-mingw.ads, system-solaris-sparc.ads, system-solaris-sparcv9.ads,
system-solaris-x86.ads, system-tru64.ads, system-vms_64.ads,
system-vms.ads, system-vms-ia64.ads, system-vms-zcx.ads,
system-vxworks-arm.ads, system-vxworks-m68k.ads,
system-vxworks-mips.ads, system-vxworks-ppc.ads,
system-vxworks-sparcv9.ads, system-vxworks-x86.ads
(Backend_Overflow_Checks): Set to True.

From-SVN: r145924

[multiple changes]

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* exp_attr.adb: Minor reformatting

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_aux.adb: Minor reformatting

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_prag.adb (Check_Precondition_Postcondition): Within a generic,
analyze the expression for a postcondition, even if the compiler mode
is Generate_Code.

From-SVN: r145923

sem_aux.adb: Minor reformatting

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_aux.adb: Minor reformatting

From-SVN: r145922

Complete previous change.

From-SVN: r145921

init.c: Change VxWorks 6 stack overflow checking for kernel apps.

2009-04-10  Ed Falis  <falis@adacore.com>

* init.c: Change VxWorks 6 stack overflow checking for kernel apps.

* system-vxworks-ppc.ads, system-vxworks-x86.ads: Update header.

From-SVN: r145920

sem_ch6.ads (Check_Subtype_Conformant): Add ??? comment for undocumented formal.

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* sem_ch6.ads (Check_Subtype_Conformant): Add ??? comment for
undocumented formal.
Minor reformatting

* a-direio.ads: Fix typo in comment

* sem_ch3.adb, g-dirope.adb, sem_type.adb, sem_ch12.adb, sem_case.adb,
errout.adb, sem_ch4.adb, sem_ch11.adb, exp_dist.adb, sem_ch13.adb:
Use uniform phrasing for comment at start of subprogram body.

* xsnamest.adb: Add note to explain why we use specific names for the
newly generated files instead of generating snames.{ads,adb,h} directly

From-SVN: r145918

[multiple changes]

2009-04-10  Sergey Rybin  <rybin@adacore.com>

* vms_data.ads:
Add qualifier for new gnatstub option '--no-exception'

* gnat_ugn.texi:
Add the description of the new gnatstub option '--no-exception'

2009-04-10  Robert Dewar  <dewar@adacore.com>

* rtsfind.adb: Minor reformatting

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* sem_disp.adb: Minor reformatting.
Add comment pointing to RM clause for the case of warning against a
(failed) attempt at declaring a primitive operation elsewhere than in a
package spec.

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_ch12.adb (Denotes_Formal_Package): Check whether the package is
an actual for a previous formal package of the current instance.

From-SVN: r145917

[multiple changes]

2009-04-10  Bob Duff  <duff@adacore.com>

* rtsfind.adb (RTE): Put implicit with_clauses on whatever unit needs
them first, rather than on the extended main unit.

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_ch6.adb (Check_Discriminant_Conformance): If discriminant
specification of full view carries a null exclusion indicator, create
an itype for it, to check for conformance with partial view.

From-SVN: r145916

[multiple changes]

2009-04-10  Bob Duff  <duff@adacore.com>

* rtsfind.ads: Minor code change: make RE_Unit_Table constant.

* rtsfind.adb: Minor comment changes, and remove useless code.

* sinfo.ads: Add ??? comment.

2009-04-10  Vincent Celier  <celier@adacore.com>

* vms_data.ads: Add missing GNAT SYNC VMS qualifiers -main= and -U

From-SVN: r145914

[multiple changes]

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* exp_attr.adb (Expand_N_Attribute_Reference, case 'Tag): If the tagged
type is a synchronized type, retrieve tag information from the
corresponding record, which has the dispatch table link.

2009-04-10  Jerome Lambourg  <lambourg@adacore.com>

* g-comlin.adb (Group_Analysis): Take care of switches that might be
decomposed afterwards, but are present as-is in the command line
configuration, and thus should be kept as-is.

From-SVN: r145913

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* gnat_rm.texi: Document that postconditions are tested on implicit
returns.

* sem_aux.adb: Minor reformatting

2009-04-10  Gary Dismukes  <dismukes@adacore.com>

* itypes.adb (Create_Null_Excluding_Itype): Apply Base_Type when
setting Etype.

* par-ch3.adb (P_Access_Type_Definition): Set new attribute
Null_Exclusion_In_Return_Present when an access-to-function type has a
result type with an explicit not null.

* sem_ch3.adb (Access_Subprogram_Definition): If a null exclusion is
given on the result type, then create a null-excluding itype for the
function.

* sem_ch6.adb (Analyze_Return_Type): Create a null-excluding itype in
the case where a null exclusion is imposed on a named access type.
(Analyze_Subprogram_Specification): Push and pop the scope of the
function around the call to Analyze_Return_Type in the case of no
formals, for consistency with handling when formals are present
(Process_Formals does this). Ensures that any itype created for the
return type will be associated with the proper scope.

* sem_ch12.adb (Analyze_Generic_Subprogram_Declaration): If a null
exclusion is given on a generic function's result type, then create a
null-excluding itype for the generic function.
(Instantiate_Object): Set Null_Exclusion_Present of a constant created
for an actual for a formal in object according to the setting on the
formal. Ensures null exclusion checks are done when the association is
elaborated.

* sinfo.ads: Add new flag Null_Exclusion_In_Return_Present on
N_Access_Function_Definition.

* sinfo.adb: Add Get_ and Set_ operations for
Null_Exclusion_In_Return_Present.

From-SVN: r145912

parser.c (cp_parser_class_specifier): Remove the unused has_trailing_semicolon.

2008-04-10  H.J. Lu  <hongjiu.lu@intel.com>

* parser.c (cp_parser_class_specifier): Remove the unused
has_trailing_semicolon.

From-SVN: r145911

re PR fortran/38709 (ICE on zero-sized array in initialization expression)

gcc/fortran/:
2009-04-10  Daniel Franke  <franke.daniel@gmail.com>

        PR fortran/38709
        * expr.c (find_array_section): Leave early on zero-sized arrays.

gcc/testsuite/:
2009-04-10  Daniel Franke  <franke.daniel@gmail.com>

        PR fortran/38709
        * gfortran.dg/zero_sized_6.f90: New.

From-SVN: r145909

Fix runtime license text.

From-SVN: r145908

exp_ch5.adb, [...]: Move the code that creates a call to the _Postconditions procedure in the case...

2009-04-10  Bob Duff  <duff@adacore.com>

* exp_ch5.adb, exp_ch6.adb, sem_ch6.adb: Move the code that creates a
call to the _Postconditions procedure in the case of implicit returns
from analysis to expansion. This eliminates some duplicated code. Use
the Postcondition_Proc to find the identity of this procedure during
expansion.

From-SVN: r145906

sem_ch6.adb: Minor code clean up.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_ch6.adb: Minor code clean up.

* einfo.ads, sem_attr.adb: Minor comment fixes.

* sem_ch8.adb: Minor reformatting

From-SVN: r145905

einfo.ads, einfo.adb (Postcondition_Proc): New attribute for procedures.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* einfo.ads, einfo.adb (Postcondition_Proc): New attribute for
procedures.

* sem_ch6.adb: Minor code clean up.

From-SVN: r145903

Update to GPL v3+

From-SVN: r145902

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* mlib-tgt-specific-xi.adb: Minor reformatting

2009-04-10  Bob Duff  <duff@adacore.com>

* einfo.ads: Minor comment fixes

2009-04-10  Vincent Celier  <celier@adacore.com>

* snames.ads-tmpl: Remove names that are no longer used in the
Project Manager.
Mark specifically those that are used only in gprbuild

2009-04-10  Eric Botcazou  <ebotcazou@adacore.com>

* init.c: Adjust EH support code on Alpha/Tru64.

2009-04-10  Bob Duff  <duff@adacore.com>

* sem_ch6.adb (Process_PPCs): Add a call to the _Postconditions
procedure on every path that could return implicitly (not via a return
statement) from a procedure.

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* exp_ch9.adb (Build_Master_Entity): An extended return statement is a
valid scope for a task declarations and therefore for a master id.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_aux.adb: Minor reformatting

2009-04-10  Vincent Celier  <celier@adacore.com>

* scn.adb (Obsolescent_Check_Flag): New Boolean flag, initialized to
True.
(Obsolescent_Check): Do nothing if Obsolescent_Check_Flag is False
(Set_Obsolescent_Check): New procedure to change the value of
Obsolescent_Check_Flag.

* scn.ads (Set_Obsolescent_Check): New procedure to control
Obsolescent_Check.

* sinput-l.adb (Load_File): Do not check for pragma Restrictions on
obsolescent features while preprocessing.

From-SVN: r145901

Make-lang.in: Stabilize new stamp-snames rule.

* Make-lang.in: Stabilize new stamp-snames rule.
Deal with relative source directories.

From-SVN: r145900

Minor clean up.

From-SVN: r145899

[multiple changes]

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* xsnamest.adb: Use XUtil to have uniform line endings (UNIX style) in
generated files on all platforms.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_aux.adb: Minor reformatting

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_ch3.adb (Access_Definition): Handle properly the case of a
protected function with formals that returns an anonymous access type.

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* sem_disp.adb: Minor reformatting

2009-04-10  Vasiliy Fofanov  <fofanov@adacore.com>

* seh_init.c: Do not use the 32-bit specific implementation of
__gnat_install_SEH_handler on 64-bit Windows target (64-bit specific
version TBD).

2009-04-10  Jose Ruiz  <ruiz@adacore.com>

* mlib-tgt-specific-xi.adb (Get_Target_Prefix): Target_Name may contain
a '/' at the end so we better use the complete target name to determine
whether it is a PowerPC 55xx target.

From-SVN: r145898

Make-lang.in: Add dependency on xutil to build xsnamest Add dependency of sem_aux.o on snames.ads.

        * Make-lang.in: Add dependency on xutil to build xsnamest
        Add dependency of sem_aux.o on snames.ads.

From-SVN: r145897

darwin-vecsave.asm: Remove extra "*/".

2009-04-10  H.J. Lu  <hongjiu.lu@intel.com>

* config/rs6000/darwin-vecsave.asm: Remove extra "*/".

From-SVN: r145896

Update comments.

From-SVN: r145895

sem_eval.adb: Minor reformatting

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* sem_eval.adb: Minor reformatting

From-SVN: r145894

Makefile.in: (stamp-tools): Add handling of snames.ad[sb]

2009-04-10  Thomas Quinot  <quinot@adacore.com>

gnattools/
* Makefile.in: (stamp-tools): Add handling of snames.ad[sb]

ada/
* snames.h, snames.ads, snames.adb: Remove files, now generated from
templates.

* snames.h-tmpl, snames.ads-tmpl, snames.adb-tmpl: Templates for the
above.

* xsnamest.adb: New file.

* gcc-interface/Make-lang.in: New target for automated generation of
snames.ads, snames.adb and snames.h

From-SVN: r145893

2009-04-10  Manuel López-Ibáñez  <manu@gcc.gnu.org>

PR  c++/20118
cp/
* parser.c (cp_parser_check_template_parameters): Take a
cp_declarator parameter.
(cp_parser_elaborated_type_specifier): Update to
cp_parser_check_template_parameters.
(cp_parser_class_head): Likewise.
(cp_parser_check_declarator_template_parameters): Likewise.
(cp_parser_check_template_parameters): Handle first the non-error
conditions. Give more accurate diagnostics if a declarator is
given.
testsuite/
* g++.dg/parse/pr20118.C: New.
* g++.dg/template/spec16.C: Update.

From-SVN: r145892

Makefile.in, utils.c: Include "rtl.h" to avoid compile time warnings.

2009-04-10  Tristan Gingold  <gingold@adacore.com>

* gcc-interface/Makefile.in, gcc-interface/utils.c: Include "rtl.h" to
avoid compile time warnings.
Do not add gcc/config in include search list while compiling the RTS.
Pragma Thread_Local_Storage is available on any target.

From-SVN: r145891

sem.ads, [...]: Minor comment fixes.

2009-04-10  Bob Duff  <duff@adacore.com>

* sem.ads, par.adb, sem_ch6.adb, sem_ch8.adb: Minor comment fixes.

From-SVN: r145890

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_prag.adb: Minor reformatting

* exp_util.adb (Make_Non_Empty_Check): New function
(Silly_Boolean_Array_Not_Test): Add call to Make_Non_Empty_Check
(Silly_Boolean_Array_Xor_Test): Use Make_Non_Empty_Check

2009-04-10  Arnaud Charlet  <charlet@adacore.com>

* make.adb, gnatlink.adb: Rename JGNAT toolchain.

2009-04-10  Jose Ruiz  <ruiz@adacore.com>

* mlib-tgt-specific-xi.adb (Get_Target_Prefix): Insert the appropriate
tool prefix for AVR and PowerPC 55xx targets.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_warn.adb (Within_Postcondition): New function
(Check_Unset_Reference): Use Within_Postcondition to stop bad warning

From-SVN: r145889

init.c: Install signal handler on Darwin.

2009-04-10  Tristan Gingold  <gingold@adacore.com>

* init.c: Install signal handler on Darwin.

From-SVN: r145888

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_warn.adb: Minor reformatting

* make.adb: Minor reformatting.

2009-04-10  Gary Dismukes  <dismukes@adacore.com>

* exp_ch7.adb (Find_Final_List): When creating a finalization-chain
entity and the scope is a subprogram, retrieve the Sloc of the
subprogram's body rather than using the sloc of the spec, for better
line-stepping behavior in gdb.
(Wrap_Transient_Declaration): For the Sloc of nodes created with a list
controller, use the Sloc of the first declaration of the containing list
rather than that of the node that triggered creation of the list
controller.

From-SVN: r145887

[multiple changes]

2009-04-10  Vincent Celier  <celier@adacore.com>

* prj-nmsc.adb (Check_Naming_Schemes): Initialize local variable Casing
to avoid gcc warning.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* g-comlin.adb: Add ??? comment

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_warn.adb (Check_Unused_Withs): Do not emit message about
unreferenced entities for a package with no visible declarations.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* exp_ch9.adb: Minor reformatting

From-SVN: r145886

[multiple changes]

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* sem_prag.adb: Minor reformatting

2009-04-10  Vincent Celier  <celier@adacore.com>

* prj-nmsc.adb:
(Check_Library_Attributes): For a project qualified as a library project
that is not a library project, indicate in the error message which
attributes are missing (Library_Dir and/or Library_Name).

From-SVN: r145885

Minor reformatting.

From-SVN: r145884

exp_ch5.adb, [...]: Avoid use of No_Position in Sloc of generated nodes...

2009-04-10  Bob Duff  <duff@adacore.com>

* exp_ch5.adb, exp_ch9.adb: Avoid use of No_Position in Sloc of
generated nodes, because it might confuse various circuits in the FE.

From-SVN: r145883

[multiple changes]

2009-04-10  Ed Schonberg  <schonberg@adacore.com>

* sem_prag.adb (Analyze_Pragma, case Task_Name): Do not expand argument
of pragma. It will be recopied and analyzed when used in call to
Create_Task.

* sem_res.adb (Resolve_Call): Clarify use of secondary stack within
initialization operations and recognize use of it in procedure calls
within init_procs.

* exp_ch9.adb (Make_Task_Create_Call): Copy full tree of Task_Name
argument, because it may have side-effects.

* exp_ch2.adb: Remove obsolete comments on default functions

2009-04-10  Jose Ruiz  <ruiz@adacore.com>

* adaint.c (RTX section): Do for RTX the same thing as we do for
Windows (include ctype.h and define a fallback ISALPHA if IN_RTS).

From-SVN: r145882

sem_aux.ads, [...] (Nearest_Current_Scope): New function.

2009-04-10  Robert Dewar  <dewar@adacore.com>

* sem_aux.ads, sem_aux.adb (Nearest_Current_Scope): New function.

* sem_res.adb (Resolve_Call): Fix test for
Suppress_Value_Tracking_On_Call (was wrong for the case of a call from
a non-dynamic scope).

From-SVN: r145881

[multiple changes]

2009-04-10  Robert Dewar  <dewar@adacore.com>

* make.adb: Add comment.
Minor reformatting

2009-04-10  Nicolas Setton  <setton@adacore.com>

* s-osprim-darwin.adb: New file.

* s-osinte-darwin.adb, s-osinte-darwin.ads: Fix binding to timespec.

2009-04-10  Thomas Quinot  <quinot@adacore.com>

* g-socket.ads: Add comment clarifying alignment requirement for Fd_Set

From-SVN: r145880

io.h (move_pos_offset): Remove prototype.

2009-04-10  Janne Blomqvist  <jb@gcc.gnu.org>

        * io/io.h (move_pos_offset): Remove prototype.
        * io/transfer.c (formatted_transfer_scalar_read): Use sseek
        instead of move_pos_offset.
        * io/unix.c (move_pos_offset): Remove.

From-SVN: r145876

PR libfortran/39665 libfortran/39702 libfortran/39709

2009-04-10  Janne Blomqvist  <jb@gcc.gnu.org>

PR libfortran/39665 libfortran/39702 libfortran/39709
* io/io.h (st_parameter_dt): Revert aligned attribute from u.p.value.
        * io/list_read.c (read_complex): Read directly into user pointer.
        (read_real): Likewise.
        (list_formatted_read_scalar): Update read_complex and read_real calls.
        (nml_read_obj): Read directly into user pointer.

From-SVN: r145875

re PR testsuite/35621 (gcc.target/i386/ssse3-*.c violates C/C++ aliasing rules)

2009-04-09  H.J. Lu  <hongjiu.lu@intel.com>

PR testsuite/35621
* gcc.target/i386/avx-vpabsb-1.c: Add -fno-strict-aliasing.
* gcc.target/i386/avx-vpabsb-1.c: Likewise.
* gcc.target/i386/avx-vpabsd-1.c: Likewise.
* gcc.target/i386/avx-vpabsw-1.c: Likewise.
* gcc.target/i386/avx-vpalignr-1.c: Likewise.
* gcc.target/i386/avx-vphaddd-1.c: Likewise.
* gcc.target/i386/avx-vphaddsw-1.c: Likewise.
* gcc.target/i386/avx-vphaddw-1.c: Likewise.
* gcc.target/i386/avx-vphsubd-1.c: Likewise.
* gcc.target/i386/avx-vphsubsw-1.c: Likewise.
* gcc.target/i386/avx-vphsubw-1.c: Likewise.
* gcc.target/i386/avx-vpmaddubsw-1.c: Likewise.
* gcc.target/i386/avx-vpmulhrsw-1.c: Likewise.
* gcc.target/i386/avx-vpshufb-1.c: Likewise.
* gcc.target/i386/avx-vpsignb-1.c: Likewise.
* gcc.target/i386/avx-vpsignd-1.c: Likewise.
* gcc.target/i386/avx-vpsignw-1.c: Likewise.

From-SVN: r145871

Daily bump.

From-SVN: r145870

PR libstdc++/39629 (final bits)

2009-04-09  Paolo Carlini  <paolo.carlini@oracle.com>

PR libstdc++/39629 (final bits)
* include/bits/random.h (subtract_with_carry_engine<>::_S_modulus):
Remove, consistently use __detail::_Shift<_UIntType, __w>::__value.
* include/bits/random.tcc (shuffle_order_engine<>::operator()()):
Fix formula for __j.
* include/bits/random.tcc (operator==(const normal_distribution<>&,
const normal_distribution<>&): Remove.
* include/bits/random.h (operator==(const uniform_int_distribution<>&,
const uniform_int_distribution<>&), operator==(const
uniform_real_distribution<>&, const uniform_real_distribution<>&),
operator==(const normal_distribution<>&, const normal_distribution<>&),
operator==(const lognormal_distribution<>&,
const lognormal_distribution<>&), operator==(const
chi_squared_distribution<>&, const chi_squared_distribution<>&),
operator==(const cauchy_distribution<>&, const cauchy_distribution<>&),
operator==(const fisher_distribution<>&, const fisher_distribution<>&),
operator==(const student_t_distribution<>&,
const student_t_distribution<>&), operator==(const
bernoulli_distribution<>&, const bernoulli_distribution<>&),
operator==(const binomial_distribution<>&, const
binomial_distribution<>&), operator==(const geometric_distribution<>&,
const geometric_distribution<>&), operator==(const
negative_binomial_distribution<>&, const
negative_binomial_distribution<>&), operator==(const
poisson_distribution<>&, const poisson_distribution<>&),
operator==(const exponential_distribution<>&,
const exponential_distribution<>&), operator==(const
gamma_distribution<>&, const gamma_distribution<>&), operator==(const
weibull_distribution<>&, const weibull_distribution<>&),
operator==(const extreme_value_distribution<>&,
const extreme_value_distribution<>&), operator==(const
discrete_distribution<>&, const discrete_distribution<>&),
operator==(const piecewise_constant_distribution<>&,
const piecewise_constant_distribution<>&), operator==(const
piecewise_linear_distribution<>&, const
piecewise_linear_distribution<>&)): Remove.
* testsuite/26_numerics/random/uniform_real_distribution/operators/
serialize.cc: Remove.
* testsuite/26_numerics/random/piecewise_constant_distribution/
operators/serialize.cc: Likewise.
* testsuite/26_numerics/random/chi_squared_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/normal_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/uniform_int_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/poisson_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/bernoulli_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/discrete_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/weibull_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/negative_binomial_distribution/
operators/serialize.cc: Likewise.
* testsuite/26_numerics/random/cauchy_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/gamma_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/fisher_f_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/exponential_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/lognormal_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/binomial_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/extreme_value_distribution/
operators/serialize.cc: Likewise.
* testsuite/26_numerics/random/piecewise_linear_distribution/
operators/serialize.cc: Likewise.
* testsuite/26_numerics/random/student_t_distribution/operators/
serialize.cc: Likewise.
* testsuite/26_numerics/random/geometric_distribution/operators/
serialize.cc: Likewise.

* include/bits/random.h (linear_congruential_engine<>::seed(_Gen&,
true_type), linear_congruential_engine<>::seed(_Gen&, false_type),
mersenne_twister_engine<>::seed(_Gen&,
true_type), mersenne_twister_engine<>::seed(_Gen&, false_type),
subtract_with_carry_engine<>::seed(_Gen&,
true_type), subtract_with_carry_engine<>::seed(_Gen&, false_type)):
Remove, unused.
* include/bits/random.tcc (linear_congruential_engine<>::
seed(_Gen&, false_type): Likewise.

From-SVN: r145867

re PR target/39678 (complex type isn't passed correctly)

gcc/

2009-04-09  H.J. Lu  <hongjiu.lu@intel.com>

PR target/39678
* config/i386/i386.c (classify_argument): Handle SCmode with
(bit_offset % 64) != 0.

gcc/testsuite/

2009-04-09  H.J. Lu  <hongjiu.lu@intel.com>

PR target/39678
* g++.dg/torture/pr39678.C: New.
* gcc.dg/compat/struct-complex-2.h: Likewise.
* gcc.dg/compat/struct-complex-2_main.c: Likewise.
* gcc.dg/compat/struct-complex-2_x.c: Likewise.
* gcc.dg/compat/struct-complex-2_y.c: Likewise.
* gcc.dg/torture/pr39678.c: Likewise.
* gcc.target/i386/pr39678.c: Likewise.

* gcc.dg/compat/struct-complex-1_x.c: Add -Wno-psabi.
* gcc.dg/compat/struct-complex-1_y.c: Likewise.

* gcc.target/x86_64/abi/test_passing_structs.c: Include
<complex.h>.  Add tests for structure with complex float.

From-SVN: r145865

invoke.texi (Optimize Options): Add cross-reference to -Q --help=optimizers examples.

2009-04-09  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/invoke.texi (Optimize Options): Add cross-reference to
-Q --help=optimizers examples.

From-SVN: r145864

configure.host: Restore match for darwin9 or later.

2009-04-09  Jack Howarth  <howarth@bromo.med.uc.edu>

* configure.host: Restore match for darwin9 or later.
* configure.ac: Likewise.
* configure: Regenerate.

From-SVN: r145861

configure.ac: Restore match for darwin9 or later.

2009-04-09  Jack Howarth  <howarth@bromo.med.uc.edu>

* configure.ac: Restore match for darwin9 or later. Use double
brackets since regeneration eats one pair.
* configure: Regenerate.

From-SVN: r145860

re PR target/36800 (va_arg for _Decimal128 on 32-bit Power mishandled in certain cases)

PR target/36800

PR target/36800
* config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Do not set
regalign for the reg == fpr and TDmode case.

testsuite/
* gcc.dg/dfp/pr36800.c: New.

Co-Authored-By: Joseph Myers <joseph@codesourcery.com>
From-SVN: r145859

re PR objc/29200 (%s substituted with "methods"/"selectors" can't be properly translated)

gcc/
2009-04-09  David Ayers  <ayers@fsfe.org>

PR objc/29200
* objc/objc-act.c (warn_with_method): Remove helper function.
(check_duplicates): Call warning and inform directly.
(really_start_method): Likewise.

From-SVN: r145857

ChangeLog: Fix ChangeLog.

* gcc/ChangeLog: Fix ChangeLog.
* gcc/testsuite/ChangeLog: Ditto.

From-SVN: r145855

Fix entry

From-SVN: r145853

re PR libfortran/39665 (Fortran IO using unaligned accesses to read/write doubles.)

2009-04-09  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/39665
* io/io.h (st_parameter_dt): Add aligned attribute to u.p.value.
* io/read.c (convert_real): Add note about alignment requirements.

From-SVN: r145852

help.exp: New file.

gcc/testsuite/
* gcc.misc-tests/help.exp: New file.
* lib/options.exp: New file.

Co-Authored-By: Rafael Avila de Espindola <espindola@google.com>
From-SVN: r145851

re PR libobjc/36610 (objc_msg_sendv is broken for targets which pass argument via registers)

PR libobjc/36610
* objc/execute/forward-1.x: New.

From-SVN: r145849

expmed.c (expand_divmod): Always use a comparison for a division by a large unsigned integer.

2009-04-09  Paolo Bonzini  <bonzini@gnu.org>

* expmed.c (expand_divmod): Always use a comparison for a division
by a large unsigned integer.

        * fold-const.c (tree_single_nonzero_warnv_p): Always treat decls
        for things others than variables or functions as nonzero.

testsuite:
2009-04-09  Paolo Bonzini  <bonzini@gnu.org>

        * gcc.dg/pr27150-1.c: Change to a link test.

From-SVN: r145846

Move gcc ChangeLog entry in revision 145415 to gcc/ChangeLog.

From-SVN: r145844

Change spaces to tab.

From-SVN: r145843

re PR c++/10039 ([macosx] gcc hangs at a c++-Constructor with switch "-Os")

2009-04-09  H.J. Lu  <hongjiu.lu@intel.com>

PR gas/10039
* configure.ac: Require texinfo 4.7.
* configure: Regenerated.

From-SVN: r145842

Licensing changes to GPLv3 resp. GPLv3 with GCC Runtime Exception.

From-SVN: r145841

re PR testsuite/39696 (gcc.dg/tree-ssa/ssa-ccp-25.c scan-tree-dump doesn't work on Linux/ia32)

2009-04-09  Richard Guenther  <rguenther@suse.de>

PR testsuite/39696
* ssa-ccp-25.c: Adjust pattern.

From-SVN: r145839

[multiple changes]

2009-04-09  Nicolas Setton  <setton@adacore.com>

* s-osinte-darwin.ads: Fix wrong binding to struc timeval.

* s-osinte-darwin.adb (To_Timeval): Adapt to fixed implementation of
struct_timeval.

2009-04-09  Bob Duff  <duff@adacore.com>

* exp_ch5.adb, exp_ch9.adb: Correct miscellaneous Slocs in
internally-generated nodes related to select statements to avoid
confusing the debugger.

2009-04-09  Pascal Obry  <obry@adacore.com>

* make.adb: Ensure that all linker arguments are duplicated.

From-SVN: r145838

[multiple changes]

2009-04-09  Robert Dewar  <dewar@adacore.com>

* sem_ch5.adb: Minor reformatting

2009-04-09  Vincent Celier  <celier@adacore.com>

* vms_data.ads:
Change GNAT CHECK qualifier /DIAGNOSIS_LIMIT to /DIAGNOSTIC_LIMIT
New qualifier /LEXPAND_SOURCE=nnn for -gnatGnnn
New qualifier /LXDEBUG=nnn for -gnatDnnn
For H820-010

* gnat_ugn.texi:
Update documentation for VMS qualifiers equivalent to -gnatGnn and
-gnatDnn

2009-04-09  Nicolas Setton  <setton@adacore.com>

* s-osinte-darwin.ads: (Pad_Type): Make this an array of unsigned_long,
to match layout of siginfo_t in sys/signal.h.

* gcc-interface/Makefile.in: Add section for x86_64 darwin.

2009-04-09  Thomas Quinot  <quinot@adacore.com>

* g-socket.ads: (Fd_Set): Use Interfaces.C.long alignment.

2009-04-09  Nicolas Setton  <setton@adacore.com>

* s-oscons-tmplt.c: Allow long lines in the generated spec.
Add generation of Darwin-specific constants needed when binding to the
pthread library.

From-SVN: r145837

2009-04-09  Robert Dewar  <dewar@adacore.com>

        * checks.adb:
        (Insert_Valid_Check): Avoid unnecessary generation of junk declaration
        when no invalid values exist, Avoid duplicate read of atomic variable.

        * cstand.adb (Build_Signed_Integer_Type): Set Is_Known_Valid
        (Standard_Unsigned): Set Is_Known_Valid

        * sem_ch3.adb (Analyze_Subtype_Declaration): Copy Is_Known_Valid on
subtype declaration if no constraint.
        (Set_Modular_Size): Set Is_Known_Valid if appropriate
        (Build_Derived_Numeric_Type): Copy Is_Known_Valid if no constraint

From-SVN: r145836

Minor editing.

From-SVN: r145835

[multiple changes]

2009-04-09  Robert Dewar  <dewar@adacore.com>

* switch-c.adb, gnat_ugn.texi, vms_data.ads, switch.adb,
switch.ads: for numeric switches, an optional equal sign is always
allowed.

2009-04-09  Vincent Celier  <celier@adacore.com>

* prj-nmsc.adb (Get_Unit): Do not consider Casing on platform where
the case of file names is not significant.

From-SVN: r145834

New file.

From-SVN: r145833

[multiple changes]

2009-04-09  Vincent Celier  <celier@adacore.com>

* errout.adb: Remove dependency on package Style

* style.ads, styleg.adb, styleg.ads (RM_Column_Check): Remove function,
moved to Stylesw.

* stylesw.ads, stylesw.adb (RM_Column_Check): New function, moved from
Styleg.

* errutil.adb, par.adb: Import Stylesw

2009-04-09  Arnaud Charlet  <charlet@adacore.com>

* opt.ads: Fix typos.

2009-04-09  Robert Dewar  <dewar@adacore.com>

* einfo.adb: Minor reformatting

From-SVN: r145832

Make-lang.in, [...]: Reorganize style units.

2009-04-09  Robert Dewar  <dewar@adacore.com>

* gcc-interface/Make-lang.in, style.ads, style.adb: Reorganize style
units.

* styleg-c.ads, styleg-c.adb: Removed, no longer used.

From-SVN: r145831

[multiple changes]

2009-04-09  Robert Dewar  <dewar@adacore.com>

* g-comver.adb: Minor reformatting.

2009-04-09  Thomas Quinot  <quinot@adacore.com>

* lib-load.ads (Load_Unit): Update documentation.

2009-04-09  Ed Schonberg  <schonberg@adacore.com>

* lib-load.adb (Load_Unit): When loading the parent of a child unit
named in a with_clause, retain the with_clause to preserve a
limited_with indication.

From-SVN: r145829

i386.md (cmpcc): New.

2009-04-09  Paolo Bonzini  <bonzini@gnu.org>

        * config/i386/i386.md (cmpcc): New.
        * config/i386/sync.md (sync_compare_and_swap*): Set FLAGS_REG.
        (sync_compare_and_swap_cc*): Delete.

        * config/s390/s390.c (s390_compare_emitted): Remove.
        (s390_emit_compare): Handle MODE_CC s390_compare_op0 like
        s390_compare_emitted used to be handled.  Assert that modes match.
        (s390_emit_compare_and_swap): Use s390_emit_compare, do not
        refer to sync_compare_and_swap_ccsi.
        * config/s390/s390.h (s390_compare_emitted): Remove.
        * config/s390/s390.md (seq): Look for MODE_CC s390_compare_op0
        instead of s390_compare_emitted.
        (stack_protect_test, sync_compare_and_swap_cc): Set s390_compare_op0
        instead of s390_compare_emitted.
        * config/s390/s390.md (cmpcc): New.
        (sync_compare_and_swapqi, sync_compare_and_swaphi): Clobber
        CC_REGNUM, do not pretend it's set.
        (sync_compare_and_swap_cc*): Delete.
        * config/s390/predicates.md (cc_reg_operand): New.

        * expr.c (sync_compare_and_swap_cc): Delete.
        * optabs.h (sync_compare_and_swap_cc): Delete.
        * optabs.c (prepare_cmp_insn): Ignore which specific CCmode
        is being used with can_compare_p.
        (emit_cmp_and_jump_insn_1): Likewise when looking in the optab.
        (find_cc_set): New.
        (expand_bool_compare_and_swap): Do not use sync_compare_and_swap_cc,
        look for a MODE_CC set instead.  Use emit_store_flag.
        (expand_compare_and_swap_loop): Likewise, with some additional
        complication to avoid a force_reg when useless.  Use
        emit_cmp_and_jump_insns.
        * genopinit.c (optabs): Delete sync_compare_and_swap_cc.
        * doc/md.texi (sync_compare_and_swap_cc): Merge with
        sync_compare_and_swap documentation.

java:
2009-04-09  Paolo Bonzini  <bonzini@gnu.org>

        * builtins.c (compareAndSwapLong_builtin,
        compareAndSwapInt_builtin, compareAndSwapObject_builtin,
        VMSupportsCS8_builtin): Do not look at sync_compare_and_swap_cc.

From-SVN: r145825

Minor reformatting.

From-SVN: r145824

sem_ch7.adb, [...]: Code clean up...

2009-04-09  Robert Dewar  <dewar@adacore.com>

* sem_ch7.adb, sem_ch10.adb, sem_prag.adb, sem_ch12.adb, sem_util.adb,
exp_ch13.adb, sem_ch6.adb, exp_disp.adb, sem_ch8.adb, sem_warn.adb,
sem_cat.adb: Code clean up: use Is_Package_Or_Generic_Package where
possible to replace an OR of two separate tests.

From-SVN: r145823

binderr.adb, [...]: New circuitry for handling Maximum_Messages.

2009-04-09  Robert Dewar  <dewar@adacore.com>

* binderr.adb, errout.adb, errutil.adb: New circuitry for handling
Maximum_Messages.

* erroutc.adb, erroutc.ads (Warnings_Suppressed): Now tests global
warning status as well.

* opt.ads (Maximum_Messages): New name for Maximum_Errors.

* switch-b.adb, switch-c.adb: Change name Maximum_Errors to
Maximum_Messages.

* bindusg.adb, usage.adb: Update line for -gnatm switch

* gnat_ugn.texi: Update documentation for -gnatmnn compiler switch and
-mnn binder switch.

* sem_ch10.adb: Minor reformatting.

From-SVN: r145822