gcc.git
16 years agore PR fortran/32600 ([ISO Bind C] C_F_POINTER w/o SHAPE should not be a library function)
Christopher D. Rickett [Mon, 15 Oct 2007 19:58:55 +0000 (19:58 +0000)]
re PR fortran/32600 ([ISO Bind C] C_F_POINTER w/o SHAPE should not be a library function)

2007-10-15 Christopher D. Rickett <crickett@lanl.gov>

        PR fortran/32600
        * trans-expr.c (gfc_conv_function_call): Generate code to inline
        c_associated.
        * symbol.c (get_iso_c_sym): Preserve from_intmod and
        * intmod_sym_id
        attributes in the resolved symbol.
        * resolve.c (gfc_iso_c_sub_interface): Remove dead code.

2007-10-15 Christopher D. Rickett <crickett@lanl.gov>

        PR fortran/32600
        * libgfortran/intrinsics/iso_c_binding.c: Remove c_associated_1
        and c_associated_2.
        * libgfortran/intrinsics/iso_c_binding.h: Ditto.
        * libgfortran/gfortran.map: Ditto.

From-SVN: r129367

16 years agore PR tree-optimization/33136 (wrong code due to alias with allocation in loop)
Jakub Jelinek [Mon, 15 Oct 2007 18:29:54 +0000 (20:29 +0200)]
re PR tree-optimization/33136 (wrong code due to alias with allocation in loop)

PR tree-optimization/33136
* opts.c (decode_options): Don't enable flag_ipa_type_escape.

* gcc.c-torture/execute/20070824-1.c: New test.
* gcc.dg/pr33136-1.c: New test.
* gcc.dg/pr33136-2.c: New test.
* gcc.dg/pr33136-3.c: New test.

From-SVN: r129366

16 years agotls.m4 (GCC_CHECK_TLS): Rename have_tls to gcc_cv_have_tls.
Maciej W. Rozycki [Mon, 15 Oct 2007 18:09:32 +0000 (18:09 +0000)]
tls.m4 (GCC_CHECK_TLS): Rename have_tls to gcc_cv_have_tls.

config/:
2007-10-15  Maciej W. Rozycki  <macro@linux-mips.org>

* tls.m4 (GCC_CHECK_TLS): Rename have_tls to gcc_cv_have_tls.
(GCC_CHECK_CC_TLS): Rename have_cc_tls to gcc_cv_have_cc_tls.

libgomp/:
2007-10-15  Maciej W. Rozycki  <macro@linux-mips.org>

* configure: Regenerate following changes to ../config/tls.m4.

libjava/:
2007-10-15  Maciej W. Rozycki  <macro@linux-mips.org>

* configure: Regenerate following changes to ../config/tls.m4.

libmudflap/:
2007-10-15  Maciej W. Rozycki  <macro@linux-mips.org>

* configure: Regenerate following changes to ../config/tls.m4.

From-SVN: r129364

16 years agore PR tree-optimization/33735 (verify_stmts failed: missing PHI def)
Alexandre Oliva [Mon, 15 Oct 2007 17:07:20 +0000 (17:07 +0000)]
re PR tree-optimization/33735 (verify_stmts failed: missing PHI def)

gcc/ChangeLog:
PR tree-optimization/33735
PR tree-optimization/33572
* tree-inline.c (update_ssa_across_abnormal_edges): Revert
2007-10-09's change.
* except.c (duplicate_eh_regions): Don't look for prev_try
beyond ERT_ALLOWED_EXCEPTIONS with an empty list.
gcc/testsuite/ChangeLog:
PR tree-optimization/33735
* g++.dg/torture/pr33735.C: New.

From-SVN: r129356

16 years agore PR middle-end/33706 (gcc_assert failure in verify_eh_edges)
Alexandre Oliva [Mon, 15 Oct 2007 17:05:19 +0000 (17:05 +0000)]
re PR middle-end/33706 (gcc_assert failure in verify_eh_edges)

gcc/ChangeLog:
PR middle-end/33706
* tree-inline.c (copy_bb): Use bsi_replace to replace a
__builtin_va_arg_pack-containing call stmt.
gcc/testsuite/ChangeLog:
PR middle-end/33706
* gcc.dg/va-arg-pack-2.c: New.

From-SVN: r129355

16 years agomatrix-reorg.c (gate_matrix_reorg): Don't comment out whole program flag.
Razya Ladelsky [Mon, 15 Oct 2007 15:16:25 +0000 (15:16 +0000)]
matrix-reorg.c (gate_matrix_reorg): Don't comment out whole program flag.

2007-10-15  Razya Ladelsky  <razya@il.ibm.com>

        * matrix-reorg.c (gate_matrix_reorg): Don't comment out whole
        program flag.

From-SVN: r129351

16 years agore PR tree-optimization/33619 (TER breaks some inline-asm code (again))
Jakub Jelinek [Mon, 15 Oct 2007 15:14:46 +0000 (17:14 +0200)]
re PR tree-optimization/33619 (TER breaks some inline-asm code (again))

PR tree-optimization/33619
* tree-ssa-ter.c (is_replaceable_p): Return false for all
calls.

* gcc.dg/pr33619.c: New test.

From-SVN: r129350

16 years agolib/target-supports.exp (check_vmx_hw_available): Add -maltivec flag for AIX.
David Edelsohn [Mon, 15 Oct 2007 15:00:49 +0000 (15:00 +0000)]
lib/target-supports.exp (check_vmx_hw_available): Add -maltivec flag for AIX.

lib/target-supports.exp (check_vmx_hw_available): Add -maltivec
flag for AIX.
(check_effective_target_powerpc_altivec_ok): Only skip before AIX 5.3.

From-SVN: r129349

16 years ago* config.gcc (powerpc-ibm-aix5*): Install altivec.h
David Edelsohn [Mon, 15 Oct 2007 14:52:43 +0000 (14:52 +0000)]
* config.gcc (powerpc-ibm-aix5*): Install altivec.h

From-SVN: r129348

16 years agofold-const.c (extract_array_ref): Remove.
Richard Guenther [Mon, 15 Oct 2007 14:49:55 +0000 (14:49 +0000)]
fold-const.c (extract_array_ref): Remove.

2007-10-15  Richard Guenther  <rguenther@suse.de>

* fold-const.c (extract_array_ref): Remove.
(fold_comparison): Handle POINTER_PLUS_EXPR with the
generic address expression comparison folding.  Remove
the folding that used extract_array_ref.

From-SVN: r129347

16 years agore PR libfortran/33055 (Runtime error in INQUIRE unit existance with -fdefault-integer-8)
Jerry DeLisle [Mon, 15 Oct 2007 14:03:52 +0000 (14:03 +0000)]
re PR libfortran/33055 (Runtime error in INQUIRE unit existance with -fdefault-integer-8)

2007-10-15  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

PR libfortran/33055
* gfortran.dg/inquire_11.f90: New test.
* gfortan.dg/negative_unit_int8.f: New test.

From-SVN: r129346

16 years agoResync.
Arnaud Charlet [Mon, 15 Oct 2007 14:00:38 +0000 (16:00 +0200)]
Resync.

From-SVN: r129345

16 years agore PR libfortran/33055 (Runtime error in INQUIRE unit existance with -fdefault-integer-8)
Jerry DeLisle [Mon, 15 Oct 2007 13:59:02 +0000 (13:59 +0000)]
re PR libfortran/33055 (Runtime error in INQUIRE unit existance with -fdefault-integer-8)

2007-10-15  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

PR libfortran/33055
* io/inquire.c (inquire_via_unit):  If inquiring by unit, check for
an error condition from the IOSTAT variable and set EXIST to false if
there was a bad unit number.

From-SVN: r129344

16 years agoerrout.ads: Comment clarification
Robert Dewar [Mon, 15 Oct 2007 13:58:20 +0000 (15:58 +0200)]
errout.ads: Comment clarification

2007-10-15  Robert Dewar  <dewar@adacore.com>

* errout.ads: Comment clarification

* exp_ch4.adb (Expand_N_Allocator): Code cleanup.
(Expand_N_Op_Eq): Improve handling of array equality with -gnatVa

* lib.ads: Comment update

* init.c: Minor reformatting.

* sem_attr.adb: Minor formatting

* osint-b.ads: Minor reformatting

* sem_ch9.adb: Implement -gnatd.I switch

* g-comlin.adb: (Start): Fix handling of empty command line.

* gnatcmd.adb (GNATCmd): Do not put the -rules in the -cargs section,
even when -rules follows the -cargs section.

From-SVN: r129343

16 years agog-soccon-vms.ads: Fix value of MSG_WAITALL.
Thomas Quinot [Mon, 15 Oct 2007 13:58:10 +0000 (15:58 +0200)]
g-soccon-vms.ads: Fix value of MSG_WAITALL.

2007-10-15  Thomas Quinot  <quinot@adacore.com>

* g-soccon-vms.ads: Fix value of MSG_WAITALL.

* gen-soccon.c:
Update documentation to note that OpenVMS 8.3 or later must be used
to generate g-soccon-vms.ads.

* atree.adb: Add ??? comment

* exp_util.adb: Minor reformatting.
Add ??? comment in Kill_Dead_Code.

From-SVN: r129342

16 years agognat_ugn.texi: Add documentation for switches --version and --help for the GNAT tools...
Vincent Celier [Mon, 15 Oct 2007 13:57:59 +0000 (15:57 +0200)]
gnat_ugn.texi: Add documentation for switches --version and --help for the GNAT tools gnatbind...

2007-10-15  Vincent Celier  <celier@adacore.com>
    Robert Dewar  <dewar@adacore.com>

* gnat_ugn.texi: Add documentation for switches --version and --help
for the GNAT tools gnatbind, gnatlink, gnatmake, gnatchop, gnatname,
gnatxref, gnatfind, gnatls, and gnatclean.
Document -gnatw.o.
Mention attribute Excluded_Source_Dirs
Replace obsolescent attribute Locally_Removed_Files with attribute
Excluded_Source_Files.
Improve documentation of -u (gnatbind)
Document how to do reliable stack checking for the environmental task
on iVMS.

* gnat_rm.texi: Rewrite section about No_Implicit_Dynamic_Code.
Document attribute Excluded_Source_Files and indicate that attribute
Locally_Removed_Files is obsolescent.

From-SVN: r129341

16 years agoxref_lib.adb (Get_Full_Type): Add support for the 'h' entity type, ie interfaces.
Emmanuel Briot [Mon, 15 Oct 2007 13:57:46 +0000 (15:57 +0200)]
xref_lib.adb (Get_Full_Type): Add support for the 'h' entity type, ie interfaces.

2007-10-15  Emmanuel Briot  <briot@adacore.com>

* xref_lib.adb (Get_Full_Type): Add support for the 'h' entity type, ie
interfaces.

* xr_tabls.adb (Add_Reference): Add support for the new 'R' reference
type, for dispatching calls.

From-SVN: r129340

16 years agotrans.c (gnat_to_gnu): Remove the padding structure more often.
Tristan Gingold [Mon, 15 Oct 2007 13:57:36 +0000 (15:57 +0200)]
trans.c (gnat_to_gnu): Remove the padding structure more often.

2007-10-15  Tristan Gingold  <gingold@adacore.com>

* trans.c (gnat_to_gnu): Remove the padding structure more often.
This optimize assignment to over-aligned record.

From-SVN: r129339

16 years agotb-alvms.c (tb_entry_t, [...]): Store a frame pointer instead of a procedure value...
Olivier Hainque [Mon, 15 Oct 2007 13:57:26 +0000 (15:57 +0200)]
tb-alvms.c (tb_entry_t, [...]): Store a frame pointer instead of a procedure value in each traceback entry.

2007-10-15  Olivier Hainque  <hainque@adacore.com>

* tb-alvms.c (tb_entry_t, __gnat_backtrace): Store a frame pointer
instead of a procedure value in each traceback entry.

* g-trasym-vms-alpha.adb (Symbolic_Traceback): Pass frame pointer
instead of procedure value to TBK$SYMBOLIZE.

* s-traent-vms.adb (PV_For): Rename as FP_For and access the proper
field.
(TB_Entry_For): Account for the PV/FP renaming.

* s-traent-vms.ads (Traceback_Entry): Rename PV component into FP and
add comment.
(Null_TB_Entry): Account for change of component name.
(PV_For): Rename as FP_For.

From-SVN: r129338

16 years agosem_disp.adb (Find_Controlling_Arg): Examine the call node before examining its origi...
Ed Schonberg [Mon, 15 Oct 2007 13:57:17 +0000 (15:57 +0200)]
sem_disp.adb (Find_Controlling_Arg): Examine the call node before examining its original form...

2007-10-15  Ed Schonberg  <schonberg@adacore.com>

* sem_disp.adb (Find_Controlling_Arg): Examine the call node before
examining its original form, to handle properly operator calls that
have been rewritten.

From-SVN: r129337

16 years agosem_ch6.adb (Find_Corresponding_Spec): If the previous entity is a body generated...
Ed Schonberg [Mon, 15 Oct 2007 13:57:06 +0000 (15:57 +0200)]
sem_ch6.adb (Find_Corresponding_Spec): If the previous entity is a body generated for a function with a controlling...

2007-10-15  Ed Schonberg  <schonberg@adacore.com>

* sem_ch6.adb (Find_Corresponding_Spec): If the previous entity is a
body generated for a function with a controlling result that is a null
extension, discard the generated body in favor of the current explicit
one.

From-SVN: r129336

16 years agosystem-vms-ia64.ads: New file.
Tristan Gingold [Mon, 15 Oct 2007 13:56:55 +0000 (15:56 +0200)]
system-vms-ia64.ads: New file.

2007-10-15  Tristan Gingold  <gingold@adacore.com>

* system-vms-ia64.ads: New file.

* system-vms_64.ads: Minor comment fix.

From-SVN: r129335

16 years agosem_ch4.adb: Minor code and comment reformatting.
Hristian Kirtchev [Mon, 15 Oct 2007 13:56:46 +0000 (15:56 +0200)]
sem_ch4.adb: Minor code and comment reformatting.

2007-10-15  Hristian Kirtchev  <kirtchev@adacore.com>

* sem_ch4.adb: Minor code and comment reformatting.
(Analyze_Allocator): When the designated type of an unconstrained
allocator is a record with unknown discriminants or an array with
unknown range bounds, emit a detailed error message depending on the
compilation mode and whether the designated type is limited.

From-SVN: r129334

16 years agosem_ch3.adb (Build_Derived_Concurrent_Type): Set the Is_Constrained flag of derived...
Gary Dismukes [Mon, 15 Oct 2007 13:56:36 +0000 (15:56 +0200)]
sem_ch3.adb (Build_Derived_Concurrent_Type): Set the Is_Constrained flag of derived concurrent types...

2007-10-15  Gary Dismukes  <dismukes@adacore.com>

* sem_ch3.adb (Build_Derived_Concurrent_Type): Set the Is_Constrained
flag of derived concurrent types, taking into account the flag setting
on the parent subtype and any new set of discriminants.

From-SVN: r129333

16 years agosem_ch12.adb: Minor code reformatting.
Hristian Kirtchev [Mon, 15 Oct 2007 13:56:26 +0000 (15:56 +0200)]
sem_ch12.adb: Minor code reformatting.

2007-10-15  Hristian Kirtchev  <kirtchev@adacore.com>

* sem_ch12.adb: Minor code reformatting.
(Check_Generic_Child_Unit): Iterate over the homonym chain in order to
find the parent package which may have been hidden by local
declarations.

From-SVN: r129332

16 years agosem_aggr.adb (Resolve_Record_Aggregate): In the case of a box association for an...
Gary Dismukes [Mon, 15 Oct 2007 13:56:15 +0000 (15:56 +0200)]
sem_aggr.adb (Resolve_Record_Aggregate): In the case of a box association for an access component...

2007-10-15  Gary Dismukes  <dismukes@adacore.com>

* sem_aggr.adb (Resolve_Record_Aggregate): In the case of a box
association for an access component, add an association with null as
the expression. Remove testing for array subtypes and the setting in
that case of Ctyp to the array component type, which prevented proper
inclusion of an association for null-initialized arrays. Collapse
condition that tests for array subtypes into just a test of
Is_Partially_Initialized_Type (which already covers arrays anyway).

From-SVN: r129331

16 years agortsfind.adb: (Load_RTU): Turn off style checks for Load call
Robert Dewar [Mon, 15 Oct 2007 13:56:05 +0000 (15:56 +0200)]
rtsfind.adb: (Load_RTU): Turn off style checks for Load call

2007-10-15  Robert Dewar  <dewar@adacore.com>

* rtsfind.adb: (Load_RTU): Turn off style checks for Load call

From-SVN: r129330

16 years agosnames.adb, snames.ads: Add new standard name runtime_library_dir
Vincent Celier [Mon, 15 Oct 2007 13:55:54 +0000 (15:55 +0200)]
snames.adb, snames.ads: Add new standard name runtime_library_dir

2007-10-15  Vincent Celier  <celier@adacore.com>

* snames.adb, snames.ads: Add new standard name runtime_library_dir

* prj.ads (Language_Config): Add new component Runtime_Library_Dir

* prj-attr.adb: Add project level attribute Runtime_Library_Dir

* prj-env.adb (Create_Mapping_File): Do not put an entry if the path of
the source is unknown.

* prj-ext.adb: Spelling error fix

* prj-nmsc.adb (Check_Ada_Name): Reject any unit that includes an Ada
95 reserved word in its name.
(Process_Project_Level_Array_Attributes): Process new attribute
Runtime_Library_Dir.

* prj-part.adb (Parse_Single_Project): Do not check the name of the
config project against the user project names.

* prj-proc.adb (Expression): In multi-language mode, indexes that do
not include a dot are always case insensitive.
(Process_Declarative_Items): Ditto
(Process_Project_Tree_Phase_1): Set Success to False in case an error is
detected.

* prj-util.adb (Value_Of (In_Array)): When Force_Lower_Case_Index is
True, compare both indexes in lower case.

From-SVN: r129329

16 years agore PR libfortran/33055 (Runtime error in INQUIRE unit existance with -fdefault-integer-8)
Jerry DeLisle [Mon, 15 Oct 2007 13:55:47 +0000 (13:55 +0000)]
re PR libfortran/33055 (Runtime error in INQUIRE unit existance with -fdefault-integer-8)

2007-10-15  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

PR fortran/33055
* trans-io.c (create_dummy_iostat): New function to create a unique
dummy variable expression to use with IOSTAT.
(gfc_trans_inquire): Use the new function to pass unit number error info
to run-time library if a regular IOSTAT variable was not given.

From-SVN: r129328

16 years agosem_case.adb, [...]: Replace use of Heap_Sort_A (passing'Unrestricted_Access of neste...
Bob Duff [Mon, 15 Oct 2007 13:55:27 +0000 (15:55 +0200)]
sem_case.adb, [...]: Replace use of Heap_Sort_A (passing'Unrestricted_Access of nested subprograms...

2007-10-15  Bob Duff  <duff@adacore.com>

* sem_case.adb, sem_ch13.adb, lib-sort.adb: Replace use of Heap_Sort_A
(passing'Unrestricted_Access of nested subprograms to Sort) with use of
the generic Heap_Sort_G, in order to avoid trampolines.

From-SVN: r129327

16 years agomlib-tgt-specific.adb (Support_For_Libraries): New function...
Vincent Celier [Mon, 15 Oct 2007 13:55:17 +0000 (15:55 +0200)]
mlib-tgt-specific.adb (Support_For_Libraries): New function...

2007-10-15  Vincent Celier  <celier@adacore.com>

* mlib-tgt-specific.adb (Support_For_Libraries): New function,
returning None, used when there is no platform specific body for
MLib.Tgt.Specific.

From-SVN: r129326

16 years agoexp_dist.adb (Add_RACW_Primitive_Declarations_And_Bodies): Do not attempt to generate...
Thomas Quinot [Mon, 15 Oct 2007 13:55:07 +0000 (15:55 +0200)]
exp_dist.adb (Add_RACW_Primitive_Declarations_And_Bodies): Do not attempt to generate stubs for hidden primitive operations.

2007-10-15  Thomas Quinot  <quinot@adacore.com>

* exp_dist.adb (Add_RACW_Primitive_Declarations_And_Bodies): Do not
attempt to generate stubs for hidden primitive operations.

From-SVN: r129325

16 years agoexp_ch9.adb (Actual_Index_Expression): When the expansion occurs inside a generic...
Hristian Kirtchev [Mon, 15 Oct 2007 13:54:57 +0000 (15:54 +0200)]
exp_ch9.adb (Actual_Index_Expression): When the expansion occurs inside a generic body...

2007-10-15  Hristian Kirtchev  <kirtchev@adacore.com>

* exp_ch9.adb (Actual_Index_Expression): When the expansion occurs
inside a generic body, retrieve the full view of the entry family
discrete subtype if available.

From-SVN: r129324

16 years agoexp_ch3.adb (Build_Init_Procedure): Keep separate the initialization of tagged types...
Ed Schonberg [Mon, 15 Oct 2007 13:54:47 +0000 (15:54 +0200)]
exp_ch3.adb (Build_Init_Procedure): Keep separate the initialization of tagged types whose ultimate ancestor is a...

2007-10-15  Ed Schonberg  <schonberg@adacore.com>

* exp_ch3.adb (Build_Init_Procedure): Keep separate the initialization
of tagged types whose ultimate ancestor is a CPP type.
(Freeze_Array_Type): For a packed array type, generate an initialization
procedure if the type is public, to handle properly a client that
specifies Normalize_Scalars.

From-SVN: r129323

16 years agoexp_attr.adb (Expand_N_Attribute_Reference): Case Access, Unchecked_Access, and Unres...
Javier Miranda [Mon, 15 Oct 2007 13:54:33 +0000 (15:54 +0200)]
exp_attr.adb (Expand_N_Attribute_Reference): Case Access, Unchecked_Access, and Unrestricted_Access.

2007-10-15  Javier Miranda  <miranda@adacore.com>

* exp_attr.adb (Expand_N_Attribute_Reference): Case Access,
Unchecked_Access, and Unrestricted_Access. Cleanup code that takes
care of access to class-wide interface types plus removal of bizarre
conversion of tagged object to access type (reported by Gary
Dismukes). After this patch there is no need to perform any
additional management on these nodes in Expand_Interface_Actuals.

* exp_disp.adb (Expand_Interface_Actuals): Code cleanup. Remove code
that handles use of 'Access and 'Unchecked_Access applied to
actuals covering interface types. Such code is now
centralized in Expand_N_Attribute_Reference.

From-SVN: r129322

16 years agodebug.adb: Document use of -gnatd.a and -gnatd.I
Tristan Gingold [Mon, 15 Oct 2007 13:54:21 +0000 (15:54 +0200)]
debug.adb: Document use of -gnatd.a and -gnatd.I

2007-10-15  Tristan Gingold  <gingold@adacore.com>

* debug.adb: Document use of -gnatd.a and -gnatd.I

* layout.adb: On OpenVMS -gnatd.a disables alignment optimization.

From-SVN: r129321

16 years agoali.ads, ali.adb (Scan_ALI): Initialize XE.Tref to a known default value.
Emmanuel Briot [Mon, 15 Oct 2007 13:54:12 +0000 (15:54 +0200)]
ali.ads, ali.adb (Scan_ALI): Initialize XE.Tref to a known default value.

2007-10-15  Emmanuel Briot  <briot@adacore.com>

* ali.ads, ali.adb (Scan_ALI): Initialize XE.Tref to a known default
value.
(Xref_Record): Change type for Line, since in the case of a reference to
a predefined entity (as happens for array index types), the line is set
to 0.
Add support for parsing multiple array index types info, or
multiple inherited interfaces info. This information cannot be stored
in Xref_Entity_Record, which only supports a single instance of Tref_*,
and is therefore stored in the list of references instead. It has a
special treatement later on in tools that use this information.

From-SVN: r129320

16 years agoadaint.c, [...]: Initial port to arm-mentor-nucleus.
Geert Bosch [Mon, 15 Oct 2007 13:54:02 +0000 (15:54 +0200)]
adaint.c, [...]: Initial port to arm-mentor-nucleus.

2007-10-15  Geert Bosch  <bosch@adacore.com>

* adaint.c, socket.c, cal.c: Initial port to arm-mentor-nucleus.

* expect.c: Initial port to arm-mentor-nucleus.
Use kill for __gnat_kill() on VMS.

From-SVN: r129319

16 years agos-taprop-solaris.adb, [...]: Minor reformatting.
Robert Dewar [Mon, 15 Oct 2007 13:53:48 +0000 (15:53 +0200)]
s-taprop-solaris.adb, [...]: Minor reformatting.

2007-10-15  Robert Dewar  <dewar@adacore.com>

* s-taprop-solaris.adb, s-taprop-vms.adb, s-taprop-mingw.adb,
s-taprop-vxworks.adb, s-taprop-posix.adb, a-calend-vms.adb,
a-calend.adb, a-nuflra.adb, a-tigeau.adb, a-wtgeau.adb,
checks.adb, bindgen.adb, eval_fat.adb, exp_fixd.adb, fmap.adb,
freeze.adb, g-awk.adb, g-calend.adb, g-diopit.adb, g-expect.adb,
gnatchop.adb, gnatlink.adb, g-spipat.adb, g-thread.adb, make.adb,
mdll.adb, mlib.adb, mlib-prj.adb, osint.adb, par-ch3.adb, prj.adb,
prj-makr.adb, sem_prag.adb, sem_type.adb, s-fatgen.adb, s-fileio.adb,
sinfo.ads, sinput-d.adb, s-taasde.adb, s-tasdeb.ads, s-tasren.adb,
s-tassta.adb, s-tpobop.adb, s-tposen.adb, stylesw.adb, types.ads,
uintp.adb, validsw.adb, makegpr.adb, a-rbtgso.adb, a-crbtgo.adb,
a-coorse.adb, a-convec.adb, a-coinve.adb, a-cohama.adb, a-ciorse.adb,
a-cihama.adb, a-cidlli.adb, a-chtgop.adb, a-cdlili.adb, a-cdlili.adb,
a-coormu.adb, a-ciormu.adb, a-cihase.adb, a-cohase.adb, a-ciorma.adb,
a-coorma.adb, a-ztgeau.adb, symbols-vms.adb, a-crdlli.adb,
a-calari.adb, a-calfor.adb, s-os_lib.adb, s-regpat.adb, a-ngrear.adb:
Minor reformatting.
Add Unreferenced and Warnings (Off) pragmas for cases of
variables modified calls where they are IN OUT or OUT parameters and
the resulting values are not subsequently referenced. In a few cases,
we also remove redundant code found by the new warnings.

* ug_words, vms_data.ads, usage.adb, sem_util.adb, sem_util.ads,
sem_warn.adb, sem_warn.ads, sem_res.adb, sem_ch7.adb, sem_ch8.adb,
sem_ch5.adb, opt.ads, lib-xref.adb, lib-xref.ads, exp_smem.adb,
sem_ch11.adb, exp_ch6.adb, einfo.ads, einfo.adb: implement a new
warning controlled by -gnatw.o that warns on cases of out parameter
values being ignored.

From-SVN: r129318

16 years agos-osinte-tru64.adb: (Hide_Yellow_Zone): Add On parameter.
Eric Botcazou [Mon, 15 Oct 2007 13:53:30 +0000 (15:53 +0200)]
s-osinte-tru64.adb: (Hide_Yellow_Zone): Add On parameter.

2007-10-15  Eric Botcazou  <ebotcazou@adacore.com>

* s-osinte-tru64.adb: (Hide_Yellow_Zone): Add On parameter.
Set the protection status of the guard page based on the value of On.

* s-osinte-tru64.ads: (Hide_Yellow_Zone): Add On parameter.

* s-taprop-tru64.adb: (Enter_Task): Pass True to Hide_Yellow_Zone.
(Exit_Task): Pass False to Hide_Yellow_Zone.

From-SVN: r129317

16 years agoPR libstdc++/33773 (cont)
Paolo Carlini [Mon, 15 Oct 2007 13:43:33 +0000 (13:43 +0000)]
PR libstdc++/33773 (cont)

2007-10-15  Paolo Carlini  <pcarlini@suse.de>

PR libstdc++/33773 (cont)
* testsuite/21_strings/headers/cwchar/macros.cc: Guard with
_GLIBCXX_USE_WCHAR_T, instead.

From-SVN: r129316

16 years agore PR target/33133 (ICE in try_ready, at haifa-sched.c:2958 with -O2/-O3)
Maxim Kuvyrkov [Mon, 15 Oct 2007 10:30:13 +0000 (10:30 +0000)]
re PR target/33133 (ICE in try_ready, at haifa-sched.c:2958 with -O2/-O3)

PR target/33133

* haifa-sched.c (process_insn_forw_deps_be_in_spec): Check if
speculation type of insn can be changed before trying to do that.

* gcc.c-torture/compile/pr33133.c: New test.

From-SVN: r129315

16 years agore PR libstdc++/33771 (FAIL: 17_intro/headers/c++1998/all.cc (test for excess errors))
Paolo Carlini [Mon, 15 Oct 2007 09:34:56 +0000 (09:34 +0000)]
re PR libstdc++/33771 (FAIL: 17_intro/headers/c++1998/all.cc (test for excess errors))

2007-10-15  Paolo Carlini  <pcarlini@suse.de>

PR libstdc++/33771
PR libstdc++/33773
* testsuite/21_strings/headers/cwchar/macros.cc: Guard test with
_GLIBCXX_HAVE_WCHAR_H.
* testsuite/21_strings/headers/cwctype/macros.cc: Likewise with
_GLIBCXX_HAVE_WCTYPE_H.
* testsuite/17_intro/headers/c++200x/all.cc: Guard inclusions
of <wchar.h> and <wctype.h>.
* testsuite/17_intro/headers/c++200x/all_multiple_inclusion.cc:
Likewise.
* testsuite/17_intro/headers/c++1998/all.cc: Likewise.
* testsuite/17_intro/headers/c++1998/all_multiple_inclusion.cc:
Likewise.

From-SVN: r129314

16 years agore PR libstdc++/33771 (FAIL: 17_intro/headers/c++1998/all.cc (test for excess errors))
Paolo Carlini [Mon, 15 Oct 2007 09:34:49 +0000 (09:34 +0000)]
re PR libstdc++/33771 (FAIL: 17_intro/headers/c++1998/all.cc (test for excess errors))

2007-10-15  Paolo Carlini  <pcarlini@suse.de>

PR libstdc++/33771
PR libstdc++/33773
* testsuite/21_strings/headers/cwchar/macros.cc: Guard test with
_GLIBCXX_HAVE_WCHAR_H.
* testsuite/21_strings/headers/cwctype/macros.cc: Likewise with
_GLIBCXX_HAVE_WCTYPE_H.
* testsuite/17_intro/headers/c++200x/all.cc: Guard inclusions
of <wchar.h> and <wctype.h>.
* testsuite/17_intro/headers/c++200x/all_multiple_inclusion.cc:
Likewise.
* testsuite/17_intro/headers/c++1998/all.cc: Likewise.
* testsuite/17_intro/headers/c++1998/all_multiple_inclusion.cc:
Likewise.

From-SVN: r129313

16 years agodse.c (struct insn_info): Add 'frame_read' field.
Eric Botcazou [Mon, 15 Oct 2007 07:41:28 +0000 (09:41 +0200)]
dse.c (struct insn_info): Add 'frame_read' field.

* dse.c (struct insn_info): Add 'frame_read' field.
(scan_insn): For the call to a const function, set frame_read if
reload has been run.
If the insn reads the frame, kill the frame related stores.
(scan_reads_nospill): Likewise.

From-SVN: r129312

16 years agotree-eh.c (optimize_double_finally): Don't assume that the cleanup we're duplicating...
Jason Merrill [Mon, 15 Oct 2007 05:21:37 +0000 (01:21 -0400)]
tree-eh.c (optimize_double_finally): Don't assume that the cleanup we're duplicating is only one statement.

        * tree-eh.c (optimize_double_finally): Don't assume that the
        cleanup we're duplicating is only one statement.

From-SVN: r129311

16 years agore PR libfortran/33672 (Additional runtime checks needed for namelist reads)
Jerry DeLisle [Mon, 15 Oct 2007 01:36:13 +0000 (01:36 +0000)]
re PR libfortran/33672 (Additional runtime checks needed for namelist reads)

2007-10-14  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

PR libfortran/33672
* gfortran.dg/namelist_40.f90: New test.

From-SVN: r129310

16 years agore PR libfortran/33672 (Additional runtime checks needed for namelist reads)
Jerry DeLisle [Mon, 15 Oct 2007 01:33:16 +0000 (01:33 +0000)]
re PR libfortran/33672 (Additional runtime checks needed for namelist reads)

2007-10-14  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

PR libfortran/33672
* io/list_read.c (nml_parse_qualifier): Add character specific error
messages.  Check for proper form of sub-string qualifiers.  Return the
parsed_rank flag indicating a non-zero rank qualifier.
(nml_get_obj_data):  Count the instances of non-zero rank qualifiers.
Issue an error if more that one non-zero rank qualifier is found.

From-SVN: r129309

16 years agoDaily bump.
GCC Administrator [Mon, 15 Oct 2007 00:17:21 +0000 (00:17 +0000)]
Daily bump.

From-SVN: r129307

16 years agore PR target/33700 (FAIL: 17_intro/headers/all_pedantic_errors.cc (test for excess...
John David Anglin [Sun, 14 Oct 2007 22:22:25 +0000 (22:22 +0000)]
re PR target/33700 (FAIL: 17_intro/headers/all_pedantic_errors.cc (test for excess errors))

PR target/33700
* inclhack.def (hpux11_pthread_const): New.
* fixincl.x: Regenerate.
* tests/base/sys/pthread.h: Update.

From-SVN: r129304

16 years agostl_algo.h (is_sorted, [...]): Add.
Paolo Carlini [Sun, 14 Oct 2007 21:17:23 +0000 (21:17 +0000)]
stl_algo.h (is_sorted, [...]): Add.

2007-10-14  Paolo Carlini  <pcarlini@suse.de>

* include/bits/stl_algo.h (is_sorted, is_sorted_until): Add.
* include/bits/algorithmfwd.h: Add.
* include/ext/algorithm: Adjust.
* testsuite/25_algorithms/is_sorted/requirements/
explicit_instantiation/2.cc: New.
* testsuite/25_algorithms/is_sorted/requirements/
explicit_instantiation/pod.cc: Likewise.
* testsuite/25_algorithms/is_sorted/1.cc: Likewise.
* testsuite/25_algorithms/is_sorted_until/requirements/
explicit_instantiation/2.cc: Likewise.
* testsuite/25_algorithms/is_sorted_until/requirements/
explicit_instantiation/pod.cc: Likewise.
* testsuite/25_algorithms/is_sorted_until/1.cc: Likewise.
* testsuite/25_algorithms/headers/algorithm/synopsis.cc:
Add is_sorted and is_sorted_until.

* include/bits/stl_heap.h (is_heap_until): Add concept and
debug-mode checks.

From-SVN: r129303

16 years agore PR fortran/33745 (-fbounds-check: Bogus out-of-bounds run-time error for assumed...
Tobias Burnus [Sun, 14 Oct 2007 20:24:20 +0000 (22:24 +0200)]
re PR fortran/33745 (-fbounds-check: Bogus out-of-bounds run-time error for assumed-size array)

2007-10-14  Tobias Burnus  <burnus@net-b.de>

        PR fortran/33745
        * trans-array.c (gfc_conv_ss_startstride): Fix dimension check.
        (gfc_trans_array_bound_check, gfc_conv_array_ref,
        gfc_conv_ss_startstride): Simplify error message.
        * resolve.c (check_dimension): Fix dimension-type switch;
        improve error message.

2007-10-14  Tobias Burnus  <burnus@net-b.de>

        PR fortran/33745
        * gfortran.dg/bounds_check_11.f90: New.

From-SVN: r129302

16 years agoconfigure.ac: Don't run config-ml.in directly.
H.J. Lu [Sun, 14 Oct 2007 18:17:14 +0000 (18:17 +0000)]
configure.ac: Don't run config-ml.in directly.

2007-10-14  H.J. Lu  <hongjiu.lu@intel.com>

* configure.ac: Don't run config-ml.in directly.
(multilib_arg): New.
* configure: Regenerated.

From-SVN: r129299

16 years agore PR c++/30303 (ICE with invalid constructor definition)
Andrew Pinski [Sun, 14 Oct 2007 18:15:35 +0000 (11:15 -0700)]
re PR c++/30303 (ICE with invalid constructor definition)

2007-10-14  Andrew Pinski  <pinskia@gmail.com>

        PR c++/30303
        * decl.c (grokfndecl): Return NULL after the "definition of
        implicitly-declared" error happened.

2007-10-14  Andrew Pinski  <pinskia@gmail.com>

        PR c++/30303
        * g++.dg/other/ctor1.C: New test.
        * g++.dg/other/ctor2.C: New test.
        * g++.dg/other/dtor1.C: New test.

From-SVN: r129298

16 years agoc-objc-common.h (LANG_HOOKS_TREE_INLINING_CANNOT_INLINE_TREE_FN): Remove.
Kazu Hirata [Sun, 14 Oct 2007 13:03:10 +0000 (13:03 +0000)]
c-objc-common.h (LANG_HOOKS_TREE_INLINING_CANNOT_INLINE_TREE_FN): Remove.

* c-objc-common.h (LANG_HOOKS_TREE_INLINING_CANNOT_INLINE_TREE_FN):
Remove.
* c-tree.h: Remove the prototype for c_cannot_inline_tree_fn.

From-SVN: r129296

16 years agobounds_check_10.f90: Fix testcase.
Tobias Burnus [Sun, 14 Oct 2007 10:07:07 +0000 (12:07 +0200)]
bounds_check_10.f90: Fix testcase.

2007-10-14  Tobias Burnus  <burnus@gcc.gnu.org>

        * gfortran.dg/bounds_check_10.f90: Fix testcase.

From-SVN: r129293

16 years agoc-common.h: Remove the prototype for c_expand_body.
Kazu Hirata [Sun, 14 Oct 2007 02:15:39 +0000 (02:15 +0000)]
c-common.h: Remove the prototype for c_expand_body.

* c-common.h: Remove the prototype for c_expand_body.
* c-tree.h: Remove the prototype for c_disregard_inline_limits.
* tree.h: Remove the prototype for fold_build_call_expr.

From-SVN: r129292

16 years agofixed-bit.c, [...]: Fix comment typos.
Kazu Hirata [Sun, 14 Oct 2007 01:36:18 +0000 (01:36 +0000)]
fixed-bit.c, [...]: Fix comment typos.

* config/fixed-bit.c, config/i386/cpuid.h, config/i386/i386.c,
config/i386/i386.md, config/i386/sse.md, function.c, jump.c,
modulo-sched.c, ra-conflict.c, toplev.c, tree-eh.c, tree-sra.c,
tree-ssa-dse.c, tree-vect-analyze.c, tree-vect-patterns.c,
tree-vect-transform.c: Fix comment typos.
* doc/extend.texi: Fix a typo.

From-SVN: r129291

16 years agoDaily bump.
GCC Administrator [Sun, 14 Oct 2007 00:17:15 +0000 (00:17 +0000)]
Daily bump.

From-SVN: r129289

16 years agore PR fortran/33254 (Diagnose different string lengths in array constructors at run...
Tobias Schlüter [Sat, 13 Oct 2007 21:43:49 +0000 (23:43 +0200)]
re PR fortran/33254 (Diagnose different string lengths in array constructors at run time)

2007-10-13  Tobias Schlueter  <tobi@gcc.gnu.org>
    Paul Thomas  <pault@gcc.gnu.org>

PR fortran/33254
PR fortran/33727
fortran/
* trans-array.c (get_array_ctor_var_strlen): Check upper bound for
constness instead of lower bound.
(get_array_ctor_strlen): Add bounds-checking code.
testsuite/
* bounds_check_10.f90: New.

Co-Authored-By: Paul Thomas <pault@gcc.gnu.org>
From-SVN: r129286

16 years agoaix53.h: New file.
David Edelsohn [Sat, 13 Oct 2007 21:40:11 +0000 (21:40 +0000)]
aix53.h: New file.

        * config/rs6000/aix53.h: New file.
        * config/rs6000/aix{41,43,51,52}.h (TARGET_ALTIVEC): Define to 0.
        (TARGET_ALTIVEC_ABI): Same.
        * config/rs6000/aix.h (TARGET_ALTIVEC): Delete.
        (TARGET_ALTIVEC_ABI): Delete.
        * config.gcc (powerpc-ibm-aix5*): Rename to aix5.2.  Add new
        stanza defaulting to aix5.3.

From-SVN: r129285

16 years agoparity-1.c: POWER5 feature, not POWER6.
David Edelsohn [Sat, 13 Oct 2007 15:03:07 +0000 (15:03 +0000)]
parity-1.c: POWER5 feature, not POWER6.

        * gcc.target/powerpc/parity-1.c: POWER5 feature, not POWER6.
        * gcc.target/powerpc/popcount-1.c: Same.

From-SVN: r129284

16 years agoDaily bump.
GCC Administrator [Sat, 13 Oct 2007 00:17:06 +0000 (00:17 +0000)]
Daily bump.

From-SVN: r129280

16 years agoi386.md (SI_REG, DI_REG): New constants.
Nathan Froyd [Fri, 12 Oct 2007 21:21:28 +0000 (21:21 +0000)]
i386.md (SI_REG, DI_REG): New constants.

* config/i386/i386.md (SI_REG, DI_REG): New constants.
(strmov): Use defined constants.
(cmpstrnsi): Likewise.
* config/i386/i386.c (decide_alg): Use defined constants.
(ix86_expand_strlen): Likewise.

From-SVN: r129276

16 years agore PR c++/26698 (g++ accepts const-incorrect code due to conversion function)
Simon Martin [Fri, 12 Oct 2007 18:43:33 +0000 (18:43 +0000)]
re PR c++/26698 (g++ accepts const-incorrect code due to conversion function)

gcc/cp/

2007-10-12  Simon Martin  <simartin@users.sourceforge.net>

PR c++/26698
* call.c (build_user_type_conversion_1): Do not consider conversion
functions to convert a (possibly cv-qualified) object to the (possibly
cv-qualified) same object type (or a reference to it), to a (possibly
cv-qualified) base class of that type (or a reference to it).

gcc/testsuite/

2007-10-12  Simon Martin  <simartin@users.sourceforge.net>

PR c++/26698
* g++.dg/conversion/op4.C: New test.

From-SVN: r129274

16 years agopr33572.C (main): Allow argc to be zero.
Richard Sandiford [Fri, 12 Oct 2007 17:21:26 +0000 (17:21 +0000)]
pr33572.C (main): Allow argc to be zero.

gcc/testsuite/
* g++.dg/torture/pr33572.C (main): Allow argc to be zero.

From-SVN: r129272

16 years agodse.c (find_shift_sequence): Reinstate "<= UNITS_PER_WORD" condition.
Richard Sandiford [Fri, 12 Oct 2007 16:54:38 +0000 (16:54 +0000)]
dse.c (find_shift_sequence): Reinstate "<= UNITS_PER_WORD" condition.

gcc/
* dse.c (find_shift_sequence): Reinstate "<= UNITS_PER_WORD" condition.
* var-tracking.c (micro_operation_def): Update comment on u.loc.
(mode_for_reg_attrs, var_lowpart): New functions.
(add_uses): Consider recording a lowpart of LOC for MO_USE.
(add_stores): Likewise MO_SET and MO_COPY.  If the source of a set
or copy is known, set LOC to the SET that performs the set, instead
of the destination.
(find_src_status, find_src_set_src): Remove LOC parameter.
Replace INSN with the source value.
(compute_bb_dataflow, emit_notes_in_bb): Check for a SET u.loc when
handling MO_SET and MO_COPY.  Update the calls to find_src_status
and find_src_set_src.

From-SVN: r129269

16 years agore PR fortran/33542 (gfortran does not detect ambigious specific names if they are...
Paul Thomas [Fri, 12 Oct 2007 16:51:53 +0000 (16:51 +0000)]
re PR fortran/33542 (gfortran does not detect ambigious specific names if they are the same as generic names)

2007-10-12  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/33542
* resolve.c (resolve_actual_arglist): If the actual argument is
ambiguous, then there is an error.

2007-10-12  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/33542
* gfortran.dg/ambiguous_specific_1.f90: New test.

From-SVN: r129268

16 years agore PR fortran/33664 (crash on invalid program)
Paul Thomas [Fri, 12 Oct 2007 16:45:46 +0000 (16:45 +0000)]
re PR fortran/33664 (crash on invalid program)

2007-10-12  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/33664
* expr.c (gfc_specification_expr): If a function is not
external, intrinsic or pure is an error.  Set the symbol pure
to prevent repeat errors.

2007-10-12  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/33664
* gfortran.dg/impure_spec_expr_1.f90: New test.
* gfortran.dg/char_result_7.f90: Remove illegal test.

From-SVN: r129267

16 years agostl_heap.h (__is_heap_until): Add.
Paolo Carlini [Fri, 12 Oct 2007 16:26:03 +0000 (16:26 +0000)]
stl_heap.h (__is_heap_until): Add.

2007-10-12  Paolo Carlini  <pcarlini@suse.de>

* include/bits/stl_heap.h (__is_heap_until): Add.
(__is_heap(_RandomAccessIterator, _Distance),
__is_heap(_RandomAccessIterator, _Compare, _Distance)):
Adjust, call the latter.
(is_heap, is_heap_until): Add, call the above.
* include/bits/algorithmfwd.h: Add.
* testsuite/25_algorithms/is_heap/requirements/
explicit_instantiation/2.cc: New.
* testsuite/25_algorithms/is_heap/requirements/
explicit_instantiation/pod.cc: Likewise.
* testsuite/25_algorithms/is_heap/1.cc: Likewise.
* testsuite/25_algorithms/is_heap_until/requirements/
explicit_instantiation/2.cc: Likewise.
* testsuite/25_algorithms/is_heap_until/requirements/
explicit_instantiation/pod.cc: Likewise.
* testsuite/25_algorithms/is_heap_until/1.cc: Likewise.
* testsuite/25_algorithms/headers/algorithm/synopsis.cc:
Add is_heap and is_heap_until.

From-SVN: r129266

16 years agore PR rtl-optimization/11001 (global register %edi versus string builtins)
Nathan Froyd [Fri, 12 Oct 2007 16:12:45 +0000 (16:12 +0000)]
re PR rtl-optimization/11001 (global register %edi versus string builtins)

gcc/
PR 11001
* config/i386/i386.md (strmov): Check for esi and edi usage.
* config/i386/i386.c (decide_alg): Check whether we can use a
rep prefix and adjust algorithm choice accordingly.
(ix86_expand_strlen): Check for eax, ecx, and edi usage.

gcc/testsuite/
PR 11001
* gcc.target/i386/pr11001-strlen-1.c: New testcase.
* gcc.target/i386/pr11001-strlen-2.c: New testcase.
* gcc.target/i386/pr11001-strlen-3.c: New testcase.
* gcc.target/i386/pr11001-memset-1.c: New testcase.
* gcc.target/i386/pr11001-memset-2.c: New testcase.
* gcc.target/i386/pr11001-memset-3.c: New testcase.
* gcc.target/i386/pr11001-memcpy-1.c: New testcase.
* gcc.target/i386/pr11001-memcpy-2.c: New testcase.
* gcc.target/i386/pr11001-memcpy-3.c: New testcase.

From-SVN: r129265

16 years agouser.cfg.in: Scan tr1_impl/hashtable.
Benjamin Kosnik [Fri, 12 Oct 2007 16:11:41 +0000 (16:11 +0000)]
user.cfg.in: Scan tr1_impl/hashtable.

2007-10-12  Benjamin Kosnik  <bkoz@redhat.com>

* docs/doxygen/user.cfg.in: Scan tr1_impl/hashtable.
* include/tr1_impl/hashtable: Correct comment.

* include/ext/hash_map: Remove extraneous public markers.
* include/ext/hash_set: Same.

* testsuite/25_algorithms/headers/parallel_algorithm.cc: Move
inside algorithm directory.
* testsuite/25_algorithms/headers/algorithm_parallel_mode.cc: Same.
* testsuite/25_algorithms/headers/parallel_algorithm_mixed1.cc: Same.
* testsuite/25_algorithms/headers/parallel_algorithm_mixed2.cc: Same.
* testsuite/25_algorithms/headers/algorithm/parallel_algorithm.cc:
To this.
* testsuite/25_algorithms/headers/algorithm/
algorithm_parallel_mode.cc: Same.
* testsuite/25_algorithms/headers/algorithm/
parallel_algorithm_mixed1.cc: Same.
* testsuite/25_algorithms/headers/algorithm/
parallel_algorithm_mixed2.cc: Same.

* testsuite/21_strings/basic_string/requirements/
explicit_instantiation/debug.cc: Remove exception text.

* testsuite/23_containers/vector/bool/modifiers/insert/31370.cc:
Fix for debug mode testing.
* testsuite/23_containers/vector/bool/capacity/29134.cc: Same.

From-SVN: r129264

16 years agoc4x.md (mulqi3, mulhi3): Use optab_libfunc.
Kaveh R. Ghazi [Fri, 12 Oct 2007 15:24:35 +0000 (15:24 +0000)]
c4x.md (mulqi3, mulhi3): Use optab_libfunc.

* config/c4x/c4x.md (mulqi3, mulhi3): Use optab_libfunc.

From-SVN: r129263

16 years agoglobal.c (build_insn_chain): Formatting fixes.
Kenneth Zadeck [Fri, 12 Oct 2007 14:36:47 +0000 (14:36 +0000)]
global.c (build_insn_chain): Formatting fixes.

2007-10-12  Kenneth Zadeck <zadeck@naturalbridge.com>

* global.c (build_insn_chain): Formatting fixes.

From-SVN: r129262

16 years agotree-ssa-forwprop.c (forward_propagate_into_cond): For combining both operands requir...
Richard Guenther [Fri, 12 Oct 2007 14:21:43 +0000 (14:21 +0000)]
tree-ssa-forwprop.c (forward_propagate_into_cond): For combining both operands require either both have single uses or combining...

2007-10-12  Richard Guenther <rguenther@suse.de>

* tree-ssa-forwprop.c (forward_propagate_into_cond): For
combining both operands require either both have single uses
or combining to a constant.

From-SVN: r129261

16 years ago[patch]: Fix native bootstrap failure for x86_64-pc-mingw32
Kai Tietz [Fri, 12 Oct 2007 11:54:16 +0000 (11:54 +0000)]
[patch]: Fix native bootstrap failure for x86_64-pc-mingw32

* configure.ac: x86_64 mingw target uses twoprocess.
* mkfixinc.sh: Add x86_64 mingw target.
* configure: Regenerate.

From-SVN: r129260

16 years ago* ChangeLog: Fix typo.
Uros Bizjak [Fri, 12 Oct 2007 10:25:43 +0000 (12:25 +0200)]
* ChangeLog: Fix typo.

From-SVN: r129259

16 years ago20_riemann_zeta/check_value_neg.c (MAX_ITERATIONS): Add limit when target is simulator.
Jesper Nilsson [Fri, 12 Oct 2007 10:23:55 +0000 (10:23 +0000)]
20_riemann_zeta/check_value_neg.c (MAX_ITERATIONS): Add limit when target is simulator.

2007-10-12  Jesper Nilsson  <jesper.nilsson@axis.com>

       * testsuite/tr1/5_numerical_facilities/special_functions/
       20_riemann_zeta/check_value_neg.c (MAX_ITERATIONS): Add limit
       when target is simulator.
       * testsuite/tr1/5_numerical_facilities/special_functions/
       20_riemann_zeta/check_value_pos.c (MAX_ITERATIONS): Likewise.

From-SVN: r129258

16 years agopt.c (tsubst): Use template_parm_level_and_index.
Paolo Carlini [Fri, 12 Oct 2007 09:53:27 +0000 (09:53 +0000)]
pt.c (tsubst): Use template_parm_level_and_index.

2007-10-12  Paolo Carlini  <pcarlini@suse.de>

* pt.c (tsubst): Use template_parm_level_and_index.

From-SVN: r129257

16 years agore PR middle-end/26198 (Unfolded comparison after cfg_cleanup)
Richard Guenther [Fri, 12 Oct 2007 08:42:13 +0000 (08:42 +0000)]
re PR middle-end/26198 (Unfolded comparison after cfg_cleanup)

2007-10-12  Richard Guenther  <rguenther@suse.de>

PR middle-end/26198
* tree-ssa-forwprop.c (can_propagate_from): Do not propagate from
a rhs with side-effects or which is a load.
(forward_propagate_into_cond): Also try combining both operands.

* gcc.dg/tree-ssa/forwprop-3.c: New testcase.
* gcc.c-torture/execute/20071011-1.c: Likewise.
* gcc.dg/tree-ssa/ssa-pre-9.c: Adjust.

From-SVN: r129256

16 years agore PR tree-optimization/33742 (Segfault in vectorizable_operation)
Uros Bizjak [Fri, 12 Oct 2007 08:37:17 +0000 (10:37 +0200)]
re PR tree-optimization/33742 (Segfault in vectorizable_operation)

PR tree-optimization/33742
* tree-vect-transform.c (vectorizable_operation): Return false
if get_vectype_for_scalar_type for scalar_dest can't be determined.
(vectorizable_call): Same for rhs_type and lhs_type.

testsuite/ChangeLog:

PR tree-optimization/33742
* gcc.dg/pr33742.c: New testcase.

From-SVN: r129255

16 years agore PR tree-optimization/33645 (undefined static variable in vortex for -fno-unit...
Jakub Jelinek [Fri, 12 Oct 2007 07:10:22 +0000 (09:10 +0200)]
re PR tree-optimization/33645 (undefined static variable in vortex for -fno-unit-at-a-time)

PR tree-optimization/33645
* tree-ssa-live.c (mark_all_vars_used): Add data argument,
pass it to walk_tree.
(mark_all_vars_used_1): Pass data through to mark_all_vars_used.
When calling set_is_used on a VAR_DECL, if data is not NULL and
its DECL_UID is in the bitmap, call mark_all_vars_used on its
DECL_INITIAL after clearing the bit in bitmap.
(remove_unused_locals): Adjust mark_all_vars_used callers.
Instead of removing unused global vars from unexpanded_var_list
immediately record them in bitmap, call mark_all_vars_used on
all used global vars from unexpanded_var_list and only purge
global vars that weren't found used even during that step.

* gcc.dg/pr33645-1.c: New test.
* gcc.dg/pr33645-2.c: New test.
* gcc.dg/pr33645-3.c: New test.

From-SVN: r129254

16 years agore PR c++/32121 (C++ front-end accepts invalid __label__ declarations)
Jakub Jelinek [Fri, 12 Oct 2007 07:07:46 +0000 (09:07 +0200)]
re PR c++/32121 (C++ front-end accepts invalid __label__ declarations)

PR c++/32121
* parser.c (cp_parser_compound_statement): Handle label-declarations
at the beginning of the compound statement.
(cp_parser_block_declaration): Issue diagnostics about __label__
not at the beginning of a block.

* g++.dg/ext/label4.C: Adjust error regexp.
* g++.dg/ext/label6.C: Adjust error regexp.
* g++.dg/ext/label7.C: New test.
* g++.dg/ext/label8.C: New test.
* g++.dg/ext/label9.C: New test.

From-SVN: r129253

16 years agopr33676.c: Move from gcc.dg/torture directory.
Uros Bizjak [Fri, 12 Oct 2007 05:56:56 +0000 (07:56 +0200)]
pr33676.c: Move from gcc.dg/torture directory.

* gcc.dg/pr33676.c: Move from gcc.dg/torture directory.

From-SVN: r129252

16 years agoMakefile.def (dependencies): Add all-gdb -> all-libdecnumber.
Daniel Jacobowitz [Fri, 12 Oct 2007 02:23:19 +0000 (02:23 +0000)]
Makefile.def (dependencies): Add all-gdb -> all-libdecnumber.

* Makefile.def (dependencies): Add all-gdb -> all-libdecnumber.
* Makefile.in: Regenerate.

From-SVN: r129251

16 years agoDaily bump.
GCC Administrator [Fri, 12 Oct 2007 00:17:18 +0000 (00:17 +0000)]
Daily bump.

From-SVN: r129249

16 years agogthr-posix.h (__gthread_active_init): Create detached instead of joinable thread...
John David Anglin [Fri, 12 Oct 2007 00:09:10 +0000 (00:09 +0000)]
gthr-posix.h (__gthread_active_init): Create detached instead of joinable thread when...

* gthr-posix.h (__gthread_active_init): Create detached instead of
joinable thread when testing whether threads are active on hppa-hpux.
* gthr-posix95.h (__gthread_active_init): Likewise.

From-SVN: r129246

16 years agoacinclude.m4 (GLIBCXX_CHECK_SYSTEM_ERROR): Add EOVERFLOW for Tru64.
Roger Sayle [Thu, 11 Oct 2007 23:22:22 +0000 (23:22 +0000)]
acinclude.m4 (GLIBCXX_CHECK_SYSTEM_ERROR): Add EOVERFLOW for Tru64.

* acinclude.m4 (GLIBCXX_CHECK_SYSTEM_ERROR): Add EOVERFLOW for Tru64.
* configure: Regenerate.
* config/os/generic/error_constants.h: Guard.
* testsuite/19_diagnostics/headers/system_error/
types_std_c++0x.cc: Same.

From-SVN: r129245

16 years agore PR rtl-optimization/33676 (libgfortran bootstrap failure: selected_int_kind.f90...
Kenneth Zadeck [Thu, 11 Oct 2007 22:31:55 +0000 (22:31 +0000)]
re PR rtl-optimization/33676 (libgfortran bootstrap failure: selected_int_kind.f90:22: Segmentation fault, wrong code with -fomit-frame-pointer)

2007-10-11  Kenneth Zadeck <zadeck@naturalbridge.com>

PR middle-end/33676
* global.c (build_insn_chain): Include insn that occur between
basic blocks.

2007-10-11  Kenneth Zadeck <zadeck@naturalbridge.com>

PR middle-end/33676
* gcc.dg/torture/pr33676.c: New.

From-SVN: r129244

16 years agoConvert ext/hash_map, set to tr1/functional, tr1/unordered_map, set.
Benjamin Kosnik [Thu, 11 Oct 2007 21:48:54 +0000 (21:48 +0000)]
Convert ext/hash_map, set to tr1/functional, tr1/unordered_map, set.

2007-10-11  Benjamin Kosnik  <bkoz@redhat.com>

Convert ext/hash_map, set to tr1/functional, tr1/unordered_map, set.
* include/ext/rope: Use tr1/functional hash instead of ext/hash_fun.h.
* include/ext/pb_ds/detail/standard_policies.hpp: Same.
* include/precompiled/extc++.h: Remove hash_map, hash_set.
* testsuite/util/testsuite_abi.h (__gnu_cxx): Use unordered_map
instead of hash_map. Remove hash specialization for string.
* testsuite/util/testsuite_abi.cc: Include <vector>.
* testsuite/util/native_type/assoc/native_hash_set.hpp: Remove
hash_set code.
* testsuite/util/native_type/assoc/native_hash_map.hpp: Remove
hash_map code.
* testsuite/util/native_type/assoc/native_hash_multimap.hpp:
Change hash_multimap to unordered_multimap.
* testsuite/util/testsuite_common_types.h: Remove hash_map, hash_set.
* testsuite/ext/headers.cc: Same.
* testsuite/17_intro/headers/all_c++200x_compatibility.cc: Remove
hash_map.h, hash_set.h.
* testsuite/17_intro/headers/all.cc: Same.
* testsuite/17_intro/headers/all_pedantic_errors.cc: Same.

* include/ext/pb_ds/detail/standard_policies.hpp: Format.
* testsuite/util/native_type/assoc/native_set.hpp: Same,
update comment.
* testsuite/util/native_type/assoc/native_map.hpp: Same.
* testsuite/util/native_type/assoc/native_multimap.hpp: Same.
* testsuite/util/performance/assoc/timing/
multimap_insert_test.hpp: Same.
* testsuite/util/performance/assoc/timing/multimap_find_test.hpp: Same.

From-SVN: r129243

16 years ago* gengtype-yacc.y: Delete.
Tom Tromey [Thu, 11 Oct 2007 20:40:41 +0000 (20:40 +0000)]
* gengtype-yacc.y: Delete.

From-SVN: r129240

16 years agostl_heap.h (__push_heap, [...]): Use _GLIBCXX_MOVE.
Chris Jefferson [Thu, 11 Oct 2007 17:56:31 +0000 (17:56 +0000)]
stl_heap.h (__push_heap, [...]): Use _GLIBCXX_MOVE.

2007-10-11  Chris Jefferson  <chris@bubblescope.net>
    Paolo Carlini  <pcarlini@suse.de>

* include/bits/stl_heap.h (__push_heap, push_heap, __adjust_heap,
make_heap): Use _GLIBCXX_MOVE.
(__pop_heap): Likewise, adjust signature.
(pop_heap): Adjust __pop_heap call.
* include/bits/stl_algo.h (__heap_select): Likewise.
* testsuite/25_algorithms/heap/moveable.cc: Remove dg-require-rvalref.
* testsuite/25_algorithms/partial_sort/moveable.cc: Likewise.

Co-Authored-By: Paolo Carlini <pcarlini@suse.de>
From-SVN: r129237

16 years agoi386.md (AX_REG, [...]): New constants.
Uros Bizjak [Thu, 11 Oct 2007 17:51:01 +0000 (19:51 +0200)]
i386.md (AX_REG, [...]): New constants.

        * config/i386/i386.md (AX_REG, BX_REG, CX_REG): New constants.
        * config/i386/i386.c (ix86_function_arg_regno_p, function_arg_32,
        function_value_32, function_value_64, function_value_ms_64,
        setup_incoming_varargs_64, ix86_expand_prologue, ix86_expand_call,
        legitimize_tls_address, x86_this_parameter, x86_output_mi_thunk):
        Use new constants.

From-SVN: r129236

16 years agoFixed omission in changelog.
Kenneth Zadeck [Thu, 11 Oct 2007 13:14:18 +0000 (13:14 +0000)]
Fixed omission in changelog.

From-SVN: r129234

16 years agore PR libstdc++/33734 (iconv_t must not be assumed to be a pointer)
Paolo Carlini [Thu, 11 Oct 2007 10:50:09 +0000 (10:50 +0000)]
re PR libstdc++/33734 (iconv_t must not be assumed to be a pointer)

2007-10-11  Paolo Carlini  <pcarlini@suse.de>

PR libstdc++/33734
* include/ext/codecvt_specializations.h (encoding_state::good,
init, destroy): Use cast notation instead of reinterpret_cast.

From-SVN: r129232

16 years agomoveable.cc: Fix and extend.
Paolo Carlini [Thu, 11 Oct 2007 10:04:55 +0000 (10:04 +0000)]
moveable.cc: Fix and extend.

2007-10-11  Paolo Carlini  <pcarlini@suse.de>

* testsuite/25_algorithms/heap/moveable.cc: Fix and extend.

From-SVN: r129231

16 years agogcse.c (hash_scan_set): If the next nonnote insn is in a different basic block insert...
Richard Guenther [Thu, 11 Oct 2007 09:46:04 +0000 (09:46 +0000)]
gcse.c (hash_scan_set): If the next nonnote insn is in a different basic block insert the set.

2007-10-11  Richard Guenther  <rguenther@suse.de>

* gcse.c (hash_scan_set): If the next nonnote insn is in
a different basic block insert the set.

From-SVN: r129230

16 years agore PR c++/31441 (ICE with variadic template and specialization)
Paolo Carlini [Thu, 11 Oct 2007 09:39:41 +0000 (09:39 +0000)]
re PR c++/31441 (ICE with variadic template and specialization)

/testsuite
2007-10-11  Paolo Carlini  <pcarlini@suse.de>

PR c++/31441
* g++.dg/cpp0x/variadic83.C: New.

/cp
2007-10-11  Paolo Carlini  <pcarlini@suse.de>

PR c++/33461
* pt.c (coerce_template_parameter_pack): Do not pass error_mark_node
to convert_template_argument.
(coerce_template_parms): Return error_mark_node after fixed-length
error.
(tsubst_decl): Check for error_mark_node the return value of the
first tsubst in 'case VAR_DECL'.

/testsuite
2007-10-11  Paolo Carlini  <pcarlini@suse.de>

PR c++/33461
* g++.dg/cpp0x/variadic81.C: New.
* g++.dg/cpp0x/variadic82.C: Likewise.

From-SVN: r129229

16 years agore PR middle-end/33724 (Type checking error with address-of and ref-all pointer type)
Richard Guenther [Thu, 11 Oct 2007 08:58:28 +0000 (08:58 +0000)]
re PR middle-end/33724 (Type checking error with address-of and ref-all pointer type)

2007-10-11  Richard Guenther  <rguenther@suse.de>

PR middle-end/33724
* tree-cfg.c (one_pointer_to_useless_type_conversion_p): New function.
(verify_gimple_expr): Use it to verify pointer-to types for
ADDR_EXPRs.

* gcc.dg/pr33724.c: New testcase.

From-SVN: r129228

16 years agore PR c/33726 (Type checking error with address-of and volatile and arrays)
Richard Guenther [Thu, 11 Oct 2007 08:11:18 +0000 (08:11 +0000)]
re PR c/33726 (Type checking error with address-of and volatile and arrays)

2007-10-11  Richard Guenther  <rguenther@suse.de>

PR c/33726
* c-typeck.c (build_array_ref): Do not strip qualifiers from
the array element type.

* gcc.dg/pr33726.c: New testcase.

From-SVN: r129227