Jakub Jelinek [Mon, 29 Dec 2008 18:25:19 +0000 (19:25 +0100)]
re PR middle-end/36191 (can't use complex in a eh region if non-call-exceptions is enabled)
PR c++/36191
* tree-complex.c (expand_complex_libcall): Call
maybe_clean_or_replace_eh_stmt and gimple_purge_dead_eh_edges
instead of passing true as 3rd argument to gsi_replace.
* gcc.c-torture/execute/builtins/fprintf.x: Disable test for
freestanding targets.
* gcc.c-torture/execute/pr37573.x: Disable test for 16-bit targets.
* gcc.c-torture/execute/20081117-1.x: Disable test for 16-bit targets.
* gcc.c-torture/compile/limits-stringlit.c: Disable test for picochip.
* gcc.dg/array-quals-1.c: Disable test for picochip.
* lib/target-supports.exp: Add support for freestanding ports and
disable picochip for profiling tests.
Jakub Jelinek [Mon, 29 Dec 2008 08:32:21 +0000 (09:32 +0100)]
re PR driver/38381 (-b option seems to be broken)
PR driver/38381
* gcc.c (process_command): Accept also -b with configuration name
in the next argument.
* doc/invoke.texi (-b): Document that no hyphen is required if
configuration name is in the next argument after -b.
Jakub Jelinek [Mon, 22 Dec 2008 23:41:17 +0000 (00:41 +0100)]
re PR middle-end/31150 (Not promoting an whole array to be static const)
PR middle-end/31150
* dse.c (struct store_info): Add const_rhs field.
(clear_rhs_from_active_local_stores): Clear also const_rhs.
(record_store): Try also cselib_expand_value_rtx to get a constant.
(find_shift_sequence, get_stored_val): Use const_rhs instead of
rhs if worthwhile.
* cselib.c (cselib_record_sets): If !cselib_record_memory and
there is just one set from read-only MEM, look at REG_EQUAL or
REG_EQUIV note.
* dse.c (struct store_info): Add redundant_reason field.
(record_store): When storing the same constant as has been
stored by an earlier store, set redundant_reason field
to the earlier store's insn_info_t. Don't delete cannot_delete
insns.
(find_shift_sequence): Remove read_info argument, add read_mode
and require_cst arguments. Return early if require_cst and
constant wouldn't be returned.
(get_stored_val): New function.
(replace_read): Use it.
(scan_insn): Put even cannot_delete insns with exactly 1 store
into active_local_stores.
(dse_step1): Don't delete cannot_delete insns. Remove redundant
constant stores if contains_cselib_groups and earlier store storing
the same value hasn't been eliminated.
(dse_step6): Renamed to dse_step7. New function.
(dse_step7): Renamed from dse_step6.
(rest_of_handle_dse): Call dse_step6 and dse_step7 at the end.
* cselib.c (cselib_expand_value_rtx): Don't wrap CONST_INTs
into CONST unless really necessary. Handle SUBREG, unary,
ternary, bitfield and compares specially, to be able to simplify
operations on constants.
(expand_loc): Try to optimize LO_SUM.
* dse.c (get_call_args): New function.
(scan_insn): Don't handle BUILT_IN_BZERO. For memset, attempt
to get call arguments and if successful and both len and val are
constants, handle the call as (mem:BLK) (const_int) store.
* dse.c (struct store_info): Add is_large bool field, change
positions_needed into a union of a bitmask and bitmap + count.
(free_store_info): Free bitmap if is_large.
(set_usage_bits): Don't look at stores where
offset + width >= MAX_OFFSET.
(set_position_unneeded, set_all_positions_unneeded,
any_positions_needed_p, all_positions_needed_p): New static inline
functions.
(record_store): Handle BLKmode stores of CONST_INT, if
MEM_SIZE is set on the MEM. Use the new positions_needed
accessor inlines.
(replace_read): Handle reads from BLKmode CONST_INT stores.
(check_mem_read_rtx): Use all_positions_needed_p function.
(dse_step1): Free large positions_needed bitmaps and clear is_large.
* dse.c (struct store_info): Change begin and end types to
HOST_WIDE_INT.
* dse.c (record_store): Fix check for unused store.
* expr.c (block_clear_fn): No longer static.
* expr.h (block_clear_fn): Declare.
* dse.c (scan_insn): Memset and bzero can just read their
arguments.
Jakub Jelinek [Mon, 22 Dec 2008 23:34:07 +0000 (00:34 +0100)]
re PR target/38488 (x86_64 generates much larger and slightly slower code for memset)
* config/i386/i386.c (expand_setmem_via_rep_stos): Add ORIG_VALUE
argument. If ORIG_VALUE is const0_rtx and COUNT is constant,
set MEM_SIZE on DESTMEM.
(ix86_expand_setmem): Adjust callers.
PR target/38488
* expr.h (get_mem_align_offset): New prototype.
* emit-rtl.c (get_mem_align_offset): New function.
* config/i386/i386.c (expand_movmem_via_rep_mov): Set MEM_SIZE correctly.
(expand_constant_movmem_prologue, expand_constant_setmem_prologue):
New functions.
(ix86_expand_movmem): Optimize if COUNT_EXP
is constant, desired_align > align and dst & (desired_align - 1)
is computable at compile time.
(ix86_expand_setmem): Likewise.
* builtins.c (get_memory_rtx): Try to derive MEM_ATTRS from not yet
resolved SAVE_EXPR or POINTER_PLUS_EXPR.
* gcc.target/mips/atomic-memory-2.c: Use dg-options instead of
dg-mips-options. Use isa>=2 instead of -mips32.
* gcc.target/mips/branch-cost-1.c: Use dg-options instead of
dg-mips-options. Use isa>=4 instead of -mips64.
* gcc.target/mips/branch-cost-2.c: Likewise.
* gcc.target/mips/cache-1.c: Use dg-options instead of
dg-mips-options. Add isa>=3 and NOMIPS16 attributes.
* gcc.target/mips/call-saved-1.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option. Use isa_rev=0 instead of -mips2.
* gcc.target/mips/call-saved-2.c: Likewise.
* gcc.target/mips/call-saved-3.c: Likewise.
* gcc.target/mips/clear-cache-1.c: Use dg-options instead of
dg-mips-options. Use isa_rev>=2 instead of -mips32r2.
* gcc.target/mips/ext_ins.c: Likewise.
* gcc.target/mips/code-readable-1.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option. Use addressing=absolute too.
* gcc.target/mips/code-readable-2.c: Likewise.
* gcc.target/mips/code-readable-3.c: Likewise.
* gcc.target/mips/dmult-1.c: Use dg-options instead of
dg-mips-options. Remove the mips16_attribute directives.
* gcc.target/mips/gcc-have-sync-compare-and-swap-4.c: Likewise.
* gcc.target/mips/dpaq_sa_l_w.c: Use dg-options instead of
dg-mips-options. Remove target restrictions from dg-do.
Use -mgp32 instead of -mips32r2.
* gcc.target/mips/dpsq_sa_l_w.c: Likewise.
* gcc.target/mips/dsp-ctrl.c: Use dg-options instead of
dg-mips-options. Remove target restrictions from dg-do.
Remove the !__mips_dsp code and add -mdsp -mgp32 to dg-options
instead. Add NOMIPS16 attributes.
* gcc.target/mips/dspr2-MULT.c: Use dg-options instead of
dg-mips-options. Replace -march=mips32r2 with -mgp32.
* gcc.target/mips/dspr2-MULTU.c: Likewise.
* gcc.target/mips/ext-1.c: Use dg-options instead of
dg-mips-options. Use isa_rev>=2 instead of -mips64r2.
* gcc.target/mips/fixed-scalar-type.c: Use dg-options instead
of dg-mips-options. Remove target restrictions from dg-do.
Remove -march=mips32r2.
* gcc.target/mips/fixed-vector-type.c: Likewise.
* gcc.target/mips/mips32-dsp-run.c: Likewise.
* gcc.target/mips/mips32-dspr2.c: Likewise. Add NOMIPS16 attributes.
* gcc.target/mips/fpr-moves-7.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option. Remove -msoft-float.
* gcc.target/mips/fpr-moves-8.c: Likewise.
* gcc.target/mips/int-moves-1.c: Likewise.
* gcc.target/mips/int-moves-2.c: Likewise.
* gcc.target/mips/gcc-have-sync-compare-and-swap-1.c: Use dg-options
instead of dg-mips-options. Use isa>=2 instead of -mips2.
Add -mgp32.
* gcc.target/mips/gcc-have-sync-compare-and-swap-3.c: Likewise.
* gcc.target/mips/ins-1.c: Use dg-options instead of
dg-mips-options. Use "isa_rev>=2 -mgp32" instead of -march=mips32r2.
* gcc.target/mips/loongson-muldiv-1.c: Use dg-options instead of
dg-mips-options. Use isa=loongson instead of -march=loongson2e.
* gcc.target/mips/loongson-muldiv-2.c: Likewise.
* gcc.target/mips/loongson-simd.c: Remove mips_loongson
target requirement and use isa=loongson instead. Add -mhard-float,
-mno-mips16 and -flax-vector-conversions.
* gcc.target/mips/lazy-binding-1.c: Use dg-options instead of
dg-mips-options. Remove target restrictions from dg-do.
Add NOMIPS16 attributes.
* gcc.target/mips/long-calls-pg.c: Use dg-options instead of
dg-mips-options. Remove -march=mips32 and -fno-pic.
Add NOMIPS16 attributes.
* gcc.target/mips/madd-3.c: Use dg-options instead of
dg-mips-options. Use isa_rev>=1 instead of -mips32.
* gcc.target/mips/maddu-3.c: Likewise.
* gcc.target/mips/msub-3.c: Likewise.
* gcc.target/mips/msubu-3.c: Likewise.
* gcc.target/mips/madd-4.c: Use dg-options instead of
dg-mips-options. Remove -mips32r2.
* gcc.target/mips/maddu-4.c: Likewise.
* gcc.target/mips/msub-4.c: Likewise.
* gcc.target/mips/msubu-4.c: Likewise.
* gcc.target/mips/mips-ps-5.c: Use dg-options instead of
dg-mips-options. Remove -mips64.
* gcc.target/mips/mips-ps-type.c: Likewise.
* gcc.target/mips/mips-ps-7.c: Use dg-options instead of
dg-mips-options. Replace -mips32r2 with -mgp32.
* gcc.target/mips/mips-ps-type-2.c: Use dg-options instead of
dg-mips-options. Use "isa_rev>=2 -mgp32" instead of -mips32r2.
* gcc.target/mips/mips16-attributes.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option.
* gcc.target/mips/mips16-attributes-2.c: Likewise.
* gcc.target/mips/r10k-cache-barrier-14.c: Likewise.
* gcc.target/mips/scc-3.c: Likewise.
* gcc.target/mips/mips16e-extends.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option. Use isa_rev>=1 instead of -march=mips32.
Add -mlong32.
* gcc.target/mips/mips32-dsp.c: Use dg-options instead of
dg-mips-options. Replace -march=mips32 with -mgp32.
* gcc.target/mips/mips32-dsp-type.c: Use dg-options instead of
dg-mips-options. Remove -march=mips32.
* gcc.target/mips/mips32-dspr2-type.c: Use dg-options instead of
dg-mips-options. Remove -march=mips32r2.
* gcc.target/mips/mips32r2-mxhc1.c: Use dg-options instead of
dg-mips-options. Add NOMIPS16 attributes.
* gcc.target/mips/movcc-1.c: Use dg-options instead of
dg-mips-options. Use isa>=4 instead of -mips4.
* gcc.target/mips/movcc-2.c: Likewise.
* gcc.target/mips/movcc-3.c: Likewise.
* gcc.target/mips/octeon-bbit-1.c: Use dg-options instead of
dg-mips-options. Declare foo as a NOMIPS16 function.
Use foo instead of g.
* gcc.target/mips/near-far-1.c: Use dg-options instead of
dg-mips-options. Replace the nonpic directives with an
addressing=absolute option.
* gcc.target/mips/near-far-2.c: Likewise.
* gcc.target/mips/near-far-3.c: Likewise.
* gcc.target/mips/near-far-4.c: Likewise.
* gcc.target/mips/nmadd-1.c: Use dg-options instead of
dg-mips-options. Use isa=4 instead of -mips4.
* gcc.target/mips/nmadd-2.c: Likewise.
* gcc.target/mips/nmadd-3.c: Likewise.
* gcc.target/mips/rsqrt-1.c: Likewise.
* gcc.target/mips/rsqrt-2.c: Likewise.
* gcc.target/mips/rsqrt-3.c: Likewise.
* gcc.target/mips/save-restore-1.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option. Use isa_rev>=1 instead of -mips32r2.
* gcc.target/mips/save-restore-2.c: Likewise.
* gcc.target/mips/save-restore-3.c: Likewise.
* gcc.target/mips/save-restore-4.c: Likewise.
* gcc.target/mips/save-restore-5.c: Likewise.
* gcc.target/mips/scc-1.c: Use dg-options instead of
dg-mips-options. Replace the mips16_attribute directives with
a (-mips16) dg-option. Use isa_rev>=1 instead of -mips32.
* gcc.target/mips/timode-2.c: Remove target restrictions from dg-do.
Use -mgp64.
mips-dsp.md (mips_lbux): Turn into a define_expand, changing operand 1 to a pmode_register_operand.
gcc/
* config/mips/mips-dsp.md (mips_lbux): Turn into a define_expand,
changing operand 1 to a pmode_register_operand.
(mips_lhx, mips_lwx): Likewise.
(mips_lbux_<mode>, mips_lhx_<mode>, mips_lwx_<mode>): New patterns.
* config/mips/mips.c (mips_prepare_builtin_arg): Get the mode of
the value from the argument expression.
Jerry DeLisle [Sun, 21 Dec 2008 21:23:52 +0000 (21:23 +0000)]
re PR fortran/38398 (g0.w edit descriptor: Update for F2008 Tokyo meeting changes)
2008-12-21 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libfortran/38398
* io/io.h (st_parameter_dt): Add new bit to keep track of when to
suppress blanks for g0 formatting.
* io/transfer.c (formatted_transfer_scalar): Always call write_real_g0
for g0 formatting.
* io.c (write.c): Do not use ES formatting and use new bit to suppress
blanks.
* io/write_float.def (output_float): Adjust the location of setting the
width so that it can be adjusted when suppressing blanks. Set number of
blanks to zero when dtp->u.p.g0_no_blanks is set. Do some minor code
clean-up and add some white space for readability.
Arjen Markus [Sun, 21 Dec 2008 18:45:17 +0000 (19:45 +0100)]
re PR fortran/37605 (Remarks on user manual for Gfortran)
2008-12-21 Arjen Markus <arjen.markus@wldelft.nl>
Daniel Kraft <d@domob.eu>
PR fortran/37605
* gfortran.texi: Fixed some typos and some minor style improvements.
* intrinsic.texi: Some clarifications and typo-fixes.
* invoke.texi: Better documenation of the behaviour of the
-fdefault-*-8 options and some other fixes.
Co-Authored-By: Daniel Kraft <d@domob.eu>
From-SVN: r142866
Paolo Carlini [Sun, 21 Dec 2008 15:56:22 +0000 (15:56 +0000)]
re PR libstdc++/38596 (tr1_impl/functional incompatible with -fno-rtti)
2008-12-21 Paolo Carlini <paolo.carlini@oracle.com>
PR libstdc++/38596
* include/tr1_impl/function (function<>::target,
function<>::target_type): Provide only when __GXX_RTTI is defined.
(_Function_base::_M_manager, _Ref_manager<>::_M_manager,
_Function_handler<void(_ArgTypes...), _Member _Class::*>::_M_manager):
Adjust.
Jan Hubicka [Sun, 21 Dec 2008 13:28:26 +0000 (14:28 +0100)]
i376.md (UNSPEC_MS_TO_SYSV_CALL): New constant.
* i376.md (UNSPEC_MS_TO_SYSV_CALL): New constant.
(call_1_rex64_ms_sysv, call_value_0_rex64_ms_sysv,
call_value_1_rex64_ms_sysv): New patterns.
* i386.c (function_arg_ms_64): Pass magical value of -2 as callarg.
(ix86_expand_call): Emit extra clobbers for ms->sysv ABI calls.
Co-Authored-By: Kai Tietz <kai.tietz@onevision.com>
From-SVN: r142859
Janis Johnson [Fri, 19 Dec 2008 18:12:40 +0000 (18:12 +0000)]
revert: re PR libfortran/24685 (real(16) formatted input is broken for huge values (gfortran.dg/default_format_2.f90))
Revert:
2008-12-12 Janis Johnson <janis187@us.ibm.com>
PR libgfortran/24685
* gfortran.dg/default_format_denormal_2.f90: Change XFAIL to check
for size of long double.
Richard Earnshaw [Fri, 19 Dec 2008 17:31:12 +0000 (17:31 +0000)]
re PR target/38548 (bootstrap broken on arm-linux-gnu (not gnueabi))
PR target/38548
* arm/t-linux (LIB1ASMFUNCS): Add _arm_addsubdf3 and
_arm_addsubsf3.
* arm/lib1funcs.asm (clzsi2): Use RET macro for return
instruction.
Jakub Jelinek [Fri, 19 Dec 2008 14:57:29 +0000 (15:57 +0100)]
re PR libgcj/38396 (ecj1 linked against both -lgcj and -lgcj_bc)
PR libgcj/38396
* configure.ac (use_libgcj_bc): Set to no if not enable_shared.
(LIBGCJ_SPEC): Use -lgcj instead of -lgcj_bc even for -static
or -static-libgcj.
* Makefile.am (ecjx_SOURCES): Add ecjx.cc.
(ecjx_LDADD): Don't add libgcj.la when
NATIVE && USE_LIBBGCJ_BC.
* ecjx.cc: New file.
* Makefile.in: Regenerated.
* configure: Regenerated.
Jakub Jelinek [Fri, 19 Dec 2008 14:55:42 +0000 (15:55 +0100)]
re PR bootstrap/37739 (bootstrap broken with core gcc > gcc-4.2.x)
PR bootstrap/37739
* config.host: For powerpc*-*-linux* host with 32-bit GCC,
use rs6000/x-linux-relax snippet if ld is new enough,
otherwise use rs6000/x-linux-O1.
* config/rs6000/x-linux-relax: New file.
* config/x-cflags-O1: New file.
Jack Howarth [Thu, 18 Dec 2008 21:08:28 +0000 (21:08 +0000)]
21674.cc: Add xfail on all targets except darwin.
2008-12-18 Jack Howarth <howarth@bromo.med.uc.edu>
* testsuite/21_strings/basic_string/element_access/wchar_t/21674.cc:
Add xfail on all targets except darwin.
* libstdc++-v3/testsuite/21_strings/basic_string/element_access/char/
1674.cc: Same.
* libstdc++-v3/testsuite/22_locale/ctype/is/char/2.cc: Correct syntax
for xfailing on all targets except linux.
Ralf Wildenhues [Thu, 18 Dec 2008 20:04:55 +0000 (20:04 +0000)]
Backport from upstream Libtool:
* libltdl.m4 (_LT_SYS_DYNAMIC_LINKER)
(_LT_LINKER_SHLIBS): Add cache variables to tests that
require the linker to work. For shlibpath_overrides_runpath,
this also changes the semantics to let the result from the C
compiler take precedence.
Jan Hubicka [Thu, 18 Dec 2008 13:48:36 +0000 (14:48 +0100)]
i386.h (CONDITIONAL_REGISTER_USAGE): Initialize for current function ABI.
* i386.h (CONDITIONAL_REGISTER_USAGE): Initialize for current function
ABI.
* i386.c (ix86_call_abi_override): Do not trigger target re-init and
do not try to modify call used regs.
(ix86_maybe_switch_abi): New function.
(TARGET_EXPAND_TO_RTL_HOOK): New macro.
Co-Authored-By: Kai Tietz <kai.tietz@onevision.com>
From-SVN: r142810
PR rtl-optimization/37922
* dse.c (bb_info): Added regs_live field.
(look_for_hardregs): New function.
(replace_read): Added regs_live parameter and code to check that
shift sequence does not clobber live hardregs.
(check_mem_read_rtx): Added parameter to replace_read.
(dse_step1): Added regs_live bitmap and initialize it.
(rest_of_handle_dse): Added DF_NOTES problem and earlier call to
df_analyze.
* df-problems.c Renamed to
df_simulate_initialize_backwards.
(df_simulate_one_insn): Renamed to
df_simulate_one_insn_backwards.
(df_simulate_artificial_refs_at_top): Renamed to
df_simulate_finalize_backwards.
(df_simulate_initialized_forwards,
df_simulate_one_insn_forwards,
df_simulate_finalize_backwards): New functions.
* df.h (df_simulate_artificial_refs_at_end): Renamed to
df_simulate_initialize_backwards.
(df_simulate_one_insn): Renamed to
df_simulate_one_insn_backwards.
(df_simulate_artificial_refs_at_top): Renamed to
df_simulate_finalize_backwards.
(df_simulate_initialized_forwards,
df_simulate_one_insn_forwards,
df_simulate_finalize_backwards): New functions.
* ra-conflict.c (global_conflicts): Renamed
df_simulate_artificial_refs_at_end to
df_simulate_initialize_backwards.
* sel-sched.c (propagate_lv_set): Renamed df_simulate_one_insn to
df_simulate_one_insn_backwards.
* ifcvt.c (dead_or_predicable): Renamed
df_simulate_artificial_refs_at_end to
df_simulate_initialize_backwards. Renamed df_simulate_one_insn to
df_simulate_one_insn_backwards.
* recog.c (peephole2_optimize): Ditto.
* rtl-factoring (collect_pattern_seqs, clear_regs_live_in_seq): Ditto.
Daniel Kraft [Thu, 18 Dec 2008 10:05:54 +0000 (11:05 +0100)]
re PR fortran/31822 (Missing run-time bound checks for character pointer => target)
2008-12-18 Daniel Kraft <d@domob.eu>
PR fortran/31822
* gfortran.h (gfc_check_same_strlen): Made public.
* trans.h (gfc_trans_same_strlen_check): Made public.
* check.c (gfc_check_same_strlen): Made public and adapted error
message output to be useful not only for intrinsics.
(gfc_check_merge): Adapt to gfc_check_same_strlen change.
* expr.c (gfc_check_pointer_assign): Use gfc_check_same_strlen for
string length compile-time check.
* trans-expr.c (gfc_trans_pointer_assignment): Add runtime-check for
equal string lengths using gfc_trans_same_strlen_check.
* trans-intrinsic.c (gfc_trans_same_strlen_check): Renamed and made
public from conv_same_strlen_check.
(gfc_conv_intrinsic_merge): Adapted accordingly.
2008-12-18 Daniel Kraft <d@domob.eu>
PR fortran/31822
* gfortran.dg/char_pointer_assign_2.f90: Updated expected error message
to be more detailed.
* gfortran.dg/char_pointer_assign_4.f90: New test.
* gfortran.dg/char_pointer_assign_5.f90: New test.
Jakub Jelinek [Thu, 18 Dec 2008 07:54:43 +0000 (08:54 +0100)]
re PR middle-end/38533 (tree-ssa-reassoc.c increases register pressure several times)
PR middle-end/38533
* tree-ssa-reassoc.c (remove_visited_stmt_chain): New function.
(rewrite_expr_tree): Add moved argument, move stmts together if
needed. Call remove_visited_stmt_chain.
(linearize_expr_tree): Don't move stmts here.
(reassociate_bb): Call remove_visited_stmt_chain if num ops is 1.
Adjust rewrite_expr_tree caller.
* gcc.dg/tree-ssa/pr38533.c: New test.
* gcc.c-torture/execute/pr38533.c: New test.
Jason Merrill [Wed, 17 Dec 2008 19:08:14 +0000 (14:08 -0500)]
semantics.c (describable_type): New function.
gcc/cp:
* semantics.c (describable_type): New function.
(finish_decltype_type): Use it for dependent exprs.
* cp-tree.h: Declare it.
* mangle.c (write_type) [DECLTYPE_TYPE]: Set skip_evaluation.
(write_expression): If skip_evaluation, use type stubs.
* tree.c (cp_tree_equal): Handle PARM_DECLs from different
declarations of a function.
* init.c (build_new): Do auto deduction if type is describable.
* decl.c (cp_finish_decl): Likewise.
* parser.c (cp_parser_omp_for_loop): Likewise.
gcc/testsuite:
* g++.dg/cpp0x/auto6.C: Test more stuff.
* g++.dg/cpp0x/auto12.C: New test.
libiberty:
* cp-demangle.c (d_expression): Handle rvalue stubs too.
[DEMANGLE_COMPONENT_CAST]: Update mangling.
(d_print_comp): Avoid extra ", " with empty template argument packs.
Remove handling for obsolete T() mangling.
Richard Earnshaw [Tue, 16 Dec 2008 12:03:41 +0000 (12:03 +0000)]
re PR target/37436 (arm-cross-g++. internal compiler error: in extract_insn, at recog.c:1990)
PR target/37436
* arm.c (arm_legitimate_index): Only accept addresses that are in
canonical form.
* predicates.md (arm_reg_or_extendqisi_mem_op): New predicate.
* arm.md (extendqihi2): Use arm_reg_or_extendqisi_mem_op predicate
for operand1.
(extendqisi2): Likewise.
(arm_extendqisi, arm_extendqisi_v6): Use arm_extendqisi_mem_op
predicate for operand1.