1 2009-01-09 Diego Novillo <dnovillo@google.com>
3 * gimple.h (struct gimple_statement_base)<uid>: Document
4 the restrictions on its use.
6 (gimple_set_uid): Tidy.
8 2009-01-09 Jakub Jelinek <jakub@redhat.com>
10 * config/i386/i386.c (ix86_expand_movmem, ix86_expand_setmem): Add
11 zero guard even if align_bytes != 0 and count is smaller than
14 2008-01-09 Vladimir Makarov <vmakarov@redhat.com>
16 PR rtl-optimization/38495
17 * ira-emit.c (print_move_list, ira_debug_move_list): New
19 (add_range_and_copies_from_move_list): Print all added ranges.
20 Add ranges to memory optimized destination.
22 2009-01-09 Jakub Jelinek <jakub@redhat.com>
26 * config/i386/i386.c (override_options): Reject
27 -mstringop-strategy=rep_8byte with -m32.
28 (ix86_expand_movmem): For size_needed == 1 set epilogue_size_needed
29 to 1. Do count comparison against epilogue_size_needed at compile
30 time even when count_exp was constant forced into register. For
31 size_needed don't jump to epilogue, instead just avoid aligning
32 and invoke the body algorithm. If need_zero_guard, add zero guard
33 even if count is non-zero, but smaller than size_needed + number of
34 bytes that could be stored for alignment.
35 (ix86_expand_setmem): For size_needed == 1 set epilogue_size_needed
36 to 1. If need_zero_guard, add zero guard even if count is non-zero,
37 but smaller than size_needed + number of bytes that could be stored
38 for alignment. Compare size_needed with epilogue_size_needed instead
39 of desired_align - align, don't adjust size_needed, pass
40 epilogue_size_needed to the epilogue expanders.
43 * c-pretty-print.c (pp_c_expression): Handle GOTO_EXPR like
46 2009-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
48 * pa.c (last_address): Change to unsigned.
49 (update_total_code_bytes): Change argument to unsigned. Don't
50 check if insn addresses are set.
51 (pa_output_function_epilogue): Set last_address to UINT_MAX if insn
52 addresses are not set.
53 (pa_asm_output_mi_thunk): Handle wrap when updating last_address.
55 2009-01-09 Nick Clifton <nickc@redhat.com>
57 * config/sh/symbian.c: Replace uses of DECL_INLINE with
58 DECL_DECLARED_INLINE_P.
60 2009-01-09 Jakub Jelinek <jakub@redhat.com>
63 * dojump.c (do_jump_by_parts_zero_rtx): Use mode instead of
64 GET_MODE (op0) in operand_subword_force calls.
67 * fold-const.c (fold_unary): For COMPOUND_EXPR and COND_EXPR,
68 fold_convert arg0 operands to TREE_TYPE (op0) first.
70 2009-01-08 Vladimir Makarov <vmakarov@redhat.com>
72 * params.def (ira-max-conflict-table-size): Decrease default value
75 2009-01-08 Jakub Jelinek <jakub@redhat.com>
77 PR tree-optimization/37031
78 * lambda-code.c (lambda_collect_parameters): Call pointer_set_destroy
80 (build_access_matrix): Reserve correct size for AM_MATRIX vector,
81 allocate it using gc instead of heap, use VEC_quick_push instead of
83 * graphite.c (build_access_matrix): Allocate AM_MATRIX vector using gc
84 instead of heap, use VEC_quick_push instead of VEC_safe_push.
85 * tree-data-ref.h (struct access_matrix): Change matrix to gc
86 allocated vector from heap allocated.
87 * lambda.h: Add DEF_VEC_ALLOC_P for gc allocated lambda_vector.
88 * tree-loop-linear.c (linear_transform_loops): Allocate nest
89 vector only after perfect_loop_nest_depth call.
91 2009-01-08 Sebastian Pop <sebastian.pop@amd.com>
92 Jan Sjodin <jan.sjodin@amd.com>
94 PR tree-optimization/38559
95 * graphite.c (debug_value, copy_constraint,
96 swap_constraint_variables, scale_constraint_variable, ): New.
97 (get_lower_bound, get_upper_bound): Removed.
98 (graphite_trans_bb_strip_mine): Clean up this code that works
99 only for constant number of iterations. Fully copy upper and
100 lower bound constraints, not only the constant part of them.
101 * graphite.h (debug_value): Declared.
103 2009-01-08 Ira Rosen <irar@il.ibm.com>
105 PR tree-optimization/37194
106 * tree-vect-transform.c (vect_estimate_min_profitable_iters):
107 Don't add the cost of cost model guard in prologue to scalar
108 outside cost in case of known number of iterations.
110 2009-01-07 Nathan Froyd <froydnj@codesourcery.com>
111 Alan Modra <amodra@bigpond.net.au>
113 * config/rs6000/rs6000.c (rs6000_legitimize_address): Check for
114 non-word-aligned REG+CONST addressing.
116 2009-01-07 Uros Bizjak <ubizjak@gmail.com>
119 * config/alpha/alpha.c (alpha_end_function): For TARGET_ABI_OSF, call
120 free_after_compilation when outputting a thunk.
121 (alpha_output_mi_thunk_osf): Assert that we are processing a thunk.
122 Do not call free_after_compilation here.
124 2009-01-07 Uros Bizjak <ubizjak@gmail.com>
126 * config/i386/i386.c (ix86_target_string): Use ARRAY_SIZE.
127 (ix86_valid_target_attribute_inner_p): Ditto.
129 2009-01-07 Jan Sjodin <jan.sjodin@amd.com>
131 PR tree-optimization/38492
132 PR tree-optimization/38498
133 * tree-check.c (operator_is_linear, scev_is_linear_expression): New.
134 * tree-chrec.h (scev_is_linear_expression): Declared.
135 * graphite.c (graphite_cannot_represent_loop_niter): New.
136 (scopdet_basic_block_info): Call graphite_cannot_represent_loop_niter.
137 (graphite_loop_normal_form): Use gcc_assert.
138 (scan_tree_for_params): Use CASE_CONVERT.
139 (phi_node_is_iv, bb_contains_non_iv_scalar_phi_nodes): New.
140 (build_scop_conditions_1): Call bb_contains_non_iv_scalar_phi_nodes.
141 Use gcc_assert. Discard scops that contain unhandled cases.
142 (build_scop_conditions): Return a boolean status for unhandled cases.
143 (strip_mine_profitable_p): Print the loop number, not its depth.
144 (is_interchange_valid): Pass the depth of the loop nest, don't
145 recompute it wrongly.
146 (graphite_trans_bb_block): Same.
147 (graphite_trans_bb_block): Print tentative of loop blocking.
148 (graphite_trans_scop_block): Do not print that the loop has been
150 (graphite_transform_loops): Do not handle scops that contain condition
153 2009-01-07 H.J. Lu <hongjiu.lu@intel.com>
155 AVX Programming Reference (December, 2008)
156 * config/i386/avxintrin.h (_mm256_stream_si256): New.
157 (_mm256_stream_pd): Likewise.
158 (_mm256_stream_ps): Likewise.
160 * config/i386/i386.c (ix86_builtins): Add IX86_BUILTIN_MOVNTDQ256,
161 IX86_BUILTIN_MOVNTPD256 and IX86_BUILTIN_MOVNTPS256.
162 (ix86_special_builtin_type): Add VOID_FTYPE_PV4DI_V4DI.
163 (bdesc_special_args): Add __builtin_ia32_movntdq256,
164 __builtin_ia32_movntpd256 and __builtin_ia32_movntps256.
165 (ix86_init_mmx_sse_builtins): Handle VOID_FTYPE_PV4DI_V4DI.
166 (ix86_expand_special_args_builtin): Likewise.
168 * config/i386/sse.md (AVXMODEDI): New.
169 (avx_movnt<mode>): Likewise.
170 (avx_movnt<mode>): Likewise.
171 (<sse>_movnt<mode>): Remove AVX support.
172 (sse2_movntv2di): Likewise.
174 2009-01-07 Richard Guenther <rguenther@suse.de>
177 * fold-const.c (extract_muldiv): Remove obsolete comment.
178 (fold_plusminus_mult_expr): Undo MINUS_EXPR
179 to PLUS_EXPR canonicalization for the canonicalization.
181 2009-01-07 Gerald Pfeifer <gerald@pfeifer.com>
183 * doc/install.texi (alpha*-dec-osf*): Remove note on 32-bit
184 hosted cross-compilers generating less efficient code.
186 2009-01-06 Richard Sandiford <rdsandiford@googlemail.com>
188 * function.h (rtl_data): Add a dbr_scheduled_p field.
189 * reorg.c (dbr_schedule): Set it.
190 (gate_handle_delay_slots): Check it.
191 * config/mips/mips.c (mips_base_delayed_branch): Delete.
192 (mips_reorg): Check flag_delayed_branch instead of
193 mips_base_delayed_branch.
194 (mips_override_options): Don't set mips_base_delayed_branch
195 or flag_delayed_branch.
197 2009-01-06 Richard Sandiford <rdsandiford@googlemail.com>
199 PR rtl-optimization/38426.
200 * ira.c (ira): Set current_function_is_leaf earlier.
202 2009-01-06 Jakub Jelinek <jakub@redhat.com>
204 PR rtl-optimization/38722
205 * combine.c (try_combine): Don't modify PATTERN (i3) and notes
206 too early, only set a flag and modify after last possible
209 2009-01-06 Janis Johnson <janis187@us.ibm.com>
212 * ginclude/float.h: Rename DECnn_DEN to DECnn_SUBNORMAL_MIN.
213 * real.c (decimal_single_format): Correct values of emin and emax.
214 (decimal_double_format): Ditto.
215 (decimal_quad_format): Ditto.
216 * c-cppbuiltin.c (builtin_define_decimal_float_constants): Adjust
217 computation of DECnn_MIN and DECnn_MAX for corrected values of
218 emin and emax. Define __DECnn_SUBNORMAL_MIN__ instead of
219 __DECnn_MIN__, and adjust its computation for the corrected value
222 2009-01-06 Jan Hubicka <jh@suse.cz>
225 * i386.c (ix86_expand_call): Use ARRAY_SIZE.
227 2009-01-06 Gerald Pfeifer <gerald@pfeifer.com>
229 * doc/contrib.texi (Contributors): Slightly adjust the end note.
230 Add Robert Clark to the list of testers.
232 2009-01-06 Jan Hubicka <jh@suse.cz>
233 Kai Tietz <kai.tietz@onevision.com>
235 * i386.md (*msabi_syvabi): Add SSE regs clobbers.
236 * i386.c (ix86_expand_call): Add clobbers.
238 2009-01-06 Jan Hubicka <jh@suse.cz>
239 Kai Tietz <kai.tietz@onevision.com>
241 * i386.h (CONDITIONAL_CALL_USAGE): SSE regs are not used for w64 ABI.
242 * i386.c (struct ix86_frame): Add padding0 and nsseregs.
243 (ix86_nsaved_regs): Count only general purpose regs.
244 (ix86_nsaved_sseregs): New.
245 (ix86_compute_frame_layout): Update nsseregs; set preferred alignment
246 to 16 for w64; compute padding and size of sse reg save area.
247 (ix86_emit_save_regs, ix86_emit_save_regs_using_mov): Save only
248 general purpose regs.
249 (ix86_emit_save_sse_regs_using_mov): New.
250 (ix86_expand_prologue): Save SSE regs if needed.
251 (ix86_emit_restore_regs_using_mov): Use only general purpose regs.
252 (ix86_emit_restore_sse_regs_using_mov): New.
253 (ix86_expand_epilogue): Save SSE regs if needed.
255 2009-01-06 Jan Hubicka <jh@suse.cz>
256 Kai Tietz <kai.tietz@onevision.com>
258 * i386.h (ACCUMULATE_OUTGOING_ARGS): Enable for MSABI
259 * i386.c (init_cumulative_args): Disallow calls of MSABI functions
260 when accumulate outgoing args is off.
262 2009-01-06 H.J. Lu <hongjiu.lu@intel.com>
265 * ira-color.c (ira_reuse_stack_slot): Check ENABLE_IRA_CHECKING
266 before using pseudos_have_intersected_live_ranges_p.
268 * ira-int.h (ira_assert): Always define.
270 2009-01-06 H.J. Lu <hongjiu.lu@intel.com>
272 AVX Programming Reference (December, 2008)
273 * config/i386/avxintrin.h (_mm_permute2_pd): Removed.
274 (_mm256_permute2_pd): Likewise.
275 (_mm_permute2_ps): Likewise.
276 (_mm256_permute2_ps): Likewise.
277 * config/i386/i386.md (UNSPEC_VPERMIL2): Likewise.
278 * config/i386/sse.md (avx_vpermil2<mode>3): Likewise.
280 * config/i386/i386.c (ix86_builtins): Remove
281 IX86_BUILTIN_VPERMIL2PD, IX86_BUILTIN_VPERMIL2PS,
282 IX86_BUILTIN_VPERMIL2PD256 and IX86_BUILTIN_VPERMIL2PS256.
283 (ix86_builtin_type): Remove V8SF_FTYPE_V8SF_V8SF_V8SI_INT,
284 V4DF_FTYPE_V4DF_V4DF_V4DI_INT, V4SF_FTYPE_V4SF_V4SF_V4SI_INT
285 and V2DF_FTYPE_V2DF_V2DF_V2DI_INT.
286 (bdesc_args): Remove __builtin_ia32_vpermil2pd,
287 __builtin_ia32_vpermil2ps, __builtin_ia32_vpermil2pd256 and
288 __builtin_ia32_vpermil2ps256.
289 (ix86_init_mmx_sse_builtins): Updated.
290 (ix86_expand_args_builtin): Likewise.
292 2009-01-05 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
294 * pa.c (output_call): Relocate non-jump insns in the delay slot of
295 long absolute calls when generating PA 2.0 code.
297 2009-01-05 Vladimir Makarov <vmakarov@redhat.com>
299 PR rtl-optimization/38583
300 * params.h (IRA_MAX_CONFLICT_TABLE_SIZE): New macro.
302 * params.def (ira-max-conflict-table-size): New.
304 * doc/invoke.texi (ira-max-conflict-table-size): Decribe.
306 * ira.h (ira_conflicts_p): New external definition.
308 * ira-conflicts.c (build_conflict_bit_table): Do not build too big
309 table. Report this. Return result of building.
310 (ira_build_conflicts): Use ira_conflicts_p. Check result of
311 building conflict table.
313 * ira-color.c (fast_allocation): Use num instead of ira_allocnos_num.
314 (ira_color): Use ira_conflicts_p.
316 * global.c: Include ira.h.
317 (pseudo_for_reload_consideration_p, build_insn_chain): Use
320 * Makefile.in (global.o): Add ira.h.
322 * ira-build.c (mark_all_loops_for_removal,
323 propagate_some_info_from_allocno): New.
324 (remove_unnecessary_allocnos): Call
325 propagate_some_info_from_allocno.
326 (remove_low_level_allocnos): New.
327 (remove_unnecessary_regions): Add parameter. Call
328 mark_all_loops_for_removal and remove_low_level_allocnos. Pass
329 parameter to remove_unnecessary_regions.
330 (ira_build): Remove all regions but root if the conflict table was
331 not built. Update conflict hard regs for allocnos crossing calls.
333 * ira.c (ira_conflicts_p): New global.
334 (ira): Define and use ira_conflicts_p.
336 * reload1.c (compute_use_by_pseudos, reload, count_pseudo,
337 count_spilled_pseudo, find_reg, alter_reg, finish_spills,
338 emit_input_reload_insns, delete_output_reload): Use ira_conflicts_p.
340 2009-01-06 Ben Elliston <bje@au.ibm.com>
342 * gengtype-lex.l (YY_NO_INPUT): Define.
344 2009-01-05 Andrew Pinski <andrew_pinski@playstation.sony.com>
347 * c-common.c (handle_vector_size_attribute): Also reject
350 2009-01-05 Sebastian Pop <sebastian.pop@amd.com>
352 PR tree-optimization/38492
353 * graphite.c (rename_map_elt, debug_rename_elt,
354 debug_rename_map_1, debug_rename_map, new_rename_map_elt,
355 rename_map_elt_info, eq_rename_map_elts,
356 get_new_name_from_old_name, bb_in_sese_p): Moved around.
357 (sese_find_uses_to_rename_use): Renamed sese_build_livein_liveouts_use.
358 (sese_find_uses_to_rename_bb): Renamed sese_build_livein_liveouts_bb.
359 (sese_build_livein_liveouts): New.
360 (new_sese, free_sese): New.
361 (new_scop): Call new_sese.
362 (free_scop): Call free_sese.
363 (rename_variables_from_edge, rename_phis_end_scop): Removed.
364 (register_old_new_names): Renamed register_old_and_new_names.
365 (register_scop_liveout_renames, add_loop_exit_phis,
366 insert_loop_close_phis, struct igp,
367 default_liveout_before_guard, add_guard_exit_phis,
368 insert_guard_phis, copy_renames): New.
369 (translate_clast): Call insert_loop_close_phis and insert_guard_phis.
370 (sese_add_exit_phis_edge): Renamed scop_add_exit_phis_edge.
371 (rewrite_into_sese_closed_ssa): Renamed scop_insert_phis_for_liveouts.
372 (scop_adjust_phis_for_liveouts): New.
373 (gloog): Call scop_adjust_phis_for_liveouts.
375 * graphite.h (struct sese): Documented. Added fields liveout,
377 (SESE_LIVEOUT, SESE_LIVEIN, SESE_LIVEIN_VER, SESE_NUM_VER): New.
378 (new_sese, free_sese, sese_build_livein_liveouts): Declared.
379 (struct scop): Added field liveout_renames.
380 (SCOP_LIVEOUT_RENAMES): New.
382 2009-01-05 Harsha Jagasia <harsha.jagasia@amd.com>
384 PR tree-optimization/38510
385 * graphite.c (recompute_all_dominators): Call mark_irreducible_loops.
386 (translate_clast): Call recompute_all_dominators before
388 (gloog): Call recompute_all_dominators before graphite_verify.
390 2009-01-05 Harsha Jagasia <harsha.jagasia@amd.com>
391 Jan Sjodin <jan.sjodin@amd.com>
393 PR tree-optimization/38500
394 * graphite.c (create_sese_edges): Call fix_loop_structure after
397 2009-01-05 Joel Sherrill <joel.sherrill@oarcorp.com>
399 * config.gcc: Add m32r*-*-rtems*.
400 * config/m32r/rtems.h: New file.
402 2009-01-05 Ben Elliston <bje@au.ibm.com>
404 * Makefile.in (.po.gmo): Use mkinstalldirs, not test -d || mkdir.
406 (po/gcc.pot): Likewise.
408 2009-01-04 David S. Miller <davem@davemloft.net>
410 * config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED_RTX): Delete.
411 (STARTING_FRAME_OFFSET): Always set to zero.
413 2009-01-04 Richard Sandiford <rdsandiford@googlemail.com>
415 * tree.def (LSHIFT_EXPR, RSHIFT_EXPR): Add commentary.
416 * tree-cfg.c (verify_gimple_assign_binary): Allow shifts of
417 fixed-point types, and vectors of the same.
419 2009-01-04 Richard Sandiford <rdsandiford@googlemail.com>
421 * config/mips/sync.md (*mb_barrier): Rename to...
422 (*memory_barrier): ...this.
424 2009-01-04 Jonathan Wakely <jwakely.gcc@gmail.com>
426 * doc/extend.texi (Function Attributes): Move @cindex after @item
427 for 'artificial' and 'flatten'. Fix grammar for 'externally_visible'
428 and put in alphabetical order. Fix 'target' name and put in order.
429 * doc/invoke.texi (-Wstrict-null-sentinel, -fipa-matrix-reorg): Fix
432 2009-01-04 Uros Bizjak <ubizjak@gmail.com>
434 * config/s390/s390.md (UNSPEC_MB): Rename from UNSPECV_MB.
435 (memory_barrier): Expand as unspec instead of unspec_volatile.
436 Remove mem:BLK from insn operands. Use Pmode scratch register.
437 (*memory_barrier): Define as unspec instead of unspec_volatile.
438 Use (match_dup 0) as input operand.
440 * config/sparc/sparc.md (UNSPEC_MEMBAR): Rename from UNSPECV_MEMBAR.
441 * config/sparc/sync.md (memory_barrier): Expand as unspec instead of
442 unspec_volatile. Remove mem:BLK from insn operands. Use Pmode
443 scratch register. Remove operand 1.
444 (*stbar): Define as unspec instead of unspec_volatile.
445 Use (match_dup 0) as input operand, remove (const_int 8).
446 (*membar): Define as unspec instead of unspec_volatile.
447 Use (match_dup 0) as input operand, remove input operand 2.
449 * config/xtensa/xtensa.md (UNSPEC_MEMW): Rename from UNSPECV_MEMW.
450 (memory_barrier): Expand as unspec instead of unspec_volatile.
451 Remove mem:BLK from insn operands. Use Pmode scratch register.
452 (*memory_barrier): Define as unspec instead of unspec_volatile.
453 Use (match_dup 0) as input operand.
455 * config/ia64/sync.md (memory_barrier): Redefine as expander pattern.
456 Remove mem:BLK from insn operands. Use Pmode scratch register.
457 Set volatile flag on operand 0.
458 (*memory_barrier): New insn pattern.
460 * config/rs6000/sync.md (memory_barrier): Remove mem:BLK from
462 (*memory_barrier): Use (match_dup 0) as input operand.
464 * config/mips/sync.md (memory_barrier): Redefine as expander pattern.
465 Remove mem:BLK from insn operands. Use Pmode scratch register.
466 Set volatile flag on operand 0.
467 (*mb_internal): New insn pattern.
469 * config/alpha/sync.md (*memory_barrier): Rename from *mb_internal.
471 2009-01-04 Steven Bosscher <steven@gcc.gnu.org>
474 * function.c (struct temp_slot): Move to the section of the file
475 that deals with temp slots. Remove field 'address'.
476 (temp_slot_address_table): New hash table of address -> temp slot.
477 (struct temp_slot_address_entry): New struct, items for the table.
478 (temp_slot_address_compute_hash, temp_slot_address_hash,
479 temp_slot_address_eq, insert_temp_slot_address): Support functions
481 (find_temp_slot_from_address): Rewrite to use the new hash table.
482 (remove_unused_temp_slot_addresses): Remove addresses of temp
483 slots that have been made available.
484 (remove_unused_temp_slot_addresses_1): Call-back for htab_traverse,
485 worker function for remove_unused_temp_slot_addresses.
486 (assign_stack_temp_for_type): Don't clear the temp slot address list.
487 Add the temp slot address to the address -> temp slot map.
488 (update_temp_slot_address): Update via insert_temp_slot_address.
489 (free_temp_slots): Call remove_unused_temp_slot_addresses.
490 (pop_temp_slots): Likewise.
491 (init_temp_slots): Allocate the address -> temp slot map, or empty
492 the map if it is already allocated.
493 (prepare_function_start): Initialize temp slot processing.
495 2009-01-04 Steven Bosscher <steven@gcc.gnu.org>
498 * cfgexpand.c (estimate_stack_frame_size): Simplify the estimate:
499 Calculate the size of all stack vars assuming no packing of stack
500 vars will happen, replacing a quadratic algorithm with a linear one.
502 2009-01-03 Jakub Jelinek <jakub@redhat.com>
505 * expmed.c (store_bit_field_1): Don't modify op0 if movstrict insn
508 2009-01-03 Diego Novillo <dnovillo@google.com>
510 * doc/contrib.texi: Update contributions.
512 2009-01-03 Jakub Jelinek <jakub@redhat.com>
515 * builtins.c (fold_builtin_memory_op): Give up if either operand
516 is volatile. Set srctype or desttype to non-qualified version
520 * builtins.c (fold_builtin_expect): Only check DECL_WEAK for VAR_DECLs
523 2009-01-02 Kenneth Zadeck <zadeck@naturalbridge.com>
525 PR rtl-optimization/35805
526 * df-problems.c (df_lr_finalize): Add recursive call to resolve lr
527 problem if fast dce is able to remove any instructions.
528 * dce.c (dce_process_block): Fix dump message.
530 2009-01-02 Mark Mitchell <mark@codesourcery.com>
533 * tree-ssa-pre.c (compute_antic): Correct loop bounds.
535 2009-01-02 Jakub Jelinek <jakub@redhat.com>
538 * tree-flow.h (op_code_prio, op_prio): New prototypes.
539 * tree-pretty-print.c (op_code_prio): New function.
540 (op_prio): No longer static. Use op_code_prio.
541 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs):
542 Use op_prio and op_code_prio to determine if () should be
543 printed around operand(s) or not.
545 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
546 dump_gimple_call, dump_gimple_switch, dump_gimple_cond,
547 dump_gimple_label, dump_gimple_try, dump_symbols, dump_gimple_phi,
548 dump_gimple_mem_ops, dump_bb_header, dump_bb_end, pp_cfg_jump): Use
549 pp_character instead of pp_string for single letter printing.
551 2009-01-02 Richard Sandiford <rdsandiford@googlemail.com>
553 * doc/extend.texi: Fix '#pragma GCC option' typo.
555 2009-01-02 Richard Guenther <rguenther@suse.de>
557 * doc/install.texi (--enable-checking): Mention different
559 (--enable-stage1-checking): Document.
561 2009-01-01 Andrew Pinski <pinskia@gmail.com>
564 * tree-cfg.c (verify_expr): Add INDIRECT_REF case. Change MODIFY_EXPR
567 2009-01-02 Ben Elliston <bje@au.ibm.com>
569 * config/fp-bit.h (pack_d): Constify argument.
570 * config/fp-bit.c (makenan): Constify return type. Remove casts.
571 (isnan): Constify argument.
575 (_fpadd_parts): Constify return type.
576 (_fpmul_parts): Likewise.
577 (_fpdiv_parts): Likewise.
579 2009-01-01 Jakub Jelinek <jakub@redhat.com>
582 * c-typeck.c (add_pending_init): Add IMPLICIT argument. Only
583 warn about overwriting initializer with side-effects or
584 -Woverride-init if !IMPLICIT.
585 (output_init_element): Likewise. Pass IMPLICIT down to
587 (process_init_element): Add IMPLICIT argument. Pass it down
588 to output_init_element.
589 (push_init_element, pop_init_level, set_designator): Adjust
590 process_init_element callers.
591 (set_nonincremental_init, set_nonincremental_init_from_string):
592 Adjust add_pending_init callers.
593 (output_pending_init_elements): Adjust output_init_element callers.
594 * c-tree.h (process_init_element): Adjust prototype.
595 * c-parser.c (c_parser_initelt, c_parser_initval): Adjust
596 process_init_element callers.