2004-08-29 Andrew John Hughes <gnu_andrew@member.fsf.org>
* java/util/AbstractCollection.java, java/util/AbstractList.java,
java/util/AbstractMap.java, java/util/AbstractSequentialList.java,
java/util/ArrayList.java, java/util/Arrays.java,
java/util/BitSet.java, java/util/Calendar.java,
java/util/Collection.java, java/util/ListIterator.java,
java/util/Map.java, java/util/SortedSet.java:
Added additional exceptions to documentation, along
with some additions and corrections.
From-SVN: r86730
2004-08-29 Steven G. Kargl <kargls@comcast.net>
Paul Brook <paul@codesourcery.com>
* intrinsics/bessel.c: New file.
* intrinsics/erf.c: New file.
* Makefie.am: Add intrinsics/bessel.c and intrinsics/erf.c.
* configure.ac: Test for C99 Bessel and Error functions.
* Makefile.in: Regenerate.
* config.h.in: Regenerate.
* configure: Regenerate.
From-SVN: r86728
* config/mips/mips.md (divide_condition): New mode attribute.
(div[sd]f3, *div[sd]f3): Use it. Redefine using :SCALARF.
(sqrt[sd]f3): Redefine using SCALARF.
(*recip[sd]f3, *rsqrt<mode>[ab]): Likewise. Name formerly unnamed
patterns.
From-SVN: r86721
* config/mips/mips.h (MIPS_CMP_MOVT, MIPS_CMP_MOVF): Delete.
* config/mips/mips.c (mips_prepare_builtin_arg): New function.
(mips_prepare_builtin_target): New function.
(mips_expand_builtin): Use them. Update calls to subroutines.
(mips_expand_ps_cond_move_builtin): Use the new functions. Replace
cmp_choice with a boolean argument to select between movt and movf.
Take the comparison's icode as an argument, not the function's fcode.
(mips_expand_compare_builtin): New function. Rename existing function
to mips_expand_scalar_compare_builtin.
(mips_expand_scalar_compare_builtin, mips_expand_4s_compare_builtin)
(mips_expand_ps_compare_builtin): Use the new functions. Take the
comparison's icode as an argument, not the builtin's fcode.
From-SVN: r86716
* config/mips/mips.c (print_operand): Remove '%Y' and '%y'.
(mips_expand_4s_compare_builtin): Redefine CCV4 to be 0 if all
registers are false and -1 if all registers are true.
(mips_expand_ps_compare_builtin): Likewise CCV2. Use subregs for
MIPS_CMP_UPPER and MIPS_CMP_LOWER.
* config/mips/mips.md (mips_cond_move_tf_ps): Use %Q rather than %y.
(bc1any4t, bc1any4f): Adjust for new CCV4 interpretation.
(bc1any2t, bc1any2f): Likewise CCV2.
(bc1upper2t, bc1lower2t, bc1upper2f, bc1lower2f): Delete.
From-SVN: r86715
* config/mips/mips.md (UNSPEC_MOVE_TF_PS): New.
* config/mips/mips-ps-3d.md (mips_cond_move_tf_ps): Express as an
UNSPEC rather than an IF_THEN_ELSE.
* config/mips/mips.c (mips_expand_ps_cond_move_builtin): Emit
mips_cond_move_tf_ps by name.
From-SVN: r86714
2004-08-29 Chao-ying Fu <fu@mips.com>
James E Wilson <wilson@specifixinc.com>
* config/mips/mips-modes.def (V2SF, CCV2, CCV4): New modes. Give CCV2
8 byte size and alignment. Give CCV4 16 byte size and alignment.
* config/mips/mips-ps-3d.md: New file.
* config/mips/mips.c (TARGET_VECTOR_MODE_SUPPORTED_P,
TARGET_INIT_BUILTINS, TARGET_EXPAND_BUILTIN): Define.
(mips_const_insns): Handle CONST_VECTOR the same as CONST_DOUBLE.
(mips_output_move): Emit "mov.ps" for V2SFmode.
(mips_arg_info): Add checks for VECTOR_FLOAT_TYPE_P and
MODE_VECTOR_FLOAT.
(override_options): Check if TARGET_MIPS3D and there was an explicit
!TARGET_PAIRED_SINGLE_FLOAT. Set MASK_PAIRED_SINGLE, if TARGET_MIPS3D.
Check if TARGET_FLOAT64 and TARGET_HARD_FLOAT are both true, when
TARGET_PAIRED_SINGLE_FLOAT is true. Check if the ISA supports
TARGET_PAIRED_SINGLE_FLOAT. Allow MODE_VECTOR_FLOAT in fp registers,
and allow CCV2 and CCV4 to occupy two and four CC registers.
(print_fcc_operand): New function.
(print_operand): Add %Y, %y, %V, %v, %Q for CCV2, CCV4, CC. Modify
%Z to call print_fcc_operand.
(mips_function_value): Handle MODE_VECTOR_FLOAT.
(mips_class_max_nregs): Check for ST_REGS, and handle CCmodes.
(mips_vector_mode_supported_p): New function.
(mips_hard_regno_nregs): Check for ST_REG_P, and handle CCmodes.
(builtin_description): New struct.
(mips_bdesc): New array of struct builtin_description.
(mips_expand_builtin, mips_init_builtins,
mips_expand_ps_cond_move_builtin, mips_expand_4s_compare_builtin,
mips_expand_compare_builtin, mips_expand_ps_compare_builtin): New
functions.
* config/mips/mips.h (MASK_PAIRED_SINGLE, MASK_MIPS3D): New defines.
(TARGET_PAIRED_SINGLE_FLOAT, TARGET_MIPS3D): New defines.
(TARGET_CPU_CPP_BUILTINS): Add __mips_paired_single_float.
(TARGET_SWITCHES): Added "-mpaired-single", "-mno-paired-single",
"-mips3d", and "-mno-mips3d".
(ASM_SPEC): Map -mips3d to -mips3d in gas.
(EXTRA_CONSTRAINT_Y): New macro.
(EXTRA_CONSTRAINT_STR): Renamed from EXTRA_CONSTRAINT. Add new
multi-letter constraint 'Y'.
(CONSTRAINT_LEN): New macro.
(enum mips_builtins): New for MIPS builtin functions.
(enum mips_function_type): New for the types of MIPS builtin functions.
(enum mips_cmp_choice): New for the MIPS comparison builtin functions.
* config/mips/mips.md: New constants for paired single and MIPS-3D
instructions. Include the new mips-ps-3d.md file.
(addv2sf3, subv2sf3, mulv2sf3, absv2sf2, negv2sf2, movv2sf,
movv2sf_hardfloat_64bit): New named patterns.
(madd.ps, msub.ps, nmaddv2sf, nmaddv2sf_fastmath, nmsubv2sf,
nmsubv2sf_fastmath, ldxc1_v2sf_si, ldxc1_v2sf_di, sdxc1_v2sf_si,
sdxc1_v2sf_di): New unnamed patterns.
* config/mips/predicates.md (const_0_operand, const_1_operand): Add
const_vector support.
(const_0_or_1_operand): New predicate.
* doc/invoke.texi (MIPS Options): Add -mpaired-single and -mips3d.
Co-Authored-By: James E Wilson <wilson@specifixinc.com>
From-SVN: r86713
2004-08-29 Paolo Carlini <pcarlini@suse.de>
* include/bits/locale_facets.tcc (time_get<>::_M_extract_via_format,
case 'S'): Allow for at least one leap-second (as per C99, 7.23.1
and 7.23.3.5), two if !_GLIBCXX_USE_C99.
* testsuite/22_locale/time_get/get_time/char/4.cc: New.
* testsuite/22_locale/time_get/get_time/wchar_t/4.cc: Likewise.
From-SVN: r86712
* Makefile.in (OBJS-common): Add tree-ssa-propagate.o
(tree-ssa-propagate.o): New rule.
(GTFILES): Add tree-ssa-propagate.c.
* tree-flow.h (struct stmt_ann_d): Remove field
in_ccp_worklist.
* tree-ssa-propagate.c: New file.
* tree-ssa-propagate.h: New file.
* tree-ssa-ccp.c: Re-write to use the routines from
tree-ssa-propagate.c.
From-SVN: r86711
[gcc/ChangeLog]
2004-08-28 Ziemowit Laski <zlaski@apple.com>
* c-common.h: Update comments about ObjC/ObjC++ entry points.
(objc_is_reserved_word): New prototype.
(get_current_scope, objc_mark_locals_volatile): Move prototypes to
separate section; these are call-backs.
* c-lex.c (c_lex_with_flags): Call objc_is_reserved_word() to detect
ObjC/ObjC++ "@" keywords.
* c-tree.h (get_current_scope, objc_mark_locals_volatile): Remove
prototypes; they already live in c-common.h.
* stub-objc.c: Update copyright notice.
(objc_is_reserved_word): New stub.
[gcc/cp/ChangeLog]
2004-08-28 Ziemowit Laski <zlaski@apple.com>
* Make-lang.in (CXX_OBJS): Split up into CXX_OBJS and
CXX_AND_OBJCXX_OBJS.
(CXX_C_OBJS): Include in CXX_AND_OBJCXX_OBJS instead of listing
separately on the link line.
[gcc/objc/ChangeLog]
2004-08-28 Ziemowit Laski <zlaski@apple.com>
* objc-act.c (objc_is_reserved_word): New function.
From-SVN: r86702
PR libfortran/17195
* libgfortran.h (rtoa): Remove prototype.
* runtime/error.c (rtoa): Remove.
* io/write.c (calculate_G_format): Don't add blanks if E format is
used. Add correct number of blanks when exponent width is specified.
(output_float): Rewrite.
testsuite/
* gfortran.dg/edit_real_1.f90: New test.
From-SVN: r86701
* error.c (gfc_error_init_1): Remove blank line in front of
function body. Add missing blank.
(gfc_buffer_error, error_char, error_string): Remove blank line in
front of function body.
(show_locus): Add comma in comment.
(gfc_clear_warning, gfc_warning_check, gfc_clear_error,
gfc_push_error, gfc_pop_error): Remove blank line in front of
function body.
(gfc_get_errors): Typo fix in comment in front of function. Remove
blank line in front of function body.
From-SVN: r86691
PR c++/13684
* cp/decl.c (expand_static_init): Use thread-safety API.
(register_dtor_fn): Return the call, don't expand it.
* cp/tree.c (add_stmt_to_compound): New fn.
(stabilize_call): Use it.
* gimplify.c (gimplify_cleanup_point_expr): Handle CLEANUP_EH_ONLY.
(gimple_push_cleanup): Add eh_only parm.
(gimplify_target_expr): Pass it.
* c.opt (-fno-threadsafe-statics): New option.
* c-opts.c (c_common_handle_option): Handle it.
* c-common.h (flag_threadsafe_statics): Declare it.
* c-common.c (flag_threadsafe_statics): Record it.
* doc/invoke.texi: Document it.
* tsystem.h (_GNU_SOURCE): Define.
* gthr-posix.h (__gthread_recursive_mutex_t): New typedef.
(__GTHREAD_RECURSIVE_MUTEX_INIT): New macro.
(__GTHREAD_RECURSIVE_MUTEX_INIT_FUNCTION): New macro.
(__gthread_recursive_mutex_init_function): New fn.
(__gthread_recursive_mutex_lock): New fn.
(__gthread_recursive_mutex_trylock): New fn.
(__gthread_recursive_mutex_unlock): New fn.
* gthr-solaris.h, gthr-single.h, gthr-dce.h: Likewise.
* gthr-win32.h, gthr-vxworks.h: Likewise.
* gthr.h: Document.
* libsupc++/guard.cc (static_mutex): Internal class implementing a
recursive mutex which controls initialization of local statics.
(__gnu_cxx::recursive_init): New exception class.
(__cxa_guard_acquire): Deal with locking and recursion detection.
(acquire_1, __cxa_guard_abort, __cxa_guard_release): Likewise.
From-SVN: r86687
* configure.ac: Handle --enable-gc-debug.
* configure: Regenerate.
* include/config.h.in (LIBGCJ_GC_DEBUG): Add.
* boehm.cc: Include gc_mark.h, javaxfc.h, but no GC private files.
Rearrange include file order.
(GC_DEBUG): Set if LIBGCJ_GC_DEBUG is set.
(GC_finalize_all, GC_debug_generic_malloc): Don't declare.
(disable_gc_mutex): Delete along with all references.
(_Jv_MarkObj, _Jv_MarkArray): Use public types,
adjust for debug header size.
(_Jv_AllocObj, _Jv_allocPtrFreeObj): Define out of line for
debug case.
(_Jv_AllocArray): Declare min_heap_addr only if needed.
(gcj_describe_type_fn): New.
(_Jv_InitGC): Use GC_new_free_list, GC_new_proc, and GC_new_kind.
Register gcj_describe_type_fn.
* include/boehm-gc.h:
(_Jv_AllocObj, _Jv_allocPtrFreeObj):
Don't define, but declare, for debug case.
* java/lang/natObject.cc:
(GC_DEBUG): Define if LIBGCJ_GC_DEBUG is set.
From-SVN: r86686
* config/rs6000/rs6000.c (rs6000_override_options): Increase
maximum skip to 15. Set function alignment, jump alignment, and
loop alignment to 16 for processors that form dispatch groups.
(rs6000_rtx_costs): Outer code is used.
From-SVN: r86670
* rtl.h (read_rtx): Change prototype.
* read-rtl.c (read_rtx): Provide the caller with both an rtx and a
line number. Return true on success.
* gensupport.c (process_include, init_md_reader_args_cb): Adjust
callers accordingly.
From-SVN: r86657
2004-08-27 Bud Davis <bdavis9659@comcast.net>
PR fortran/16597
* io/io.h: created typedef for unit_mode.
* io/io.h (gfc_unit): added mode to unit structure.
* io/transfer.c (data_transfer_init): flush if a write then
read is done on a unit (direct access files).
* io/rewind.c (st_rewind): Used unit mode instead of global.
* gfortran.dg/pr16597.f90: New test.
From-SVN: r86654
