* configure.in: Rename to configure.ac.
* configure.ac: Renamed from configure.in; make minimum necessary
changes for autoconf 2.5x.
* aclocal.m4: Make minimum necessary changes for autoconf 2.5x.
* configure: Regenerate with autoconf 2.57.
PR bootstrap/11932, PR bootstrap/11933
(I don't know if it will fix either of them, but it relates
to them.)
* configure.in: Don't use shared config.cache for target
directories.
* configure: Regenerate.
cp-demangle.c (d_encoding): When DMGL_PARAMS is not set...
* cp-demangle.c (d_encoding): When DMGL_PARAMS is not set, strip
CV-qualifiers from D_COMP_LOCAL_NAME right subtree.
* cplus-dem.c (demangle_arm_hp_template): Set DMGL_PARAMS when
demangling template parameters.
* testsuite/test-demangle.c (fail): New static function.
(main): Support new options in input file: --no-params,
--is-v3-ctor, and --is-v3-dtor.
* testsuite/demangle-expected: Add --no-params to most tests, and
add the correct result when parameters are not demangled. Add
some simple tests for V3 constructor/destructor recognition.
Jan Hubicka [Fri, 2 Jan 2004 18:06:16 +0000 (19:06 +0100)]
cgraphunit.c (cgraph_optimize_function): Call optimize_inline_calls when there is nothing to inline but warnings are...
* cgraphunit.c (cgraph_optimize_function): Call optimize_inline_calls
when there is nothing to inline but warnings are requested.
(cgraph_decide_inlining): Fix memory leak.
PR debug/13367
* cgraph.c (cgraph_function_possibly_inlined): Even with
flag_really_no_inline we inline always_inline functions.
* cgraphunit.c (cgraph_analyze_function): Clear inlinable flag
for non-always_inline functions when there is flag_really_no_inline.
(cgraph_decide_inlining): Limit work done when not inlining.
(cgraph_decide_inlining_incrementally): Likewise.
(cgraph_optimize_function): Check whether something got inlined.
* c-objc-common.c (c_disregard_inline_limits): Do not always inline
extern inline functions when not inlining.
Richard Kenner [Wed, 31 Dec 2003 23:02:44 +0000 (23:02 +0000)]
print-rtl.c (print_rtx): For hard register...
* print-rtl.c (print_rtx): For hard register, write out register
number and register name instead of calling PRINT_REG.
* defaults.h (PRINT_REG): Deleted.
* config/i386/i386.c (print_reg): Remove handling of CODE of -1.
Move comments here from i386.h.
(print_operand, print_operand_address): Call print_reg directly.
* config/i386/i386.h (PRINT_REG): Deleted.
Joseph Myers [Wed, 31 Dec 2003 21:19:51 +0000 (21:19 +0000)]
texinfo.tex: Update to version 2003-12-21.10.
* doc/include/texinfo.tex: Update to version 2003-12-21.10.
* doc/gcc.texi, doc/gccint.texi: Don't set font for
@def... commands.
* doc/invoke.texi: Don't use empty @opindex.
Mark Mitchell [Wed, 31 Dec 2003 18:41:28 +0000 (18:41 +0000)]
c-common.c (c_expand_expr): Remove code to return a value different from that returned by expand_expr.
* c-common.c (c_expand_expr): Remove code to return a value
different from that returned by expand_expr.
* expr.c (store_expr): Use the validity of a target MEM, rather
than checking DECL_RTL (exp), to figure out if a copy is
required.
Michael Koch [Wed, 31 Dec 2003 11:27:09 +0000 (11:27 +0000)]
2003-12-31 Michael Koch <konqueror@gmx.de>
* gnu/java/net/protocol/http/Connection.java
(sendRequest): Dont encode output in default character encoding,
add correct version number to HTTP user agent string.
Guilhem Lavaux [Wed, 31 Dec 2003 11:04:21 +0000 (11:04 +0000)]
2003-12-31 Guilhem Lavaux <guilhem@kaffe.org>
* java/io/ObjectOutputStream.java
(putFields): Reindented. Fixed behaviour: currentPutField should be
null
before calling this method.
(writeFields): Likewise.
(markFieldsWritten): Fixed the exception message.
(callWriteMethod): Ensure currentPutField is null.
(getBooleanField): Translate IllegalArgumentException into
InvalidClassException.
(getByteField): Likewise.
(getCharField): Likewise.
(getDoubleField): Likewise.
(getFloatField): Likewise.
(getIntField): Likewise.
(getLongField): Likewise.
(getShortField): Likewise.
(getObjectField): Check the type code before returning the object.
(getField): Translate NoSuchFieldException into InvalidClassException
directly.
Paolo Carlini [Wed, 31 Dec 2003 08:28:10 +0000 (08:28 +0000)]
locale_facets.tcc (num_get::_M_extract_int, [...]): According to 22.2.2.1.2...
2003-12-31 Paolo Carlini <pcarlini@suse.de>
* include/bits/locale_facets.tcc (num_get::_M_extract_int,
num_get::_M_extract_float): According to 22.2.2.1.2, p8-9,
_first_ look for thousands_sep, then for decimal_point and
finally for digits.
(num_get::_M_extract_float): After the decimal_point or
'e'/'E', decimal_point and thousands_sep just break out the
parsing loop.
* testsuite/22_locale/num_get/get/char/11.cc: Add tests.
* testsuite/22_locale/num_get/get/wchar_t/11.cc: Likewise.
Michael Koch [Tue, 30 Dec 2003 19:56:49 +0000 (19:56 +0000)]
[multiple changes]
2003-12-30 Guilhem Lavaux <guilhem@kaffe.org>
* java/util/GregorianCalendar.java
(computeFields): Reported by Ito Kazumitsu <kaz@maczuka.gcd.org>.
Fixed the computation of DAY_OF_WEEK_IN_MONTH.
(computeTime): 12:00 midnight is AM and 12:00 noon is PM.
2003-12-30 Michael Koch <konqueror@gmx.de>
* testsuite/libjava.mauve/xfails: Removed the following testcase
because it passes now:
FAIL: gnu.testlet.java.text.SimpleDateFormat.Test: parse() strict
(number 1)
Roger Sayle [Tue, 30 Dec 2003 17:21:54 +0000 (17:21 +0000)]
fold-const.c (fold): Don't require strict type equality...
* fold-const.c (fold) <COND_EXPR>: Don't require strict type
equality, instead just prevent replacing a COND_EXPR of non-void
type by one of its operands of void type.
Michael Koch [Tue, 30 Dec 2003 13:43:24 +0000 (13:43 +0000)]
2003-12-30 Michael Koch <konqueror@gmx.de>
* gnu/java/net/protocol/http/Connection.java
(outputStream): New field.
(bufferedOutputStream): New field.
(connect): Initialize outputStream and bufferedOutputStream.
(sendRequest): Create PrintWriter object from outputStream,
support HTTP 1.1, send missing HTTP headers and buffered output data
for POST method.
(getOutputStream): Set request method to POST if output stream is
used, return bufferedOutputStream.
(setRequestMethod): Allow HEAD and POST methods.
This fixes libgcj PR/6302 and libgcj PR/7752.
Michael Koch [Tue, 30 Dec 2003 12:02:47 +0000 (12:02 +0000)]
2003-12-30 Michael Koch <konqueror@gmx.de>
* gnu/java/net/protocol/http/Connection.java
(requestProperties): New field.
(addRequestProperty): New method.
(getRequestProperty): New method.
(setRequestProperty): New method.
(getRequestProperties): New method.
Nathan Sidwell [Tue, 30 Dec 2003 10:07:13 +0000 (10:07 +0000)]
re PR c++/13494 (ICE on bidimensional array subscription in template function.)
cp:
PR c++/13494
* tree.c (build_cplus_array_type_1): Only build a minimal array
type for dependent types or domains.
testsuite:
PR c++/13494
* g++.dg/template/array2-1.C: New test.
* g++.dg/template/array2-2.C: New test.
Andrew Pinski [Tue, 30 Dec 2003 01:53:50 +0000 (17:53 -0800)]
re PR target/13456 (ICE in extract_insn: , at recog.c:2061)
2003-12-30 Jan Hubicka <jh@suse.cz>
PR target/13456
* i386.md (allocate_stack_worker): Use different pattern for pre and
post reload expansion.
(allocate_stack_worker_1, allocate_stack_worker_rex64): Use
match_scratch.
(allocate_stack_worder_1_postreload,
allocate_stack_worker_rex64_postreload): New.
Jan Hubicka [Tue, 30 Dec 2003 01:21:50 +0000 (02:21 +0100)]
i386.md (allocate_stack_worker): Use different pattern for pre and post reload expansion.
* i386.md (allocate_stack_worker): Use different pattern for pre and
post reload expansion.
(allocate_stack_worker_1, allocate_stack_worker_rex64): Use
match_scratch.
(allocate_stack_worder_1_postreload,
allocate_stack_worker_rex64_postreload): New.
Roger Sayle [Mon, 29 Dec 2003 16:16:39 +0000 (16:16 +0000)]
re PR fortran/12632 ([g77 only] -fbounds-check ICE)
PR fortran/12632
* fold-const.c (fold) <COND_EXPR>: Don't fold a constant condition,
if the type of the selected branch doesn't match its' parent.
* com.c (ffecom_subscript_check_): Take as an extra argument the
(possibly NULL) decl of the array. Don't create unnecessary tree
nodes if the array index is known to be safe at compile-time.
If the array index is unsafe, force the array decl into memory to
avoid RTL expansion problems.
(ffecom_array_ref_): Update calls to ffecom_subscript_check_.
(ffecom_char_args_x_): Likewise.
* coverage.c (read_counts_file): Better error messages; cause corrupted
profiles to produce hard errors, not just warnings
(get_coverage_counts): Similarly.
Michael Koch [Sun, 28 Dec 2003 21:30:45 +0000 (21:30 +0000)]
2003-12-28 Michael Koch <konqueror@gmx.de>
* gnu/java/net/protocol/http/Connection.java
(inputStream): Made it a DataInputStream.
(requestProperties): Removed.
(hdrHash): Removed.
(hdrVec): Removed.
(headers): New field to store headers.
(connect): Initialize inputStream.
(receiveReply): Merged from classpath. The new algorithm is line based
instead of character based.
(getHeaderField): Use headers.
(getHeaderFields): Use headers.
(getKey): Removed.
(getField): Removed.
* gnu/java/net/HeaderFieldHelper.java: New file.
* Makefile.am (ordinary_java_source_files):
Added gnu/java/net/HeaderFieldHelper.java.
* Makefile.in: Regenerated.
* sbitmap.c (sbitmap_union_of_diff_cg, sbitmap_a_and_b_cg,
sbitmap_a_xor_b_cg): Accumulate "changed" properly.
(sbitmap_not): Zero all bits past n_bit.