summaryrefslogtreecommitdiffstats
path: root/gnu
AgeCommit message (Collapse)AuthorFilesLines
2004-04-19fix mysql bug: http://bugs.mysql.com/bug.php?id=1442etoh1-2/+3
changed the order of array relocation with the depth first order of variable sco pe. ok pvalchev@
2004-04-19fix mysql bug: http://bugs.mysql.com/bug.php?id=1442etoh1-2/+3
changed the order of array relocation with the depth first order of variable scope. ok pvalchev@
2004-04-16- a malicious server may send path names that translate out of theotto2-0/+21
local cvs tree on the client, enabling the server to overwrite files on the client. - a client may read files outside the repository using the -p flag with the checkout command. ok deraadt@
2004-04-09cvs does not need a private version of getopt(3), our in-tree getopt(3)millert2-12/+2
has compatible semantics.
2004-04-08In GO_IF_LREGITIMATE_ADDRESS(), use RTX_OK_FOR_BASE_P() to check for amiod1-11/+5
valid base register for indexed addressing, rather than providing an incomplete inline version. This allows constructs such as (plus:SI (reg:SI (subreg:SI (reg:DI 1)))) to properly pass through the reload phase if optimization is enabled. Such a construct can be generated with buf[index + index2] = value; if either index or index2 are long long. Fixes the ICE in zsh Src/Zle/computil.c.
2004-04-07merge local changes into perl-5.8.3millert144-4414/+7331
2004-04-07perl 5.8.3 from CPANmillert188-2221/+9191
2004-04-06When optimizing a logical arithmetic operation, which operands are bothmiod1-6/+6
comparisons using different operators, make sure to generate an insn with the correct mode. This means that constructs like if (c != -1 & i < 9999) and if ((i<=7) | (value != 0)) as found in transfig and teTeX, respectively, will now compile with optimization. Note the single & or | in these constructs - this kind is rare enough for this bug to have remained unnoticed until now...
2004-04-05Switch mvme88k to -O2 optimization by default.miod1-2/+2
(the perl chunk discussed with millert@ and pushed upstream)
2004-03-31byteswap strtab length on writing w/o affecting variable'smickey2-6/+6
byte order as it seems to be used later on again. does not affect native linker at all. also fix includes order to avoid picking up native machine/ includes instead of the target (for cross-ld build).
2004-03-3010 X's for mktemp(1);jmc1-6/+6
adjust list -widths accordingly; based on a diff from STeve Andre'
2004-03-19Disable caller-save (which is enabled at -O2 or -Os) as it suffers from amiod1-2/+4
few bugs in the 2.95 codebase, one of them being fatal to m88k. A real fix will come after 3.5. ok deraadt@
2004-03-19Stricter file permissions upon install.miod1-1/+3
ok espie@
2004-03-08Do not build binutils on arm. ARM needs at least binutils 2.14.drahn1-2/+5
Allows others to build OpenBSD/cats, use binutils from devel/binutils/stable Requested by deraadt@
2004-03-03Do not segfault writing an empty history list to a file. Resolves PR 3690.otto1-2/+2
ok tdeval@ markus@
2004-03-01Correctly compute pc-relative relocation addend for amd64.kettenis1-3/+19
ok deraadt@
2004-03-01Build libf2c on gcc3 architectures, ok deraadt@drahn1-3/+3
2004-02-29Build this properly on OpenBSD.drahn11-987/+12451
2004-02-29Import of libf2c from gcc3.3.2drahn257-0/+31949
2004-02-27cope with MACHINE_ARCH changing from x86_64 to amd64. For now, we wantderaadt5-5/+16
the toolchain to remain at the name x86_64. Tested on all other architectures that care. Done as very small snippets, so that espie does not blow a fuse.
2004-02-27Teach gcc the proper "set" syntax for __attribute__ ((alias)) to workmiod2-0/+11
on mvme88k. ok deraadt@ espie@
2004-02-24(process_mips_specific): Print conflictsno as an unsigned long.brad1-1/+2
From: binutils CVS HEAD branch, based on rev 1.231 Provided by: Daniel Lucq <daniel at lucq dot org> ok fgsch@ drahn@
2004-02-24(display_debug_info): Cast for 64bit hosts.brad1-2/+2
From: binutils CVS HEAD branch, based on rev 1.113 Provided by: Daniel Lucq <daniel at lucq dot org> ok fgsch@ drahn@
2004-02-23fix a seg fault issue with objdump on sparc64.brad1-4/+12
From: binutils CVS HEAD branch, based on rev 1.8 Reported by: Daniel Lucq <daniel at lucq dot org> ok miod@
2004-02-21kill old copies of gcc-localespie2-572/+0
2004-02-21lose gcc-local as well.espie1-2/+2
2004-02-21zap cccp.1, install phantom info.espie1-4/+36
ok theo
2004-02-20do not install a ccpp.1 linkderaadt1-4/+2
2004-02-18Be sure to pass a third argument to open() when used with O_CREAT.miod2-3/+3
ok espie@
2004-02-17qmake on sparc64 uses a character array with the size of 1 and pp moves its location in the stack. then, pp breaks the memory alignment of every local variables.etoh1-6/+5
This fixes the problem not to change the location when the array is not aligned based on word boundary. This also changes SUSPICIOUS_BUF_SIZE from zero to eight. This is the original pp protection. ok espie@, sturm@, drahn@, brad@
2004-02-16fix gcc3 i386 strange behavior about extracting insns from PARALLEL insn. this patch takes out inside insns from the parallel insn only when the parallel insn is not valid.etoh1-19/+24
ok drahn@, brad@, sturm@, espie@
2004-02-12remove the rev 1.2 fix on every architecuters except ARM due to propolice breakage on i386. This is a temporary fix until the breakage will be resolved.etoh1-2/+2
ok espie and dale
2004-02-11enable shared libs on amd64; from drahnderaadt1-1/+1
2004-02-10configs for amd64; brad@ okmickey8-3/+27
2004-02-10remove HARD_QUAD from sparc64, as was done in gcc 2.95sturm1-2/+1
ok jason@, pval@ no objection espie@
2004-02-10force jump tables into .text for they get screwed in .rodata for -fpic; espie@ okmickey1-0/+3
2004-02-09Corrected lib_specs for OPENBSD/arm shared libraries. copied from alpha.drahn1-2/+7
2004-02-09Do not compile libstdc++ with --enable-shared on NOPIC archs. ok espie@drahn1-2/+7
2004-02-08shave build time.espie1-12/+30
okay millert@
2004-02-07Copy arm/t-netbsd removing the shared libgcc pieces, suggested by espiedrahn2-1/+8
2004-02-05fix the conversion of the insn that comes from gen_load_multiple. ARM processor defines HAVE_load_multiple and uses this gen_load_multiple.etoh1-0/+85
ok pvalchev, drahn
2004-02-03new file is needed to get define early enough for other headers, adddrahn4-10/+48
OpenBSD to the comments, adjust target code to be same as netbsd, now working.
2004-02-03grok VIA C3 xmove-* and xcrypt-* instructions; niklas okderaadt1-2/+38
2004-02-01Add support for xmove-rng and xcrypt-{ecb,cbc,cfb,ofb} instructions foundderaadt1-0/+7
in VIA C3 cpus. Very simple encoding: 0f a7 {c0, c8, d0, e0, d8}. This change is placed into the public domain, so someone get it back to FSF
2004-02-01Permit '-' in a mnemonic. This is needed for the new VIA C3 xmove-rngderaadt1-0/+1
and xcrypt-{ecb,cbc,cfb,ofb} instructions. This chagne is placed into the public domain, so someone get it back to FSF
2004-02-01add paper.txt target;jmc1-0/+4
2004-01-31fix a few #if MACRO -> #if defined(MACRO) (-Wsystem-headers clean-up)espie32-116/+116
naddy, sturm, brad: tests, comments, fixes... thanks.
2004-01-31Provide correct start files and link option specifications on sparc;miod1-0/+17
borrowed from openbsd64.h. "Go ahead, can't hurt" espie@
2004-01-31add documentation (straight copy from egcs)espie2-3/+399
2004-01-31closer to gcc2: bring back protoize.espie2-10/+10
fake more files, to please theo.