[Bug gold/6767] program linked with gold seg faults

2008-07-24 Thread ian at airs dot com
--- Additional Comments From ian at airs dot com 2008-07-25 04:29 --- The powerpc support in gold is not complete. I'm not sure whether it works for any program. -- http://sourceware.org/bugzilla/show_bug.cgi?id=6767 --- You are receiving this mail because: --- You are on th

[Bug gold/5990] running out of file descriptors when linking >1024 files

2008-07-24 Thread ian at airs dot com
--- Additional Comments From ian at airs dot com 2008-07-25 04:28 --- I finally got back to this and committed a patch. Let me know if you have any trouble with it. -- What|Removed |Added --

[Bug gold/6767] New: program linked with gold seg faults

2008-07-24 Thread bje at sources dot redhat dot com
The gengenrtl program seg faults on startup when trying to build GCC on a powerpc64-linux system with gold as the linker: $ build/gengenrtl Segmentation fault Program received signal SIGSEGV, Segmentation fault. 0x1000faf4 in ?? () (gdb) bt #0 0x1000faf4 in ?? () #1 0x100016f8 in __libc_csu_in

[Bug gold/6716] gettext macros mangle system include files and causes compilation errors.

2008-07-24 Thread ian at airs dot com
--- Additional Comments From ian at airs dot com 2008-07-24 21:47 --- Thanks for the bug report. I just committed a patch which should fix this problem. -- What|Removed |Added --

[Bug gold/6048] gold segfaults when used with g++ 2.95.3

2008-07-24 Thread ian at airs dot com
--- Additional Comments From ian at airs dot com 2008-07-24 21:25 --- Thanks for the bug report. I just committed a patch which fixes the problem. * ehframe.cc (Eh_frame::add_ehframe_input_section): Check whether this->eh_frame_hdr_ is NULL before using it. --

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread hjl dot tools at gmail dot com
--- Additional Comments From hjl dot tools at gmail dot com 2008-07-24 17:59 --- (In reply to comment #18) > Is 2.18.50.0.8 the release where you reported success? It should work. -- http://sourceware.org/bugzilla/show_bug.cgi?id=6753 --- You are receiving this mail because: --

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread funtoos at yahoo dot com
--- Additional Comments From funtoos at yahoo dot com 2008-07-24 17:58 --- Is 2.18.50.0.8 the release where you reported success? -- http://sourceware.org/bugzilla/show_bug.cgi?id=6753 --- You are receiving this mail because: --- You are on the CC list for the bug, or are wa

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread hjl dot tools at gmail dot com
--- Additional Comments From hjl dot tools at gmail dot com 2008-07-24 17:46 --- You can download the Linux binutils source from http://www.kernel.org/pub/linux/devel/binutils/ and compare ChangeLogs. -- http://sourceware.org/bugzilla/show_bug.cgi?id=6753 --- You are receiving

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread funtoos at yahoo dot com
--- Additional Comments From funtoos at yahoo dot com 2008-07-24 17:37 --- (In reply to comment #15) > (In reply to comment #14) > > > > Is there a CHANGELOG for cvs version I can inspect with respect to > > 2.18.50.0.7? > > There are ChangeLog files in each subdirectory. You can do a

[Bug ld/6766] New: Should "--entry foo" imply "-u foo"?

2008-07-24 Thread hjl dot tools at gmail dot com
[EMAIL PROTECTED] entry]$ cat foo.c void foo () { } [EMAIL PROTECTED] entry]$ make cc-c -o foo.o foo.c ar rv libfoo.a foo.o ar: creating libfoo.a a - foo.o ld --entry foo -o foo libfoo.a ld: warning: cannot find entry symbol foo; not setting start address ld -u foo --entry foo -o bar libfoo.a n

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread hjl dot tools at gmail dot com
--- Additional Comments From hjl dot tools at gmail dot com 2008-07-24 17:23 --- (In reply to comment #14) > > Is there a CHANGELOG for cvs version I can inspect with respect to > 2.18.50.0.7? There are ChangeLog files in each subdirectory. You can do a diff on them. -- http://sou

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread funtoos at yahoo dot com
--- Additional Comments From funtoos at yahoo dot com 2008-07-24 17:13 --- (In reply to comment #13) > Still works for me with Today's CVS: > > [EMAIL PROTECTED] 6753]$ make > gcc -m32 -B/export/home/hjl/bugs/binutils/6753/usr/bin/ -I. -fPIC -c -o > mysymbol.o mysymbol.c > mkdir -p /e

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread hjl dot tools at gmail dot com
--- Additional Comments From hjl dot tools at gmail dot com 2008-07-24 16:56 --- Still works for me with Today's CVS: [EMAIL PROTECTED] 6753]$ make gcc -m32 -B/export/home/hjl/bugs/binutils/6753/usr/bin/ -I. -fPIC -c -o mysymbol.o mysymbol.c mkdir -p /export/home/hjl/bugs/binutils/675

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread funtoos at yahoo dot com
--- Additional Comments From funtoos at yahoo dot com 2008-07-24 15:52 --- you did not specify -L/usr/lib on the link line. The point is that $PREFIX/usr/lib loses out to /usr/lib when specified before it. -- What|Removed |Added --

[Bug ld/6761] abbrev_offset in DWARF2 info is set incorrectly (zero) by ld

2008-07-24 Thread nickc at redhat dot com
--- Additional Comments From nickc at redhat dot com 2008-07-24 15:12 --- Hi Bjarne, This a GDB problem, not a binutils one. The reason that the abbrev offset field is not initialised in the object file is that its final value is not known. Instead there is a relocation that is resp

[Bug binutils/6765] [PATCH] support symbol unbinding in objcopy

2008-07-24 Thread srk31 at srcf dot ucam dot org
--- Additional Comments From srk31 at srcf dot ucam dot org 2008-07-24 14:17 --- Created an attachment (id=2839) --> (http://sourceware.org/bugzilla/attachment.cgi?id=2839&action=view) Patch to objcopy adding --unbind-symbol option -- http://sourceware.org/bugzilla/show_bug.cgi?id=

[Bug binutils/6765] New: [PATCH] support symbol unbinding in objcopy

2008-07-24 Thread srk31 at srcf dot ucam dot org
When manipulating object code, it's often useful to separate symbol references from symbol definitions -- for example in conjunction with the --wrap option to ld. Ordinarily --wrap doesn't work within the scope of a single object file, since the assembler has already resolved references to symbols

[Bug ld/6753] In a prefixed install, -L order is not respected when searching for -l, but -T fixes it

2008-07-24 Thread hjl dot tools at gmail dot com
--- Additional Comments From hjl dot tools at gmail dot com 2008-07-24 14:21 --- (In reply to comment #10) > (In reply to comment #9) > > (In reply to comment #8) > > > > > > > > You have a very strange setup. You better come up with a testcase > > > > which doesn't require me to copy t

[Bug gold/6501] Linux kernel (32 bit) doesn't build: upset at vdso offsets

2008-07-24 Thread ian at airs dot com
--- Additional Comments From ian at airs dot com 2008-07-24 07:28 --- Thanks for reporting the problem and providing the config. I just committed a patch to fix this problem. * dynobj.cc (Versions::Versions): If there is an soname, use it as the base version name. --