--- Additional Comments From cvs-commit at gcc dot gnu dot org 2009-06-24
05:09 ---
Subject: Bug 10156
CVSROOT:/cvs/src
Module name:src
Changes by: i...@sourceware.org 2009-06-24 05:09:11
Modified files:
gold : ChangeLog layout.cc
Log message:
--- Additional Comments From ian at airs dot com 2009-06-24 05:07 ---
I apologize for my caution, but there just aren't any good endgames for me or my
employer if somebody gets sticky about my looking at proprietary code. I'm not
going to download the whole zip file. Can you attach the
--- Additional Comments From ian at airs dot com 2009-06-24 05:03 ---
Thanks for the bug report. I fixed these cases. Rather than worry about the
internal data structures when using -r with a shared library, I just made it a
fatal error.
--
What|Removed
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2009-06-24
05:01 ---
Subject: Bug 10237
CVSROOT:/cvs/src
Module name:src
Changes by: i...@sourceware.org 2009-06-24 05:01:17
Modified files:
gold : ChangeLog gold.cc options.cc options.
--- Additional Comments From llucax at gmail dot com 2009-06-24 03:53
---
You can download the zip and just extract the dmd/linux/lib/libphobos2.a file.
I can successfully reproduce the bug doing:
ld -o x hello.o libphobos2.a
If you don't want to download the zip file at all, I can u
--- Additional Comments From ian at airs dot com 2009-06-24 03:33 ---
Thanks for the description. I'm not willing to download source code that is not
free as in speech, especially not a compiler and tools that create object files.
If you could perhaps put together a tarball of objects f
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-24
02:22 ---
archive:file specifiers in a linker script are just a pattern to match against.
They do not cause the corresponding file to be loaded. So you need to arrange
for the archive member to be extracted by ref
--- Additional Comments From ian at airs dot com 2009-06-24 00:39 ---
Thanks for the bug report. The GNU linker links .stab and .stabstr sections,
and dump_syms was relying on that to find the .stabstr section. I just
committed a patch to gold to link the sections in the same way, which
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2009-06-24
00:38 ---
Subject: Bug 10219
CVSROOT:/cvs/src
Module name:src
Changes by: i...@sourceware.org 2009-06-24 00:38:00
Modified files:
gold : ChangeLog layout.cc layout.h
Log me
Conditions: Host computer: Debian sid/i386, Crosscompiling to arm-none-eabi for
a arm Cortex-M3 microcontroller. No target operating system, just a binary to
flash.
- There are two object files which contain some code (interrupt service rutines,
reset rutines, etc): file1.o file2.o
- In file1.o t
seems to be a regression between 2.17 and 2.18
with binutils 2.17:
$ rm -f f
$ touch f
$ objcopy f
objcopy: error: the input file 'f' is empty
but with 2.18+ (including today's CVS HEAD), there is no output at all, just an
exit code of 1
--
Summary: objcopy silently exits with empty
--- Additional Comments From peter dot kovar at gmail dot com 2009-06-23
17:57 ---
Fixed. Thank you for your efforts.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10105
--- You are receiving this mail because: ---
You are on the CC list for the bug, or are watching some
--- Additional Comments From chris at seberino dot org 2009-06-23 17:43
---
It looks like a file named gas/testsuite/gas/arm/arm-it-auto-2.d is missing from
binutils-2.19.51 so I can't apply the patch. What version did you apply this
patch against?
cs
--
http://sourceware.org/bugzi
--- Additional Comments From ian at airs dot com 2009-06-23 14:40 ---
Thanks for the detailed information. I committed a patch which I hope will fix
this problem. Please let me know if it does not.
--
What|Removed |Added
-
--- Additional Comments From vapier at gentoo dot org 2009-06-23 14:36
---
do you need me to put together a new test case ? you should be able to
reproduce by downloading the latest alsa-lib and a normal configure && make ...
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10147
--- Additional Comments From llucax at gmail dot com 2009-06-23 14:14
---
Seeing the original bug description, I now remember how to reproduce it (it's
there =).
You just have to download the DMD compiler and try to compile/link the trivial
program in the bug description. The compiler i
--- Additional Comments From llucax at gmail dot com 2009-06-23 14:06
---
Ok, I reopen the bug then.
The internal error appeared in a big build, I'll try to see if I can reproduce
it again and comment the steps to do it.
--
What|Removed |Added
--
--- Additional Comments From ian at airs dot com 2009-06-23 13:26 ---
That's true, the internal error should be fixed, if it still happens. I wasn't
able to recreate it with the single object file you sent, though.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10126
--- You
--
What|Removed |Added
Status|REOPENED|ASSIGNED
http://sourceware.org/bugzilla/show_bug.cgi?id=10147
--- You are receiving this mail because: --
--- Additional Comments From ian at airs dot com 2009-06-23 13:20 ---
Reopening since not fixed.
--
What|Removed |Added
Status|RESOLVED|REOPEN
It should be possible to specify the end address of a section in the linker
script. Specifically, statements like the following should be possible in a
linker script:
.foo 16K - SIZEOF(.foo) :
{
PROVIDE (_foo_begin = .);
*(.foo)
}
PROVIDE (_foo_end = .);
In this example _foo_end would al
--- Additional Comments From llucax at gmail dot com 2009-06-23 12:58
---
Ok, thanks for taking the time to analyze this. I'll pass this information to
the compiler writer.
What about the internal error in Gold? Even if it's triggered by a problem in
the object file, it's not a bug? If
--- Additional Comments From nickc at redhat dot com 2009-06-23 12:20
---
Hi Chris,
OK, please try out this new patch. (You will need to remove the old one
first). The patch also includes the (current set of) changes for 10288.
Cheers
Nick
--
http://sourceware.org/bugzilla/
--- Additional Comments From nickc at redhat dot com 2009-06-23 12:19
---
Created an attachment (id=4016)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4016&action=view)
Add comments for immediate values outside of the range +-32.
--
What|Removed
--- Additional Comments From vapier at gentoo dot org 2009-06-23 11:35
---
this was a reduced test case. the original one was using a version script.
perhaps i reduced it too much and instead triggered a different bug ? ;) i was
going by the internal error message being the same.
i68
--- Additional Comments From t dot sailer at alumni dot ethz dot ch
2009-06-23 09:47 ---
Created an attachment (id=4015)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4015&action=view)
offending wpcap.dll.debug
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10316
--
i686-pc-mingw32-objdump -x wpcap.dll.debug
wpcap.dll.debug: file format pei-i386
wpcap.dll.debug
architecture: i386, flags 0x013b:
HAS_RELOC, EXEC_P, HAS_DEBUG, HAS
--- Additional Comments From peter dot kovar at gmail dot com 2009-06-23
08:58 ---
configure:6868: checking whether basename is declared
configure:6892: g++ -c -g -O2 conftest.cc >&5
conftest.cc: In function 'int main()':
conftest.cc:87: error: overloaded function with no contextual typ
--- Additional Comments From ian at airs dot com 2009-06-23 07:07 ---
I committed a patch to optimize the section name pool, since that is a case
where matches are likely and the string pool is normally small.
--
What|Removed |Added
---
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2009-06-23
07:04 ---
Subject: Bug 10133
CVSROOT:/cvs/src
Module name:src
Changes by: i...@sourceware.org 2009-06-23 07:04:10
Modified files:
gold : ChangeLog layout.cc stringpool.cc str
--- Additional Comments From ian at airs dot com 2009-06-23 06:58 ---
Can you look in the gold/config.log file for the test for the basename
declaration? In a working case, it looks something like this:
configure:6868: checking whether basename is declared
configure:6892: g++ -c -g -O2
--- Additional Comments From ian at airs dot com 2009-06-23 06:49 ---
Sorry for not getting back to this sooner. Thanks for attaching the object
file. The object file you attached does indeed have an e_ehsize field with a
value of 36. The e_ehsize field is supposed to hold the size of
--- Additional Comments From ian at airs dot com 2009-06-23 07:21 ---
Thanks for the bug report. The problem was an unexpected empty .gnu.warning
section. I committed a patch to fix this in the development sources.
The link still fails, but now it gets a plausible error message:
ld: e
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2009-06-23
07:19 ---
Subject: Bug 10147
CVSROOT:/cvs/src
Module name:src
Changes by: i...@sourceware.org 2009-06-23 07:19:35
Modified files:
gold : ChangeLog object.cc
Log message:
34 matches
Mail list logo
