Re: Bug#246547: amd64 support for glibc 2.3.2.ds1-14
On 04-Jul-08 23:07, GOTO Masanori wrote:
> Please test this version:
Thank you for trying my amd64-glibc-patches.
I tried to download your debs but I get the following error:
'Forbidden - You don't have permission to access
/~gotom/debian/glibc/amd64/libc6_2.3.2.ds1-14_amd64.deb on this server.'
> BTW, is this dpatch written by only you? I need to describe the
Yes, I wrote the amd64-lib.dpatch.
> BTW, I concerned gcc multilib + gcc 3.4 support. This may be not
> happened in sarge. I wait to put the modification of #252720 until
> the request is come. If we want to release sarge soon, and if we want
> to put amd64 into sarge, then it's good idea to keep gcc 3.3, IMHO.
> What's the merit of pushing gcc 3.4 into sarge on amd64?
gcc-3.4 has much better support for amd64 than gcc-3.3 (better
performance, better support for large data structures, fewer internal
compiler errors, etc.). Some packages do no build at all with gcc-3.3
(e.g. gnustep-base, boost and a few others) and other packages do not
work if compiled with gcc-3.3 (e.g. some recent mozilla* packages).
There is a patch for gcc-3.3 ('hammer-branch') which seems to solve some
amd64 problems but gcc-3.3 with this patch miscompiles the linux kernel.
I am just recompiling my own private amd64 archive using gcc-3.4.1 and
it seems to work quite well so far. I compiled about 2000 packages
up to now and only 14 of those needed simple patches to make gcc-3.4.1
However, there is a problem with gcc-3.4 and glibc. The 'nptl'
pass fails when using gcc-3.4 to compile glibc and the pthread library
(libpthread-0.10.so) seems to be miscompiled by gcc-3.4.1
(every binary linked to it segfaults). I will try to investigate this.
Maybe you have an idea where these problems could come from?