[12:19] <jbailey> Does ccache invalidate the cache based on compiler version?
[12:55] <lamont> jbailey: yes
[05:42] <infinity> doko : make binutils_2.16 stop hating pike7.6 (or vice versa).  kthxbye.
[05:46] <jbailey> http://people.ubuntu.com/~lamont/buildLogs/p/pike7.6/7.6.27-1ubuntu2/ shows all successful, provide a better bug report.
[05:46] <jbailey> I know I packed the lart around here somewhere.
[05:46] <jbailey> Even better..
[05:47] <jbailey> Enya piano music.
[05:47] <jbailey> infinity: What problem are you seeing?
[06:08] <infinity>  /build/buildd/pike7.6-7.6.27/src/interpret.c:1287: error: PIC register 'ebx' clobbered in 'asm'
[06:08] <infinity> Looks like it's been happening ever since the binutils 2.16 upgrade.
[06:09] <infinity> (Last successful build was with gcc 4.0.0 and binutils 2.15)
[06:09] <infinity> Only on i386.  The other 3 arches are fine.
[06:09] <fabbione> morning
[06:10] <infinity> While it's entirely possible that "write better inline asm" is the real answer, I don't speak x86 asm..
[06:11] <infinity> jbailey : Also, there's a failed build log there now (and there's a matching one for the previous version)
[06:12] <jbailey> Eh, looks from the error like they're not setting up their register spilling correctly.
[06:13] <jbailey> Recent gcc's are a bit stricted about making sure things mark all their asm constructs correctly.
[06:13] <jbailey> I'd be surprised if there wasn't a fix in pike's cvs or whatever for it.
[06:13] <jbailey> If doko hasn't looked by tomorrow, I'll try to do so.
[06:15] <elmo> infinity: regardless, could you look at de-piking-swig?
[06:15] <elmo> I think it'd be good to drop the stupid thing
[06:15] <elmo> and/or should I file a bug
[06:16] <infinity> Bug.  Too bys right now to remember anything that isn't written down.
[06:16] <infinity> busy, too.
[06:26] <infinity> elmo : While you're running spiffy tools, can you tell me what would fall out of the archive if libmysqlclient-lgpl went away?
[06:31] <infinity> Meh, still a few things, apparently.  I should really recompile those on the sly when no one's looking, so I can drop the vile libmysqlclient10
[12:20] <chmj> doko: ping 
[12:21] <doko> chmj: pong
[01:35] <jbailey> # 1 "main.c"
[01:35] <jbailey> # 1 "/usr/include/x86_64-linux-gnu/stdio.h" 1 3
[01:35] <jbailey> Luvly.
[01:35] <jbailey> doko: Around?
[01:37] <doko> yes, only short, have to go to my computer shop ...
[01:37] <doko> broken amd mainboard
[01:44] <jbailey> Ouch
[01:44] <jbailey> doko: Is there anything else you need in glibc i386? 
[01:44] <jbailey> If no, I'll upload this.
[01:46] <doko> do you build it with 4.0 ?
[01:46] <doko> ahh, no, you don't need the biarch includes for the glibc _build_
[01:47] <jbailey> Right. =)
[01:47] <doko> one thing ... please let libc6-dev-amd64 depend on lib64gcc1, same for powerpc, sparc, s390
[01:48] <doko> and libc6-dev-i386 on lib32gcc1
[01:48] <jbailey> 'kay.  Should it be (= ?? )
[01:48] <jbailey> or >= ?
[01:48] <jbailey> Or just as long as it's installed?
[01:49] <jbailey> Hmm, sparc64 already has lib64gcc1 in there.
[01:49] <jbailey> Oh, not dev, but libc6-sparc64
[01:49] <jbailey> All of them ought to.
[01:50] <jbailey> Package: libc6-amd64
[01:50] <jbailey> Depends: libc6 (= 2.3.5-1ubuntu9), lib64gcc1
[01:50] <doko> yes, maybe  >= 4.0.1 would be ok?
[01:50] <jbailey> Are you seeing it not get pulled into sometimes, though?
[01:51] <doko> yes, but the -dev package as well.
[01:51] <doko> no, by what?
[01:51] <doko> you need the libgcc for development
[01:51] <jbailey> libc6-amd64-dev depends on libc6-amd64, which depends on lib64gcc1
[01:51] <doko> gcc-X.X can't depend on it, because it suckes in the biarch stuff unconditionally ...
[01:51] <jbailey> This has been in here for a while, it looks like.
[01:52] <jbailey> BenC added it on 15 Jul 2005
[01:54] <doko> ok. and maybe the libc6-amd64-dev can provide a lib64c-dev virtual package
[01:55] <doko> I don't like b-d's like: libc6-dev-s390x [s390] , libc6-dev-sparc64 [sparc] , libc6-dev-i386 [amd64] , libc6-dev-amd64 [i386] 
[01:56] <jbailey> Then it will be lib64c-dev [s390 sparc i386]  lib32c-dev [amd64]  ?
[01:56] <jbailey> That's not much better. =(
[01:58] <doko> well, I would like to avoid a gcc64 package
[02:00] <jbailey> Let's think about it and do it for a different upload.
[02:00] <jbailey> It would be nice to find something elegantish.  like libc-biarch-dev
[02:00] <jbailey> And have it be a dummy packages on archs without biarch.
[02:01] <doko> yep, that could be nice as well ...
[02:02] <doko> then talk with gotom about it, I persuaded him to add lib64c-dev ...
[02:10] <jbailey> doko: Do you know when you will backport the includedir path to gcc-3.4?
[02:10] <jbailey> I'd like to fix l-k-h do use those paths.
[02:10] <doko> done
[02:11] <jbailey> Cool!
[02:11] <doko> could you upload glibc first, so that I can depend on it?
[02:11] <jbailey> doko: Yup, lemme push the upload now, then.
[02:11] <doko> ok, I'm away now for two hours
[02:11] <jbailey> 'k
[02:12] <jbailey> I have an appointment this morning too.
[02:14] <chmj> erm 
[02:14] <jbailey> chmj: ?
[02:14] <chmj> how do I install glibc  ?
[02:14] <chmj> hi jbailey 
[02:15] <jbailey> What? =)
[02:15] <jbailey> If you don't have glibc installed, you likely aren't doing to install anything else. =)
[02:15] <chmj> erm, right 
[02:15] <chmj> issit it suppose to come with libc.mo ? 
[02:16] <chmj> in debian that is 
[02:16] <jbailey> For i18n?
[02:16] <jbailey> Oh, in Debian?
[02:16] <jbailey> No idea off hand.
[02:16] <jbailey> What are you trying to do?
[02:17] <chmj> find the package that install libc.mo's in /usr/share/locale/(.*)/LC_MESSAGES/libc.mo
[02:18] <jbailey> On Debian, it's locales
[02:22] <chmj> oh right 
[04:45] <jbailey_> Feh, i386 failed.  /me checks
[04:52] <jbailey_> forced unwind support unavailable?  wtf?
[07:59] <elmo> what are our chances of dropping gcc-3.3 from main?
[08:00] <elmo> grub, libpng, libpng3 and openoffice.org seem to be the only things in main that explicityly b-d on it
[08:11] <jbailey> I thought doko had said okay to that before, with the goal of 3.4/4.0 only for this release.
[08:11] <jbailey> (so also dropping 2.95)
[08:13] <elmo> grub at least looks like it's not going to be >> 3.3 ready
[08:13] <elmo> the changelog talks about 0.97 being the first to support it
[08:13] <elmo> I  know doko was working on libpng, dunno if it's fixed yet
[09:04] <jbailey> lamont, lamont-away: ping?
[09:04] <jbailey> lamont: I think the gcc wrapper might be confused.
[10:05] <elmo> lamont: ia64 has a kernel now - reminder about daily d-i builds
[10:15] <jbailey> doko: I seem to get an FTBFS on gcc-4.0:
[10:16] <jbailey> The following requested languages were not found: ada
[10:16] <jbailey> Missing build-dep?
[10:45] <lamont> elmo: thanks
[10:48] <lamont> elmo: cronjob enabled, and one kicked manually
[11:18] <doko> elmo: hmm, looks that I missed the libpng uploads. couldn't reproduce the failures anymore
[11:18] <doko> jbailey: build log?
[11:24] <jbailey> doko: Locally with the new glibc in a clean chroot.
[11:24] <jbailey> I can post it for you, though.
[12:00] <doko> jbailey: the log would be nice. I'll have my computer probably back on Monday
[12:03] <jbailey> doko: 'kay.  I'll get to it before that.