[12:03] <jbailey> I was checking out the gcc-4.0 ftbfs that filed in Debian.
[12:04] <doko> which one?
[12:04] <jbailey> Making sure I could close it in Ubuntu when I discovered that it was ftbfs here.
[12:04] <jbailey> http://bugzilla.ubuntu.com/show_bug.cgi?id=13660
[12:08] <doko> ah, yes, that one should be fixed with the biarch include dirs?
[12:10] <jbailey> I was hoping so, or that it didn't affect us at all (since we were able to build gcc-4.0 before just against the glibc)
[12:10] <jbailey> But then I had that ftbfs.
[12:10] <jbailey> So... =)
[12:12] <doko> the 3.4 with biarch include dirs is in the archives. 
[12:34] <jbailey> Cool!
[09:10] <lamont-away> doko: fwiw, gcc-3.4's logwatch process never exits, until you kill it... :-( [hppa] 
[09:16] <doko> lamont-away: but it does terminate for 4.0?
[09:19] <lamont-away> dunno - I'll make a note of it next time there's a 4.0 upload
[09:19] <doko> jbailey: the glibc build did fail on i386 ...
[09:20] <lamont-away> doko: -9 or -10?
[09:21] <doko> -9
[09:21] <lamont-away> (-9 is known/expected, -10 should build...)
[09:21] <lamont-away> he was grumbling about it earlier
[09:21] <doko> I don't see a -10
[09:21] <lamont-away> he was going to upload it sometime "soonish"
[09:21] <lamont-away> -9 had missing build-depends, iirc
[09:21] <lamont-away> but it's 0121, I should really crawl into bed
[09:23] <doko> good night, really :)
[09:24] <doko> jbailey: if you move around the headers in l-k-h, the headers for the default arch will stay where they are?
[10:09] <jbailey> ROAR I hate sleeping alone in such a big place.  Noises keep waking me up.
[10:10] <jbailey> doko: I was testing that gcc-4.0 still built correctly with the new glibc and the build deps in a clean chroot while I was at it.
[10:12] <doko> heh, think you're in Montreal?
[10:12] <jbailey> Err..
[10:12] <jbailey> Is this one of those twilight zone moments where you tell me that I'm really in another universe?
[10:13] <doko> anyway, do you have a l-k-h package which I can test?
[10:13] <doko> you didn't move from Toronto to some French province?
[10:13] <jbailey> Yup.  The province of Qubec, city of Montral.
[10:15] <jbailey> Hey, I see that the massive failures in gcc-4 biarch went away except in Java and mudflap, cool.
[10:17] <doko> we disabled mudflap for biarch
[10:17] <jbailey> doko: But in Toronto we lived on the 11th floor surrounded by concrete.
[10:17] <jbailey> And it was a 1-room appartment.
[10:17] <jbailey> Here, there's 4 rooms, and it's woodframe.
[10:17] <jbailey> So things creak and groan, and they're too far away to just look at from bed.
[10:17] <doko> and libjava isn't built as well, so the failures are ok
[10:18] <jbailey> I think what woke me up just now was the door to the outside storage swingning a bit (It's broken and needs to be repaired so that it will stay shut)
[10:18] <jbailey> But I woke up basically to the sound of a door openning.
[10:18] <jbailey> *sigh*
[10:19] <jbailey> glibc ubuntu10 uploaded.
[10:19] <doko> do you have a l-k-h for testing somewhere?
[10:19] <doko> thanks
[10:20] <jbailey> I don't.  I was trying to decide if it's worth it for breezy, or if it's too featurish.
[10:20] <jbailey> Right now it's setup the same way that ppc/ppc64, sparc/sparc64 and s390/s390x is.
[10:20] <doko> if we want to build the libs from amd64-libs, it's needed
[10:20] <jbailey> Why?  I think the current biarch stuff should cover it
[10:21] <doko> no, that' what drow did mention.
[10:21] <doko> i.e. ncurses includes the wrong headers. wait ...
[10:21] <jbailey> Debian's lkh is only vaguely related to Ubuntu's. =(
[10:22] <jbailey> Lemme double check.
[10:22] <jbailey> Oh, hmm, I'm on crack it seems.
[10:22] <jbailey> I thought I had biarch'd i386
[10:23] <doko> ncurses in my home on chinstrap
[10:24] <jbailey> Ah, I have done so.
[10:24] <jbailey> Right.  the asm directories, not the linux/ directory which is common.
[10:24] <doko> shouldn't it be safe, if we just move the headers for the non-default arch to /usr/include/<arch> ?
[10:25] <jbailey> It should be, but I think Matt considers those things to be features.
[10:25] <jbailey> And it should be safe to do it the way it's already done, since that's how it's been for other biarchs in the Sarge release.
[10:26] <doko> so we have to keep amd64-libs, but drop glibc from it?
[10:26] <jbailey> No idea.  I wouldn't have considered those features.
[10:27] <doko> well, one goal was to remove at least amd64-libs