[01:48] <jbailey> Oh, I see, it's his sparc fun.
[01:48] <jbailey> Yeah, I should try to fix that this week.  At least see what the symbol conflict is.
[01:48] <jbailey> There should be enough of the right people at OLS if I can't answer it myself/
[04:51] <daniels> doko: http://gcc.gnu.org/ml/gcc/2005-07/msg00699.html
[06:27] <fabbione> morning
[06:28] <fabbione> jbailey: you around?
[09:47] <doko> daniels: yep, I'd like to have a test bot which builds some apps with such a modified compiler ...
[09:47] <doko> lamont, infinity: any news about the 4.0 amd64 build?
[01:24] <infinity> doko : Hrm, did -2ubuntu2 work around the bootstrap issues?
[01:24] <infinity> doko : I kicked off a manual bootstrap build on yellow, and noticed that an automatic build on crested was going just fine..
[01:52] <doko> infinity: no, it cannot
[02:08] <fabbione> sh[22008] : segfault at 0000000000000000 rip 0000000000000000 rsp 00007fffffe00860 error 14
[02:08] <fabbione> ld[11285]  general protection rip:2aaaaaaac180 rsp:7fffffbc04f0 error:0
[02:08] <fabbione> YAY
[02:28] <fabbione> kiko: thanks a lot
[02:28] <fabbione> ops
[02:41] <infinity> doko : Interesting, cause it's well into the testsuite by now.
[02:42] <infinity> doko : And the chroot wasn't pre-seeded.
[02:48] <doko> hmm, which lib32gcc1 was installed?
[02:49] <infinity> 4.0.0-12ubuntu2
[02:49] <doko> strange
[02:49] <infinity> Rather, yes.  Either way, I see no obvious failures, and the suite is so far quite happy.
[02:50] <infinity> I assume the gcc summary should be half failures if -m32 was broken?
[02:51] <infinity> 34629 expected passes, 2 unexpected failures, 1 unexpected success, 97 expected failures, 28 untested, 433 unsupported.
[02:51] <infinity> Oh, wait.  That was just for unix.  unix/-m32 is still in progress.  But I don't see any failures output there yet.
[02:52] <infinity> I'm going to let the manual build run through, just in case the automagic one dies somewhere, but it looks like it'll probably finish, get signed, and upload without my intervention.
[02:53] <infinity> (Granted, I had a hand earlier in the day in fixing up all the other build-deps, but that was X breakage, not lib32gcc1...)
[02:58] <doko> looks fine
[03:12] <jbailey> fabbione: I'm here now for a few minutes.
[04:31] <doko> infinity: what's the state of the amd64 gcc-4.0 build?
[07:25] <lamont> Moved gcc-3.3_1:3.3.6-7 to upload-breezy
[07:42] <lamont> hrm.. gonna have to bludgeon that upload a bit, since it has 'libstdc++5' in it...
[07:42] <lamont> (and my scripts strip those from the upload list)
[07:42] <lamont> and _finally_ gcc-4.0 is trying to build
[07:45] <lamont> actually, it appears that nothing in gcc-3.3 Depends: libstdc++5.  kewl
[08:04] <doko> lamont: desperate hope
[08:05] <doko> I'll have to do that, because seb128 did upload an API incompatible libcairo :-/
[08:06] <lamont> I have libcairo1_0.5.1-0ubuntu2 - is the build going to b0rk (and I'll just kill it now), or will it succeed?
[08:06] <lamont> and if it'll succeed, will it do so before you upload -2ubuntu3...
[08:07] <doko> the new gcc-4.0 upload will require 0.5.2
[08:07] <lamont> and besides, you want to wait for infinity to finish amd64, so he doesn't have to hurt you...
[08:07] <lamont> and you're sleepy... :)
[08:08] <doko> yep, that's what I'm waiting for
[08:08] <lamont> still rather early in his day for him to be awake
[08:10] <lamont> he was fighting with some underlying libraries as of last night when I went to bed...  dunno where he was before he went to bed.
[08:11] <doko> he said, testsuite was runnign
[08:12] <lamont> so does that mean that you expect to upload gcc-4.0 w/in 8 hours?
[08:16] <doko> if infinity wakes up before, yes.
[08:19] <lamont> but no gcc-3.3 upload for an hour or 2, ok???
[08:19] <lamont> :-)
[08:32] <doko> no, no 3.3 anymore
[09:31] <lamont> doko: looking to see what c++ apps I can build on hppa... do you have a list of the build-depended c++libs for each c++app?
[09:32] <doko> lamont: no
[09:33] <doko> looking at debian-devel, mfurr did post somethings like this, didn't look, if he was sorting it by lib-dependencies.
[10:02] <lamont> doko: what failure mode would we expect if the apps were built before the libs were?
[10:08] <doko> libgcj doesn't build, so you would have to disable java
[10:42] <lamont> doko: it occurred to me that all of hppa is built after the transition, so if I can meet build-deps with pure breezy, it's OK to build.