doko | jbailey, lamont: gcj-4.1 on dapper shows 500 test failures in libjava, not seen in current unstable. | 01:13 |
---|---|---|
doko | see chinstrap:~doko/gcj-4.1* | 01:13 |
doko | nearly all interpreted test cases fail | 01:14 |
=== doko [n=doko@dslb-088-073-092-101.pools.arcor-ip.net] has joined #ubuntu-toolchain | ||
=== jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-toolchain | ||
doko | jbailey: in which ways differ the stack unwinding bits in glibc in unstable and dapper on hppa? | 04:55 |
=== doko wonders why lamont is not grumbling anymore ... | ||
jbailey | doko: Eh? No idea off hand. | 05:40 |
jbailey | AFAIK, everything's sync'd now. | 05:40 |
lamont | jbailey: so what did doko break in gcj-4.1 then? :-) | 05:42 |
doko | just verified, that the same source works on unstable, but not dapper. so either lamont did touch the kernel, or jbailey glibc ;-) | 05:44 |
lamont | hrm.. could signals compatibility be part of it? | 05:45 |
doko | lamont: maybe, the stacktrace I did send, started somewhere in the garbage collector | 05:54 |
lamont | hrm | 07:32 |
doko | jda cannot reproduce it either, but maybe my chroot is corrupt | 07:37 |
lamont | your dapper chroot, or your sid chroot? | 07:38 |
doko | dapper | 07:38 |
lamont | it's possible that dapper will debootstrap from ports.u.c right now... | 07:41 |
infinity | lamont: Well, the chroots are completely built with packages from ports.u.c now, if that's any consolation. debootstrap *should* work. | 07:52 |
lamont | infinity: cool | 07:54 |
lamont | yeah - it was more a question of the transient nature of things while it's playing catchup | 07:54 |
infinity | lamont: No, I meant I *just* rebuilt the hppa chroot with ports.u.c, like, an hour ago, so really, it SHOULD work. :) | 07:55 |
lamont | oh, rock. | 07:55 |
infinity | Nothing toolchain-wise is coming from rockhopper anymore, just some -desktop stragglers. | 07:56 |
infinity | I hope to turn off the rockhopper repo completely in a day or two, once GNOME is fully bootstrapped again. | 07:56 |
infinity | (It was a big help, BTW, thanks...) | 07:57 |
lamont | n ice | 08:05 |
lamont | well, when debhelper isn't installable, life just isn't fun | 08:05 |
infinity | Generally not, no. | 08:06 |
infinity | GNOME's bootstrapping issues are just as un-fun, when you're behind too. | 08:06 |
infinity | Took me a week to untangle ia64, so the fact that hppa Just Works (because it was mostly up to date, and just building new versions againast themselves) was a big help. | 08:07 |
infinity | Heck, even powerpc needed untangling, and it wasn't "broken" to start with. Boy, I love new GNOME versions. :) | 08:07 |
doko | jbailey: ping | 08:33 |
doko | jbailey, lamont: it's missing glibc patch | 08:34 |
lamont | doko: coolness. | 08:34 |
infinity | doko: If it's a straighforward patch that's in Debian already, smack it at Malone, and I can apply it. | 08:42 |
infinity | doko: In theory, jbailey's backing off glibc maintenance... In theory. :) | 08:42 |
doko | infinity: currently building on hppa ... | 08:43 |
doko | wondering why jbailey did rename all the debian patches ... | 08:46 |
infinity | Patch reorg, meant to get pushed back to Debian, hasn't been yet. I need t odiscuss it with some Debina folk. | 08:48 |
infinity | Debian, too. | 08:48 |
doko | lamont: how can I ask about libunwind on ia64? | 09:12 |
doko | s/how/who/ | 09:12 |
doko | should stop for today ... | 09:12 |
lamont | doko: can you tell me which patch, and I'll upload a new glibc (or even check with jbailey...) | 09:45 |
lamont | libunwind... hrm... | 09:45 |
doko | lamont: chinstrap:~doko | 09:46 |
doko | eh-frame-pointer | 09:47 |
lamont | doko: uh... ENOFILE | 09:55 |
doko | on chinstrap? | 09:57 |
lamont | doko: yeah - was looking on chinstrap in ~doko | 10:24 |
lamont | no file named 'eh-frame-pointer' | 10:25 |
lamont | nor anything named *-frame-pointer* anywhere under ~doko | 10:25 |
jbailey | infinity: They've basically agreed to it. I need to switch us to quilt first. | 10:25 |
doko | lamont: glibc* | 10:25 |
lamont | ah, the diff of glibc. got it | 10:26 |
lamont | doko: and that diff/dsc are uploadable? | 10:27 |
lamont | rather, they have that fix, and just that fix? | 10:27 |
lamont | and we know it'll fix things. | 10:27 |
lamont | ? | 10:27 |
doko | lamont: my build is still running, currently rebooting m yi386 machine witht this glibc, please don't be IMPATIENT! | 10:29 |
lamont | ah, ok. | 10:29 |
lamont | doko: I'll let you/jbailey work it out, no hurry here | 10:30 |
doko | yes, I will ask jbailey before uploading ... | 10:31 |
jbailey | doko: Yeah yeah sure. Just don't break other things ;) | 10:31 |
jbailey | My pending glibc uploads are all for various -updates. | 10:31 |
jbailey | Thinking of which, they've probably finished by now. | 10:31 |
doko | btw: lamont: | 10:31 |
doko | <doko> lamont: who can I ask about libunwind on ia64? | 10:32 |
jbailey | So hmm. | 10:32 |
doko | jbailey: for dapper as well? | 10:32 |
jbailey | Should I put out a call for testers of the new glibc for warty/hoary/breezy, or should I do the usual chroot test and upload it if it works? =) | 10:32 |
jbailey | doko: No. Dapper already has the fix. | 10:32 |
lamont | doko: mossberger seems to be the consensus | 10:33 |
doko | ok | 10:34 |
doko | jbailey: who do you ask? there are not many people on the channel? | 10:35 |
jbailey | Who do I ask for what, sorry? | 10:37 |
doko | <jbailey> Should I put out a call for testers of the new glibc for warty/hoary/breezy, or should I do the usual chroot test and upload it if it works? =) | 10:38 |
jbailey | Oh, I see. | 10:39 |
doko | or maybe that was a rethoric question | 10:39 |
jbailey | More talking to myself but announcing it in case someone had an opinion. | 10:39 |
=== lamont considers saying "just don't break other things :-)" | ||
jbailey | Well, that's more what I'm worried about here. | 10:41 |
jbailey | Since it's a timezone update and we haven't done one of those before, I can't guess what it might break. | 10:41 |
jbailey | It *should* be nothing. | 10:41 |
doko | ok, i386 did survive the eh-framepointer reboot | 10:41 |
lamont | oh. that fix | 10:43 |
lamont | xgcc: Internal error: Segmentation fault (program as) | 10:45 |
lamont | that'd be breezy-security's gcc-3.4 on hppa | 10:45 |
doko | lamont: try to configure dash as sh on hppa and see some nice segfaults building glibc | 10:47 |
lamont | I see | 10:51 |
lamont | the normal medical suggestion in such a situation is "don't do that, then" | 10:52 |
lamont | doko: re libunwind - have you tried pinging the pkg maintainer/ | 11:02 |
lamont | >? | 11:02 |
doko | hmm, no, it's more the gcc related patches that fc and opensuse has | 11:05 |
lamont | still, I think the debian maintainer is probably well connected to upstream... | 11:05 |
lamont | and therefore a reasonable resource to involve in the search | 11:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!