/srv/irclogs.ubuntu.com/2012/09/05/#ubuntu-x.txt

RAOFOh, fun. mesa FTBFS with -j5.04:16
tjaaltonyeah it can have those issues.. though haven't seen it fail in quite a while04:17
tjaaltonRAOF: building the unreleased git version?04:17
RAOFtjaalton: Both that and the archive version fail in the same way.04:18
tjaaltonah04:18
RAOFTrying to generate radeonsi source files.04:18
tjaaltonhuh, built it several times yesterday and they were all fine (-j8 though)04:19
RAOFtjaalton: Hah. And now git fails to build in the final stage because glu.h exists in debian/tmp. What?!04:45
RAOFOh, that reminds me. Who wants to review apitrace? It's ready in git, I think we should have it in Ubuntu, and we don't need to wait for Debian. jcristau's obviously too busy to review it there.04:46
tjaaltonRAOF: that's.. weird, don't see it here04:49
tjaaltonI can have a look at apitrace today04:49
tjaaltonhmm, wonder why cloning it drops it in a nonexistant master branch04:52
tjaaltonoh right04:53
tjaaltonsince there's no debian-unstable branch :)04:53
tjaaltonwhich is where HEAD is pointing at04:53
RAOFHeh.04:58
tjaaltonwhich one should be fixed, the remote HEAD or the branch name(s)?05:20
tjaaltons/fixed/changed/05:20
RAOFtjaalton: Are you talking about apitrace or mesa here?05:22
tjaaltonapitrace?05:22
tjaaltonuh05:23
tjaalton-?05:23
tjaaltonthe convention has been to add -unstable(/-experimental) there, that's why setup-repository made apitrace.git/HEAD look like that05:24
RAOFAh, sorry.05:25
tjaaltonno worries, easy to change either way :)05:26
RAOFAh.05:26
RAOFRight. And I didn't notice because I don't clone that!05:26
tjaaltonexactly05:26
tjaaltonbeen there, done that..05:26
RAOFI guess it should follow pkg-xorg convention and be named debian-unstable.05:26
tjaaltonok, so I renamed refs/heads/{debian,upstream}05:28
tjaaltonnow when you fetch you have both the old and new branches under as origin/*05:29
tjaalton-under05:29
RAOFCool05:29
tjaaltonjust reset the local branch to follow the correct one05:29
tjaaltondunno what is the best way to remove the old remote branches05:31
tjaaltonah, git branch -r -d origin/foo05:32
RAOFAlready done that.05:32
tjaaltoncool :)05:32
tjaaltonso we're set05:32
RAOFtjaalton: Mesa build confusion resolved! I was gentarball-ing with a seriously out of date upstream-experimental.06:18
tjaaltonhah :)06:19
tjaaltonthat would explain glu.h, yes06:19
RAOFYeah, I was wondering about that, given the damn thing has been deleted from the mesa tree!06:22
dupondjenext Xorg version will support Optimus?07:13
dupondjeAny roadmap available? thats for 13.04 ?07:13
ml|vacationdupondje: kernel support is missing07:16
dupondjewhat do we need exactly for it in kernel?07:16
ml|vacationsure you can share buffers, but for it to be actually useful you need to synchronize them to07:17
ml|vacationtoo*07:18
dupondjeOptimus seems to be a mess to handle :(07:18
dupondjethings on the roadmap for it in kernel?07:18
ml|vacationso that is what I've been working on, kernel tree is somewhere but i wont restart until next monday :p07:19
tjaaltondupondje: hm, I thought I asked if it was a hybrid setup.. noticed the bug discussion now07:19
dupondjeml|vacation: its planned for 13.04 or will it still make in 12.10 ?07:21
tjaalton13.0407:22
dupondjeor xorg-edgers ^^07:23
tjaaltonwell the userspace bits are nearly there07:24
tjaaltonso you can just use a mainline kernel in addition to 12.1007:24
dupondjecan I follow the status somewhere? So I'm up-to-date ? :)07:25
tjaaltonhttps://blueprints.launchpad.net/ubuntu/+spec/desktop-q-hybrid-graphics probably07:25
tjaaltonit's missing some detail though07:25
dupondjesubscribed :)07:30
tjaaltonRAOF: apitrace doesn't have a gentarball target, shouldn't it?07:38
ml|vacationdupondje: I'll try to get the upstream pieces inside quantal, but it's less useful without the kernel parts07:38
dupondjeusing a daily kernel isnt an issue :)07:40
RAOFtjaalton: It's got a pristine-tar branch07:42
RAOFtjaalton: I guess it could also have a gentarball target, though.07:44
tjaaltonRAOF: oh right, it does07:48
RAOFgit-buildpackage. It's pretty good :)07:49
tjaaltondebian/gbp.conf needs an update :)07:50
tjaaltonoh you fixed it07:51
tjaaltonhmm, why is the lintian warning suppressed? for backports?07:53
RAOFHm. Does that no longer apply?07:55
RAOFIt's there because when I first packaged it that warning showed up; is debhelper 9 stable now?07:55
tjaaltonyeah07:56
tjaaltonso you can bump it to >= 907:57
dupondjenouveau bug followup is sweet :)07:59
dupondjein less then 24h issue is closed :)07:59
dupondjethe reason we love open source!07:59
tjaaltonI should probably forward some..08:00
tjaaltonRAOF: policy is at 3.9.308:00
dupondjetry it out :) Ben does a great job!08:00
tjaaltonI should probably try quantal on my hybrid t420s soon..08:01
tjaaltonprecise finally getting stable on it :)08:02
dupondjethen its getting boring ;)08:02
tjaaltondang, didn't forward the compiz gpu-hungness yet08:02
tjaaltonRAOF: don't know how to run the beast, but it builds and looks clean, so I'm ok with uploading it to quantal08:04
RAOFtjaalton: Install the frontend package and you can run qapitrace (it'll be in the dash), which allows you to trace stuff, or you can run “apitrace trace $app” to generate a trace.08:05
RAOFBut given that I've built it in a clean chroot here, and it works, you probably don't have to do much testing :)08:05
tjaaltonRAOF: yeah, I'll give it a whirl at some point08:06
tjaaltonoh right, quantal laptop still gpu-hung, hehe..08:07
RAOFWith the multiarch fixes in there you should even be able to trace wine apps reasonably!08:07
tjaaltonhope I never need to go down that path..08:08
tjaaltonbut actually, would this help with the gpu hang I have? or is it more for debugging performance issues?08:09
tjaaltonah, read the description now :)08:09
RAOFYeah.08:10
RAOFIf you have some app that triggers the hang you could trace that, then the trace can be replayed on a dev box.08:11
RAOF...which should reproduce the hang, making it easy to test :)08:12
tjaaltonoh cool08:15
tjaaltonI'll try it out then08:15
tjaaltonhmm, should have a way to reproduce it more quickly08:45
dupondje[ 9560.052783] [drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung19:15
dupondje[ 9560.052788] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state19:15
dupondjeoutch19:16
tjaaltondupondje: so file a bug and attach it there19:25
tjaaltoni915_error_state19:25
dupondjeon what package?19:32
tjaaltonubuntu-bug xorg19:34
dupondjehttps://bugs.launchpad.net/ubuntu/+source/xorg/+bug/104650519:39
ubottuLaunchpad bug 1046505 in xorg (Ubuntu) "i915_hangcheck_hung" [Undecided,New]19:39
dupondjethere :)19:39
tjaaltonthanks19:50
tjaaltondupondje: can you reproduce it?19:59
dupondjetjaalton: nope, unable to reproduce it20:44
dupondjenever had it before afaik20:44

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!