=== koolhead17 is now known as koolhead17|zzZZ === dendro-afk is now known as dendrobates === aaron is now known as Guest7801 === Guest7801 is now known as albrigha [00:26] barry: I see bug #911124 is assigned to you; did you sync sphinx from Debian (pulling python-support back into main ;)? [00:26] Launchpad bug 911124 in sphinx (Ubuntu) "Please sync python-sphinx 1.1.2 from Debian experimental" [Wishlist,Confirmed] https://launchpad.net/bugs/911124 [00:26] slangasek: i'm on it already :) [00:26] ok :-) === dendrobates is now known as dendro-afk === JackyAlcine is now known as ThisCode === ThisCode is now known as JackyAlcine [04:36] Good morning [04:38] tomreyn: no, it's in: update-manager | 1:0.152.25.6 | oneiric-proposed | source, all [04:48] barry, SpamapS: still here? [04:50] morning pitti === koolhead17|zzZZ is now known as koolhead17 [05:36] pitti: morning! [05:36] hey mbiebl [05:36] mbiebl: another early riser :) [05:37] you've been up since an hour ago :-) [05:37] pitti: just had a look a the pygobject bug [05:37] looks like the upload wasn't bug-fix only [05:37] http://git.gnome.org/browse/pygobject/commit/?id=ee62df4d2fc0cc63c2f29d3ad9b47b875dbd5f89 [05:37] this requires glib 2.31 [05:38] oh, how's that? [05:38] I'm wondering how it built... [05:39] GValue isn't exactly new? [05:39] pitti: the _schar functions were added after .30 [05:39] /build/buildd-pygobject_3.1.0-1-i386-WbEc0E/pygobject-3.1.0/gi/pygi-property.c:124:13: warning: implicit declaration of function 'g_value_get_schar' [-Wimplicit-function-declaration] [05:39] right, that's the one [05:39] mbiebl: indeed; I'll revert that bit then, it doesn't work for chars yet anyway [05:40] mbiebl: thanks for spotting [05:40] np [05:41] looks like python on the kbsd buildds is borked [05:41] at least the test-suite runs into a timeout [05:41] the same for pygtk which I uploaded yesterday [05:41] I want to backport some more fixes from trunk anyway [05:41] yeah, saw :/ [05:42] mbiebl: still curious how that linked.. [05:43] I first thought you had built in on precise [05:43] when I realised that the other buildds had built it, too [05:43] hm, it's a loadable module [05:44] mbiebl: so, sorry for that blunder; will fix right away [05:44] nah, np [05:50] pitti: #659184 is the corresponding bug report [05:50] ah, thanks; I just had a look, seems it's not in the PTS yet [06:08] mbiebl: uploaded [06:12] pitti: thanks [07:56] jpds: ping. [07:56] jpds: would you be interested in getting ubumirror into debian? [08:18] jibel: I believe the current alternate/server failure galore is due to a missing linux-meta upload for -15; I can't prove it, though, there are no useful logs [08:18] apw, smb`: ^ I can haz -meta for -15? [08:19] I'd upload it myself, but that would disrupt your git [08:19] pitti, good morning. right kernel version change. [08:20] pitti, colin uploaded d-i 20101020ubuntu107 [08:20] yes, and the seeds are fine, too; just linux-meta missing === smb` is now known as smb [08:21] jibel: seems oneiric-universe is again due to bug 917153 [08:21] Launchpad bug 917153 in libreoffice (Ubuntu Precise) "failed to upgrade from oneiric to precise: /usr/lib/libreoffice/program/unopkg.bin: error while loading shared libraries: libicule.so.48: cannot open shared object file: No such file or directory" [High,Triaged] https://launchpad.net/bugs/917153 [08:21] * smb looks slightly sleepy [08:21] wassa? [08:22] * pitti hands smb a cup of steaming coffee [08:22] pitti, Thanks, got one in front of me... now I just need to drink it :) [08:23] pitti, re kernel version, i386 passed, isn't it just a matter of respining amd64 to pull latest packages ? [08:23] smb: do you think it's responsible to do a changelog-only package upload at this hour? :-) [08:23] jibel: no, it didn't? https://jenkins.qa.ubuntu.com/view/Precise/job/precise-upgrade-oneiric-universe/lastFailedBuild/ARCH=i386,LTS=non-lts,PROFILE=universe,label=upgrade-test/ [08:23] pitti, I meant alternate/server install failure [08:23] jibel: oh, sorry, wrong failure [08:24] pitti, Certainly not. Not sure _when_ I did the last upload of a kernel package [08:24] jibel: hm, curious how it passed -- it shouldn't [08:24] Eh I man I can try again :) [08:24] pitti, packages are not the same version on both arch [08:24] If I finally am able to read correctly [08:24] jibel: I figure if we respin now, i386 will fail as well [08:25] jibel: I think I'll wait on smb's -meta upload, and respin when that published [08:25] pitti, k [08:27] jibel: ah, seems lucid->precise got a little further after the qtdbus fix; at least we now have two errors in one log, yay for slight parallelization :) [08:27] :) [08:31] jibel: are you interested in tracking the bugs (just filed 929381, for example), or don't you care? [08:34] pitti, Now lets see how much worth that coffee was... Think I uploaded some meta [08:34] smb: AFAIK it's just bumping the version in changelog, no other changes [08:35] pitti, It is. And there was something already in git with that [08:35] smb: but as it's in git, us mere mortals can't easily do that [08:35] Probably just held back until the compile [08:35] dholbach, ping [08:35] smb: ah, uploaded in ogasawara's name :) thanks [08:35] diwic, pong [08:36] pitti, Yeah, did all the real work anyway :) [08:36] dholbach, two things: First, I vaguely remember me signing up for "writing a launchpad script that summarises ppa owners" [08:37] pitti, I am interested if they are found with automated tests. [08:37] jibel: yes, both are [08:37] (that's where I got them from) [08:37] dholbach, the idea was to contact ppa owners to see if they wanted to step up and maintain the official thing [08:37] jibel: they seem to be at least partially fallout from the kvm envirionment [08:37] jibel: bug 929381, bug 929382 [08:37] Launchpad bug 929381 in cgroup-lite (Ubuntu Precise) "init script fails to start" [High,New] https://launchpad.net/bugs/929381 [08:37] Launchpad bug 929382 in lxc (Ubuntu Precise) "package fails to install: SIOCSIFNETMASK: Cannot assign requested address" [High,New] https://launchpad.net/bugs/929382 [08:38] cgroup-lite installs fine on my real workstation, tryign in kvm now [08:38] dholbach, but I'm overwhelmed with audio bugs, and it feels more efficient if I do the audio bugs and someone familiar with launchpad hacking would do that script [08:38] dholbach, and I can't even find the blueprint. Do you remember which one it was? [08:40] dholbach_, did you get all of that? === dholbach_ is now known as dholbach [08:40] diwic, pong [08:40] dholbach, repeating in PM [08:40] no, I'm sorry [08:40] ok great [08:41] jibel: hm, is /proc/cgroups nonempty in the environment of the upgrader? [08:42] jibel: oh, I figure it's running under a lucid kernel; I'll try reproducing that [08:45] pitti, looking [08:45] jibel: hold on for now, I'll debug this locally first [08:45] pitti, k [08:49] pitti, Just from my random playing around with things. Usually when you looked at lxc you were told in the past that you need to mount cgroups somewhere. So I would not find it completely unexpected that some prople will have it in the fstab already. [08:49] smb: ah, so perhaps lxc failed because cgroup-lite failed [08:50] pitti, Might be. I remember having seen something (if I could remember the symptoms) coming from automount when in precise cgroups gets mounted before and you still have it in fstabs [08:51] Not sure that was detected in the installer... [08:51] I mean the problem [08:52] Think the upgrade went ok, but on reboot there was something wrong until I cleaned up fstab [08:52] right, I can reproduce on a lucid live system [08:53] Hi, are there plans to switch thunderbird to the rapid release cycle too, so that we have thunderbird 10, 11 etc in Lucid? Or that was only for firefox? [08:54] I think it's the plan [08:56] Thank you :) [08:56] micahg or chrisccoulson will know better [09:11] jibel: hm, I cannot reproduce bug 929382 in a lucid live system [09:11] Launchpad bug 929382 in lxc (Ubuntu Precise) "package fails to install: SIOCSIFNETMASK: Cannot assign requested address" [High,New] https://launchpad.net/bugs/929382 [09:12] smb, did you do the meta upload? else will do it shortly [09:13] apw: yes, it's in [09:13] apw, done [09:13] cool [09:13] https://launchpad.net/ubuntu/+source/linux-meta/3.2.0.15.15 [09:15] pitti: d-i failed to build on amd64 which didn't help. I've just uploaded a fix. I haven't looked at the logs but I doubt linux-meta had much to do with it. [09:16] cjwatson: ah, ok; but it should have independently failed due to the API mismatch, I guess? [09:17] that's exactly what the failure was [09:17] cjwatson: thanks [09:17] arguably the CD build should have failed but it would still have been a failure :) [09:17] not sure why that didn't happen, no time to look now [09:19] cjwatson: np [09:24] join #evince === yofel_ is now known as yofel [10:12] how to remove seat once created by lightdm? [10:27] infinity: it seems https://launchpad.net/ubuntu/+source/libnih/1.0.3-4ubuntu7/+build/3197190 is stuck (didn't change in over an hour) [10:27] could that be killed, and we'll try it again? [10:46] cjwatson, jibel: d-i and linux-meta are published, rebuilding alternate/server === _salem is now known as salem_ === tkamppeter_ is now known as tkamppeter [11:16] pitti: Stuck how? Are you sure you're not just being impatient? :) [11:17] infinity: "stuck" in the sense of https://launchpad.net/ubuntu/+source/libnih/1.0.3-4ubuntu7/+build/3197190 not changing for two hours [11:17] pitti: It historically takes about 4.5 hours on a machine of that class. I wouldn't complain until it passes that. [11:17] the build should take 1:15 in total [11:17] (the last build did, anyway) [11:17] infinity: but 2 hours to compile a single .c file? [11:17] pitti: The last build was probably on a faster machine... [11:17] infinity: well, if that's "normal", I can certainly just wait further [11:18] I'm willing to believe it's broken, but there are also some things that, yes, will take hours on a single file. [11:18] So, until it passes 5ish... [11:18] https://launchpad.net/ubuntu/+source/libnih/1.0.3-4ubuntu2/+build/2571092 <-- 4.5h [11:19] ack [11:19] infinity: thanks! [11:19] pitti: The good news is that Very Soon Now, all those old machines are going away. [11:44] mvo, hi, did you ever see ? [11:51] Hi all [11:51] I make a design for let users to choose want to have full screen with luncher or without it. where should i send it ? [12:00] stgraber: I would appreciate if you could take a look at bug 929382 and see whether my proposal would be acceptable [12:00] Launchpad bug 929382 in lxc (Ubuntu Precise) "package fails to install: SIOCSIFNETMASK: Cannot assign requested address" [High,New] https://launchpad.net/bugs/929382 [12:00] hallyn: ^ you, too, as the last uploader [12:01] if it is, I'd like to do an upload [12:14] mpt: hello! no I didn't, but I'm also not involved with apt-mirror [12:15] oh, ok [12:27] infinity: ah, done now \o/ [13:00] sh: 1: /usr/bin/gdbus: not found === MacSlow is now known as MacSlow|lunch [13:00] lamont, install glib? [13:01] seb128: I'm more curious as to why the natural order of things didn't install it for me... [13:01] current precise, mythbuntu box [13:01] lamont, what do you run to get that error? [13:01] apt-get update [13:02] Hit http://us.archive.ubuntu.com precise-updates/universe Translation-en [13:02] sh: 1: /usr/bin/gdbus: not found [13:02] Reading package lists... Done [13:02] to place it in context [13:03] which appears to be apt.conf.d/20packagekit just blindly doing its thing [13:05] I shall file the packagekit bug [13:05] yeah [13:05] what an idea to use that ;-) [13:07] yeah, hopefully it's a more valid bug that my last one. :( [13:07] E: Internal Error, No file name for libc6 [13:08] uh... wut? [13:08] (apt is mad about libc6 deps, apt-get -f install fails as above) === Ursinha` is now known as Ursinha === dendro-afk is now known as dendrobates [13:44] hi === MacSlow|lunch is now known as MacSlow [14:00] I am actually editing a page of the wiki containing infos for build requests for Openbox 3.5.0. I will now add imlib2 devel which is an optional lib. [14:00] could someone tell me what is the right name for this package at Ubuntu ? [14:01] http://openbox.org/wiki/Help:Installing [14:02] imlib2 devel when used as a build request allows getting icons in the openbox menus (since 3.5.0) such as here: [14:02] http://i.minus.com/ipNlKuTlCJr25.png [14:02] this is why I am adding it to this page now. [14:03] cking, ? AaronMT ? do you know about imlib2 devel package name please ? [14:03] Please dont ping random people. [14:04] you just arrived. sorry (oops) [14:04] :) [14:04] are you angry ? [14:05] melodie: libimlib2-dev === davmor2_ is now known as davmor2 [14:05] I thought it was important to let some members of the devel team I will edit this page [14:05] "apt-cache search imlib dev" [14:05] thanks pitti [14:06] -dev is the standard name for development packages in Debian/Ubuntu [14:06] pitti, this is not my distro. I did same at fedora devel chan. just giving info about this edit while looking for the right package name. thanks [14:06] mvo: do you know what's wrong with https://jenkins.qa.ubuntu.com/view/Precise/job/precise-softwarecenter-amd64/ and what it's supposed to test? [14:06] melodie: you can use the same for Debain [14:06] Debian [14:07] pitti: its the software-center testsuite [14:07] melodie: (and Mint, etc.) [14:07] I was just wondering about it : thanks ! [14:07] mvo: ah, like running "make check" in a checked out tree? === dendrobates is now known as dendro-afk [14:07] mvo: is it supposed to fail, or is that due to running in the autotest environment? [14:08] pitti: its the environment I think [14:08] mvo: well, "supposed to fail" is certainly bogus, I meant "fail locally as well" [14:08] done : http://openbox.org/wiki/Help:Installing#Dependencies_in_Ubuntu_and_Debian [14:08] pitti, it does a make check [14:08] have a nice day, bye now ! [14:09] pitti, and probably failing because a test requires and external resources blocked by a fw [14:09] https://jenkins.qa.ubuntu.com/view/Precise/job/precise-softwarecenter-amd64/45/console is a bit of a messy read [14:09] Unable to find the server at recommender.ubuntu.com [14:09] yeah, seems like it [14:10] DatabaseOpeningError: Couldn\'t stat \'/var/lib/apt-xapian-index/inde [14:10] x [14:10] that seems solvable, though [14:11] pitti: yeah === dendro-afk is now known as dendrobates [14:21] hum [14:21] https://bugs.launchpad.net/bugs/929384 [14:21] Ubuntu bug 929384 in nux (Ubuntu) "unity_support_test crashed with SIGSEGV" [High,Confirmed] [14:21] the libc update broken nvidia drivers [14:22] "Happens here too, when running the nvidia-current driver from precise. [14:22] It also makes gnome-shell crash on startup. [14:22] this is caused by the eglibc update (2.13-24ubuntu4 -> 2.15~pre6-0ubuntu10). I reverted it, it's fine now." [14:26] seb128: I haven't updated the nvidia driver for a while now so eglibc is more likely to be the culprit [14:26] pitti: that (bug 929382) sounds fine [14:26] Launchpad bug 929382 in lxc (Ubuntu Precise) "package fails to install: SIOCSIFNETMASK: Cannot assign requested address" [High,Triaged] https://launchpad.net/bugs/929382 [14:27] pitti: i think it's what i meant to do [14:28] tseliot, right, I just wonder if the nvidia drivers are doing something weird that lead to issues with the new libc [14:28] or if that's a bug with libc itself [14:28] cjwatson, slangasek, pitti, skaet: ^ nvidia drivers broken by the libc update today [14:28] that let quite some users without a working computer [14:28] not sure what to do about it but I think it should be raised [14:28] that's on precise [14:29] seb128: maybe we should revert the upload and then I can discuss this issue with Nvidia [14:30] tseliot, not easy as fabian wrote [14:30] "(but reverting this libc6 is not easy as more packages are rebuild for it, leading to failing GLIBC_2.15 version checks everywhere)" [14:30] seb128: that was already with the 2.15~pre version, I take it? [14:30] because 2.15 final is still building [14:31] pitti, yes === elif is now known as elif_lunch [14:34] Is there a way to run sbuild but with an extra local repository added to resolve build dependencies from? === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [14:41] rickspencer3, is there any official way to raise precise brekages? [14:43] seb128, I'm not sure what you mean [14:43] rickspencer3, the libc update from today break nvidia binary drivers and let precise user without a working machine [14:44] rickspencer3, I mentioned it on the channel but nobody seems to be around [14:44] rickspencer3, is there any official way to flag that we have an issue and should deal with it? [14:44] seb128, can you raise it with the +1 maintenance team? [14:44] rickspencer3, should that be broadcasted in some way so users don't upgrade? [14:44] oh [14:44] rickspencer3, where,who,how? [14:45] is barry barry warsaw? [14:45] rickspencer3: the one and only [14:45] seb128, also, can you tell skaet, she can use the twitter/identica account to broadcast a "do not upgrade" message [14:46] rickspencer3, I did a 25 minutes ago, waiting for her to be around or reply [14:48] can you get it blocked on the mirrors? [14:48] Laney, I would like to know ;-) [14:48] heh [14:49] but seems none of the rt people are around atm [14:49] #-release informs me that skaet might be in mumble [14:49] go gatecrash :P [14:49] seb128, rickspencer3 - on a call on 10.04.4, reading backscroll. Will get on it. [14:50] skaet, thanks [14:52] I'm not sure what we can do about libc - as noted, reverting it would cause widespread chaos [14:52] do we know precisely why nvidia breaks? [14:52] (getting it blocked on the mirrors would be very bad too IMO) [14:53] cjwatson, no, just limited to what is in the bug [14:53] I don't have an nvidia card myself [14:53] but olli and some other people on IRC ran into the bug [14:54] there doesn't seem to be a useful stacktrace in the bug [14:54] no :-( [14:54] yeah, we already rebuilt libnih on the new libc and then uploaded a new upstart, so reverting it would likely break most if not all existing systems. [14:54] I'm not sure we have debug symbols for nvidia binary drivers [14:56] I suspect it needs somebody with an affected system who's competent to debug it interactively, at least as far as diagnosing e.g. an affected libc symbol or something [14:56] tseliot, ^ [14:56] tseliot, do you have any precise nvidia box, any chance you could look at it? [14:56] hey. i'm in need of some help. bug 929523. [14:56] Launchpad bug 929523 in bacula (Ubuntu) "bacula-director does not start, dummy libbaccats " [Undecided,New] https://launchpad.net/bugs/929523 [14:56] if we're lucky we just need to rebuild something [14:57] I have an nvidia desktop box that wasn't updated yet, running precise 32bits [14:57] mvo: the point release process says "Ping mvo to update meta-release file on http://changelogs.ubuntu.com/" [14:57] seb128: yes, I do and I'll have a look at it [14:57] mvo: this is "release - 6 days", which would be tomorrow [15:02] pitti: hey, can we chat sometime today about +1? [15:02] seb128: unfortunately my 1st machine with nvidia died (I don't know what hardware issue that is but I can't even access the BIOS). Fortunately I also have a laptop that I can use but it will probably take a while... [15:02] pitti: ok, I guess ideally this should be done on the day of the release instead, its just chaning the number from 10.04.x to x+1 [15:03] sorry.. back with a bit more information on that bug now. (in the last comment). [15:03] i'm running into dh_shlibdeps complaining but it seems that bacula is doing what it intends to do. [15:03] mvo: ah, thanks [15:03] mvo: I'll update the wiki page [15:03] cjwatson, I have an un updated nvidia box, if you want me to break it [15:05] pitti: thanks! === dendrobates is now known as dendro-afk [15:08] barry: In a meeting, and need to run out in ~ 20 mins, but yes, let's have a quick chat in a sec [15:08] pitti: cool. i'm over in #ubuntu+1-maint === dendro-afk is now known as dendrobates === elif_lunch is now known as elif [15:13] could an archive admin give gst-plugins-bad a push, it's in NEW due to libs&dev being split [15:15] tjaalton: done [15:15] pitti: thanks! now on to gstreamer-vaapi.. === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [15:36] Is there a way to run sbuild but with an extra local repository added to resolve build dependencies from? [15:39] rbasak: when I've had to do that I resorted to editing sources.list in the source schroot and doing the build, then reverting the change once done. [15:40] rbasak: I have no idea how it would be done otherwise -- anyway, sbuild doesn't seem to have a switch to do it. [15:41] Thanks cyphermox - maybe I'll patch sbuild to add a switch one day, but I'll use your hack in the meantime [15:41] cyphermox: did you use an http repo somewhere, or a local file one? And if a local file one, did you add that to the source chroot or something? [15:41] it might be useful, if say you want to enable universe for just one build, test builds against a particular PPA, etc. [15:42] I'm trying to do a test rebuild of a dependent package - it fails in a PPA, but I can't figure out why. [15:42] rbasak: I was using reprepro to manage a local repository hosted in on my system with apache, in ~/public_html/ [15:42] (it succeeded for me doing it manually in a chroot) [15:42] rbasak, sbuild --pre-build-commands [15:42] I wonder if it cannot add a new apt sources file in / [15:43] janimo, the part I don't understand is how to get my local repository into the chroot [15:43] rbasak, ah the files themselves from a local filesystem? [15:43] yes [15:44] rbasak: with that --pre-build-commands I guess you could echo a file in /etc/apt/sources.list.d for your local repo [15:44] chroot-setup-commands too [15:44] I have a pile of debs. Usually I "dpkg-scanpackages . /dev/null|gzip -9>Packages.gz" and then add "file:///path/to/dir /" in sources.list [15:44] not sure how to copy files in there though [15:44] rbasak: oh, that's why I was using a local apache [15:44] yeah that's the bit where I'm stuck :) [15:45] * janimo used sbuild a week ago for the first time === dendrobates is now known as dendro-afk [15:45] otherwise if you use the security team's sbuild howto to bind mount /home you could leave the files there? [15:45] Ah, a bind mount would be awesome [15:46] rbasak: https://wiki.ubuntu.com/SecurityTeam/BuildEnvironment ; point number 5 [15:50] thanks cypermox - but what are ddebs? [15:52] debugging symbols packages === dendro-afk is now known as dendrobates [15:53] I see - so irrelevant to me? [15:58] rbasak: probably [16:03] did the google talk plugin break in precise for anyone else? [16:06] brendand: intermittently and possibly related to the eglibc upgrade today [16:07] slangasek, audio works for first few seconds and then stops [16:07] yes, that's the issue I've seen [16:07] did anyone Pilot today? [16:07] brendand: however, I wasn't able to reproduce it again after shuffling packages, so I haven't raised a bug [16:07] slangasek, what did you do? [16:08] brendand: I downgraded eglibc, then upgraded it again, and then I couldn't reproduce the bug... so it may have actually been a bug somewhere in the audio stack [16:09] slangasek, what's the older version? [16:09] brendand: 2.13- [16:14] slangasek, by eglibc you mean libc-bin, libc6, or all of them? [16:15] slangasek, brendand: maybe that's the bug which was in yesterday's alsa update and fixed today? [16:15] the configs were installed in the wrong dir [16:15] TheMuso fixed that today [16:16] seb128, nope i applied the workaround today. besides that didn't affect the googletalk plugin [16:16] dunno then [16:21] brendand: broken for me too [16:27] seb128: according to ricotz the new Nvidia driver seems to work with the new eglibc. I'll test it here and upload it (if successful) today [16:27] seb128: Sarvatt brought this to my attention [16:29] tseliot, ok, great, thanks [16:31] hi all [16:32] ehy tseliot ! how are you? [16:32] quadrispro: hey, I'm still a little sick but fine otherwise, you? [16:35] tseliot, here the same, the flu has gone away right 2 days ago. I'm not used to this deep cold weather :/ [16:36] quadrispro: yes, I guess that's the problem.. [16:36] brendand: all of them, they have to be upgraded and downgraded in step [16:46] brendand: note that at this point, due to rebuilds against the new version of glibc, you would also have to downgrade cups, bluez, pulseaudio, and libnih1 (used by upstart) === dendrobates is now known as dendro-afk [16:59] slangasek, hmm, I don't have any nvidia hw :-/ [17:00] barry, ping [17:00] doko: I only assigned the eglibc task to you, if that's what you mean [17:00] ahh, ok [17:01] barry ping === deryck is now known as deryck[lunch] === arges is now known as arges_ [17:38] seb128: unfortunately I'm unable to reproduce the problem on my laptop with the new eglibc... === dduffey is now known as dduffey_afk === dduffey_afk is now known as dduffey [17:43] tseliot: every dupe i've opened so far is i386, that may explain why you can't reproduce on amd64 === dduffey is now known as dduffey_afk === sabdfl_ is now known as sabdfl [17:45] Sarvatt: good point, I was starting to suspect i386... [17:49] bdmurray, why did you assign that gnome-screenshot upstream behaviour change to our team? it's not really a bug... [17:49] it's not a regression either, it's a on purpose behaviour change [17:50] seb128: I think its a huge usability issue [17:50] bdmurray, it's not really [17:50] How are people supposed to find their screenshots? [17:51] bdmurray, the new behaviour is how any phone or tablet out there will behave (yeah, they are different devices but still) [17:51] bdmurray, how do they find them on phones and tablets? [17:51] I don't think you take screenshots with a phone rather you take pictures and the photo software shows you the picture [17:52] bdmurray, well android tablets have a screenshot button on their 4 standard buttons which behaves like gnome-screenshot does, sound and visual effect and storing in the images directory [17:54] bdmurray, also the non interactive mode is only on screenshots, if you run gnome-screenshot from unity you will get an ui as before [17:55] seb128: part of assigning the bug to your team was to get an authoritative statement about the bug / workflow. I'd imagine that bug will receive quite a few duplicates and people are interested in this information. [17:56] bdmurray, ok, I'm commenting on it, I just disagree with the regression tagging I think ;-) [17:56] though I'm not sure we will keep the new behaviour I think we should not hurry in reverting but rather see how the new workflow works for users [17:56] I can see how regression- tags can be subjective in design chagnes like this. [17:57] Its still not clear to me how I am supposed to find the screenshot. Where do I navigate to? [17:58] View Photos in the dash? [17:59] bdmurray, they should show in the dash "recent files" on the dash home screen [17:59] bdmurray, if they don't atm we will fix that (need to check if that works) [18:00] bdmurray, the xdg Image folder is also in the bookmarks and easily accessible, if you know screenshots go there it's obvious to find them [18:00] (which is what phone and tablets do as well, you just need to know how to access your image dir) [18:01] bdmurray, in fact that works, they show there [18:11] hey, is it possible to change dch --increment behaviour? it always adds ubuntuX suffix which isn't really what I always want [18:12] kklimonda: what's the use case? [18:12] kklimonda: and welcome back :) [18:13] ah yes, I haven't been around in quite a while.. hey everyone :) [18:13] micahg: I maintain some packages that are not part of Ubuntu - they are either for debian or for a project based on Ubuntu. [18:15] ah, I can pass --distributor to override that [18:15] yeah === deryck[lunch] is now known as deryck [18:15] just found that in the source === azeem_ is now known as azeem [18:18] kklimonda: you can also pass -U, which is shorter [18:18] great, I can even use DEB_VENDOR to override that [18:19] bryceh, can you confirm that an updated glibc will fix the nvidia issue? [18:20] slangasek: hmm, doesn't seem to work [18:20] ah [18:20] doko, Sarvatt and tseliot say there is an nvidia driver update which will resolve it [18:20] sorry, I've misunderstood you [18:20] it works just fine :) [18:20] bryceh: it seems that the latest nvidia (beta) driver solves the problem ^^ [18:21] bryceh: argh [18:21] bryceh: I take back that. Apparently it doesn't [18:21] bryceh: https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/929384/comments/16 [18:21] Ubuntu bug 929384 in nvidia-graphics-drivers (Ubuntu) "nvidia drivers broken by the recent libc update on i386 arch" [High,Confirmed] [18:34] doko, which glibc update? [18:35] bryceh, there's only one, 2.15 [18:35] has anyone managed to get valgrind output? [18:36] cjwatson: do you suspect some major leak on i386? [18:37] probably not a leak, but maybe some memory smashing that valgrind could pick up on [18:37] right [18:37] interesting [18:37] Is https://bugs.archlinux.org/task/27737 possibly related? [18:39] possibly, but I don't see any of those problems here on amd64 [18:39] oh, linked to nvidia [18:39] I don't see the delays with firefox and chromium [18:39] sorry, had a different eglibc issue on the brain :P [18:39] ah, here we go [18:40] https://bugzilla.redhat.com/show_bug.cgi?id=737223 [18:40] bugzilla.redhat.com bug 737223 in glibc "gl applications (using libGL.so) crash with glibc-2.14.90-8 when using proprietary NVIDIA driver" [High,Closed: errata] [18:40] http://sourceware.org/git/?p=glibc.git;a=commitdiff;h=0c95ab64 [18:40] doko: ^- [18:41] cjwatson, the fix is in the package [18:41] oh, already? [18:41] valgrind output: http://pastebin.ubuntu.com/835568/ [18:43] doko: I don't see it ... [18:44] looking in lp:ubuntu/eglibc [18:44] hmm, only part of it: [18:44] /usr/share/sounds/speech-dispatcher/test.wav [18:45] if (copy == NULL) [18:45] _dl_fatal_printf ("out of memory\n"); [18:45] l->l_libname->name = l->l_name = memcpy (copy, dsoname, len); [18:45] } [18:45] I wonder, should this bug be reproducible without nvidia hardware just by installing the lib [18:45] incidentally, lp:ubuntu/eglibc is not up to date with precise [18:45] ahh, I'll commit [18:45] slangasek: if this patch is what fixes it, that looks quite plausible [18:46] it's very similar to that rh bug [18:46] yep, testing now [18:46] interesting LD_DEBUG options [18:46] ahasenack: please try running whatever you were running with 'LD_DEBUG=version' [18:47] it says the option "version" is unknown, I tried help and there are others [18:47] all quite verbose [18:48] ahasenack: er, right, =versions [18:48] standard shell redirection to a file doesn't work [18:48] 2> [18:48] i.e. LD_DEBUG=versions glxinfo 2>x [18:48] oh, duh [18:48] slangasek: some of the bug reporters mentioned it still happened when they switched the gpu to intel but still had nvidia installed, yep it should [18:48] nm [18:50] cjwatson: versions: http://pastebin.ubuntu.com/835583/ [18:50] looks very much like the same bug to me. [18:50] and with LD_DEBUG=all [18:50] http://pastebin.ubuntu.com/835584/ [18:51] cf. https://bugzilla.redhat.com/show_bug.cgi?id=737223#c12 [18:51] bugzilla.redhat.com bug 737223 in glibc "gl applications (using libGL.so) crash with glibc-2.14.90-8 when using proprietary NVIDIA driver" [High,Closed: errata] [18:52] got the symbol lookup error too [18:52] buried in those logs [18:54] slangasek: any luck? [18:54] cjwatson: still working on it, just found that my i386 chroot didn't have eglibc upgraded yet [18:54] yep, got it [18:55] doko: it's reproducible in an i386 chroot with nvidia-current+mesa-utils installed, without needing nvidia hardware [18:56] I'm updating my nvidia hw test box, but will be a few more minutes [18:56] bryceh: reproducible in any i386 chroot [18:56] hey... [18:56] slangasek: did you say you saw audio problems as well? I wonder if that's https://bugzilla.redhat.com/show_bug.cgi?id=769421 [18:56] bugzilla.redhat.com bug 769421 in glibc "NPTL issues with 2.14.90-25 - mostly pulseaudio but others as well" [Unspecified,Closed: errata] [18:56] cf. http://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/glibc&id=ead7c0f93b194f6e3a100f249edd76c826debaef [18:56] i'm looking at bacula, and its not getting correct mysql libs because it is explictly checking (in configure) for certain paths that the lib might exist in [18:57] cjwatson: yes; bug #929713 [18:57] Launchpad bug 929713 in eglibc (Ubuntu Precise) "audio unreliable with eglibc 2.15" [Critical,Confirmed] https://launchpad.net/bugs/929713 [18:57] and that does not include /usr/lib/x86_64-linux-gnu [18:57] configure looks like this: http://paste.ubuntu.com/835598/ [18:57] smoser: doesn't it make sense to patch that in as an include path? [18:57] I have those audio problems too, FWIW [18:57] should i just patch it? [18:57] it seemed wrong to patch a 'configure' [18:58] is the configure script generated using autoconf? [18:58] smoser: are there any .in files? [18:58] if so, you should patch the generating files [18:58] the modern form is .ac [18:58] ah yeah, right [18:59] cjwatson. yes, but i dont think they're in the upstream tarball [18:59] although it might well be in .m4 somewhere [18:59] smoser: crazy upstream [18:59] if they don't ship the autoconf input, you don't have a lot of choice *shrug* [18:59] yes, it's wrong, but they were wrong first [18:59] playground logic [18:59] so cjwatson you were saying 'configure.ac' ? [18:59] yes [19:00] yeah. not there. [19:00] smoser: I don't think you gain much by trying to do it 'right' [19:00] smoser: sure it is, in the autoconf/ subdirectory. [19:00] configure.in [19:00] ah.. /me retreats. [19:01] autoconf/bacula-macros/db.m4 is probably what you want to change. [19:02] cjwatson: the "NPTL issues" backtrace is not dissimilar to my quodlibet backtrace; checking the other apps where I'm seeing it [19:04] cjwatson, thank you. [19:08] how did that NPTL reversion not get into eglibc 2.15 upstream? that bug is almost two months old [19:20] elmo, nvidia might not be as high a priority for them? [19:21] bryceh: nvidia is the rtld.c change - the broken NPTL patch affects audio on machines with no nvidia [19:21] bryceh: the NPTL reversion is not about nvidia [19:22] I don't have an answer for why the Fedora patch fixing nvidia isn't upstream either, mind [19:22] * cjwatson -> pub, sod it [19:23] slangasek: I tentatively added a Fedora bug task for the audio bug [19:23] cjwatson: seen, thanks === bdrung_ is now known as bdrung === salem_ is now known as _salem [19:27] the pthread_cond_wait issue is debian/patches/i386/local-pthread_cond_wait.diff, isn't it? [19:27] cjwatson, ^^^ [19:31] doko: I'm seeing the NPTL issue on amd64 [19:32] me too [19:33] hmm, reverted in Debian: [19:33] * patches/amd64/cvs-pthread_cond_wait.diff: remove as it seems to cause [19:33] some issue with some kernels. Closes: #651746. [19:34] the remaining problem doesn't appear to be kernel-related [19:34] and the proposed fix isn't to remove the assembly version altogether, but to patch it? === aaron_ is now known as albrigha [19:44] doko, I've got 2 systems (one Nvidia, one Intel) reproducing the nvidia bug. Both have libc 2.15 installed. [19:47] slangasek, do you know if somebody has been looking at upgrading valgrind's suppr files for multiarch? I've been adding local ones to clean my logs but I noticed some are in the default.suppr but the paths don't match because of multiarch dirs [19:48] seb128: I haven't heard this mentioned before now, no [19:48] slangasek, hum, ok [19:49] slangasek, i.e /usr/lib/valgrind/default.supp has [19:49] zlib-1.2.x trickyness (1b): See http://www.zlib.net/zlib_faq.html#faq36 [19:49] Memcheck:Cond [19:49] obj:/*lib*/libz.so.1.2.* [19:49] seb128: seems like it should be a simple bugfix though, right? [19:49] slangasek, yeah, it's just adding * to the lib paths in a file :p [19:49] slangasek, I was just wondering if somebody was working on it or if that was at least reported [19:49] slangasek, I will open a bug in the bts if there isn't one yet [19:49] well, I'm not subscribed to the valgrind bugs :) [19:49] ok [19:50] slangasek, let me check [19:50] slangasek, btw I sent you automake1.11 multiarch diff to the bts, it got merged but they credited me even if I mentioned the patch was yours [19:50] slangasek, sorry for stealing credit from you ;-) on the positive side it allowed us to sync automake1.11 again [19:50] seb128: no worries, my name is in enough changelogs :-P [19:51] slangasek: Since we started trimming changelogs, I feel much less cool. [19:51] slangasek: That recursive grep in /usr/share/doc isn't what it used to be. [19:51] :-) === dduffey_afk is now known as dduffey [20:13] bryceh, slangasek, cjwatson, elmo: test packages (amd64) on http://people.canonical.com/~doko/tmp/eglibc-2.15/ [20:23] and i386 === dendro-afk is now known as dendrobates [20:33] doko, looks good. installed those debs on the broken nvidia box, rebooted, and unity came right up [20:34] interestingly, it seems to still be failing in my chroot [20:34] ah, no [20:34] it works, I just had to install libc-bin with it so the package actually got configured [20:34] verified NVIDIA loaded, no X errors, dmesg is clean [20:35] bryceh, amd64, or i386? [20:35] doko, i386 [20:35] doko: glxinfo i386 nvidia chroot test succeeded here [20:35] testing the audio regression now [20:35] bryce@porlock:~$ uname -a [20:35] Linux porlock.bryceharrington.org 3.2.0-15-generic-pae #24-Ubuntu SMP Tue Feb 7 23:30:35 UTC 2012 i686 i686 i386 GNU/Linux [20:35] bryce@porlock:~$ apt-cache policy libc6 [20:35] libc6: [20:35] Installed: 2.15-0ubuntu2 [20:35] Candidate: 2.15-0ubuntu2 [20:36] network sucks here, apt-get update taking now over 15min ... === funkyHat_ is now known as funkyHat [20:38] doko, what's in that 0ubuntu2 version? [20:41] bryceh, the two patches identified by cjwatson [20:49] doko, also verified it fixes the unity_support_test crash on the intel box [20:50] doko: audio is testing out ok [20:50] jdstrand: care to test the updated libc from http://people.canonical.com/~doko/tmp/eglibc-2.15/ with me in G+? [20:51] slangasek: sure. a browser restart should be enough I'm assuming, correct? [20:52] jdstrand: no, the gtalk plugin starts its own server process, which is the one that hangs; so you want 'killall -9 GoogleTalkPlugin' [20:52] jdstrand: possibly with a 'killall plugin-container' thrown in for flavor [20:52] heh, ok. downloading-- it'll probably be a few minutes [20:52] * slangasek nods [20:57] slangasek: ok, ready [20:57] joining [20:58] * jdstrand adjusts hair [20:59] hyperair, Do you have thoughts on this patch? https://code.launchpad.net/~ballogy/libindicate/fix-mono-location/+merge/92313 [20:59] hyperair, I always forget where things are supposed to go with Mono :-) [20:59] jdstrand, do you see core dumps without adjusted hair? ;-P [20:59] worked great :) [21:00] doko: hehe-- maybe that was my problem before :) [21:01] tedg: if you gacutil, don't bother installing it in assemblydir. [21:01] tedg: afaik gacutil does that for you [21:01] er i think [21:02] hyperair, How do I know if I gacutil? ;-) It's used several times, I'm not sure which usages are important. [21:03] tedg: gacutil -i installs into the GAC. [21:03] tedg: if you do that, installing it into assemblydir like what it was before the patch was applied, and even after the patch is applied, is redundant. [21:04] tedg: take a look at taglib-sharp's src/Makefile.am [21:04] hyperair, So the patch is from the arch guy, in theory it helps their packaging. Would you say it's a no-op for Ubuntu then? [21:05] tedg: we'll probably just ignore things found in /usr/lib/mono [21:05] debian policy is to stick things into /usr/lib/cli in the .deb, and then gacutil -i it in postinst [21:05] whereas upstreams should just gacutil -i straight. [21:06] tedg: is ballogy around on irc somewhere? [21:06] tedg: also i really need to sleep. [21:06] hyperair, Not sure, I've not talked with him/her on IRC === fenris_ is now known as Guest25990 [21:07] No IRC listed on LP either. [21:07] tedg: well i'm pretty sure the Right Way™ for upstreams is what taglib-sharp is doing [21:07] so please look at how it's done there and follow suite. [21:07] * hyperair → bed [21:07] it's 5am and i need to wake up early tomorrow [21:07] hyperair, Okay, thanks! [21:07] 'night [21:07] * tedg doesn't want to be technical, but it already is early tomorrow :-) [21:10] tedg: you don't *want* to be technical, you just find that it's an innate aspect of your makeup? :) [21:16] slangasek, I blame society [21:24] hey folks [21:24] is sound still broken in Precise for flash? [21:24] still having issues with playing flash videos and using G+ [21:25] jono: eglibc 2.15-0ubuntu2 will fix it; building now [21:25] (separate issue than the flash one yesterday) [21:25] slangasek, awesome, thanks [21:25] oh I see [21:44] slangasek: is that eglibc build the fix for the omg nvidia issuse? [21:44] dobey: yes [21:44] yay [21:45] was it an issue in __memcpy_ia32 btw? i'm hitting an odd crash in it right now, and just wondering if it's related [21:50] dobey: no, it was in the linker [21:50] ah ok [21:51] then perhaps it's time for some bourbon, if i'm going to have to actually debug this crash in gnome-keyring, in introspection, in python :-/ === bladernr` is now known as bladernr_afk === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [23:55] [6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~[6~OQhi [23:55] does someone know is reproductible hangs of chromium and firefox are due to eglibc 2.15 or other recent upgrade? [23:56] hrw: haven't seen any reports of browser hangs due to eglibc 2.15; there are two critical bug fixes in eglibc 2.15-0ubuntu2 [23:56] slangasek: here, I run chromium or firefox, enter address and it hangs - have to xkill [23:57] hrw: arch? [23:57] amd64 [23:57] hmm [23:57] I've been running firefox w/ eglibc 2.15 for a week, no issues [23:59] hrw: Possibly flash being crazy?