[00:00] kirkland: probably means that the script does some other things in shell before forking the main process, and upstart follows the first fork [00:00] even 'if [ ... ]' can cause that === mnepton is now known as mneptok [00:01] Yeah, it’s very easy to break current Upstart’s fork following code. 0.10 will fix that issue. [00:01] cjwatson: http://pastebin.ubuntu.com/352593/ [00:01] cjwatson: modeled off of ssh.conf [00:01] well, that looks ok, maybe libvirt forks twice? [00:01] libvirtd that is [00:01] cjwatson: must be [00:02] if so, 'expect daemon' [00:02] * slangasek nods [00:02] cjwatson: would that be causing my problems stopping the job? [00:02] certainly [00:02] yes, there's a bug where you can completely bugger upstart to the point where you have to reboot [00:02] fwiw, libvirtd does write its own pid file [00:03] upstart doesn't have any secret ways to stop jobs, it uses the same PID information that it reports to you with 'status' :) [00:03] I don't believe upstart cares about pid files [00:03] indeed not [00:03] indeed the general philosophy of upstart is that pid files are obsolete [00:03] slangasek: okey doke [00:03] cjwatson: How *does* one get Upstart to a state you need a reboot to fix? :-) [00:04] cjwatson: this is the pull command: ssh target_address dd if=remotefile | dd of=localfile [00:04] slangasek: cjwatson: \o/ [00:04] ion: I forget, but I did it while writing ssh.conf and Keybuk said "oh, yeah, known bug" [00:04] "expect daemon" worked like a champ [00:05] douglasawh-work: well, grub2 doesn't store anything anywhere magic, it just puts it on the disk with everything else ... [00:05] cjwatson: so your comment in ssh.conf, about editing the exec line, rather than /etc/default/* ... i mimic'd that.... is that correct, or lazy? [00:05] douglasawh-work: I don't understand how your *encryption* is laid out, though. when you get time, could you please file a bug with full details of what you're trying to make work? [00:05] kirkland: it's following strong preference from Keybuk [00:05] cjwatson: yeah, that's what I thought, but it doesn't work. I'm just using the alternate cd full disk encryption [00:06] kirkland: besides, after starting to use upstart's oom feature, there really wasn't that much left in /etc/default/ssh [00:06] cjwatson: yeah, there's nothing of value (except for "-d" == deamonize") in /etc/default/libvirt-bin [00:06] cjwatson: i'm on board ;-) [00:07] douglasawh-work: (I'm partly asking for a bug because unfortunately I don't really have time to look into this right now) [00:08] cjwatson: I'll file a bug. I just mainly wanted to make sure it wasn't already fixed in lucid before submitting [00:09] douglasawh-work: not to my knowledge, but then I didn't know it was broken :) [00:09] so it's always possible I missed something [00:10] cjwatson: slangasek: what do I need to do to get the server ISO build cronjob changed? File an RT? [00:10] ion: bug 406397 [00:10] Launchpad bug 406397 in upstart "init: job stuck with expect fork/daemon when parent reaps child" [Low,Triaged] https://launchpad.net/bugs/406397 [00:11] kirkland: RT is no use at all [00:11] (for this!) [00:11] cjwatson: heh [00:11] kirkland: if you want to file a bug, file it on the ubuntu-cdimage project, but it's probably quicker to just ask somebody in the ubuntu-cdimage team ... [00:11] cjwatson: send you chocolate, then? [00:11] kirkland: close your eyes, click your heels three times, and 'bzr update' to reveal that it's already been done [00:11] slangasek: ah, send slangasek chocolate :-) [00:11] assuming you mean 'run ubuntu-server earlier' :) [00:11] slangasek: yeah, thanks [00:12] cjwatson: Ah, that. I managed to get out of it without needing a reboot (as a reboot would have been inconvenient with users on the box in question). :-) [00:21] hi anyone know how can i get involved with ubuntu developmenting? [00:27] Shiran: http://wiki.ubuntu.com/ContributeToUbuntu, http://wiki.ubuntu.com/UbuntuDevelopment [00:28] yea yea i don't get it [00:28] i registered myself in launchpad [00:28] and i dont get how to find stuff to work on in there [00:30] Shiran: you won't find that an effective way to get started, no. It's much more effective to find something that you actually want to work on (e.g. a bug that's annoying you personally) and work on that [00:30] Launchpad has *everything*, it's a bit like wandering into a field and looking for an interesting blade of grass [00:30] you can try https://bugs.launchpad.net/ubuntu but it will still be overload :) [00:31] i wanna get involved with programming [00:31] isn't there a way to find teams that work on stuff that need more people? [00:32] it is much more effective to decide what you want to work on and then look for a relevant team [00:32] "programming" is a bit too vague for this purpose, try to narrow it down :) [00:32] how can i narrow it down? [00:32] based on your interests [00:33] my interest is writing code [00:33] yes, try to be a bit more specific :) [00:33] im not sure i can :P [00:34] c++ [00:34] is that more specific? [00:34] not really, but it may help you to find (for example) some programs already installed on your system that are written in that language that you might like to help with [00:35] distributions, by their nature, are rather more involved in integrating existing code from elsewhere than they are in writing lots of new code, although of course some of the latter does happen [00:36] so if you want to work in the context of a distribution, you're generally talking about either maintaining existing code (bug-fixes, feature enhancements, etc.) or finding one of the areas where the distribution is trying to distinguish itself [00:36] isn't there a way to find teams of programmers? [00:37] say i wanna work on KDE [00:37] http://wiki.ubuntu.com/Teams [00:41] Shiran: Kubuntu developers are in #kubuntu-devel, we do occational bits of c++ but mostly its working with upstream KDE code. KDE developers are in #kde-devel === Shiran_ is now known as Shiran === kirkland` is now known as kirkland === dendrobates is now known as dendro-afk [01:41] hi :) === dendro-afk is now known as dendrobates [02:38] cjwatson: installed grub1 and the problem is resolved, so thanks for that suggestion. At some point I'll test whether the problem exists in Lucid at all...I just know it's a problem in karmic. after I test that, I'll file a bug === Shiran_ is now known as Shiran [03:27] hey could anyone help a newbie apply a patch? http://www.mail-archive.com/hydrogen-devel@lists.sourceforge.net/msg00083.html this one particularly [03:27] i eed it for the hydrogen source code to compile === jam1 is now known as jam === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [06:39] TheMuso: Hi, Ubuntu doesn't work with libsdl1.2debian-alsa, but it does work with the pulseaudio variants. The same is true for LTSP (thin clients). Could we make it so that either libsdl1.2debian-pulseaudio or libsdl1.2debian-all are installed when an Ubuntu user installs an SDL-using app? [06:40] For libsdl1.2debian-all to work with pulse, SDL_AUDIODRIVER=pulse needs to be in the environment, but than can be taken care of by e.g. LTSP. [06:48] alkisg: The simple way to do that is to just adjust the relative dependencies of apps. My memory is that we spent a lot of time changing to use -alsa some time back, and those changes can probably be reverted. [06:49] persia: then those apps wouldn't work in kubuntu... [06:49] (and also, they're 387 on them using sdl in my current installation) [06:49] I thought the phonon pulse backend was first-class now. [06:49] Is the vlc backend still default? [06:50] I don't use Kubuntu or Xubuntu, but I was told that those still need -alsa [06:51] I haven't looked at it for lucid, but I believe the blocker for Xubuntu was xfce4-mixer running into bugs with gstreamer0.10-pulseaudio, which I thought were fixed. [06:52] For Kubuntu, the issue is all about having a fully-featured pulse backend for phonon so that users don't end up without sound. I don't know the status on that. [06:52] OK, so the plan is to switch all derivatives to using pulseaudio? That's fine, but why then do we have libsdl1.2debian depend on libsdl1.2debian-alsa first? Shouldn't we switch that to -pulseaudio? [06:52] But I don't know of any way (other than including in the default install) of preferring alternate libsdl implementations based on flavour, especially for people who use subsets of flavours. [06:53] My personal opinion is that everything should use pulse, and we should make that switch. I don't imply that this is an official plan. [06:53] It's just the easiest way to get out of the mess. [06:54] persia: the packaging problem can be solved with libsdl1.2debian-all, as this contains support for all backends [06:54] I know there was discussion about it at UDS Karmic, but I don't remember follow-up discussion at UDS Lucid. [06:54] alkisg: Well, except that then each flavour has to somehow handle setting SDL_AUDIODRIVER in some way. [06:55] Unless you want to change libsdl1.2debian-all to be smart enough to use pulse if it is available and fall back to alsa if it isn't. [06:55] Right [06:56] I think that's the best plan, it would work out of the box in all cases, and it would still enable some extreme user to select his own output [06:56] Right. [06:56] ...and it's already in main, while libsdl1.2debian-pulseaudio isn't. [06:58] Well, I think that if we're going to default to pulse if it's available, it's worth putting together an MIR for libsdl1.2debian-pulseaudio. [06:59] Actually, looking at it, it doesn't even need an MIR. It just needs to get pulled with a dependency. The source is already in main. [06:59] I was going to write one, but I don't think that's needed, -all is still the best plan even if we want the default output to pulse, as it enables users to select their own output [06:59] Well, except we need -all to recommend -pulseaudio if we want the default to be pulse [07:00] No - it conflicts with -pulseaudio [07:00] Ah, and it already includes the pulse bits. [07:00] Right [07:00] * persia hasn't actually looked at libSDL for far too long [07:01] * alkisg is a teacher and unfortunately he gets complains for all sdl-based edu* apps :( [07:01] Well, that's a good thing. Helps highlight the issue. [07:01] Of course, it quickly becomes a bad thing if we can't fix it. [07:04] Also, crimsun has said that -alsa with be patched to work even when pulseaudio is installed (I don't know the details). In that case, all workstation users will be satisfied, as SDL apps will work out of the box. But LTSP users wouldn't be satisfied because they'd require pulseaudio output as the sound goes through the network. In that case, LTSP could set SDL_AUDIODRIVER=pulse, and everyone is satisfied again. [07:04] Well, except it's a bit of a hack doing it that way. [07:04] Yes, ALSA can route clients into pulse which then routes back into ALSA which then routes to ears. [07:04] But it'd be cleaner if the apps talked to pulse directly. [07:05] Right, the correct way would be for Ubuntu to use libsdl1.2debian-all, and for -all to use pulse if it's available, else fall back to alsa. [07:06] So, 2 things need to be done: a small change in libsdl1.2debian dependencies, and a code fix in -all. [07:06] Well, that's the least invasive way to do it. I still think everything should use pulse :) [07:08] In that case, libsdl1.2debian should me modified to depend on -pulseaudio first. But we proposed that already, and it looked like it wasn't going to happen (because of K|Xubuntu), at least in the Lucid time frame... [07:08] https://bugs.launchpad.net/ubuntu/+source/libsdl1.2/+bug/203158 [07:09] Right, but that requires testing/changes to phonon and xfce4-mixer [07:09] If that work is done, then the bug can proceed. [07:09] The detection code in libsdl1.2debian-all is just a workaround [07:12] There were also complains from people not wanting to use pulseaudio at all... wouldn't libsdl1.2debian-all give them a way to do that? [07:13] Well, yeah, but most of those complaints come from people who ran into bugs and blame pulse (even when pulse wasn't the cause) [07:13] Mind you, I have a number of use cases where I *really* don't want pulse, but in those cases, I'm not using SDL audio anyway. [07:15] I'm a pulseaudio fan myself, as I use LTSP a lot, but "preferring" pulse instead of "forcing" it seems better imo... [07:15] I guess I see both as expressions of preference, but I agree forcing is wrong. [07:16] Anyway, both of them would solve my problem. So I'll stop nagging people and wait for K|Xubuntu to be fixed so they are able to use pulseaudio before I start nagging again :D === micahg1 is now known as micahg [07:16] Thanks a lot persia for your time & advice :) [07:16] alkisg: Well, if you're up for trying to do autodetection, that gives us a nice workaround. [07:17] Alternately, if you're up for testing the xfce4-mixer pulse interaction, that should be in good shape, and strong feedback that it is (or bugs that the devs can fix) would probably sway Xubuntu. [07:17] I know that upstream xfce4-mixer devs would like it to work properly, but don't have many user reports. [07:18] I'm less knowledgeable about the situation with phonon: I think that is still a bit of a mess, because there's just so much needed to make a working backend. [07:18] But I'd be glad to be wrong about that. [07:18] ISP has been up and down all day. Reddit is down. just ate the last Kasugai gummy. life bites. [07:18] I could give the autodetection thing a try. But for xfce4-mixer testing, as I never use Xubuntu, I wouldn't like to go ahead and report that "it works OK when ran from the live cd"... :) [07:18] good morning [07:20] Ah, a last argument for -all just came to me, that it would enable users with different needs to use different outputs even *on the same pc* [07:21] (e.g. on a multiuser workstation or on some kind of a server where they don't have administrative rights) [07:22] So user A logs on and wants pulse => OK, user b logs on and wants alsa => he just sets an environment variable. [07:22] Well, OK, but let's analyse that. esound and arts are deprecated. pulse supplies many of the core nas features. alsa was breaking stuff, and the latest OSS drivers aren't shipped by default. [07:22] arts isn't just deprecated. It's dead. [07:23] We've removed it from the archive and upstream has closed the bug tracker. [07:23] ScottK: I thought there was still support for KDE3 for some time (maintenance mode) [07:23] Ah, then it's very dead. [07:23] persia: There is, but not with arts. [07:23] ScottK: Do you happen to know the current state of phonon & pulse? [07:24] There were patches done to integrate them somewhat in KDE 4.4. We have them in Lucid. [07:24] I don't know the details. [07:24] So it's just a matter of testing at this point, or should be? [07:24] Should be. [07:24] Kmix needs some work too and that's not complete. [07:25] There is currently some back and forth about breaking feature freeze to squeeze that in. [07:25] Mandriva (where the work is being done) will ship it as a distro patch if it doesn't go upstream for 4.4. [07:25] We didn't consider it yet. [07:25] Regardless, Kubuntu will not ship pulse in Lucid. [07:25] That's colin guthrie's work? [07:26] I don't recall. [07:26] persia: Yes. [07:26] ScottK: Kubuntu will work fine with libsdl1.2debian-all with no modifications, right? [07:26] Current state of play: http://colin.guthr.ie/2010/01/mix-it-up/ [07:27] *would [07:27] alkisg: I don't know. [07:27] It's also 2:30 AM here, so anything I pretend to know, unless i give references, be skeptical. [07:27] So, if Kubuntu won't ship pulse in Lucid, and if I test all derivatives to see if they work with -all, could we make the dependency switch? [07:27] ScottK: Thanks for the references: this is looking promising (albeit for Lucid+1) [07:28] alkisg: Didn't crimsun say that was just a workaround and the wrong way to go about it? [07:28] alkisg: From -alsa to -all? That should be easy enough, so long as -all is configured to be sane about which backend it chooses. [07:28] ScottK: he said that for workstations, but when I mentioned that LTSP *needs* pulseaudio, I got the feeling that he accepted that. [07:29] persia: yes. [07:29] alkisg: When it comes to stuff like this, I'm personally going to go with what crimsun says. [07:29] So get him to agree and I'm all for it then. [07:30] OK, so /me writes down the steps he needs to make: (1) verify that all derivatives work with -all, (2) talk to crimsun about it, (3) [if everyone seems to agree] nag until the dependency is changed to -all first. :) [07:31] Ah and (4) try to get -all to use pulse by default when it's available [07:31] If you get 1 and 2 done, 3 will be easy. [07:31] Well, except don't nag. Just file a bug with the conclusions of your verifications and discussions. [07:31] And make sure to have tasks for any packages that need modification. [07:32] persia: sure, I was just kidding with that word. I don't go around nagging people... Sorry for my lack in English language. [07:32] alkisg: No worries. Just wanted to be clear for the logs :) [07:32] :) === dendrobates is now known as dendro-afk [08:14] Good morning [08:19] Good morning [08:25] hello, any recent change in GDM code? After today upgrade, GDM doesnt start [08:25] (lucid here) [08:35] jjardon: gdm didn't change, but yesterday we got a lot of new gnome libs [08:36] pitti, maybe that is the problem: I can see the "x" cursor, but GDM doesnt boot [08:37] jjardon: does it work if you boot with "text", log into a VT, and "sudo start gdm"? [08:37] cjwatson: there still is an issue with UEC installer not using the preseed file [08:37] (that's how I have to boot my machine these days) [08:37] cjwatson: https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/504157 [08:37] Launchpad bug 504157 in eucalyptus "Lucid UEC installer: CC install is not preseeded when separated from CLC" [High,Confirmed] [08:38] cjwatson: wget.gnu seems to work ok now, so it comes from something else [08:38] I can switch to a VT but sudo start gdm does nothing [08:38] ttx: I'll need the full syslog [08:38] ok, just a sec [08:38] ttx: ideally with DEBCONF_DEBUG=developer [08:38] pitti, I just joined but the changes from yesterday should not make a difference for login... [08:38] hm [08:39] cjwatson: where do I specify that ? [08:41] pitti, any other idea? your machine works ok? [08:41] jjardon: yes (except for gdm starting too early) [08:42] cjwatson: existing installer syslog attached to the bug [08:42] cjwatson: looking into how to regenerate it with DEBCONF_DEBUG=developer [08:45] ttx: on the boot command line [08:45] cjwatson: I'm on it [08:46] pitti, I'm going to restart and try some things. Wish me luck! [08:52] weee, I'm very lucky! I've just executed locale-gen and now all works again [08:52] strange [08:54] ah, yesterday's glibc broke all locales for some reason [08:55] I did locale-gen as well [08:55] but right after dist-upgrading [09:04] cjwatson: syslog with DEBCONF_DEBUG=developer uploaded to bug [09:12] seb128: thanks for updating my rhythmbox bug. Is there a better way of looking for bugs cause I looked it up before filing and couldn't find it :( bug 503887 (for your info). I'm assuming it down to the keywords you use etc [09:12] Launchpad bug 503887 in rhythmbox "Icon on the task bar is wrong (dup-of: 497095)" [Low,Invalid] https://launchpad.net/bugs/503887 [09:12] Launchpad bug 497095 in rhythmbox "App indicator icon is broken" [High,Triaged] https://launchpad.net/bugs/497095 [09:12] davmor2, you're welcome, I just looked for bugs with the name "icon" [09:12] sorting by most recent first... [09:13] there is not too many of those [09:13] seb128: cool ta for the info :) [09:17] since yesterday's dist-upgrade, I experience segfaults in gdm/X: http://paste.ubuntu.com/352787/ - Any idea what this could come from? is there perhaps some bug on this open already? [09:18] siretart`: The consensus is that the glibc update destroed all locales, and running locale-gen fixes things. I think. [09:18] let's try... [09:19] * alkisg is also having gdm problems, but locale-gen didn't solve them. startx works, though... [09:20] pitti: maybe the locale definition format changed? [09:20] jmarsden: excellent hint, locale-gen regenerated *only* the german locales (6), the english ones where already up-to-date. thanks a bunch! [09:20] in which case the installer is going to need an update ... [09:20] siretart`: You're welcome. [09:21] ttx: hmm. could I see the preseed file as well, which it allegedly downloaded? something isn't adding up here [09:24] cjwatson: I can't find anything in /tmp/eucalyptus-udeb*, does it delete it after "successfully loading" it ? [09:24] ttx: I mean the one at the URL you cited [09:25] cjwatson: ah ok [09:25] yes, it deletes it [09:25] might be worth double-checking that this shows something sensible when run in the installer: wget.gnu -q --no-check-certificate -O - $url [09:26] from the syslog, it looks rather as if it downloaded an empty file [09:27] * alkisg has el_GR.UTF-8 as the system locale, and while locale-gen didn't solve his gdm problems, locale-gen --purge did [09:28] cjwatson: the command gets the right contents, let me pastebin what it gets [09:30] maybe $url is not what we think it is [09:31] I'll make it syslog the URL [09:31] preseed contents is http://people.canonical.com/~ttx/preseed.conf [09:35] ok, so it definitely didn't actually load that [09:37] cjwatson: I think I got it [09:37] clouds="$(euca_find_component cloud)" [09:37] cloud="${clouds#* }" [09:37] cloud_preseed "https://$cloud:8443/preseed/preseed.conf" [09:38] euca_find_component returns "CLC 192.168.0.151:8773" [09:38] so it does [09:39] and wget.gnu doesn't really mid trying to fetch from 192.168.0.151:8773:8443, apparently [09:39] willfix [09:40] cjwatson: assigning to you then. Some more logging could come in handy :) [09:41] already added :) [09:41] cjwatson: heh :) [09:41] node->cluster preseed fetching has a similar bug [09:41] actually no it doesn't [09:42] no, it works [09:42] ... yes it does. COFFEE [09:42] it has it in at least some modes, I'm pretty sure. The code is cognate [09:42] right but maybe :8773:8773 just works [09:42] oh, you're right, actually the cluster code never appends a port [09:43] the CLC+Walrus+CC+SC / NC topology works [09:43] those bugs come from my testing of CLC+Walrus / CC+SC / NC [09:43] Then more bugs might come from CLC / Walrus / CC / SC / NC :) [09:44] ttx: fix committed, want me to upload? [09:44] cjwatson: sure, I could use an ISO with that [09:45] I still haven't had time to get into your netboot magic [09:45] this week is crazy (I should get used to it) [09:46] gah, eucalyptus bzr is out of sync with the archive [09:46] beh [09:46] which branch ? [09:46] ~trout kirkland [09:46] arh [09:46] lp:~ubuntu-core-dev/eucalyptus/ubuntu [09:47] kirkland: use 'bzr bind' and then you won't forget to push ... [09:48] ttx: should I wait for kirkland to get back and merge whatever he has locally? [09:48] bzr is missing: [09:48] cjwatson: he uploaded to archive what he had locally, so in theory we could reapply it to branch [09:48] + * debian/control, debian/eucalyptus-nc.upstart: (LP: #446036, #452572) [09:48] + - add a versioned depends for eucalyptus-nc on a new version [09:48] + of libvirt-bin that starts using upstart [09:48] + - start eucalyptus-nc on started libvirt-bin [09:49] yes, if you'd like me to do that, I can [09:49] cjwatson: if you don't mind, I really could use those hours before he wakes up for more testing [09:49] can some archive admin let the uboot-imx binary out of NEW (it changed the binary name to -to3 from -to2 with teh new upstream) [09:51] * ttx tries netboot magic in the mean time [09:57] argh [09:58] pitti, irc went mad for you, to? [09:58] doko__: does the new glibc disable that ipv6 patch again? my internet connection sucks, and now I found out that apparently that DNS bug is back [09:58] * pitti puts back the old workaround [10:01] aah, sanity again [10:17] it always scares me when I accidentally type 'bzr di' in (say) a git branch AND IT WORKS ANYWAY [10:18] oh, you got bzr-git working? [10:20] pitti: it worked for basic working tree stuff well before you could use it to branch [11:03] doko__: why has my en_US.UTF-8 locale broken on eglibc upgrade? [11:26] pitti: it's still enabled [11:27] slangasek: didn't see that, how is this broken? [11:28] doko__: bug #504198 [11:28] Launchpad bug 504198 in eglibc "locale support broken on upgrade to latest eglibc" [Critical,New] https://launchpad.net/bugs/504198 [11:28] caused at least one maintainer script crash on upgrade [11:33] slangasek: in gnome-menus? [11:33] (I fixed that this morning) [11:34] well fixing the crash doesn't fix the fact that locales are broken [11:34] ie everything is in english [11:34] of course [11:34] pitti: yes [11:35] at least libc6's script should call locale-gen to rebuild them [11:35] that's not enough to fix the locale there [11:35] I don't think missing locales are the problem, it looks to me like glibc is getting the path lookup wrong [11:36] or at least, when it finds /usr/lib/locale/en_US.utf8/LC_CTYPE, it decides it's no good and keeps looking [11:36] caused a software-center maintainer script crash for me [11:36] pitti, I did run locale-gen and it says fr_FR.UTF-8 is still up-to-date but I get warnings on running gtk applications saying the locale is no supported by the C library... [11:37] hm, I did that, and a reboot, and everything is working again [11:37] not there [11:38] oh, "up-to-date"? that's wrong [11:38] it should regenerate it [11:38] I think it did on the first run [11:38] now it says that [11:38] sudo locale-gen --purge [11:38] no hurry it's the mini config [11:38] I use it only for bootchart [11:39] I will let the system in state to verify that the fix work when there will be one === binitamshah is now known as binitamshah|away [11:40] ah; so indeed, locale-gen regenerates my non-English locales with success [11:40] but says the English ones are up-to-date [11:57] slangasek: is locale-gen --purge just enough to fix the problem? [11:57] doko__: yes [11:59] hey, I'm sorry to be asking questions here, but there doesn't seem to be anyone around on #ubuntu+1 [11:59] my keyboard stopped working with X [11:59] since yesterday [11:59] can anyone give me some pointers on how I can fix it? [11:59] this is lucid, btw [12:01] slangasek: I can confirm your observations, only my non-english locales were regenerated. - however, the effects were more grave: it made gdm and X segfault on the next reboot. [12:02] oh, I haven't gone so far as to reboot yet :/ [12:03] that was fun. espc since the segfault is pretty unobviously hidden in /var/log/syslog [12:34] pitti, cjwatson, slangasek: I could use a server CD respin to test the eucalyptus 1.6.2~bzr1120-0ubuntu5 [12:34] ttx, cjwatson, slangasek: triggering [12:34] ok [12:35] * ttx needs a local archive mirror to use netboot in a satisfying way :) [12:54] It seems hal is back in the default desktop. [12:54] ubuntu-desktop recommends pitivi, pitivi recommends hal. [12:55] it doesn't want to die [13:03] running apt-get install --fix-policy wants to remove grub-pc and replace it with the old one.. [13:03] among other scary things === mcasadevall is now known as NCommander [13:06] --fix-policy? [13:08] pitti: ^^ who can look at fixing pitivi's recommends on hal? === MacSlow is now known as MacSlow|lunch [13:09] slangasek: I discussed that with upstream, but it's not trivial; pitivi is in python, and there are no python gudev bindings yet (and probably won't be), and the python gobject-introspection support still needs to land [13:09] slangasek: but I changed hal to not start on boot, but on demand (dbus activation) [13:09] so it's not reintroduced into the boot path [13:10] pitti: beside other breakage, is xorg broken? Seeing this in lucid: http://pastebin.com/m3cc1d9d5 [13:10] in fact I always feared that we had to do that anyway to not break brightness settings for hardware without xbacklight [13:11] sebner: bug 504198 [13:11] Launchpad bug 504198 in eglibc "locale support broken on upgrade to latest eglibc" [Critical,In progress] https://launchpad.net/bugs/504198 [13:11] pitti: ohhh, my hero :D [13:11] pitti: ok, ack [13:12] slangasek: do I need to do a X restart after locale-gen --purge ? [13:13] ion: there's an upstream bug for pitivi, FYI [13:13] sebner: shouldn't be? [13:13] slangasek: to install all recommends [13:13] ion: https://bugzilla.gnome.org/show_bug.cgi?id=605920 [13:14] Gnome bug 605920 in general "Switch to (g)udev for device detection/support" [Normal,New] [13:14] slangasek: I'll try xserver reboot because your "workaround" doesn't fix it for now [13:14] seems to be undocumented though [13:14] tjaalton: hmm, indeed [13:15] slangasek: ahh, sudo ;) [13:15] sebner: hem, yes... :) [13:16] slangasek: wondering why xserver reconfig isn't working but at least the warnings are gone ^^ [13:16] slangasek: any chance you use nvidia-glx? [13:17] reconfig? it's been removed [13:17] there is no xorg.conf anymore (by default= [13:17] ) [13:18] tjaalton: ahh, right, just need it for haXX0ring the nvidia driver evidently (which is now broken for me) [13:18] tjaalton: I don't see anything in lucid with wrong Recommends on grub; sounds to me like the undocumented option is broken... === robbiew1 is now known as robbiew [13:18] sebner: nvidia-glx> not at all [13:18] slangasek: kk, thx anyways [13:19] slangasek: it's probably something recommending grub, and it conflicts with grub-pc [13:19] tjaalton: nothing in lucid recommends grub [13:19] slangasek: huh, ok then :) [13:19] everything correctly recommends grub-pc | grub [13:20] mvo: ^^ [13:33] tjaalton: hello, could you please add "-o Debug::pkgDepCache::AutoInstall=true -o Debug::pkgProblemResolver=true" and send me the output? [13:33] tjaalton: btw, it looks like my xorg patch is in the xorg-devel moderation queue [13:33] cjwatson: CC/SC picks up CLC preseed perfectly now, thanks ! [13:37] !summon Keybuk [13:40] ttx: excellent [13:42] ok xserver really looks b0rken, nv segfaults X (gdm starts, black screen but mouse is visible, also loginable) and nvidia-glx wants to remove the entire X stack :( [13:43] tjaalton: now I can't use vesa because I removed my xorg.conf :P [13:44] sebnerr: yes, nv is broken but mvo has fixed it partially (rest is on the way) :) [13:44] and tseliot has a ppa for the blob [13:44] mvo: ahah, got the error.."Installing grub as Recommends of linux-image-2.6.30-5-generic" :P [13:45] sebnerr: if you try https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-nv/+bug/494627 that would be helpful [13:45] Launchpad bug 494627 in xserver-xorg-video-nv "nv driver crashing with segmentation fault in libpthread.so.0" [High,In progress] [13:45] mvo: I'll ask people to push it through the queue [13:45] many thanks tjaalton [13:46] mvo: thanks for letting me see the error :) [13:46] * tjaalton runs the cleanup app [13:47] mvo: when did you send the patch? [13:47] to the list [13:47] tjaalton: yesterday in the afternoon, I can give you the exact time, give me a sec [13:48] tjaalton: Message-ID: <20100106135338.GL2603@tas> [13:48] mvo: it's on the list [13:48] tjaalton: Date: Wed, 6 Jan 2010 14:53:38 +0100 [13:48] tjaalton: oh, cool [13:48] jcristau accepted it [13:48] tjaalton: sorry for the noise then [13:48] sweet [13:48] no replies yet ;) [13:48] ion: do you have a link to your bug report or patch for mountall again? [13:49] mvo: ohhhhhhhh, testbuilding xserver? Sure I'll try :D /me removed -nv btw to let vesa start somehow ^^ [13:50] hmm computer-janitor crashes, need to clean up manually [13:50]