=== Brumle__ is now known as Brumle === asac_ is now known as asac [05:39] Good morning [05:40] morning pitti [05:42] hey jibel, bonjour [05:42] good morning [05:42] pitti: simple-scan has a ubuntu-desktop branch [05:44] can anyone point out what is wrong with this "bzr mu --version 0.0.3 ~/bzr/unity-greeter/unity-greeter-0.0.3.tar.gz" - it just updates the changelog but doesn't pull in any of the new files [05:45] jbicha: ah, pushing there, sorry [05:45] robert_ancell: looks fine [05:45] morning, pitti [05:45] hey jasoncwarner_ [05:45] * robert_ancell scratches his head [05:46] pitti: if you update, you should get a new compiz...been running it all day and it has been much more stable for me [05:46] jasoncwarner_: already updated last night [05:46] :) [05:46] I lost my theming (back to Ambiance) [05:46] robert_ancell: I've had headaches with bzr mu, I just reported bug 832496 about a different issue today [05:47] Launchpad bug 832496 in bzr-builddeb "[bzr-mu] New version was not found in " [Undecided,New] https://launchpad.net/bugs/832496 [05:47] and using dash with keyboard si still a nuisance [05:47] but at least it didn't crash yet or destroy the apport-gtk popups :) [05:47] jbicha, ah, cheers [05:47] pitti: yeah, dash + keyboard aren't working well right now, but hoping unity today fixes those (have to and check the changelog) [05:47] and hopefully, this unity bugfix is more substantial than last week's :-) [05:49] jbicha: ah, I think I ignored it because the previous oneiric upload wasn't there either [05:49] * pitti commits that as well, though [05:49] yay, amd64 retracer chugging away happily now \o/ [05:51] jbicha: branch updated [05:55] Hey pitti. [05:56] hey RAOF, how are you? [05:57] Pretty good! [05:58] I'm off for about an hour for a run [05:58] I think I've got all the security complaints on the colord MIR resolved, so once that's processed g-c-c and g-c-m can get built. [05:58] Have running fun! === htorque_ is now known as htorque [06:26] I'm having some big problems with gconf but I need help figuring it out [06:27] one example: I just reinstalled yesterday & cleared my gconf settings, I went to gnome-terminal to disable the Alt shortcuts [06:27] to switch tabs so that I could use that in irssi, but the settings don't actually get written to gconf [06:28] the other bigger problem is in the metacity keybindings, run_command_terminal is supposed to be Ctrl-Alt-T by default in Ubuntu [06:29] and show_desktop should be Super-D but this isn't what's actually happening [06:31] ah, I think this is the first issue: https://bugzilla.gnome.org/show_bug.cgi?id=656873 [06:31] Gnome bug 656873 in general "gconf-editor has sync issue to set value" [Major,Unconfirmed] [06:48] morning mvo [06:50] mvo, https://bugs.launchpad.net/ubuntu/+source/aptdaemon/+bug/831965 do you have got any idea where this exception could be tiggered? [06:50] Ubuntu bug 831965 in aptdaemon "aptd crashed with SystemError: E:Sub-process /usr/bin/dpkg returned an error code (1) (dup-of: 799982)" [Undecided,New] [06:50] Ubuntu bug 799982 in aptdaemon "aptd crashed with SystemError: E:Sub-process /usr/bin/dpkg returned an error code (2)" [High,Confirmed] [06:51] mvo, actually we should catch all SystemError exceptions in Worker._apply_changes [06:52] and map them to a TransactionFailed one [07:06] maybe my keybindings problems are compiz's fault; does compiz mess with the metacity settings? [07:06] glatzor: yeah, that sounds sensible [07:07] glatzor: note that we use DPKG_UNTRANSLATED_MESSAGES (patches/01_add_dpkg_untranslated_messages) so the terminal log will be english [07:09] RAOF: I replied to bug 785828 [07:09] Launchpad bug 785828 in libdbusmenu "memory leak in label_notify_cb" [Undecided,Fix committed] https://launchpad.net/bugs/785828 [07:09] RAOF: I tried to build the source three times, but that single extra line was about as much as I could reduce the noise [07:10] pitti: That's rather annoying :/ [07:10] mvo, we already map the error. so I don't know why the exception is not already handled properly [07:11] pitti: In that case, I'll wave it through. [07:11] mvo, every SystemException from apt.Cache.commit() is mapped to a TransactionFailed error with the code ERROR_PACKAGE_MANAGER_FAILED [07:18] RAOF: you'll fish it out of rejected? [07:18] pitti: [07:18] pitti: Unless you can easily just upload it again. [07:19] RAOF: no, I'd have to re-download and upload, out of rejected is faster [07:19] RAOF: I can accept it, too [07:19] mvo: guten Morgen [07:20] glatzor: lp keeps timing out for me on the master bug [07:20] mvo: http://paste.ubuntu.com/673267/ > I've seen that happen when you mix static and GI bindingd [07:20] pitti: guten morgen! [07:20] mvo: what I did is to edit /usr/lib/python2.7/dist-packages/gobject/constants.py [07:20] pitti: oh, ok. I will hunt some more, I thought that all mixing was gone for good [07:20] mvo: and add these lies to the top: [07:20] "lines" [07:20] morning pitti [07:20] import traceback [07:20] print '--------- static module import! --------------' [07:20] traceback.print_stack() [07:20] print '-------------------------------------' [07:20] pitti: Ok, done. Sorry about that. [07:20] mvo: that makes it rather easy to see through which deep dependency chain it gets pulled in [07:21] hey glatzor, servus! [07:21] mvo, yeah, the timeout is because of the high number of duplicates. I already talked to people at #launchpad [07:21] RAOF: no need to be sorry at all -- it's good that you spot and complain about such things [07:21] bah, and there we are again -- LP timing out in the retracers :( [07:21] mvo, but the master bug doesn't provide any additional information [07:21] makes perfect sense to timeout for bugs that affect a lot of people [07:22] mvo: I had some of these as well -- I used /+text for them [07:22] that usually works [07:24] thanks pitti, the traceback helped [07:24] mvo: you can of course also just raise an exception there [07:25] glatzor: hrm, I had hoped that there would be a hint somewhere. I have seen sometimes that the defered eat the exception handling, [07:33] hi huats [07:33] hey rickspencer3 ! [07:33] morning everyone [07:33] hey rickspencer3 [07:33] bonjour huats [07:34] bonjour pitti [07:34] bonjour mes amis [07:35] :) [07:35] mvo, pitti ... obviously SC is broken in natty, only 1 person bought Photobomb so far, and no one reviewed it [07:35] I guess the system got overwhelmed by it's popularity and knocked off line [07:35] it's the only possible explenation [07:38] yeah, I've seen it installed on thousands of machines, something must be wrong [07:38] rickspencer3: there was an earthquake over here, give us a few weeks to recover! [07:42] mvo, but un-reproducable? [07:44] rickspencer3: just look at the popcon data instead, we need to use 64bit INTs there now for the first time [07:44] heh [07:45] just doing my job [07:50] hiya [07:51] do you have a bug report about a not vanishing alt-tab dialog already? :) [07:55] good morning everyone [07:56] ok, back to unity-2d - I can see anything with the alt-tab thing open all the time [07:56] hey chrisccoulson [07:56] hi pitti, how are you? [07:57] chrisccoulson: pretty well, thanks! happy that retracers are back up (mostly) [07:57] it was a 13 hour hackfest, but oh well :) [07:57] excellent, it's good that they are working again :) [07:58] amd64 is currently catching up; haven't started i386 yet, but will soon [08:01] jbicha: what is the last word on aisleriot? I thought the license problem was ok now? [08:01] pitti: upstream forgot to include their COPYING.LGPL in their Makefile so it wasn't in the tarball [08:02] jbicha: right, but I thought it would be in the next release, and until then debian/copyright would suffice? [08:02] morning [08:02] hey rodrigo_ [08:03] pitti: well we talked to them a while ago and they did improve their licensing stuff significantly [08:03] bonjour seb128 [08:03] I think we should put it in Ubuntu now and I'm waiting for a sponsor [08:04] jbicha: I agree; I'll upload now [08:04] pitti: thanks [08:04] hey desktopers [08:04] hey pitti, how are you? [08:04] hey jbicha [08:04] seb128: happy :) [08:05] seb128: good morning [08:05] seb128: the retracers didn't do anything over night, LP kept timing out on searchTasks() [08:05] seb128: but I found a cheesy workaround [08:05] pitti, something making you happy? [08:05] great ;-) [08:05] seb128: (only grabbing bugs since August 1, and once these are done, working backwards) [08:05] hi pitti, jbicha, seb128 [08:05] seb128: and I fixed some more remaining crashes which happen on old/weird bugs, and it doesn't break permanently (keep lock file) on LP timeouts any more [08:06] seb128: hacking until midnight paid off :) [08:06] hey rodrigo_ [08:06] pitti, that's most excellent! [08:06] hi seb128, how are you? [08:06] rodrigo_: howdy [08:06] pitti, well, packaging until midnight paid off as well, we got a compiz update out :p [08:06] seb128: congrats about that, great job! [08:06] how is compiz working for everybody today? [08:06] seb128: stable, but I lost my Radiance theme :/ [08:06] it's dark now [08:06] hey chrisccoulson, I'm great thanks, what about yourself? [08:07] and using the windows key/keyboard is still utterly broken [08:07] but it stopped disrupting the screen on popup dialogs like apport-gtk [08:07] and hasn't crashed yet [08:07] pitti, yeah, I don't know what's going on with the theme, I'm getting that as well this morning [08:07] pitti, keybindings is unity rather than compiz [08:07] seb128, yeah, not too bad thanks [08:07] $ gsettings list-recursively|grep theme [08:07] org.gnome.desktop.interface gtk-theme 'Radiance' [08:08] seb128: right, but I thought so is the theme [08:08] i'm a bit frustrated that i can't get the new thunderbird theme to apply on upgrades :/ [08:08] chrisccoulson, oh, still not? [08:08] pitti, are you sure it's the compiz update that broke the theme? [08:08] that seems weird [08:08] seb128: I'm not [08:09] seb128: I just dist-upgraded yesterday, and after restarting my session Radiance went AWOL [08:09] pitti, hum [08:09] seb128, no. even though the preference is set correctly, the new theme isn't loaded (because only one theme extension can be loaded by the addon manager at a time, and upgraders already have an extensions.sqlite DB with the old theme marked as active) [08:10] not sure how to fix that on upgrade [08:10] :-( [08:11] i might have to just patch the default theme, which is how it will be implemented when it all lands upstream anyway [08:11] the metacity keybindings that are used by gnome shell aren't being set right either [08:12] however, there are concerns that the new icons don't work so well with high-contrast themes, so using a separate theme extension has an advantage here that enabling accessibility disables the custom theme [08:12] i'd lose that if i patched the default theme [08:13] Sweetshark: guten Morgen! [08:13] chrisccoulson: don't we just patch the default theme for firefox? [08:13] jbicha, no [08:13] that theme is upstream [08:13] Sweetshark: do you have access to the porter machines, so that you can more easily investigate the FTBFSes? [08:14] brb === rickspencer3_ is now known as rickspencer3 [08:17] rodrigo_: have you considered building g-s-d with packagekit support? [08:19] I don't imagine many Ubuntu users use gnome-packagekit but without the gsettings schema from g-s-d, the prefs gui won't even start === tkamppeter__ is now known as tkamppeter [08:20] jbicha: thanks for the simple-scan update! sponsoring this as well [08:20] it's probably not worth my time messing with it except that it's in the archives [08:20] jbicha: you are unstoppable [08:22] well this gives me something useful to do with my time [08:24] re [08:24] pitti, ok, it's mterry's and robert_ancell's fault [08:25] it's the unity-greeter g-s-d which doesn't exit and prevent the session one to work [08:25] you can stop the running g-s-ds and restart your user one [08:25] or use the gtk standard greeter as a workaround [08:25] uh, what? how does lightdm's session affect the user session? [08:26] seb128: oh, g-s-d crashes [08:26] (gnome-settings-daemon:19784): power-plugin-WARNING **: Failed to get brightness: GDBus.Error:org.freedesktop.DBus.Error.UnknownMethod: Method "GetBrightness" with signature "" on interface "org.freedesktop.UPower.KbdBacklight" doesn't exist [08:26] jbicha, hmm, which prefs gui you mean? [08:26] Speicherzugriffsfehler (Speicherabzug geschrieben) [08:26] chrisccoulson: ^ didn't you do something with the backlight yesterday? [08:27] pitti, same issue than the one we had with gdm, there can only be one xsettings handler by display [08:27] pitti - yeah, but it hasn't built yet (waiting on the colord MIR) [08:27] i changed the backlight OSD to use notify-osd [08:28] pitti, the greeter one is running in :0, the greeter switch to the session but the session g-s-d can't start since there is already an handler set [08:28] that message doesn't look like anything i've touched ;) [08:28] ah [08:28] well, maybe the segfault is unrelated to that [08:28] so probably due to what seb128 says [08:28] $ grep xsettings .xsession-errors [08:28] ** (gnome-settings-daemon:1867): WARNING **: You can only run one xsettings manager at a time; exiting [08:28] pitti, that warning is there since gpm moved to gsd [08:28] hm, why does the lightdm session keep running in the first place? [08:29] that I need to check with robert_ancell [08:29] rodrigo_: http://paste.ubuntu.com/673627/ it's part of gnome-packagekit [08:29] pitti, it doesn't I think [08:29] pitti, ps ax | grep greeter -> nothing [08:29] seb128: check "ps aux|grep lightdm" [08:29] pitti, it's just that it doesn't clean as it should before starting the user session [08:30] I have gnome-settings-daemon, dbus-launch, notifyucation-daemon, gconfd, gelclue, gvfsd, syndaemon [08:30] jbicha, right, the problem is that we don't have pk in main [08:30] pitti, right, the session doesn't keep running, just the processes it started [08:30] seb128: that sounds like a lot to run even while lightdm is running [08:30] pitti, I bet it will make robert_ancell want systemd and cgroup handling ;-) [08:31] seb128, I might have a question regarding the gtksourceview3 packaging. When I build the package it says that there is a new symbol that appears... I have understood that I need to add the new symbol in the symbols file, but which release number do I need to associate with ? the first packaged release where this symbols appears or the first upstream release where it appears ? [08:31] pitti, those are mostly spawned by indicators that get loaded in the greeter, mterry is looking at cleaning that [08:31] right, what I mean is, we don't need notification-daemon or syndaemon in lightdm at all, but after starting the user session they should all go [08:31] didn't robert rant against running all these bits in the greeter and wrote lightdm for this? :-) [08:31] ah [08:31] huats, we usually use the first packaged version [08:32] huats, without the revision [08:32] huats, is 3.1.4 in your case I think [08:32] ok [08:32] pitti, well the gtk greeter run nothing of that, it's a bit trickier with the unity one due to indicators and other things design added for it [08:32] thanks ! [08:32] pitti, but "known bug, being worked" in short [08:32] seb128: thanks [08:33] then I should finish it today, it was a blocker for me almost since the begining, but I wanted to understand it clearly by myself :) [08:33] pitti, we had a bit of a discussion about using g-s-d in the greeter, it makes some useful things [08:33] pitti, only a couple of options are turned on in the greeter g-s-d, theme, xrandr, power basically [08:39] I can see retracing \o/ [08:41] seems that the amd64 one is chugging away happily, so I'll setup an i386 one now [08:42] seb128: I'm holding back the lucid/maverick/natty ones FYI, they are less time sensitive, and I haven't extensively tested that yet [08:42] pitti, wfm [08:47] seb128: ok, i386/amd64 are cron'ed now and both work [08:47] i386 should start in 20 mins, amd64 in 5 [08:47] the amazing pitti! [08:47] pitti, hug, you rock ;-) [08:47] seb128: can you have a quick look at osageorange ~ubuntu-archive/ ? [08:47] seb128: thanks :) [08:48] seb128: just want to give you a quick walkthrough [08:48] pitti, is osageorange the same box that the old retracers? [08:48] seb128: yes [08:48] porter-i386 aka porter-amd64 [08:48] ok, I'm there [08:49] pitti, yeah, I use the porter alias, I'm bad with names :p [08:49] seb128: so, ~/apport/ is a checkout of the no-chroot branch [08:49] ok [08:49] seb128: it doesn't need any particular setup, just checking out the source [08:49] I'll merge it into trunk after some final cleanups [08:49] seb128: please do a bzr diff in there [08:50] created_since= [08:50] nice [08:50] seb128: the "created_since='2011-07-01'" stuff is my workaround for the LP timeouts [08:50] we need to slowly move back until we caught the backlog, and LP doesn't time out yet [08:50] seb128: the other diff is that I temporarily disabled consolidation, I'll debug that next (dupe checking has almost caught up) [08:51] seb128: apport-retracer-{i386,amd64} are the old retracers and are now largely obsolete/useless [08:51] seb128: I moved all our hack scripts into the "scripts/" subdir [08:51] good idea [08:51] seb128: cache-{i386,amd64} is for the new retracers, mostly to keep the apt tree and *.deb there, doesn't need any maintenance [08:52] seb128: if you ever have to reinstall this stuff, you can just rm -r and mkdir them [08:52] seb128: logs are now put into log/: dupcheck.txt, i386.txt, amd64.txt [08:52] seb128: the retracers are now not release specific any more, in theory they can retrace stuff from any release [08:53] seb128: the only release specific thing that they need is an apt sources.list for the built sandboxes (unpacking of the required debs and ddebs) [08:53] seb128: these are in config//sources.list [08:53] seb128: they are shared between the i386/amd64 one [08:54] seb128: e. g. check config/Ubuntu\ 11.10/sources.list [08:55] pitti, should that sources have natty-proposed and natty-updates ddebs? [08:55] ups, it has [08:55] ignore me ;-) [08:55] seb128: yes, you remember the reason? [08:55] seb128: I also prep'ed a config/Ubuntu\ 11.04.disabled/sources.list [08:56] pitti, the ddebs are not archive copied to oneiric? [08:56] once it has caught up with oneiric, we can rename it and activate retracing for natty [08:56] ok [08:56] and copy/paste/adjust accordingly for lucid/maverick [08:56] seb128: correct [08:56] seb128: the retracer should now crash a lot less on LP timeouts, it'll just retry the next time [08:57] is there a lock of some way? [08:57] seb128: and finally, ~ has the lock files now: lock.dupcheck, lock.amd64, lock.i386 [08:57] seb128: snap :) [08:57] ;-) [08:57] seb128: the dup checker just finished [08:57] it fully caught up [08:57] I see the amd64 lock [08:57] those seem much easier to maintain \o/ [08:57] https://bugs.launchpad.net/ubuntu/+bugs?field.tag=need-duplicate-check [08:57] seb128: they are indeed; no chroots, no fakechroot, no separate instances per release, etc [08:58] seb128: finally, please check crontab -l [08:58] it's like the best thing since sliced bread! ;-) [08:58] *beam* [08:58] yeah, it's so much easier, and more robust, too [08:59] seb128: amd64 started, so you can watch tail -f log/amd64.txt [08:59] why did you keep the older retracer entries commented? [08:59] any idea of a reason the monitor would lose signal, apart from a hardware problem? [08:59] rodrigo_, xorg or video driver bug? [08:59] i.e output turning off [08:59] I have this monitor connected to 2 computers, and seems to only happen on my oneiric box [08:59] seb128: just in case we need them back [08:59] seb128, I guess so [08:59] rodrigo_, try asking on #ubuntu-x maybe [08:59] yeah [08:59] seb128: you'll see that I let the dupe retracer run in the lucid dchroot, as it's very stable [09:00] seb128: I let the i386/amd64 ones run in oneiric, as this has the most complete multiarch ld.so configuration [09:00] ok, makes sense [09:00] seb128: I haven't yet tested retracing oneiric crashes in a lucid dchroot; my suspicion is that gdb looks for the ld.so configuration to find libraries [09:00] but I haven't checked that [09:01] but in general it seems good to use the latest gdb etc. [09:01] well if that works in the oneiric environmment no reason to use lucid [09:01] right [09:01] I think natty should work as well, if the oneiric one ever goes bad [09:01] natty had the multiarch ld.so config already [09:01] seb128: ok, any other questions? [09:02] * pitti goes to try the consolidation [09:02] pitti, not so far [09:02] let's see after having them running for some time [09:02] but that seems well architectured and easy to use [09:02] so I'm happy ;-) [09:02] * pitti hugs seb128 === mpt_ is now known as mpt [09:02] * seb128 hugs pitti back [09:02] seb128: it's also quite a lot faster now [09:03] let's hope it catches up with the backlog in a few days [09:03] pitti, we should maybe change the "created_since='2011-07-01'" to 08-15 I think [09:03] 0815 sounds nice [09:04] pitti, i.e get the retracing that will work and are useful first [09:04] seb128: WFM [09:04] then we can do all the old ones [09:04] those will fail because version changed anyway [09:04] done [09:05] ok, let's see how far the consolidation gets [09:05] I want this working again as well [09:05] otherwise we won't catch regressions [09:06] mvo, the "aptd crashed with SystemError: E:Sub-process /usr/bin/dpkg exited unexpectedly" bug seems to go crazy today. [09:07] mvo, we get several duplicates per minute! [09:08] welll hello Apport retracer [09:15] glatzor: urgh, all against oneiric? [09:16] glatzor, mvo: the retracers were down for some weeks so the dupchecker is catching up on weeks of backlog [09:16] glatzor, mvo: i.e those are not filled from today [09:16] glatzor, mvo: if that's of any consolation ;-) [09:20] ah ok [09:21] thanks seb128 for the clarification [09:21] yw [09:31] pitti, good work !! [09:32] ronoc, oh btw I found a bug in indicator-session [09:32] brilliant [09:32] ronoc, it displays "software up to date" wrongly after doing some upgrades [09:32] seb128, whats up ? [09:33] ronoc, like I selected only 5 packages to upgrade in the 90 available [09:33] i.e there are still 85 to upgrade [09:33] but my indicator session says "software up to date" [09:33] seb128, yeah I have seen this before. will check after beta freeze, [09:33] ok, no hurry [09:34] ronoc, seems to be easy to trigger by right clicking in update-manager, selecting only a few things to upgrade and upgrading those [09:34] ronoc, it acts like "update-manager is done running, the system must be uptodate" ;-) [09:35] seb128, yeah it seems to work fine at startup but after that it never reacts to another instance of apt running [09:35] (somewhat a corner case, but I like to select updates I want or need and install those first) [09:39] seb128: it used to always say "Updates Available" so at least they're trying different things :-) [09:40] hehe [09:40] close a bug, open a new one... [09:42] jbicha: Fail! You're meant to close one and open 2 your hearts just not in it is it ;) [09:43] opening new bugs gets tiring! [09:47] jbicha, write a lplib script to do it for you ;-) [09:49] hmm, I'm just barely a programmer [09:49] but it could be Python... [10:00] seb128, since clutter update in oneiric, when building empathy I get that: /usr/lib/libclutter-gst-1.0.so.0: undefined reference to `glGetIntegerv' [10:00] it's a link-time actually [10:00] xclaesse, seems like a clutter-gstreamer issue, it maybe needs to be updated? [10:01] ricotz, ^ do you know? [10:01] no update available atm... is migration still ongoing? [10:01] seb128, note that empathy does not link on clutter itself, dunno what pulls that [10:01] xclaesse, I've no clue about clutter-gst, we don't use it [10:02] let's wait if ricotz knows about the issue [10:02] ok :) [10:03] Something's not linking to libGL correctly, it seems. [10:04] likely clutter since that's the only thing that changed [10:04] or clutter-gst was relying on clutter to bring the symbol [10:05] That seems likely. [10:05] that's one issue, the other is why does empathy try to link on clutter at all [10:06] xclaesse, because of the new call ui [10:06] I guess [10:06] it's using clutter-gst [10:06] clutter is in optional deps, but I'm not building with that [10:06] ok, dunno then [10:06] and it compiles fine, the error is at linking [10:07] well something tries to bring clutter in [10:07] I don't even have libclutter-gst-dev installed [10:07] it doesn't happen on our builds for sure so it ought to work without it [10:10] hmmm, it's time to make the firefox apport hook attach submitted crash ID's to bug reports [10:10] seb128, oh, i just went to sponsor gmime [10:10] denied ;) [10:10] chrisccoulson, thanks [10:10] doh, stupid from me to ask you [10:11] the reason it needed sponsoring is because mathieu failed to upload it... [10:11] heh [10:11] pitti, ^ can you sponsor it? ;-) [10:11] (will do a bit latter otherwise, I'm on a call now) [10:12] xclaesse, this looks like clutter-gst misses to link against libGL since it uses this symbol [10:12] should probably be in the desktop packageset really, seeing as kubuntu-dev can upload it [10:12] oh [10:12] actually [10:12] they can't ;) [10:12] but could at some point [10:13]