[00:01] bryce RAOF TheMuso robert_ancell meeting time! https://wiki.ubuntu.com/DesktopTeam/Meeting/2011-06-14 [00:03] exciting! [00:03] Good morning. [00:03] morning....hope everyone is having a good penultimate week before the rally! [00:04] anyway... bryce or RAOF here? [00:04] [TOPIC] X.org update [00:05] Ok, moving on. robert_ancell can you update us on LightDM? [00:07] still coming along. There were some crashers that people are getting, which some have been fixed in an update. I plan on releasing 0.4.0 today or tomorrow which has some more fixes. The apport crash retracer doesn't seem to be working, which is making it harder to notice and decode these crashes - bug reports remain private so I'm not seeing them [00:07] talked with the design team last night, and they have a draft design so I can start working on a prototype greeter [00:08] awesome. [00:08] * bryce waves [00:08] is it on by default now? MIR approved? [00:08] hey bryce [00:09] yes, so the number of bug reports has gone up quite significantly [00:09] robert_ancell: :) [00:09] awesome [00:09] I think you still need to opt-in when you upgrade [00:09] robert_ancell: thanks...good work. can't wait to see the protoype greeter as well :) [00:10] bryce: want to update on Xorg now? [00:10] sure [00:11] the new xorg package is merged in; this moves the failsafe-x bits and apport hooks over the the new xdiagnose package. [00:11] I've been working on getting a MIR in for that, migrate it to dh_python2, etc. Hopefully should be done this week [00:12] I know RAOF is working on mesa but don't know what the status is for this week. 7.10.3 came out this week, but I think in oneiric we're moving to a git snapshot anyway [00:13] been lots of user questions about graphics/input bugs this week too [00:13] jasoncwarner, think that covers it [00:13] bryce: yeah...and I've been hearing about the general state of drivers (fwiw)... [00:13] thanks, bryce [00:14] TheMuso: care to give quick update on Accessibility and Qt specific stuff? [00:15] jasoncwarner: Well didia is dealing with the QT patch integration side, I am working with upstream to get a few license issues solved for the qt-at-spi bridge. Hope to get that done today, push the patches upstream and get them integrated, as I was asked to fix up the license stuff due to lack of time on upstream's part. Hope to upload to the archive by EOW. [00:15] Didier even [00:15] TheMuso: :) [00:16] TheMuso: awesome...thanks! I know didrocks is going to a Qt conf end of this week to sort some remaining issues out... [00:16] TheMuso: thanks! [00:16] Cool. [00:16] np [00:16] anyone else have anything they wanted to discuss? [00:17] jasoncwarner, yeah one thing [00:17] I'm interested in gathering ideas on improvements for launchpad that would help us. [00:17] if anyone has ideas, please shoot them to me [00:18] jasoncwarner, there's a coming feature for Launchpad that promises to help make doing merges easier [00:18] bryce: very cool! thanks for the update :) [00:18] (not that I do merges, but I've read about them....in a book) [00:20] Ok...sounds like a wrap...anyone else (last call)? [00:21] alrighty then.... [00:21] thanks everyone! [00:21] [END OF MEETING] [00:34] BAH! That'll learn me for relying on technology. Apparently the meaning of “Enable USB Tethering” has been changed to “Ensure my mobile will not connect to a mobile network” behind my back. [00:34] Sorry guys. [00:34] lol [00:34] Shit happens. [00:35] heya RAOF [00:36] bryce: mesa is now blocked on bug #790204 [00:36] Launchpad bug 790204 in llvm-2.9 "[MIR] libllvm-2.9" [Undecided,New] https://launchpad.net/bugs/790204 [00:37] RAOF, that's for the 7.11 branch, or still .10? [00:38] It's for 7.10 [00:38] I *could* turn off the llvm support and upload 7.10.3. [00:38] may as well, unless you think the mir is likely to go through soon [00:39] My thinking was to turn on llvm early, since we're going to *have* to turn it on soon anyway. [00:40] Yeah. I guess I'll undo the llvm bits and upload. [00:42] RAOF, btw if you get bored, take a look at the mesa bugs on http://www.bryceharrington.org/Arsenal/Reports/ubuntu-x-swat/workqueue-natty.html; only about half a dozen, but a couple are -nouveau bugs that maybe could use looked at since we're flipping on 3d now [00:42] (if you click on the Package header, it makes it easier to spot the mesa bugs) [00:46] Ah, yeah. There are one or two actual bugs in there. [00:59] kenvandine: Is there anywhere I can fetch a test copy of the Gwibber GTK3 port? I'd like to have a look and track its progress. [01:33] Is all my technology complaining about the cold or something? Come on, sbuild, I know you can start! [01:41] lol [01:41] RAOF: Whats weather like in Tassy these days? My mother is actually going down there to spend the weekend with one of her sisters. I expect its rather cold. [01:42] It's not that cold; it's just about 2C in the morning, and my office isn't well heated by the gas fire. [01:43] It gets up to 12ish, so not too cold. [01:44] Ok nothing new there, sometimes on the mountains it doesn't get above 5-6 degrees. [01:44] So she'll be used to it. [01:44] Yeah. It won't be as cold as the Blue Mountains in winter. [01:47] Yup seems so. [01:51] TheMuso, sure [01:51] lp:~ken-vandine/gwibber/libgwibber-o-client/ [01:51] you need libdee from my gwibber ppa too, ppa:ken-vandine/gwibber [01:52] build and run ./service/gwibber-streams and ./client/gwibber-client [01:57] kenvandine: Thanks. === asac_ is now known as asac [05:28] Good morning [05:30] good morning [05:31] didrocks: oh, you're up early! [05:31] Good morning pitti, didrocks! [05:31] pitti: yeah, enough sleep today :-) [05:32] hey RAOF [05:34] robert_ancell: btw, on yesterday's daily I still got the lightdm crash at start; it happens as soon as I press Enter, then a subsequent session works fine [05:35] robert_ancell: this sounds a lot like the "wrong vt" but that we used to have, i. e. it starts on vt1 first and then pressing enter is caughht by the underlying getty [05:35] pitti, are you building from bzr? [05:35] Yay vt fighting! [05:35] robert_ancell: do you have a mechanism that forces lightdm to start on vt7 the first time? [05:35] robert_ancell: no, I was just testing the daily [05:35] pitti, yes, set vt=7 in the config [05:35] robert_ancell: locally I'm currently running gdm3, as I'm still working on it [05:35] pitti, is the daily working? [05:36] robert_ancell: with the "first session crashes" bug above, otherwise yes [05:36] it shows a "GNOME" session which isn't installed, but I guess you know that already [05:36] robert_ancell: "yes, set vt=7 in the config" -> is that done by default now? [05:37] also, does it only apply to the first-ever X.org start? [05:37] pitti, it defaults to vt=active, which I think it should only do if it detects plymouth [05:37] Good morning pitti, didrocks. [05:38] hey TheMuso! [05:38] pitti, so where are you getting daily builds from? I didn't get them to work [05:38] didrocks: I am working on the qt at-spi packaging, however there are some license issues that need solving, mainly that the upstrea git repo doesn't have all the required license documents/headers present. [05:38] TheMuso: I've already done the packaging [05:38] robert_ancell: cdimage.u.c., as usual [05:38] I'm working with upstrea to fix that. [05:38] pitti, oh, so you're running 0.3.7 [05:39] TheMuso: hum, who asked you to do that? I think we need to be more synced. People asked me to backport a11y to Qt hence the work on it :) [05:39] didrocks: hrm ok, I have started to maintain it in debian pkg-a11y git. Have you got something somewhere that I can have a look at? The eventual plan is for the debian pkg-a11y team to maintain it. [05:39] TheMuso: there was already a package and it's in the ~kubuntu-ppa [05:39] didrocks: Right, but I thought that was only Qt proper. [05:39] didrocks: Yes, but it needs work. [05:40] TheMuso: I know about the licence and already pinged upstream about it, we need to avoid duplication there to not confuse them :) [05:40] didrocks: Ok. [05:40] TheMuso: then, you can backport to the pkg-a11y git I guess once we are happy with it, let's not depend on that right now [05:41] Yup sounds good. [05:41] nice :) [05:42] TheMuso: FYI, it's very crashy, but qt-at-spi seems to work from my tests [05:43] didrocks: Yeah from what I tested I had the same experience. [05:44] pitti: robert_ancell: about the "GNOME" session, it's because lightdm doesn't test the TryExec key [05:44] didrocks, do you know why we have xsession files installed but not have the session to run? [05:45] robert_ancell: because all "default debian session" are packed in gnome-session [05:45] robert_ancell: so I'm not really eager to move them, import translation and such, in other packages [05:46] didrocks, what is the .desktop file that causes the "GNOME" session? [05:46] robert_ancell: /usr/share/xsessions/gnome-shell.desktop I would say [05:46] let me check [05:47] robert_ancell: yeah, that one [05:47] oh, I'm still on natty so I don't see it [05:47] shouldn't it really be a part of the gnome-shell package? [05:47] oh, it's the shell? [05:48] I had assumed it was the panel fallback one [05:48] robert_ancell: well, as previously said, we have upstream translations [05:48] robert_ancell: so, moving it to another source would mean to move translations, sync everytime there is a change upstream [05:48] not sure we want to play that game [05:48] anyway, there is the TryExec key for that [05:48] didrocks, I guess I mean, "why don't they put it in the gnome-shell package" [05:48] lightdm should just test it [05:49] seems a bit of a roundabout way of doing things. Yeah, I need to add it, it is a standard freedesktop option [05:49] robert_ancell: GNOME always put their default sessions in gnome-session, can be something raised upstream as before, there was not really the notion of "session" [05:49] (like "session type") [05:51] pitti: it's the shell, at least, the fact that you see the panel fallback means that the fallback is still working ;) [05:51] didrocks: I don't see the fallback [05:52] didrocks: this is from a daily live CD [05:52] oh? [05:52] we don't install fallback by default [05:52] oh ok :) [05:52] yeah [06:10] session restart, brb [06:34] chrisccoulson: I followed up and subscribed [06:49] pitti, do you happen to know how to run a test d-bus? [06:57] robert_ancell: session bus is easy, just run the program under dbus-launch [06:57] you can't launch a system bus as user [06:57] you have to change programs to connect to the session bus when testing [06:57] pitti, but wont that conflict with the existing session bus? [06:57] that's what I did with e. g. upower [06:58] robert_ancell: no, it just changes $DBUS_SESSION_BUS_ADDRESS for the program you run under dbus-launch [06:58] you can run as many in parallel as you like [06:58] but of course they are isolated from each other [06:58] pitti, looking at upower source, where are the tests? [06:58] so you can't use it to test functinoality for a different program (like talking to the running gnome-session) [06:58] sure [06:59] robert_ancell: http://cgit.freedesktop.org/upower/tree/src/Makefile.am#n126 [06:59] -> call of the tests [07:00] http://cgit.freedesktop.org/upower/tree/src/linux/integration-test [07:00] -> actual tests [07:06] pitti, ta [07:11] this doesn't seem useful: http://blogs.gnome.org/dcbw/2011/06/14/networkmanager-and-dual-stack-addressing/ [07:13] yeah, read that this morning [07:13] yeah, it somehow turns the story upside down [07:14] most apps shouldn't care about ipv4 vs. ipv6 [07:14] if I can reach my server, why do I need to care about the "how" [07:15] and 60 sec is an awful long time to wait to reconnect after waking my computer from sleep [07:15] they still need to do some sane error handling, of course [07:15] i. e. being online doesn't imply that I can talk to all servers in the world [07:15] they might be down, or only on ipv6, etc. [07:17] agreed, if applications need to go to that level of handling for all services provided on the desktop… [07:23] pitti, yay, it worked! Thanks again! [07:26] robert_ancell: rockin' [07:26] robert_ancell: test suite for lightdm? [07:28] pitti, yeah, runs from bzr-buildpackage finally [07:29] actually, now I need to test in a pbuilder... [07:29] right, dbus-launch works wonders there, and it works on the buildds [07:29] pitti, hey, I also have to move the config file in the package, any tips? [07:29] guessing I need to make a preinst [07:29] robert_ancell: hang on [07:30] robert_ancell: there's http://www.dpkg.org/dpkg/ConffileHandling standard recipes, but they are obsolete now [07:30] I guess the head explain the new method [07:30] there is a new dpkg-maintscript-helper method now [07:30] The maintscript method is nice and easy. [07:30] http://wiki.debian.org/DpkgConffileHandling [07:30] pitti: robert_ancell ^ [07:31] didrocks wins the google race! [07:31] :-) [07:31] robert_ancell: man dpkg-maintscript-helper explains it quite nicely [07:31] * didrocks is used to google for "moving conffile" [07:31] works wonder, first result :) [07:31] the header has some world on dpkg-maintscript-helper [07:32] robert_ancell: do not forget to pre-depends on dpkg 1.15.7.2 for lucid -> next LTS upgrade [07:33] (otherwise, you can have an issue in the preinst) [07:34] didrocks, is there a package you can point me at that uses the new method? All the packages I have installed seem to use the old one [07:34] robert_ancell: sure, one sec [07:35] grep dpkg-maintscript-helper /var/lib/dpkg/info/* [07:35] robert_ancell: ^ plenty [07:35] ntpdate, for example [07:35] /var/lib/dpkg/info/ntpdate.preinst, /var/lib/dpkg/info/ntpdate.postinst, /var/lib/dpkg/info/ntpdate.postrm [07:35] aha! [07:36] uses mv_conffile [07:36] yeah, I found a rm_conffile, but there are examples on mv_conffile :) [08:01] robert_ancell: so, lightdm just crashed but I still see the session on the screen :( [08:01] micahg, can I haz stacktrace? [08:01] nope :( [08:02] X is gone too... [08:03] pitti - thanks [08:03] the user has got the updated language pack btw [08:03] you can tell that from the list of extensions he has installed :) [08:03] robert_ancell: ah, I have an X stacktrace though :) [08:04] One for the RAOF department then! [08:04] micahg: Pastebin! [08:04] RAOF: http://pastebin.ubuntu.com/627128/ [08:05] RAOF: interesting snippet from the xorg log: http://pastebin.ubuntu.com/627129/ [08:06] micahg: Ra raw! (Incidentally, pastebining the full Xorg.0.log is generally a much better idea). How reproducible is this? [08:06] Pastebins are cheap. *Always* paste full logs! [08:06] (Unless you have security concerns about their content) [08:06] * micahg always has security concerns ;) [08:06] ok, I'll catch up the plane for the Qt contributor summit. Will not be really online since I'm back on Saturday night, but you still can drop me an email [08:07] RAOF: [ 98665.710] (WW) intel(0): first get vblank counter failed: Invalid argument [08:07] see you guys! [08:07] micahg: Aaah, right. robert_ancell, you're up! You haven't set up the VT correctly :) [08:07] 5565 of those messages [08:08] I don't think there's ever anything particularly sensitive in Xorg.0.log. [08:08] * micahg is still wondering why he sees a screenshot of everything before it crashed [08:08] chrisccoulson: erm, so he does have firefox-locale-nl? [08:08] If I had the full log, I believe that we'd find that X can't claim the VT, resulting in the intel driver not being able to claim drm master, resulting in explosions. [08:08] RAOF, please explain [08:08] chrisccoulson: but it doesn't actually show in Dutch? [08:09] pitti - no, he's got the updated language-pack-nl-base, but no firefox-locale-nl [08:09] ok [08:09] but they are both in the very same archive, pocket, and even component [08:09] i. e. the same Packages.gz [08:09] I don't see how you can have one without the other [08:09] micahg: *Could* I get the full Xorg.0.log? [08:09] * micahg skims quick [08:10] micahg, you are running lightdm from upstart right? [08:10] robert_ancell: I believe so [08:10] micahg, have you got a lightdm log? [08:11] pitti - his extension list shows he's missing any firefox language pack extensions entirely, and because he has multiple firefox profiles, you can also tell that he used to have the firefox 4 language packs installed by looking at the profiles which he hasn't used since the upgrade [08:11] pitti - " /usr/lib/firefox-4.0.1/extensions/langpack-nl@firefox.mozilla.org.xpi does not exist (old profile?)" [08:12] RAOF: http://paste.ubuntu.com/627136/ [08:12] robert_ancell: yes, a minimal one [08:12] micahg: Thanks. Told you there was nothing security critical in there ;) [08:13] robert_ancell: what are you looking for out of it? [08:13] micahg, anything, just checking it's behaving as normal [08:14] everything seems fine until the end and it just records a signal 111 [08:14] *11 [08:14] micahg, from the X server? [08:14] yep [08:14] Hm. No, that's not what I thought it was. [08:16] the X server was killed from under lightdm, so lightdm quit gracefully :) [08:16] micahg: Could I have a (filtered, if you want ☺) dmesg corresponding to that Xorg.0.log? [08:17] It looks like something permission-related is screwed up and the intel driver is having a hissy fit. [08:17] micahg, well, it should have started another X server, but that's not implemented yet [08:17] RAOF: 2 segfaults, will pastebin [08:18] I'm particularly interested in any {drm,i915} messages, obivously. [08:18] http://pastebin.ubuntu.com/627139/ [08:19] RAOF: ah, I have something else interesting: [ 1765.960027] [drm] GMBUS timed out, falling back to bit banging on pin 6 [i915 gmbus reserved] [08:20] 71 of those about 30 minutes after boot [08:20] If your monitors are on and have the right resolution that message probably isn't interesting. [08:21] it's a laptop using its internal screen [08:21] Hm. Was this crash at the lightdm login screen, or during regular use and just happened to additionally take down lightdm? [08:23] pitti - if you do "apt-get upgrade", then the recommends don't get installed [08:23] this was after being up for a day [08:23] i just tried it here [08:23] chrisccoulson: correct, that's its purpose [08:23] chrisccoulson: and apt-get upgrade is pretty broken really [08:23] folks need to use dist-upgrade, or update-manager [08:24] upgrade wouldn't work with a Depends: either, it would just hold back the packages [08:24] pitti - oh, i didn't realise that with the Depends: case [08:24] pitti - "I dont recall exactly if I used apt-get update/upgrade or update-manager" [08:24] i guess he used apt-get upgrade ;) [08:24] *nod* [08:24] micahg: So it was within a session, rather than at the greeter. This would make more sense, what with the intel driver apparently spontaneously going crazy. [08:25] yep [08:25] machine's been running hot too since the 3.0 kernel [08:26] There's nothing else drm-related in dmesg around the 95,000 second mark? I'll have to wander through the code to find where Invalid Argument could be returned. [08:27] nope [08:27] * micahg checks again [08:28] It's often annoying that i915 isn't more verbose when things go wrong. [08:30] RAOF: no, just apparmor denial messages (which I'll probably fix at the rally if I don't get to them before that) [08:32] Yeah. Silly i915. Why would you bother logging failed syscalls? [08:33] to fill up your disk? [08:33] :) [08:35] * RAOF reboots to test stuff. [08:38] robert_ancell: RAOF: do either of you need anything else from this session? I'm going to reboot the machine if not [08:39] micahg, no, I think I'm ok [08:46] hey desktopers [08:48] micahg: No, reboot away. [08:53] hey seb128 [08:56] hey chrisccoulson, how are you? [08:56] seb128, yeah, good thanks. how are you? [08:57] quite fine I think, let me finish my coffee before being sure ;-) [08:59] I assume the gnome-settings-daemon crash is well-known by now? [09:00] bonjour seb128 [09:00] Mmm, featureless grey. [09:00] WFM [09:01] Ah, ok. I'll see if it's a local problem then. [09:01] hey RAOF, which one? [09:01] After dinner, though. [09:01] hey pitti [09:02] seb128: Heh! Got lots? This one is apparently the one that I could very happily reproduce in gdb just five minutes ago but that now works. :) [09:03] RAOF, some over the cycles, one known race on login where the gdm instance didn't exit yet when the session one tries to start [09:03] It *might* be that it only happens in a gnome-shell session; it was segfaulting in gtk_ensure_icon_for_name or somesuch. [09:03] some other random or hard to debug ones [09:04] RAOF, hum, stacktrace would be interesting, could be the gnome-icon-theme split and the g-s session trying to load an icon pitti didn't put in the standard set [09:04] I don't have g-i-t-full either, and it's working fine here [09:04] seb128: That was my first thought; I've got gnome-icon* installed, though. [09:05] pitti, you are not using gnome-shell though? [09:05] Anyway, I'll log back into a gnome-shell session and post a full backtrace on launchpad for you. After dinner. :) [09:05] RAOF: oh, you are? [09:05] he's a traitror :p [09:05] RAOF: gnome-shell is missing a dependency to gnome-icon-theme-full [09:05] I have gonme-icon-theme-full installed. [09:05] RAOF: we just didn't add it yet to continue autosyncs for a while still [09:05] ok, so that's not it [09:05] ok, then it's not my bug :) [09:06] pitti, btw feel free to drop tomboy from the seeds if you want, I emailed the lists about it [09:06] FWIW, gdm 3 and gnome-screensaver already look like gnome-shell :) [09:06] seb128: ok [09:06] lightdm doesn't ;-) [09:06] robert_ancell, hey! [09:06] I was seeing if the Xorg pointer-barriers patch was working properly; only gnome-shell actually uses it. [09:06] seb128, hello [09:06] The shell looks pretty swanky. ;) [09:06] Even with a borken theme! [09:06] robert_ancell, how are you? [09:07] same old [09:07] ;-) [09:07] robert_ancell, can you fix bug #794315 [09:07] Launchpad bug 794315 in lightdm "lightdm forgets to source /etc/profile and ~/.profile" [Medium,Triaged] https://launchpad.net/bugs/794315 [09:07] yes [09:07] should be trivial and I would get my ~/bin back in the path :p [09:07] seb128: kicked from the seeds [09:07] * RAOF → dinner [09:07] pitti, danke [09:07] robert_ancell, thanks [09:18] pitti, hi, can i tell apport-retrace to stop trying to search for the dbgsyms? it's totally broken in oneiric and it's painfully slow [09:18] fta: just run it as non-root [09:20] pitti, i do, it's still taking several minutes spitting hundreds of WARNINGS [09:20] pitti, http://paste.ubuntu.com/627166/ [09:21] gedit is painfully slow to load on my machine [09:21] firefox even starts faster than gedit [09:21] fta: you currently can't disable the warnings, yes [09:21] chrisccoulson, same here [09:22] gedit works fine there [09:22] fta, i'm glad it's not just me then ;) [09:22] did it start being slow recently for you? [09:22] well it's less than 1 second to start here [09:22] seb128, ever since we got the new version [09:22] so dunno if that's slow for you [09:25] evo3 crashes in gtk3 while opening a meeting invite: http://paste.ubuntu.com/627168/ [09:37] seb128: hm, I just rebooted, and g-s-d is now not running any more either [09:37] seems it's due to the very latest updates from this morning [09:37] pitti, stacktrace? [09:37] I can run it manually in a terminal [09:38] and did so now [09:38] seb128: I'll investigate [09:38] grep settings .xsession-errors? [09:38] morning pitti seb128 and mvo [09:38] pitti, ok [09:38] hey glatzor, how are you? [09:38] already checked, nothing standing out [09:38] hey glatzor! [09:38] fine and yourself, pitti and seb128 ? [09:38] pitti, can you pastebin it? [09:38] glatzor, I'm fine thanks ;-) [09:39] seb128: need to restart my session, it's cluttered with several g-s-d restarts right now [09:39] pitti, don't bother [09:40] we can see that later [09:40] no problem, I just want to submit my updated patches to bugzilla, then I'm done with udisks and get to this [09:40] grrr, thanks to mutt, i discovered i'm late for a meeting [09:43] hi desktop, morning's crash bug 797498 [09:43] Launchpad bug 797498 in gtk+3.0 "nm-applet crashed with SIGSEGV in gtk_icon_source_set_pixbuf()" [High,Triaged] https://launchpad.net/bugs/797498 [09:43] not specific to nm-applet since it also affect gnome-settings-daemon [09:46] oh, perhaps that's the one we are currently looking at in g-s-d [09:47] same thing, I didn't have it running, but running from a terminal works [09:47] jibel, looks like my evolution crash too [09:47] * pitti blames new GTK :) [09:47] seb128, jibel: I'll restart my session with apport enabled, to see whether that's the same issue RAOF and I have [09:50] ok [09:50] hey jibel [09:50] hey rodrigo_, how are you? [09:50] Good morning rodrigo_ [09:50] hi seb128 [09:51] jibel, so, about the g-s-d's crash you had yesterday, whats the contents of /usr/lib/gnome-settings-daemon-3.0/gtk-modules, only the atk module? [09:52] bah, totally useless trace :/ [09:52] hmm, there's no -dbg package for at-spi2-atk [09:52] rodrigo_, yes [09:53] brb [09:56] jibel, can you install libgtk-3-0-dbg and see if you get a better backtrace? [10:00] jibel, and libglib2.0-0-dbg [10:00] rodrigo_, sure [10:01] re [10:01] jibel, pitti: works fine there but I don't dist-upgrade often, I tend to select updates, I've the new gtk though [10:01] I try to downgrade gtk to 3.1.4 and compare [10:03] pitti, you get it at every login? [10:04] have a stacktrace would be useful [10:04] I did some updates, let me restart [10:04] jibel, do you have glib 2.28 or 2.29 installed? [10:05] seiflotfy: yes, consistently [10:05] sorry, seb128 ^ [10:05] rodrigo_, there is a better trace in duplicate bug 796497 and bug 795392 [10:05] Launchpad bug 796497 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in __run_exit_handlers() (dup-of: 788710)" [Undecided,New] https://launchpad.net/bugs/796497 [10:05] Launchpad bug 788710 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in exit()" [Medium,In progress] https://launchpad.net/bugs/788710 [10:05] Launchpad bug 795392 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in __run_exit_handlers() (dup-of: 788710)" [High,Triaged] https://launchpad.net/bugs/795392 [10:06] rodrigo_, 2.29.8-0ubuntu1 [10:06] here as well; latest oneiric [10:07] downgrading to libgtk-3-0_3.1.4-0ubuntu1_amd64.deb and keeping everything else up to date fixes it [10:07] i. e. also the other binaries of gtk [10:07] jibel, no, both of those traces don't show where the actual crash is, see #0 in Thread 1, that's the actual atexit handler crashing [10:08] ok, so 2.29.x has the atexit call commented because of a crash, which might be this one, checking when that was commented... [10:09] rodrigo_: did the previous gtk 3.1.4 not use the new glib atexit handler then? [10:09] rodrigo_: want me to try current gtk with an older glib? (right now I tried older GTK 3.1.4 with current glib) [10:11] pitti, you get the same crash? [10:11] rodrigo_: yes, you don't? [10:11] no [10:11] g-s-d and nm-applet don't start automatically [10:11] although I don't have glib 2.29 [10:11] but running either in a terminal works [10:11] rodrigo_: dist-upgrade to current oneiric? [10:11] ii libglib2.0-0 2.29.8-0ubuntu1 GLib library of C routines [10:11] yes, will do [10:11] rodrigo_: shall I try downgrading glib? to which version? [10:12] current gtk doesn't seem to have any atexit handler in the code [10:12] pitti, yes, please [10:12] rodrigo_: did the previous gtk 3.1.4 not use the new glib atexit handler then?https://launchpadlibrarian.net/73542141/Stacktrace.txt applies to an atexit handler? [10:12] it looks like a fairly "standard" crash to me? [10:13] rodrigo_: trying 2.29.6 then [10:18] rodrigo_: glib 2.29.6 and gtk 3.1.6 does not fix it [10:18] downgrading glib makes no difference here [10:18] and above crash really looks like being in GTK only [10:18] rodrigo_, I can't get the 1rst symbol being resolved. I still get #0 0x019e5a70 in ?? () Here is the list of dng packages installed on this system: http://paste.ubuntu.com/627199/ any idea ? [10:19] rodrigo_: are we perhaps talking about two different crashes here? [10:19] s/dng/dbg [10:19] rodrigo_: I mean bug 797498 , which also applies to g-s-d [10:19] Launchpad bug 797498 in gtk+3.0 "nm-applet crashed with SIGSEGV in gtk_icon_source_set_pixbuf()" [High,Triaged] https://launchpad.net/bugs/797498 [10:19] pitti, yes it's a different crash [10:20] pitti, yes, the stacktace the crash I talk about is about an atexit handler, bug #788710 [10:20] Launchpad bug 788710 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in exit()" [Medium,In progress] https://launchpad.net/bugs/788710 [10:20] the gtk_icon_source_set_pixbuf started this morning, the at_exit crash 2 weeks ago [10:20] ah, that explains it [10:20] I don't get 788710 then, it seems to apply to the live CD only, or other setups? [10:21] re [10:21] seb128: wb [10:21] ok, after several restart with different gtk versions I don't get g-s-d issues or similar [10:21] seb128: so for bug 797498, downgrading just libgtk-3-0 to 3.1.4 helps [10:21] Launchpad bug 797498 in gtk+3.0 "nm-applet crashed with SIGSEGV in gtk_icon_source_set_pixbuf()" [High,Triaged] https://launchpad.net/bugs/797498 [10:21] weird [10:22] why does it work for me?! [10:22] is anybody else on i386 or are you all on amd64 there? [10:22] the atexit call in glib has always been commented, so it might be another lib using an atexit handler [10:23] jibel, looking at the libs used in g-s-d, to see if its in some other -dbg package [10:23] pitti, is that happening only on login? [10:23] my evo crash in i386 [10:23] +is [10:23] seb128: yes, apparently so; as I said, I can launch g-s-d and nm-applet from a terminal just fine [10:24] which makes this "fun" to debug :/ [10:24] indeed [10:25] pitti, did you manage to get a stacktrace with libgtk-3-0-dbg installed? [10:25] seb128: I have gtk, glib, g-s-d debug symbols installed, but totally useless trace [10:25] pitti, what is extra "fun" is that there is 5 weeks of active commits between the 2 gtk versions [10:25] I guess I'll try again [10:26] pitti, it's in gtk code? or atk or gail or something in the a11y stack? [10:26] if someone has a way to reproduce in a running system, then bisecting would be easy [10:26] seb128: I just downgraded libgtk-3-0, not gail, not atk, not gir, etc. [10:26] ok [10:26] you are on i386, right? perhaps that's the reason [10:27] well, I'm also running gdm 3, but other people who get this don't [10:27] and the LP bug is i386 as well, hmm [10:27] well fta says he gets a similar issue on i386 [10:28] i'm also using gdm, not lightdm [10:28] (not sure why it would matter for evo though) [10:29] fta: does evo crash for you in a running system? or just e-d-s on sessino start? [10:29] e-d-s is fine here [10:29] evo starts fine here even with gtk 3.1.6 [10:30] i just have a 100% reproducible crash while opening meeting invites from outlook [10:30] pitti, there is a recent commit in trunk: http://git.gnome.org/browse/gtk+/commit/gtk/gtkimage.c?id=c903ece96be6b307eb756430f0b0b9acfca4dbae [10:30] not sure if that could be it [10:30] my stack: http://paste.ubuntu.com/627168/ [10:30] oooh [10:30] seb128: the trace goes through that [10:31] #1 0x002d803b in ensure_stated_icon_from_info (image=, info=0x0) at /build/buildd/gtk+3.0-3.1.6/./gtk/gtkimage.c:1528 [10:31] pitti, right, I was looking to recent fixes to gtkimage.c [10:31] seb128: shall I build a GTK with that patch and check? [10:31] l1528 is gtk_icon_source_set_pixbuf (source, destination); [10:32] so that would fit with the destination being not properly initialized [10:32] pitti, if you want sure [10:32] doing [10:32] pitti, want I usually do is commenting the udeb and static flavor in the rules to win some build time [10:32] if you want to build a package [10:33] you need to clean the corresponding lines in the .install and .install.in but that's easy, just grep for static *install* and udeb *install* [10:33] ok, thanks [10:34] pitti, well, I'm pretty sure it's that commit, should I just backport and upload? [10:34] yes, that might be it, it tries to g_object_unref the destination in some code path where it might not be initialized correctly [10:34] seb128: I have it locally in bzr now [10:35] pitti, ok, I let you deal with it then ;-) [10:35] seb128: and building [10:35] seb128: but if you want to do it, sure [10:35] seb128: ah, ok :) [10:35] no need to duplicate work [10:35] seb128: letting it build while I'm talking to Jason then :) [10:36] jibel, about your crash, I just narrowed it down to, maybe, the X libs, so are you fully up to date? [10:36] * rodrigo_ upgrades everything related [10:37] rodrigo_, Yes I am, me and my system [10:37] heh [10:37] ok, I'll upgrade and try to replicate it myself [10:40] chrisccoulson: silly question, but is thunderbird using evo-data-server for the addressbook? [10:40] mvo, it's going to [10:41] mvo, not yet [10:41] for oneiric? [10:41] mvo, that's one of the blockers for the tb by default [10:45] hey mvo [10:45] mvo: happy belated birthday! didn't see you around yesterday, were you partying? :-) [10:45] * pitti hugs mvo [10:46] hi [10:46] oh, mvo is getting older again? [10:46] * seb128 hugs mvo [10:46] hey hrw [10:46] pitti: haha, many thanks! I was indeed :) [10:46] thanks seb128! [10:46] rodrigo_, for info I'm using lightdm, if it makes a difference. [10:47] * mvo hugs seb128 [10:47] jibel, shouldn't [10:48] is it normal that after installation of gnome-control-center/oneiric it segfaults on some components? [10:49] no, which ones? what stacktrace? [10:50] 'background' for example - nevermind that it does not have sense running it on xfce desktop [10:50] I need to install some debugs first [10:51] http://pastebin.com/ebK9R2rw - maybe it is because of Oxygen icon theme [10:52] it has been creating issues before, try without it [10:53] Humanity icon theme will be fine? [10:53] nope [10:55] same after switching to full gnome 3.0.2 icon theme [10:56] http://pastebin.com/XaH1mtca [10:56] oh, that's a gtk bug [10:56] we were discussing it half an hour ago, pitti is trying a built with a git patch to fix it [10:56] thanks [10:56] yw [10:57] I am trying to find a way to setup theme for gtk3 apps [10:57] as xfce takes care of gtk2 ones [10:58] install gnome-tweak-tools [10:58] or use dconf-editor [10:58] dconf-editor is in which package? [10:59] 11:59 hrw@puchatek:~$ LC_ALL=C gnome-tweak-tool 2>&1 |pastebinit [10:59] http://paste.ubuntu.com/627220/ [11:01] gnome-tweak-tool should depends on gnome-shell since it uses a schemas from it [11:01] dconf-editor is in dconf-tools [11:03] gsettings-desktop-schemas is what gnome-tweak-tools depends on [11:05] right, well the clock schemas is a gnome-shell one so it should depends on it [11:05] \o/ new GTK working like charm [11:05] that's the way gsettings works [11:05] pitti, \o/ [11:05] pitti: cool [11:09] seb128: nice, GTK built exactly as long as the 27 minutes I was talking to Jason :) (I didn't bother to disable stuff in rules) [11:10] ;-) [11:10] if you want longer phonecalls, get arm :) [11:10] it's taking less than 10 minutes if you hack the rules [11:12] seb128: with gnome-shell installed gnome-tweak-tool works [11:12] seb128: ^ could we apply the gnome-shell dependency to tweak tool in Debian? [11:13] seb128: then we'd get rid of our delta, as g-i-t-full would then be transitively pulled in from g-shell [11:13] hrw, seb128, this is bug 795084 [11:13] Launchpad bug 795084 in gnome-tweak-tool "gnome-tweak-tool fails to start .. org.gnome.shell.clock not installed - should depend on gnome-shell" [Undecided,New] https://launchpad.net/bugs/795084 [11:13] seb128: I think I'll just apply the dep to g-shell right now, there seem to be enough people stumbling over this [11:14] but at least we'd only have the delta in one place then [11:14] thx jibel [11:15] jibel, thanks [11:15] pitti, ok, well gnome-tweak-tools is in pkg-gnome I think so feel free to add the depends there [11:15] seb128: right [11:16] seb128: oh, and I think I'll add some dpkg-vendor --is ubuntu magic to gnome-shell to add the g-i-t-full dep only for Ubuntu [11:16] ok, taking bug 795084 then and doing above [11:16] Launchpad bug 795084 in gnome-tweak-tool "gnome-tweak-tool fails to start .. org.gnome.shell.clock not installed - should depend on gnome-shell" [Undecided,New] https://launchpad.net/bugs/795084 [11:17] pitti, thanks [11:18] heh - https://lists.ubuntu.com/archives/ubuntu-devel/2011-June/033476.html :) [11:19] yay chrisccoulson [11:19] i'm not going to spend any more time on that now [11:25] pitti - i closed one of my WI's on the default-email blueprint now ;) [11:25] i had a chat with one of the lightning developers last night [11:25] nice! [11:25] we've got a plan for how to do calendaring [11:26] splendid, g-shell currently FTBFS [11:26] oh, how come? nothing to do with spidermonkey i hope ;) [11:26] /usr/include/clutter-1.0/clutter/clutter-actor-meta.h:95:1: error: unknown type name 'G_CONST_RETURN' [11:26] oh, I think that might be a deprecated one now [11:26] oh, that's ok then [11:26] yeah, it is [11:26] -DG_DISABLE_DEPRECATED [11:28] pitti, the clutter update from yesterday should have sorted that for clutter, did you update today? [11:29] yes, I dist-upgraded some minutes ago [11:30] I have libclutter-1.0-dev 1.6.16-0ubuntu1 [11:31] pitti, yeah, seems like there is still some issues [11:31] any dbus daemon package that shuts down during upgrades, to see as an example for accountsservice? [11:31] pitti, ok, that's fixed in git but not in the current tarball, I guess build without the deprecation errors until the next update [11:31] *nod* [11:32] rodrigo_, what do you mean? [11:32] seb128, see https://bugs.launchpad.net/ubuntu/+source/accountsservice/+bug/785680 , there's a comment saying " I see no way for the daemon to shut down during a package upgrade (and the associated postinst to perform that)." [11:32] Ubuntu bug 785680 in accountsservice "[MIR] accountsservice" [Undecided,Incomplete] [11:33] hum [11:33] so I want to have a look how this is done in other packages [11:33] pitti, ^ do you know of any? [11:33] the couple ones I've looked don't do anything like that [11:33] rodrigo_, seb128: yes, upower and udisks do it [11:33] although, is it really a good idea to shut down during an upgrade? [11:33] pitti, ah, ok, cool [11:33] /var/lib/dpkg/info/udisks.postinst [11:34] pitti, danke [11:34] pitti, so, it needs the postrm, prerm and postinst, right? [11:35] it shouldn't? [11:35] rodrigo_: postinst is the most important bit; prerm is good for cleaning up, postrm shouldn't be necessary [11:35] ok, thanks pitti [11:46] seb128, is it worth keeping debian/control in bzr if it is autogenerated from control.in? I see it is there in the clutter package [11:46] seb128: ok, http://anonscm.debian.org/viewvc/pkg-gnome?view=revision&revision=28452 should do it [11:47] I tested the build in Debian in both ways [11:47] janimo, why not? it's easier to bzr add debian than to filter files [11:47] pitti, danke [11:47] seb128, ok, just that it seems an unpacked source has a different control file than the one kept in LP, as the latter is the old version [11:48] janimo, I usually do debuild clean before commiting to have it updated [11:48] but I guess it's not always the case [11:48] janimo, does it create any issue in practice? [11:48] seb128, probably not, I was wondering if a spurious file will get added to the debdiff because of this [11:49] no [11:49] seb128, I modified the bzr package branch, and when copying it over to the src package before upload, it changes control too [11:49] ok [11:49] bzr bd will do a clean before building the source so the control will be updated [11:49] janimo, why do you copy over? [11:49] seb128, probably because I do not fully understand how UDD works :) [11:50] janimo, debian only != udd ;-) [11:50] janimo, just run "bzr bd --source" [11:50] I made packagiung changes and now am about to do an update of the 'real' package then dput [11:50] it will build the source for you [11:50] or bzr bd to build with the binaries [11:50] or bzr bd-do to edit a patch [11:51] bzr bd-do gives you an unpacked source and copy over the changes you do in the debian dir when you exit 0 [11:51] you usually do bzr bd-do; quilt push, edit; quilt refresh; exit 0 [11:51] then bzr bd [11:52] oh quilt too? for every package? [11:52] well those using quilt [11:52] you can edit-patch [11:53] well that's for editing a patch [11:53] if you just need tweaks to the debian dir just edit files in the checkout and use bzr bd and bzr bd --source === MacSlow is now known as MacSlow|lunch [11:53] is this written up? The only link I found in the wiki was about packages under bzr (debcommit, etc) which is nice, but apparently not all I need [11:53] janimo, https://wiki.ubuntu.com/DesktopTeam/Bzr [11:53] seb128, thanks [12:01] seb128, bzr bd is nice indeed. I never used it before. When getting a package which is tracked in bzr it would be helpful to point to a wikipage as well and make a stronger case for actually using that instead of classical dput only. [12:01] I must have missed that last time, so my clutter changes got dropped on last sync [12:04] janimo, right, sorry about the dropping [12:04] pitti, is this ok -> https://code.launchpad.net/~rodrigo-moya/ubuntu/oneiric/accountsservice/add_postinst/+merge/64664 ? seems to work for me, for both upgrades and removals [12:04] seb128, it's ok, I missed to put them in bzr, and the rules for syncing were unclear to me [12:05] rodrigo_: looks great to me, thanks! [12:05] * pitti -> lunch, bbl [12:05] pitti, ok, I'll upload then [12:06] hmm, although not sure I have permissions for accountsservice? [12:06] we are on sync with debian so ideally we would get those fix there [12:07] the debian maintainer is responsive as well [12:07] so maybe open a bug on the bts with the diff [12:07] seb128, ah, ok, I'll submit it to debian then [12:07] thanks [12:07] lunch, bbl [13:12] * rodrigo_ -> lunch [13:32] argh, silly apt === MacSlow|lunch is now known as MacSlow [13:36] pitti, did you remove half of your install in a dist-upgrade? ;-) [13:36] no, but apt-cache now fails to work under fakechroot [13:36] this happened since the last fakechroot update [13:36] currently submitting a debian bug, I have a workaround now [13:36] (it made pkgbinarymangler fail on the buildds) [13:47] jibel: argh, forgot to add the bug# to gtk-3; thanks for closing! [13:49] pitti, no worries. Thanks for fixing it! [13:52] rodrigo_, now that the gtk_icon_source_set_pixbuf bug is fixed, the exit() crash is back on boot. === m_conley_away is now known as m_conley [13:58] hey seb128 - if tomboy doesn't make it we could always just put it in Featured Apps [14:01] jcastro, sure [14:02] salut seb128 [14:09] lut kinouchou === cking_ is now known as cking [14:17] bonjour groupe! [14:19] mpt, ping about Date & TIme [14:20] hey mterry [14:20] hi cyphermox [14:20] hey seb128 [14:20] how are you? [14:20] seb128, hello! [14:21] I wish I could sleep in my bed. the bedroom is still not fixed yet, they're about to start the works though [14:21] what happened to your bedroom? [14:24] pitti, btw I reassigned a bunch of bugs to the teams, I'm not sure how to keep a list of "desktop things that somebody should work on" [14:24] seb128: that's fine [14:25] I tend to cross some when reading my bug emails but I'm not sure how to deal with them so they don't get lost [14:28] mterry, Wed Jun 15 14:28:18 BST 2011 :-) [14:29] mpt, :) I'm working on the gtk3 port of indicator-datetime [14:29] cool [14:29] mpt, and the unlock button has changed appearances, wanted to know how you wanted it to look [14:29] i've got a screenshot [14:29] It will be difficult to get it to look like the mockups (lock button, label label) [14:30] Now lock and label are inside a button [14:30] mpt, http://mterry.name/time.png [14:30] Obviously, I can play with the width and justification [14:30] oh dear [14:31] seb128, i have a list created for that [14:32] pedro_, "that"? [14:32] http://people.canonical.com/~pedro/desktop/ [14:32] check 'assigned bugs' sort by assigned [14:32] oh, I noticed on the wiki yesterday [14:32] \o/ [14:32] but I forgot to ask how you were building it [14:32] mterry, it's good that the label is clickable now, but the border is overdoing it (compare the labels for checkboxes and radio buttons) [14:32] mterry, so, how about you unjustify it and left-align it, and then Cimi can fix the border in the theme. [14:32] mpt, I could probably drop the border, but it would be for the whole thing [14:33] Does that make sense? [14:33] seb128, notice also that i'm listing 'bad bugs' like ones having a good quantity of dups or users affected , so you guys can look and start grabbing some from the list [14:33] mterry, that way we won't be doing something special just for this particular unlock button. [14:33] seb128, i'm not looking much into 'heat' cause sometimes that is not accurate , bugs with 0 users affected gets a high heat and the opposite for bugs with a bunch of users affected ,etc [14:34] mpt, yeah, I can do that. OK. Shall I drop Cimi a line? [14:34] You'd like the whole border gone? [14:34] seb128, i'm using launchpadlib for building the whole thing, is a script that runs every two hours [14:34] (for the general case, not for me to do it in this one case) [14:35] mterry, no, just from the label. I think it makes sense for the icon to still have a border, like checkboxes and radio buttons do. [14:35] pedro_, great list, very useful, thanks! [14:35] pedro_, you should perhaps email the ubuntu-desktop list about it [14:35] pedro_, btw what about moving it next to version to ~platform/desktop rather? [14:35] seb128, you're welcome if you have any feedback just ping me and i can improve the search for your area of interest [14:35] mpt, agreed, but I'll be interested to see how easy that is for Cimi :) [14:35] hmm, seems like nm-applet doesnt autostart anymore since the last upgrade of my recent oneric arm install [14:36] pedro_, that will avoid people highlighting you went copying the url :p [14:36] ogra: what libgtk-3-0 version? [14:36] seb128, sure i just need to do some factoring on the code cause is a 'bit' ugly :-P [14:37] pedro_, well, no hurry, let me know if you need help to move it to the platform location [14:37] seb128, oki doki, thanks :-) [14:37] pedro_, well first fix your code and when you are happy with it we can do that ;-) [14:37] seb128, 3.1.6-0ubuntu1 [14:37] ogra: upgrade to 3.1.6-0ubuntu2 once it's built on your slow arch ;-) [14:38] upgraded 5 min ago... firing up nm-applet through alt+f2 works fine [14:38] ah, k [14:38] waiting for the buildd then, thanks :) [14:38] yw [14:40] jibel, right, my upgrade has finished, so testing and debugging now [14:40] pedro_, can you can a bug number or date column? [14:41] pedro_, I would like to be able to sort new bugs first [14:42] seb128, all the columns are sortable [14:43] pedro_, right, but there is no bug number one :p and sorting by summary doesn't seem to work [14:43] ah ok, yes i can split that part [14:43] pedro_, it's neither sorted by number nor alphabetically, not sure what it's doing [14:43] i was just following the lp way to do it, but that's easy [14:43] * pedro_ takes note [14:43] pedro_, thanks [14:44] thank you for the feedback :-9 [14:44] pedro_, well or get the summary sorting to act on the numbers [14:44] pedro_, otherwise great work, I see it lists all the bug I reassigned to the team recently so I could just have worked from that list ;-) [14:44] pedro_, it's using the desktop-bugs subscriptions right? so we should make sure that's updated for the new GNOME3 components [14:44] i've noticed some really old bugs still assigned to folks on the team [14:44] like bugs from 2007 or so [14:45] so doing a sanity check would be nice [14:45] yeah, we do that every cycle usually [14:45] the issue is that there is no way to make the difference between individual pet bugs and bugs assigned that need work [14:46] seb128, yes it's using desktop-bugs subscriptions and i'm also building special list for bugs not being there, like the telepathy stack, gwibber, mozilla packages , etc [14:46] at the other packages section you'll find more packages not being included on desktop-bugs but that we're maintaining [14:47] and compiz / unity have a separate section cause they're having way more bugs and the list was really not lookable [14:49] pedro_, ok, thanks === zyga is now known as zyga-food [14:50] pitti, did you try to get oneiric retracers running? [14:52] not yet [14:52] seb128: they keep crashing even on natty :/ [14:53] pitti, on what errors? [14:53] not really sure why the amd64 one crashed, though [14:53] the i386 seems a bad bug, untagging [14:53] pitti, I will try to get an oneiric environment on and running at least manually if that's ok with you [14:53] seb128: oh, sure! [14:53] seb128: take the natty chroot and dist-upgrade it? [14:53] robert_ancell mentioned he needs it for lightdm [14:53] pitti, yes [14:53] s/take/copy/ [14:54] I fixed the bug and restarted [14:54] hm, why does a freshly created user have a custom dconf setting for org.gnome.desktop.background picture-uri!? [14:56] I guess the gconf migration might get into the way [14:56] pitti, how custom setting? [14:57] it's set to warty-final-ubuntu [14:57] in ./.config/dconf/user [14:57] i. e. you can't change it any more in the system default [14:57] we could perhaps take the opportunity that there is a migration to change the filename btw [14:59] how could the conversion help us there? [14:59] picture-uri = /desktop/gnome/background/picture_filename [14:59] that just assings the old value to a new key; can you specify value transformations there? [14:59] pitti, well, there is code in g-s-d that converts that to an uri [14:59] hm, I'll try to set both gconf and dconf defaults then [15:00] ooh, that might be it then [15:00] i. e. it's not gsettings-desktop-schemas.convert, but that g-s-d code which writes a non-customized system default into the user's dconf db [15:00] pitti, http://launchpadlibrarian.net/73156984/gnome-settings-daemon_3.0.2-1ubuntu1_3.0.2-1ubuntu2.diff.gz [15:01] pitti, it's likely yes [15:01] hm, no, it doesn't touch gconf [15:01] so I guess the migrator just takes the old gconf system default [15:06] could be, I didn't check if the .convert migrate user keys only or if they migrate key values without making a difference between default and user ones [15:06] yes, I just confirmed it [15:06] rodrigo_, ^ do you know? [15:06] gsettings-data-convert --verbose --dry-run|grep background [15:06] set key 'picture-uri' to string '/usr/share/backgrounds/warty-final-ubuntu.png' [15:07] so I need to set the gconf default as well [15:07] pitti, is that for your custom CD spec? [15:07] it'd be easier if it would just iterate over the user's actual gconf customizations [15:07] seb128: yes [15:07] instead of assuming that it would need to migrate each and every gconf system default [15:07] seb128, hmm, I think it migrates key values, whatever they come from [15:07] pitti, feel free to add a workitems for "just iterate over the user's actual gconf customizations" to the gnome3 spec [15:08] it would certainly help to avoid other cases like that [15:08] it shouldn't be really hard to do [15:08] ok, will do [15:08] and it will avoid getting stucked later because we got system default set as user settings [15:08] which will not reflect default changes [15:09] right [15:09] added [15:09] danke [15:10] the fun thing is that it skips other keys, such as [15:10] Skipping GConf key '/apps/gnome_settings_daemon/plugins/background/priority', no user value [15:16] hggdh, hey, no clue about bug #797710 then, it seems similar to the issue pitti was having earlier which got fixed in 0ubuntu2 [15:16] Launchpad bug 797710 in gtk+3.0 "nautilus crashed with SIGSEGV in gtk_icon_source_set_pixbuf()" [Low,New] https://launchpad.net/bugs/797710 [15:16] did you restart since the update? [15:16] nice, works [15:16] seb128: yes, I had rebooted to guarantee a clean env [15:16] ok, dunno then [15:17] it happens every time you try to run nautilus? [15:17] hggdh: dpkg -l libgtk-3-0 ? [15:17] this looks too similar to be a mere coincidence [15:17] pitti: ii libgtk-3-0 3.1.6-0ubuntu2 GTK+ graphical user interface library [15:17] and apport also reported the same versions [15:18] ok, thanks [15:18] seb128: no, it does not happen every time. I am still trying to zero in [15:18] seb128: what if we go ahead and close it pending a repeat? [15:19] well set it to incomplete maybe and wait to see if it happens again [15:19] k [15:19] hggdh, did it happen with the new version or did apport run with the new version but maybe reporting a segfault from the previous session? [15:21] seb128: it *might* have happened. I usually clean up /var/crash before rebooting; it might have happened in between [15:28] jibel, so, you get the crash in g-s-d at login? [15:29] rodrigo_, yes, so it may crash on shutdown. but it doesn't crash if I launch it manually once the session started. [15:29] rodrigo_: additional yes from here ;-) === zyga-food is now known as zyga-afk [15:47] pitti, "shrug" [15:47] ERROR: connecting to Launchpad failed: [Errno 185090050] _ssl.c:340: error:0B084002:x509 certificate routines:X509_load_cert_crl_file:system lib [15:47] pitti, did you see that before? [15:47] erk [15:47] seb128: not that I remember; try purging the cache? [15:47] ok [15:49] seb128: seeing that with pull-lp-source ? [15:49] Daviey, no, seeing that with the retracers ;-) [15:49] seb128, oneiric? [15:50] *sigh* [15:50] but they use launchpadlib [15:50] hmm [15:50] james_w, no, it's a natty chroot [15:50] seb128: well i'm seeing that consistently with pull-lp-source.. so most certainly a major issue. [15:50] fta saw that and it was fixed by downgrading python-httplib2 [15:51] seb128, bug 797281 [15:51] james_w, ok, I'm lying [15:51] Launchpad bug 797281 in python-httplib2 "LP API broken in oneiric with python-httplib2 0.7.0-1" [Critical,Triaged] https://launchpad.net/bugs/797281 [15:51] seb128, ^ [15:52] james_w, it's likely the issue comes from the retracing environment which is oneiric [15:52] thanks jibel [15:52] let me downgrade it in the chroot [15:52] thanks james_w jibel [15:57] jibel, hggdh: doesn't happen here, so you might have something different [15:57] rodrigo_: we certainly have something different -- bad luck ;-) [15:57] well that's a segfault on session closing [15:57] :) [15:57] rodrigo_: I will delete the crash report, and reboot/relogin, and see what happens [15:57] we should really filter those out from apport in some way [16:07] chrisccoulson: any chance you could do https://code.launchpad.net/~pitti/ubufox/default-homepage-override/+merge/63121 soon? I'm slowly getting to the point where I need it [16:07] pitti - sure, no problem [16:07] i've got another merge to look at too [16:10] mpt: offering capplets in a unity lens> ooh, my hero! [16:10] mpt: that's exactly what I was whining about as well [16:10] mpt: the shell bolted into the session menu has always felt like a hack to me [16:10] Daviey, pitti, jibel, james_w: thanks, downgrading python-httplib2 worked! [16:11] pitti, to be fair, I think he/she was looking at g-c-c in 11.04 and thinking that 11.10 would also just be launchers, not embedded settings windows [16:11] pitti, capplets in a lens? [16:11] having capplets in a lens doesn't really solve the problem of our system settings being a mass of random dialogs though does it? [16:12] they are not in oneiric [16:12] they are embedded in the gnome-control-center ui [16:14] bug #797443 [16:14] Launchpad bug 797443 in gedit "[Oneiric] gedit crashed with SIGSEGV in g_type_check_instance_is_a()" [Medium,New] https://launchpad.net/bugs/797443 [16:14] oneiric retracing! [16:14] \o/! [16:15] bring it on ;-) [16:18] seb128: somewhere in unity, yes [16:18] seb128: wohoo! thanks! === stdin is now known as Guest80131 [16:20] seb128, Thanks! [16:20] pitti, jibel: yw ;-) [16:21] pitti, wasn't mpt's comment on this spec rather suggesting just adding a launcher to start the g-c-c binary rather than a lens? === Guest80131 is now known as ts2 [16:23] well, I might have misunderstood it then [16:24] pitti, or I did ;-) [16:24] in any case let's see what they come with [16:24] pitti, ok, I will pin the old python-httplib in the retracer [16:40] Good night everyone! time for sports and dinner [16:40] 'night pitti [17:32] I was thinking about going out for some fresh air, but outside it's like an oven :( [17:32] but anyway, stopping now for a bit, later all [17:34] good evening [17:59] hey didrocks [17:59] salut seb128 :) [17:59] didrocks, how was your day? [17:59] you joined after most people leaved or took a break [17:59] so low replies ;-) [18:00] mvo, I've got a weird problem. USC 4.0.3 is just starting up with a blank grey window [18:00] seb128: was nice, just arrived in Berlin. Got some work done in ubuntu one sso connection for oneconf [18:00] I now start some fake syncing :) [18:02] didrocks, great [18:02] seb128: how was your day? [18:03] didrocks, good, nothing special [18:03] anyone know why gsettings can't access schemas only installed in .local/share/glib-2.0/schemas ? [18:03] you didn't find 20 MB of additional on the CD by accident? :) [18:03] didrocks, did a few GNOME updates, started looking to 3.1, and I set up oneiric retracers this afternoon [18:04] oh retracers -> nice! [18:04] didrocks, no... [18:04] kenvandine, desrt would know, ping him on #gnome-hackerS? [18:06] seb128: https://launchpad.net/ubuntu/+source/gnome-python-desktop dropping python-gnomeapplet is breaking revelation :/ [18:07] bigon, not dropping it would have broken it as well [18:07] bigon, the new gnome-panel is on gtk3 and can't load gtk2 applets [18:07] seb128, ok... i think it used to work :) [18:07] the code needs to be ported to use gir [18:08] :/ [18:10] I guess the applet part of revelation could be removed [18:12] seb128: the problem here is that python-gnomeapplet is not migrated to dh_python2 and then breaks lot of things on existing installation [18:12] well lot == pkg that use python-gnome [18:12] bigon, didn't pitti fix that? [18:13] well of pkg sharing the same namespace MUST use the same helper system [18:13] bigon, http://packages.qa.debian.org/g/gnome-python/news/20110608T151738Z.html [18:13] " * debian/python-gnome2.postinst: Remove old __init__.py{,c} cruft from [18:13] pysupport, which causes bad imports after upgrades. (LP: #790613)" [18:14] bigon, that was supposed to fix the python-gnomeapplet issue [18:14] * didrocks waves goodnight [18:14] see you tomorrow (half connected probably) [18:14] well just got hit now on my oneiric system [18:14] purging python-gnomeapplet fix that [18:15] hum [18:15] the fix might fix leftover issues once uninstalled [18:15] we should get gnome-panel to conflicts on python-gnomeapplet or something [18:15] or Breaks it [18:15] yeah maybe [18:15] ok, I will add a Breaks [18:16] btw there are still files installed in /var/lib/python-support/python2.7/gtk-2.0 [18:18] looks like it is because $HOME/.local/share isn't in $XDG_DATA_DIRS [18:18] anyone know if that has changed? [18:18] i am pretty sure it used to be [18:23] kenvandine, blame it on robert_ancell [18:23] random guess but seems a login manager environment reading sort of thing [18:23] perhaps [18:24] in fact no [18:24] not important... just took me a bit to figure out why the gsettings work i had done was failing... [18:24] I have an "env" from gdm and lightdm in case I needed to diff [18:24] i think it is a xsession thing [18:24] XDG_DATA_DIRS=/usr/share/gnome:/usr/local/share/:/usr/share/ [18:24] I think that's normal [18:24] I don't think user dirs have ever been in the variable [18:25] it was, we had documented that for making unity launcher changes [18:26] but it was maverick/early natty time frame [18:26] anyway, not important... just confused me :) [18:26] my code works when installed :) [18:27] annoying to have to install with root though to run from a checkout :/ [18:27] * kenvandine grabs some food [18:27] well .local should work without relying on the environment [18:28] what happened to the ppa:Ubuntu-desktop/gnome3-builds repo? it seems to be gone now. any other repos for gnome3 on maverick? === zyga-afk is now known as zyga [21:47] hi guys... anyone know how can I disable automatic logging via gconftool? [21:48] you can't [21:48] seb128: any way to do it from command line? [21:48] the daemon use a key file on disk not gconf for its config [21:48] edit custom.conf [21:48] sed or whatever else the command line gives you [21:49] seb128: ok thanks! [21:50] yw === m_conley is now known as m_conley_away