=== duflu_ is now known as duflu [05:30] Good morning [06:52] hi. I have a paper-cuts issue [06:53] I love Docky, even if it's mono and crashes often (and on every upgrade) [06:53] the issue I'm having is it doesn't hide anymore [06:54] i must have changed some setting in CCSM and now hiding options like intellihide, window-dodge just don't do their thing [07:00] you may wish to try #docky [08:26] Hello, world! [08:28] sarnold, thx [08:28] sarnold, but there's no-one in there [09:04] hey [09:05] stgraber: I'd file a bug and assign it to Robert [09:05] hey Laney, how are you? [09:05] good morning desktopers! [09:09] hey seb128, I'm good thank you [09:10] played board games in the pub yesterday including http://boardgamegeek.com/boardgame/131357/coup which is a fun one ;-) [09:10] you? [09:11] I'm good thanks [09:12] "board games in the pub" sounds like a nice evening ;-) [09:13] my evening was boring compared to that :p some house-cleaning and videogame playing [09:15] heh [09:15] found some people recently who are into games so we get together every week or so to play a couple [09:17] nice! ;-) [09:17] houses need cleaning too :P [09:18] lol, indeed :p [09:18] hum [09:18] daily upgrade is unhappy [09:19] Do we have some glib masters here who can look at bug 1256961? [09:19] Launchpad bug 1256961 in gnome-panel (Ubuntu) "new glib (2.39.1) causes some indicators & nautilus to not load promptly" [Undecided,Confirmed] https://launchpad.net/bugs/1256961 [09:19] mitya57, glib master -> desrt [09:19] though that has "indicator" in the name so maybe larsu can help as well [09:20] I looked at that towards the end of last year after darkxst pinged but I didn't find out what was up [09:20] gnome-session stays in the intialisation phase for ages [09:22] interesting... [09:23] sounds like indicator-applet needs a long time to init and gnome-panel waits for that? [09:24] Maybe... [09:25] Can it be caused by moving indicators to Upstart jobs? [09:26] You see the keyboard and messaging indicator [09:26] then , then the rest and the session is initialised [09:27] Laney: hm. It shouldn't block when loading the indicators [09:27] also, session shouldn't take that long [09:27] is there something different about the first two? [09:27] anyway, try gnome classic in a VM or something and you'll see it [09:28] session does more initialisation [09:28] otherwise I can't think of a difference right now [09:28] they are all ported to upstart, right? [09:29] what indicators are running or not shouldn't make a difference to the loader/panel right? [09:29] ya [09:29] "shouldn't" [09:29] but looking at the last commits in -applet, it does some weird strcmps on indicator names [09:29] hmm, maybe they are the ones that are non-upstart [09:29] the problem still happens if you don't launch under upstart though [09:33] Laney: indeed, messages and keyboard don't have upstart scripts === _thumper_ is now known as thumper [09:33] the other ones do (I checked sound and power) [09:35] nod [09:37] * larsu finds memory leaks [09:37] unlikely to be the problem though :) [09:37] or _a_ problem, for that matter [09:43] Laney: does gnome-panel or the applet emit that weird upstart signal? [09:43] I don't know, probably not [09:43] well ya [09:43] been staying away from that stuff ... [09:43] there's your problem [09:43] me too :) [09:43] what's the timeout? [09:44] it's really really dumb [09:44] Laney: what timeout? [09:44] the indicators all appear after some delay [09:44] well, if the signal is not emited those indicators should just not load [09:44] right [09:44] nothing should be blocked or delayed [09:44] what happens to make that work? [09:44] seb128 finds holes in my theory [09:44] ;-) [09:45] however [09:45] but I can't find "indicators-loaded" or "indicator-services-start" in indicator-applet [09:45] hmm, I think indicators could be a red herring [09:45] if you emit the signal manually then the session still doesn't initialise [09:45] but the indicators do appear [09:46] ah [09:46] maybe the signal is emitted after whatever is causing the delay in gnome-sesion [09:46] Laney, what's the issue? what is blocking? [09:46] gnome-session doesn't get into the "Running" phase [09:46] not sure if that is the cause or a symptom [09:47] you see that nautilus isn't handling the desktop yet, no indicators and can't logout via gnome-session-quit [09:47] if you can reproduce, drop indicator-applet and see if that's still happening? [09:47] also gnome-session --debug log would be useful [09:47] I bet anyone can, try flashback in a vm [09:48] Laney: wfm in a guest session [09:48] does it matter which flashback? [09:48] I have a compiz and a metacity one [09:49] I'm using the metacity one [09:50] that's what I tried as well [09:50] hm, I have glib 2.39.2 though, hope that's not a problem [09:52] no, that's what I have [09:53] wtf [09:53] guest session always gives me unity [09:55] yeah, same here, that's not new [09:55] wfm?! [09:55] I didn't have the spare cycles to investigate though [09:55] * larsu feels special these days [09:55] haha [09:55] larsu, that's probably another thing you hacked up and don't remember ;-) [09:56] anyway, try with a real user [09:56] seb128: I would never do that! [09:56] sssssure ;-) [09:59] okay I have both issues after rebooting [10:01] http://paste.ubuntu.com/6719886/ [10:04] Laney: it seems to think it's in shutdown mode?! [10:04] or does that log start with an old session [10:05] the start is me logging out [10:05] the new one begins around 325 [10:07] hm, this happens for some indicators: gnome-session[12583]: WARNING: Could not parse desktop file indicator-sound.desktop or it references a not found TryExec binary [10:07] that's because I removed them [10:07] but it keeps the xdg autostart file [10:08] ah okay [10:20] guys, that's the bug i've been complaining about for weeks [10:23] Laney: remember i told you the workaround for it? [10:23] if you mean the no watchers thing then it's not that [10:24] ali1234, did you debug it by chance? ;-) or just complained about it? [10:24] ali1234, I've an endless list of bugs I would like to complain about, that's not going to solve them though :/ [10:25] i debugged it, yes [10:25] then i wrote a fix [10:25] then i sent a MR to tedg [10:26] that was before christmas [10:26] neat. Where is it? [10:26] spoke to him about it yesterday or the day before [10:26] didrocks: can I make a release of session-migration? [10:26] the MR is attached to the "indicators should use upstart" bug report [10:26] Laney: depends, do you need it quickly? I was wondering to not put it to dailies [10:26] if it's urgent, please do ;) [10:26] the reason that people see indicators not loading is because they will quit if nothing claims them within 500 msec [10:26] only in as much as I would like to move on with life [10:27] ahah ;) [10:27] so if your gnome-panel takes 1.5 seconds to load up, then yeah, half the indicators will disappear [10:27] ok, feel free, just tell me ;) [10:27] I'll do it, /and/ you can put it under daily release for the next changes [10:27] :-) [10:27] ali1234: this? https://code.launchpad.net/~a-j-buxton/libindicator/remove-timeout/+merge/198070 [10:27] larsu: that's the one yes [10:28] you can also disable the timeout by setting that environment variable [10:28] without patching i mean, if you just want to confirm this is the issue [10:29] ali1234, seems orthogonal to the bug Laney is seeing, in his case the session doesn't fully init (no nautilus bg either, closing session doesn't work)à [10:29] this is a separate issue to the bug at hand [10:30] ali1234: there's not much left using IndicatorService. Are you sure that's the issue we're seeing here? [10:30] also, I fail to see how this would block gnome-panel [10:30] hm, indicator-application maybe? [10:30] yeah i guess you're right. it's a separate bug that is having a knock on effect. if desktop startup is delayed you *will* see this bug too [10:31] I think one of the first things I tried was setting that variable :P === davmor2_ is now known as davmor2 [11:35] Laney: looking at this again. Is gnome-panel started twice? [11:42] larsu: Not sure [11:42] that 'failed to register' message might be the clue [11:42] I'm pretty sure it is [11:43] and if it's started twice, then the second instance never inits [11:43] because it tries to get the name on the bus synchronously [11:46] * larsu is just theorizing while makeing some food. Might be all wrong [11:58] I tried rolling back to the saucy versions of gnome-panel, g-s-d and gnome-session but it's the same [11:58] maybe it really is some glib change === tkamppeter__ is now known as tkamppeter [12:21] Laney: thinking about session-migration, this is working on the destkop as well (through upstart session), right? [12:21] yes [12:21] so, maybe I should remove the hack in gnome-session [12:21] already changed that [12:21] oh great ;) [12:22] I didn't remove it, because then you wouldn't get migration if you aren't using upstart [12:22] Laney: well, I think only our session is really using session-migration TBH [12:23] was easy to keep it flexible [12:23] how do you do it? just for my interest ;) [12:23] checking for UPSTART_whatever [12:23] ah ok, just env variable checking [12:23] nod [12:24] hmm [12:24] do I need to specify the deps for the migration script manually? [12:24] not if you are using the dh_migrations [12:25] like the python libraries it's using [12:25] you have a sequence as well dh --with migrations if needed [12:25] yeah, using that [12:25] ah, the script itself [12:25] yeah, session-migration just exec it, doesn't care about the langage or what the script does [12:26] so you need to have your script ensuring it can run (so bringing its own dep) [12:26] Laney, is there any gnome-session based desktop not using upstart? [12:27] gnome-xmonad :-) [12:27] now that's crazy talk :p [12:27] yeah, who really uses that :P [12:28] I'm surprised how long I wrote session-migration, and it's almost 2 years after that I see a lot of people interesting using it :) [12:28] didrocks, we have been using it for a few things in previous cycle, so in has already been useful ;-) [12:28] (e.g migrating the nautilus launcher icon in unity, when the desktop name changed) [12:29] it's definitely interesting for some phone cases too [12:29] yeah, but not as much, I'm happy anyway. It was just a couple of days, but well invastigated :) [12:29] investigated* [12:29] invested* [12:29] grrr :) [12:30] hehe [12:30] indeed [12:30] Laney: btw, good news, you are going to be credited again for the system-settings uploads :) [12:30] credited in what way? [12:30] even people committing to trunk directly will be credited, we shouldn't ;) [12:30] in Changed-By? [12:30] your name in debian/changelog [12:30] like the [ foo ] [12:31] oh [12:31] as you had with daily release [12:31] did that stop working? [12:31] well, on purpose, yesterday [12:31] with the new system [12:31] fair enough [12:31] I didn't notice ;-) [12:31] ahah ;) [12:31] it was on purpose because of… it's complex in case someone brings a branch merging multiple ones [12:31] and you have to compare with potential direct commit to trunk [12:31] and so on… [12:32] but now that we changed slightly the procedure, I rewrote that and it's possible [12:33] Laney, I submitted a "add the cmake.user to bzrignore" before noticing you sneaked in a similar change in an merge request which had nothing to do the topic ... sorry about the dup, but what do you want to do about that? should I reject mine, or do you want to rebase yours on my branch so we have logical changesets/commits? ;-) [12:34] approve yours [12:34] and I'll re-merge it [12:35] Laney, https://code.launchpad.net/~seb128/ubuntu-system-settings/bzrignore-cmake-user/+merge/200988 if you want to +1 it [12:36] done, thanks [12:36] the new world order when approved merges land only after upload is going to be annoying for rebases and stuff though :/ [12:36] I can just merge with the branch instead of trunk [12:36] right [12:36] but you have to know which ones to do I guess [12:36] hmm [12:37] btw for those who didn't notice, we have unity-control-center in universe [12:38] and gnome-control-center's binary is that same source is now a wrapper that call u-c-c if available or g-c-c.real [12:39] wooot [12:39] ubuntu-themes/unico landed, the CI team is rolling nowadays ;-) [12:40] larsu, ^ your fixes from yesterday are already uploaded to trusty, yeah for working dnd again and nautilus stable columns ;-) [12:41] that was quick! === Sweetsha1k is now known as Sweetshark [12:44] isn't it? :) (thanks to Mirv) === Trevinho_ is now known as Trevinho [13:17] :) === alan_g is now known as alan_g|lunch === dobey_ is now known as dobey === alan_g|lunch is now known as alan_g [14:48] Laney: ok, will do [15:01] whoa, 15 degrés, c'est comme le printemps ! [15:01] Laney, kenvandine, tedg, charles, attente: want to do a settings meeting? [15:01] pitti, ou l'automne :/ [15:01] il pleut ici aujourd'hui [15:01] it seems between the US and Europe the average temp is just fine, but in reality it's totally wrong everywhere [15:02] seb128: non, non, nous avons du soleil ! [15:02] I just did a really nice short walk to think about how to fix these apport popups after logging in [15:02] how come we still get some of those? [15:03] didn't you add code to stop prompting about issues from other sessions? [15:03] seb128: for processes that crash during logout, apparently there are still a lot of them [15:03] or did we just flag them with some tag? [15:03] seb128: don't mind, can do if you want [15:03] Laney, is your mic working in 2014 ? ;-) [15:03] I looked at tests before the holidays and this week have been working on AS [15:03] seb128, i can if others want to [15:03] seb128: so far I just have code to stop attaching xsession-errors to reports from other sessions [15:03] [15:03] umm, didn't change anything there ;-) [15:03] unless google fixed it for me [15:03] seb128: balloons poked me about bug 1033932, related to bug 1067646 [15:04] Launchpad bug 1033932 in apport (Ubuntu) "Error alerts appear on login for problems in previous session" [Medium,Triaged] https://launchpad.net/bugs/1033932 [15:04] Launchpad bug 1067646 in Apport "should report if crash happened at logout/shutdown" [Medium,Triaged] https://launchpad.net/bugs/1067646 [15:04] * kenvandine has done nothing more on uss than a few reviews this week :) [15:04] Laney, kenvandine: ok, let's have a quick one, just to resume/keep the rhythm, otherwise we are going to keep skipping them as we did before holidays [15:04] what happened with the person that was supposed to take over wifi/bluetooth? [15:04] oh okay! [15:04] what? [15:04] Laney, bluetooth is still charles afaik [15:05] cyphermox, hey, what what? [15:05] hahaha [15:05] trigger word [15:05] Laney: you trying to tell me I got fired? :) [15:05] Laney, Wellark is supposed to do wifi [15:06] I thought it was both [15:06] but ok [15:06] cyphermox, did you just volunteer to finish the touch system settings for bluetooth? [15:06] I didn't think I did [15:06] tbh, I'm not sure I have the time to do that too [15:06] seb128, i'm all alone... [15:06] kenvandine, stupid google keeps logging me out every time I close firefox [15:06] huh, weird, I had a spam event on my gcal [15:06] cyphermox, that's what we were discussing ;-) [15:06] never seen that before [15:07] Laney, tedg, charles, attente: anyone coming? [15:07] it's the seb128/kenvandine show :) [15:07] Sorry, sure [15:07] Though, I don't wnat to interrupt that :-) [15:08] that's fine, Laney already did [15:08] But everyone knows that hangouts never work for Laney [15:08] he keeps trying though [15:08] you have to admire the motivation [15:09] Heh, insanity is trying the same thing... ;-) [15:10] .. this time it's different! [15:12] you guiz [15:17] kenvandine, https://code.launchpad.net/~unity-team/ubuntu-system-settings/welcome-wizard/+merge/186862 if you want to help reviewing [15:17] Laney, did you want to block on the upstart stuff? [15:17] not as such [15:19] wah, I looked at it again [15:19] maybe xnox or one of the guys can advise on a better way [15:20] hm? [15:20] we should get him to update the branch from latest trunk to get another CI build [15:20] so we can get debs to test [15:22] Laney: seb128: that job looks odd, as it will unconditionally run welcome wizard, and if that exits non-zero it will block starting the session / unity8. [15:22] mterry_, ^ [15:23] i guess the latest debs aren't that outdated from trunk, but there has been some change [15:23] Laney: seb128: if it's not ready now, it's best to have pre-start check which checks for a magic file /to/ run wizards, instead of current logic of running it unconditionally. [15:24] Laney: seb128: or otherwise guarantee that our base images / initramfs does "touch ~/.config/ubuntu-system-settings/wizard-has-run" [15:24] Laney: seb128: and those who are testing the wizard, would remove that file and reboot. [15:24] does the dconf bridge work? [15:24] Laney: it does, but nobody used it yet =) [15:25] could use that [15:25] i think i prefer the file myself [15:25] anyway, I'm more concerned about the start_xsession() behaviour [15:25] Laney: so no idea if does what one expects. tedg tried to use it, but it wasn't what he wanted / needed. [15:25] yeah "feature-flag" file is easier and can be controlled from e.g. recovery / adb /etc [15:26] right [15:26] and phablet-flash could tweak it easily too [15:27] or one of the phablet tools [15:27] for automated testing [15:28] xnox, I'm confused about the changes you want. You want something outside the job to create the wizard-has-run file? [15:30] mterry_: i want pre-start script to have: [ -f ~/.config/yes-i-want-to-try-welcome-wizard ] || { stop; exit 0 } [15:31] mterry_: cause at the moment if welcome wizard crashes, or exists non-zero, the session/unity8 will not start. [15:31] Laney, start_xsession has two parts: 1) stopping maliit server which is a workaround for a maliit bug and presumably will go away and 2) starting xsession manually which was done so that the wizard can stay running / animating while session prepares itself in background. Presumably that will go away too once unity-system-compositor gains nice animating behavior between sessions [15:31] Laney, so while I agree that start_xsession is hackish, it's a temporary hack [15:32] xnox, wouldn't upstart just continue with xsession-init then? [15:34] mterry_: let me test here to confirm. [15:36] mterry_: seems to work, never mind me then. [15:36] xnox, also, my branch doesn't even install the upstart file right now, because this is all a work in progress [15:36] seb128: Laney: never-mind, if welcome crashes / exits non-zero, xsession-init will start as far as i can tell here. I got confused. [15:37] are there bugs for these issues? [15:38] Laney, for unity-system-compositor, yes... For the maliit one, I think I just talked about it on IRC at the time with the devs. Don't think I filed one [15:39] mterry_: would you be able to file one so that it's tracked? [15:40] Laney, uh, I'll add a note to test again and file if it's still an issue, sure [15:40] would be nice [15:40] This was early days of maliit, hopefully it's better now [15:44] seb128: Hmm [15:45] rvr, hey Victor [15:45] seb128: http://paste.ubuntu.com/6721361/ Is that also what you get? [15:45] rvr, did you figure those autopilot issues out? [15:45] StateNotFoundError: State not found for class '*' and filters {'objectName': 'aboutPage'} [15:46] looks right [15:47] rvr, what Laney said [15:47] It's strange. Using the introspector tool, I can see the aboutPage object === alan_g is now known as alan_g|tea [15:47] http://paste.ubuntu.com/6721373/ [15:48] > objectName: aboutPage [15:48] It's there === gatox is now known as gatox_lunch === greyback is now known as greyback|away === alan_g|tea is now known as alan_g [16:07] slomo: reckon I could get added to pkg-gstreamer? [16:08] Laney: you are not yet? [16:08] nope [16:08] cyphermox, btw since your reacted to the bluetooth comment earlier ... Chad's patch for gnome-bluetooth is in the sponsoring queue ;-) [16:08] Laney: sure, what do you want to break? :) [16:08] Laney: what's your alioth username? [16:08] laney [16:08] We've got some autoreconf-related changes that could go to Debian [16:09] Laney, I was pondering fwd that earlier, but I'm not sure it's non-buggy [16:09] which one? [16:09] is Debian's libtool including the --as-needed stuff? [16:09] Laney: ah those. i don't think they're correct [16:09] no it doesn't [16:09] :/ [16:09] you need to run dh_autoreconf with --as-needed [16:09] Laney, the issue is that autoreconf is going to replace the ltmain.sh hack for as-needed [16:09] does that work? [16:09] it does the patch for you [16:10] oh? that's nice :) [16:10] how does that work? is dh-autoreconf including the ltmain.sh hack? [16:10] it ships the patch for you [16:10] so they can update it centrally [16:10] nice [16:11] I didn't know about that [16:11] ok, so then only be careful to not break translations ;) [16:11] because we apply a patch after gettextize in gstreamer [16:11] mmkay [16:11] fixed in 1.3 finally, now plain autoreconf works [16:11] in which package is this? [16:13] all of them :) [16:13] I see, not a debian patch then [16:13] gstreamer, gst-plugins-base,good,bad,ugly,*, gst-libav, etc [16:13] yes, not a debian patch [16:13] that's why export AUTOPOINT=true is in debian/rules [16:13] so keep that [16:14] i added you btw [16:14] excellent, thanks [16:14] thanks for merging these patches :) [16:14] np [16:14] please make sure that change is in all the packages [16:14] not only the ones that were important enough to get a bugreport ;) [16:15] I abused some people to forward them [16:15] then decided JFDI is more efficient sometimes [16:16] yes, or telling me about that --as-needed feature of dh-autoreconf would've been sufficient too ;) [16:16] hrhr [16:16] but now you please change that, i'm busy with other things :) [16:24] seb128: ack [16:48] oh no [16:48] glib started ftbfsing [16:48] seb128: FYI, http://bazaar.launchpad.net/~apport-hackers/apport/trunk/revision/2746 [16:48] Laney, :-( [16:48] seb128: apport popups for logout problems be gone [16:48] pitti, \o/ [16:48] seb128: I almost never get those, but it seems a lot of people do, so I wasn't aware that it's such a big problem [16:48] so let's ditch them for the lTS [16:51] pitti, it's quite frequent in some of "services" (indicators, g-s-d, gvfs) [16:52] is/was, we fixed some of those bugs [16:52] seb128: well, do you think they are interesting enough to re-enable them in trusty+1? [16:52] like indicator used to g_error when the session bus was missing, which happens on logout [16:52] not atm [16:53] ok; it's a simple change to re-enable, but at least the logic is there now [16:53] it'll break all over again once /sys/fs/cgroup/ becomes unreadable, but that's still some way out [16:53] (and once it does, we need to talk to the cgroup service instead) [16:53] or logind directoy [16:53] directly === gatox_lunch is now known as gatox [16:57] bonne nuit tous le monde ! [16:57] "tout" [16:57] au revoir [16:58]