=== skaet is now known as skaet__ === pitti_ is now known as pitti [03:50] Good morning [07:01] good morning everyone [07:05] hey chrisccoulson, how are you? [07:05] pitti - yeah, not too bad thanks. how are you? [07:06] i'm battling with a bit of an unreliable connection today :( [07:06] chrisccoulson: je suis bien, merci! j'ai ecrit une test de smb pour gvfs [07:07] actually, un test [07:07] excellent, good stuff :) [07:07] i'm going to have to start learning french ;) [07:09] j'apprends le français avec duolingo.com -- c'est bon! [07:09] chrisccoulson: I had thought that seb128 forced everyone to do that now :0 [07:09] heh, i think he'd like to do that ;) [07:55] seb128: bonjour mon ami, ça va? [07:56] hey desktopers [07:56] pitti, salut, ca va bien, et toi ? [07:56] hating compiz! :) [07:56] je suis bien [07:58] I got bug 1042041 again, and it keeps forgetting my keybindings [07:58] Launchpad bug 1042041 in compiz "1:0.9.8+bzr3319-0ubuntu1 regression: wrong auto-raise, cannot be disabled" [Undecided,New] https://launchpad.net/bugs/1042041 [07:59] pitti, je "vais" bien btw ;-) [08:00] seb128: pardon, d'accord! [08:01] pas de soucis ;-) [08:02] smspilla|z, ^ could you look at this compiz bug when you are around? [08:03] pitti, ctrl-alt-t got fixed in gnome-control-center,gnome-settings-daemon by jbicha this w.e, did you restart your session yesterday? [08:04] seb128: yes, twice; and today three times more [08:04] hum [08:04] fun, this time ctrl+alt+t works, but I get that autoraise misbehaviour [08:04] can you set it in the g-c-c ui? [08:04] I can always set back the keybinding for "lower window behind all others", but it keeps forgetting it [08:05] seb128: g-c-c doesn't allow configuring FFM or autoraise [08:09] pitti, I was speaking about ctrl-alt-t specifically [08:09] seb128: ah, that one always seems to be set [08:09] it forgets about my alt+b binding for "put window to the back" and also show it as disabled in g-c-c, but hte ctrl+alt+t launcher is always shown (but doesn't work often) [08:10] seems there's a weird bug in the gsettings port [08:11] * pitti runs gsettings set org.gnome.desktop.wm.preferences auto-raise-delay 10000, which does seem to work [08:11] pitti, do you know what's the name the key for "put window to the back" ? [08:12] org.gnome.desktop.wm.keybindings lower ['b'] [08:13] pitti, is your issue only with keys under org.gnome.desktop.wm.keybindings? [08:13] or do you have issues with compiz schemas keys? [08:13] j'ai ecrit un nouveau test de smb:// pour le gvfs [08:14] . o O { is gvfs male or female? } [08:14] "pour gvfs" we would say :p [08:14] seb128: windows+up/down are also broken right now; is that a compiz key? [08:14] seb128: what? no article? :-) [08:15] no, I'm trying to think what the rules is [08:15] like you would say "pour Canonical" [08:15] those things as proper name don't have an article [08:15] le français n'est past facile [08:15] non, il ne l'est pas ;-) [08:15] seb128: ah, merci [08:18] pitti, I'm trying to figure what those actions are [08:18] pitti, that's the maximize,unmaximize,minimize key right? [08:18] right [08:18] org.gnome.desktop.wm.keybindings minimize ['KP_0'] [08:18] that looks strange [08:19] org.gnome.desktop.wm.keybindings maximize ['Up'] [08:19] that too [08:19] pitti, ctrl-super-up,down works here [08:20] but isn't it suposed to be win+up/down? [08:20] at least I configured it that way when they changed the default [08:20] and sometimes it works, and sometimes not [08:21] pitti, where did you configure it? [08:22] in g-c-c [08:23] pitti, what's the name of the action you use? [08:23] pitti, those are the max,restore win? [08:25] seb128: right; they were shown as ctrl+super+up/down in g-c-c, I just reset them back to super+up/down and that works now [08:25] (until I restart once or twice) [08:25] pitti, do you run another session than unity sometimes? [08:25] not for my user [08:25] pitti, those keys are common ones, I wonder if gnome-shell or something change them [08:25] hum, k [08:25] well, if compiz was making wrong use of the keys I would say it's a compiz bug [08:26] but it seems the value of those keys change on disk [08:26] so I'm a bit unsure what's going on [08:26] otherwise, https://launchpad.net/ubuntu/+source/compiz/1:0.9.7.4-0ubuntu1 [08:26] * debian/patches/ubuntu-config.patch: [08:26] Ctrl + Super + Cursor up Maximises the current window [08:26] Ctrl + Super + Cursor down Restores or minimises current window [08:26] [08:27] not sure if that's what we stayed on for precise or if that's one we reverted [08:27] * seb128 checks the source [08:29] I think ctrl+super is the current default [08:29] pitti, yeah, we have ctrl-super for those [08:29] so the issues is that sometimes my settings get reverted [08:29] so, something reset your user config sometime, will be fun to figure the something ... [08:29] and that autoraise behaviour is strange, as we do not defualt to autoraise [08:30] which one is that? [08:30] and unlike for the keybindings, the gsettings key for autoraise looks correct [08:30] err [08:30] or so it was yesterday [08:30] org.gnome.desktop.wm.preferences auto-raise true [08:30] org.gnome.desktop.wm.preferences auto-raise-delay 10000 [08:30] now it seems it got set to true [08:30] #)*$# [08:30] :-( [08:30] so your bug is basically "something plays with gsettings keys and reset them" [08:30] well, at least that makes it more consistent [08:31] seb128: yes, except for autoraise [08:31] ? [08:31] * pitti resets autoraise and a-r-delay now and will check at next login [08:31] I though you just said the key got resetted as well? [08:31] seb128: no, the default for aut-raise is false (as it should be), and something set it to true [08:32] seb128: can you tag is with "gsettings" ? [08:32] smspilla|z, sure === smspilla|z is now known as smspillaz|class [08:33] pitti: Is it bug 1022743 ? [08:33] Launchpad bug 1022743 in unity "Shortcuts in CCSM reset after getting changed." [Low,In progress] https://launchpad.net/bugs/1022743 [08:33] not the same thing no [08:33] probably a thinko when I ported the integration code :1~ [08:34] pitti: right, compiz is playing with those keys :] [08:34] pitti, auto-raise is false in a guest session for me [08:34] tag the bug gsettings and I'll fix it [08:34] seb128: right [08:34] I wasn't able to write an autotest suite for the integration code [no time] [08:34] smspillaz|class, why would it write those keys at all if you don't use ccsm? [08:35] seb128: so next time I'll check if it really resets the key bindings, or just sets them to a different value [08:36] seb128: as for the autoraise key, that can't be a reset, it must actually set the key to true (for whatever strange reason) [08:36] could it be that there is some permanent gconf migration going on, and I have some gconf settings somewhere? [08:36] pitti, I get you can edit the schemas and put a weird default value (and run the glib-compile-schemas) helper for that [08:37] seb128: it just synchronizes keys [08:37] smspillaz|class, synchronize with what? [08:37] seb128: there are some gnome keys that it synchronizes with each other for non-trivial reasons [08:37] but let me have a look into it [08:37] ok [08:38] seb128: anyways, please tag any such regressions with "gsettings" [08:38] this fix is also key-related and needs approval: https://code.launchpad.net/~mc-return/unity/unity.merge.fix-hardcoded-keys-part1/+merge/121545 [08:38] I did use ccsm in the past for some bits (configuring FFM mostly) [08:38] smspillaz|class: I'll tag my bug [08:38] I expected there might be a few, I had to rush the integration code [08:38] smspillaz|class, done [08:38] pitti: btw, FFM is not a supported usecase :] [08:38] smspillaz|class: well, but changing key bindings in g-c-c certainly is? [08:39] sure, just keep that in mind [08:39] the bug priority will likely be downgraded [08:39] smspillaz|class: FFM is the one thing that never broke during this, just my keybindings and enabling autoraise [08:39] sure [08:39] pitti, gsettings-data-convert --dry-run --file /usr/lib/compiz/migration/compiz-profile-active-Default.convert [08:39] just letting you know :1~ [08:39] :] [08:39] pitti, does that work or hit a schemas bug? [08:40] smspillaz|class, btw are the .convert upstream or ubuntu packaging stuff? [08:40] seb128: no error [08:40] seb128: and it seems to get the right values (after I set them in g-c-c) [08:40] Set key 'maximize' to string 'Up' [08:40] Set key 'unmaximize' to string 'Down' [08:40] Set key 'lower' to string 'b' [08:41] pitti, so it's not likely the migration running over again and screwing your values [08:41] pitti, does that include auto-raise? [08:42] seb128: no, the only hit of "raise" is the 'raise' keybinding [08:43] pitti, ok, in fact auto-raise is migrated by /usr/share/GConf/gsettings/wm-schemas.convert [08:43] which I guess is writen as migrated in .local/share/gsettings-data-convert for you [08:44] e.g has been migrated and will not be again [08:45] confirmed [08:46] * pitti loves the word "téléchargement" [08:46] ;-) [08:46] pitti, running your desktop in french? [08:47] oui [08:47] err [08:47] j'aime le mots "téléchargement" [08:47] "le mot" [09:05] hey seb128, how are you? [09:08] chrisccoulson, hey, good thanks you, how are you? no rain today yet? :p === larsu_ is now known as larsu [09:08] seb128, heh, it's sunny at the moment [09:08] chrisccoulson, can't you use your phone as a modem btw? [09:08] not sure how long that's going to last though ;) [09:09] seb128, i tried that too, and that also didn't work [09:09] it seems everything is regressing in quantal ;) [09:09] my touchpad doesn't work properly either [09:09] although, that is bug 1041594 [09:09] Launchpad bug 1041594 in linux "Edge scrolling on touchpad broken since the upgrade to 3.5.0-11" [Medium,Incomplete] https://launchpad.net/bugs/1041594 [09:11] chrisccoulson, boot a precise kernel? [09:12] chrisccoulson, you probably still have it installed since we never clean those :p [09:14] seb128, i'm currently using the 3.5.0-10 quantal kernel, which works fine === seb128_ is now known as seb128 === smspillaz|class is now known as smspillaz === MacSlow is now known as MacSlow|lunch === tkamppeter_ is now known as tkamppeter [11:50] seb128: "13:47 <@pmladek> caolan: 3.6.1 should go out tomorrow if nothing happens; so it is basically done" <- we have a ppa build for that, but I will prepare one for quantal-proposed today, okay? [11:51] Sweetsha1k, works for me [12:02] awesome, i've got all of the ffox/tb branches all ready for the next set of releases, by lunchtime :) [12:02] can i haz 6 weeks off now? ;) [12:02] if you dont need them to upload through your 3G :) [12:02] ogra_, chinstrap ;) [12:02] haha [12:02] cheater ! [12:02] chrisccoulson, are you done with your workitems? do we have overlay scrollbars yet? ;-) [12:03] i've got the next set of releases all ready before we've even got the current one out ;) [12:03] seb128, hah :) [12:03] i knew there was something else!" [12:03] ;-) [12:32] whaaaaaaaaaaaaaaaaaaaaaaa [12:32] the upgrade wants to remove vim! [12:32] chrisccoulson, welcome to the world of emacs! [12:32] * pitti gets the torches ready [12:33] heh [12:34] chrisccoulson, it's firefox's fault right? [12:34] ;-) === MacSlow|lunch is now known as MacSlow [12:37] heh [12:37] of course ;) [12:56] seb128, you have sent out the reminder but it seems that you have forgotten to create https://wiki.ubuntu.com/DesktopTeam/Meeting/2012-08-28 [12:57] does http://summit.ubuntu.com/uds-r/sponsorship/review/ work for anyone else? [12:57] i just keep getting looped back to the login page [13:00] chrisccoulson, same for me. [13:02] chrisccoulson, tkamppeter: it's not supposed to work, jasoncwarner made an error with his email, it's only available to ~uds-organizers [13:03] aha [13:03] you should tell jasoncwarner_ (or me) if you want somebody added or voted [13:03] tkamppeter, oh yeah, I forgot, seems like kenvandine just did it [13:03] ? [13:03] hey kenvandine, mterry, good morning ;-) [13:04] create the page? [13:04] good morning [13:04] kenvandine, seb128, you have sent out the reminder but it seems that you have forgotten to create https://wiki.ubuntu.com/DesktopTeam/Meeting/2012-08-28 [13:04] :-D [13:04] * kenvandine took matters in his own hands [13:04] :-p [13:04] thanks man ;-) [13:04] np [13:04] seb128, i think our bugs are related [13:04] kenvandine, I told you! ;-) [13:04] or rather the one you found is causing mine [13:05] but i think there is a bug in LMM too :) [13:05] kenvandine, is that xchat being buggy, did you figure out? [13:05] so here's what i think is happening [13:05] * seb128 listen [13:05] when i focus the chat it calls remove_source on one of the indicators [13:05] for the one xchat says is focused [13:05] at the same time, it is clearing the indicator i clicked on [13:05] so removing that one [13:06] oh, that disconnect the signal? [13:06] i think the one i am not clicking on, is sometimes not getting removed [13:06] or that [13:06] that makes more sense though [13:06] so i think that is why when we click on the other one [13:06] we don't get the callback [13:06] because of xchat [13:06] i can't figure that part out though [13:07] i don't see how that could have anything to do with our code [13:07] puzzling [13:07] yeah, me neither [13:07] but you said it was working before those updates? [13:08] yes [13:08] i checked last night [13:08] in precise [13:08] seb128, under the sponsorship applications should be Kai-Uwe Behrmann, one of the two most important people (together with Richard Hughes) for color management under Linux. I would very much like if he gets funded to get to UDS. [13:08] i queued up 5 indicators [13:08] and went through them [13:08] tkamppeter, ok, thanks [13:08] they all worked [13:08] kenvandine, thanks for setting up the page. [13:08] bo [13:08] np [13:08] :-D [13:11] kenvandine, on quantal how often does it not work? does the "not work" only concerns the last "active" channel (e.g the one that xchat consider selected wrongly and which leads to a source being cleared when it should not)? [13:12] at least 50% [13:12] seb128, but i really can't see why [13:12] it's very bazaar [13:13] kenvandine, indeed :-( [13:13] ok, channel, if you feel adventurous: [13:13] https://launchpad.net/~unity-team/+archive/release updated, testing welcome although do note there are already a bunch of known bugs so check for duplicates [13:14] [13:14] that's the ffe unity,compiz-gles candidate stack [13:14] works fine out of minor details for me [13:14] and webapps right? [13:14] :-D [13:14] but having extra testing would probably be welcome [13:14] kenvandine, you wish :p [13:16] * ogra_ would love to help testing if there were any arm packages :/ [13:16] ogra_, do you have a non virtual ppa to throw those sources at? [13:16] especially since i just uploaded the (hopefully) final fix for our driver [13:17] seb128, oh, indeed i should just be able to copy them to the canonical-arm ppa [13:17] ogra_, you should yes [13:17] ogra_, that would be great if you could try that ;-) [13:21] seb128, argh [13:21] if the packager would have actually allowed any arm arches in debian/control this would actually have worked :( [13:22] ogra_, unity you mean? [13:22] * ogra_ glares at https://launchpad.net/~canonical-arm-dev/+archive/ppa/+builds?build_state=pending .... nux, unity, compiz and bamf are all waiting for amd64 and i386 builds now [13:22] ogra_, I guess that's part of the workaround that was used by then when gles was not ready yet and nobody wanted to update the compiz patch [13:22] yep [13:22] might be [13:23] or they just got tired of getting failure mails for the meissing arches [13:23] *missin [13:23] bah [13:24] ogra_, I'm pretty sure Didier did that before a2 or a3 because he needed to land the updates and nobody wanted to update the gles patch [13:25] well, cant test then ... at least not without fiddling with the pakages which i dont have time for right now [13:26] ogra_, right, I will get that fixed and ping you for a new ppa copy [13:27] seb128, great, else i can do it tomorrow myself [13:27] just not today [13:31] ogra_, ok, let's see how things go [13:39] oh, fun amd the copied packages actually FTBFS on i368 [13:39] ogra_, yeah, their build-depends are not well versioned, you need the new bamf and nux [13:40] which are in the same PPA (and copied) ... but if they arent versioned it wont o into dep-wait indeed [13:40] right [13:40] hmpf, somehow my g key doesnt like me on that new kbd [13:59] hmmm, bugger, just updated my thunderbird nightly and it crashes every time i click on a source in the messaging indicator [13:59] that didn't happen this morning [14:02] chrisccoulson, :-( [14:03] chrisccoulson: btw if you want I can help making your 3G dongle work ;) [14:03] seb128, this one is going to be an absolute joy to debug: http://paste.ubuntu.com/1171850/ [14:03] crashes deep in the JS engine ;) [14:03] :-( [14:03] cyphermox, oh, i'm not sure what's going on with that. i've rebooted again in to the 3.5.0-10 kernel, and it worked again :/ [14:03] it seems quite random [14:04] ok [14:04] did you use a different usb port? [14:04] cyphermox, it's an internal 3G card [14:04] I have at least one dongle that refuses to work unless I use my USB 3.0 port [14:04] ah [14:04] chrisccoulson: internal 3G from the thinkpads? the Gobi horror? [14:09] cyphermox, it's one of these: http://search.dell.co.uk/1/2/128267-mobile-broadband-internal-dell-wireless-5540-card-3g-hsdpa-sim-not-included-kit.html [14:10] oh, man, there is a huge black cloud moving over here [14:11] chrisccoulson: do tell, how is fiber supposed to be affected by rain? :) [14:13] ogra_: what did you discuss about compiz gles on arm? (I only joined after) [14:13] cyphermox, the bit from the cabinet to my house is still copper ;) [14:13] ogra_: I'm compiling atm on arm without the patch just to see if the current problem is merely packaging [14:13] (but the cabinet is pretty much a stone-throw away from the house) [14:14] Mirv, seb128 called for testers, and i'm just about to be done to have the pandaboard GLES driver ready for quantal ... [14:14] so i would have tested [14:14] chrisccoulson: but it's wires? is it raining directly on the exposed wire? [14:14] ogra_: ah, ok, so no testing yet. if it compiles for me, I'll update the packaging so that also staging & friends will work again on ARM [14:14] awesome, thanks [14:14] seb128: still looking at pango for the lucid->precise upgrade [14:15] cyphermox, ok [14:15] seb128: I think I found the issue, but I'm not sure how to fix it [14:15] what is it? [14:15] cyphermox, i'm not sure where the rain gets in yet, although the cable that goes from my house across the road to the telegraph pole is probably a good bet :) [14:16] seb128: the libpango1.0-0.modules file moves from a non-multiarch dir to a multiarch dir [14:16] chrisccoulson: wow :( [14:17] seb128: problem is the actual execution; I'll do one more upgrade test to look at what happens to those files and figure out if it could just be a symlink added from /usr/lib/pango to /usr/lib/$arch/pango maybe [14:20] bonne nuit! [14:20] pitti, bonne après-midi [14:20] (or copying the new one in /etc/pango/pango.modules maybe) what I don't understand is why nobody noticed or filed a similar bug for another release upgrade between lucid and quantal [14:20] pitti, c'est pas encore la nuit [14:20] pitti: bonne nuit! [14:20] pitti, ou "bonne soirée" [14:21] cyphermox, is evolution the only app having that issue? didn't the hook you did to restart the service work? [14:21] nah [14:21] *any* application started after libpango1.0-0 would be affected [14:22] but it also needs to be something that wasn't upgraded yet, or it needs to happen between the time that pango is updated and something else [14:22] (or maybe a new dialog from a running app?) [14:23] ok [14:32] hmmmm, i'm gonna have to get bisecting === mterry_ is now known as mterry [15:06] pitti: I addressed the aptdaemon points you raised, would be great if you could have a look at the MP again at some point :) [15:30] kenvandine, chrisccoulson, Ursinha, Laney, mlankhorst, cyphermox, mterry, Sweetshark, tkamppeter, robru: it's meeting time if anyone has a topic (none on the wiki so far), also please update https://wiki.ubuntu.com/DesktopTeam/Meeting/2012-08-28 with things you worked on this week for those who didn't write anything yet [15:43] That was a quick meeting! :) [15:45] mterry, I was wondering if I was still online, or if anyone was working today [15:45] usually I get at least a bunch of "hey, no topic from me" replies [15:45] hey, no topic from me [15:45] seb128, work? nah [15:46] mterry, ;-) [15:46] seb128, barry and I are working hard on this port, there's a lot done but also a lot left to do. [15:46] robru, yeah, what is this work thing everybody keeps talking about, right? ;-) [15:46] ;-) [15:48] seb128: my brother works in climate simulation (in cars, planes etc.). One day his boss called him, he recognized the number and answered the phone with: "$COMPANYNAME, we simulate work!' [15:48] lol [15:48] le boss was not amused ... and my brother made the perfect trollface. [15:58] ricotz: you're working on the gdm changelog? [15:58] seb128: are we sticking with glew 1.8 for quantal? [16:00] micahg, I would think so, would you prefer to go for 1.9? [16:01] Sweetsha1k, lol [16:01] seb128: no, no preference, but I won't bother with 1.5, 1.6, and 1.7 rebuilds if we're not sticking with 1.8 :) [16:01] we do plan to stick with 1.8 [16:01] ok, sounds good, thanks [16:02] ricotz: or I could just diff fta's changelog & credit you with what's changed [16:03] jbicha, i updated the changelog (just drop the file which sneaked in) [16:03] jbicha, but give is some testing before you consider pushing it [16:03] *it [16:05] ricotz: yeah I've been testing it, except for the one major issue with gdm-fallback-mode I emailed you about, things are working well [16:06] it also took me a bit of effort to even trigger fallback mode, but I'm going to try reporting that bug to GNOME [16:06] jbicha, you emailed me before i uploaded it, and your patch wasnt right if you used it [16:08] I've been using your 0827 ricotz0 build this morning [16:09] jbicha, ok [16:09] didnt see that issue [16:13] jbicha, did you forgot to push your g-c-c update? [16:13] jbicha, it's UNRELEASED in the vcs [16:14] quantal is unreleased too :P [16:14] seb128: done, GNOME wasn't interested in my patch by the way [16:15] jbicha, why not? too technical? [16:15] they don't want terminal getting special treatment [16:15] ok [16:15] well I'm unsure about the ui bit for it [16:16] it might be enough to have it in g-s-d [16:30] seb128: hmmm, the new libreoffice package is just as bad -- if not worse -- wrt unitymenus. [16:31] Sweetsha1k, "new", you got an update? what changed in good or bad? [16:34] lololololol @ https://twitter.com/paul_irish/status/240479498112475136 [16:35] seb128: well, afernandez asked be for an updated build. However, I see lots of menus having 'EMPTY STRING' as label now, and bug 1041354 isnt solved at all. [16:35] Launchpad bug 1041354 in libreoffice "unity-panel-service since yesterday uses ~100% CPU when libreoffice-gtk is installed and enabled" [High,Confirmed] https://launchpad.net/bugs/1041354 [16:36] desrt, travelling I guess? [16:49] it's typical. when i want it to rain, it stays dry [16:50] never mind, the forecast for tomorrow is rain for the whole working day [16:50] want to end your day early eh ? [16:50] i'd best make sure my 3G is working properly then :) [16:50] ogra_, no, i want to see if our line actually sounds noisier when it rains. my ISP wants to know ;) [16:50] chrisccoulson, start complaining louder to your ISP [16:51] pee on it [16:51] lol [16:51] "sorry, i won't be working today. i got arrested overnight for indecent exposure after following advice from ogra" [16:51] *g* [16:53] jbicha, you guys go the opposite direction from us, rather than being conservative you go for git snapshots? ;-) [16:53] gdm (3.5.90+git20120827.b558e179-0ubuntu1) quantal; urgency=low [16:53] " * TODO: the following Ubuntu patches need to be ported:" [16:53] not cool :-( [16:53] seb128: just be glad we're not responsible for nautilus [16:54] jbicha, lol [16:54] jbicha, I wouldn't be woried, you couldn't break it over what upstream has done :p [16:58] guest session really ought to be built into GNOME [16:58] oh, nautilus is still such a mess? I remember it from back in the days ... [16:59] guest session was mentioned this morning as a workaround for https://bugzilla.gnome.org/show_bug.cgi?id=676577 [16:59] Gnome bug 676577 in general "There is not a place for the user to set Privacy or Security related options" [Enhancement,New] [17:16] kenvandine, I was thinking of rolling a new light-themes (I committed a unity-greeter-specific change there). Any objection? === ogra_ is now known as ogra === ogra is now known as ogra_ [18:02] mterry, nope [18:02] go for it [18:24] kenvandine, done [18:35] ricotz: did you know that the gdm login screen looks different if you were still using gnome-shell 3.5.4? [18:35] it also looks like gdm won't load if gnome-shell isn't installed (probably part of why Debian force gdm-fallback-mode) [18:37] so I'm going to bump gnome-shell to a depends [18:38] that'll be a fun upgrade :), if gdm isn't removed on upgrade from lucid -> precise, when you upgrade to quantal you get gnome-shell [18:38] s/unity-2d/gnome-shell/ :) [18:39] micahg: yeah I don't think the gdm developers have tried running gdm without gnome-shell :| [18:41] getting gnome-shell after the upgrade wouldn't be the end of the world ;) [18:44] seb128, oh, the crash i see is definitely a thunderbird regression (in nightly), and not my addon. so it doesn't affect quantal (yet) ;) [18:44] chrisccoulson, \o/ [18:45] it's taking agest to bisect though because there are so many commits which touch the build system and trigger a full rebuild [18:45] which is a pain [18:45] ricotz: it looks like lightdm works with gnome-shell as long as gdm is installed, Switch Session won't work and if you lock your screen you can't unlock it... [18:45] Switch Session does nothing [18:45] jbicha, you should talk to robert_ancell about getting lightdm to support those [18:46] * micahg is happy that his latest round of Firefox/Thunderbird crashes appear to have been due to a bad stick of TAM [18:46] *RAM [18:46] it seems like the GNOME guys are settings for a "GNOME or nothing" which is a bit annoying [18:46] I guess at some point it will be impossible to get GNOME for most distros [18:49] s/GNOME/Fedora/ [18:49] yeah, there is a bit of that... [18:50] not sure if that's a good move being done for GNOME as a project and community though [18:50] they move to "let's fix the stack" to "we define the stack and everybody else can go play somewhere else" [18:50] it seems a bit overzealous, ideological, and utopian; the whole "GNOME OS" thing [18:51] things like lock screen or login manager should be a fdo interface [18:51] gah, ffs @ bug 1042894 [18:51] Launchpad bug 1042894 in thunderbird "Thunderbird hangs at startup only on laptop" [Undecided,New] https://launchpad.net/bugs/1042894 [18:51] 2 copies of eds being loaded? [18:51] they moved from ... to* [18:51] O.o? [18:52] chrisccoulson, aka the pre-upgrade version is still running and user didn't restart? [18:53] seb128, actually, looking at https://launchpadlibrarian.net/113859523/Extensions.txt, i suspect that the couchdb addon loads an old eds ABI [18:53] i should just kill that entirely [18:53] it's about time ;-) [18:55] ah, i bet that's the reason for bug 1040839 too [18:55] Launchpad bug 1040839 in evolution-data-server "Thunderbird hangs accessing eds on startup" [Undecided,New] https://launchpad.net/bugs/1040839 [18:58] yay @ the third item on http://www.mozilla.org/en-US/thunderbird/15.0/releasenotes/ ;) [18:58] \o/ [19:09] bryceh, hello [19:10] bryceh, I believe I've been hitting bug 932900, could you enlighten me with ideas on how to figure out if it's the same problem? [19:10] Launchpad bug 932900 in mesa "Unity freeze, crashes when trying to restart - intel_do_flush_locked failed: Input/output error" [Undecided,Fix released] https://launchpad.net/bugs/932900 [19:21] kenvandine: how are you deciding which account plugins should be recommends and which should be suggests? [19:21] will there be a way in the UI for users to install additional plugins? [19:22] just with software center i guess [19:22] ok, should IRC be installed by default? [19:22] seb128 and i did a quick brain storm to pick what is there [19:22] no, telepathy-idle isn't [19:23] jbicha, @IRC: no [19:23] if users want IRC they are better served getting an IRC client from s-c [19:23] im client doing IRC are poor choices [19:23] mho [19:23] imho [19:23] chrisccoulson, nice re: U1 in thunderbird [19:24] so connecting to #ubuntu requires installing extra software? [19:24] so far yes, is #ubuntu an user recommended contact point? [19:24] we should maybe figure a way to web proxy there... [19:25] it's listed at http://www.ubuntu.com/support [19:25] IRC stays a really specialized medium, look at the number of people on IRC compared to e.g forums or askubuntu [19:25] imho we should better direct user to those sites [19:26] (just my opinion [19:26] ups [19:27] * micahg would still be using pidgin for IRC if it wasn't for the freenode flood issue I could never track down [19:28] Ursinha, anything in /sys/kernel/debug/dri/0/i915_error_state attached ? [19:28] there's http://webchat.freenode.net/ also [19:29] seb128: maybe a .desktop file linking to ^^ if #ubuntu is useful [19:29] or whatever the URL with the channel in it would be [19:30] * mterry is going afk for a bit [19:30] jbicha as being a documentation team member is probably better placed that me to reply [19:30] seb128, where is the packaging branch for unity staging? [19:30] but I don't think IRC is a good end user recommendation [19:30] kenvandine, ~/staging/unity/ubuntu [19:30] kenvandine, :-p [19:31] kenvandine, there is none, what's up? [19:31] i am having to patch unity, and want to base it off staging for now [19:31] since it has all the previews stuff and all [19:34] bryceh, there's nothing in /sys/kernel/debug [19:34] (sorry, was answering the door) [19:35] Ursinha, are you able to reproduce the bug deliberately? [19:35] bryceh, whenever I try to unity --replace, it gives me that error message [19:36] I'm right now using metacity --replace to be able to talk to you here [19:36] ok good [19:37] Ursinha, ssh into your system from another one and then do unity --replace to lock the system up. Then, from ssh collect /sys/kernel/debug/dri/0/i915_error_state and post to the bug [19:38] if you have 'IPEHR: 0x7a000002' in that file, then you probably have the same bug [19:39] it doesn't exactly lock, it becomes impossible to move windows or use alt-tab [19:39] Ursinha, ok interesting. Still, reproduce that failure case and we can verify whether or not it's a gpu lockup [19:39] ok, a moment [19:42] bryceh, here's the output of unity --replace: http://paste.ubuntu.com/1172430/ [19:43] there you go, the debug file is there [19:43] let me look for the specific thing you said [19:44] bryceh, no, I see only 'IPEHR: 0x00000000' in i915_error_state [19:45] do you want me to attach it to that same bug, or should I open another? [19:46] Ursinha, there's several IPEHR's in the debug file; try grep IPEHR /sys/kernel/debug/dri/0/i915_error_state [19:47] if that still results in just zeros, then yeah let's get a new bug filed about that. [19:47] I did that, it returned three, all of them like this one [19:47] hm, ok [19:49] bryceh, ok, so how do I file the perfect bug? :) [19:58] bryceh, against which package should I file this bug? [20:02] Ursinha, for now file it against xorg as usual. ubuntu-bug xorg. Include your unity --replace output, the i915_error_state file, and dmesg [20:03] I see some stacktraces on dmesg [20:03] will do, just a moment [20:03] ho ho! that's a good sign [20:03] :) [20:04] Ursinha, fwiw I'm not familiar with the "intel_do_flush_locked failed" error message but it sort of sounds like a generic "I couldn't talk to the gpu" message, in which case that may be misleading you to think you have bug #932900 when it could be something different entirely [20:04] Launchpad bug 932900 in mesa "Unity freeze, crashes when trying to restart - intel_do_flush_locked failed: Input/output error" [Undecided,Confirmed] https://launchpad.net/bugs/932900 [20:05] right [20:05] do you know what is that IPEHR value? [20:07] IPEHR: 0x7******* errors typically indicate a fault in the renderer (mesa) [20:08] so Giovanni's bug looks to be mesa related. Yours is sounding more like the kernel, if you're getting traces in dmesg [20:09] right, this is cool [20:11] Ursinha, oh I forgot to ask, what video card you are on? [20:15] bryceh, it's an intel generic one, I have a lenovo x220 [20:16] * Ursinha is looking for specifics [20:19] Ursinha, ok intel generic's sufficient to know [20:19] Ursinha: is this after you resume from suspend? [20:20] tjaalton, nope, I was moving windows from one side to another and it happened [20:20] I meant, the first crash, the unity --replace issue I can reproduce consistently at this moment [20:20] ok, bug 966744 was the one I was after [20:20] Launchpad bug 966744 in xserver-xorg-video-intel "[i965] Resume from suspend leaves me with black screen or a screen of the desktop before it suspended. Compiz hung in intel_update_renderbuffers() from intel_prepare_render() from brw_draw_prims()" [Critical,Confirmed] https://launchpad.net/bugs/966744 [20:21] nevermind then :) [20:21] :) [20:24] bryceh, maybe it's worth mentioning the first crash happened when I was using an external monitor in a dock? [20:24] I'm undocked now [20:36] hmm. [20:36] bryceh, there are two folders in /sys/kernel/debug/dri/: 0 and 64 [20:37] I just noticed the i915_error_state in the 0 one is empty [20:37] Ursinha, 0 is the one to use [20:38] the 64 one has the IPEHR info, the other one is empty... do you know why is that? [20:38] Ursinha, docking is worth mentioning, although hard to say at this point if it's relevant [20:39] I'm not sure why there is a 64 dir. [20:41] these crashes seem rather old. none of these files were touched today, but Aug. 21 -- I've been having these for a while, was lucky to find that bug on a google search [20:44] Rhythmbox seems broken [20:44] due to U1 plugin... [20:51] bryceh, bug 1042947 [20:51] thx [20:51] thanks for the pointers, bryceh [20:53] sure [20:56] Ursinha, Uname: Linux 3.4.0-030400rc4-generic - do you have a non-standard kernel installed by chance? [20:57] bryceh, I installed this a while ago because of another bug that I don't recall right now, it was asked to test with a newer kernel package than the default... will have to dig a bit [21:01] Ursinha, hmm [575708.158540] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung [21:01] [575708.158545] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state [21:01] Ursinha, so seems you *are* having a gpu lockup but guess it's resetting the gpu and continuing. [21:02] the i915_drv.c:398 gen6_gt_check_fifodbg warnings seem pertinent [21:02] bryceh, hm. why that error file is empty I don't know [21:02] bryceh, is there anything I could do to help debugging the issue? [21:03] installing a package, a procedure or something [21:07] Ursinha, when the gpu is reset, that file is zero'd out [21:07] oh, right [21:07] the kernel will try to reset the gpu when it notices it gets locked, which appears to be what's happening here [21:08] Ursinha, there is a gpu lockup detector udev rule which is turned off in precise that you could try turning back on [21:08] sure [21:09] SUBSYSTEM=="drm", ACTION=="change", ENV{RESET}=="1", RUN+="/usr/share/apport/apport-gpu-error-intel.p [21:09] y" [21:09] that's the rule that's needed [21:10] in xdiagnose-2.5.2/debian/xdiagnose.udev it is commented out [21:10] grab that package, uncomment the line, rebuild, install, (reboot?) and that should theoretically enable the gpu error collector [21:12] all right, will do that now [21:16] Ursinha, since you're able to repro this fairly easily, another thing I'd like you to try is boot to the current stock precise kernel and repro it with that. [21:16] right [21:16] if this ends up just being a flaw in one of the interim mainline kernels, might not be worth your time to do a lot of debugging on it [21:16] makes sense [21:25] okay, the rule is there, I'll let you know if anything happens bryceh [21:25] thanks again! [21:25] great, good luck [21:25] thanks === jalcine is now known as jacky [21:54] kenvandine, if I have a machine with the webapps PPA, and I distupgrade it to Quantal, I should be fine right? don't need to do anything special first? [21:56] rickspencer3, not sure about quantal, ken said he had unity segfaulting when updating to unity staging the other dya [21:56] day [21:56] rickspencer3, the webapp abi changed slightly between their ppa and the official version that landed apparently [21:57] seb128,well, a dist-upgrade should remove the PPA version and install the one in archives, right? [21:57] rickspencer3, if the archive version are > ppa [21:57] hmmmm [21:57] which usually they are not [21:57] otherwise you couldn't install the ppa packages [21:57] well, at some point the archive will pass the PPA, and I'll be good [21:58] yeah, "some point" ;-) [21:58] well, I'm on Precise on this machine [21:58] I am going to move it to Quantal now [21:58] not sure what's their versioning in thar regard [21:58] I should keep one on Precise, I suppose, but I can't help myself [21:58] but I would recommend to ppa-purge first [21:58] ok [21:58] will do [21:59] keeping a machine on precise is good btw ;-) [21:59] I upgraded recently but I miss my precise :p [21:59] hmmm [21:59] maybe I should just keep this one Precise [21:59] * rickspencer3 ponders [22:52] morning everyone. bryceh TheMuso RAOF robert_ancell meeting reminder. https://wiki.ubuntu.com/DesktopTeam/Meeting/2012-08-28 please update with your items and add any agenda items you want to talk about. [22:52] \o/ [22:54] bryceh, you are way too happy about meetings [22:54] robert_ancell, heh [22:54] robert_ancell, at least it's a nice break from looking at X crashes and GPU lockups [22:56] bryceh, :)