=== \vish is now known as vish [07:34] Good morning! [07:38] good morning [07:45] heloi [07:45] hello [07:53] good morning pitti [07:53] hey al-maisan and baptistemm :) [07:55] bonjour didrocks, how are you? [07:55] hey al-maisan! want to become a desktop hacker now? :-) [07:56] pitti: very fine, it's update day, again \o/ [07:56] pitti: you? :) [07:57] I read ubuntu won't provide version 2.30 for evolution, right? what is the reason for that? [07:59] pitti: why not :) [08:28] good morning desktopers [08:32] salut seb128 :) [08:33] lut didrocks [08:36] seb128: you didn't go to bed too late yesterday? :) [08:37] I'm having a look to the gtk update [08:37] didrocks, no that's ok, 0:30 or something [08:37] hey seb128 [08:37] didrocks, I just did the glib update which too a bit since there is new apis and a new gio query tool to run [08:37] hey pitti [08:37] ok, I'm finishing the file-roller one. There is an issue with tomboy in configure failing, didn't find it yet. [08:38] seb128: yes, I've read the changelog. The new gio query tool can be useful for any debugging bit? [08:40] didrocks, no, it's to make a cache, it's not an user tool [08:40] I need to ask to alex how it works, for now the cache is empty [08:40] seb128: hum, ok [08:45] meh, firefox just dies and does nothing any more with my profile [08:48] seb128: gnome-doc-utils hasn't been updated for a while, is it intended? [08:48] didrocks, no [08:48] ok, doing it so [08:48] thanks [08:50] hrm, is LP a bit on the slow side these days? [08:51] mvo, I've not tried today yet [08:51] mvo, I had to switch to non-edge yesterday because edge was timeouting on buglists [08:51] mvo, which one do you use? [08:51] edge, slow and timeouts :/ [08:51] I try the non-edge now [08:52] timeout issues on edge is a known issue [08:52] ok, thanks [08:56] * mvo celebrates commit r50 to the gnome-terminal packaging branch [08:57] good morning everyone [08:57] hey hey chrisccoulson [08:57] hey seb128, how are you? [08:57] hello chrisccoulson [08:57] good! [08:57] you? [08:58] hate g-s-d [08:58] the thing keep crashing every morning [08:59] and the "don't use the pc speaker" doesn't work after restarting it... [09:00] to give some context I suspend the laptop undocked [09:00] and tend to wake it up docked [09:00] and it crashes when trying to active the external screne [09:00] screen [09:00] seb128 - yeah, i'm good thanks [09:00] you're having problems with g-s-d? [09:01] yes [09:01] what I just described [09:01] it crashes on me every morning [09:01] for one thing wake the laptop docked with lid closed [09:01] and you get no screen [09:01] it should activate the monitor [09:01] ah, that's a pain. perhaps i could have a look at that [09:01] I've to open the lid [09:02] and use the magic xrandr key [09:02] then g-s-d crashes [09:02] i've got a laptop and docking station on the way now, so i should be able to reproduce all these issues that people have [09:02] then I start gnome-appearance-capplet which restart it [09:02] and from then I get this loud and annoying speaker [09:02] which forces me to restart my session [09:02] chrisccoulson, oh, you are a fast decider :-) [09:03] the crash after hitting the xrandr key after the display configuration changes seems similar to what slangasek has, which seems to be triggered by a xorg issue [09:03] chrisccoulson, I took me ages to pick a laptop when I bought mine [09:03] which one did you take? [09:04] pitti, did you already do a une chart today? [09:04] i got a latitude e5500 in the end. it looks a bit bulky, but then, i need something i can work on every day. and it has all intel hardware [09:04] i also got a docking station and 20" monitor too [09:04] seb128: I did, just not copied to my server; you need one? [09:04] pitti, I'm just curious to see the difference with yesterday's one [09:04] I did a clean reinstall with today's daily ISO this morning [09:04] I cleaned my cache yesterday [09:04] hey pitti - i updated my packages in http://people.ubuntu.com/~chrisccoulson/desktop-startup-speed/ last night, and rebased them on the current ubuntu versions [09:05] feel free to test them at some point :) [09:05] pitti, ok, so it will be slower than yesterday's one you did right? [09:05] pitti, ie you had your wncksync caching in use and other changes [09:06] yes, it has the normal wncksync again [09:06] the other changes should all be in [09:06] pitti, how much does xdg-user-dirs-gtk-update takes? [09:06] seb128: uploaded (daniel-lucid-20100126-1.png) [09:06] danke [09:07] seb128: it literally disappeared, AFAICS [09:07] who is doing tracker these days? I just ran across bug #474324 and wonder if its ok to drop the /usr/lib/deskbar-applet/deskbar-applet dir from the package [09:07] pitti, ok, good, so that worked ;-) [09:07] Launchpad bug 474324 in tracker "package libdeskbar-tracker 0.6.95-1ubuntu3 failed to install/upgrade: trying to overwrite '/usr/lib/deskbar-applet/deskbar-applet', which is also in package deskbar-applet 0:2.26.1-0ubuntu1" [High,Confirmed] https://launchpad.net/bugs/474324 [09:07] * pitti hugs seb128 [09:07] * seb128 hugs pitti [09:07] hey mvo - that's probably me (sort of) [09:07] mvo, you? [09:07] mvo, it's all yours [09:08] seb128: my wncksync merge was approved, but nobody did it yet [09:08] mvo, official reply otherwise is "nobody" or "chrisccoulson when he has time but we prefer him to focus on default desktop things" [09:08] ;-) [09:08] pitti, ok [09:08] * pitti tries to rescue his firefox profile [09:09] mvo - i have some people interested in helping out with tracker packaging here: https://launchpad.net/~tracker-team [09:09] i don't know whether any of them would be interested to look at that [09:10] they're mainly focused on 0.7.x at the moment [09:10] chrisccoulson: its seems like its a very simple thing, wrong path or something [09:10] mvo - i can look at that, it shouldn't take me long [09:10] feel free to assign it to me [09:10] seb128: does "nobody" mean that I should stop caring too? [09:10] mvo - yes [09:10] :P [09:10] mvo, no, it means you are welcome to step and take over it ;-) [09:11] mvo, it's just a "too much to do and not being a priority for us" issue [09:12] mvo, joke aside we are just too busy to look at it but that bug seems something easy to fix, I can have a look [09:13] chrisccoulson: thanks, I give it 5min, I think I have a idea what is wrong [09:14] seb128: its ok, I know everybody is busy [09:14] mvo - don't worry about it too much. i'm just in #tracker speaking to a potential volunteer ;) [09:14] chrisccoulson: oh, nice [09:14] mvo, how come directories create conflicts? [09:14] its a dir in one package and a file in the other [09:14] oh [09:14] yeah [09:14] I think its just a bug in the .install file somwhere [09:15] or upstream which didn't expect libexecdir to be there [09:15] it should not be a double deskbar-applet dir, then it should work [09:16] didrocks, looking to the g-s-d update [09:16] ah, never mind, my potential volunteer prefers to stay focused on the 0.7 version at the moment [09:17] session restart brb [09:17] mvo - so, i can take a look at that when i get a spare 5 minutes [09:17] taking yelp update [09:18] seb128: taking yelp update [09:18] good luck ;-) [09:19] I think we synced the debian version using the webkit codebase [09:19] not sure how that would apply to current [09:19] don't bother too much with it if it's not trivial [09:20] seb128: it's not :-) [09:21] ok, delaying this one [09:21] seb128: btw, my gnome-doc-utils upload was rejected [09:21] (permission) [09:21] I start hating those team uploads [09:21] I can sponsor it for you [09:21] chrisccoulson: nice, thanks. I added the info I gathered into the bugreport [09:22] seb128: ~ubuntu-desktop/gnome-doc-utils/ubuntu [09:22] mvo - thanks [09:22] thanks :) [09:22] you guys should apply for main upload [09:22] it makes http://people.ubuntu.com/~mvo/automatic-upgrade-testing/current/ look unhappy, this is why I care [09:22] mvo, how come, we should have neither deskbar nor tracker installed [09:22] especially not on server... [09:22] seb128: it comes from a hardy->lucid one [09:23] I don't think we had tracker by default in hardy, did we? [09:23] seb128: the upgrade tester tests a bunch of different setups, server is doing fine, desktop not currently, I fixed a bug in g-t [09:23] seb128: I don't know, I use whatever apt-get install ubuntu-desktop gives me [09:24] that sucks [09:24] seb128: even then, its still a bug and would show up in the "main-all" profile otherwise [09:24] hu? that we have bugs? [09:24] it means people upgrading will get that buggy ressource sink on karmic? [09:24] or that we have tools to discover them ,) [09:24] can't we fix that? [09:24] taking gnome-games, testing will be a little relaxing :) [09:24] no, to have tracker on upgrade [09:24] manybe we should get it out of main if it sucks? [09:24] it's eating ios and power for no use [09:24] I mean, removing it is trivial [09:24] +1 [09:25] heh :) [09:26] I check why its there (well, I added to check on my todo list) [09:40] re [09:40] got disconnected [09:53] seb128_, didrocks: am I ok to touch gnome-control-center? [09:53] pitti: sure [09:53] * pitti has an one-second boot speed present [09:55] pitti, if you touch it you do the version update too :-p [09:56] pitti, urg? gnome-control-center is supposed to not be running at all at login [09:56] it's only the capplets nowadays [09:56] seb128_: no, but adding an AutostartCondition to /etc/xdg/autostart/gnome-at-session.desktop works wonders :) [09:56] it avoids calling gconftool and a shell script which just checks a gconf key and exits [09:56] pitti - what phase does that start in? [09:57] chrisccoulson: normal app phase [09:57] pitti - thats ok then [09:57] who cares about accessiblity after all [09:57] ;-) [09:57] http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100126-1.png [09:57] pitti - did you manage to test the packages in http://people.ubuntu.com/~chrisccoulson/desktop-startup-speed/ ? [09:57] it's the gnome-at-visual/gconftool-2 pair [09:58] chrisccoulson: not yet, sorry [09:58] pitti, it's weird [09:58] took me a bit to repair firefox, and I'm still fighting with glib [09:58] pitti - no worries [09:58] seb128_: why? [09:58] pitti - you're testing the UNE sessionn aren't you? [09:58] pitti, I would have though that AutostartCondition would be as expensive [09:58] if that's about reading a gconf key too [09:58] seb128_: right, but this executes a shell script plus gconftool-2, as in the program [09:58] seb128_ - with my changes, gconf should be ready by the time the autostart condition is read [09:59] seb128_: I expect that gnome-session just uses a library call, and has gconf connection already [09:59] and other desktop files have an AutostartCondition already, so one more check should be trivial [09:59] part of my gnome-session patch delays reading the value for the autostart condition until necessary [10:00] pitti, I'm just surprised it costs that much but good ;-) [10:00] I'm checking there [10:00] last time I tried to drop all the autostart it was winning less than 1 second there [10:00] the autostart*s* [10:00] ie, /etc/xdg/autostart [10:02] http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100126-1-gnome-at-autostartcond.png [10:02] oh, I see [10:02] it moved the "I'm done" bar much more to the front [10:02] pitti, commenting it in the session capplet makes no win there [10:02] it was 0.2 seconds slower in fact [10:03] but that's random difference between boots [10:03] weird; saving a shell script and yet another external program (gconf) should be expensive [10:03] s/saving/calling/ [10:03] hm [10:03] I do another one, to compare [10:03] in the 0.1s magnitude expensive I would say [10:04] we have lot of those ran [10:04] as chrisccoulson said yesterday [10:04] all those .desktop with a sleep [10:04] seb128_: so you wouldn't like to see the change? [10:05] pitti, I do, I was just surprised that it won so much to you because I tried to disable the autostarts there and it was not making such a difference [10:05] still seem a good idea [10:05] seb128_: as I said, it wasn't actually winning a second; it just lightened the load a bit, and bootchart considered the startup done earlier [10:05] and should be a bit less expensive even if it's not a one second win [10:05] right [10:05] * seb128_ hugs pitti [10:05] ok [10:06] seb128_: I'll do the new version, too, then [10:06] next on my list is to identify this python process [10:06] pitti, don't bother [10:06] the only two python things that we start are the printer applet and jockey [10:06] but both have a sleep before [10:06] seb128_: bother> about the python one? [10:06] you know what it is? [10:06] pitti, the sleep one is the scriptto add indicator applet [10:07] add-indicator-applet.py [10:07] pitti - do we need to delay starting those with a sleep? [10:07] seb128_: right, but that also has sleep [10:07] if so, i can implement a gnome-session to add a autostart delay key to the desktop files [10:07] seb128_: (but right, that's the third python one) [10:07] to avoid spawning a shell [10:07] chrisccoulson, would be nice [10:07] chrisccoulson: that would be great indeed [10:07] cool, i'll work on that then [10:11] seb128_: right, but on our charts there's a python process which starts right away; so you don't know what that is either? [10:12] pitti, no [10:15] * didrocks marks this day as being the easiest update of gnome-games ever done (nothing to touch to the packaging \o/). Just had to trigger the build twice because of a xml2po call failing [10:15] hehe [10:15] chrisccoulson, the gnome-session capplet still fails to store change when closed to quickly [10:15] I keep getting the issue [10:16] seb128_ - did we ever backport the change from git to fix that? [10:16] dunno [10:16] i don't think we did, and there hasn't been a release this cycle [10:16] * seb128_ looks [10:17] didrocks: Is there a Vcs-Bzr for wncksync? [10:17] chrisccoulson, you are right [10:17] thanks [10:17] wncksync is in universe when it was NEWed? I was thinking it was promoted to main in the first shot [10:17] lool: it's directly lp:ubuntu/wncksync [10:18] strange ... since a few days my irssi refuses to autojoin #ubuntu-desktop #ubuntu-motu and #ubuntu-bugs ... every other channel seems to be fine :/ [10:19] you need to be registered I think [10:19] seb128_ - http://git.gnome.org/browse/gnome-session/commit/?id=dd9df169196441717498ac09ee8c437fb552116e [10:19] they got some spambot issues or something [10:19] when did tha tchange? [10:19] asac - at the weekend i think [10:19] oh right #ubuntu-devel has the same problem somewhat [10:19] i am registered [10:19] guess its a race [10:19] e.g. first join then register or something [10:21] who is leading this anti spambot effort? [10:22] * asac finds this kind of a drastic measure [10:22] if permanently [10:25] asac: Perhaps they lowered the number of channels one is allowed to join? [10:25] I know I've hit that limit a bunch of time [10:25] https://help.ubuntu.com/community/AspireOne/Ubuntu9.10 <-- i'm thinking it might be a good idea to include the pm-utils sleep hook script on that page in our pm-utils package. [10:25] I have to identify before /join-ing [10:25] does anyone have an opinion? [10:25] pitti, python seems to come from the print applet, dunno why though [10:25] seb128_: weird; it also has a sleep [10:25] lool: hmm maybe [10:25] anyway, I'll track it down [10:26] seb128_: I finished wrestling with glib, so this has my full attention again :) [10:26] didrocks: In wncksync changelog, - change libglib2.0-dev version dependency to get glib fix [10:26] didrocks: Could you expand on that? [10:26] didrocks: Which fix is that? [10:26] lool, 71_gio... [10:26] lool, it's a patch which adds a feature wncksync needs [10:26] Ok; thanks [10:27] pitti, did you manage to improve glib? === seb128_ is now known as seb128 [10:28] seb128: Is there a Vcs-Bzr for glib2.0 or do you use lp:ubuntu/glib2.0 as well? [10:28] lool, no [10:29] You don't use any? [10:29] lool, we are in sync with debian most of the time [10:29] so I didn't bother setting a vcs to have extra work when doing syncs [10:29] I guess we could use lp:ubuntu/glib2.0 [10:29] seb128: the assert msg patch caused some problems, fixed now [10:30] seb128: Ok [10:30] pitti, the upstream version? [10:30] seb128: right [10:30] seb128: Ah glib2.0 isn't imported [10:30] pitti, ok, sorry about that [10:31] pitti, how did you notice? [10:31] pitti, just curious to improve my glib testing before upload ;-) [10:31] seb128: mclasen pinged me about https://bugzilla.gnome.org/show_bug.cgi?id=594872 [10:31] Gnome bug 594872 in general "Support storing assertion messages into core dump" [Enhancement,New] [10:31] oh ok [10:31] seb128: the private glibc symbol caused some linking problems on Fedora apparently [10:31] seb128: and "make distcheck" was failing (due to the out-of-tree build) [10:32] seb128: don't worry, they didn't come and say "OMGyoubrokeithowcanyoubesobad", they just asked me whether I can look into it :) [10:32] ok, good [10:33] hm, g-c-c 2.29.6 has a NEWS file from 2.29.4 [10:33] don't they get updated usually? [10:35] ah, http://git.gnome.org/browse/gnome-control-center/log/ is small enough [10:37] seb128: Is it ok to push glib2.0 (just a trivial packaging change) [10:37] Or do you have lots of GNOME stuff in the pipe which might FTBFS? [10:38] lool, talk to pitti, he was just working on it [10:38] pitti: ^ [10:38] lool, we will deal with build retry for GNOME updates [10:38] lool: I have nothing urgent for glib2.0 [10:39] It's not urgent here either, just doing it now if I can forget about it ;) [10:39] what I just talked about was purely an upstream issue, no hurry to get that in [10:39] lool: do it :) [10:39] lintian is on crack today [10:39] W: glib2.0 source: patch-system-but-direct-changes-in-diff libglib2.0-0.install and 1 more [10:40] lool, did you upload already? [10:40] No [10:41] lool, would be nice to rebase on debian if you want to do it ;-) [10:41] lool, slomo did the update and did the gio-query call as a trigger [10:41] the only change we have should be the gio change for wncksync [10:43] Ok [10:43] thanks [10:54] Guess what -- lintian was not on crack! [10:55] glib2.0-2.23.2/libglib2.0-0.install [10:55] glib2.0-2.23.2/libglib2.0-dev.install [10:55] In the patch [10:55] urg [10:58] didrocks, you can do the gvfs update btw [10:58] the gio lazy init might win some speed [11:00] seb128: ah, I know now; it seems that bootchart retroactively labels the shell process (which first ran sleep) "python", since that execs [11:01] so it's not really running python [11:01] ok [11:01] same issue than the ssh-agent one [11:02] to which locations can i put .desktop files to make them appear in the gnome menu? [11:02] (user locations) [11:04] seb128 - i saw you just update g-s-d. i'll have a g-s-d change for you later, to fix the volume notification overshoot which doesn't work [11:04] (when it stops crashing with my change in anyway) [11:05] asac, .local/share/applications [11:05] chrisccoulson, ok [11:07] njpatel: good morning [11:07] njpatel: my wncksync branch was approved; do I need anything to merge it? (I can't do it myself) [11:08] seb128: do you know if its possible to overwrite a system .desktop file in that way? e.g. by using the same name or something? [11:08] pitti, will get dbo to merge and make a release, we have another thing that needs to be fixed too, so we can roll your work and that together in a release [11:09] njpatel: thanks [11:09] asac, yes it is [11:09] njpatel: other question, did you ever test when mutter actually starts loading the plugins? [11:09] i basically want to ship a generic .desktop file (like mail) and then based on the user decision change the name/icon slightly [11:09] e.g. evo tbird, gmail etc. [11:09] njpatel: IOW, how much of the 7.5 second mutter startup is due to the plugin, and how much due to mutter init? [11:09] seb128: cool so if i put the same file name there it will replace the system one? [11:10] asac, I think it should [11:10] * asac tries [11:10] njpatel: (sorry, it's more like 6.2 s; mislooked) [11:10] seb128: ok, putting gvfs on my queue :) (fixing some une related stuff now) [11:10] ok thanks [11:11] yay [11:11] that really works [11:11] !! [11:11] ;-) [11:12] why is everything so easy ... boring ;) [11:12] asac, speaking of desktop files btw [11:12] asac, bug #512692 known issue? [11:12] Launchpad bug 512692 in gnome-panel "Firefox icon missing from panel after 3.5->3.6 upgrade" [Undecided,New] https://launchpad.net/bugs/512692 [11:12] asac, did you rename or drop the desktop entry? [11:13] pitti, I don't think we know exactly when it starts loading plugins [11:13] seb128: should be still firefox.desktop ... i think i even tried that [11:13] hmm [11:13] so the icon name might have changed [11:13] guess if someone sets up a launcher manually that would break [11:13] njpatel: ok; I think I'll test that then [11:14] is our default launcher setup manually? [11:14] pitti, cool, that would be useful info [11:14] njpatel: btw, looking at the large gaps in mutter, can it be that the plugin does parts of the initialization in an idle loop? [11:15] njpatel: I suppose it doesn't have lots of sleeps in it, or sth like that :) [11:15] pitti, yes, it does [11:15] (I thought I mentioned that :) [11:15] asac, no, it's part of the gnome-panel profile [11:15] seb128: what i verified was to add a launcher with right click in the menu for 3.5 and then upgraded ...killall gnome-panel etc. [11:15] njpatel: ah; that explains why it's so scattered [11:15] njpatel: right, ISTR that you talked about it, but wanted to confirm [11:15] seb128: the gio-querymodules is the new glib stuff, right? [11:15] asac, + firefox.desktop [11:15] seb128: right. but does that refer to the .desktop file? [11:15] hmm [11:15] ok [11:15] thats ok [11:15] asac, I will try there and ping you back [11:15] didrocks, yes, don't bother about it [11:15] thanks ... [11:16] didrocks, new glib will have a trigger for those [11:16] pitti, we try and hand back the main-loop to mutter asap, so it can present the desktop to the user. The rest we do as in idle once that is done. [11:16] seb128: ok, thanks :) [11:16] pitti, that's what I meant that the time bootchart shows will be a bit different than the time it really 'feels' like the desktop is loaded [11:17] njpatel: that's true [11:17] I'm afraid there's a mismatch between what an user considers as "done" and what bootchart does.. [11:18] Sure, but I don't think we should be initiall blocking the WM from loading too much (we may still push more things out of the _init). Just because we're uncessarily blocking (the user doesn't need the places straight away) [11:23] didrocks: Could someone from DX subscribe to the wncksync bugs in launchpad? or at least the ubuntu desktop bug contact? [11:24] lool: I can suscribe, it makes sense, or DBO, maybe <- njpatel ? [11:25] didrocks, canonical-dx-team for now, I think [11:25] I just would like someone to tend to it for the MIR [11:25] (I mean I'd like to be able to write that someone will care for it) [11:25] dbo is part-time, so not always around if something critical pops up [11:25]