=== rickspencer3_ is now known as rickspencer3 === Amaranth__ is now known as Amaranth [06:29] morning [07:12] eh, ppa builders have seven days long queue [08:04] davidbarth: morning :) [08:05] greetings everybody [08:05] davidbarth: I have some more patches + something that might fix the stacking bug and a fix for bug#t682781 [08:06] bug 682781 [08:06] Launchpad bug 682781 in Unity "empathy buggy list vanish when opened twice using the indicator" [High,Triaged] https://launchpad.net/bugs/682781 [08:07] MacSlow: guten morgen [08:07] hey smspillaz [08:07] :) [08:07] typing german makes me really want to go back to germany, I was there last year [08:08] didrocks: morning [08:08] smspillaz, wait with that until it's warm again here :) [08:08] hey smspillaz, good morning! [08:08] hey didrocks [08:08] Guten Morgne MacSlow [08:08] didrocks: I think I have a fix for this gconf issue (can't reproduce it but it is "theoretical") I will mail you a patch [08:09] it err, *cough* breaks the ABI though [08:09] smspillaz: oh nice [08:09] oh "nice" [08:09] ;) [08:09] smspillaz: just wait a second [08:09] sure :p [08:09] smspillaz: so, you know what happens? I now see the unity interface before the hang [08:09] MacSlow: do you still have snow? [08:09] smspillaz: if I click on a launcher just before it hangs (it happens 2s after), I can get it crash [08:09] smspillaz: interested in the stacktrace? [08:10] didrocks: well, the crash happend in Glib::Mainloop::Run - and that is created after gconf is loaded, so maybe it doesn't like that [08:10] kvalo, thank god no [08:10] didrocks: that sounds like a completely different issue [08:10] smspillaz: yeah, and that's the one I'm seeing reliably now [08:10] hrm [08:10] ISTR it crashing earlier on [08:10] give me a stacktrace for this one I guess [08:11] smspillaz: ok, let me install some debug symbols first [08:11] ok [08:11] I'd like to merge glib into upstream before the release too [08:11] so then we can just rebase on upstream for A2 [08:13] didrocks: also I have a fix for empathy vanishing too [08:13] smspillaz: excellent :) [08:19] smspillaz: ok, let me upgrade without the workaround first to ensure we are on the same basis with the right debug symbols [08:19] smspillaz: I can give it to you without them though [08:20] smspillaz: http://paste.ubuntu.com/559423/ [08:28] didrocks: #2 0x00163ce0 in Glib::Source::unreference (this=0xb6e11380) at main.cc:765 we're doing something wrong here, although it shouldn't impact startup [08:29] smspillaz: well, I clicked on a launcher as said before the interface got frozen [08:29] strange [08:29] maybe something is creating a glib main loop and then trying to quit it [08:29] and of course, there's this bug where we crash on exit [08:30] right [08:30] that I haven't really gotten around to fixing yet [08:31] yeah, so maybe something wrong happen at initialization and that doesn't work well… [08:34] didrocks: ok, I'll mail this patch to you [08:34] (haven't tried it with the unity plugin yet, but it should work) [08:35] didrocks: I have been checking all my commits with git diff now for cleanliness :) [08:35] so no more random whitespace changes [08:35] smspillaz: nice! :) [08:35] smspillaz: can you just try it with the unity plugin first? [08:36] smspillaz: because as it's an ABI break, I don't want to speak 1h30 of rebuilding for nothing :) [08:36] didrocks: this is true [08:36] thanks! ;) [08:36] didrocks: I'll mail it to you for now (so you can build in the background) and test it too [08:36] sure [08:36] because I'm pretty sure it should work [08:38] smspillaz: well, the issue I'm rather seeing now is the other hang TBH [08:38] yeah [08:38] oh brilliant [08:38] I did a merge and forgot to commit right after the merge [08:38] smspillaz: if you change the way glib mainloop it handled maybe it can impact it [08:38] so now my local changes are stuck in alongside the merge [08:39] argh [08:39] *cough* I wonder how evil it is to just have that [08:39] since I don't know how to fix it [08:40] smspillaz: well, the other changes are already in the package? [08:40] I think so [08:41] thats the problem [08:41] well, I have everything open [08:41] maybe I can just reset, then git stash, then commit the merge, then merge the stashed branch [08:42] not sure about git :) [08:49] I think I'll need to just get the diff, reset, pull from master and then do edit the diff till it applies >.> [08:50] smspillaz: good luck ;) [08:57] packaging issue? [08:57] dbarth_: nah, me forgetting to do something with git issue :) [08:57] dbarth_: what? [08:58] trying something for the hang, brb === _LibertyZero is now known as LibertyZero [09:02] smspillaz: hold on, I know from where the hang is coming! [09:02] urgh, ok, applying the whole thing manually [09:02] go on [09:02] it's only if I have unity-places-* installed [09:02] dbarth_: ^^ [09:02] omer26 rocks! [09:02] ah yes [09:03] didrocks: grep unity-places stuff for g_main_loop_quit [09:03] smspillaz: I'll, first need to thanks omer26 :) [09:03] :) [09:03] thanks omer26! [09:06] hi njpatel [09:06] (on this bug report, FYI: https://bugs.launchpad.net/ubuntu/+source/unity/+bug/683328) [09:07] hey smspillaz [09:07] smspillaz: no ABI break for the empathy patch? [09:07] didrocks, smspillaz, using latest compiz and latest defaults, excellet job guys :) [09:08] njpatel: I'll change two more settings for alpha2 btw, so you will need to reset soon to get the latest and greatest :) [09:08] heh, will do :) [09:08] compiz (core) - Error: Plugin 'core' has ABI version '20110126', expected ABI version '20110113'. [09:09] why do I always get ABI snafu whenever I update_ [09:09] kamstrup: any reason you don't use the packages? [09:09] I'm not sure you need compiz trunk :) [09:10] seems duplicating the effort of rebuilding everything and such [09:10] (furthermore, you don't have the same default than us) [09:10] didrocks: so you suggest that we just submit branches without compiling or testing them? I can do that [09:10] kamstrup: did you submit branch of compiz? [09:11] didrocks: no, and I didn't install it from source either [09:11] kamstrup: oh ok, it's unity bailing then, I was thinking it was because you rebuild compiz in jhbuild :) [09:11] kamstrup: my bad then, yeah, ABI is changing often [09:12] hah compiz jhbuild [09:12] njpatel: btw - is the compiz branch dbarth_ imported into LP the "official" one for Natty? [09:12] kamstrup: just think than everytime the ABI is changing, I have to rebuild 5 packages (and hate sam of course!) ;) [09:13] didrocks: i just reinstalled *nux* *compiz* and *unity* packages, and I still get the ABI break... I think that maybe a source install of nux or unity overrides something the packages doesn't touch? [09:14] kamstrup: yeah, do you have the unity plugin in ~/compiz-1 ? [09:14] source instal of anything => eternal doom [09:14] didrocks: lemme check [09:14] kamstrup, maybe i don't help you..but i get your same error yesterday...i've reinstalled also plugins-main and plugin-extra [09:14] and now works [09:14] kamstrup: I have the same issue there and I rm -rf /usr/local/***, that's not good :/ [09:15] kamstrup, i believe so, but we always carry distropatches so better to use what's in natty [09:15] kamstrup: ~/.compiz-1/plugins/libunityshell.so [09:15] ~/.compiz-1/metadata/unityshell.xml for the referential [09:16] then, you should have leftovers in /usr/local/share/unity/, /usr/local/lib/*unity*, /usr/local/bin/unity [09:16] (and also the locales of courses… in /usr/local/share/locale/…) [09:17] kamstrup: I'll make a script "remove/destroy my local unity" for next release, that will be easier [09:17] didrocks: indeed I appeared to have the local plugin - i need to kill someone for doing that when I specifically request another install dir... [09:18] njpatel: so who do I need to kill for lp:unity installing into my homedir even when i set CMAKE_INSTALL_PREFIX? [09:18] kamstrup: well, I know who played a lot with compiz cmake… ;-) [09:18] kamstrup: I think that's compiz to detect the plugins TBH [09:18] kamstrup, smspillaz's neck is especially nice to hold during cmake issues [09:19] all the rest is installed elsewhere [09:19] kamstrup, it's autodetected from compiz's cmake stuff [09:19] just not the metadata and the plugin [09:19] kamstrup, -DCOMPIZ_PLUGIN_INSTALL_TYPE=local is what defines installation of the plugin [09:20] didrocks: nope [09:20] * kamstrup needs to restart [09:20] smspillaz: you mean, you can install the metadata and the plugin wherever you want and compiz can detect it? [09:20] sounds "magical" :) [09:20] didrocks: either to ~/.compiz or to $COMPIZ_PREFIX [09:21] didrocks: I could make an install manifest file that goes in ~/.compiz or $COMPIZ_PREFIX (depending on the user?) so that you can install plugins anywhere you want [09:21] but that sounds like more trouble than its worth [09:21] smspillaz: or we can maybe play with symlinks for now :) [09:22] didrocks: I think in the CMakeLists.txt we should make COMPIZ_PLUGIN_INSTALL_TYPE default to package [09:22] smspillaz: hum, doesn't that set the prefix to /usr instead of /usr/local? [09:23] didrocks: it sets it to CMAKE_INSTALL_PREFIX [09:23] kamstrup: -DCOMPIZ_PLUGIN_INSTALL_TYPE=package [09:23] we'll make it the default :) [09:23] smspillaz: oh? so that can be interested, as long as it doesn't install in /usr by default, I'm all for it :) [09:24] kamstrup: does vala have pop() for string[]? [09:25] didrocks: sure [09:30] kvalo: nope, strings and arrays are immutable afaic [09:30] afaik [09:31] wow, xchat has some really messed up rendering since the last round of x/compiz/theworld updates [09:35] didrocks, njpatel, ping ;) [09:35] hey aquarius_ [09:37] didrocks, I'm having unity problems (specifically, it isn't starting :)) I don't have unity-place-applications or -files installed (I would have thought that an upgrade should pull them in, but it doesn't seem to have done). njpatel said I should come and talk to you [09:38] aquarius_: do you see the unity interface still or nothing at startup? [09:39] didrocks, no unity interface at all. [09:39] aquarius_: ok, can you pastebin ~/.xsession-errors? [09:40] didrocks, I killed compiz, metacity --replace, gnome-terminal, then unity --replace from inside that gnome-terminal... [09:40] or the output :) [09:40] and the terminal window just went solid grey and nothing else happened :) [09:40] no output. Will get xsession errors [09:40] aquarius_: ok, try to start your session, then go a tty, and paste the xsession-erros :) [09:41] didrocks, http://paste.ubuntu.com/559455/ [09:42] hum, weird, unityshell is loading, you don't have any ABI issue [09:42] aquarius_: do you have a .crash in /var/crash/ ? [09:42] $ ls /var/crash [09:42] compiz-gnome.0.crash [09:42] _usr_bin_evince-thumbnailer.1001.crash [09:42] _usr_lib_at-spi_at-spi-registryd.1001.crash [09:42] _usr_lib_couchdb_bin_couchjs.1001.crash [09:42] _usr_lib_evolution_e-addressbook-factory.1001.crash [09:43] compiz-gnome. Hm. [09:44] except that I believe that the compiz-gnome crash was on *installation*, not on running, and I fixed that with apt-get -f install. I think. [09:44] yeah, it's not that one [09:45] so… can you still [09:45] dpkg -l compiz, dpkg -l unity, dpkg -l compiz-fusion-plugins-main [09:48] smspillaz: look at the htorque's comment on https://bugs.launchpad.net/bugs/682781 [09:48] ii compiz 1:0.9.2.1+glib OpenGL window and compositing manager [09:48] ii compiz-fusion- 0.9.2.1-0ubunt Compiz Fusion plugins - main collection [09:48] ii unity 3.2.16-0ubuntu Interface for Ubuntu Desktop Edition [09:48] didrocks, ^ [09:50] aquarius_: apt-cache policy or bigger terminal for the two first one, please :) [09:51] haha! oops :) [09:51] compiz - 1:0.9.2.1+glibmainloop4-0ubuntu5, compiz-fusion-plugins-main 0.9.2.1-0ubuntu8 [09:52] all sounds fine then [09:52] aquarius_: and so, you still see compiz running at start? [09:52] aquarius_: do you have the same issue in the classic session? [09:53] didrocks, don't know if I see compiz or not. I get my desktop background and that's it [09:53] haven't tried the classic session... [09:53] aquarius_: can you try the classic session first? (that will help) [09:54] didrocks, I can once I've finished all the other conversations I'm having about work and so can log out ;) [09:54] aquarius_: ok ;) [09:56] didrocks: I'll look into it [09:59] oh, I should mail you this patch [09:59] I haven't tested it with unity yet (fixing another bug) but it should work [10:01] smspillaz: i need your divine insight for a bit... [10:01] smspillaz: can you off the bat tell which threads there are in compiz? [10:01] smspillaz: what patch? [10:02] the empathy one? [10:06] kamstrup: there is only one [10:06] smspillaz: ok glad to hear :-) [10:07] smspillaz: I'm tracking down a most vicked deadlock and the less threads to track the better :-) [10:07] kamstrup: ah yes :) [10:07] didrocks: I mailed you the glib one. it is untested against unity though [10:08] smspillaz: didn't I ask you to test it? as It will take me 1h30 to rebuild everything for the ABI break for maybe nothing? [10:08] don't have time right now :) I'm fixing another stacking bug [10:08] same here [10:08] heh [10:08] ok, it can wait then :p [10:08] I'll just mail it to you for now [10:08] yeah, as we can't reproduce it anymore [10:09] yeah [10:09] well I mean, the backtrace you gave me last night was crashing in Glib::MainLoop::Run, so that gave me a good idea of what is going on [10:09] in any case, this patch does not break anything [10:09] well, we don't know with unity [10:10] yeah :) [10:10] I will test that [10:10] and it's an ABI break again, so rebuilding everything :) [10:10] :p [10:10] not sure it worth risking this now for alpha2 [10:10] as long as we don't have bug report about people getting it [10:10] didrocks: the workaround we've got now increases startup time [10:10] that's the thing :/ [10:10] smspillaz: it's been dropped since yesterday, you were there :) [10:11] we dropped the workaround ? [10:11] there is still the other one for dconf and unity [10:11] right [10:11] O.o === MacSlow is now known as MacSlow|cops [10:12] urhg, ok, cannot figure out what's going with gnome-terminal [10:12] * smspillaz drops that issue [10:16] smspillaz: hi, about the empathy thing (bug 682781): the buddy list wasn't visible before applying the patch, so this could very well be a separate bug [10:16] Launchpad bug 682781 in Unity "empathy buggy list vanish when opened twice using the indicator" [High,Triaged] https://launchpad.net/bugs/682781 [10:17] smspillaz: it also seems not to happen after every workspace switch [10:19] htorque: it is a similar bug [10:19] I know what it is [10:19] I'm just looking into why it happens now [10:19] htorque: seems that the decoration window is being destroyed [10:20] incorrectly [10:25] htorque: will have a fix in 1 - 1.5h [10:26] (if the bug isn't complicated, which I don't think it is) [10:26] smspillaz: nice :-) [10:26] I really need to write a workspace switcher for the 2D mode, that would make my life so much easier [10:46] didrocks, I confirm that I have the same problem with the classic desktop, so that suggests it's compiz. Also, I got a compiz crash complaint from apport, and so filled out https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/709161 [10:46] aquarius: oh nice, let's wait for a retrace then :) [10:46] didrocks, what's a retrace? [10:48] didrocks, is that where one of you guys runs a thing over my core dump? Or do I need to do that? [10:49] aquarius: exactly, you need to do nothing if the report is fine, let me look [10:49] didrocks, I'm happy to provide any more information you guys might need, as long as you don't try and make me actually fix the problem since my C-fu is weak ;) [10:50] aquarius: can you subscribe me and smspillaz to the bug report please? I don't know why I can't have access [10:50] didrocks, done. I think that might be because it said "this will contain your gdm logs, which might be sensitive information" [10:51] yeah, but I should as a core-dev have access to it… weird [10:51] didrocks, what sensitive information might it contain? If it's not *really* sensitive, I'm happy to mark the bug public. [10:51] aquarius: not sure, #ubuntu-x would know, I never saw sensitive information apart from your login name :) [10:52] hum, that's a new bug [10:53] smspillaz: maybe something to take into account quite seriously ^^ [10:53] smspillaz: see the Stacktrace.txt file [10:54] didrocks, smspillaz, that file says that I don't have symbol info: would it be helpful if I installed some -dbg package or other and recreated the bug? Or do you have enough info from the bug already? [10:54] htorque: try the patch I just attached on the bug report [10:54] * smspillaz prepares for a long night [10:55] didrocks: seems that one of my recent patches has reintroduce the stacking bug :/ tracing which one it is [10:55] aquarius: depends, do you have some time right now? :) [10:55] dinner first though === smspillaz is now known as smspillaz|dinner [10:55] smspillaz: sure, enjoy [10:55] didrocks, I have a little time; having a working desktop is important, so I can devote a small amount of time to it :) [10:56] aquarius: do you still have the .crash file in /var/crash/ ? [10:56] aquarius: you should still have the classic session (No Effect) as a backup :) [10:56] (it launches metacity) [10:56] didrocks, I have a _usr_bin_compiz.1001.crash file in /var/crash, yes [10:56] ok, nice [10:58] aquarius: so, apt-get install libglibmm-2.4-1c2a-dbgsym libglib2.0-0-dbg compiz-core-dbgsym compiz-plugins-dbgsym compiz-gnome-dbgsym libstartup-notification0-dbgsym [10:59] then apport-unpack /var/crash/_usr_bin_compiz.1001.crash /tmp/compizcrash [10:59] cd /tmp/compizcrash [10:59] gdb /usr/bin/compiz CoreDump [10:59] apply all thread bt full [10:59] and paste the output :) === MacSlow|cops is now known as MacSlow [11:00] (apport will do that for you on launchpad, but we can stil win some time doing it locally) [11:01] didrocks, none of those packages exist. :( [11:01] do I need some extra repo enabled? [11:01] aquarius: oh sorry, you need the ddeb repository [11:01] one sec [11:01] aquarius: https://wiki.ubuntu.com/DebuggingProgramCrash [11:02] echo "deb http://ddebs.ubuntu.com $(lsb_release -cs) main restricted universe multiverse" | [11:02] sudo tee -a /etc/apt/sources.list.d/ddebs.list [11:02] then apt-get update… [11:08] didrocks, (gdb) apply all thread bt full [11:08] Undefined command: "apply". Try "help". [11:08] didrocks: I am hitting this hang too now [11:08] let me try and apply this patch [11:08] smspillaz|dinner: this == aquarius' one? [11:09] aquarius: hang on, it's thread apply all bt full [11:11] didrocks, http://paste.ubuntu.com/559475/ is the output I think [11:11] aquarius: excellent :) [11:11] didrocks: its the one where removing the workaround makes compiz hang on start [11:11] smspillaz|dinner: ^^ [11:12] you mean, we are heading to add a third time my ugly workaround? :) [11:12] didrocks: no, I'm going to try my solution [11:13] smspillaz|dinner: ok, then maybe we can propose it to aquarius to confirm it works (I'll prepare the packages in a ppa) [11:13] yeah [11:13] didrocks: let me see if it works here first! :) [11:13] sure [11:14] so to add a new patch to series, I need to bzr bd-do; export QUILT_PATCHES=debian/patches; quilt-push -f 100_fix_mainloop.patch; quilt refresh; exit 0 right? [11:14] hooray, I shall be running not *just* alpha packages from natty, but also not-even-released PPA packages! I am so cool. OMG I'M GOING SO FAST OH F*** [11:14] :-) [11:14] with patches from the upstreams themselves [11:14] your request is important to us :) === smspillaz|dinner is now known as smspillaz [11:14] smspillaz: nooo! [11:15] I didn't really have dinner, that was a false alarm from lurking parents :p [11:15] aquarius: well, oneconf is crashing because of desktopcouch right now, do you want to trade? :p [11:15] smspillaz: ahah :) [11:15] smspillaz: so [11:15] smspillaz: bzr bd-do; export QUILT_PATCHES=debian/patches; [11:15] didrocks, touché ;) [11:15] quilt new 100_fix_mainloop.patch ; quilt add [11:16] quilt refresh [11:16] exit 0 [11:16] aquarius: héhé ;) [11:16] didrocks: by "every files you will touch, you mean every file in the src right? [11:16] smspillaz: no, the file that will be affected by your patch [11:16] sure [11:16] * smspillaz does that [11:16] (do that before applying your patch!) [11:17] there is another solution with edit-patch as a wrapper but I don't want to trouble you with yet-another-way :) [11:23] didrocks: I'm getting dpkg-source: error: cannot read compiz-0.9.2.1+glibmainloop4/debian/patches/100_fix_mainloop.patch: after bzr bd [11:23] did I miss something? [11:24] smspillaz: bzr add? [11:24] didrocks: ah, right :P [11:25] :) [11:25] do I need to do that in bzr bd-do ? [11:26] no, after exit 0 [11:30] argh :/ even when applying this patch using quilt it still fails to apply [11:31] smspillaz: on all files? [11:31] yeah [11:31] that's because quilt is expecting -p1 [11:31] not -p0 [11:31] didrocks: right [11:31] and git diff generate -p0 [11:31] so how do I fix that? :) [11:32] I'm using git format-patch [11:32] quite easily, sed to prepend "foo" [11:32] yeah, git format-patch is doing the same [11:32] well, prepend "foo/" even :) [11:32] my bash skills are not exactly brilliant :) [11:33] so sed -e foo/ 100_fix_mainloop.patch ? [11:33] hum… no :) [11:33] can you pastebin the patch? [11:33] sure [11:34] http://paste.ubuntu.com/559484/ [11:35] smspillaz: it's already in -p1 format [11:35] smspillaz: so, it should work without any issue [11:36] hrm ok, it doesn't seem to apply [11:36] smspillaz: ok, just bzr revert [11:36] then cp debian/patches/100_fix_mainloop.patch [11:36] echo 100_fix_mainloop.patch >> debian/patches/series [11:37] bzr add [11:37] bzr bd [11:37] just to ensure it doesn't apply first :) [11:37] yeah, it doesn't apply [11:38] ok, so now [11:38] bzr bd-do [11:38] quilt push -f 100_fix_mainloop.patch [11:38] it fails [11:38] do some chunks applies or it's rejected everywhere? [11:38] it's rejected everywhere [11:38] hum, weird, let me try there [11:39] ok [11:39] smspillaz: can you pastebin to pastebin.com or somewhere I can download the file without credentials? (hard in wget…) [11:44] smspillaz: ? [11:47] sure [11:50] didrocks: http://pastebin.com/raw.php?i=0VsiM4Mk [11:50] smspillaz: excellent, thanks! :) [11:52] ahah, I needed to remove some html, but it's fine :) === bregma_ is now known as bregma [11:53] smspillaz: it doesn't fail on every file [11:53] smspillaz: just some chunks don't apply [11:53] smspillaz: so, it's as yesterday, just reapply them manually [11:53] bzr bd-do [11:53] quilt push -fa [11:54] -> reject [11:54] look at all the .rej files [11:54] reinject the changes in each file [11:54] quilt refresh [11:54] exit 0 [11:54] what do you mean by "reinject the changes in each file? " [11:54] that was at 1am for me, I don't remember much :) [11:55] smspillaz: like, look at the .rej file, see what should be added back to the file [11:56] and then? [11:56] smspillaz: it's reject because it can't apply the patch (something in the context changed) [11:56] so edit the files to add them [11:56] once all changes are back in the files [11:56] quilt refresh [11:56] exit 0 [11:56] bzr bd [11:56] since ah right! [11:56] *ah right! [11:56] :) [11:56] so quilt will automatically fix the patch [11:56] not automatically? [11:56]