=== bjf[afk] is now known as bjf === asac_ is now known as asac === smspillaz|asleep is now known as smspillaz [04:53] Good morning [04:54] chrisccoulson: indeed, hasn't crashed once yesterday; I'd say you fixed it, great work! [05:27] good morning pitti [05:28] pitti, i just added libunity launcher integration to xchat-indicator http://ubuntuone.com/p/cVu/ [05:28] :) [05:29] hey kenvandine [05:29] kenvandine: oh, cool! [05:29] i'll do empathy this week too [07:00] Morning! [07:00] hey Sweetshark [07:17] * Sweetshark starts to try his luck on libreoffice-3.3.1.1 === oubiwann_ is now known as oubiwann [07:25] pitti: Any idea when I can get a hold on doko? LO-3.3.1.1 has been released and I am currently merging in changes from debian ... [07:25] Sweetshark: may the source be with you! [07:25] Sweetshark: hm, no idea I'm afraid; ask robbiew? (foundations team lead) [07:25] he's in the US, so is probably asleep now [07:30] * Sweetshark is reconsidering thorsten wasnt joking back then with OOo on hg, LO on git, ppas on bzr and lots of diffs flying around in the build repo: http://blog.thebehrens.net/2009/04/01/how-openofficeorg-will-switch-to-a-dscm/ [07:38] good morning [07:45] bonjour didrocks [07:45] Guten Morgen pitti, how are you? [07:46] didrocks: I'm great! just woke up a little early (5:30) [07:47] urgh, why? [07:47] couldn't sleep more? [07:48] was awake enough [07:52] woah, I would need lots of tea to keep me going at that time of the night^Wmorning :) [07:53] morning all [07:58] hey dpm [07:58] mvo: cold shower did the job, too :) [07:58] heya pitti [08:02] meh, i guess the driver is called nouveau because it like a nouveau boot ... [08:06] It also likes to be given regular walks, and to have a stick thrown to it every now and then. [08:28] didrocks: just to avoid double work, I'm reviewing https://code.launchpad.net/~gunnarhj/gdm/language-menu/+merge/45681 now [08:28] didrocks: (in case you stumble over it in sponsoring) [08:29] pitti: ok, I'm more doing the sync review right now :) [08:29] pitti: in any case gdm | grep language -> pitti :) [08:33] morning [08:38] pitti, hey, do you think bug 284443 is worth fixing? [08:38] Launchpad bug 284443 in gdm "gdmsetup: Don't offer autologin for ecryptfs users" [Low,Triaged] https://launchpad.net/bugs/284443 [08:39] robert_ancell: I think so, as if you set autologin, you'll just run into a totally broken desktop session, which makes it difficult to clean up again [08:40] I thought initially that the issue was the filesystems were being decrypted without a password (i.e. a security problem), but the only issue appears to be that your ~/Private directory doesn't work [08:40] pitti, is the encryption on the whole of ~ or just ~/Private? [08:40] robert_ancell: no, it's not about ~/Private, we encrypt the entire home by default now [08:40] robert_ancell: I mean if you select "encrypted" in the installer [08:40] pitti, is that decrypted in the pam conversation? [08:40] you'll just get a 20 second timeout, then a nasty error message, and then the session crashes [08:41] robert_ancell: yes [08:41] it seems to me (not tested) that /etc/pam.d/gdm-autologin should force the password, and GDM should launch the greeter to request this [08:42] does this make sense? [08:42] if that works, sure; I'm not that familiar with how autologin works in gdm [08:42] robert_ancell: I liked your suggestion of just pre-selecting the user and asking the password [08:42] because you also have the fun situation that you might not even have a home directory until PAM completes, so GDM/gdmsetup can't guarantee that it can tell if the filesystem will be encrypted [08:42] robert_ancell: i. e. not trying to fix it in gdmsetup [08:43] robert_ancell: you mean for cases where pam mounts your encrypted home dir over the net? [08:43] yeah [08:43] yes [08:44] ok, I think I know the path forward then [08:44] robert_ancell: nice! [08:46] pitti, so, being marked as low priority we could defer this to 11.10 right? ahem *lightdm* *lightdm* [08:46] ;) [08:47] robert_ancell: you mean this bug? [08:47] yup [08:47] robert_ancell: as I said, if you don't have time to work on it, I'll find someone else (or me) [08:47] as it's a "shoot yourself into the foot" option, I'd rather have it fixed earlier [08:48] I'm working on the webkit one first. Having trouble reproducing it, you didn't reproduce either right? [08:48] hey rodrigo_, robert_ancell [08:48] robert_ancell: I did [08:48] didrocks, het [08:48] hey [08:48] robert_ancell: but only in the installer with the slideshow, on amd64 [08:48] hi didrocks [08:48] hey robert_ancell [08:48] robert_ancell: happens reliably every time there, including kvm [08:48] pitti, could you reproduce from your desktop install? [08:48] robert_ancell: I haven't found a way yet to reproduce in a running system; perhaps one can start the ubiquity slideshow standalone somehow [08:49] ubiquity crashes when I try and run it, I have the A2 image, but haven't tried it yet [08:49] * pitti plays around with this [08:50] hm, so how would I open a HTML page with webkit.. [08:50] /usr/lib/webkitgtk-1.0-0/libexec/GtkLauncher doesn't do it for me (doesn't show any of the images, just the navigation buttons). Firefox renders it fine [08:50] pitti, ^^ [08:50] oh, cool [08:50] but not seeing the crash or any errors in valgrind/gdb [08:52] $ for f in *.html; do echo $f; /usr/lib/webkitgtk-1.0-0/libexec/GtkLauncher $f; done [08:52] hm, they all work [08:52] apparently ubiquity launches that differently [08:52] no-one recorded what version of webkit it was, do you happen to know? [08:52] hey [08:52] robert_ancell: Hi, are you aware that your totem-pl-parser upload is in DEPWAIT on libquvi-dev which needs a MIR? [08:52] hey pitti robert_ancell [08:52] geser, no, I didn't notice that, thansk [08:53] libwebkitgtk-1.0-0 1.3.10-1ubuntu1 [08:53] robert_ancell: according to http://cdimage.ubuntu.com/releases/natty/alpha-2/natty-desktop-amd64.manifest [08:53] pitti, you reproduced with that, or that is what you have now? [08:53] got hit 3 times in a row by that compiz crashing in sn_... today, it's fun things start but you just see you background and have no way to fix it out of restarting the session [08:53] hey seb128 [08:53] robert_ancell: that was on the alpha-2 CDs, and reproduced [08:53] ok [08:53] salut seb128 [08:53] robert_ancell: but it's still current [08:53] pitti, I gtg, thanks for the infor [08:54] robert_ancell: sleep well! [09:01] didrocks: do you know if there's already a compiz bug to bring back proper "snap to screen edge" for window movement? [09:01] if you don't know off-hand, I'll search/create one [09:01] pitti: there is a bug about it, but only for nvidia driver [09:01] as smspillaz was telling it was because of the nvidia one [09:01] these "windows stick into the next screen by 2 pixels" is really annoying [09:02] pitti: yeah, for the additional pixels, it's fixed and will be uploaded next week [09:02] awesome [09:02] but the resitance on edge isn't working for me [09:02] do you have it working, [09:03] doesn't work for me either [09:03] didrocks, oh, compiz update will only be next week? [09:04] didrocks: no, it doesn't; even worse, it seems to actively move the windows so that they don't fit on the scren any more [09:04] seb128: sam has to make dist, he doesn't want to do it because fixing the invisible window bug and he didn't focus on fixing it yet [09:04] smspillaz: see, I'm not the only one not having the edge resistance working ^^ [09:04] my gsession script (which does all the geometry arrangement) hasn't changed in ages, and I even modified it to put the terminals 10 pixels to the left; but it still moves them around [09:04] so no a twinview only issue [09:05] and yes, the option is still activated by ccsm :) [09:05] * didrocks really wonders if grid can be the cause of that not working [09:06] didrocks: likely not [09:06] didrocks: edge resistance worksfinehere[tm] [09:07] didrocks: its only enabled for screen edges by default though [09:07] smspillaz: I know it's only enblaed for screen edge, I did it :) [09:07] ok [09:07] so what doesn't work then? [09:08] it seems to work for the left side [09:08] except that it stops ~ 10 pixels before [09:08] smspillaz: screen edge, as told yesterday :) [09:08] but not for the right side [09:08] pitti: that's a problem with the new decorator [09:08] pitti: works fine here on the right size [09:08] ah, "screen" != "virtual desktop" [09:08] *right side [09:08] pitti: oh, so it's working for you? [09:08] didrocks: on the left side (badly), but not on the right or bottom [09:08] it works fine here [09:08] seb128: do you have the same? (take into account the new decoration side) [09:09] pitti: the offset by 10px will be fixed in the next update [09:09] smspillaz: it doesn't work, even taking into account the decoration side [09:09] pitti: its because the new decoration frame window is that much bigger [09:09] didrocks: sure its enabled ? [09:09] ah [09:09] I think I was confused by the large gap [09:09] pitti: yeah [09:09] it actually does seem to work on the right side, too [09:09] smspillaz: I didn't change from yesterday when I told you I checked it's enabled :) [09:09] pitti: that also affected placement as well [09:10] smspillaz: perhaps that's what keeps making my windows spill over into the next screen then [09:10] pitti: it is [09:10] coolio then [09:10] didrocks, no, doesn't work here [09:10] smspillaz: thanks for the heads-up! nice to see things getting fixed so quickly [09:10] smspillaz: so, doesn't work either for seb128 and me [09:10] but again my system seems to trigger bugs [09:11] like the click-alt-dnd still doesn't work on focussed dialogs [09:11] seb128: yeah I hit that one [09:11] seb128: I'm going to fix it after I do the invisible window one [09:12] seb128: I know what it is, its that we had to add this stupid workaround to make GTK correctly and that is hard to implement properly [09:12] k [09:12] *make GTK work correclty [09:12] smspillaz, well, one thing at the time, get the invisible one fixed and roll a tarball [09:12] then we can talk about the next round ;-) [09:12] invisible - > alt -> tar [09:12] also I have a fix to make the switcher faster [09:23] hello mvo and pitti [09:24] hey glatzor [09:24] smspillaz, you can always make another fix before rolling a tarball [09:24] smspillaz, I would say fix the invisible dialogs one and roll a tarball [09:25] yeah [09:25] smspillaz, we need that one out to users, the alt-click and switcher issues can wait next roll [09:25] ok [09:25] smspillaz, thanks! [09:25] hi seb128 ! [09:26] hi pitti - you were up early today weren't you? [09:26] chrisccoulson: yeah, couldn't sleep any more [09:26] hey glatzor [09:29] chrisccoulson, it's a german thing I guess, not sleeping, he learnt that from asac [09:29] chrisccoulson, hey btw, how are you? [09:31] hi seb128, yeah, i'm good thanks [09:31] a little tired though [09:31] how are you? [09:32] hey glatzor [09:32] did your daughter woke you up early again? [09:32] I'm fine thanks [09:32] chrisccoulson: its the normal state of a parent, no? being tired? at least in the first years ;) [09:33] mvo - yeah, i always seem to be tired at the moment ;) [09:33] i guess i should get an early night! [09:36] I can see the gdm update next cycle being fun [09:37] they dropped most of the language and keyboard selection from it, it seems [09:37] they use the user account new things instead or something [09:38] seb128 - we'll be using lightdm won't we? ;) [09:38] if it works ;- [09:38] ;-) [09:38] heh [09:39] will we upgrade to the new control-center too? [09:39] next cycle? yes [09:39] lot of fun ;-) [09:40] indeed! [09:40] and there will probably be 2 firefox transitions too ;) [09:40] if mozilla stick to schedule [09:40] or you mean we will switch to chromium? ;-) [09:40] heh ;) [09:45] nobody has had problems upgrading firefox this morning? [09:45] the packaging layout changed a fair bit yesterday [09:49] chrisccoulson, I didn't try yet, you failed to build on armel btw [09:49] still? [09:49] i thought i worked around it ;) [09:49] http://people.canonical.com/~platform/desktop/versions.html says so [09:50] yeah, i just looked too [09:50] failed quite late in the build [09:50] 13 hours! [09:50] "cp: cannot stat `./debian/tmp/usr/include/firefox-4.0b11/vorbis': No such file or directory" [09:50] it failed after the build [09:50] in the dh_install [09:50] hmmmm :/ [10:22] bratsche: ping [10:52] http://mkestner.blogspot.com/2011/02/signs-of-life.html might interest some of you ;-) [10:52] seb128: ^ [10:52] Laney, thanks [10:53] good to have news! [10:55] session restart, brb === MacSlow is now known as MacSlow|lunch [11:03] pitti: did you forget to bzr push your dch -r "" on gtk? [11:03] checking, hang on [11:03] oops, apparently so; done [11:03] pitti: thanks :) [11:03] sorry about that [11:03] I keep forgetting to push the dch -r [11:03] usually I bzr bd -S, upload, and then push [11:04] I should probably do that the other way round [11:04] pitti: no worry :) yeah, we should have a wrapper maybe [11:06] bah, firefox crashes on http://sourceforge.net/projects/flac/files/ [11:06] well b10 does, I need to upgrade and test with b11 [11:07] seb128 - doesn't crash here ;) [11:07] seb128 - did you submit the crash report? [11:07] heh, crashes here, too [11:07] Version: 4.0~b11+build3+nobinonly-0ubuntu2 [11:08] chrisccoulson, no, I don't like much sending browser crashes [11:08] since the browser tend to have sensible infos, password, etc [11:08] * pitti submits [11:08] pitti - could you send me the crash ID once you're done? [11:08] I'm also using an outdated version [11:08] chrisccoulson: how do I know when it's done? [11:08] speaking of which is there an url to watch current beta tarballs? [11:09] chrisccoulson: the "restart" thing has never ever worked for me [11:09] pitti - you don't get a confirmation when it's done, but you can look in about:crashes [11:09] chrisccoulson: https://crash-stats.mozilla.com/report/index/bp-4780dc92-9a7e-4e55-af0b-b335a2110210 [11:10] seb128 - there's an ftp server [11:10] (and also mercurial) [11:10] wow, this thing is clever [11:10] seb128 - ftp://ftp.mozilla.org/pub/firefox/releases/ [11:10] it refers to https://bugzilla.mozilla.org/show_bug.cgi?id=632867 [11:11] Mozilla bug 632867 in General "Firefox Beta 11 crash [@ libc-2.12.2.so@0x33b45]" [Critical,New] [11:11] thanks [11:11] yeah, it's good there's a bug already [11:12] hmmm, not sure why the traces are like that :/ [11:12] * chrisccoulson checks our symbols are on the server [11:14] yeah, they're all there [11:15] pitti - mind getting a GDB backtrace with --sync? [11:15] it looks like it's an X error [11:16] chrisccoulson: gdb --args firefox --sync ? [11:16] hm, apparently not [11:16] pitti - firefox -g --sync should do it [11:17] we have some logic in the shell script to set up the right environment ;) [11:18] chrisccoulson: http://paste.ubuntu.com/565324/ [11:18] not too helful, I guess -- do I need more -dbg packages? [11:18] ###!!! ABORT: X_GLXVendorPrivateWithReply: BadLength (poly request too large or internal Xlib length error): file /build/buildd/firefox-4.0~b11+build3+nobinonly/build-tree/mozilla/toolkit/xre/nsX11ErrorHandler.cpp, line 190 [11:19] _XError+0x000000F7 [/usr/lib/libX11.so.6 +0x00040F97] [11:19] this probably? [11:19] pitti - yeah, it looks like firefox-dbg, and whatever you need for mesa and libx11 too [11:19] pitti - yeah, the abort message is on your crash report too [11:19] singe it's sigabort, it sounds like an assertion failure [11:20] firefox-dbg> ugh, 109 MB [11:21] chrisccoulson: hm, some trouble with the mesa dbg ones [11:21] apparently I'm still newer than natty, I have glx from xorg-edgers [11:21] * pitti tries to sort that out first [11:22] pitti - are you on nvidia btw? [11:22] intel [11:22] oh [11:22] now that's weird [11:23] it should be impossible to create GL context on != nvidia, as other drivers were blacklisted in firefox [11:23] (for this reason) [11:23] hmmmm :/ [11:23] I downgraded a lot of mesa stuff now, but libdrm-nouveau1a is still uninstallable [11:23] oh, libdrm perhaps [11:24] ah, there [11:30] pitti, chrisccoulson: I'm on intel as well [11:30] but stock natty, no xorg-edger or anything like that [11:30] * Sweetshark is on nouveau. [11:30] Sweetshark, is http://sourceforge.net/projects/flac/files/ crashing your firefox? [11:31] Are you guzs offering an opportunity to crash my desktop? [11:31] * Sweetshark tries [11:31] seb128 - if you run firefox from a console, do you see "[GLX] currently only allowing the NVIDIA proprietary driver, as other drivers are giving too many crashes. To bypass this, define the MOZ_GLX_IGNORE_BLACKLIST environment variable." when you access the website that crashes it? [11:31] (you should do) [11:31] seb128: no [11:31] i see it here, which is why i don't see the crash :/ [11:31] chrisccoulson: http://paste.ubuntu.com/565327/ [11:31] but it seems that firefox thinks you're on nvidia ;) [11:31] * Sweetshark starts ff from term [11:32] pitti - excellent, thanks [11:32] will attach to the bug report [11:32] chrisccoulson, no, just the abort lines [11:32] Sweetshark: every crash opportunity is good to take :) [11:32] seb128 - oh, that's the problem then :) [11:32] chrisccoulson, did you try on your latitude? [11:32] * Sweetshark is GLX blacklisted [11:32] seb128 - yeah. i get the console message here about being blacklisted [11:32] weird [11:33] but the fact that you don't seems to be the real bug ;) [11:34] chrisccoulson, how is firefox detecting the video card and driver in use? [11:35] seb128 - http://hg.mozilla.org/mozilla-central/file/afb62e95311a/gfx/thebes/GLContextProviderGLX.cpp#l228 [11:35] not sure how it gets that information yet [11:41] pitti / seb128 - i guess that neither of you are setting MOZ_GLX_IGNORE_BLACKLIST in the environment are you? ;) [11:41] no [11:41] env | grep MOZ gives nothing [11:41] that's good. i had to ask ;) [11:41] ;-) [11:42] chrisccoulson: right [11:45] the correllations stuff on crash-stats is quite nice: https://crash-stats.mozilla.com/report/index/9ba1f89e-55db-4cdb-a270-e84cc2110208 [11:45] 71% of people with that crash also have ubufox ;) [11:45] not that that is of any use [11:45] and 71% are on an amd64 with 2 cores ;) [11:47] do you stats on those who are on intel? ;-) [11:47] video driver speaking [11:47] seb128 - no, i don't think that information is submitted. i'll talk to the guy working on breakpad and ask him if he wants a patch to submit that [12:13] pitti / seb128 - OOI, what does glxinfo say for "server glx vendor string"? [12:14] hmm, switch user is broken, it seems [12:14] or I'm missing some update [12:14] firefox is using the same string there [12:14] $ glxinfo | grep -i vendor [12:14] server glx vendor string: SGI [12:14] chrisccoulson, ^ [12:15] seb128 - oh, that's the same as me then [12:15] hmmm :/ [12:15] also, gdm doesn't start when rebooting [12:15] well, it starts, but doesn't show up the login screen, anybody's seen that? [12:15] rodrigo_, is that new? when did you upgrade? do you get any error? [12:16] seb128, I'm not full up-to-date, just rebooted an hour ago or so [12:16] rodrigo_, no but i didn't upgrade or restart yet today [12:16] what video card and driver? [12:16] nvidia [12:16] k, so might be it [12:16] wqith startx it works, running an X session now [12:16] well the binary drivers are not available for the new xserver abi [12:16] hum, dunno about this issue [12:17] I haven't upgraded any of the xserver* pacvkages [12:17] did you try restarting? [12:17] i.e is that a one time thing? [12:17] yes, that¡s why I was 1 hour offline :) [12:17] I guess my system doesn't like being half-upgraded [12:18] * rodrigo_ looks at upgradable packages [12:18] seb128, apart from the gdm issue at restart, does user switching work for you? [12:18] seb128 or anyone else? [12:19] will it screw my current session if I try? ;-) [12:19] rodrigo_, well it relies on gdm so if you used startx it will not work [12:19] seb128, maybe, it just does nothing for me, so try later if you can't afford losing your current session :) [12:19] seb128, right, I mean before restarting, it didn't work, that's why I restarted after upgrading some packages [12:20] now, with startx, of course it doesn't work [12:20] rodrigo_, well I just tried a guest session, that worked [12:20] ok [12:20] so, it's me [12:20] seb128, are you fully up-to-date? [12:21] no, I didn't dist-upgrade since monday [12:21] I did pick some selected upgrades though [12:21] yes, that's what I do [12:21] but I guess most people here are uptodate [12:22] ok, I'll sacrifice a goat and then dist-upgrade [12:22] rodrigo_, did you restart only once? [12:22] seb128, several times [12:22] ok [12:22] what did you upgrade before getting the issue? [12:22] gnome stuff mainly, from the gnome3 ppa [12:22] rodrigo_, well don't dist-upgrade if you are on nvidia and need unity [12:22] and some other packages [12:22] ah, right [12:23] try doing a sudo gdm in your session maybe and see if it complains about something [12:23] yes, also, I see I've got a gdm update [12:23] or look at the /var/log/gdm logs [12:23] trying both [12:24] you can try downgrading gdm... [12:25] ok, brb === MacSlow|lunch is now known as MacSlow [12:34] ok, found it [12:35] rodrigo_, oh, what was it? [12:35] seb128, the gdm logs show a "GTK3 symbols found on GTK2 process" error, so my gdm is loading some GTK3 stuff, not sure what [12:36] weird [12:36] hmm, 1st thought was g-screensaver, which I upgraded yesterday with a new version I uploaded to the gnome3 ppa [12:36] but seems that's not the issue [12:42] hm. AttributeError: 'gi.repository.Gtk' object has no attribute 'FILE_CHOOSER_ACTION_OPEN'. How can I find out what it's called in the New World Order of g-i? [13:14] seb128, btw, I'll be gone for an extended time during my midday today, if you try to ping me and I don't answer [13:14] hey mterry [13:14] mterry, ok, no worry [13:14] mterry, I've another bug for you if you run out of things to do ;-) [13:15] seb128, dammit :) [13:15] lol [13:15] well appmenu is solid so great work on that ;-) [13:15] mterry, the g-s-d keyboard layout indicator is buggy, the label doesn't update when the layout is changed [13:16] not sure if that's the g-s-d patch or an indicator issue [13:16] well I think I already assigned it to you on launchpad, not sure if you noticed [13:18] seb128, yeah I did, but kenvandine said people were investigating, so I held off [13:18] seb128, seems like there's more agreement now about the issue [13:21] mterry, well that bug is a mix [13:23] there is 3 issues [13:23] 1- the indicator label doesn't update [13:23] like if you pick an another layout in the menu or use the keyboard to switch [13:24] 2- if you start with 1 layout, add another and delete it the menu stays instead of hidding [13:24] but it's in broken state, i.e it stops working [13:24] 3- if you stop the indicator-application service when being in 2- gsd crashes [13:26] mterry, bug #711749 [13:26] Launchpad bug 711749 in indicator-application "layout indicator items are not being updated and not working" [High,Confirmed] https://launchpad.net/bugs/711749 [13:27] mterry, bug #712635 is the crash [13:27] Launchpad bug 712635 in libappindicator "gnome-settings-daemon crashes with SIGSEGV in watcher_owner_changed()" [Undecided,New] https://launchpad.net/bugs/712635 [13:27] bug #712581 [13:27] Launchpad bug 712581 in indicator-application "keyboard layout label is not updated" [Undecided,New] https://launchpad.net/bugs/712581 [13:27] mterry, check with kenvandine if he's working on those maybe [13:27] but ted seems busy with features until feature freeze so I doubt he will [13:27]