[00:53] hmm...I seem to have lost quite a few icons in my system settings after latest update and reboot [00:53] anyone else have same issue? [00:54] wait, nm [00:54] they are back? weird [01:42] RAOF: for what it's worth, i didn't do a whole lot of review outside the scope of my complaint, and i don't really speak GL [01:43] That's ok. I'm confident in the GL bit, it's the associated stuff that I was concerned about. [01:44] Thanks for the review. [01:44] (and i was particularly sensitive to the multithread issue, because (a) libpcsclite caused problems for me by doing that and (b) i'm currently writing code to use libgnome-desktop from a multithreaded app, even though it won't be under unity so this code shouldn't trigger in the first place :-P) [01:46] :) [01:47] Hm, performance optimisation! Only check for GL_MAX_TEXTURE_SIZE when running under unity. [01:49] oh, huh - i guess i would have hit it, wouldn't i? :) [01:49] Yup :) [01:50] Anyway, it now fork()/exec()s a small helper instead. [01:50] i am using compiz, so i need to do something about this case. i haven't decided yet whether i want to muck with the environment to force your check or enable sam's plugin to split up the textures [01:52] You should be able to just enable that plugin; Sam swears blind that it works :) [01:53] this is the copytex plugin, right? [01:58] That is my understanding, yes. [01:59] RAOF: did you try that? [02:00] enabling it and plugging in a monitor on your netbook [02:00] It doesn't work for Unity. I guess I could try it in a straight compiz session... [02:00] Sadly, that means I need to get off this nice warm sunny verandah and go inside my office. [02:01] nah thats not important [02:01] its not like we have any straight compiz sessions anymore [02:02] thats what i was curious about though, unity is still broken with it enabled [02:02] i have hardware i can try it on here, though i need to grab an oneiric build - i only have natty easily accessible === SoulShadow is now known as Shadows`sleep [04:10] Good morning [04:14] Morning pitti. [04:16] hey pitti good morning! [04:17] hey pitti [04:18] pitti: re: the NM bug, I was able to confirm it's really an issue with libnl, not NM. libnl3 fails, and if I rebuild against libnl1 things work just fine [04:20] cyphermox: ah, nice; is libnl so hard to debug? [04:20] painful [04:20] at least to me it is ;) [04:20] cyphermox: you could create a small isolated test case with just the call(s) that NM does to set a route, and then send that to the upstream bug and use it for git bisect to find the commit that broke it [04:21] I just sent an email with my findings so far the the libnm mailing list [04:21] nice idea [04:21] yeah, I think that should be fairly easy to do [04:21] that's my standard approach if I don't know the upstream code and it's not obvious where in it the bug happens [04:21] haven't really looked at libnl source so much yet [04:22] I only just tried updating libnl3 to 3.1 or 3.2.1 (we have 3.0 now), in case it's alrady been fixed, but no luck yet [04:50] freaking libc. Why is our software all dependant on this pile of crap? setenv has a bug, which means g_setenv has a bug, which means indicator-datetime has a bug, which means language support is randomly broken in unity-greeer [04:50] greeter [04:52] *setenv* has a bug? [04:52] RAOF, yup, setenv with NULL value partially works [04:52] except for a crucial \0 character that is missed [04:53] Whoops! [04:53] so indicator-datetime does a x=getenv("LANGUAGE");usetenv("LANGUAGE");setenv("LANGUAGE",x); to workaround the crapness of gettext and it all goes horribly wrong when LANGUAGE is undefined [04:55] :( [04:56] jbicha: hello [04:56] jbicha: hm, seems ubuntu:mutter is at 3.1.92-0ubuntu2 while archive is at 3.1.92-0ubuntu3 [04:56] jbicha: but it seems you already manually folded that into your merge? [04:58] jbicha: right, seems you did; thanks [05:46] good morning [05:46] bonjour didrocks [05:46] guten morgen pitti [05:57] didrocks: something else with bug 856884 a brand new user also gets those gconf settings on first login to Unity [05:57] Launchpad bug 856884 in unity "Running unity --reset breaks metacity keyboard shortcut defaults" [Low,In progress] https://launchpad.net/bugs/856884 [05:57] so it's not unity reset that's broken [05:57] GAK. For future reference, when testing a new libgnome-desktop CRTC restriction, you need to both restart gnome-display-properties *and* gnome-settings-daemon. [05:58] jbicha: it is the fact to reset and change compiz keys [05:58] as I explained :) [05:58] RAOF: in fact, g-display-properties doesn't do much by itself, just pokes g-s-d, right? [05:58] jbicha: btw, did you succeed in testing your patch? [05:58] RAOF: did you find out how to start g-s-d in the foreground in a terminal? I remember it to be a bit of fiddling [05:58] pitti: It *does* query libgnome-desktop for the configuration, which is why I was confused when it actually set the mode. [05:59] Eh, I just log out and log back in again :) [05:59] right, but IIRC it just writes monitors.xml and lets g-s-d pick up the file change and act on it [05:59] RAOF: ah, I found it more convenient to kill it and watch its debug output in a terminal [05:59] didrocks: no, I've been swamped this week, but...my approach wasn't going to fix new user login [06:00] jbicha: it's the same, compiz is copying a bunch of key in /apps/compiz-1/ [06:00] jbicha: so I guess it has the same side-effect [06:00] pitti: Yeah. Whereas I was finding it more convenient to watch gnome-display-properties write out the debug messages into the terminal. And then wondering why it was still setting the mode it quite clearly knew was invalid :) [06:00] So, convenient. Just not correct! [06:04] pitti, are you ok with bug 860246 - there's no string changes so it shouldn't affect the translators [06:04] Launchpad bug 860246 in unity-greeter "UIFE: Fade username labels to give carousel effect" [Wishlist,Confirmed] https://launchpad.net/bugs/860246 [06:05] robert_ancell: FWIW, I think we have bigger fish to fry, but it was acked, yes [06:06] pitti, yeah, I just rolled it in because of the indicator bar change (the code has been there for some time, but it's just been disabled) [06:11] There we go, merge proposal resubmitted. [06:16] pitti, do you have any ConsoleKit debugging tips? shutdown/restart doesn't work in the greeter anymore, are there any logs that might say why? [06:17] robert_ancell: did you try running the CK daemon in the foreground console? it's normally not logged anywhere [06:17] pitti, just killing it and running it as root? [06:18] robert_ancell: yes; NB that this will kill all your existing user CK sessions, so better do it on a VT [06:18] robert_ancell: if you want to do this more often, you can also edit /usr/share/dbus-1/system-services/org.freedesktop.ConsoleKit.service [06:18] robert_ancell: and append a "> /tmp/ck.log 2>&1" [06:19] robert_ancell: whoops, you need to wrap that into a /bin/sh -c "..." [06:19] robert_ancell: but for one-off debugging, running in foreground works fine [06:19] ok, I may or may not be back soon then :) [06:19] robert_ancell: oh, you can leave your session running [06:19] robert_ancell: but after killing CK you won't be able to do hardware stuff any more like mounting USB devices and the like [06:19] np [06:19] brb [06:28] pitti, aha. It's failing to shutdown if you have more than one session open. My user can do that (must have the appropriate PK permissions). Do all users have that? Or is this an indicator bug that it doesn't detect it's not allowed to shutdown? [06:32] robert_ancell: IIRC all admin users can do that. [06:32] RAOF, yeah, that's what I'm guessing. [06:34] auth_admin_keep [06:34] So, if lightdm had a polkit frontend running, you could shutdown after entering an admin password. [06:35] RAOF, right. (it doesn't) [06:35] I just tried a non administrator account, and that just returned me to the login scree [06:35] n [06:38] pitti: for u-l-m, if we have the confirmation that the fix needed is in banshee for the "I'm not already logged in" case, are you happy with acking the FFe? [06:42] does anybody have a copy of matlab OR happen to know a crazy java application which displays this bug: https://bugs.launchpad.net/unity/+bug/857201 [06:42] Ubuntu bug 857201 in unity "Java application windows cut-off/truncated/not displayed properly" [High,Confirmed] [06:45] didrocks: how does knowing where the bug is improve the user experience? [06:45] didrocks: is it realistic to actually fix this in banshee by release? [06:45] pitti: they already fixed those cases in banshee I set as criteria, one sec: [06:45] didrocks: I expect that I don't actually have the option to say no :) [06:47] pitti: you know I'm not the one pushing the crack, but on that case, I'm pretty confident if they confirm the issue is in banshee [06:47] so, pitti, I defined 3 bugs I had with previous version to get u-l-m proposed to oneiric [06:47] bug #840758 which is a crash in unity (fixed with latest release) [06:47] Launchpad bug 840758 in unity "compiz crashed with SIGSEGV in dee_model_get_tag()" [High,Fix released] https://launchpad.net/bugs/840758 [06:47] bug #856542 and bug #856547 [06:48] Launchpad bug 856542 in banshee "when selecting a track in the music store scope, banshee is started, but no switch to the store" [Medium,Fix released] https://launchpad.net/bugs/856542 [06:48] Launchpad bug 856547 in unity-lens-music "when selecting a track in the music store scope, banshee started, the right track in the music store view isn't selected" [Medium,Fix released] https://launchpad.net/bugs/856547 [06:48] the two previous ones were issues in banshee (even when logged in to u1ms) [06:48] they fixed in banshee [06:49] I guess the "not logged case" (at least being driven to the u1ms scope in banshee) can be dealt as well, isn't it? [06:49] I sure hope so; I think the u1 login dialog crash is independent from this as well (but still spoiling the user experience, of course) [06:51] didrocks: you where involved in the zg-datahub 20s delay for the autostart... I still see datahub starting immediately on login here...' [06:51] pitti: is it a crash or does the sso login dialog vanish? [06:51] kamstrup: really? Are you sure you don't have any local version installed? [06:51] didrocks: yes [06:52] didrocks: just tested it again, this time the U1 dialog stayed around, so not reproducible [06:52] kamstrup: need a bootchart I guess [06:52] pitti: I bet it's because banshee music store scope uses the old API [06:52] didrocks: so I guess the remaining bug is banshee not switching to the store for results [06:52] didrocks: perhaps... or ted's magical dbus-logger for the login sequence [06:53] kamstrup: seb128 did a bootchart after my fix, was working [06:53] hmmm, ok [06:53] didrocks: when does the 20s start from? When I hit enter on the password? [06:53] pitti: indeed, when you are not logged in [06:53] maybe my tests are just bad [06:53] pitti: if you are logged in, it works, isn't it? [06:54] kamstrup: just a little bit after (when gnome-session starts the application phase) [06:54] didrocks: ok, let me try and bump the delay to 60s which should give me ample time to verify this [06:54] kamstrup: but you can count from there [06:54] kamstrup: indeed [06:54] kamstrup: otherwise, maybe something new do some weird dbus activation? [06:55] didrocks: yeah, my fear is that we dbus activate it indirectly somehow [06:55] didrocks: testing now, just rebooted to get everything [06:55] * kamstrup restarts session [06:56] pitti: sure [06:56] pitti: that should work with the latest version, at least, it works there :) [06:56] didrocks: no, doesn't work for me [06:56] you have latest banshee, isn't it? [06:56] ah, now; it just takes some 10 seconds to switch [06:56] yeah [06:56] it's slow [06:56] as always with banshee… [06:56] and there's no throbber [06:57] broder: If you'd like to re-review the gnome-desktop merge, it's refreshed. [06:57] kamstrup: oh, I synced the new zg since, can that have any influence? [06:57] didrocks: meh, and "preview" playing doesn't work either -- I wonder if anyone ever tests the banshee store :( [06:57] didrocks: unlikely [06:58] meh, and there crashes banshee again [06:58] pitti: urgh? I didn't test it, just check that if banshee closed/opened, or in another view, we have the banshee store switched to the right album [06:58] pitti: as all of that was not working before and I got some "we can't fix that", so had to persuade them… [06:58] I must say that I didn't test the music store itself (buying, preview…) [06:59] I guess we need to file bugs to the u1 team? [06:59] yeah [07:00] * didrocks waits on the music store to refresh… [07:00] pitti: preview works there [07:02] pitti: tried on various albums/tracks and it works [07:02] you don't get crashes either? it crashes for me after three or four clicks [07:02] apport doesn't catch it as it forks, getting trace from stderr [07:03] no, let me try with my personal music [07:03] well, it's slow, really slow… [07:04] didrocks: call of the alarm :-) in an epic stroke of irony, the way I detected zg was by clicking the launcher and starting a gnome-terminal... this in turn dbus activates zeitgeist, logging the app usage ;-) [07:04] talk about a "quantum bug" [07:04] the act of measuring the bug causes it! [07:04] :-D [07:04] pitti: works as well for a local playlist and works [07:05] kamstrup: ahah, excellent story! :-) you finally had a bug :p but it's your head itself [07:05] didrocks: so the question is - if no one is around to observe a bug, is it still a bug?! [07:05] kamstrup: ctrl + alt + T FTW! :) [07:06] didrocks: preview works in a guest session; might be some old banshee db laying around for me [07:06] kamstrup: hum, interesting question, when you close the door and through the key, is the cube in next room white? is there even a cube? :-) [07:07] didrocks: can you try in a guest session, search something, and then login with your account? [07:07] didrocks: I get a 404 page for the music store then [07:07] pitti: yeah, let me try, /!\ guest session plays bad with nvidia hardware, so 5/10 minutes to be able to act again on this session [07:08] just pushing Qt before as bzr bd -S takes 20 minutes as it's almost finished. If the guest session makes my X swapping and crashing, I won't be happy :) [07:08] good morning all and sorry to interrupt: is "suspend after 30 min. inactivity" supposed to be enabled by default? this seems to be a recent change as i almost made this a bug report against the -12 kernel. [07:10] didrocks: I tried it again in guest session, reproduced; filing bug now [07:12] pitti: seems the initial u1ms account log in experience isn't great… [07:12] pitti: lintian still running in Qt source package, will confirm then [07:12] I made some screenshots, hope Unity doesn't change too much now :) [07:12] ok, after removing all banshee config/cache files it works again [07:12] but not handling old config/cache gracefully is also not great [07:13] who on earth wanted banshee to be the default :( [07:13] jbicha: nothing that would need retaking screenshot, making some color adjustement to make the text more readable on white wallpaper background, but that's it (this doesn't need screenshots) [07:13] chrisccoulson: For your information, the GL_MAX_TEXTURE_SIZE check adds 150 msec to g-s-d's xrandr plugin's startup (and approximately nothing to startup under Unity 2D) on my atom. [07:14] pitti: TBH, I really have the impression it regressed since the UNE time [07:14] pitti: it was really snappy on netbook a little bit more than a year ago [07:16] didrocks: bug 861185 [07:16] Launchpad bug 861185 in banshee "Ubuntu one music store: 404 page after logging in" [Undecided,New] https://launchpad.net/bugs/861185 [07:16] Qt pushed, trying guest session [07:16] didrocks: good morning! did you see bug #859632 ? did show up in the auto upgrade tester this morning [07:16] Launchpad bug 859632 in compiz-plugins-main "package compiz-plugins-main-default 1:0.9.5.94 bzr20110919-0ubuntu1 failed to install/upgrade: trying to overwrite '/usr/share/compiz/grid.xml', which is also in package compiz-plugins 1:0.9.4 bzr20110606-0ubuntu1~natty2" [Undecided,New] https://launchpad.net/bugs/859632 [07:17] mvo: oh interesting, I though I dealt with this, but maybe my Replaces: doesn't deal with the SRU, will fix it (a new compiz is scheduled today), thanks! [07:19] didrocks: yw :) [07:19] didrocks: bug updated [07:19] hey mvo, good morning [07:21] meh, last unity upload made the Win+Number assignment unstable; /me finds reproducer in guest session and files [07:21] hey pitti! good morning [07:25] pitti: I can't confirm it because i can't log in sso [07:25] pitti: same with software-center [07:25] didrocks: oh? [07:25] I have "one moment, please wait", with the spinner [07:26] I waited for a minute [07:26] it's basically instant here [07:26] weird… I still connected yesterday for some software-center tests [07:26] maybe a temporary glitch? [07:26] pitti: however, it starts banshee for me, and go to u1ms [07:27] and then ask for the prompt… [07:27] (slowly of course, but it does) [07:30] mvo: fix pushed to bzr, thanks! [07:32] ta [07:34] agateau, hi [07:35] tkamppeter: hi [07:35] agateau, any news about the hp-systray problem? [07:35] tkamppeter: I updated a Qt patch to fix it, it has been pushed this morning [07:36] tkamppeter: and I am preparing a release of sni-qt with the fix for the empty menu [07:38] tkamppeter: when both hit the archive, bug 860395 should be history [07:38] Launchpad bug 860395 in sni-qt "sni-qt and hp-systray interact badly under GNOME Unity" [High,Confirmed] https://launchpad.net/bugs/860395 [07:42] agateau, OK, I see on the BZR that Qt 4.7.4-0ubuntu5 is prepared, I assume that it is building currently. [07:43] tkamppeter: I think that's it [07:43] (reviewing Qt from unapproved ATM) [07:46] morning [07:46] hey rodrigo_ [07:47] agateau, bug 860395 has received the "Fix Released" message for Qt a minute ago. [07:47] Launchpad bug 860395 in sni-qt "sni-qt and hp-systray interact badly under GNOME Unity" [High,Confirmed] https://launchpad.net/bugs/860395 [07:47] pitti: thanks for approving, hopefully latest Qt :-) [07:47] tkamppeter: good! [07:48] agateau, thanks for the Qt patch and didrocks, thanks for applying and uploading it. [07:48] yw tkamppeter [07:48] * didrocks now looks at agateau for a sni-qt release [07:48] * agateau is busy running manual tests for sni-qt to please didrocks [07:49] * didrocks is happy that agateau is doing manual tests :-) [08:04] hey [08:05] salut seb128 [08:05] bonjour seb128 [08:05] hey didrocks pitti, how are you? [08:05] I'm fine, thanks, you? [08:06] bit tired but fine, I stayed up late to catch up with robert_ancell [08:06] i.e stayed on IRC until 2am [08:06] I've been running lightdm trunk for a few days and had some issues I wanted to make sure he has on his list before rolling the new version [08:06] hi seb128 [08:06] seb128: urgh, late evening indeed :) [08:07] hey rodrigo_ [08:10] how do I install a png I have in debian/ so that it's added to the package? binary-install/...: doesn't seem to do it [08:10] Aloha [08:10] hi czajkowski [08:12] rodrigo_: Add it to the .install file? [08:12] rodrigo_: howdy [08:12] RAOF, yes, already added, but I want to copy it from debian/ to the correct place [08:13] didrocks: your sni-qt release is ready for your packaging pleasure, sir [08:13] rodrigo_: you can do that in the .install file [08:13] agateau: that's so qt :-) [08:13] :) [08:14] micahg, just adding it there and it knows where the source file is? [08:14] doesn't seem to work [08:14] rodrigo_: no, debian/path/to/png /path/on/system [08:14] mvo, ping about logos in USC [08:14] micahg, ah === ara is now known as Guest99538 [08:16] aquarius: pong [08:18] rodrigo_, btw do you know how to add a png to the debian dir? [08:18] rodrigo_, if not don't waste time on it just ask there [08:18] rodrigo_, it's a bit non trivial since it's a binary and can't be diffed [08:18] so dpkg will complain [08:19] seb128, hmm, bzr add doesn't work? [08:19] mvo, I've just been asked by someone this question: they went to opera.com and downloaded the opera deb, which opened in software centre (yay), but USC didn't show an opera logo on the install screen, and it looks like it should do, top left - instead it shows a "generic" icon. (A cardboard box with rulers and books in.) Does the icon for a package come in the deb itself, or from somewhere else? [08:19] seb128, oh, not sure what you mean [08:19] rodrigo_, dpkg-buildpackage will fail to create a diff.gz [08:19] oh [08:19] rodrigo_, you will get a "can't diff that binary" [08:20] aquarius: for debs we are not really good about screenshot/icons at this point, we could try to be more clever and inspect the deb for common icon locations, but that is currently not implemented. so the best is to lobby to get opera into partner :) [08:20] seb128, so, how do I add it then? [08:20] mvo, aha, right, OK. Opera used to be in partner, didn't it? [08:21] rodrigo_, you are lucky it's easier in source v3 which g-c-c use, before we used to ship an uuencoded png and decode it from the rules ;-) [08:21] heh [08:21] aquarius: skype is, I don't think that opera is [08:21] aquarius: or was [08:22] aquarius: but I may be wrong, it fills up on the stable release which I don't run usually so I miss apps often [08:22] rodrigo_, I'm trying to find it back on google, you need to add the png to a file in debian/source [08:22] mvo, so at the moment we've got a Big List Of Icons For Packages We Support somewhere? [08:23] aquarius: there are various way, we support downloading icons on demand but for the stuff in the archive the icons are part of the app-install-data-ubuntu package [08:23] aquarius: actually I think if opera would include xb-appinstall-icon-url in its debian/control file it may even work with a deb [08:23] mvo, ah, ok -- and stuff in partner gets into that list? Where are on-demand icons downloaded from? (That is: would it be easier for Opera to get into that list?) [08:24] mvo, aaah, that sounds interesting. How would I confirm that debian/control thing? (Is it easier to just ask you? ;)) [08:24] aquarius: IMO the best would be to get them into partner, we have a app-install-data-partner package. the trouble with debs is always that they don't get updates (unless they ship their own sources.list.d snippet) [08:25] didrocks, mvo, pitti: help, what's the file you need to list pngs you add to the debian dir to? [08:25] mvo, yeah, but I am 62% sure that opera *used* to be in partner, which suggests that it got pulled for a reason. (Maybe they should go through MyApps?) [08:25] rodrigo_ needs it for g-c-c [08:25] with source v3 [08:26] something in debian/source/... iirc? [08:26] seb128: debian/source/include-binaries [08:26] didrocks, thanks [08:26] rodrigo_, ^ [08:26] aquarius: hrm, so AppInstall-Icon-Url is currently not supported for deb files, but that should really be simple to implement [08:26] so just adding the name of the file there is enough? [08:27] aquarius: and I think you are right, it used to be there a good while ago [08:27] rodrigo_, should be yes [08:27] rodrigo_: yeah, the relative path of it [08:27] mvo, OK. So, I'll talk to them about partner (or maybe the ARB?) [08:28] didrocks, seb128: ok, thanks [08:28] aquarius: http://developer.ubuntu.com/ got updated, that should hopefully make it easier to give them the required info. how does opera work, do they use license-keys for purchase? or do they have a free and a pay version? [08:29] mvo, neither: it's free but closed-source. [08:29] mvo, there's no paid version [08:29] oh, g-c-c already has a include-binaries, and it lists missing files [08:30] aquarius: ok, so http://developer.ubuntu.com/publish/ should hopefully cover it then [08:30] rodrigo_, that's probably from debian [08:31] seb128, ok, so I leave them? there seems to be no error from them missing [08:31] mvo, thanks! [08:31] hey aquarius [08:31] RAOF: are you able to reproduce that _XFreeEventCookies crash ? [08:31] heya rodrigo_ [08:33] rodrigo_, no, you can clean them, that's buggy [08:33] rodrigo_, it doesn't break anything but still buggy ;-) [08:33] seb128, ok [08:35] rodrigo_, we have extra bugs waiting for you btw :p [08:35] seb128, g-c-c related? [08:35] rodrigo_, the g-c-c filter is broken, you can type anything you want in it and you get an empty grid [08:35] ok, bug #? [08:36] I won't upload my g-c-c package then [08:36] bug #860952 [08:36] Launchpad bug 860952 in gnome-control-center "[Oneiric] gnome-control-center filter/search bar does not work" [Undecided,New] https://launchpad.net/bugs/860952 [08:36] although I already debcommit -r [08:36] bug #860950 [08:36] Launchpad bug 860950 in gnome-control-center "Search filter in System Settings in not working" [Undecided,New] https://launchpad.net/bugs/860950 [08:36] so IO guess I'd better upload it [08:36] rodrigo_, do an upload with the current fixes, we can do another one later or tomorrow [08:36] yeah [08:36] there are 4/5 fixes in this one [08:37] * rodrigo_ uploads [08:37] \o/ [08:37] rodrigo_: yes, please; tomorrow is hard code freeze, so tomorrow it's much harder to argue for fixes going in [08:38] waoh [08:38] waouh [08:38] lightdm 1.0 \o/ [08:38] pitti, ok, I'll do another upload with the other fixes [09:24] nedd to buy some food, bbiab [09:45] rodrigo_, seb128, didrocks: has the pad been unstable for you as well recently? [09:45] should we perhaps create one on pad.ubuntu.com? [09:45] +1 [09:45] +1 [09:46] yes, I keep being disconnected [09:46] ok, will do that and put a note on the old pad [09:46] with chars sometimes being dupped or text that didn't get commited [09:46] pitti, danke [09:50] http://reports.qa.ubuntu.com/reports/boot-speed/dell-vostro-3400/index.html looks like awesome progress [09:50] pitti, yeah, somewhat ureadahead started working recently [09:50] main progress here is the xrandr [09:51] well the < beta2 charts are buggy [09:51] seems like ureadahead didn't work they were hitting ios a lot [09:51] didrocks, agateau, I have the next problem with the hp-systray icon: bug 861270 [09:51] Launchpad bug 861270 in hplip "systray.py crashed with SIGSEGV in gtk_widget_style_get()" [Medium,New] https://launchpad.net/bugs/861270 [09:52] tkamppeter: didn't you say that you tested agateau's fixes? [09:52] pitti, http://reports.qa.ubuntu.com/reports/boot-speed/dell-latitude-e6510/index.html is back to maverick times [09:52] better than natty [09:52] desktop is especially better [09:52] didrocks: tkamppeter: sounds like infamous overlay-scrollbar bug [09:52] but plumbing got quite a hit [09:53] agateau: indeed [09:53] didrocks, yes, I tested them, yesterday the one for getting the menu, and yesterday I could open the Settings window, and today the appearing of hp-systray on login. [09:53] didrocks, agateau, so the bugs of yesterday are all fixed. [09:54] tkamppeter: can you try starting hp-systray with "LIBOVERLAY_SCROLLBAR=0 hp-systray" ? [09:54] didrocks, agateau, the bug of today probably occured this night. [09:54] tkamppeter: does it fix it? [09:56] agateau, that's it! Thank you very much! [09:57] tkamppeter: that's the diagnostic, not the fix :/ [09:57] agateau, with "LIBOVERLAY_SCROLLBAR=0" it works, without it segfaults. [09:57] agateau, much better than taking down the whole system ... [09:58] tkamppeter: I expected that, but this is bug #805303 which was supposed to be fixed, but seems to be more resilient than I expected [09:58] Launchpad bug 805303 in overlay-scrollbar "Gtk-CRITICAL **: IA__gtk_widget_style_get: assertion `GTK_IS_WIDGET (widget)' failed with the default qt4 gui" [Undecided,Confirmed] https://launchpad.net/bugs/805303 [09:58] agateau, so either the scrollbar story needs to get completely fixed for Oneiric or somehow hp-systray needs to get started with LIBOVERLAY_SCROLLBAR=0. [09:59] didrocks, ^ ^ [10:00] * agateau beats overlay scrollbars with a club [10:02] agateau, didrocks, bug 809009 seems to be the same ... [10:02] Launchpad bug 809009 in hplip "systray.py crashed with SIGSEGV in g_return_if_fail_warning()" [Medium,Confirmed] https://launchpad.net/bugs/809009 [10:02] tkamppeter: indeed, it is [10:02] agateau, I think that's what crashes metacity too, I get bug 848808 very easily with vlc [10:02] Launchpad bug 848808 in metacity "metacity crashed with SIGABRT in __kernel_vsyscall()" [High,Triaged] https://launchpad.net/bugs/848808 [10:03] jibel: yes, I am on that [10:03] agateau, ok, thanks [10:03] * agateau is surrounded by evil overlay scrollbars bitting his lovely Qt [10:03] jasoncwarner_: your second and third chart look pretty cool (seems on the first you took rather long to enter your password?); 12.8 seconds with interactive login and dropbox isn't bad at all [10:03] pitti: yeah, the first one I think I left it sitting there for a bit as I made tea ;) [10:04] jasoncwarner_: so, 10 vs. 12 seconds is a matter of pride, but 30 vs. 12 a matter of usability [10:04] so this is really good now [10:04] jasoncwarner_, pitti: where are those? [10:04] seb128: I sent them to pitti via email. [10:05] jasoncwarner_: mind if I put it on people? [10:05] where can I upload for public consumption? [10:05] pitti: no prob [10:05] please do [10:05] Ok. I've got one review for https://code.launchpad.net/~raof/gnome-desktop/workaround-bug-824099/+merge/77097 ; if no one complains, I'll merge it and upload gnome-desktop3 now. [10:05] didrocks: for a python program, g_get_prgname() will return "python", right? [10:05] seb128: http://people.canonical.com/~pitti/tmp/xdub-oneiric-20110928-3.png [10:05] If so I am afraid we can't blacklist hp-systray from liboverlay-scrollbar [10:06] (assuming blacklisting all python program is not an option :)) [10:06] RAOF, please do [10:06] RAOF, we know where to find you if it breaks ;-) [10:06] RAOF, oh, and local rules imply that breakage must be compensated by beers at the next UDS [10:06] agateau: It should be possible for hp-systray to set it's program name so you can blacklist it; does it do so? [10:06] RAOF, just as you know ;-) [10:07] seb128: And if it works flawlessly? :) [10:07] RAOF, you are the one getting the beers ;-) [10:07] RAOF, as long as there are beers it's fine ;-) [10:09] RAOF: good point, it does not set its program name, I am looking at a more generic fix, but I'll investigate this solution if it doesn't work [10:10] argh "you're not channel operator" [10:11] pitti, ^ === pitti changed the topic of #ubuntu-desktop to: Welcome to #ubuntu-desktop | https://wiki.ubuntu.com/DesktopTeam | If you want to help out, check out http://pad.ubuntu.com/ubuntu-desktop | For support please join #ubuntu [10:12] seb128: cheers; I'm an IRC n00b === pitti changed the topic of #ubuntu-desktop to: Welcome to the Desktop Team | https://wiki.ubuntu.com/DesktopTeam | If you want to help out, check out http://pad.ubuntu.com/ubuntu-desktop | For support please join #ubuntu [10:12] (no reason to state the channel name again) [10:12] pitti, dunno if you are a channel admin, I can use /msg chanserv op #ubuntu-desktop [10:12] when authentificated [10:13] pitti, you can try if it works for you [10:13] how do I un-op? [10:13] pitti, /unop [10:13] I think [10:13] ah, /deop [10:13] agateau: I guess so, never tried [10:13] ok, I was going to suggest that if unop doesn't work ;-) [10:14] seb128: no, I'm not authorized to op [10:14] seb128: anyway, thanks! pad copied [10:14] agateau: hp-systray can export any needed variable itself [10:14] I would go that way TBH [10:14] pitti, thanks for sorting the pads ;-) [10:15] didrocks: it seems the variable trick does not work here, otherwise we wouldn't have the problem [10:15] oh [10:15] didrocks: it does not work anyway for vlc, so an alternative is needed [10:15] pitti, op> ok, dunno who made me op and when, just ping me if you need to be oped [10:16] agateau, didrocks, also note that hp-systray forks into three threads. Only one is the GUI. [10:16] tkamppeter: ok [10:17] didrocks: I am considering disabling the init_once part of the overlay scrollbar init code. There is some generic code in liboverlay-scrollbar to disable itself when run within a Qt application, but I think this code is call too early to detect Qt [10:18] agateau: should I be prepared to a new sni-qt then? [10:18] or just overlayscrollbar? [10:18] didrocks: no, maybe a new... gtk patch :/ [10:18] ok, see that with seb128 :) [10:18] (assuming there is a gtk patch to put overlay scrollbars in) [10:18] Qt, gtk… in 2 days! :) [10:18] heh [10:18] pitti: have you engaged upstream pygobject with https://bugs.launchpad.net/unity/+bug/855402 - or have anything to report? Otherwise I am going all in on it now. [10:18] Ubuntu bug 855402 in unity "Python lenses segfault with pygobject3" [Critical,In progress] [10:18] I am getting warm [10:19] kamstrup: sorry, didn't get to this one yet [10:19] kamstrup: so if you want to discuss with them now, please go ahead [10:19] pitti: ok, i'll engage [10:20] agateau, you are looking to make friends there ;-) [10:20] agateau, did you talk to Cimi about the issue? [10:20] seb128: indeed :) [10:20] seb128: we talked when I first discovered the bug, that's when he told me there was code to detect Qt [10:21] seb128: but for some reason this code did not work so I ended up patching Qt with a crude workaround. Which does not work for some apps like vlc... so I am now considering fixing the Qt detection code from overlay scrollbars [10:21] but the fix is going to be in the gtk overlay scrollbar patch [10:23] seb128: do you know why we test for overlay scrollbar 3 times at init? [10:25] agateau, no, I never looked at the overlay scrollbar code ;-) [10:25] seb128: I mean in the gtk patch [10:25] didrocks, smspillaz: oh noes! http://people.canonical.com/~pitti/tmp/unity-layering.png [10:26] no, I reviewed the patch when he landed but it's mostly Cimi and loicm's work [10:26] smspillaz, didrocks: is there any state I should capture? it consistently happens right now when I press windows [10:26] (or move mouse to the corner) [10:27] pitti: it's known, the new compiz I'm testing should fix the remaining issues [10:27] moreover, launcher doesn't go away by itself [10:27] didrocks: oh, good! [10:27] thanks [10:30] pitti: which applications are you running [10:31] pitti: note that, there is a case now where there are some windows which *should* go on top of the launcher and panels, but they have to have a special hint set to do so [10:31] (so namely, just onboard) [10:31] smspillaz: on this desktop, four gnome-terminal windows [10:31] smspillaz: (are there other apps? :) ) [10:31] hm ok [10:32] smspillaz: desktop 2: firefox, desktop 3: terminal with mutt/empathy, gtimelog, desktop 4: nothing [10:32] at least the case I've fixed is with certain applications that misbehave, but there could be other cases I don't know about [10:32] oh ok, firefox misbehaves [10:32] i keep hearing that, but i've never seen this issue in mutter or metacity [10:32] (doesn't matter which desktop they are on, the window stack is global [10:32] (with firefox) [10:33] smspillaz: I just closed firefox, still teh same problem [10:33] chrisccoulson: that's because metacity handles some of the crazy things firefox does [10:33] uh chrisccoulson? aren't you supposed to be on holidays? [10:33] pitti: the issue will remain even if you close the applications [10:33] pitti: basically, it's once you make one mistake its game over [10:33] and the fun part is [10:33] pitti - yeah, i'm supposed to be ;) [10:33] often the main mistake is invisible [10:34] and then it compounds itself [10:36] chrisccoulson, hey ;-) [10:59] seb128: FYI, new ubuntu-mono just built, which should avoid reading the bg image 6 times [10:59] great [11:00] hey seb128 === MacSlow is now known as MacSlow|lunch [11:36] bah, robert_ancell! [11:36] he didn't fix the "unity-greeter g-s-d is still running" issue [11:41] he didn't move the conffile as well :) I just pushed a fix for this [11:41] didrocks, ok, I was just commented on bug #861316 and starting to fix it [11:41] Launchpad bug 861316 in lightdm "Uncleaned conffile in lightdm-gtk-greeter" [High,Confirmed] https://launchpad.net/bugs/861316 [11:41] seb128: done then :p [11:41] didrocks, well he did apply the patch that was provided, it's just that slangasek did an upload in between and nobody updated the version [11:42] seb128: yeah, I noticed that. I explained why the change in theh changelog [11:44] if someone has a g-c-c update to do, please don't upload, I've got a fix in bzr and another coming [11:46] rodrigo_, it's all yours [11:47] :) [11:47] so, hard code freeze is tomorrow, right? and at what time? [11:47] rodrigo_, I guess that will all the crazy requests you didn't have time to look at the "never" option for the screen? [11:48] rodrigo_, Final Freeze[2]: 2011/09/29 at 2100 UTC [11:48] seb128, hmm, no, got lost on my todo list, so including it for the next upload, if it's easy [11:49] rodrigo_, well evaluate it at least ;-) [11:49] chrisccoulson, pitti, others: did you check if the "suspend after 30min idle" bug is fixed? [11:50] yes, doing it now [11:51] rodrigo_, thanks [11:51] rodrigo_, don't forget to eat! ;-) [11:53] yes, I won't, I bought some nice stuff to cook! :) [11:54] seb128: I don't think it is; I just manually set it to "never" in control-center right before lunch [11:54] to avoid it suspending over lunch break [11:54] seb128: but I'll reset my dconf key now [11:55] pitti, do it via the g-c-c panel please [11:55] pitti, and let's check if it suspends [11:55] rodrigo_: how do I reset a key with control-center? [11:55] I thought once you set one, you'll keep that forever [11:55] pitti, set it to don't suspend on the panel [11:55] rodrigo_: right, I did right before lunch [11:55] ah, and it still suspended? [11:56] rodrigo_: I thought seb128 meant that the default was set to 30 mins [11:56] rodrigo_: no, it didn't [11:56] rodrigo_: it behaved right, just the default is still bad [11:56] pitti, yes, the default is 30mins, the panel wasn't taking into account the key to not suspend, just checking the timeout [11:56] and that's fixed in the last upload I did for g-c-c [11:56] ah, I didn't get that yet [11:56] right [11:57] uh, I just see that the default for "button-sleep" is "hibernate" [11:57] rodrigo_, pitti: should be change the default to not suspend on idle? [11:57] and "button-hibernate" is also "hibernate" [11:57] YAY! I only have one Ubuntu One icon in my system settings :) no, I'm going to bed ;) [11:57] jasoncwarner_, 'night [11:57] seb128: yeah, at least on AC; that's what we had in previous releases [11:58] seb128, it is already, just that the UI wasn't taking into account the key for it [11:58] it's reaally unexpected on AC [11:58] rodrigo_, it's not an ui thing [11:58] seb128, hmm, but it's not suspending, or is it? [11:58] rodrigo_, people there got their plugged laptop to suspend over lunch yesterday [11:58] so it was not an ui bug [11:58] ok, looking at g-s-d, maybe it doesn't read the key neither [11:58] but g-s-d which was suspending the box [11:58] this is sleep-inactive-ac-timeout [11:59] or maybe they opened the panel and g-c-c set the key to a buggy value due to the bug you fixed? [11:59] ok, I reset all plugins.power keys to the defaults now [12:00] it defaults to suspend after 30 minutes when idle and on AC [12:00] (in both dconf and c-c) [12:00] pitti, the boolean key is to false by default there [12:00] but both yesterday as well as today the actions were correct according to the settings [12:00] i.e sleep-inactive-ac [12:00] yesterday it said "suspend after 30" and it did; today I set it to "never" in c-c, and it didn't suspend [12:00] ok, so it respect the actions [12:01] so now to figure if the default behaviour is 30min or never [12:01] seb128: oh, there is a timeout _and_ a boolean? it set the timeout to "0" when I said "never" [12:01] * pitti watches the dconf keys [12:01] org.gnome.settings-daemon.plugins.power sleep-inactive-ac false [12:01] org.gnome.settings-daemon.plugins.power sleep-inactive-ac-timeout 1800 [12:01] ^ default [12:01] right [12:01] org.gnome.settings-daemon.plugins.power sleep-inactive-ac false [12:01] org.gnome.settings-daemon.plugins.power sleep-inactive-ac-timeout 0 [12:01] ^ set to "never" [12:01] not sure how to boolean is used [12:01] org.gnome.settings-daemon.plugins.power sleep-inactive-ac true [12:01] org.gnome.settings-daemon.plugins.power sleep-inactive-ac-timeout 1800 [12:01] ^ set to "30 mins" [12:02] it's confusing to have both a boolean and a value rather than using 0 as never [12:02] *nod* [12:03] yeah [12:03] Pici, updated with the in-progress aptch [12:03] Pici, sorry [12:03] pitti, ^ updated the uife with the in-progress patch [12:04] njpatel: which one? (sorry, I've looked at dozens recently) [12:04] https://bugs.launchpad.net/unity/+bug/732016 [12:04] Ubuntu bug 732016 in unity-2d "UIFe: Desktop should be titled" [Low,Confirmed] [12:04] yes, it sets it to 0 when don't suspend is selected, so g-s-d checks the timeout, and if it's 0, it doesn't suspend [12:04] so I think it should be fixed with my last upload [12:05] rodrigo_, pitti: ok, so in a guest session with today's g-c-c update the combo are "30 min" "never" [12:05] njpatel: ah, so many domains to choose from :) [12:05] indeed :) [12:05] njpatel: (and they all have different translations, too!) [12:05] woops [12:05] rodrigo_, pitti: do we want to suspend on idle on battery? we got some complain about it [12:05] and the patch removes a null check so far, ignore that, have already added that back [12:05] seb128, hmm, I think it's a sane default, isn't it? [12:06] njpatel: nautilus WFM, it is for nautilus after all [12:06] rodrigo_, well some people complained that it stopped their download, I don't have a strong opinion, let's keep it this way [12:06] having the battery drain will also stop their download :) [12:06] so yes, not 100% sure what's the best [12:06] njpatel: bug updated [12:07] thanks [12:07] rodrigo_, it's late, let's keep it this way for this cycle, we can sru a fix if there is a consensus that users don't like it for some valid reason [12:07] seb128: I have no strong opinion on that one, but I'd opt for consistency with previous releases here; I think we disabled autosuspend on battery [12:07] rodrigo_, thanks for all the fixes btw ;-) [12:07] seb128: but I'm happy to boot natty and triple-check [12:08] seb128, ok [12:08] pitti, I'm pretty sure we didn't ... do you think we should go back to that? [12:08] * pitti boots natty [12:08] seb128: if "not suspend" is consistent with both natty and AC, I think that'd be best [12:08] principle of least surprise [12:08] automatic suspends are generally quite unexpected IMHO [12:09] wow, natty's control-center was quite different [12:09] indeed ;-) [12:11] /apps/gnome-power-manager/timeout [12:11] sleep_computer_ac = 0 [12:11] sleep_comptuer_battery = 1800 [12:11] so it seems we did default to 30 seconds in natty when on battery [12:12] seconds -> minutes [12:12] but yeah [12:12] so let's keep it this way [12:12] erm, yes :) [12:14] meh, apparently I suck at bzr pushing [12:14] I should probably move to bound branches [12:17] ok, cooking/lunch time, bbl [12:18] rodrigo_, enjoy === Amaranthus is now known as Amaranth [12:23] mvo_, hey [12:23] mvo_, do you have a minute to chat about update-notifier, unity? ;-) [12:23] mvo_, did you have any success with porting to an indicator? [12:26] pitti, I guess it's late now to turn sound events off for Oneiric right? [12:26] pitti, i.e to avoid login sound [12:26] pitti, we should rather do it next cycle? [12:26] seb128: late, yes; but I'm not really unbiased there, so maybe let's ask in #u-release [12:27] and do an UIFE bug [12:28] pitti, I think we pushed enough late UIFe [12:28] we can live with another cycle with it [12:28] let's change that early next cycle [12:28] yeah [12:28] for the Long Term Silence release [12:28] and let's add a QA test case that it may never ever come back! [12:29] oh, disabling it by default? [12:29] yay [12:31] seb128: is someone on packaging the new indicator-session? [12:31] ken probably? [12:31] pitti, I was going to look at it soon [12:32] seb128: I have three bug fixes to work on, but they should all be fine for post-freeze [12:32] so looking for remaining package updates [12:32] can do pangomm in debian and sync [12:32] seb128: yes and yes [12:32] pitti, shared-mime-info maybe as well? [12:33] seb128: lp:~mvo/update-notifier/app-indicator-port - but its still wip [12:33] seb128: yeah, although that needs a merge [12:33] seb128: at it [12:33] mvo_, hey [12:33] seb128: but I think I got it mostly working, just need to do the dull and boring legwork now [12:33] mvo_, is that planned to land in oneiric? [12:33] seb128: but got interrupted a couple of times [12:33] rodrigo_: you are still working on glib-networking? [12:33] rodrigo_: (you have the grab for a day or two now) [12:33] seb128: well, its targeted, I'm not sure how wise it is at this point anymore though [12:34] mvo_, ok, so my issue is that update-notifier in its default config uses the systray for "broken apt state" [12:34] mvo_, I just got it while doing a dpkg -i on a deb without having the corresponding ddeb [12:34] mvo_, added to bug #856125 it's a bit of an issue [12:34] Launchpad bug 856125 in unity "update-notifier systray icons showed and the wrong place and size" [Medium,Triaged] https://launchpad.net/bugs/856125 [12:34] seb128: but I think it can be done today, I would feel better if there was a way to ensure I use app-indicator *only* if there is a indicator area to minimize risk [12:35] mvo_, though I guess most users don't run in a "broken apt state" often [12:35] seb128: yeah [12:35] heh :) i have the misplaced icon too [12:35] mvo_, not sure about that, maybe check with ted [12:35] sometime the sizes are all funny too [12:35] the idea is that libappindicator will fallback for you [12:36] so not sure they handle the case where you don't want it ;-) [12:36] right, I just don't want to take chances on a system like e.g. xubuntu [12:36] I would hate to break it for them last minute [12:37] indeed [12:42] mvo_, I will not be popular with slangasek but I'm close to suggest we drop the whitelist again and stay on what we had in natty for oneiric :p [12:42] seb128: no, I want touch that ever ever ever again! :) [12:43] slangasek had some security concerns in fact [12:43] yeah, I know, I discussed for a while with him before he opened his bug [12:43] and 15 minutes to find the wmclass typo, wasn't fun! [12:43] :-) [12:43] he said that it basically lead people who turned of the auto launch key in old version to not know about updates [12:43] aha [12:43] [12:44] which is a valid point... [12:44] seb128: well, we have a big issue in natty with unity then, people never upgraded then! :) [12:44] anyway, I will just state it's all mvo_'s fault :-) [12:44] eh, why is dpms enabled on my screen ? it turns it off after 10min [12:44] didrocks, well the people who tweaked a gconf key should be able to figure how to update :-) [12:45] seb128: exactly, that was my point during the discussion :) [12:45] xset q tells m ethat at least, is that new? [12:45] mvo_, because that's what you configured in g-c-c? [12:45] mvo_, screen capplet [12:45] what is the "turn the screen off after" value? [12:45] mvo_, next you are going to complain that GNOME dropped the "never" option (design decision)? ;-) [12:46] seb128: that is set to 1h [12:46] seb128: but when I change set, xset q reports the same values [12:46] mvo_, do you use a docked laptop? [12:46] no [12:46] seb128: pangomm uploaded to Debian, but the only visible change is an extra .m4 file installed, which we apparently don't need; I think it's not worth the trouble for oneiric [12:46] a laptop, but no dock [12:46] mvo_, ok, I get a similar issue on my laptop but not my desktop [12:47] pitti, ok, let's claim that we are done with updates and focus on bugs? [12:47] slangasek: ping? [12:47] seb128: haven't looked at s-mime-info yet [12:47] seb128: I have xset q reporting on my desktop that its set to 600 too [12:47] pitti, oh there was a gnome-menus new tarball to work with current pygobject today, did somebody package that? [12:47] I don't think so [12:47] seb128: I'll have a look [12:47] mvo_, could you open an upstream bug on bugzilla? or ask on #control-center on irc.gnome.org? [12:48] mvo_, I've a similar issue on my latitude but not on the 10v [12:48] mvo_, but I didn't have time to debug or bring it upstream yet [12:48] pitti, thanks [12:49] has anyone noticed icons not being displayed in the history menu in firefox until opening the menu a second time? [12:52] seb128: oh, that's just for the menu editor; we don't even ship that ATM [12:54] seb128: pushed to bzr, but not worth uploading IMHO === MacSlow|lunch is now known as MacSlow [12:55] taking gnome-themes-standard, for the new wallpaper (upstream gnome experience) [12:56] seb128: we should still get glib-networking to 3.0 at least; waiting for rodrigo_ to respond [12:57] -standard> heh, no; 3.2.0.2 just reverts 3.2.0.1 [12:57] pitti, oh, we stopped shipping the editor? [12:58] needs to be packaged separately [12:59] (separate binary, that is) === m_conley_away is now known as m_conley [13:00] pitti, oh ok [13:04] nice, I think it's the first time in years that screen dim and goes back to the value you were using when you use the mouse and keyboard again rather than going back to the default [13:06] seb128: shared-mime-info is rather intrusive; I think we should postpone that to P [13:06] pitti, wfm [13:06] ok, so glib-networking, and c'est la oneiric? [13:07] pitti, seems so! [13:07] nice [13:07] pitti, I will wait a bit still for ken to show up before starting on indicator-session since ted sent a merge request his way yesterday and I'm not sure if he started on it yet [13:09] pitti, can you review https://code.launchpad.net/~indicator-applet-developers/indicator-session/ubuntu/+merge/77243 to see if the diff seems fine to you? [13:10] pitti, there is a string change but I discussed it with ted and he sed-ed the po files to update the msgid and keep the translations [13:10] pitti, the string is the same, just some case changes [13:10] i.e "Restart to complete updates" -> "Restart to Complete Updates" [13:11] so the translations should still be valid [13:11] oh, speaking of kenvandine [13:11] kenvandine, hey [13:11] hey [13:12] seb128: oh, "Switch Off..." -> "Shut Down…" ? [13:12] kenvandine, will you do the indicator-session update or do you want me to look at it? [13:12] that looks rather backwards to me [13:12] hey kenvandine [13:12] i will [13:12] string change? [13:12] pitti, where? [13:13] pitti, I think it's just noise, the po files where outdated, ted did a new export to be able to sed the msgid and keep translations [13:13] where->were [13:13] seb128: if it was properly sed'ed, fine for me; it's mixed with translation updates, so it's very hard to read [13:13] ah, whew :) [13:16] pitti, seems to be correctly seded to me, I've done some checking [13:23] mvo_, so, don't be shy :-) [13:23] seb128: Hi Sebastien, can you possibly help with a lightdm question, or suggest somebody (besides Robert, who is sleeping) who might know? [13:24] hey GunnarHj [13:24] GunnarHj, better to just ask your question [13:24] then we can reply or redirect you ;-) [13:24] seb128: Ok. I'd like to understand this item in the lightdm changelog: "Set LANG variable based on the user language". I'm wondering because LANG is currently designated for regional formats in Ubuntu, not display language. [13:24] Or is it only applicable if you enable the language chooser? [13:25] http://bazaar.launchpad.net/+branch/lightdm/revision/1181 [13:26] seb128: I did not enter into any confusion when I tested, but I'd like to be sure. [13:26] pitti: Hi Martin, will take a look. [13:26] what pitti said [13:26] locales are too hard :-) [13:27] NB I just looked up the patch, but I didn't read it yet (busy with something else) [13:27] sigh, mono [13:28] how do i make mono not totally screw up a backtrace from within C/C++ code? [13:29] Locales are hard only because there are too many ideas on how to deal with them. [13:29] Excellent area for some kind of standard. [13:33] seb128, pitti: Found what I needed in a comment: "In the case of Ubuntu these will be overridden by setting these variables in ~/.profile". In other words, it happens before lightdm-session is sourced. [13:33] Thanks for the help! [13:34] yw [13:57] chrisccoulson, hi [13:57] rickspencer3, hey [13:57] rickspencer3, he's on holidays this week and next one [13:57] I'm with someone who is having a weird effect where firefox doesn't seem to have window decorations [13:57] hi seb128 [13:58] rickspencer3, hum [13:58] seb128, have you seen this at all? [13:58] rickspencer3, is that maximized firefox and the space is between the unity panel and firefox? [13:58] seb128, it opens maxmized, they unmaximize it, no window decoratos [13:59] rickspencer3, not seen that but it seems a compiz bug [13:59] the decorations are purely compiz, no application side [13:59] seb128, the weird thing is it is only Firefox [13:59] and also weird, it's not happening to me [13:59] is it happening in a constant way? [13:59] or is that a random bug? [13:59] agateau, bug 858484 looks like a dupe of bug 861125. [13:59] Launchpad bug 858484 in hplip "systray.py crashed with SIGSEGV in g_type_module_use()" [High,Confirmed] https://launchpad.net/bugs/858484 [13:59] seb128, constant [13:59] Launchpad bug 861125 in hplip "systray.py crashed with SIGSEGV in gtk_widget_style_get()" [Medium,Confirmed] https://launchpad.net/bugs/861125 [14:00] tkamppeter: indeed [14:01] rickspencer3, bug #861143 [14:01] ? [14:01] Launchpad bug 861143 in unity "unity 3d firefox window borders disappear" [Undecided,New] https://launchpad.net/bugs/861143 [14:03] rickspencer3, well in any case check with smspillaz but my guess is that firefox does something compiz doesn't like, still a compiz bug [14:03] yeah [14:03] thanks seb128 [14:03] rickspencer3, yw [14:03] they are "me tooing it" [14:03] ok, great [14:04] pitti, oh, sorry about glib-networking, I just moved to other stuff, so if you want to do it, please go ahead [14:04] rodrigo_: okay [14:08] didrocks: seb128: I think you're going to "love" this: I have a clean fix for bug 805303 which consists of two patches, one for Qt and one for GTK... I guess I am going to keep the patches for an SRU? [14:08] Launchpad bug 805303 in overlay-scrollbar "Gtk-CRITICAL **: IA__gtk_widget_style_get: assertion `GTK_IS_WIDGET (widget)' failed with the default qt4 gui" [Undecided,Confirmed] https://launchpad.net/bugs/805303 [14:08] agateau, gtk2 or gtk3 or each? ;-) [14:09] seb128: ah no, I was lazy, it only affects gtk2 :) [14:09] agateau: you are very lucky, I have an incoming Qt upload schedule! [14:09] agateau: twice in 2 days that you are so lucky and avoid being killed, that's weird :p [14:10] * agateau makes a note not to try this too often [14:10] agateau, do they need to go in together? [14:10] seb128: to fix the bug, yes: but they won't break builds if they are not applied together [14:11] agateau: is that an update of your sni-qt patch? [14:11] agateau, ok, try convincing kenvandine to take the gtk2 patch then I guess ;-) [14:11] agateau: can I make a rant in the changelog as well? :) [14:11] agateau, open a bug with the patches (or use an open one) anyway [14:11] didrocks: no, it is overlay-scrollbar related [14:12] didrocks: you are free to rant, no problem :) [14:12] \o/ [14:12] seb128: ok [14:13] * agateau uploads patches [14:13] agateau, send me the patch in a bug report and i'll look [14:13] kenvandine: ok thanks [14:13] building Qt, story of my life, lalala :-) [14:14] I use over 50% of my CPU time building unity in an average day :P [14:14] agateau, didrocks, kenvandine, seb128: I vote for letting all fixes regarding overlay scrollbar directly into Oneiric and not only as SRU. It is too annoying. [14:14] DBO, good, you have margin to do twice the work you are doing ;-) [14:15] kenvandine: on the gtk2 side, it is actually going to be an update of the overlay scrollbar patch [14:15] seb128, I [14:15] ... [14:15] DBO, should I run away from you when we cross at next UDS? ;-) [14:15] pfft, if [14:16] I suspect they will lock me in a small room and tell me to code [14:16] I will not complain at people coding [14:16] ;-) [14:16] DBO: and the air conditioning will freeze you to death :/ [14:17] oh right, orlando, air conditioning… [14:17] bring your snow coat! [14:18] rodrigo_, oh btw did you see that diwic commented on the g-s-d segfault assigned to you with a tentative explanation and patch prototype for the issue? [14:18] seb128, yes, answerting him now [14:18] seb128, it's just a fix for the sympto, not the real problem [14:22] rodrigo_, ok [14:22] rodrigo_, he's diwic on IRC if you want to reply there [14:22] yeah, I know him :) [14:23] GunnarHj: yeah, at this point the LC_* stuff is P cycle matter indeed; I think what we have right now is working mostly, so let's not tear it apart so close to release; ok? [14:23] rodrigo_, ok, great [14:25] man it takes forever to get a stack trace from banshee with all the dbg packages installed [14:27] pitti: Do you mean not unsetting them for lightdm users? Well, I thought that you were more concerned over the sent env. variables via ssh than I am, so if you prefer to keeping it as is, I have no problem with it. [14:28] GunnarHj: it's not ideal, but at least not a regression from natty [14:28] and implementing this properly (either with your or with my proposal) is far from trivial as far as I can see, and it seems that whenever this stuff is touched it breaks something else [14:28] so let's rather keep the known bugs for now :) (which don't hurt too much) [14:29] pitti: Ok. [14:30] pitti: Then there is only the l-s MP left from me. I suppose you'll upload it before final freeze? [14:31] GunnarHj: the language-selector fix for the signal handler, etc., right? [14:31] https://code.launchpad.net/~gunnarhj/language-selector/oneiric/+merge/77109 [14:31] GunnarHj: yes, that's still on my radard [14:31] I got the chinese CD fixed, finally debugged jockey, doing glib-networking now, and then this [14:33] pitti: I realize that you are loaded these days. ;-) [14:33] if people stop sending UIFEs and FFEs to review, we might actually get this thing done :) [14:33] GunnarHj: c'est la release time ;) === AfC is now known as AfC|zzz [14:42] is there an easy way to get list of packages which were uploaded within a specific date range in ubuntu? [14:43] nevermind [14:46] dobey: the oneiric-changes@ archives should be a close approximation [14:46] dobey: (it shows when a package gets accepted, but usually they don't stay in unapproved for that long) [14:48] GunnarHj: pong [14:48] pitti: i think i just found the culprit [14:51] libwebkitgtk-1.0-0-dbg is insanely huge though [14:52] anyone here with the excessively slow blur perf? [14:53] DBO: everything seems excessively slow on my laptop with oneiric; seems like it is a fan of disk i/o [14:53] dobey, I am looking for particular issues when something is performing an active blur [14:59] pedro_: ok. the banshee crashing in u1ms on scroll is DEFINITELY webkit :) [15:00] dobey, lets reassign that then , thanks for looking at it :-) [15:01] seb128, mvo_: as far as security concerns, if we really can't have libappindicator support this cycle (which I think we ought to still push for), we could simply have update-notifier ignore the setting and always pop the window up? I just don't want us to have users losing all their update notifications [15:01] pedro_: just did. i will do better than that even :) [15:01] slangasek: I give it another go after the meeting, I think I hvae the foundations ready [15:02] mvo_: ok [15:02] err, launchpad hates me :( [15:03] hate it back ! [15:03] dobey, it hates everybody :-P [15:04] it likes to give me timeout errors when i try to relocate bugs in Ubuntu === ara is now known as Guest78845 [15:09] i don't know why amz store isn't also crashing though. it's quite odd [15:10] GunnarHj: still here? [15:10] pitti: Sure. [15:10] + if 'kde-' in sys.argv[0]: [15:10] GunnarHj: that's mean to catch kde-language-selector, right? [15:11] pitti: Yes. [15:11] GunnarHj: does that actually work also when being called from the control-center? is it a separate process? [15:12] pitti: Don't know. But since l-s isn't integrated in g-c-c, I suppose it's a separate process. [15:12] pitti: Or is there a better way to distinguish between KDE and others? [15:14] didrocks: kenvandine: patches uploaded here https://bugs.launchpad.net/ubuntu/+source/qt4-x11/+bug/805303 [15:14] Ubuntu bug 805303 in overlay-scrollbar "Gtk-CRITICAL **: IA__gtk_widget_style_get: assertion `GTK_IS_WIDGET (widget)' failed with the default qt4 gui" [Undecided,Confirmed] [15:14] GunnarHj: just asking because kde-language-selector doesn't have a main method, so it looks like systemsettings would just use it as a library [15:15] GunnarHj: oh, how about we revert this? [15:15] agateau, thx [15:15] GunnarHj: not revert, I mean "if 'gnome-' not in sys.argv[0]" [15:15] agateau: thanks, prepping the Qt part [15:15] pitti: sure, do you want me to make that change? [15:15] GunnarHj: I can do it during merging [15:16] pitti: Ok [15:16] mdeslaur, ping [15:16] kenvandine: the gtk side could be more elegant: the new code should move to its own .h/.c file, but deadline is approaching [15:16] rodrigo_: yes? [15:16] GunnarHj: btw, it was glade which removed the relief property; it now removes properties which have default values, so it didn't actually change something [15:16] but having it doesn't matter, that's fine [15:16] GunnarHj: thanks! [15:17] mdeslaur, seems you forgot to add 49_fix_suspend_media_key.patch to g-s-d's bzr branch [15:17] mdeslaur, can you please add it? or send me the file and I'll add it myself [15:17] ? [15:17] rodrigo_: argh, hold on a sec [15:17] rodrigo_, hi, could you make libgoa-dev depend on gir1.2-goa-1.0? [15:17] pitti: adding back relief did make a difference. Without it you did only see the label except when hovering. [15:18] ricotz, yes, sure [15:18] rodrigo_, thanks [15:18] GunnarHj: oh, interesting; sounds like a glade bug then [15:21] rodrigo_: pushed [15:21] rodrigo_: sorry about that [15:21] mdeslaur, thanks! [15:24] agateau: wooo! introspection in C++! :) [15:25] agateau, your gtk patch doesn't include any of the changes to the .h files [15:26] so it drops them [15:26] kenvandine: the previous version of the patch used to include changes to the .h, but the new version does not [15:26] ok [15:27] didrocks: Qt actually does have decent introspection, but nothing beats dlopen :) [15:27] resolve() ? === mvo_ is now known as mvo [15:28] good night everyone! time for having a break and some sports, my head is exploding [15:31] 'night pitti [15:35] have a good night pitti! [15:35] see ya pitti, goodnight to you! === m_conley is now known as m_conley_away [15:56] desrt, ping [15:58] rodrigo_: hey [15:58] hey desrt, a quick gdbus question [15:58] desrt, if I call g_bus_get with a GCancellable and then, before the callback is called, I call g_cancellable_cancel, g_bus_get_finish will return NULL? [15:58] rodrigo_: and error will be set [15:59] desrt, ok cool, you rock ! :) [15:59] rodrigo_: in the case the callback has already been dispatched, there's no promise, though [15:59] rodrigo_: ie: sent to the mainloop for an idle dispatch [15:59] the call may still succeed in that case [16:00] desrt, hmm, so would you recommend adding a check for the GCancellable in the callback, just in case that happens? [16:00] rodrigo_: if it's important to you, yes [16:01] GSimpleAsyncResult has GCancellable integration [16:01] but it's probably not universally used [16:01] desrt, well, I want to avoid getting a correct bus connection if other members of the user_data being passed to the callback have been freed [16:01] rodrigo_: i see [16:02] so, I really need to make sure when I get the connection everything's ok [16:02] the members of the user_data are freed only after the cancellable is cancelled [16:02] so if the cancellable is NULL or is_cancelled, they should not be freed [16:03] s/freed/used [16:03] makes sense. i'd double-check. [16:03] ok [16:03] poke alex about what his intentions were here [16:04] the exact semantics of how the 'ideal' async call in GIO is done was never fully established [16:04] ok [16:04] we have a lot of expectations about what is right, but nobody ever wrote it down [16:10] ok, sport time! [16:10] why doesn't midori crash? :( [16:11] Maybe the new webkit fixed it? [16:12] no, the new webkit introduces a crash; but it seems to only happen when scrolling in the u1 music store in banshee [16:13] it doesn't crash in midori dong the same thing :-/ [16:20] out for a bit, bbl [16:22] i wonder if it's possible to do anything useful with webkit from vala yet; or with gobject-introspection. stuff failing to compile because of it, is lame [16:29] * cyphermox -> lunch [16:48] good night everyone [16:54] is there a way i can make apport file a bug by pointing it at a core file? [17:05] i guess that is a question for pitti [17:15] after today's update, i cannot launch byobu from the unity launcher :-( [17:15] late cycle regression :-( [17:51] pitti: ping [17:52] dobey: with some work (not all that much), you could create your work .crash file to pass the core file to apport and file a bug [17:52] work = own [17:53] hrmm [17:53] eww :) [18:09] seb128: so my workstation apparently shut down. gnome is not doing something funny like auto shuting down after 1h or 30min of inactivity? there is something like this is the gnome-control-center iirc [18:10] mvo: does your computer have trouble suspending? [18:11] i just had the same problem [18:11] both my laptop and desktop boxes suspended while i was eating lunch [18:11] and neither are configured to suspend on inactivity [18:12] jbicha: its a desktop, I would be really susprised if that works [18:13] kenvandine: yeah, sounds similar, I'm not amused given that a upgrade test was running on the box :/ [18:13] that is now gone [18:13] bummer [18:13] my desktop didn't resume either, and it usually does suspend/resume fine [18:19] mvo, it's bug #860485 [18:19] Launchpad bug 860485 in gnome-settings-daemon "bad default setting: suspend after 30min when plugged in" [High,Confirmed] https://launchpad.net/bugs/860485 [18:19] well there is a bug which makes it suspend anyway when it shouldn't added to it [18:20] rodrigo_, said he would look at it [18:20] mvo, did you mention your other issue on their channel? [18:32] pedro_, ^ could you look if you get those issues and try to upstream those? [18:32] cyphermox: or i can't; "This problem report is damaged and cannot be processed." [18:33] dobey: uh, something must be missing [18:33] seb128, sure , will have a look [18:33] you got ProblemType, Package, CoreDumpGz, etc? following the apport spec? [18:33] pedro_, basically laptop or desktop that suspend after half an hour idle when the option is set to never and screen dpms (xset q) that stay on 600 rather than respecting the setting [18:34] cyphermox: unfortunately it tells me not what is missing [18:34] yeah I know :/ [18:34] pedro_, several people mentioned it so I think we should upstream those early and try to push to get them fixed [18:34] pedro_, thanks! [18:34] no Package, but other stuff i have [18:34] kenvandine, mvo: ^ [18:34] no ProcFoo stuff though [18:34] likely upstream will want a gnome-settings-daemon --debug log for those with our version, mention also that we backport gnome-3-2 git fixes [18:34] cyphermox: i know nothing of an apport spec; where is one? [18:35] dobey: hold on I'm trying to find the link again [18:36] google is not helpful; it just gives me RPM stuff searching for "apport spec" [18:37] dobey: https://wiki.ubuntu.com/Apport?action=AttachFile&do=view&target=data-format.pdf [18:37] I've been able to create crash reports following this before [18:38] dobey: the gotcha is just the format for the base64 block of the coredump, but you can see how the python code in apport itself does this [18:48] cyphermox: it doesn't help that the core dump was 145M either [18:49] thanks a lot webkit [18:49] why didn't that get caught by apport anyway? [18:50] becasue i was running a test program under gdb i guess? this isn't from a normally installed process [18:50] hrmm; gedit really does not like editing a 25M text file [18:51] where 1 of the lines is a 25M compressed/base64 encoded core dump [18:52] of course, vim isn't happy either [18:52] and apport still being useless as ever with the "need more than one value to unpack" error [18:56] oh, the Proc* junk is required [19:00] hrmm [19:03] dobey: Proc*? [19:03] dobey: this is very roughly what you want: http://paste.ubuntu.com/698661/ [19:06] cyphermox: ProcMaps: etc [19:09] well webkit 1.6 did release on time after all [19:11] what does that mean? [19:11] it will be in oneiric? [19:12] no, I think the decision was to just put it in the PPA since it's feared that some stuff like ubiquity hasn't been tested at all with 1.6 === m_conley_away is now known as m_conley [19:13] dobey: forget the last part, the coredump bit doesn't work properly in my paste, but anyway, apport has a few functions to help you add the infor [19:13] there was a misunderstanding where we didn't think WebKit would release in time for Oneiric (no published schedule) [19:14] could I get a sponsor for https://code.launchpad.net/~ubuntu-core-doc/ubuntu-docs/oneiric [19:14] oh [19:14] i wonder if 1.6 fixes this crash :-/ [20:18] my Alt-tab Show Desktop icon background is black now === m_conley is now known as m_conley_away === Shadows`sleep is now known as SoulShadow [21:24] got this when trying to run Unity on my CR-48: http://paste.ubuntu.com/698727/ it worked just fine yesterday with Unity 3D [22:58] hells to the yeah desktop peoples [22:58] onboard + unity dash works now [23:06] DBO: ROckin! [23:06] yeah [23:06] who needs a keyboard [23:07] I'll click my way to victory [23:16] That'll make cnd's galaxy tab running Unity much more useful :) [23:18] hey RAOF saw the bug was fix committed...awesome,thanks! [23:18] Should in fact be fix released now. [23:18] oh? nice... [23:18] It got accepted last night; it shouldn't take _that_ long to build + publish ;) [23:18] Except maybe on armel :) [23:22] No, even on armel. Yay!