=== thumper is now known as thumper-afk [04:33] Good morning [07:20] morning folks [07:20] hey TheMuso [07:21] Hey willcooke. [07:50] good morning desktopers [07:56] morning seb128 [07:56] seb128, would you have time to help me with the what's new in 14.10 list today? [07:56] or tomorrow is fine [07:56] or just a short list via email [07:56] whatevs [07:56] I just don't want to be scratching around at the last minute on Monday night :) [07:57] I found the blueprints for 14.10 - but not a whole lot to go on there [07:57] willcooke, hey, I just replied to your email [07:57] oh [07:57] wow [07:57] thanks [07:57] willcooke, you are not going to like the reply much though [07:57] ha [07:57] reality is that we don't have much that changed enough to be marketing material [07:57] nah - that reply is good [07:58] that's what I thought would be the caase [07:58] *case [07:58] we basically iterated over what we have, some updates, bugfixing, polish [07:58] so now it's confirmed I can get on with it [07:58] lot of touch work [07:58] yeah [07:58] we did some updates like gtk, some GNOME components/apps [07:58] but nothing of that is significant enough/user visible enough for a marketing page [07:59] "continuing improving hidpi support" could be one item [08:06] hey [08:07] Laney, good morning, wie gehts? [08:08] hi Laney! [08:08] oh, I'm named weirdly [08:08] keep it, it makes you look important === larsu_ is now known as larsu [08:09] it's going well thanks, had a board games night last night [08:09] we didn't defeat wu feng, BUT! [08:09] realised that every time we played that game we've played this one rule wrong which makes it way harder than it should be [08:09] so next time he's going down [08:10] thx seb [08:11] what's up larsu et seb128 et willcooke? [08:11] it's all going on today [08:11] Laney, what game is that? [08:11] it's called Ghost Stories [08:11] and it's by a french guy! [08:11] haha [08:12] http://www.amazon.com/Asmodee-5511826-Ghost-Stories/dp/B001J77UUQ [08:13] that's it [08:13] it has good reviews [08:13] Laney: not much, feeling unimportant now that you mentioned it [08:13] Laney: I'm off for a short vacation at the baltic sea after today [08:13] yeah it's a fun game [08:14] larsu: why unimportant? :( [08:14] I saw there's a public holiday in .de tomorrow [08:14] Tag der Deutschen Einheit! [08:15] not important because he lost the tail after his nick? [08:15] Laney: "keep it, it makes you look important" [08:15] Laney: yep, gives us a nice long weekend :) [08:15] ah, forgot about that already :p [08:15] you're always important to me ♥ [08:15] oh, speaking about long w.e [08:15] Laney: :) [08:15] * seb128 needs to put a vac day on monday [08:15] ;) [08:15] lol [08:16] * willcooke prepares his REJECT button [08:16] hmm [08:16] * seb128 walks out and let Laney and larsu together [08:16] I've got some bug in the last day [08:16] where my background doesn't load properly like 50% of the time [08:16] http://people.canonical.com/~laney/weird-things/what.png [08:16] I love that url... [08:16] also do I normally have those icons? [08:17] Laney: there's supposed to be a bg image? [08:17] hum [08:17] yes [08:17] I don't even know what that gradient is [08:17] Laney: text shadow on the icons is wrong [08:17] it's like nautilus was failing to read its config [08:17] ya [08:17] happens loads [08:17] Laney: you can set a gradient in bg settings [08:17] do yo have anything wrong in the logs? [08:17] Laney: in GNOME... [08:17] I was ignoring it yesterday because I was logging in all the time to fix that unity bug [08:17] larsu, in unity as well [08:18] seb128: ah, in the combo box [08:18] * larsu didn't see that at first [08:20] well, wrong bg + icons displayed when they should ... looks like to me that nautilus is not reading the correct gsettings config or something [08:20] running nautilus doesn't bring it back [08:20] well [08:20] oh [08:20] OH [08:21] is the file in org.gnome.desktop.background.picture-uri correct? [08:21] what is handling your desktop? [08:21] click on one of those icons [08:21] and do help->about [08:21] I bet it's one of the nautilus fork you installed [08:21] like the cinnamon one [08:21] or nemo [08:21] oh [08:21] does that happen? [08:21] yes [08:21] wtf [08:21] and the first to start wins the desktop [08:21] WTF! [08:21] I did install the MATE one [08:22] there you go [08:22] yeah that's totally running [08:22] this seems a bit shitty (how surprising) [08:22] and it's starting before nautilus half of the time [08:22] sigh [08:23] how do these things get started again? [08:23] /etc/xdf/autostart [08:23] xdg even [08:23] okay, I'll file a bug on caja to fix its file [08:24] to be OnlyShowIn? [08:24] yep [08:24] right, makes sense [08:24] caja is their name for the nautilus fork? [08:25] it is [08:25] yet another nautilus fork :/ [08:25] couldn't those guys work together on e.g nemo? [08:25] the old "can't we work together" argument never worked [08:26] seb128: when are we switching? :P [08:26] larsu, to which one? ;-) [08:26] we should try one every cycle! [08:27] midnight commander [08:27] seb128: lol! [08:27] Laney: ts ts [08:27] larsu, hum, theme weirdness that might be new to your recent changes/landing [08:28] larsu, in gedit, the "search" box (the one you get with ctrl-f) has its borders transparents, I don't think it did before [08:29] which is not specific to o-s for once [08:31] uh oh [08:32] seb128: white bg for me - I guess it's supposed to be @theme_bg? [08:32] shrug [08:32] I wonder why it's buggy for me :-( [08:32] I don't know, it used to be white for me as well [08:32] AH! [08:32] it's transparent for me as well [08:33] I wonder what changed that... [08:34] you just said it was white? [08:34] yes, because the text view is white as well [08:34] putting some text there showed me it's actually transparent [08:44] larsu, oh, I saw it easy because I oped a new document at the first line was selected/with a grey bg for me [08:44] I've also the 80 columns margin displayed [08:45] seb128: this is what I'm using for testing now :) [08:45] :-) [08:46] oh, that was easy [08:48] seb128: @theme_base_color is not defined in our themes (it's just called @base_color) [08:48] but some software use that name? [08:48] seb128: got copied in with the latest gedit-specific tweaks in apps/gedit.css [08:48] I'll MR after some more testing [08:49] ah, there's more @theme_* [08:49] larsu, danke [08:52] larsu, can you look at https://bugs.launchpad.net/ubuntu/+source/gnome-system-monitor/+bug/1376432 as well? [08:52] Ubuntu bug 1376432 in gnome-system-monitor (Ubuntu) "GNOME System Monitor Resources graphs show black borders" [Low,Triaged] [08:53] ya [08:53] danke [08:57] seb128: https://code.launchpad.net/~larsu/ubuntu-themes/fix-gedit/+merge/236830 [08:58] larsu, wfm, but if apps/other themes use e.g @theme_base_color, would it make sense to define that variable in our theme? [08:59] seb128: we've always called it @base_color in our theme. Not sure if we'd break anything if we went with Adwaita's definitions now [08:59] larsu, can't we add an alias? [08:59] just @theme_base_color pointing to @base_color [08:59] so both would work? [08:59] I'd rather not, there's quite a few of those [08:59] see colors.css [08:59] k [08:59] wfm in any case, thanks [08:59] sorry, gtk-main.css [09:00] seb128: in general we discourage applications from reusing variables from the theme [09:00] instead, they should just give their widgets the right classes [09:00] i.e, ".background" in this case [09:01] k [09:25] bah, GNOME bugzilla is slooow [09:25] ya :/ [09:26] hosted by RH not Canonical ;-) [09:53] pitti, hey [09:55] pitti, https://launchpad.net/ubuntu/+source/language-pack-touch-fr doesn't include indicator-sound.mo anymore, do you know why? [09:57] hey seb128 [10:00] seb128: there's indeed no trace of it, but it didn't disappear in the last upload [10:01] http://launchpadlibrarian.net/185102146/language-pack-touch-fr_1%3A14.10%2B20140909_1%3A14.10%2B20140916.diff.gz [10:01] it disappeared here [10:01] pitti, it's in the current rtm image [10:01] yes [10:01] pitti, said differently "please don't copy those updates to rtm" :-) [10:01] no, I stopped doing that, we now build RTM specifig langpacks [10:01] there should be an update todya [10:02] I wonder if that glib ever went in [10:02] Laney, speaking of glib, should be sync the stable version from Debian? [10:02] Laney, https://launchpad.net/ubuntu-rtm/+source/glib2.0 looks like it didn't :/ [10:03] yeah I was waiting because desrt said he was going to do a new stable without property deprecation warnings and removing the mutex abort [10:03] but that didn't come yet so maybe we should take this one [10:03] Laney, https://git.gnome.org/browse/glib/commit/?h=glib-2-42&id=b12bd1c3dcfbb398d2462dcf584a1f6d5173ca9a [10:03] we can at least backport that one if we want [10:04] there should be a release [10:04] k [10:04] let's see what he says later, can decide then [10:04] the mutex things is going to be removed? [10:04] just the abort I think [10:04] or asked differently, do we need to backport https://git.gnome.org/browse/gtk+/commit/?h=gtk-2-24&id=fbf38d16bcc26630f0f721d266509f5bc292f606 [10:04] the glib fix would have the same effect [10:04] desrt, ^ [10:04] k [10:05] if we are sure we are getting the glib change let's not backport the gtk one then [10:05] that gtk one is kind of a hack/workaround [10:06] right, but that's better than nothing [10:06] well if glib stop aborting even better [10:07] yes [10:07] seb128: so, I have not the slightest idea; the logs are already gone as we had two more updates after that [10:09] seb128: the next automatic update will fix it again [10:09] (I restored the files) [10:09] pitti, danke === thumper-afk is now known as thumper [10:23] so good having media keys back [10:23] I was actually getting used to scroll wheeling on the indicator [10:24] we didn't fix the direction of scroll wheel scrolling in u-c-c -> sound? [10:24] larsu looked at it but couldn't reproduce [10:25] seems to depend of the type of pointing device in use [10:25] didn't the indicator have it before? [10:25] because that's right for me now [10:26] yeah, not sure what's the difference :/ [10:26] it's buggy on most u-c-c sliders for me [10:26] like the keyboard repetitions, a11y, etc as well [10:27] seb128, Laney: gtk slider scroll direction is different from all sound sliders [10:27] same [10:27] afaik [10:27] you mean u-c-c overrides it to be the other way? [10:27] only for the sound slider, yes [10:27] well, seb128 just pointed out that other panels have this too [10:28] it's backwards for the touchpad on my laptop as well [10:28] * larsu is confused [10:28] two finger scrolling [10:29] indeed, two finger scrolling is the opposite direction [10:30] Laney: xev shows me that different buttons get activated... [10:30] ah, natural scrolling only affects two-finger scroll on the touchpad, not the scroll wheel one [10:31] Laney: if you turn natural scrolling off, is it consistent? [10:31] I don't have that setting on [10:32] me neither [10:32] still sound is the wrong way around [10:32] scroll up make it go lower [10:32] okay there are two issues here: [10:32] (1) the sound sliders react in the opposite way from all other sliders on the system [10:32] (2) when natural scrolling is turned on, two-finger scroll behaves in the opposite way from scroll wheel scroll [10:33] seb128: for me, it's the other way around. Which buttons does xev report for scroll up/down? [10:33] (3) the sound slider in u-c-c doesn't react the right way [10:33] seb128: this is (1), no? [10:33] no [10:33] for me it reacts the same way as the other sliders [10:34] the keyboard panel ones in u-c-c do the same [10:34] the a11y ones as well [10:34] they are all backwards in u-c-c for me [10:34] same here [10:34] scroll up xev button 4 [10:34] down button 5 [10:34] why do I see a difference then? [10:34] good question [10:34] I'm using an usb mouse [10:34] with a wheel [10:35] same for me, up 4, down 5 on the trackpoint and two-finger scroll [10:35] (without natural scrolling) [10:35] sound slider in u-c-c is up/forward, every other slider is up/backwards [10:36] k, for me sound is the same as every other slider [10:36] seb128: maybe you have the mate control center installed? [10:36] I'll look into it [10:36] * seb128 slaps larsu with a trout [10:36] haha [10:36] what's the gtk3 widget factory thing called? [10:36] gtk3-widget-factory [10:36] lol [10:37] oh yeah [10:37] wow [10:37] that has screwed up theming ....... [10:37] those behave the same as u-c-c sliders [10:37] up = move to the left [10:37] Laney: hm? [10:37] transparent background [10:37] Laney: apart from non-resizable window and no shadow, it should be fine [10:38] but indeed it is also backwards [10:38] Laney, bug #1285783 might be the same [10:38] bug 1285783 in gtk+3.0 (Ubuntu) "Right panel has a transparent background" [Undecided,Confirmed] https://launchpad.net/bugs/1285783 [10:38] there is a fix there we forgot to land :/ [10:38] want to test it? [10:39] kay [10:39] wfm [10:39] did he fwd it? [10:40] but I seem to be running a different os from you guys :/ [10:40] are you using the archive gtk? [10:40] yes [10:41] larsu, do you have all the sound panels doing right on up? [10:41] like the one from the effects tab as well? [10:41] http://people.canonical.com/~laney/weird-things/wowzers.png [10:41] Laney, https://git.gnome.org/browse/gtk+/commit/?id=02ccfb5d62430590b11b2d3e34b6100e4a6ddcda [10:41] seb128: yes, all the sound sliders go right on up [10:42] thanks, I'll try the upstream one [10:42] Laney, yeah, better, the upstream bug states that the downstream one had an issue [10:42] thanks [10:43] Laney: ah, o-s [10:43] larsu, k, I don't understand what's different between our setups :/ [10:43] * larsu reproduces [10:43] we should just drop those [10:43] * larsu runs [10:43] (again) [10:43] 15.04 man [10:44] I need a bigger SSD so I don't have to delete checkouts all the time [10:45] or git could grow a "give me this commit from this remote as a .patch" command ... [10:47] cgit does that... [10:48] or that could give you it with a nice filename, indeed [10:51] bah [10:51] I don't understand why https://launchpadlibrarian.net/184385786/buildlog_ubuntu-utopic-i386.evolution-data-server_3.12.6-0ubuntu2_UPLOADING.txt.gz has its translations in "list-missing" [10:51] they are installed in debian/tmp [10:51] (from the log) [10:52] e-d-s-common.install lists the directory [10:54] or that's pkgstriptranslations [10:54] but then I don't get why the .mo don't show up on https://translations.launchpad.net/ubuntu/utopic/+source/evolution-data-server/+imports?field.filter_status=all&field.filter_extension=all [10:54] pitti, ^ do you have any idea? [10:55] it makes https://translations.launchpad.net/ubuntu/utopic/+source/evolution-data-server show 3.12 as untranslated [10:55] hum, maybe I should have accepted that template as "evolution-data-server" rather than -3.12? [10:56] where is dpm when have launchpad translation questions? :-) [11:01] seb128: that patch doesn't fix the widget factory [11:03] seb128: wisely hiding as you have launchpad translation questions? [11:15] Laney, not even combined with the theme fix in the same lp bug? [11:15] davmor2, seems so [11:15] oh, didn't notice that [11:29] nah, still doesn't work :( [11:34] :-( [11:34] did you try gnome-tweak-tools to see if that one is fixed? === MacSlow is now known as MacSlow|lunch [11:34] it breaks when I change tab [11:35] Laney: already removed the properties notices. didn't remove the mutex assert -- i think that will stay, actually [11:35] should backport the gtk fix i guess [11:35] kay, up to you [11:36] you doing a release? [11:36] i can [11:36] also, hi :) [11:36] seems relatively minor at this point, though [11:36] may just wait to do it on schedule [11:37] morning desrt! [11:37] good morning [11:37] i will leave my house in 20 minutes [11:37] to drive to downtown toronto [11:37] seb128: I've MRed a backport for gnome-system-monitor, which fixes it for all themes (Adwaita wasn't affected, but High Contrast etc. were)( [11:37] being that it's 8am, this will take approximately an hour [11:38] desrt: but ... why? [11:38] and is something that, under normal circumstances, i would never ever do in a car [11:38] because, seriously... [11:38] but... [11:38] ya [11:38] it turns out that today is a special day [11:38] i have a friend who is being released from the hospital and he needs a ride home [11:38] awesome \O/ [11:38] yay! [11:38] nice [11:38] great news [11:38] larsu, thanks, looking at that, so it was not the theme itself for once? [11:39] seb128: yep. The background was drawn incorrectly. I had a fix and then realized there was a (better) one upstream already, so I backported that [11:39] larsu, great, thanks [11:39] we shoudl really update g-s-m... we're on 3.8 [11:40] yeah, I think the issue was gtkheaderbar [11:40] going to look at updating those with patches similar to evince [11:40] but that's for next cycle [11:40] * larsu makes a note of that [11:41] larsu, looks good, bonus point next time if you do (lp: ) in the changelog (doing that for you on this one) [11:41] seb128: oops, sorry [11:41] * larsu never remembers _all_ the things [11:41] no worry [11:46] larsu, on updating apps topic, next fun is going to be those annoying .desktop renames [11:46] the one to match the dbus name [11:46] I wonder what to do about that [11:46] I guess desrt is not going happy if we revert the renames [11:46] what's the problem? [11:46] users launcher settings getting screwed around? [11:47] yes [11:47] is that the only issue? [11:47] yes [11:47] but it's an annoying one [11:47] do we care about users? ;) [11:47] * seb128 slaps desrt with a trout [11:47] i have a plan... ish [11:47] we need a big map [11:47] no [11:47] we need [11:47] [Desktop Entry] [11:48] AliasFor=org.gnome.newname.desktop [11:48] as a separate file? [11:48] yes [11:48] it's the only sane way [11:48] how would that work? [11:48] hello stat! [11:48] larsu: no. [11:48] then you need to patch all the .desktop parsers [11:48] we'd love to call you more often [11:48] stat stat stat [11:48] * larsu sings [11:48] larsu: what are you even talking about? [11:49] seb128: other choice is symlinks [11:49] that leads to duplicate items in e.g the unity dash [11:49] seb128: unless we update all the desktop file parsers :) [11:50] :-/ [11:50] let's not rename [11:50] one way or another something has to give [11:50] you can't not rename :p [11:50] well, those had a name let's keep it [11:50] bet I can? [11:50] the problem will just get bigger with time [11:50] and a lot of these apps are depending on dbus activation [11:50] keep the old one with NoDisplay [11:50] Laney: NoDisplay would be implied by Alias= [11:50] doesn't require fixing parsers [11:51] with the added benefit that old parsers wouldn't even understand it at all [11:51] if we don't care about that, then ya [11:51] Laney, would e.g nautilus show launchers with NoDisplay? [11:51] desrt: I'm talking about not having a desktop file cache yet [11:51] seb128: when browsing /usr/share/applications, yes [11:51] desrt, well, what if I've a nautilus link on my desktop to /usr/share/applications/gedit.desktop [11:51] and we add NoDisplay there [11:52] yes [11:52] that should still be shown properly [11:52] yes what? [11:52] * seb128 tries [11:52] we do want the explicit aliasing functionality though [11:52] because we want to give a mechanism by which things like the launcher can update themselves [11:53] ie: check that the desktop file name is what we think that it is [11:53] and update it if it doesn't match [11:53] we could do that with symlinks as well, but it would involve more stat() calls ;) [11:53] seems like a reasonable feature [11:53] yeah [11:53] you could do both as a bridge [11:54] shame that things are done backward again though :/ [11:54] rename first [11:54] deal with migration later [11:54] meanwhile users get screwed [11:54] we could also just do a map [11:54] "map"? [11:54] * larsu coughs [11:54] possibly populated via the first mechanism [11:54] or something like it [11:54] "we" [11:54] we is me :) [11:54] s/we/desrt [11:55] would that work for e.g KDE? [11:55] no [11:55] or would they need to do a map as well [11:55] k, what I though [11:55] no matter what we're looking at changes to desktop file parsers [11:55] shrug [11:55] not with the symlinks and NoDisplay [11:55] symlinks are the only way that we could completely avoid that [11:55] but they're only nice if you enjoy double-vision [11:55] well [11:56] having all the main parsers updated is going to take time [11:56] that should have been done before we start renaming [11:56] if you're really upset, i can give you glib by the weekend :p [11:56] no, I'm not, we are one cycle behind GNOME [11:56] anyway... if we do a map file, we can populate it via the desktop files themselves [11:56] so we didn't hit that wall yet [11:56] from WasNamed=gedit.desktop [11:56] or such lines [11:56] oh, the Debian guys were discussing a related issue yesterday [11:57] your favorite apps handler getting screwed on upgrade [11:57] seb128: this is sort of how things work, though.... [11:57] upstream does the awesome future stuff [11:57] and the distro gets the fun work of figuring out how not to screw existing users [11:57] the mimetype associations also [11:57] welcome to the last 10 years of gnome-ubuntu relationship :) [11:57] like you upgrade and you get okular used for pdfs [11:58] seb128: the new spec should be backcompat to glib's previous behaviour [11:58] if not, i want to hear about it [11:58] because evince.desktop which is associated to pdf in your config isn't there anymore [11:58] ah [11:58] okay -- that's expected :) [11:58] "expected" [11:58] still wrong :/ [11:58] i thought you meant about the mimeapps selection process itself [11:58] oh no [11:58] seb128: stop being grumpy :p [11:58] just another place which records the .desktop names [11:58] this is our job to fix and we will [11:59] I'm grumpy because those sort of things should be figured out before the changes happen :/ [11:59] but yeah, let's discuss it in DC [11:59] that's a next cycle topic [11:59] i hope to have a fix before then :p [11:59] well your fixes are glib/GNOME specific [11:59] yup. [12:00] * desrt drives into the city [12:00] talk to you in a bit [12:00] say hi to your friend from me [12:00] desrt, see you! [12:00] Laney: will do :) [12:01] maybe talk to him about gtk/theme fixes for gnome-tweak-tool :-) [12:01] * Laney runs [12:09] seb128: you planning a gtk upload? [12:09] Laney, no, I did one this morning for gtk2 after our discussion and that was the one I had planned [12:09] k, np, taking that mutex fix [12:10] thanks [12:11] Laney, you can use bug #1376530 for tracking [12:11] bug 1376530 in gtk+2.0 (Ubuntu) "GTK apps crashing with "Attempt to unlock mutex that was not locked"" [High,Fix committed] https://launchpad.net/bugs/1376530 [12:11] bug #1374030 as well [12:11] bug 1374030 in gtk+3.0 (Ubuntu) "Dropbox crash on Ubuntu MATE Remix" [Undecided,Confirmed] https://launchpad.net/bugs/1374030 [12:12] yeah i had it, ty [12:15] hello everyone [12:16] seb128: update to https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bug/445333 [12:16] Ubuntu bug 445333 in gtk+3.0 (Ubuntu Trusty) "remember password on printing to windows printers does not work" [High,Triaged] [12:16] hey dgadomski [12:16] dgadomski, what update? I see no new comment [12:16] seb128: hello :) [12:17] seb128: I have been testing the backport for gtk2 and I was observing crashes with firefox [12:17] I investigated this issue and there is a bug in gtk (both: 2 and 3) [12:17] so this needs to be first patched before this backend may be used [12:17] oh, nice, did you file it upstream? [12:17] yes, along with a patch https://bugzilla.gnome.org/show_bug.cgi?id=737777 [12:18] Gnome bug 737777 in Printing "Applications crash randomly while printing with a password-secured SMB printer" [Major,Unconfirmed] [12:18] thanks [12:18] would you like this to be included in the backport or should I make a separate lp bug for gtk2 and gtk3? [12:19] just add it as another patch in the series [12:20] you mean a separate patch in debian/patches? [12:20] yes [12:20] ok, I will, thanks a lot! === MacSlow|lunch is now known as MacSlow [12:52] I just pushed gtk then, given that [12:52] might be best to wait a bit for upstream to review first [12:56] Laney, speaking about the printing issue for the upstream review? [12:56] yeh [12:56] k, makes sense [13:00] lunch, biab [13:17] pitti, do you know where I can see the translations.tar.gz exported from a build? [13:17] see/download to look at the content [13:24] seb128: ah, I don't think there's a UI for it indeed, you need to use lp-shell; hang on [13:35] seb128: http://paste.ubuntu.com/8479201/ [13:35] pitti, danke! [13:42] ok, so the tarball is good [13:43] I guess it's because the launchpad template info was buggy [13:43] I've fixed the server side and did a no change upload of eds, let's see how that goes [13:44] Hi all. I've run into an issue with gstreamer1.0-plugins-good package: it provides gstreamer1.0-audiosink, probably because it includes jack output plugin, but without jack server installed the plugin doesn't really work as an audiosink. Does it make sense to make a separate gstreamer1.0-jack metapackage to include good plugins and jack audioserver for that reason and have it provide the audiosink instead? [14:24] immerrr, hrm - I'm not sure, but I think TheMuso might be able to advise you. It's like 3 in the morning or something for him right now, but he'll no doubt see this message when he logs in [14:34] marga: hey, can you verify SRU for this: https://bugs.launchpad.net/unity/+bug/1324114 ? [14:34] Ubuntu bug 1324114 in Unity 7.2 "Unity kills running compiz, even if it belongs to a different Unity session" [Medium,In progress] [14:38] Trevinho, I can, but not today, how urgent is it? [14:38] marga: not much, take your time... I just asked since you were the reporter with a such configuration ready, so it might be quicker for you [14:39] Right [14:41] marga: thanks [14:49] Laney, do you fancy fixing a segfault in some u-s-s code you wrote? ;-) [14:49] it's glib code [14:50] umm [14:52] link? [14:52] also holy shit that eds fix, I was going to debug that [14:53] hehe [14:53] https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1375988 [14:53] Ubuntu bug 1375988 in ubuntu-system-settings (Ubuntu) "System settings crashes after quickly backing out of storage" [Critical,Confirmed] [14:53] Laney, see the comment I just added [14:54] not sure what to do there/why they don't stop [14:54] we can wait until the callbacks are done before exiting [14:54] but that seems suboptimal [14:56] is there a trace? [14:56] no but I had a valgrind log and easy to reproduce [14:57] if I put that in the destructor [14:57] while (!g_cancellable_is_cancelled (m_cancellable)) [14:57] g_warning("not done yet!\n"); [14:57] it fixes it as well [14:57] with outputing those "not done yet" for a while [14:58] Laney, it's an invalid free in measure_finished() [14:58] well, as said I think that's because we exit the plugin/structures are freed but the callback are still pending [14:58] when they come back they try to free datas that are gone [14:58] -> segfault [15:01] Laney, http://paste.ubuntu.com/8479699/ [15:01] valgrind === gatox_ultra is now known as gatox [15:02] 123 is [15:02] if (err != nullptr) { [15:02] if (g_error_matches (err, G_IO_ERROR, G_IO_ERROR_CANCELLED)) { [15:02] delete data->running; [15:02] the delete ^ [15:02] so the cancellable cancels after the plugin is gone [15:02] yes [15:02] 123 is the unref here [15:02] well, I put debug prints [15:02] haha [15:02] that's why I tell you what line it is :p [15:02] it's the delete [15:02] kay [15:04] Laney, do you have any idea what the right fix would be there? [15:04] waiting on the callbacks to be done to exit the destructor? [15:08] I reckon it's because the callback is called multiple times [15:08] could be [15:09] is that one for desrt? ;-) [15:22] Trevinho: hi, thanks for your comments to the merge proposal, I have corrected it according to them [15:22] dgadomski: cool, I'll review them in a bit [15:23] Trevinho: I'll appreciate it, thanks [15:24] dgadomski: (privateScreen.optionGetHsize () > 1 || privateScreen.optionGetVsize () > 1) can also be (privateScreen.optionGetHsize () * privateScreen.optionGetVsize () > 1), as you prefer :) [15:26] Trevinho: whatever you prefer, they both get the job done ;) [15:27] dgadomski: can you also fix indentation for PointCompare? I really hate this way of indenting, but that's it... :/ [15:27] dgadomski: and also [15:27] + savedFocus.erase (it); [15:27] + return w; [15:27] should be two tabs... [15:29] Trevinho: sure, fixed, please let me know if there is anything else before pushing [15:30] dgadomski: just few things, I'll post them in the revew in a sec [15:30] ok [15:31] seb128: something like http://paste.ubuntu.com/8479858/ ? [15:31] can't test it, desktop is too fast [15:32] Laney, I put a sleep(1) in the contructor [15:32] so I can click on the header while it's frozen [15:32] reproduce easily on desktop for me [15:32] oh ok [15:32] Laney, that diff ... if (data->running), you access the data structure after the plugin returned [15:32] that's likely to be problematic as well [15:33] where? [15:33] + if (data->running) { [15:33] + delete data->running; [15:33] but that callback is called after the plugin exited, I'm not sure the data structure is still allocated [15:34] dgadomski: done [15:34] oh right, yeah, it's on the heap so should be [15:34] might be another leak there actually [16:02] Trevinho: thanks, I've just updated the merge [16:20] dgadomski: thanks... I'm testing it, but for some reason that code don't run... and the correct lib is going, so let me check better :P [16:20] dgadomski: btw, there should be just another small clanup to do: if a window closes, then we should make sure that it gets removed from savedViewportFocus [16:21] dgadomski: oh, no sorry... that happens automatically in findViewportFocusCandidate [16:21] dgadomski: no need to do it, it will be delayed, but it's fine [16:22] dgadomski: ouch, no... So please move "savedViewportFocus.erase (it);" up before the "if (w)"... In this way also if we can't find a window that has that id, we erase it from the map [16:23] we only return it if found [16:35] seb128: try that [16:35] seems to wfm === gatox is now known as gatox_lunch [17:02] byesie bye! [17:03] chrisccoulson: https://www.youtube.com/watch?v=0YBumQHPAeU <- you brits certainly have the more interesting politicians ... === gatox_lunch is now known as gatox [21:31] Got an issue in 14.10 with Gnome. When logging in my mouse pointer freezes on the screen. [21:31] can still move the mouse, but the original one stays visible on the screen. Any idea's? [21:32] xnox: hey there! we're seeing an odd bug in ubiquity with xubuntu utopic with the background staying black. any idea what that could be related to? (i checked bzr and couldn't find any relevant changes and we haven't changed anything in our setup) === m_conley is now known as m_conley_away [23:15] immerrr: Makes sense to me, but I am not in charge of gstreamer, so you may want to talk to gstreamer maintainers.