=== JanC_ is now known as JanC [06:05] good morning! === duflu_ is now known as duflu [06:18] morning larsu [06:18] hi happyaron! === qengho is now known as CardinalFang === CardinalFang is now known as qengho [07:15] good morning [09:02] ahoy there [09:02] moin Laney [09:05] morning Laney [09:05] greetings larsu et didrocks [09:05] seems quiet today [09:05] indeed [09:05] so, my fix to gnomed-desktop breaks nautilus [09:05] because it handles device scale as well [09:05] sigh [09:07] :/ [09:08] this revert sucks [09:08] bah. my compiled nautilus behaves differently from the one in distro [09:09] morning desktopers [09:09] hey willcooke [09:10] morning willcooke [09:10] how's london? [09:10] Cold and wet [09:10] \o/ [09:10] hooray [09:10] not news! [09:10] you guys are selling your capital so well [09:11] you can join the club [09:11] didrocks, can you do me a massive favour and reply-all to that email re: the issues with the store? I won't have time today and I think it needs to be moved forwards asap [09:11] as a parishater [09:11] I'll never have a similar speach for Paris… oh wait! [09:11] willcooke: sure, doing in a few minutes then! [09:11] didrocks, if you're not happy to - pas de probleme [09:11] didrocks: every English person I've met that doesn't live in London hates London [09:11] morning willcooke [09:11] I dont hate London, I just dont like the weather so much :) [09:12] do you like it though? [09:12] Erm, I like to visit [09:13] Dont think I could work here everyday [09:13] or live here [09:13] didrocks: [rephrased] every English person I've met that doesn't live in London doesn't quite like London [09:13] haha! [09:13] ;) [09:14] heh :) [09:15] is noone running unitz on hidpi? [09:15] * larsu wonders why nobody noticed our atrocious background rendering [09:16] larsu: maybe people having hdpi thought it was pixel art [09:17] didrocks: I made a white test wallpaper with some black pixels in a 1×1 grid. It's all blurry :( [09:17] * larsu should complain less and fix more [09:18] does gnome get it right? [09:19] * larsu checks [09:23] Laney: hm, can't switch to a gnome session [09:23] /usr/share/xsessions/gnome.desktop exists... [09:23] can't how? [09:24] in lightdm [09:24] like no option? [09:24] ya [09:24] ah wait, those are the wrong files [09:24] * larsu ponders what to install [09:24] gnome-session [09:24] no [09:25] that's not enough [09:25] it should be, provided you have gnome-shell as well [09:26] larsu: do you have /usr/share/gnome-session/sessions/gnome.session? [09:26] wow. how many of those things do we have? [09:26] (normally, that doesn't prevent the session to shows up in lightdm though) [09:26] well [09:26] there is the session "session" [09:26] darkxst: that's what I meant by "it's not enough"... ;) [09:26] there is the gnome-session "session" [09:27] (I'm culpurit of the second) [09:27] which ones are the files lightdm looks at? /usr/share/xsession/*.desktop, right? [09:27] yeah, if you don't have the session, that's weird [09:27] should be that one [09:27] there is a condition though [09:27] let me look [09:27] why doesn't gnome-session pull in gnome-shell, for example? [09:27] larsu: TryExec=gnome-shell [09:28] it was causinga circular dep [09:28] do you have the gnome-shell bin installed? [09:28] I do now [09:28] ok, I guess restart lightdm [09:28] the TryExec is evaluated at lightdm startup IIRC [09:29] (and as you can see, this calls Exec=gnome-session --session=gnome) [09:29] unity-greeter theming seems to have broken again [09:29] didrocks: ya, working now [09:29] didrocks: thanks [09:29] yw! [09:29] and, keep in mind that unity needs gnome-shell [09:29] oupss [09:29] gnome-session [09:30] so, if gnome-session was depending on gnome-shell… [09:30] it does? it wasn't installed for me [09:30] (first, I wrote "g-s depending on g-s") [09:30] didrocks, no, unity uses gnome-session-common (or -bin) [09:30] ah, that changed [09:30] ok, I guess gnome-session should recommends gnome-shell [09:31] didrocks: I agree [09:31] darkxst: deps on one side, recommends on the other is fine [09:31] why? [09:31] Laney: the package contains /usr/share/xsessions/gnome.desktop [09:31] If you want shell, install shell [09:31] and /usr/share/gnome-session/sessions/gnome.session [09:31] and that's mostly it [09:31] so, 2 components deeply linked to gnoe-shell [09:31] so shell should depend on it [09:31] gnome-shell [09:32] Laney: to answer your question: gnome gets it wrong as well [09:32] /o\ [09:32] well, in that case, we can remove "gnome-shell" binary package [09:32] * didrocks doesn't think /etc/lightdm/lightdm.conf.d/50-ubuntu.conf should be part of this [09:32] and that's it for this package [09:33] Laney: on unity7 desktop which process handles the alt-f2? [09:33] dunno [09:33] probably unity [09:34] xnox: unity-lens-applications [09:34] things launched from there lack environment that is set via user-session upstart and/or possibly gnome-session. [09:34] didrocks: thanks. [09:34] xnox: enjoy my vala code :p [09:34] didrocks: hm, i don't have that running here. [09:34] didrocks: is it still lens in 14.04 LTS? [09:35] xnox: that was renamed commands.scope [09:35] /usr/bin/unity-scope-loader applications/applications.scope applications/scopes.scope commands.scope ? [09:35] but it's still the same package [09:35] didrocks: ok, and what launches that? [09:35] cause it's missing environment variables =( [09:35] xnox: dbus-activated through unity [09:36] IIRC [09:36] didrocks, 50-ubuntu.conf is in ubuntu-session also [09:36] gegls from outer space [09:36] didrocks: right, so i should be uploading enviromental variables into dbus.... however that is not in trusty's environemnt [09:36] Laney: there are 2 more easter eggs :) [09:36] xnox: agreed [09:38] didrocks: what about gnome-session variables? [09:39] good morning desktopers [09:39] didrocks: Laney: it's funny how ctrl-alt-T, clicking gnome-terminal has the env, and alt-f2 xterm does not. [09:39] xnox: I think simply everything that are set to the process (so, including gnome-session variables) when you control+alt+t should be into those processes [09:39] SSH_AGENT is in question here. [09:39] xnox: they got it at first, before being dbus-activated [09:39] can anyone give me a gnome-session set variable? [09:39] but yeah, once it's been transitioned, I guess, that's lost [09:40] xnox: XDG_CURRENT_DESKTOP=Unity for instance [09:40] hey seb128 [09:40] let me ensure, one sec [09:40] didrocks: ok that one is set in alt-f2 xterm. horum. [09:41] xnox: one sec, it's maybe not set by it if already present [09:41] hey didrocks ;-) [09:42] didrocks: where is this scope thing again - apps?! [09:42] xnox: yeah, the unity-lens-applications package [09:42] XDG_CURRENT_DESKTOP is set by the DM, and passed through gnome-session I think [09:43] yeah, it's only set if it wasn't set beforehand [09:43] xnox: GNOME_DESKTOP_SESSION_ID [09:43] not used anymore, but set by gnome-session [09:44] didrocks: upstart jobs set that as well, as far as i know. but let me see if i the value is different. [09:44] no it's same. [09:48] weird that we reset it [09:48] there is really no value in doing this [09:48] False [09:48] didrocks: yes there is. [09:49] didrocks: Qt looks like Windows 95 without it. [09:49] [ Iain Lane ] [09:49] * xsession-init: Set $GNOME_DESKTOP_SESSION_ID if we are launching a [09:49] gnome-session session. Some applications (Qt4) require this to be set to [09:49] any value to detect the environment in use. It was historically set by [09:49] gnome-session but now this is no longer the root of the session so not all [09:49] user processes are guaranteed to have it. (LP: #1305294) [09:49] Launchpad bug 1305294 in upstart (Ubuntu) "QT uses incorrect theme when GNOME_DESKTOP_SESSION_ID is unset" [Medium,Fix released] https://launchpad.net/bugs/1305294 [09:49] didrocks: hm..... i'm failing to see how things are actually execed from the scope =( [09:50] waow "nice" [09:51] didrocks: hm. i think /usr/bin/unity-scope-loader should be managed by upstart in the user session to have the right env [09:51] it's dbus activated at the moment as it has UPSTART_JOB=dbus env variable. [09:52] xnox: yeah, I don't know at all about those changes (I wasn't in charge of unity when they did that change), maybe ask thorst? [09:52] hm. or it can call into upstart dbus api and do "list-env" and then import that. [09:53] is even the lens actually launching things? or does it simply hand over result to shell to launch stuff?! [09:53] xnox: it was launching things at the time, let me have a look if this changed [09:53] AppInfo.launch_default_for_uri [09:53] Process.spawn_async [09:53] yep [09:54] /* Activation of standard application:// uris */ [09:55] * xnox ponders if i should just strace it.... [09:56] didrocks: to many if/else blocks.... if feels like half of the code there is unreachable. [09:57] xnox: g_desktop_app_info_launch() launches whatever's in the Exec= key or dbus-activates the application if it supports that [09:57] no clue why the scope would do anything but call that... [09:57] larsu: well i can type ~/bin/foobar in alt-f2 window. [09:58] oh, this is about alt+f2 [09:58] larsu: yes. [09:58] * larsu apologizes [09:58] larsu: app_info_launched things seem to have the right env. [09:58] xnox: right, but alt+f2 shouldn't use that [09:58] larsu: didrocks: ok this is crazy -> [09:59] Super -> "xterm" -> Enter = has SSH_AGENT [09:59] Alt+F2 -> "xterm" -> Enter = does not [10:00] xnox: I guess, really ask the unity API team, not sure if it's maybe just one data missing when they changed to activated scope [10:01] xnox: so Alt+f2 isn't passing the same env as g_desktop_app_info_launch() is [10:01] maybe the scope guys should fix that :) [10:03] bug #1433013 [10:03] bug 1433013 in upstart (Ubuntu Vivid) "Super -> exec vs Alt-F2 -> exec have different environment" [Undecided,New] https://launchpad.net/bugs/1433013 [10:05] didrocks: larsu: Laney: i guess I'll experiment with launching the scope loader via initctl start, such that it has the right env. [10:06] xnox: I assume the env is there, otherwise g_desktop_app_info_launch() wouldn't have it. I bet whatever handles alt+f2 just doesn't pass it along [10:08] larsu: i get to differ $ cat /proc/`pgrep -f unity-scope-loader`/environ | grep -a SSH [10:08] is empty [10:11] xnox: it's gotta come from somewhere thoguh [10:12] initctl list-env has it. [10:14] larsu: ok, so changing the dbus to do Exec=start unity-scope-application, and moving the process to an upstart job works. And then xterm launched via alt-f2 works and has ssh environment. [10:14] i'll make a patch for that, and will have to see if that gets accepted. [10:15] xnox: shouldn't unity-scope-application inherit the environment from dbus, which is already upstart? [10:16] jodh: http://upstart.ubuntu.com/wiki/Stanzas is not feeling good [10:16] larsu: it does, but after dbus is execed, upstart's environment is modified, and it modifies the dbus process environment table, but that doesn't propagate to dbus' children. [10:17] larsu: if dbus child is moved to be a direct upstart child, it will mangle said childs environ [10:17] when people do $ initctl set-env foo=bar [10:17] * xnox tests [10:17] xnox: that's what org.freedesktop.DBus.UpdateActivationEnvironment is for [10:17] larsu: is that in 14.04? [10:18] xnox: dunno [10:18] * xnox thought that was added only recently [10:18] it was, but I don't know when [10:18] but running 'start' in an Exec line sounds hacky to me [10:18] hm. initctl set-env didn't propagate to newly execed xterm via alt-f2..... not good. [10:19] larsu: looks like there is UpdateActivationEnvironemt.... [10:19] in trusty [10:19] \o/ [10:20] larsu: there is no get activation environment though..... =( [10:21] why do you need that? [10:21] larsu: to check what it currently is. [10:21] well yeah [10:21] but why do you want to do that [10:22] to trace the SSH_* variable =) [10:22] oh, for debugging [10:22] yeah. [10:25] xnox: just activate some process and inspect its env ;) [10:25] like a terminal... [10:32] xnox: ergh. thanks, I'll report that... [10:37] 'GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name com.canonical.Unity.Scope.Dima was not provided by any .service files' [10:37] * xnox gives up [10:47] ok everything works, as long as initctl set-env does a call to dbus to update activation environment. [10:47] i can do that at post in each job or patch upstart === MacSlow is now known as MacSlow|lunch === MacSlow|lunch is now known as MacSlow [14:19] aha [14:25] who knows where I can get the iso smoke testsuite scripts from? [14:29] seb128: I'm looking into that software-center issue now, no need for dobey to duplicate work, its under control [14:29] seb128: looks like its a server side issue [14:30] bah [14:30] jibel: yo [14:30] willcooke: ^^ [14:30] if I want to change https://jenkins.qa.ubuntu.com/view/vivid/view/Smoke%20Testing/job/vivid-desktop-amd64-smoke-default/130/ this testsuite [14:30] oh [14:30] where do I get it from? [14:30] dobey, thanks [14:35] happyaron: FYI, we approximatly gained 5M thanks to the fcitx switch and removal of ibus-pinyin, so size-wise, it's cool :) [14:35] Laney: seb128 ^ [14:35] (compressed size) [14:35] gaining 5M isn't that great [14:36] gained in the term of "less" [14:36] HAHA [14:36] Laney: I wouldn't be that enthousiast if we grew by 5M :) [14:36] that is innovative use of english [14:37] Laney: yeah, I see as well always this contradiction as well with "false positive" [14:37] willcooke: hi, happy to give you a update on the issue if you want [14:37] (see how seb and I use it) [14:37] I thought you mean that it's okay because it is only 5 megs [14:38] no no, it's ok because the image size + some manual adjustement to compensate the seed changes is approx 5M :) [14:39] good times [14:40] mvo, could you reply-all the email, keep everyone in the loop [14:41] jibel: found it! [14:41] 'usit' ... [14:42] didrocks: btw this is a machine ID thing again [14:42] did you decide to symlink the old file? [14:43] Laney: we didn't decide on anything to do on this. Do you have the case details leading to no machine ID file for dbus? [14:44] (I really think the symlink is what makes sense, just need to decide where this should be done… I would say in the .service file) [14:45] It'll fall back to using the systemd one [14:46] also dbus-uuidgen --ensure will make this symlink [14:46] willcooke: sure, just did that after I saw your mail about it [14:47] mvo, thank you! [14:47] Laney: so, what's the bug is? (if it fallbacks successfully?) [14:47] only the tests making the wrong assertion? [14:47] tests look for the old path [14:48] dunno if any software does [14:49] *hem* oneconf probably *hem* [14:50] didrocks, nice! [14:51] I hope the "nice" was for the won megs on the iso, not my last statement :p [14:51] with open('/var/lib/dbus/machine-id') as fp: [14:51] hostid = fp.read()[:-1] [14:52] Laney: yeah, we do have *excellent* software doing this ^ [14:53] happy symlinking ;-) [14:53] Laney: want me to deal with this? [14:53] sounds quite trivial [14:53] if you don't mind [14:53] although the unit is upstream [14:53] :) [14:53] sure (not right now, but will do) [14:53] ah? [14:54] they use dbus-uuidgen --ensure in the unit? [14:54] no [14:54] that is what you'd add [14:54] yep [14:54] * didrocks just need to think if we should use that or provide the symlink manually (as it's in the boot path) [14:55] I'll do some tests [14:55] the upstart job and init script just ran the program [14:55] ok, anyway with kdbus coming, maybe doesn't worth trying to optimize this now [14:56] is there a bug for those failing tests? [14:57] dunno [14:57] I'm going to fix that part anyway [14:58] yeah, sounds more future-proof [14:59] kdbus is coming? when? [15:00] whenever this canadian guy finishes writing support for it in some fringe library [15:01] bregma's supporting it in nux? [15:01] Trevinho: do you know why compiz doesn't render the background on unity? Because it doesn't handle the gradients? [15:01] larsu: i'm still waiting for a round of lkml patch review that doesn't turn into a total bloodbath :p [15:02] larsu: where? [15:02] desrt: last I heard was that it's as good as in [15:02] Trevinho: err, what? [15:02] larsu: did you hear that from lennart? :) [15:02] larsu: I mean, I'm not sure I understood what you mean [15:02] desrt: maybe‽ [15:03] Trevinho: ah sorry. Right now, nautilus renders the background wallpaper in unity. We're patching it to do that because upstream stopped [15:04] larsu: ok, right... [15:04] in any case, i'm unlikely to spend more time working on kdbus until i have an actual API to target [15:04] Trevinho: and I'm pretty sure that compiz could render the background for us. Just wondering why it doesn't [15:04] desrt: fair enough [15:05] larsu, because nobody has fixed the Compiz wallpaper plugin in the last 5 years [15:05] larsu: mh, I don't know exactly the reason.. I guess it was an historic deicsion as in the past every desktop had its "desktop-bg app", and thus to keep modularity [15:06] also, Compiz does not support placing icons on the desktop and handling them as if it were a file browser [15:06] bregma: right, nautilus can keep doing that on a transparent window [15:07] bregma: I'm not saying we should go that route. Just wondering why we aren't [15:07] yeah, in fact... can be done at app lelve... [15:07] * larsu had to fix some hidpi issues in that code today [15:07] and it's still badly broken [15:07] does it support gnome's interface? [15:07] work with solid colours and gradients and stuff? [15:07] Laney: ya, I suspect it doesn't [15:07] * Laney would guess it is problems like this [15:09] we could revivie the Compiz wallpaper plugin to do the drawing, but the code has not been touched since 2009 and has not been ported to GLES [15:10] bregma: I think fixing the remaining issue in gnome-desktop is ok for now, but we might think about that if we want more features [15:11] or if we want more performant bg transitiosn [15:11] * didrocks is particularly afraid by the extended combinations of transformations as well to support [15:11] didrocks: transformations? [15:12] larsu: the crop/scale/extends… [15:12] (basically the various cases in libgnome-desktop [15:13] didrocks: wouldn't it be awesome if we did them on graphics hardware instead of the cpu? [15:13] same for the gradients [15:13] seems you want to bother this poor GPU… :) [15:14] I lie, the COmpiz wallpaper plugin was ported to GLES (according to https://docs.google.com/a/canonical.com/spreadsheets/d/1tiCs1CLXZB386-6TtD_bJsWHeMBJRjNVr8D14VVFHcY/edit#gid=0), maybe it's worth playing with for the 15.10 timeframe [15:14] didrocks: lol [15:14] bregma: I want animated wallpapers! [15:14] and FIRE [15:14] \o/ [15:15] I want world peace and a cure for cancer [15:15] and a second pony, the first one is too small [15:15] how important is that related to animated wallpapers? [15:15] :) [15:16] larsu: world peace, or the pony? [15:16] * didrocks wouldn't say a world, been there in 98… [15:16] word* [15:17] the world was peaceful in 1998? [15:17] France won world cup [15:17] that's the only thing I remember [15:17] (and this is what people still refers here) [15:18] (sad that I hate soccer) [15:18] world cup in taking dives, right? [15:18] I got my first laptop! [15:18] desrt: pfffff [15:18] Laney: pretty sure I've seen you with a laptop before [15:19] man… it's not Friday [15:19] larsu: in 98... [15:19] that was obviously trolling [15:19] desrt: [15:19] didrocks: dunno. it's IRC. context is easily lost. [15:19] or is desrt trolling larsu in faking not understanding? [15:19] and so, I got trolled? [15:19] :) [15:19] and larsu is a nice guy. i don't think he would troll like that outside of friday. [15:19] wow [15:19] was googling for the scheme that gave me it [15:20] desrt: that was the point of the joke... [15:20] my school's gone all out for Microsoft [15:20] http://cambridgeppf.org/vision/3/stevens.pdf [15:20] Laney: planning to burn your diploma? :) [15:20] oh wow: http://www.newyorker.com/business/currency/the-gnu-manifesto-turns-thirty [15:21] https://www.tes.co.uk/article.aspx?storycode=306195 [15:22] What emerged during the first year of the pilot scheme was a renewed enthusiasm for learning. According to Valerie Winch, the students' form tutor, "their interest factor has quadrupled and their motivation has gone through the roof." [15:22] \o/ [15:23] Laney: i suspect faulty calibration of interestfactorometors [15:23] no shit [15:23] we rapidly discovered (a) instant messaging (2) console emulators [15:30] meeting time! [15:30] :o [15:30] no just wait 10 minutes, need to run! [15:30] brb [15:30] attente_, desrt, dgadomski, happyaron, FJKong, Laney, larsu, mlankhorst, qengho, Sweet5hark1, tkamppeter: hey! [15:30] hey :-) [15:31] mlankhorst: if you are not around, I'll just get to you later on [15:31] HELLO [15:31] I AM HERE FOR THE MEETING [15:31] hello [15:31] AT THE START OF IT [15:31] hey o/ [15:31] didrocks is the new boss? [15:31] #startmeeting [15:31] Meeting started Tue Mar 17 15:31:15 2015 UTC. The chair is didrocks. Information about MeetBot at http://wiki.ubuntu.com/meetingology. [15:31] Available commands: action commands idea info link nick [15:31] hi [15:31] * didrocks notes down Laney's remark :) [15:31] * larsu never experienced passive aggressive Laney [15:32] just making sure man [15:32] :D [15:32] got my ass handed to me last week [15:32] * larsu remembers [15:32] ahah, seems seb128 was really upseting you! :) [15:32] you should have subtly pinged seb128 in there [15:32] let's start [15:32] #topic attente_ === meetingology changed the topic of #ubuntu-desktop to: attente_ [15:32] attente_: hey! [15:33] stop u-s-d and u-c-c from pulling in fcitx-bin on dist-upgrade, which was causing it to become default... [15:33] i've got ibus somewhat running under mir, but it requires a lot of user intervention and the candidate window needs debugging [15:33] Laney, nice to see you are on time for the meeting for once [15:33] :-) [15:33] seb128: ya. i was surprised too. [15:33] didrocks: take control of these guys [15:34] Laney: not really, I find that funny :) [15:34] attente_: did you put on a wiki the necessary steps for setting up ibus under mir? [15:34] didrocks: no, i can write them up though. it is really not user friendly... [15:35]