=== sergiuse1s is now known as sergiusens [06:46] good morning, can someone advise me where I can get the source code for Unity 2D? [07:53] tsdgeos: hi [07:53] tsdgeos: did you see the bug I found in our jenkins setup? with the xpath query? [07:53] crazy shit :D [07:53] mzanetti: nope, didn't [07:53] what happened? [07:53] we have the builder job that builds the package [07:54] and then, when its done, the parent job does an xpath query to find the appropriate downstream job and download the .deb for further use in testing [07:55] now, when we added generic-mediumtests-saucy-touch, we also go a new builder, generic-mediumtests-builder-saucy-armhf [07:55] which rendered the xpath query for "generic-mediumtests-builder-saucy" find 2 jobs and throw an internal server error [07:56] but only simetimes, when the build number eg. 733 was also contained in the other e.g. 2733 :D [07:56] doh [07:56] pattern matching! [07:56] yeah that sucks :D [07:56] but it's fixed now [07:56] progress++ [08:01] flashing the phone is so much fun lately. always so much new stuff landing :) [08:13] jamesh: hello! [08:13] sil2100: hi [08:13] I saw your bug report [08:14] Things seem to have unblocked, and we have changed the data the test failed on [08:15] jamesh: indeed, since I see it built fine now - good! Michael Terry had also these two blocker issues with the main inclusion: [08:15] Quoting: [08:15] "* Is grilo-plugins-0.2 going to be dropped too? That is a problematic build-depend. * Needs a team bug subscriber" [08:15] I need to ask about the second thing, but I guess the first one is not possible to resolve, right? [08:16] sil2100: the default configuration uses grl-tmdb and grl-lastfm-albumart from that package [08:22] sil2100: I know there are questions about whether we want to use grl-lastfm-albumart though, and we haven't yet got an API key for themoviedb [08:22] sil2100: if we sorted out the alternative album art source and decided to drop themoviedb, then we could potentially drop the dependency [08:23] jamesh: I noticed that there were some problems with getting grilo into main before [08:24] sil2100: are there particular plugins that are problematic? [08:25] jamesh: I think the biggest problem with grilo in main is that the source does not run any tests on build, and all the tests that are in the source are out-of-date ;/ [08:25] Which is rather a blocker for the release team [08:25] We tried resolving that, but upstream doesn't care enough [08:28] that's certainly something that we can work on. [08:29] Grilo is a very strong dependency of the mediascanner, since it is the public API for the index [08:36] greyback: ping [08:36] tsdgeos: pong [08:37] i can't get the osk on screen anymore :'( [08:38] tsdgeos: at all? Even just at a clean boot? [08:38] yeah [08:38] i think maliit is core-dumping [08:38] let me check [08:38] feck [08:39] even running manually with QT_QPA_PLATFORM=ubuntumirclient [08:40] greyback: http://paste.ubuntu.com/6039570/ [08:41] installing libc-dbg to see if it gives me anything [08:42] tsdgeos: probably not. Try recompiling ubuntu-keyboard, but commenting out the ubuntu_iu-report_osk* functions in src/plugin/ubuntuapplicationapiwrapper.cpp [08:43] I swear it worked for me yesterday [08:43] hmm [08:43] this is bettar [08:44] http://paste.ubuntu.com/6039579/ [08:44] am i mismatched? [08:45] I'm not at all sure. Lots of stuff have landed in Mir the last 24 hours. Possible something changed [08:45] but yeah, ignore my above request. It's unrelated [08:46] Anyone upgraded with new trunk today on unity8? [08:47] greyback: hey, will work through your comments now. [08:47] greyback: regarding the dropping of deprecated stuff. I'm all for it, but... [08:47] mzanetti: let me know what you think [08:48] greyback: if this is too far off from the current version, I fear it will take ages until you guys are able to switch to the new API [08:48] greyback: do you think that's not the case? [08:53] hmmm [08:53] ApplicationManager REJECTED connection from app with pid 5297 as no desktop_file_hint specified [08:53] and that's the maliit thing [08:53] greyback: ↑↑↑ [08:53] mzanetti, tsdgeos need help somewhere? [08:53] ricmm's new thing? [08:54] Cimi: you bored? [08:54] mzanetti, not really, I can start packing for holidays :P [08:54] mzanetti: you're right, it could take some time to switch. I believe the only way to smoothly transition to Mir is to change the qtubuntu API to match what we want (and what will be in unity-mir) [08:55] mzanetti, I am little on hold on tasks... [08:55] greyback: do you see that happening anytime soon? [08:55] mzanetti: other option is I implement the existing (not so nice) qtubuntu apis in unity-mir [08:55] greyback: I'd definitely vote for implementing the new API [08:55] mzanetti: well Daniel and I are trying to close the delta gap right now, it is on our list [08:56] mzanetti: so I say: implement the new API, let's not waste time on the old one, and then transitioning over [08:57] mzanetti: and I'll do my very best to deliver it quickly [08:57] greyback: full ack. could we in theory just add a the new signal focusChanged (or what it is) with the new ApplicationInfo to the old one? Then I could start using it and you guys can transition the rest whenever you see fit [08:58] mzanetti: absolutely. [08:58] tsdgeos: weird, maliit should be permitted by the app manager [08:58] greyback: ok. I'll clean this up, add tests and mock implementation to it and return it to you for review [08:58] if (command.startsWith("maliit-server")) { [08:58] authorized = true; [08:58] return; [08:58] } [08:59] exactly [08:59] greyback: maybe the QPA_BL_ABLA [08:59] makes that code bork [08:59] right, try exporting hte var and trying again [09:00] * greyback thinks this authorizer more trouble that it is worth [09:00] Cimi: if you're still looking for some small fix to do, you could check out why the Popover in unity is white text on white background [09:01] greyback: i don't get that rejected stuff anymore [09:01] but the thing still crashe [09:01] s [09:02] (gdb) bt [09:02] #0 0x40e6b8a0 in ?? () [09:02] #1 0x412c7872 in ?? () [09:02] #2 0x412c7872 in ?? () [09:02] tsdgeos: sorry I can't try it here yet, phone refusing to connect to wifi. Could you try the ubuntu-keyboard recompile I mentioned abve [09:02] sure [09:04] mzanetti, I think there was a bug report for that [09:04] mzanetti, let me see [09:05] Cimi: yes, there is one [09:05] jamesh: ah, one more thing! [09:05] jamesh: regarding the unity-scope-mediascanner failing test [09:06] jamesh: yesterday I actually were able to get some debugging output from the test failure, not sure if it will be any use though [09:06] sil2100: It could help, so I'd be happy to look [09:10] mzanetti, where? [09:11] Cimi: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1201864 [09:11] Launchpad bug 1201864 in Ubuntu File Manager App "[popover] Impossible to read ListItems inside a popover when using SuruDark theme" [High,Triaged] [09:11] jamesh: let me pastebin it [09:11] thank u [09:12] Cimi: another one: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1205094 [09:12] Launchpad bug 1205094 in Ubuntu UI Toolkit "Popovers under SuruGradient theme are white-on-white" [High,New] [09:13] mzanetti, it's already assigned [09:13] Cimi: yeah, but florian is not working on it right now [09:13] mzanetti, more of an sdk thingy [09:13] Cimi: yeah, it's an SDK thing, but it doesn't seem to show up on the SDK people's priority list anytime soon [09:14] jamesh: http://paste.ubuntu.com/6039666/ [09:18] sil2100: okay, so I think this is a case of the mediascanner taking longer than expected to start up. [09:20] the timeout is inside the dbusmock wait_for_bus_object() helper, so isn't directly configurable [09:20] perhaps moving the part that can fail out of setUpClass() would do the trick [09:21] Since I would love to have that test back on, I love it [09:26] greyback: nothing, no keyboard at all [09:26] * tsdgeos keeps his :'( [09:26] tsdgeos: and useless backtrace? [09:29] yeah [09:29] three ?? [09:34] pstolowski, ping [09:34] tsdgeos: please try this one: https://chinstrap.canonical.com/~gerboland/packages/ [09:34] nic-doffay: pong [09:35] mzanetti, I'll take a look into those tomorrow if you need them sorted. [09:35] nic-doffay: ? [09:36] tsdgeos: please try this one: https://chinstrap.canonical.com/~gerboland/packages/ [09:36] think router on last legs [09:36] greyback: boom! [09:36] that one works :-) [09:36] mzanetti, those bugs you posted. [09:37] ah, ok. Cimi ^^ [09:37] mzanetti, ok [09:39] tsdgeos: can you work with that until I figure out what's wrong with trunk? [09:39] sure [09:39] thanks [09:40] greyback: now, you told me to remove some stuff and recompile, i did only recompile, so maybe why it was still crashing, want me to try to remove stuff you said? [09:41] tsdgeos: sure, just for sanity check. Those methods should not be called while on Mir. If they are called, they cause crash exactly like your BT describe (3 lines of rubbish) [09:42] sil2100: one other issue I ran into that I'm not sure who to follow up with: we don't seem to ever run xdg-user-dirs-update on the phone image, so the glib API for XDG special dirs doesn't give back useful results [09:42] do you know who to ask about that? [09:43] jamesh: ah, I think Michael pointed that bug out as well - I guess I would personally start off by poking ogra_ on #ubuntu-touch [09:43] okay [09:48] why doesn't the -I parameter on qmlscene get precedence over the plugins installed on the system? [09:49] at least it looks that way for me, doing "qmlscene -I . a.qml" does not even look at the current directory to find my plugin [09:51] * tsdgeos curses -Werror [09:51] to rebuild ubuntu-keyboard agai [09:51] that's a really useful way to spend my time [09:54] * greyback moving to office, back in ~40 [10:04] larsu: ping [10:04] dednick: hey. You're not pinging to answer my question, are you? [10:05] larsu: nope. :) sorry, i dont know that [10:06] nobody seems to, but it breaks my make check :( [10:07] * tsdgeos wants gerry want now not in 40 min! [10:09] larsu: i'm sure there must be a way to do it. dont think unity8 tests would work otherwise... [10:09] larsu: but on another note, can you take a quick look at this: https://code.launchpad.net/~nick-dedekind/qmenumodel/unitymenumodel.change.ActionStateParser/+merge/182855 [10:10] causing me issues [10:12] dednick: I think I'd prefer it if you had a setParent() in setActionStateParser() [10:13] dednick: but it doesn't matter really, because you create all of them with the model set as parent [10:13] larsu: the root action state parser is not parented by the menu. it's a sep entity so i dont want it deleted by the menu on destruction. [10:14] dednick: ah, okay. Approving :) [10:14] larsu: ta [10:18] tsdgeos, do you about plans for annotated icon in unity8? [10:18] annotated icon as in? [10:19] larsu: If you get a chance today can you take another look at the UnityMenuAction branch. I made the changes you commented on. [10:19] dednick: ah cool, I'll have a look [10:19] larsu: thanks. [10:19] larsu, try maybe asking the sdk guys for you -I question [10:19] ah, unity8 uses qmltestrunner... maybe that helps [10:19] tsdgeos, icon provider for image://anotated/... [10:20] dednick: thanks for the hint^^ [10:20] tsdgeos, but i guess that's a "no" :) [10:20] seb128: good idea, will do. It's the thing that causes your make check failure [10:20] mhr3_: yeah, not in my "space of things i'm aware of" sorry [10:22] larsu: we also use qmlscene for our "tryTest" targets. but it looks like it just uses the -I option. [10:29] dednick: in unity8? I can only find tests that use qmltestrunner in there [10:30] maybe I should just use that... [10:32] hm, same problem with its '-import' [10:38] sil2100, how is autolanding going to work after ff? merges against lp:xyz will still go to s, and lp:xyz/ver+1.0 will be t? [10:38] sil2100: on the subject of grilo tests, what is the best way forward? Is it a hard requirement that its test suite be brought up to date first, or is it something we can tackle after? [10:39] greyback: yo [10:39] have a sec? [10:40] mhr3: that is a valid question, but due to all the things happening now let me answer it with a "not sure right now!" ;p [10:40] tsdgeos: sure [10:40] sil2100, fair enough, will you send a mail once you do? [10:41] jamesh: I guess without those tests running, the release team won't let this package into Main after what I discussed with them [10:41] jamesh: so I think we need to target an FFe for mediascanner inclusion anywa [10:41] y [10:42] greyback: so, can you confirm that on the "bad landscape" you get [10:42] QUbuntuWindow::moveResize (this=0x83a888, x=262, y=0, w=506, h=1280) [10:42] ? [10:42] hi, how can I display in the unity's launcher side bar my mounted partition[/mnt/Warehouse] ? [10:43] tsdgeos: I'm just reflashing. I have these values in my notes: x=238, y=0, w=482, h=1280 [10:43] ah well [10:43] possibly our screens have different resolutions? [10:43] your screen is smaller [10:43] right [10:43] still [10:43] i'm getting QUbuntuWindow::moveResize (this=0x83a888, x=262, y=0, w=506, h=1280) [10:44] and i don't see the thing begin really 506 in width (which in this landscape case its his height) [10:44] x+w=770, which is not right [10:44] it is right [10:44] 768 [10:44] ah, /me idiot [10:44] that's the width of my screen [10:45] but with a rule i made in a paper (top tech!) [10:45] the w should be around 384 here [10:45] not 506 [10:45] tsdgeos: true, there's a chunk of hte OSK which is invisible. It's where the key-press popup is drawn for the top line of keys [10:45] ah [10:45] booo [10:46] so there goes my faith [10:46] sorry :( [10:46] greyback: i've just realized in "the good landscape" [10:46] i can't type [10:46] do you remember if you could? [10:46] tsdgeos: that's a different issue :) You can probably type on the right side of it? [10:46] right [10:46] we are not moving the punch hole [10:47] tsdgeos: here's hte problem. The OSK surface is on top of shell, and app surfaces. So it gets all input events by default - so shell and app don't get them [10:47] tsdgeos: oh you're ahead of me, cool [10:47] correct, the punch hole sometimes doens't move. Seems shell doesn't always get orientation change signals! [10:48] anyway as you said that's a different issue [10:48] * tsdgeos goes to his original one [10:48] oh [10:48] this one is interesting [10:48] if in the "bad landscape" [10:48] tsdgeos: really this whole OSK thing is a mess. I don't know how else to hack around it, aside from doing some major refactoring [10:49] i press the "123" thing [10:49] i get the "q" [10:49] so there is something that really thinks is at the bottom of the screen [10:49] yeah, so the inputs are being calculated correctly. But the draw is not in the correct place [10:59] larsu: what is being imported before? [11:00] dednick: hm? I think I'm missing context [11:00] larsu: oh hang on. i thought you ment that your imports were being imported after the system ones. [11:01] dednick: ah, yes. [11:01] larsu: but they're before? [11:01] dednick: no, it looks like it imports the system's version of gsettings-qt before it imports the local one [11:01] even though I specified -I [11:01] or -import on qmltestrunner [11:02] zsombi says it's my fault :-( [11:02] larsu: there's a env var you can set that will output the imports... cant remember it now. i'll look for it. [11:03] ah that would be cool, thanks [11:06] larsu, what command to you run to reproduce the issue? [11:07] seb128: I have a test script that I run with qmlscene -I . [11:07] seb128: zsombi is right, this works in ubunut-ui-toolkit [11:08] larsu: QML_IMPORT_TRACE=true i think [11:10] has anyone flashed their device today? is the battery brigtness slider still in the indicators? [11:10] dednick: yep that works, thanks. It shows me that it is loading the wrong plugin, so I'm not crazy :) [11:11] nevermind about battery. report was incorrect. [11:11] larsu: can you paste the comment you're using to import? [11:11] to run i mean [11:11] oh. "-I ." [11:11] dednick: qmlscene -I . a.qml [11:12] dednick: right, a.qml simply has an Item{} and a GSettings object [11:12] larsu: where is the qmldir file? [11:12] dednick: in GSettings/qmldir [11:14] ARGH [11:14] I've found it [11:14] this is stupid [11:14] dednick: the installed version is in a directory named GSettings.1.0 [11:15] which is more specific or something... when I move my local GSettings to GSettings.1.0, it works [11:15] hm. that's weird. [11:24] tsdgeos, mzanett: either of you know why the shell app is flickering like mad on desktop? [11:25] is it? [11:25] it was fine here [11:25] yesterday [11:25] dednick: what you do? [11:25] run it... [11:25] what flickers? [11:25] everything? [11:25] hand on, i'll get a vid. yeah, everything [11:25] let me pull & build [11:28] dednick: no blinking here [11:28] tsdgeos: pull down indicators? [11:28] dednick: you're on xmir? [11:28] tsdgeos: nope [11:28] ok, me neigher [11:29] dednick: still no blinking [11:30] sil2100: I'm stopping for dinner, but just put this up for review: https://code.launchpad.net/~jamesh/unity-scope-mediascanner/fix-integration-test/+merge/182877 [11:31] sil2100: that should hopefully fix the reliability problems in the test. [11:31] dednick: I get it. [11:31] tsdgeos: https://dl.dropboxusercontent.com/u/85539674/Screencast%202013-08-29%2012%3A30%3A39.mp4 [11:32] brrr [11:32] maybe the scopes are refresshing themselves? [11:32] and we reload the data? [11:32] seems that everthing is flickering though. not just the dash [11:33] dednick: you using qt5.1 PPA? [11:34] greyback: i think so. updated for it yesterday [11:34] dednick: me too. tsdgeos you? [11:34] greyback: but it's being doing it longer than that [11:34] greyback: so... I'd say this is ready for the real review: https://code.launchpad.net/~mzanetti/unity-api/application-api/+merge/182692 [11:34] really? Hmm [11:35] greyback: i think it has anyway. days merge into one another these days [11:35] greyback: no i'm on 5.0 regular [11:35] mzanetti: ack, on my list [11:35] grew a bit tho :/ [11:35] mzanetti: you on qt 5.1? [11:35] dednick: no [11:36] hm. ok, i'm going to purge that ppa and try [11:37] mzanetti: "This matches the Launcher api now very closely :)" - great, I'm happy that's the case === hikiko is now known as hikiko|lunch [11:40] greyback: just saying, the other api's in that repo are also in the same way. i.e. the "main" class is a model [11:40] mzanetti: ah, I understand you. Yep, that's good [11:41] om26er: hello :) [11:42] mzanetti, hey [11:42] did I broke something with the rename ? [11:42] om26er: no... everything perfect. [11:42] om26er: thanks for that btw [11:42] om26er: looking quite good now. there is one failure: https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-maguro/348/console [11:43] om26er: but I'm not sure if we can protect us better against this kind of stuff [11:43] om26er: maybe something like this in the beginning of a job [11:43] mzanetti, yeah, that's problematic I heard people were looking into solutions where internet would work from the server these devices were connected to so no need for wifi, but that didn't happen [11:44] through adb [11:44] while [ $? -ne 0 ]; do ping -c1 launchpad.net; done [11:44] that would wait until network is up... of course would need more checks etc... [11:45] mzanetti, In this case for example internet will only work of we try to reconnect to the network [11:45] its not like the internet was not ready its just some race where it does not work at all and needs to be retried [11:50] not exactly a walk in the park to get rid of qt5.1 :( [11:51] dednick, so it's just a spooky park with a wanted murderer lurking in there? [11:52] mhr3: i guess you could put it like that. if you really wanted to [11:53] dednick, but then you can that anything is a walk in the park ;) [11:55] mhr3: :) [11:56] generalizations ftw [11:56] greyback, tsdgeos: yeah, it seems that it's qt5.1 that was to blame [11:57] dednick: ok, that's not good. === alan_g is now known as alan_g|lunch [11:57] :) nope [11:58] Mirv: ping [11:59] i was also getting loads of errors from the TabStyle on 5.1, not getting anymore. [12:01] possibly sdk not 5.1 complient [12:01] which could potientially be the cause of flicker as well i guess [12:03] tsdgeos: now i'm getting "ShaderEffect: source or provider missing when binding textures" ! [12:08] sil2100, any idea what [12:08] FAILURE: http://jenkins.qa.ubuntu.com/job/unity-scope-home-saucy-armhf-autolanding/31/console is about? [12:08] greyback, any easy way of running unity8+SurfaceFlinger after having added that unity8-mir ppa? [12:09] dandrader: ppa-purge [12:10] tsdgeos: ah, i think that just be form missing sdk artwork. [12:18] greyback: pong [12:19] Mirv: hey, we're having some rendering strangeness with unity8 on the desktop. Might be something strange with a SDK component, am digging. === _salem is now known as salem_ [12:21] Mirv: lol, I should mention this is using the Qt5.1.1 PPA [12:21] greyback: ok. there is a known fix landing to morning's regression regarding Switch and CheckBox components. if something else, best pinging zsombi, timp or kalikiana from the SDK team [12:21] Mirv: ack [12:28] greyback: but it is Ubuntu.Application too right now. [12:28] greyback: do you really want to change that to Ubuntu.ApplicationManager ? [12:28] (imho Ubuntu.Application fits better because it holds all Application related stuff, not only the Manager - but I will change it ofc if you want me to) [12:31] mzanetti: Well, I'd change Ubuntu -> Unity, as it's a Unity only plugin. I'm not so bothered about the +Manager, you're looking through objective eyes so I'll trust your judgement === hikiko|lunch is now known as hikiko [12:32] greyback: even better. I'd vote for Unity too, just didn't expect you actually want that too :) [12:32] greyback: Unity.Application it is then, ok? [12:32] mzanetti: ok [12:44] dednick: interesting [12:45] Mirv: ↑↑↑ have you seen the flickering dednick talks about? [12:45] alex-abreu: hey, where do webapps install their .desktop files? [12:46] Wellark, atm in /usr/share/applications (for some default ones) or ~/.local/share/applications [12:47] alex-abreu: sweet [12:47] Wellark, ? :) [12:48] alex-abreu: for the not default ones [12:48] they go under ~/.local/share/applications [12:48] Wellark, u working on the launcher bits? [12:48] just as click? [12:48] greyback: ok. seems we're getting close with the application api. fixed the remaining issues [12:48] alex-abreu: so where do they install icons then? === jhodapp|afk is now known as jhodapp [12:49] mzanetti: thanks, looking [12:49] tsdgeos: I've only tried on device, and haven't noticed such [12:50] * Wellark wonders could we introduce X-Ubuntu-Application-Type key in .desktop files [12:50] click apps would specify "Click" [12:50] tsdgeos: this is starting to look good too: https://code.launchpad.net/~mzanetti/unity8/more-stable-tests/+merge/182778 [12:50] and webapps "Webapps" [12:50] that would remove the need for guessing where the icons are.. [12:50] Wellark, I haven't started in porting/integrating individual webapps to click, but otherwise /usr/share/icons [12:51] tsdgeos: 6 runs in a row without autopilot test failure (well one on the phone because wifi broke down) [12:51] alex-abreu: oh, will the webapps actually be click packages? [12:51] Wellark, "eventually maybe" ... I am not comitting to any eta, [12:51] I just need to know the rules where to look for the icons [12:52] Wellark, sure [12:52] and how to differenciate between a webapp .desktop and click .desktop under ~/.local/share/applications [12:52] Wellark, why do you need the distinction ? [12:53] for reliable icon lookup [12:53] to begin with [12:53] there might be something else also [12:54] but icon lookup is the first that comes to my mind [12:54] mzanetti: commit 98 removed all the Mocks, intended? [12:54] greyback: nooooo :D [12:54] greyback: moved ... forgot to add back :D [12:54] mzanetti: why not use "bzr mv" ? [12:55] greyback: dunno... I'll never learn that... [12:55] greyback: fixed [12:55] mzanetti: ta [12:58] Wellark, btw, actions api does not offer any way to add static actions to desktop file ... only to bind to existing/already defined actions from there right? [12:58] alex-abreu: yep. [12:58] alex-abreu: you have to write/generate your own desktop files [12:59] alex-abreu: what are you trying to do? [12:59] Wellark, yup, that's what I do === alan_g|lunch is now known as alan_g [13:01] mzanetti: approved [13:02] chers [13:02] +e [13:03] anyone knows if there is a easy way to access QSettings from QML? only needed temporary [13:03] tsdgeos maybe? ^ [13:04] nope sorry [13:04] mzanetti, there is a problem with the launcher icons (after the size was made smaller) there is a white glow on bottom part of most of the icons [13:05] and if you look closely like very closely you can see that the icon is not fully places above the title in the launcher [13:05] om26er: yeah, that's the UbuntuShape that paints it there [13:05] om26er: not placed where? [13:07] mzanetti, the icon for the phone app for example, if you look closely you see there is a little misalignment of something.. can't explain in words. let me get a screenshot [13:07]