[07:10] yay it only took 24hrs to migrate unity8... [07:13] Saviq, just in time to line a new landing then! :-) [07:14] seb128, yeah, we need some housekeeping first (we just resynced trunk with overlay) [07:14] but will have another request soon [07:27] Saviq: so overlay is officially not in use anymore? [07:27] awnt me to retarget my MRs? [07:27] tsdgeos, doing that right now [07:28] ok, it's always a bit ackward when my MRs are submitted by someoene else [07:28] UI-wise in launchpad [07:28] meh :) [07:28] at least now you don't need to copy the commit message [07:28] Morning o/ [07:29] Saviq: .po files are different, do we care? [07:29] tsdgeos, that will get fixed with the next translations import [07:29] tsdgeos, would've had to overwrite otherwise [07:29] ok, so we do not care [07:48] ok, unity-api and qtmir are trunk-only again, too [07:48] @unity, pstolowski ↑ [07:49] nice [07:50] all should have wily and vivid testing enabled [07:56] Saviq, ok, cool, i'm one step closer to have unity-scopes-shell ready for that too [08:02] Saviq: you did not top-approve MRs that were top-approved on overlay [08:02] tsdgeos, missed that, didn't notice there were any [08:03] is there a way we can tell anymore? [08:03] i'd say https://code.launchpad.net/~aacid/unity8/keepPreviewStackAround/+merge/270944 was [08:03] but maybe not [08:03] tsdgeos, mail history [08:06] biab, dr's appt === root is now known as Guest49862 [09:00] greyback: qtmir uses fontconfig? [09:01] tsdgeos: barely, I think the QPA plugin needs to link against it [09:01] weird === DanChapman_ is now known as DanChapman [09:32] whoa, touch support in gtk in wily [09:33] greyback: so it's not mir leaking the fd [09:33] since http://paste.ubuntu.com/12447270/ also gives me the extra fd [09:33] i'm thinking it may not even be a leak [09:34] just the image cache [09:34] woot, many jenkins results [09:35] tsdgeos: hmm, the image cache is a fixed size (in MB) right? So I guess after a while, those FDs would stop being created [09:35] or at least, old ones would be closed when new ones opened [09:36] yeah [09:36] so far i've only tried with the first one [09:37] need to see if past some range more fds are created [09:37] kevin said he saw like 70 [09:37] which seems a bit too many [09:37] the cache isn't watching each file in case it changes, surely? [09:37] dnotify/inotify style [09:37] nah since it's not even a file [09:38] it's memory stuff [09:38] ah [10:31] hmmm [10:31] this fakesensor tests are weird to be failing https://jenkins.qa.ubuntu.com/job/generic-deb-autopilot-runner-vivid-mako/3454/#showFailuresLink [10:32] and we have a qmluitests regression in master :S [10:32] my fault :D [10:33] we were so close to having a green qmluitests [10:40] tsdgeos, we might do another small silo today still, so let's get a fix for that [10:40] Saviq: https://code.launchpad.net/~aacid/unity8/fixTestDashFallback/+merge/271630 [10:40] that == qmluitests [10:41] or the sensors? [10:41] ack [10:41] the sensors stuff needs more investigation on why regressed [10:41] tsdgeos, see phone ML, likely image broke [10:41] qtubuntu-sensors fell off the image [10:41] ah yes [10:41] it's not required by anything anymore [10:42] those runtime deps are always hard for .deb :D [10:47] tsdgeos: i was thinking since you and i both verified that tedg's mp corrected at least one part of the fd leak prob, he/we should land that [10:48] * tsdgeos has a dejavu :D [10:48] kgunn: yes we should [10:48] * kgunn has dejavu too actually [10:48] but failed to find it in scrollback :) [10:52] greyback_: how do i enable qtmir debugging? [10:53] tsdgeos: QT_LOGGING_RULES="qtmir.*=true" [10:54] it's noisy, so can adjust to filter more stuff [10:54] gmmm [10:54] we've lost vibration on the phone too? [10:54] maybe the same sensors problem [10:54] qtubuntu-sensors fell off the image [10:55] there's a ML topic about it, fix in next image [10:57] kgunn, you'll be happy to hear unity8, qtmir and unity-api are single-trunk again, with CI running both wily and vivid testing for them [10:58] we've dropped the overlay branches/serie there [11:06] Saviq: regarding https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1497155 do we tell the news scope to not send us &8226; ? [11:06] Ubuntu bug 1497155 in unity8 (Ubuntu) "apostrophe in scopes bq aquaris e4.5 ubuntu touch" [Undecided,New] [11:07] has anyone run unity8 with valgrind on the phone recently? [11:08] tsdgeos, good question, on one hand I'd say we don't know what's coming from the data source, so we'd rather have plain UTF-8 coming to us, on the other I think we'll need to enable some rich text facilities, which might mean we need to support entities in any case === marcusto_ is now known as marcustomlinson [11:11] QObject: Cannot create children for a parent that is in a different thread. [11:11] (Parent is DashCommunicator(0x134736e0), parent's thread is QThread(0x7364f80), current thread is DashCommunicator(0x134736e0) [11:11] woot [11:15] Saviq, we might want to whitelist some of the most frequently used entities [11:16] Saviq: does our unity8 ci has real phone hw backing any of the tests? [11:16] Saviq, the Text QML element support only a very few [11:16] kgunn, yup, autopilot ones [11:16] Saviq: so do we run AP on both wily and viivd ? [11:17] kgunn, yes [11:17] wow...awesome [11:17] ltinkl, and what to do with the rest? I'd say we need to either support all entities, or none, and only allow a subset of tags, stripping out the rest [11:18] kgunn, an example report back from jenkins https://code.launchpad.net/~unity-team/unity8/autopkgtests/+merge/271609/comments/684368 [11:19] although that one's a bit pessimistic [11:19] :) [11:19] some device issues [11:19] this one's a lil' better https://code.launchpad.net/~aacid/unity8/clazy_run/+merge/271612/comments/684356 [11:19] just unstable, not failed [11:20] tsdgeos has a branch to fix qml tests, autopilot will be better once qtubuntu-sensors comes back on the image [11:20] so we're getting where we want to be [11:20] all good === alan_g is now known as alan_g|lunch [12:26] greyback_, yay, a successful qtmir CI run for both wily and vivid :) [12:27] Saviq: great stuff [12:41] kgunn: tedg: so how do we go about landing that branch? === pixel is now known as Guest76074 [12:51] tsdgeos, it's not a unity8 branch is it? I just added a silo with your fix [12:51] Saviq: no it's ual [12:51] hi, i'm on wily/unity8 and the terminal app fails to install (download or install fail) [12:52] the good thing is that it's working again with nvidia [12:52] https://code.launchpad.net/~ted/ubuntu-app-launch/lp1495871-unref-context/+merge/271322 or https://code.launchpad.net/~ted/ubuntu-app-launch/lp1495871-unref-context-15.04/+merge/271373 or both [12:52] Saviq: ↑ [12:57] oh and i get pages of QGridLayoutEngine::addItem: Cell (0, 0) already taken in unity8-dash.log while searching in ubuntu store [13:02] brb === alan_g|lunch is now known as alan_g [13:04] tsdgeos: I think just throw it in the same silo with other fixes [13:04] tsdgeos: Already got a code review from charles_ [13:05] cool === ChrisTownsend1 is now known as ChrisTownsend [13:23] tedg: so i should abandon that one silo i had for fd leak fix ? [13:24] sounds like you have another going with "other stuff" in it [13:24] I don't, but I assumed that tsdgeos did [13:24] i did? [13:24] :) [13:24] i don't have any silo [13:24] tedg: tsdgeos let's just land it in isolation then [13:24] WFM [13:25] ok [13:25] tsdgeos: I thought you had one with other fixes. [13:25] it's in silo 60 [13:25] NP, glad kgunn asked :-) [13:25] yeah misunderstanding from my side [13:25] sorry [13:25] tedg: i'm investigating another leak, i'm close to knowing where the error is [13:25] but still meh [13:25] tedg: tsdgeos i beat the shit out of it for app launch/close/phone calls etc....any other testing needed? === pixel is now known as Guest30724 [13:30] kgunn: I think that's probably more than the test plan. [13:30] kgunn: Did you make sure to do a legacy app and a click app? [13:30] kgunn: System settings is a good legacy app. [13:35] tedg: yes, did all of them [13:35] pretty much every app on the device [13:35] Cool [13:35] * tedg is now worried that kgunn has beat his Sudoku score [13:37] :) [13:42] so... can you guys install apps from ubuntu store on wily/desktop? [13:43] i know we're not outta the woods yet, but thanks alf tsdgeos & tedg for all the fd leak work [13:44] Guest30724: what exactly do you mean? are you asking about ubuntu touch apps ? [13:44] kgunn: i want to install the terminal app (the click has the x86 version) [13:45] kgunn: ubuntu store apparently downloads the click but then fails to install [13:45] i can't find the click anywhere ;> i can complile the terminal app and make a .desktop file but that is silly [13:47] i'm on a desktop using wily daily build + unity8 [13:49] kgunn: yes :D so yes, terminal-app (qml) not gnome terminal [13:54] woohoo silo 27 landed [13:54] Guest30724: how are you "using" unity8 on wily daily build? meaning that you are running the unity8-desktop-session-mir ? [13:55] kgunn: yep [13:55] hmmm, so good question... [13:55] it worked before [13:56] bregma: Saviq so if someones running unity8-desktop-session-mir how come apps can't be downloaded via store ? [13:56] oh...maybe it's just busted [13:57] most probably :D maybe the next update will fix it [14:00] kgunn, there's not a lot of apps in arch other than armhf [14:01] Saviq: yeah, but guest wanted terminal app which has a x86 version [14:02] and said it used to work...so my guess is it broke [14:02] kgunn, would have to check, hey ChrisTownsend did you manage to try unity8 in lxc? [14:03] thanks all :> [14:04] Saviq: Well, I looked at it a bit, but it's gonna take some work to get it resurrected. And I've been off working on the Libertine/Purtitine stuff. [14:05] ChrisTownsend, got it [14:05] Saviq: What I'd like to get working is to use the standard Wily LXC download template, install unity8-desktop-session-mir and modify the config to work as a LightDM login. [14:06] Saviq: However, it seems Trevinho has had success getting Unity 8 to run in a Wily LXC on a Mir-on-X session. [14:07] Saviq: So he may be a faster source of getting something up and running. [14:08] ChrisTownsend: yeah, I've not booted the full stuff, as I've just tested the basics, but it seems to do what it should :) [14:08] Guest30724, no, it never worked :) [14:08] Saviq, actually there's quite some multiarch apps now [14:08] however mir-on-x needs some work in the kb/mouse handling (i.e. it should grab, or you can mess with your host) [14:08] Trevinho: Really cool:) [14:08] Saviq, popey publishes everything in multiarch now (aka all core apps) [14:09] some things, not everything :) [14:09] and the multiarch click fails to install here on unity8/desktop as well :/ [14:10] yeah, i couldn't get that working either, i think there's apparmor bits missing maybe? [14:10] mzanetti: it worked before on a desktop next image [14:10] no [14:10] https://bugs.launchpad.net/ubuntu/+source/click/+bug/1396611 [14:10] yes [14:10] Ubuntu bug 1396611 in click (Ubuntu) "Can't install click packages with pkcon" [Undecided,Confirmed] [14:11] i had 2 terminals installed once, 2 diff versions [14:11] well, you can install it manually, but not from the store [14:11] one from ubuntu store and one from ppa [14:11] i know i can compile it :P [14:11] but i used to install it from the store [14:11] also the click can be installed manually [14:11] having a hard time to believe that [14:12] the very first window I painted on unity8 windowed mode was the store and I tried to install something [14:12] it failed, I reported that bug [14:12] and that bug is still open [14:12] i've installed all the apps from ubuntu store [14:18] mzanetti: what is the diff between willy desktop next and willy + unity8-desktop-session-mir [14:18] wily [14:18] hmm... don't know tbh [14:19] maybe there is something diff but i 101% remember beeing able to install apps from ubuntu store [14:19] maybe 102% [14:19] mines terminal-app zbird etc [14:21] on ubuntu next image [14:26] pstolowski, you can ask fginther for dual-release testing in ci [14:27] pstolowski, we already have it enabled for unity8, unity-api and qtmir [14:27] and it's working well [14:30] mzanetti: kgunn found the clicks in ~/.local/share/ubuntu-download-manager/Downloads [14:31] and some erros in syslog [14:31] interesting [14:31] thanks for help [14:44] Saviq, ack === dandrader_ is now known as dandrader === hggdh is now known as hggdh_ === hggdh_ is now known as hggdh__ === hggdh__ is now known as hggdh [14:52] josharenson, when I run unity8 on my desktop, via that lightdm unity8-mir session, unity8 seems to run in full-greeter mode. so I get the unity7 lightdm greeter followed by the built-in greeter inside unity8. what's the fix for having unit8 run in the correct mode in this situation? [14:53] dandrader: there is a bug filed for that and it will be resolved when the latest (and huge) slim greeter branch is merged [14:53] dandrader: essentially, greeter mode isn't implemented yet (outside of the branch I just mentioned) and once it is, you can just install the unity8-greeter package and it should "just work" [14:53] josharenson, this one https://code.launchpad.net/~josharenson/unity8/slim_greeter_real_lightdm/+merge/271614 ? [14:54] dandrader: yeah, and there is a "stable" version of that in ppa:josharenson/unity8-greeter [14:54] josharenson, but will unit8 work with the unity 7 greeter? [14:54] josharenson, or there's no point in supporting such a combination? [14:55] dandrader: humm, it should work fine, but since the shell is starting in full-greeter mode (by default) you will still see both the unity7 and unity8 greeters [14:56] dandrader: if you modify the right config files so that unity8 starts in "shell" mode, you should only see the unity7 greeter, but I haven't tested that [14:56] dandrader: but it technically should work [14:59] josharenson, so you have replaced the stock greeter with the unity8 one in your laptop? [14:59] dandrader: yes [14:59] josharenson, but then how do you select the session you want to log in. eg: unity 7, unity 8, gnome [15:01] dandrader: I haven't tried that yet. I think it should work, and its currently a bug that it doesnt [15:09] josharenson, but does the QML Greeter has UI code to allow selecting those different session types? [15:11] dandrader: humm good question, i thought I remembered seeing it on an ubuntu daily build a while back... but it could have just been the unity7 greeter, and I don't think I've seen it in the code... I assume, if its not there, that its something I'd have to get from design? [15:12] josharenson, yeah, I think so [15:12] josharenson, sounds like it indeed, I don't think it was ever thought about in that regard [15:13] dandrader: I'll try to look into that today so that we can give design as much time as possible [15:13] cool [15:18] tsdgeos, ping [15:18] pstolowski: hi [15:19] tsdgeos, hey, re your fix for preview leak - can you land unity8 part of the fix first, independent of shell plugin [15:19] ? [15:19] pstolowski: yes [15:20] tsdgeos, cool, then please go ahead [15:20] i mean "yes it can be landed" [15:20] not as "yes i will land it" :D [15:20] tsdgeos, but i cannot land shell plugin fix before unity8 part of it lands? [15:20] correct [15:20] oki [15:21] Saviq: can we land the leak-fix-prerequisite in that mini silo you're doing? [15:21] tsdgeos, i see some unity8 silos, can you use one of them to land it? [15:21] tsdgeos, yeah, I was almost asking that, but somehow didn't [15:21] tsdgeos, it's not unity8 is it? [15:22] Saviq: it's not the fd leak, but a leak we're fixing in previews for the dash [15:22] previewstacks and models where never deleted :D [15:22] now we need to keep a pointer to them [15:22] oups [15:22] since otherwise they'll be deleted too soon [15:22] tsdgeos, which branch/ [15:22] ? [15:23] it's qt's fault ;) [15:23] Saviq: https://code.launchpad.net/~aacid/unity8/keepPreviewStackAround [15:23] ack [15:24] tsdgeos, i'll land plugin part of the fix next week, after i land silo 24 (a new feature) [15:57] Saviq, so lp:unity-api/trunk-15.04 is not being used anymore [15:57] ? [15:58] dandrader, no, we've single trunks for unity8, qtmir and unity-api, dual landed into wily and vivid/overlay [16:07] Saviq, but we no longer target MPs at lp:unity-api/trunk-15.0 or lp:unity8/overlay [16:07] ? [16:07] dandrader, single trunks == lp:unity8, lp:unity-api [16:07] dandrader, just forget about the overlay branches [16:07] they're evil [16:08] ah, ok. I misread your "no" [16:08] the dual negation... === dandrader is now known as dandrader|afk [16:10] * greyback_ out [16:46] Saviq: i've never landed with bileto, once approved by QA..."publish" then "merge & clean" ? [16:47] camako: ^ ? [16:48] kgunn, me neither.. not sure [16:48] AlbertA ^ [16:53] kgunn: looks like it... [16:53] AlbertA: camako i just asked in ci-eng actually landers take it from there [16:54] they do publish (we don't have perms) and merge/clean is automagic [16:55] kgunn, yup [16:55] good to kno [16:55] you can m&c in case things fail, or to abandon a silo === dpm is now known as dpm-afk === alan_g is now known as alan_g|EOW [17:22] kgunn: is the compositor whats holding us back from "frosted glass"? Or is it more of a design issue? [17:31] mzanetti: still going to post to the unity blog today? === dandrader|afk is now known as dandrader [18:00] josharenson: we don't have any hard designs atm i think...and no one has done the work on the compositor to boot, so a little of both [18:00] kgunn: ack [18:27] mhall119, don't think so. but better ask josharenson, he's writing it [18:28] now that jenkins runs for vivid and wily, if something's wrong the jenkins comments look really depressing http://paste.ubuntu.com/12452366/ [18:28] mhall119: I'm working on final revisions and I'd like to post a video with it as well (there is a bit of a long story as to why the video isn't done, mostly related to my totally bricked nexus 10) [18:29] mhall119: so if it can wait until EOD monday, that would be best, but I could probably post it today w/o the video if necessary [18:45] josharenson: no rush, i prefer video [18:53] josharenson: not problem, post it when it's ready [18:53] mhall119: cool, will do [19:54] mterry: so the greeter is failing the qmltests because it seems to be loading parts of the lightdm (the integrated lightdm) and parts of the mock lightdm... It can successfully set the mock mode to "full", but then shows my real username instead of the mock usernames... I know that is super confusing, so hopefully this output helps clear up what I just said http://pastebin.ubuntu.com/12454487/ [19:55] mterry: I've triple checked all the cmake stuff and it looks ok to me... === dandrader is now known as dandrader|afk [19:56] josharenson, hrm [19:56] josharenson, I can give it a look [19:57] mterry: pushing most current changes to lp:~josharenson/unity8/slim_greeter_real_lightdm/ and I can send you built debs if you want as well [19:57] pushed* [19:58] mterry: its very odd, and im sure its something simple... when I run the tests on trunk/overlay, they pass just fine so its not an env issue [20:21] josharenson, gosh, my internet [20:21] josharenson, you still there? [20:21] mterry: yup [20:21] josharenson, I'm seeing an error "ASSERT: "uri == QLatin1String("PLUGIN_CLASSNAME")" in file /home/mike/Work/code/unity8/slim_greeter_real_lightdm/plugins/LightDM/plugin.cpp, line 62" [20:21] josharenson, PLUGIN_CLASSNAME shouldn't be in quotes if it's a symbol [20:22] josharenson, I also saw a warning "file:///home/mike/Work/code/unity8/slim_greeter_real_lightdm/tests/qmltests/Greeter/tst_Greeter.qml:403:9: QML Connections: Cannot assign to non-existent property "onLauncherOffsetChanged"" [20:22] mterry: yeah I see that same warning in trunk [20:22] * josharenson recompiles w/ symbol changed [20:24] no dice [20:27] josharenson, did you see the assert before? [20:27] mterry: I didn't but that doesn't mean it wasn't there [20:27] josharenson, with the change, you get a compile error because PLUGIN_CLASSNAME isn't be defined [20:27] ..... wondering why it built [20:27] trying a clean build [20:30] mterry: for the record, i know the rest of the tests are broken, im just trying to get make testGreeter working first [20:32] josharenson, sure. I haven't tried others yet :) [20:34] josharenson, If you don't see the same assert after a clean build... one of our environments is affecting the test [20:35] mterry: yeah I removed the quotes, did a clean build, and I see no errors [20:36] Humph... I'll try a clean build without the quotes... [20:36] josharenson, you also didn't see the assert without the change? [20:36] mterry: I don't think so... [20:36] josharenson, your top commit is 1915? [20:37] mterry: yes [20:38] mterry: just for sanity http://pastebin.ubuntu.com/12455825/ [20:38] josharenson, yup [20:39] josharenson, and I see: [20:39] /home/mike/Work/code/unity8/slim_greeter_real_lightdm/plugins/LightDM/plugin.cpp: In member function ‘virtual void IntegratedLightDMPlugin::registerTypes(const char*)’: [20:39] /home/mike/Work/code/unity8/slim_greeter_real_lightdm/plugins/LightDM/plugin.cpp:62:24: error: expected primary-expression before ‘(’ token [20:39] Q_ASSERT(uri == QLatin1String(PLUGIN_CLASSNAME)); [20:39] ^ [20:39] /home/mike/Work/code/unity8/slim_greeter_real_lightdm/plugins/LightDM/plugin.cpp:62:48: error: expected primary-expression before ‘)’ token [20:39] Q_ASSERT(uri == QLatin1String(PLUGIN_CLASSNAME)); [20:39] ^ [20:39] plugins/LightDM/FullLightDM/CMakeFiles/FullLightDM-qml.dir/build.make:146: recipe for target 'plugins/LightDM/FullLightDM/CMakeFiles/FullLightDM-qml.dir/__/plugin.cpp.o' failed [20:40] josharenson, I'm compiling on wily... but I doubt that would affect the compilation [20:40] mterry: ill go make some horrible syntax error and see if it fails [20:40] im on overlay [20:41] mterry: yeah I broke it [20:41] weird [20:42] guh [20:42] let me try on my overlay machine [20:43] mterry: I'm sure this doesn't matter, but I'm using gcc [20:43] (I use clang sometimes) [20:46] me too [20:47] mterry: really weird... if I, for example, remove the ";" at the end of the line, compilation fails... if I change PLUGIN_CLASSNAME to some nonsense, it builds just fine [20:48] ... that's not right === dandrader_ is now known as dandrader [20:52] mterry: for the sake of completeness, it builds just fine w/ clang too [20:52] i love when I can't get things to break when the are supposed to [20:59] mterry: I wonder if I, somehow, don't have QT_NO_DEBUG set [20:59] rather, unset [21:00] josharenson, I keep losing irc [21:00] last thing I said: [21:00] josharenson, same problem on overlay [21:00] assert, then fix, then compile error [21:00] 01:59:38 PM) josharenson: mterry: I wonder if I, somehow, don't have QT_NO_DEBUG set [21:00] (01:59:44 PM) josharenson: rather, unset [21:00] josharenson, it's crazy to me that you can change PLUGIN_CLASSNAME to gibberish and it compiles? [21:01] mterry: yes, which is why I think that line might not be getting compiled [21:01] ah right... [21:01] looking into QT_NO_DEBUG, but I might have to leave for a bit [21:01] apparently I've been at panera for too long [21:01] :) [21:02] Nothing beats Starbucks for remote work. Haven't found anyone as accomidating [21:02] mterry: duly noted... I know a million good coffee shops in seattle, but I'm in San Diego right now, so kind of foreign territory [21:03] anyway, back in a bit [21:03] thanks for the help