=== maclin1 is now known as maclin === andyrock_ is now known as andyrock [08:02] hi guys, 1 question I have noticed a bug in unity 7 in ubuntu 14.04 lts whenever i launch an app in the recent list, the icon disappears from the unity search window while the app(if I press super i can see the icon disappear) and reappears short after app closes [08:02] may I be assigned to fix this recent list bug? === davmor2_ is now known as davmor2 [09:27] tsdgeos, hey, looks like LazyImage still not good enough https://requests.ci-train.ubuntu.com/static/britney/vivid/landing-019/excuses.html - could you have a look please, now that we've Pass elsewhere it's going to be trouble :) [09:27] ok, i'll have a look [09:29] Saviq, we should try the other branch too [09:30] Saviq, https://code.launchpad.net/~cimi/unity8/fix-lazyImage-test-flakyness/+merge/277459 [09:30] Saviq, that one was before michael told me to use waitForRendering and simplify everything [09:31] cimi: honestly, first he todl you to not have the signal and just to find the object on the test [09:32] since the signal is not used for anything else [09:32] then you fell into the trap of removing everything and adding a waitForRendering [09:32] that's a trap! [09:34] iirc I couldnt find the animation because it was inside the transition, or I had issues with running property [09:34] so I went for good old signals [10:01] tsdgeos, Saviq I'm having a second look too [10:02] tsdgeos, if you have any idea/patch that i can try, I can reproduce the original bug on my desktop pc [10:02] oki, was finishing something for patty, will have a look now [10:02] cimi: so with https://code.launchpad.net/~cimi/unity8/fix-lazyImage-test-flakyness/+merge/277459 you can not reproduce? [10:03] tsdgeos, flakyness is fixed for me with both branches [10:03] tsdgeos, I can reproduce the original flakyness [10:04] jenkins always flaky - way to powerful! :) === marcusto_ is now known as marcustomlinson [10:26] cimi: the "transitionCount: 4" is "a bug"? [10:26] i mean a bug in the current code [10:27] why would it transition twice? [10:34] tsdgeos, that's exactly why I am having a look again [10:34] tsdgeos, when I proposed the fix, that was what I did in an afternoon iirc, thought it was more critical [10:36] your branch still fails for me [10:36] http://paste.ubuntu.com/14848794/ [10:36] oh crap [10:37] wait i left the waitforrendering, maybe i need to remove it [10:40] tsdgeos, I am wondering if we need a better cleanup() [10:40] tsdgeos, from my understanding, jenkins is so fast that the animation doesnt start [10:42] cimi: that makes no sense, why would your code fix it then? [10:43] tsdgeos, which code, former branch or the last one? [10:43] https://code.launchpad.net/~cimi/unity8/fix-lazyImage-test-flakyness/+merge/277459 [10:44] tsdgeos, that is right too, so why would it fail on a faster pc? [10:45] i don't know [10:45] jenkins is usually dead slow [10:45] not faster [11:00] tsdgeos, that's what I thought, but on my slower machine I cannot reproduce, while I can with the faster one === lool- is now known as lool [11:54] Saviq: cimi: my variant https://code.launchpad.net/~aacid/unity8/more_stable_lazy_image_test/+merge/284591 [11:54] it's similar to cimi's but withotu changing the lazyimage code itself [11:55] tsdgeos, that looks better, however I am trying to see if we can still add waitForRendering [12:02] dednick, forgot, please merge trunk qtmir test refactor [12:02] Saviq: k [12:03] dandrader, you might wanna remerge your bits on top of ↑ [12:04] dandrader, also, https://code.launchpad.net/~dandrader/qtmir/appRestart-lp1527737/+merge/281701 needs rebasing [12:04] Saviq, you planning on merging appRestart-lp1527737 ? [12:08] dandrader, at some point, I hope ;) [12:08] dandrader, wanted to add to our next silo to be able to test again and maybe give you more data [12:09] ltinkl, hey, I set up GECOS on the phablet user, but greeter still shows "phablet", do you know if that's expected? code says "realName" on the lightdm model, so I thought it would use the right thing? [12:09] or is realName stored somewhere else? [12:11] hmm doesn't seem to work on my laptop either, so we might need to change something after all [12:14] Saviq, dunno what the greeter shows... josh should know [12:16] hmm I wonder if it should use displayName or so [12:16] hmm no realName seems to be the right role [12:18] Saviq: done. [12:18] dednick, tx [12:19] tsdgeos, huh, here's a question: http://bazaar.launchpad.net/~lightdm-team/lightdm/trunk/view/head:/liblightdm-qt/usersmodel.cpp#L226 any idea how to get to Qt::DisplayRole in QML? [12:19] it'd have to be set in setRoleNames, wouldn't it [12:20] hmmm [12:20] think so [12:20] Saviq, it doesn't have to I think [12:20] Saviq, it's modelData in QML [12:20] ltinkl, is it? [12:21] you can try with [12:21] (*this)[Qt::DisplayRole] = "display"; [12:21] that is part of DefaultRoleNames [12:21] but i think that gets overwritten on setRoleNames [12:21] let me check [12:22] Saviq: yeah from looking at the code i'd say it gets overwritten, but give "display" a quick try [12:22] ack [12:23] tsdgeos, it gets initialized with the default ones [12:26] ltinkl: sure, but it gets overwriten on setRoleNames [12:27] ltinkl, modelData is not defined [12:27] bummer [12:28] tsdgeos, "display" seems to work actually, looks like lightdm is b0rked [12:28] tsdgeos, Saviq I'm curious to see if this one will pass too https://code.launchpad.net/~cimi/unity8/more_stable_lazy_image_test/+merge/284595, it's a mix of both branches [12:29] cimi, you'll ~know soon enough [12:29] dumb comma [12:29] * cimi goes groceries 5 mins I'm back [12:29] cimi, prereq? [12:29] Saviq, yeah I can [12:30] Saviq, https://code.launchpad.net/~cimi/unity8/more_stable_lazy_image_test/+merge/284596 [12:52] Saviq, tsdgeos: thinking about how to expose the keymap from the surface as a property... is a QPair accessible from QML? [12:52] * ltinkl doesn't think so === alan_g is now known as alan_g|lunch [13:10] ltinkl, other than some primitive thing like an array, custom object seems to be the only sane choice? === ssweeny` is now known as ssweeny [13:10] Saviq, yeah... I went with 2 read-only QString properties (layout+variant) and a setter that takes both [13:11] Saviq, not that we strictly need it now but once we'll want to make the indicator work, we must have a way to query the current surface's keymap [13:11] ltinkl, ack [13:12] Saviq, Mir doesn't provide it, it only has a setter... :/ [13:20] ltinkl, you'll need to rewrite history of your kbd branch https://ci-train.ubuntu.com/job/ubuntu-landing-051-1-build/42/console [13:20] start with Daniel's initialSurfaceGeom and put the diff on top, please [13:20] you'll need to --overwrite, but there's only so much we can do [13:20] when we get into criss-crosses [13:20] Saviq, yeah, working on that (and also adding a unity-api branch into the mix) [13:21] tx [13:25] dandrader, https://code.launchpad.net/~lukas-kde/unity-api/kbdLayout/+merge/284602 [13:25] dandrader, I'll make the changes to qtmir+unity8 in a minute [13:29] ltinkl, ok [13:30] dandrader, https://code.launchpad.net/~lukas-kde/qtmir/kbdLayout/+merge/284466 should be good too now === dandrader is now known as dandrader|afk [13:37] mterry, moarning [13:37] mterry, got bug #1536714 for you on the lightdm side [13:37] bug 1536714 in lightdm (Ubuntu) "Fix the user login experience on the greeter" [Critical,Triaged] https://launchpad.net/bugs/1536714 [13:38] mterry, never gives up anything other than username it seems [13:40] Saviq, ok will look [13:43] cimi: if you're removing the whole signal spy, shouldn't you remove it altogether? [13:46] dandrader|afk, https://code.launchpad.net/~lukas-kde/unity8/kbdLayout/+merge/284606 === alan_g|lunch is now known as alan_g === dandrader|afk is now known as dandrader [14:03] tsdgeos, yeah but you added it for the cleanup no? [14:03] cimi: ah right [14:03] cimi: still failed here [14:03] though [14:03] tsdgeos, I dislike the transitionCount [14:03] tsdgeos, which one failed? [14:04] the one i put in the MR [14:04] basically the same that fails all the time [14:04] the waitforrendering is not going to help there [14:05] you're basically making it fail more and not less [14:05] if the waitforrendering takes some time [14:05] the transition will start and finish and you'll still be in waitforrendering [14:05] ok [14:06] cimi: if you dislike the transition count another way is checking for the state and that the animation is finished [14:06] that will have the same effect [14:07] * cimi looks in 10 mins [14:10] Saviq, rebased lp:~dandrader/qtmir/removeUselessClass and lp:~dandrader/qtmir/initialSurfaceGeom [14:10] dandrader, oh? those were merging fine? [14:10] Saviq, you've marked that lightdm bug for 9.5? that seems ambitious [14:10] Saviq, not a hard change all around, but isn't 9.5 real soon now? [14:10] mterry, it was marked 9.5 before [14:10] Saviq, ah ok [14:10] mterry, but if we can make it ;) [14:11] Saviq, hah I'll start now [14:11] mterry, well, yeah, later today [14:11] ltinkl, you'll be interested ↑↑ dandrader rewrote history of his branches again ;) [14:11] ltinkl, I'm working on another maybe-9.5 thing right now, dandrader is reviewing the kbdLayout branches? [14:12] Saviq, this silo is turning into a madness... [14:12] mterry, yeah looks like it :) I'm fixing the mock + test atm [14:12] ltinkl, yeah, thanks bzr [14:21] mterry, not really reviewing them. just pointed out some missing "paperwork" [14:27] ltinkl, ^ [14:33] mterry, dandrader: you guys decide :) [14:34] ltinkl, I'm happy to review it, but I'm working on something else this morning. Can get to it later [14:34] mterry, sure, no rush I guess [14:36] ltinkl, mterry, yeah, won't land for OTA9.5 anyway so there's less pressure [14:48] dandrader, the 2 separate QString properties are there really only because exposing a QPair isn't possible, but they act together [14:49] dandrader, see the test_setKeymap(data) in u8/tst_DesktopStage.qml [14:53] ltinkl, I know they act together. but when you expose them as properties, you should do it the whole way [14:53] dandrader, by which you mean 2 separate signals? [14:56] ltinkl, you can setKeymap("x+y") then setKeymap("x+z"). so variant changes but layout does not. But if you say that Qt has no issues using this two-parameter signal as the notify one then it's ok [14:56] dandrader, yeah, the signal doesn't have to match [14:57] dandrader, you could as well have a keymapChanged() without any parameter for both properties [14:58] dandrader, but signalling both at once with the same signal makes writing the tests (and maybe the indicator in the future) much easier === dandrader is now known as dandrader|lunch === dandrader|lunch is now known as dandrader [15:31] dednick, got a quick one for you: https://code.launchpad.net/~dandrader/qtmir/updateMirVersion/+merge/284629 [15:32] dandrader: approved [15:32] thanks [16:35] Mirv, hey, https://requests.ci-train.ubuntu.com/#/ticket/939 seems to have failed to build, needs bumping symbols around? [16:42] cimi: i still don't understand what you don't like about my branch for lazy image tests [16:50] tsdgeos, you said it failed, and I wanted to get rid of that signal counts [16:50] cimi: yours failed [16:50] mine works [16:50] the signal counting is not terrible [16:50] there's 3 state changes so the signal count is 6 [16:51] there's others ways to fix ait for the same, but i don't see why counting the state changes is that bad [16:51] tsdgeos, you disliked my original branch when I added the transition count https://code.launchpad.net/~cimi/unity8/fix-lazyImage-test-flakyness/+merge/277459/comments/702293 [16:52] i did [16:52] and i still dislike the 4 [16:52] I thought we wanted to have a smarter approach [16:52] the 4 is a bug on the original code most certainly [16:52] the 6 is fine [16:53] it goes default -> loading -> error [16:53] yeah I saw [16:53] well, let's approve yours then [16:53] at least we dont have changes in LazyImage.qml [16:53] while tests pass [16:53] tsdgeos, you ok with that? [16:54] i am [16:55] tsdgeos, cimi will add to silo 19 tomorrow [16:56] k [16:59] +, [17:07] Saviq: any chance we get https://code.launchpad.net/~aacid/unity8/preview_audio_playlist/+merge/284624 for the MWC image ? seems patriciadavila/PM wants it there [17:08] tsdgeos, on that branch, I noticed you do on purpose two for cycles for the playlist, first from index to end of the repeater then you start again from the index 0 [17:09] tsdgeos, is this to have the playlist looping ya? [17:09] cimi: no, this is to have the playluist starting in the correct index [17:10] tsdgeos, yeah I meant that, it starts from where it should then adds the rest of the elements [17:10] ok [17:10] then yes is for that :) [17:11] tsdgeos, it's the only logic part you changed in the real code, so i wanted to ask why that and not a single for :) [17:14] tsdgeos, we can get anything for MWC, we'll be demoing rc-proposed from two-three weeks from now [17:14] Saviq: ah, awesomeness [17:18] tsdgeos, Saviq anyway that audio playlist, just approved [17:18] ack [17:18] I'm gonna approve the other flakyness too or we first try in silo 19 Saviq ? [17:18] * tsdgeos eods [17:18] * tsdgeos waves [17:18] bye [17:18] ops [17:18] too fast [17:20] cimi, approve away [17:22] this one didnt fail too though, maybe we can approve that one too https://code.launchpad.net/~cimi/unity8/more_stable_lazy_image_test/+merge/284596 I'll see with albert magnana [17:51] ltinkl, you asked in the qtubuntu kbdLayout branch, if it really needed to link with xkbcommon? Just look at jenkins failing. I also ran into it in a chroot build [17:54] ltinkl, and in the unity8 kbdLayout branch, you asked what I had in mind for one of the tests. In tests/plugins/AccountsService/client.cpp, we have some tests that just make sure "dbus -> AccountsService service -> our listener object" works for a few properties. And since we're using weird types in this case, it might be good to test the InputSources property === alan_g is now known as alan_g|EOD === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [19:17] Saviq: is there any way to test a Unity 8 session on my wily laptop? [20:32] ltinkl, I can't build your qtmir kbdLayout branch -- get errors about missing overrides for qtmir::MirSurface. "invalid new-expression of abstract class type ‘qtmir::MirSurface’" [21:47] The webpage http://unity.ubuntu.com/getinvolved doesn't load. [21:47] Reason: Error reading from remote server [21:56] mterry, still around? [21:56] Pback [22:01] ltinkl, i am [22:01] mterry, there's a new branch/MP: https://code.launchpad.net/~lukas-kde/qtmir/kbdLayout2/+merge/284643 [22:02] mterry, sorry, I had to start from scratch, bzr was too confused by the several pre-requisites [22:02] ltinkl, ah ok [22:02] mterry, this should build (at least it did in the silo) [22:04] mterry, there's also a (new) unity-api branch needed: https://code.launchpad.net/~lukas-kde/unity-api/kbdLayout/+merge/284602 [22:04] mterry, which dandrader already approved [22:04] ltinkl, yah saw that [22:10] mhall119, you may try, but wily is really not a target any more - you'd be better off with vivid+overlay or xenial [22:16] mterry, I added the explicit linking to xkbcommon in https://code.launchpad.net/~lukas-kde/qtubuntu/kbdLayout/+merge/284310 [22:37] Saviq: if I were to try on wily, what would be the best way? unity8-desktop-session-mir or unity8-lxc? [22:38] mhall119, lxc is unmaintained and known to not be working for a while, desktop session is it [22:52] well, that won't install because of unmet dependencies, is there a live ISO I can boot?