=== chef_ is now known as Guest34025 === Guest34025 is now known as `Chef === marcusto_ is now known as marcustomlinson [08:59] pstolowski: so the audio stuff landed \o/ [09:00] tsdgeos, :) [09:00] tsdgeos, that means beer in austin [09:00] pstolowski: is there any update we need to do for filters because of the landing [09:00] version numbers or something? [09:00] tsdgeos, probably, will let you know soon [09:49] Saviq: so the big silo is next, right? [09:49] tsdgeos, yes, DPR's under QA and the big one's next [09:55] tags infection! [09:55] * tsdgeos deletes [09:56] http://paste.ubuntu.com/14468126/ === boiko__ is now known as boiko [12:12] greyback, interestingly, qtmir tests failed in our CI on my first trials, in QML cache retention tests, thoughts? [12:13] Saviq: nothing obvious. Any error messages I can look at? [12:18] greyback, the failures just said the dir wasn't there, what should have created it? [12:20] Saviq: I don't know. I thought the code would, but maybe not [12:20] greyback, started the build again in https://unity8-jenkins.ubuntu.com/job/run-commands/223/console [12:20] greyback, it might be just that XDG env isn't set or something [12:21] will see if phone can recover if I delete ~/.cache/QML/Apps [12:21] yep, it can [12:28] greyback, the above job finished, there's build logs in there [12:29] /food [12:29] tsdgeos, so there's nothing holding up https://code.launchpad.net/~aacid/unity8/nodda/+merge/280814 besides https://code.launchpad.net/~aacid/ubuntu-ui-toolkit/xvfb_pixels_per_mm/+merge/282149 ? [12:30] Saviq: I suspect the QV4_ENABLE_JIT_CACHE=1 env var missing [12:30] greyback, would've failed in any other package builds then, wouldn't it... [12:32] /methinks it's rather something like $HOME missing or so [12:33] QByteArray path(qgetenv("HOME") + QByteArray("/.cache/QML/Apps/") + (qgetenv("APP_ID").isEmpty() [12:33] yes, it needs $HOME [12:35] meh, should use XDG instead [12:37] or well, Qt's path bits, even [12:37] QStandardPaths I mean [12:39] not to mention it pollutes $HOME this way [13:26] dandrader: afaik no [13:26] dandrader: there's a few failing tests, but i think that they are fixed by the mega silo === dandrader is now known as dandrader|afk [13:38] exit [14:30] Heh, this is cute: "Tofu (豆腐) is Japanese jargon for unicode replacement character "�" (U+FFFD) often displayed as replacement for unassigned or unknown characters." [14:37] mzanetti, I just noticed you couldn't make the lockscreen meeting today, want me to reschedule? [14:39] mterry, I am totally confident you can handle it. but if you'd like me to participate, then yes [14:40] mzanetti, naw it's cool, should be simple questions [14:40] yeah [14:48] cimi: pstolowski has a test scope you can use to test the other two fitler widgets you did not review [14:48] cimi: could you have a look? === vrruiz_ is now known as rvr [15:06] tsdgeos, sure [15:06] tsdgeos, I am doing sth else for gerry atm [15:10] k [15:21] tsdgeos, ltinkl, http://pastebin.ubuntu.com/14469671/ [15:21] Saviq: hmmm, what branch is that? the megasilo? [15:21] tsdgeos, subset of mega silo with just fixes https://requests.ci-train.ubuntu.com/#/ticket/854 [15:21] as we're in feature/string freeze [15:22] that's on xenial btw [15:23] Saviq: the card creator failure is weird, none of those MRs touch the cardcreator file, no? [15:23] * tsdgeos triple checks [15:23] Saviq, in mtg, gonna have a look at it after [15:24] Saviq: is there a lp branch with everything? [15:25] * tsdgeos quadruple checks [15:26] tsdgeos, as usual https://code.launchpad.net/unity8 [15:26] Saviq: https://code.launchpad.net/~lukas-kde/unity8/fixLauncherDismiss/+merge/282031 requires https://code.launchpad.net/~mzanetti/unity8/launcher-updates that is not in the silo, you aware of that? [15:27] tsdgeos, oh right, pulling that out [15:27] or rather, un-rebasing [15:28] Saviq, passes locally here (the launcher dismissal testú [15:29] ltinkl, can you please uncommit the merge with launcher updates and --overwrite, had to pull that one out [15:30] tsdgeos, maybe broken on trunk? [15:30] Saviq: think not, let me recheck [15:31] tsdgeos, fails reliably in the branch [15:31] oh, actually it is [15:31] http://paste.ubuntu.com/14469783/ this is trunk [15:31] at least here [15:31] Saviq, oh, you had to pull out the launcher-updates? [15:31] ltinkl, yeah, freeze [15:32] tsdgeos, yup [15:32] Saviq: i can provide a fix in a minute after standup [15:32] tsdgeos, ack [15:32] Saviq, ok, should be fine: https://code.launchpad.net/~lukas-kde/unity8/fixLauncherDismiss/+merge/281640 [15:32] Saviq, reverted the prereq [15:32] ltinkl, thanks === dandrader|afk is now known as dandrader [15:40] mterry, you need to update the AP test in the wizard change [15:41] mterry, http://pastebin.ubuntu.com/14469884/ [15:41] Saviq, ick right [15:43] ltinkl, http://pastebin.ubuntu.com/14469902/ reliably here [15:43] ltinkl, note that's the whole branch [15:43] lp:~ci-train-bot/unity8/unity8-ubuntu-xenial-landing-030 [15:43] in case that has an impact [15:44] Saviq, been a while since I touched the AP tests... are we close to those passing? Or do punks like me keep breaking them? [15:44] Saviq, with this silo https://requests.ci-train.ubuntu.com/#/ticket/854 ? [15:44] mterry, that's the only one broken [15:44] Saviq: https://code.launchpad.net/~aacid/unity8/fix_card_creator_test/+merge/282187 [15:44] ltinkl, yes [15:44] tsdgeos, tx [15:44] mterry, I run them every landing [15:44] Saviq, ok, on it [15:44] Saviq, but they were broken in jenkins, right? [15:45] mterry, kinda, were fine recently, finicky xenial though [15:45] +on [15:53] ltinkl, yeah, lp:~ci-train-bot/unity8/unity8-ubuntu-xenial-landing-030 fails on that test here [15:54] ltinkl, could be mouse touch update [15:54] Saviq, yup.. think so too... [15:54] Saviq, this one right https://code.launchpad.net/~dandrader/unity8/updateMouseTouchAdaptor/+merge/280718 [15:56] Saviq, here's where it fails: http://paste.ubuntu.com/14470063/ , line 15 [15:57] * Saviq bisects [15:58] oh right, it just crashes otherwise [15:59] Saviq: run without xvfb [15:59] that'll remove the crash [15:59] ac [15:59] k [16:02] ah and that test isn't there before :P [16:03] Saviq, mine? yeah that comes with the branch [16:07] tsdgeos, https://code.launchpad.net/~dandrader/unity8/nodda/+merge/282193 [16:08] dandrader: that's a weird diff :D [16:08] indeed [16:08] tsdgeos, yeah, was about to say that. web diff totally messed up [16:09] dandrader: http://bazaar.launchpad.net/~dandrader/unity8/nodda/revision/2116 is the important part, right? [16:09] tsdgeos, might be because took trunk, merged your branch and then added mine on top [16:09] tsdgeos, yes [16:09] k will have a look [16:11] tsdgeos, you can bzr merge -c 2116 [16:11] yep [16:23] Saviq, how do you run the tests? they don't seem to come close to passing on my xenial desktop. do I need to do it on the phone? [16:23] (the autopilot tests) [16:23] mterry, yes, phone only [16:24] humph [16:24] mterry, flash phone, citrain, install unity8-autopilot (might need to drop /etc/apt/preferences.d/*), run them [16:24] oh yeah [16:24] stop unity8 [16:24] phablet-test-run wasn't reliable for me recently [16:26] Saviq, I also noticed that the CMakeLists.txt only runs the unity8.shell tests in its one declare_autopilot_test() command. Do those never get used? [16:26] mterry, don't think so, we always use "autopilot run unity8", since we run from packages [16:27] Saviq, hrm. Meaning we don't use mocks? Meaning that we can't control the environment. That's probably why it fails for you -- you have a password set and it's skipping the password screen as a result [16:27] mterry, the cmake target was for running on your host, which is bitrot these days, and we should probably not support that at all since it will maybe run half the tests [16:27] mterry, sure it can use mocks [16:27] Saviq, just checking, do you want me to do something with the launcher_dismiss test failure at this point? [16:28] ltinkl, just finding where it breaks [16:28] will let you know when I do [16:28] Saviq, kk, thanks [16:28] mterry, tests themselves set QML2_IMPORT_PATH as appropriate, maybe that test is missing that bit [16:28] Saviq, yup I see other autopilot tests doing it, ok. will fix [16:28] tx [16:29] w00t, successful unity8 build in Jenkaas [16:30] Saviq, I can try merging lp:~dandrader/unity8/updateMouseTouchAdaptor into mine and see if it breaks [16:30] ltinkl, will know soon [16:30] ok [16:31] greyback, I think I did something bad with your DPR :) [16:35] Saviq, hmm, passes here as well [16:36] Saviq, but I'm tempted to do this: remove my test function since there's already one, practically the same: test_dragLeftEdgeToRevealLauncherAndTapCenterToDismiss() [16:38] ltinkl, still, we should understand what happened [16:38] Saviq, yeah... [16:40] ltinkl, it was launcher updates [16:40] Saviq, ack, so it should be ok now [16:40] ltinkl, maybe in conjunction with one of the other branches [16:41] Saviq, good to know :) also verified it works fine together with lp:~dandrader/unity8/updateMouseTouchAdaptor [16:41] ltinkl, the existing test only does it once, right? and yours did twice? [16:41] Saviq, yup, my test does it twice (specifically for the bug report) and in a slightly different way [16:42] ltinkl, ok, so leaving it in [16:42] Saviq, yup, definitely [16:59] mzanetti, hey, I’m wondering about how unity8 behaves with regards to multi-window apps, can you enlighten me? [16:59] dandrader: merged in [16:59] tsdgeos, ok, thanks [16:59] oSoMoN, what in particular? [16:59] oh, multiple surfaces per app [16:59] dandrader: ty [17:00] oSoMoN, we're working on it. we have a branch in the queue that makes them not crash [17:00] mzanetti, yes, is that supported at all, if so what constraints does the shell enforce, … [17:00] and daniel is working on another one that makes it actually work [17:00] mzanetti, is there a rough ETA? [17:00] dandrader, what's the status on multisurface apps? oSoMoN is aksing ^ [17:01] you took that card from me in the last sprint. haven't checked on it since the hols [17:01] and is there a bug report to track progress? [17:01] well, I guess I could say OTA-10 if all goes well [17:01] oSoMoN, there is our trello board [17:01] that sounds good enough [17:01] mzanetti, oSoMoN working on it. but long road ahead [17:02] ok, scratch ota 10 then :D [17:02] but well, 11 or so still seems realistic I'd say, no? [17:02] oSoMoN, https://trello.com/c/F7aFIPq4/209-13-support-multiple-top-level-windows-of-applications [17:03] mzanetti, you asking me? I don't know the OTA dates.. [17:03] you don't? you should :D [17:03] well, we have frozen OTA-9 right now [17:03] that means roughly 6 weeks to the next [17:03] mzanetti, I'm not a manager / team lead :D [17:04] we need a tool to map OTA numbers to estimated dates :) [17:04] well, you sure notice a huge traffic jam on landings every 6 months [17:04] just count them, we're at 9 atm [17:04] 6 weeks, sorry [17:04] mzanetti, dandrader: so when fully implemented, what kind of constraints will the shell be enforcing, if any? will apps be allowed to have multiple windows on a phone? [17:04] oSoMoN, yes, for the second question [17:05] interesting [17:05] oSoMoN, I think we're going towards the direction that yes, it would be allowed on the phone, but there's many design questions open still [17:05] guidance for a first step is to just hide them in staged mode [17:05] mzanetti, oSoMoN app lifecycle, screenshotting and prompt surfaces get more complicated [17:05] but the more we discuss it, the more it makes sense to have multiple browser windows on a phone too [17:06] oSoMoN, so prepare for everything but the first window is hidden in staged mode for a start. I will loop you in if new info comes up [17:06] mzanetti, I think it will be the first phone with multiple browser windowes [17:07] mzanetti, perfect, thanks [17:07] dandrader, afaik android does that too now [17:07] mzanetti, really!? [17:07] mzanetti, with chrome? on what android version? [17:07] given the whole phone is a tabbed ui, it seems quite odd to have tabs within a tab tbh :D [17:07] I don't know, Saviq told me [17:08] mzanetti, other windows would be hidden but not closed, right? [17:08] yeah, hidden I'd say [17:08] and restored when in windowed mode I guess [17:09] mzanetti, is "the first window" the first ever opened, or the most recently focused one? [17:09] yup, Chrome on android does multiple windows (even spawns another instance imo) [17:09] ltinkl, they don't do tabs at all any more [17:09] which /me likes btw [17:09] Saviq, yup [17:10] dandrader, ^^ [17:10] would much rather see all tabs separate in right edge than have to go through right + bottom edges [17:10] * oSoMoN wouldn’t mind that either, that would simplify quite a bit the browser’s code [17:11] oSoMoN, don't think that's clear enough yet (re: first or last focused window) [17:11] ok [17:11] oSoMoN, could be a UX question as well [17:12] oSoMoN, but if a browser window is focused, we should probably not close that on [17:12] e [17:12] oSoMoN, so that might be a good starting point, unless we find a problem with that [17:12] and is there gonna be some sort of signal when switching from windowed to staged (or vice-versa)? [17:12] not even sure that’d be useful, just thinking out loud [17:13] oSoMoN, there must be, I think, because you need to collect the tabs [17:13] mzanetti, hmm, maybe it's my phone skin on top of android that doesn't enable that multi window browser. will check my Nexus 10 later [17:13] oSoMoN, quite likely we won't be killing your windows even, only display the last one and let you know to kill the other ones [17:14] as closing windows "as usual" might have implications on user data [17:14] Saviq, what we would do when all other windows are hidden needs to be discussed with design, but indeed whatever we’ll do we would probably need to get notified [17:15] Saviq, I thought phone/staged mode would have multi-windows per app as well.... [17:16] too many open design questions with that atm [17:16] but yes, we will probably aim for that in the long run [17:16] dandrader, nope [17:16] at least don't build roadblocks :) [17:16] another (slightly related) question: does unity8 enforce a single instance for a given app, or is that up to the app to implement its own singleton mechanism? [17:16] Saviq, so far I've been workgin with the information that phone mode would also support multi-window apps [17:17] oSoMoN, qtmir enfoces that, yes [17:17] ubuntu-app-launch does, rather [17:17] or well, both do [17:17] dandrader, as things stand today, one window per app on phone [17:25] mterry, let me know if you think have the AP test fixeded so I can rebuild the silo [17:25] Saviq, I maybe pushed a fix -- was going to wait for jenkins to build it for me [17:25] Saviq, (before testing to confirm) [17:26] mterry, looks legit [17:27] * Saviq rebuilds then [17:28] are https://wiki.ubuntu.com/Unity/LauncherAPI#Static_Quicklist_entries supported in unity8’s launcher? [17:29] ltinkl, ↑ [17:31] oSoMoN, Saviq https://code.launchpad.net/~lukas-kde/unity8/desktopFileActions/+merge/276408 [17:31] right [17:32] oSoMoN, work in progress, but probably not until we support multi instance/window apps [17:34] thanks === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [19:00] Saviq, I'm back from lunch, which silo are you building in these days? I want to test my wizard fix (jenkins still didn't run) [19:03] ah silo 30 [20:07] Saviq, so I ran the wizard autopilot tests again, and silo 30 passes those [20:08] mterry, yup, just did that myself [20:08] Saviq, ran the whole suite just to see, and I got 8 failures, where they couldn't find 'MainView' widget [20:08] Saviq, is that new? [20:08] mterry, hmm [20:08] Ran 76 tests in 2231.028s [20:08] FAILED (failures=1) [20:08] but that one failure I expect, is because I have too many items on my launcher [20:09] mterry, that's on rc-proposed? [20:09] Saviq, yes [20:09] mako [20:10] mterry, I'm on rc-proposed@krillin and things are fine, just finishing up on devel-proposed@mako [20:11] mterry, I would maybe expect that on devel-proposed, not sure we've made sure they work with Qt 5.5 [20:12] Saviq, this was not a pure clean image beforehand -- I don't remember doing anything weird, but maybe I had [20:12] Saviq, if it worked for you on krillin, this is probably just noise then [20:16] Ran 76 tests in 3238.810s [20:16] OK [20:17] mterry, ↑ devel-proposed@mako [20:17] Saviq, nice [20:17] so actually better than I expected [20:17] not sure why that much slower, though [20:22] Saviq, one was on krillin and one on mako yeah? [20:22] mterry, yes [20:22] I thought krillin was faster, so that might make sense [20:22] but mako should be faster I'd think [20:22] is it? [20:22] right, it may be because less pixels [20:22] Saviq, I had a memory of thinking mako was our worst at one point [20:22] greyback, dandrader, are we ok with the TODO for bug #1527737? the delay to launch is significant, any reason why we can't show the app starting straight away? [20:23] bug 1527737 in qtmir (Ubuntu) "Apps do not start if restarted quickly after closing" [High,In progress] https://launchpad.net/bugs/1527737 [20:23] Saviq, it's already like that [20:23] Saviq, so the fix is not regressing [20:23] dandrader, or worse? [20:23] Saviq, it's just stating the problem clearly [20:24] Saviq, worse what? [20:24] dandrader, when it crashes [20:24] dandrader, well, ok, was just thinking since you touched that, could've made it good straight away, but then you're working on rehauling the whole thing anyway