[08:08] Saviq: our ci stuff is again borked? [08:08] https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/1518/console ? [08:09] tsdgeos, lookin' [08:09] ugh, someone added a bad hook ERROR:pbuilderjenkins:H10strip_native_depends not found in hooks [08:10] at least for that job [08:11] it's similar to the other errors i've seen [08:23] kgunn: hmmm. I'm sure I saw an option to enable that in the system settings at some point [09:06] tsdgeos, dude, ETOOMANYPHONES :P [09:06] sorry ^_^ [09:15] Saviq: do we have more details on this one now? https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1378092 [09:15] Ubuntu bug 1378092 in unity8 (Ubuntu) "Dialer-app is getting sigstopped on greeter mode." [Critical,Triaged] [09:15] should it just be closed? [09:17] let me try and repro [09:20] mzanetti, yeah, fixed [09:20] Saviq: I really wonder why though :D [09:20] Saviq: when reproducing with an older image it seems to totally mess up the window stack [09:21] i.e. parts of the right edge animation have wrong z [09:27] mzanetti: you didn't top approve the unity8 branch of dual sim unlocking.. [09:28] is there still something to do with it? [09:31] Wellark: I think I did [09:32] Wellark: now I top-approved too. forgot that before [09:34] mzanetti: thanks! [10:45] Cimi, how do I get to the welcome wizard? [10:45] dandrader, rm /home/phablet/.config/ubuntu-system-settings/wizard-has-run [10:45] dandrader, or somethings similar... [10:46] ok [10:47] tsdgeos: with a little finger mashing, I managed to get this: http://imgur.com/rW6F5oo [10:48] the entire lvwph is frozen now, but I can bring up the scopes view with a bottom edge swipe still [10:55] greyback: what branch is that? regular master? [10:57] tsdgeos: yeah. I just imaged my device with today's devel-proposed and played [10:57] greyback: ouch [10:57] greyback: so the list is totally dead? [10:58] tsdgeos: I was being a jerk though :) But the finger mashing test is good sometimes [10:58] tsdgeos: yeah [10:58] I just now tried to open the video scope via the scopes view thingy, and now the whole UI is frozen [10:58] am working on a backtrace [11:03] greyback: looks as if the fix dandrader did for touch vs mouse clicks didn't really fix everything maybe? [11:03] greyback: since the bottom edge is touch and the rest is mouse [11:03] tsdgeos: that's a possibility [11:03] dandrader: seen the comment i made on touchOwnership? [11:04] Wellark: hmm... I think I found an issue [11:06] tsdgeos, will get to it in a bit [11:07] k [11:07] dandrader: tsdgeos: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1378760 [11:07] Ubuntu bug 1378760 in unity8 (Ubuntu) "[dash] managed to get listview stuck with pull to refresh exposed" [Undecided,New] [11:10] didn't know we had this "pull to refresh" feature... [11:10] greyback, what's a "stable device"? [11:12] dandrader: anything I can do while I have the process running in this state? [11:14] greyback does the bottom edge swipe to show dash overview still work? [11:14] dandrader: no [11:15] pull to refresh does really nohing since it just piggy backs on the list dragging [11:15] hmm, then it doesn't seem to be an issue of missing touch events [11:15] greyback: where are you now at a scope? [11:16] tsdgeos: stuck in scope view now [11:16] greyback, is unity8-dash running in the first place? [11:17] (who knows, might be suspended) [11:17] dandrader: yes it was running [11:17] greyback: and left and right edges work, no? [11:17] darn I managed to kill it [11:17] tsdgeos: yes unity8 itself was fine [11:18] mzanetti: ? [11:18] * Wellark goes lalala [11:18] Wellark: https://code.launchpad.net/~unity-api-team/indicator-network/dual_sim_pin_unlock/+merge/237386/comments/582527 [11:18] mzanetti: don't call it [11:18] mzanetti: the "unlockallmodems" bug was withdrawn from this MP [11:19] Wellark: is there a replacement? [11:19] dandrader: tsdgeos: managed to reproduce it again [11:19] Wellark: the dialer app uses it afaik [11:19] if we do opacity: !hidden instead opacity: hidden ? 0.0 : 1.0 ?? [11:20] it takes a minute or two of 3 finger messing [11:20] greyback: :/ i tried and could get it to fail [11:20] can we do those js tricks or is ugly? [11:20] Hola [11:21] Cimi: ugly, it easier to understand what you're doing with the ? 1 : 0 there [11:21] Cimi: I personally don't like such things === MacSlow is now known as MacSlow|lunch [11:23] hmm, unity8-dash is continually creating & destroying a thread [11:24] mzanetti: nope, the dialer app does not use it [11:24] managed to reproduce the stuck dash too [11:24] mzanetti: as it's not working [11:24] Wellark: so what's the plan with that? [11:24] mzanetti: https://bugs.launchpad.net/indicator-network/+bug/1374082 [11:24] Ubuntu bug 1374082 in indicator-network (Ubuntu) "no API to unlock a specific sim" [High,Triaged] [11:25] Wellark: right... in unity we'd still need to unlock all though [11:25] mzanetti: I'm aware of that [11:25] ok [11:26] mzanetti: there's only one me, so I can't fix everything at the same time ;) [11:26] mzanetti: it will be there [11:28] greyback: tsdgeos: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1378760/comments/4 [11:28] Ubuntu bug 1378760 in unity8 (Ubuntu) "[dash] managed to get listview stuck with pull to refresh exposed" [Undecided,New] [11:28] Wellark: ack [11:29] mzanetti: nice [11:30] mzanetti: that has nothing to do with greyback's bug though [11:30] mzanetti: since it recovers nicely (or at least does for me) [11:30] oh... [11:31] mzanetti: does it recover for you? [11:31] yes... it recovers with more interaction.. [11:31] * mzanetti tries harder [11:31] i mean sure it's a bug that probably needs fixing [11:31] but not as bad as what greyback said [11:32] yep, I managed to make it recover too === Malsasa_ is now known as Malsasa [11:33] greyback: I guess you can tell us if its the same or if yours is different :) [11:33] mzanetti: mine is different, as I cannot make it recover [11:33] bt [11:35] heh, works horizontally too [11:36] oh... now I have it stuck too [11:54] tsdgeos, pushed your forceNonInteractive thing [11:56] tsdgeos, also used status instead of currentStatus. the latter was created just to help fill up previousStatus. [11:57] dandrader: can you merge to listOnBottomSwipe_touchOwnership and then request a build at https://code.launchpad.net/~unity-ui-team/+recipe/unity8-overviewlist ? [11:57] have to run for food now === _salem is now known as salem_ === Estilanda_ is now known as Estilanda [12:41] mzanetti: you forgot to restore the top approved: [12:41] mzanetti: https://code.launchpad.net/~unity-api-team/indicator-network/dual_sim_pin_unlock/+merge/237386 [12:42] Wellark: I'm quite sure there wasn't one before [12:42] at least I didn't remove it [12:42] mzanetti: oh, then charles forgot to add it [12:43] Wellark, mzanetti, I left it out on purpose -- I didn't know if antti had more work/testing in the queue [12:43] +1 on top approval though :) [12:47] tsdgeos, done (PPA just started building though) === dandrader is now known as dandrader|afk [12:47] charles: you could have just asked me, you know! :P [12:47] charles: just top approve it [12:48] charles: ok, seems jussi already did [12:52] dandrader|afk: tx === dandrader|afk is now known as dandrader === MacSlow|lunch is now known as MacSlow [13:22] Saviq: any chance we get the CI build fixed? [13:28] MacSlow: noticing the vol notification come up in the welcome wiz showing mute, but i don't think it's correct...did you notice as well ? [13:29] kgunn, which image? [13:29] dandrader: tsdgeos \o/ right edge so much better [13:29] MacSlow: i'm on utopic devel-proposed [13:29] kgunn, not seen yet... being deep down with the swipe-to-act button atm [13:29] MacSlow: no prob [13:29] yeah on the volume notification i'd expect a bar [13:30] not the speaker logo [13:30] kgunn: there's a new build coming in a few mins that should fix a few small annoyances with the bottom edge [13:30] kgunn: i will give it a final test and probably approve, would be cool if you could also test it [13:30] MacSlow: actually, i don't wanna make you switch if you're in the middle...but i think some quality time with the max vol notfication would be in order [13:31] kgunn, tsdgeos: welll that's because two of the three related branches still in the review-pipe [13:31] MacSlow: reason being, the max vol thing is legal requirement, whereas swipe to answer is user improvement....we can live with click-to-answer, but not w/o max vol notif [13:31] kgunn, tsdgeos: one (unity-notifications) is arleady top-approved... the other (unity8) needs another review (approval) [13:32] and i fear it'll take time to sort out exactly what we're gonna do in a short amount of time [13:32] cause i think it's gonna require alsa driver or audio DAC driver involvement [13:33] vesar: did you have time to show the new bottom list to JMulholland and the guys? any feedback? [13:33] kgunn: did you see my plethora of answers on the volume thing? [13:34] kgunn, regarding the max-volume issue, I think UI/UX is "tiny" compared to the needed foundational underpinnings making it possible [13:34] tsdgeos: not yet, was bug scrubbing & updating this morning [13:34] tsdgeos, nobody replied then, will pester [13:34] tsdgeos, actually fginther is looking into that now [13:34] kgunn: ok, basically there's 2 kinds, the ones that ask "you sure you want more" at 70/80% of volume, and the ones that do nothing [13:34] Saviq: cool [13:42] dednick, looking at your review comments for the greeter-profiles branch. I thought semicolons were passé? [13:43] We need some team style guidance on that, because we are half-and-half now [13:43] dandrader: found two qmluitest failures in https://code.launchpad.net/~dandrader/unity8/touchOwnership/+merge/236152 [13:43] mterry: hm. i've always been told to do them. [13:43] Saviq: ^ ? [13:43] tsdgeos, commented. are you sure you used the latest version of touchOwnership? [13:43] tsdgeos, I'm not getting this failure [13:43] mterry, in JS we want ; [13:43] Saviq, ! OK, my bad [13:44] Saviq, except for one-liners I assume? [13:44] dandrader: yeah, just pulled and make again [13:44] nothing came out of it [13:44] Saviq, do we have a style page for the wiki yet? That would be great [13:44] mterry, yeah, if there's no code block/scope (a binding), that's not really JS [13:44] dandrader: i can try make clean and start again [13:44] mterry, this is the closest we got to defining it https://docs.google.com/a/canonical.com/document/d/1gd87Wo_CSB0DpFWLpTKIIXQfdmFncrq0PHSr9H2PTnk/edit [13:45] Saviq, it is JS though technically right? Like you can do all the JS stuff just in one line [13:45] mterry, technically yes [13:45] dandrader: r1338, no? [13:45] tsdgeos, yes [13:46] tsdgeos, the tst_Shell failure happens also in trunk [13:47] dandrader: does it? [13:47] tsdgeos, try it [13:49] Saviq, OK linked that document from the Checklist wiki [13:51] dandrader: ok, there's a path problem [13:51] dednick, another thing -- you rightly show concern about the async line -- if you've got your test device still, try manually adding that line back in -- the delay is quite profound and I think too noticable for this to land with async enabled [13:51] dandrader: i have unity8 installed from the ppa and was "old", now that i dist-upgraded, the test passes [13:51] dednick, considering that once we have different greeter profiles, a common pattern is for a user to log in just to interact with the full indicator [13:51] mterry: ok. i'll give it a try [13:51] dandrader: should prepend the local paths somewhere [13:51] tsdgeos, you were running qmltest on the device? [13:51] dandrader: no, on the pc [13:52] mterry: perhaps we need to think about loading multiple profiles :/ [13:52] tsdgeos, so you install the PPA on your PC? [13:52] dandrader: i have it yes [13:52] never thought someone would do this :) [13:53] tsdgeos, "should prepend the local paths somewhere" <- didn't get it. [13:53] dandrader: in the tests, you should make sure the local library is used and not the system one [13:54] dandrader: i'm guessing it's the new lib you introduced [13:55] dandrader: xvfbtestShell succeeds here in master [13:56] tsdgeos, if we still get black photos in carousel (I'm checking if we do), that'd be unexpected after your sourceSize thingamajig (lol that's in my spell checking dictionary), right? [13:57] Saviq: there was a thread about someone putting wrong exif info somewhere and thus resulting in black images [13:57] not sure if it may be related [13:57] tsdgeos, right, from the camera app? [13:57] think so [13:57] tsdgeos, you sure you have the latest lp:unity8 built and all? [13:57] dandrader: pretty much [13:58] Saviq, does "make xvfbtestShell" pass for you with lp:unity8? [13:58] dandrader, checking === karni is now known as karni-snack [13:59] dandrader: are you sure you have the latest lp:unity8 and all? [13:59] meh. batter dead :( [14:00] charles: aiui, the checkable-bindings branch is needed, but doesn't completely solve the bug [14:00] per your comment there [14:01] tsdgeos, yes. but I don't have unity8 installed in my system, unlike you [14:01] kgunn, I figured as much, so I left it as a comment instead of needs-fixing [14:01] kgunn, is there a reason why you committed the last release to rtm-14.09 like this https://code.launchpad.net/~unity-team/unity8/rtm-14.09 ? [14:01] kgunn, can I overwrite with current trunk? [14:01] tsdgeos, if that makes any difference... [14:01] charles: rock on [14:01] kgunn, just as long as the problem doesn't get dropped when checkable-bindings lands [14:01] kgunn, :-) [14:01] charles: nope, dednick already on it [14:01] (i think :) [14:02] dednick, rock on [14:02] Saviq: did i do something wrong ? [14:02] kgunn, I usually just push trunk to rtm-14.09 after stuff landed in trunk [14:02] Saviq: while you were out, seb was climbing up my back side asking me to merge trunk to it [14:03] Saviq: yeah, i only did it once... [14:03] dandrader: it shouldn't but it may as we've already seen [14:03] kgunn, basically `bzr push -d lp:unity8 -r 8.00+14.10.20141006-0ubuntu1 lp:unity8/rtm-14.09` [14:03] Saviq: got it....not merge...you push [14:03] * kgunn notes for future [14:04] mzanetti: ^ [14:04] kgunn, which would take tag 8.00... from lp:unity8 and put it there [14:04] and it won't even need overwrite, as it has the same history (in general, now it will because of your merge) [14:05] sorry :-/ [14:05] Saviq: can you revert ? [14:05] kgunn, just an --overwrite away [14:05] kgunn, already done [14:05] hmm... I thought that would happen automatically when releasing to rtm using the train [14:05] kgunn, no need to be sorry [14:05] Saviq: ah so one time penalty of time [14:05] mzanetti, nope, train has no knowledge of the rtm branch [14:06] mzanetti, it will soon, when the two diverge and we'll start cherry-picking from trunk to rtm [14:06] mzanetti, but then there will be no syncing in rtm silos, but MPs as usual [14:06] Saviq: right... I guess we should start doing that now... [14:06] mzanetti, let's see when the floodgates are closed, that's when we'll start === karni-snack is now known as karni [14:23] Saviq: can i ask for some priority love on this one? figure you'll wanna review before top approving [14:23] https://code.launchpad.net/~josharenson/unity8/better_snap_decision_fix/+merge/237485 [14:23] small at least [14:25] kgunn, kk [14:28] Saviq, so, did the test pass? [14:29] dandrader, ugh, distracted, sorry, built now, running it [14:30] * Saviq had some /boot trouble [14:36] mterry: what kind of delay are we talking about here? 1-2 seconds to populate if you open IMMEDIATELY after you switch from greeter? [14:36] dednick, yeah about [14:36] dednick, but I figure that that's going to be common [14:36] dednick, like if you can't see your events or whatever [14:36] dednick, you log in with the intention of immediately going to the menu === gatox is now known as gatox_lunch [14:58] * mterry starts reviewing all the branches so only mine are left for others to review [15:00] mterry: :/ https://dl.dropboxusercontent.com/u/85539674/greeter-indicators.png [15:00] dednick, is that without the silo? [15:00] dandrader, yeah, xvfbtestShell passed: [15:01] mterry: that is with the silo [15:01] Totals: 21 passed, 0 failed, 0 skipped [15:01] dednick, indicator-transfer in utopic has a bug where it shows the datetime UI in greeter mode [15:01] dednick, but you don't have a transfer icon in that screenshot either... [15:01] dednick, I'm suspicious [15:01] mterry, which branches do you want to be looked at? [15:01] MacSlow, mine in https://code.launchpad.net/unity8/+activereviews [15:01] mterry: and there are 2 datetimes [15:02] MacSlow, well any of them. But *I* want mine looked at :) [15:02] mterry, I know that feeling :) [15:02] dednick, yes because transfer shows datetime UI in greeter mode (as a bug -- if you look at it's indicator keyfile, it has a typo and shows the wrong object path) [15:02] dednick, that's fixed in the silo [15:02] dednick, which is why I'm suspicious you are running the silo [15:03] hm. should be using silo [15:05] mterry: i think the ppa has older version than is available. [15:06] dednick, then you can edit /usr/share/unity/indicators/com.canonical.indicator.transfer manually and fix the typo [15:06] tsdgeos, ahhhh, finally found the cause of the mistery. my local lp:unity8 branch has a pending merge. and bzr pull still works fine even it you have a merge pending! [15:06] phew! [15:07] mterry: yup. did already [15:08] lol this is bad [15:08] open the clock app [15:08] clock app takes 70% [15:08] but not only that [15:08] ted, I thought the transfer indicator was fixed in your greeter profile silo? [15:08] untiy8 takes 30% cpy [15:08] why? [15:09] mterry, Should be… [15:09] ted, I think maybe the silo just needs a rebuild then to get back above utopic [15:10] mterry, Yeah, working on that. merge conflicts. [15:11] ah because unity8 is the display server [15:11] so that 30% is actually mir? === Estilanda_ is now known as Estilanda [15:19] mterry: hm. ok, there doesnt really seem to be that much lag to me, but the lag introducedin sync mode is pretty bad i think. [15:19] I think the lag with async will be reduced when my new indicator branch lands. The menu is actually at 0 height when you drag it, so i think the items are being loaded "only as visible". The new panel page is always the same height so the items will be loaded immediately. [15:22] dandrader: so it fails/works for you now? [15:24] tsdgeos, yeah, now it passes in lp:unity8. working on a fix for it in touchOwnership [15:24] dandrader: \o/ [15:24] :) [15:27] mterry: just commented on your branch [15:36] mterry having an internet day === dandrader is now known as dandrader|lunch === gatox_lunch is now known as gatox === dandrader|lunch is now known as dandrader === alan_g is now known as alan_g|EOD [17:40] dednick, I assume you didn't get my reply an hour ago -- I had awful internet [17:41] dednick, back in a cafe [17:51] Does anyone understand what is up with the qtmir-gles stuff and the "twin packages" [17:51] I wonder if its related to the struggle I am having with building qtmir-desktop on armhf in CI [17:51] ...with cmake [17:52] thats a lot of qualifiers lol [17:53] racarr: i think bregma knows [17:53] racarr: Qt has a compile time choice to use DesktopGL or GLES. The results are ABI incompatible. So we've 2 version of Qt in the repos [17:54] as result, we need separate version of qtmir to compile against the Qt with GL, and the Qt with GLES [17:57] greyback_: I guess I thought thats what qtmir-android v. qtmir-desktop? [17:57] was [17:58] is there a binary package called qtmir-gles? Not that I can see in the Ubuntu archives. [17:59] there isn't [17:59] ok so the binary package is qtmir-android [17:59] * greyback_ got confused, has to look it up again [18:00] my main problem is I don't understand how qtmir-desktop is building on armhf [18:00] as it stands. [18:00] e.g. ifyou try and build qtmir-desktop and qtmir-android fromt hesame [18:00] build dependencies [18:00] wont qtmir-desktop be linked against [18:01] Qt-gles [18:01] qtmir-desktop is built for all arches out of lp:qtmir - it's only qtmir-android that is split over qtmir & qtmir-gles [18:02] qtmir-gles builds qtmir-android for i386 and x64, as those are the situations in the emulator, where Qt has GLES but runs on a non-android system [18:03] yes qtmir-desktop will link against the qt-gles on armhf [18:03] which does sound wrong [18:03] then what is it? [18:03] yeah [18:04] I suspect nothing has ever used qtmir-desktop on armhf [18:04] yeah haha...I think its probably linked against [18:04] GLES... [18:04] the qtmir qmake files [18:04] dont explicitly use GL or GLES afaict [18:04] so its just coming from the [18:05] Qt configuration [18:05] makes sense yeah [18:05] so the problem is I dont think its possible to build [18:05] qtmir-desktop and qtmir-android from the same source package on armhf? [18:06] Unless Qt/GL and Qt/GLES are parallel installable [18:07] qtmir-desktop will compile against Qt/GLES, as that's the one on armhf usually [18:07] greyback_: I thought qtmir-desktop was supposed to be qtmir linked against Qt/GL though [18:08] what is the qtmir-desktop/android distinction? [18:08] racarr: I think it is supposed to be too, but in actuality I suspect it doesn't - but it's not used so no-one ever noticed [18:08] the main distinction is a call to eglBindAPI(DesktopGL) somewhere [18:09] right... [18:09] hmm ok [18:09] greyback_: Do you think I should just try and replicate this in qtmir/cmake or [18:10] we should get rid of qtmir-desktop on armhf [18:10] racarr: bregma might object to the latter notion. I would rather not having hacks in our cmake, just to copy qmake weirdness though [18:12] it's debian that was including that extra runtime dependence on mesa-gl1, can we force it just build-time only? [18:12] greyback_: I realized though, even then the resultingpackage [18:12] would be linked against GL, but Qt/GLES [18:13] so to really fix it you would need to make the Qts parallel installable and selectable [18:13] I thought Qt was going to switch to doing runtime detection? [18:13] rsalveti will probably say that's not gonna happen any time soon :) [18:13] bregma: that was on Windows [18:14] maybe qt 5.5? [18:14] mainly as some customer paid them to do it I suspect [18:14] let me find the qt bug [18:14] we do runtime detection in libSDL2, so we know it's possible [18:14] yeah, not much progress [18:14] https://bugreports.qt-project.org/browse/QTBUG-36829 [18:14] bregma: it's possible, but still require some work on qt [18:15] * greyback_ late for the pub [18:15] o/ [18:15] Bye! thanks [18:15] was having lots of trouble wrapping head around it myself [18:17] I think we should just delete the qtmir-desktop armhf package...I kind of doubt its ever worked. Does anyone even have an mesa-opengl (not gles) armhf device [18:17] to test if it works? [18:17] Does such a thing even exist? [18:17] But if its impossible to [18:17] delete packages [18:17] I can work around it now I guess... === karni is now known as karni-afk === karni-afk is now known as karni [19:34] I don't suppose anyone has extra review cycles floating around? === boiko_ is now known as boiko === salem_ is now known as _salem