[08:11] tsdgeos, mornin' [08:11] Saviq: hey [08:11] tsdgeos, we got 5.4 breaking a bunch of our tests ;/ [08:12] did we? [08:12] http://pastebin.ubuntu.com/10269595/ [08:12] oups [08:13] tsdgeos, runaway :P [08:13] tsdgeos, http://pastebin.ubuntu.com/10269595/ [08:14] what [08:14] clicking the link makes it vanish [08:14] maybe needs a qt 5.4 rebuild? [08:15] anyhow [08:15] yeah i'll have a look [08:16] tsdgeos, I can repro locally after a rebuild [08:17] tsdgeos, the other priority we have today is https://code.launchpad.net/~unity-team/unity8/physical-key-mapper/+merge/249890 - if Daniel shows up early, he could review, otherwise I'd like you to have a look, and mzanetti to test rtm silo 2 [08:17] 20 [08:18] Saviq: he's carnival'in [08:18] so i don't think he'll show up [08:18] tsdgeos, oh right [08:18] tsdgeos, then yeah, please review, it's not huge [08:18] sure [08:18] now clicking on that link didn't make it crash [08:18] :D [08:19] if you have an idea to simplify, let me know, it was 1am yesterday I finished that, so might've missed something [08:26] tsdgeos, network issues? or IRC acting up? [08:26] Saviq: irc client crashing because of Qt 5.4 [08:26] http://paste.ubuntu.com/10269746/ [08:27] o yay ;) [08:34] moin moin [08:34] * mzanetti starts testing silo2 [08:38] mzanetti, 20, not 2 [08:39] mzanetti, before that, you might wanna ask trainguards to update qtsystems in silo 0 [08:39] mhm [08:40] I just kicked a unity8 build [08:41] right, your qtsystems needs to be rebased on the new git snapshot and version bumped [08:42] yup [08:44] ack [08:44] Saviq: can you update me on deadlines? [08:45] silo20, today EOD, right? [08:52] mzanetti, yeah === tsdgeos_ is now known as tsdgeos [08:59] mzanetti: yes silo20 eod otday [09:51] Saviq: damn the touch stuff daniel did seems not to be working the same :/ [09:52] we really need to get this qmluitests into the packaging :/ === anpok_ is now known as anpok [10:37] Mirv: I've updated my qsysteminfo branch. should build fine agains 5.4 now [10:40] Saviq: hmm... rtm silo20 turns the screen off when I try to take a screenshot [10:41] shouldn't there be a u-s-c branch too in there? [10:43] mzanetti: it's not needed anymore with saviq's code [10:44] mzanetti: how do you take a screenshot? note it's the two volume keys now [10:44] just press them at the same time [10:44] not working here [10:44] takes a screenshot with power + voldown [10:44] but also turns the screen off [10:44] mzanetti: well you should not be pressing power :D [10:45] then no screenshots here [10:45] then rebuild the silo? [10:45] which branches are there? [10:45] only unity8 [10:45] ok, can try a rebuild... thought it would have been ready for testing (as per emails in my inbox) [10:46] yeah looked like it [10:47] mzanetti: but yeah look at https://code.launchpad.net/~unity-team/unity8/rtm-20150213/+merge/249680 [10:47] it has [10:47] } else if (event.key == Qt.Key_VolumeDown) { [10:47] if (event.isAutoRepeat && !d.ignoreVolumeEvents) root.volumeDownTriggered(); [10:47] else if (!event.isAutoRepeat) { [10:47] if (d.volumeUpKeyPressed) { [10:47] if (Powerd.status === Powerd.On) root.screenshotTriggered(); [10:47] right... says that commit is 12 hours old [10:48] so screenshot is triggered with down+up [10:48] and the last rebuild from silo is 12h too [10:48] so probably not in that build yet. building now [10:59] tsdgeos: think this is 5.4 related? https://launchpadlibrarian.net/197896968/buildlog_ubuntu-vivid-amd64.ubuntu-system-settings_0.3%2B15.04.20150217-0ubuntu1_FAILEDTOBUILD.txt.gz [11:00] yes [11:00] mirv had a patch that adds the missing include [11:00] not sure if it has landed or not [11:00] Mirv: ↑ ? [11:07] mzanetti: ok. someone else I guess uploaded it to your silo last time, but do you want me to do it now? [11:07] Mirv: yes please [11:07] mzanetti: tsdgeos: sorry, ubuntu-system-settings trunk now up-do-date. it was a patchwork of multiple branches. [11:08] ok. rebuilding systemsettings in silo0 then [11:19] Mirv: we need to add to the "things to check when moving qt" [11:19] run qmluitests in unity8 [11:20] i forgot and we have lots of broken tests regarding touch now :/ [11:21] tsdgeos: please add to https://wiki.ubuntu.com/Process/Merges/TestPlans/Qt [11:22] tsdgeos: I now added a new header "Checks to be done by certain upstreams" that could be useful [11:23] (or have qmluitests part of AP:s or something..) [11:23] mzanetti: new qtsystems building in 000 https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-000/+sourcepub/4783146/+listing-archive-extra [11:24] Mirv: thanks! [11:24] Mirv: added a line [11:24] tsdgeos: great! [11:24] Mirv: the plan is having qmluitests as part of the regular make check, but until that happens let's have that line :D [11:31] Hello ther eo/ [11:31] tedg, o/ [11:31] I managed to have the count in the dock \o/ [11:31] https://dl.dropboxusercontent.com/u/19626172/dock_count.png [11:33] tsdgeos, I didn't understand if I have to add maximumLineCount or not [11:33] tsdgeos, do we care? [11:33] Cimi: let's add it so others don't try to use \n again and complain when it breaks [11:34] tsdgeos, ok, both attributes and subtitle? [11:34] attributes has it already [11:45] pete-woods: do you remember which bug https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1422689 is duplicate of? [11:45] Launchpad bug 1422689 in unity8 (Ubuntu) "Cached (?) lock screen messages not translated" [Undecided,Confirmed] [11:45] MacSlow: is this wanted? http://i.imgur.com/cVnOPma.png [11:45] looks like a bug to me [11:45] mzanetti, certainly a bug [11:45] mzanetti: what's wrong? [11:46] tsdgeos: the notification covers the panel [11:46] tsdgeos, the notification needs to respect the top-bar [11:46] ah [11:46] mzanetti: where's that? rtm? [11:46] mzanetti, how did you force this to happen? [11:46] so... silo 20 no good yet [11:46] mzanetti: you sure that's new? [11:46] well, it doesn't happen on my dogfooding phone [11:47] an reflash to check [11:47] can [11:47] mzanetti, this used to be a bug... very long ago [11:47] mzanetti: so we want regular rtm, right? let me see [11:47] yeah [11:48] tsdgeos, https://code.launchpad.net/~cimi/unity8/fix-long-subtitle-card/+merge/249984 [11:48] mzanetti, unless the panel-height isn't messed around with, this should not happen [11:48] Cimi: ok [11:49] mzanetti: yeah seems to work here [11:49] probably some issue with cherry-picking [11:49] yeah [11:49] iirc there was a change related to panelHeight in trunk which isn't in rtm ye [11:49] t [11:49] cherry-picking of death [11:51] mzanetti, fyi... overall placement for notifications-listview happens in Shell.qml [11:55] tsdgeos: weird... I just reflashed rtm and seems I have the issue there too [11:56] hmmm [11:57] i was using image 234 [11:57] let me use 236 [11:58] mzanetti: well is on release or on autorepeat, is it really that bad? [11:58] it felt ok to me [11:58] it's not *that* bad... but still not as good as direct feedback [11:59] tsdgeos: https://bugs.launchpad.net/ubuntu-translations/+bug/1327419 [11:59] Launchpad bug 1327419 in libusermetrics (Ubuntu) "Infographic's language not updated until the metric value is changed" [High,In progress] === MacSlow is now known as MacSlow|lunch [12:00] I've marked the duplicate [12:00] pete-woods: thanks :) [12:05] mzanetti: 236 works fine here on krillin [12:05] mzanetti: weeeird [12:05] mzanetti: what do you do? [12:06] so afaict the topmostIsFullscreen property in Notifications is broken in trunk and in rtm [12:06] MacSlow|lunch: ^ [12:06] tsdgeos, need my help on anything? [12:06] still that doesn't explain the issue [12:06] must be something with panel.panelHeight [12:06] mzanetti: but but it works for me :D [12:06] tsdgeos, with this I am done, unless I start fixing the zoom with the image gallery preview [12:07] mzanetti, I'll look into it when I'm back from eating [12:07] Cimi: i'm looking at some regressions we have due to 5.4 in touch, no other thing for you in my side [12:07] Cimi: so if you have something else to fix, fix! [12:09] tsdgeos: which channel did you flash? [12:09] wat! [12:09] ubuntu-device-flash --channel=ubuntu-touch/ubuntu-rtm/14.09-proposed [12:09] rebooted, it works now [12:09] without changing anything [12:09] :S [12:09] * mzanetti installs silo20 again [12:12] ok... broken again [12:14] tsdgeos: I have a feeling the SIM pin notification breaks it [12:14] ah [12:14] that may be [12:14] i have no sim [12:15] it'd be cool if you confirm it in plain rtm [12:15] so we have a proper bug to fix :) [12:15] yep, on it [12:17] tsdgeos: yep, confirmed... removing the sim fixes it. the one time it was working before was because I'm suffering from the bug where the Nexus4 sometimes won't pick up the sim card [12:17] mzanetti: good, assign to MacSlow|lunch then :) [12:17] yep, I'll report [12:20] darg [12:20] so there's touch event compression now [12:20] breaks all our tests [12:20] question is if also breaks runtime or not [12:21] seems it doesn't [12:21] so let's force it for tests now only [12:22] * mzanetti is not exactly sure what tsdgeos is talking about :D [12:22] mzanetti: we have lots of tests failing with Qt 5.4 [12:22] because QtQuick has introduced "touch event compression", [12:22] i.e. basically adding up together move events [12:22] but that breaks some of our tests [12:22] that do: move, check, move, check, [12:22] does that mean it discards intermediate events? [12:23] but real life DDAs and stuff work [12:24] mzanetti: yes and no [12:24] basically delays it by 1 [12:24] so basically they are compressing events tha are exactly the same [12:25] i guess they were getting repeated stationary touch events or something [12:25] mhm [12:25] you want to talk to daniel, he does some fancy gesture speed calculations that are probably affected [12:26] Saviq: when you have a minute, please ping me [12:26] mzanetti: of course, that's why tests are failing [12:28] mzanetti: qtubuntu also needs rebuild in 000 [12:28] Mirv: ack, thanks [12:29] greyback_: fyi ^^ [12:29] greyback_: ^ too, since you're rebuilding qtmir atm [12:29] :) [12:38] hmm... can't swipe away the volume notification :/ [12:41] not all notifucations are swipable === MacSlow|lunch is now known as MacSlow [13:04] mzanetti, that's intentional [13:04] mzanetti, sync. notifications and expanded snap-decisions are not meant to be swipe-able [13:05] I can see how that makes sense for snap decisions, as they have buttons to get away with them [13:05] but volume up/down... [13:06] mzanetti, is the top-margin issue still valid? [13:07] MacSlow: yes: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1422711 [13:07] Launchpad bug 1422711 in unity8 (Ubuntu) "SIM pin notification breaks notification positioning" [Undecided,New] === alan_g is now known as alan_g|lunch [13:14] mzanetti, pong [13:14] mzanetti, trying to reproduce it now [13:14] Saviq: hey [13:15] mzanetti, volup+voldown for screenshot [13:15] not power after all [13:15] yeah, saw that... feels odd [13:15] the onReleased to change volume [13:15] mzanetti, why? [13:15] because it does :D you press and it doesn't change the volume [13:15] mzanetti, it odes [13:15] onRelease only [13:16] mzanetti, if you release straight away [13:16] mzanetti, which is 99% of the use cases [13:16] yeah well... just my opinion [13:16] mzanetti, and for long-press it works again [13:16] I wouldn't do it this way [13:16] mzanetti, best of both worlds simply, otherwise you'd need to press power *before* vol [13:16] so? [13:17] anyways, if you all agree the onReleased is ok... works for me I guess... [13:18] just doesn't feel polished enough IMO [13:18] another thing is that volume notifcations can't be dismissed by swiping [13:18] which according to Mirco is the expected behavior. feels odd to me too [13:19] mzanetti, ok... reproduced... and got an idea what might be causing it. [13:19] mzanetti, notifications are getting a redesign right now, so don't worry there [13:19] mhm, ok [13:19] mzanetti, if you have better ideas for the screenshot, I'm all ears, [13:20] but requiring power first, then volume, sounds not user friendly, especially on arale, where you have power on top and volume on side [13:20] thing is, you have to look it up anyways how to take screenshots as it's kinda a hidden feature, and very few people will use it. [13:21] but the volume buttons reacting too late affects everyone [13:21] mzanetti, I'm just fixing a bug now, really, there's no design for this [13:21] ack [13:21] mzanetti, we can file a bug for ubuntu-ux to actually test a few possibilities [13:21] Saviq, mzanetti: how about combined swipe-guestures instead of hw-buttons? [13:21] as I said... just my 2 cents... you asked me to test, here's my opinion [13:22] Saviq, mzanetti: it would also avoid the always present volume notification on resulting screenshots [13:22] MacSlow, that's what my branch fixes [13:22] MacSlow: you're not up to date :) [13:23] mzanetti, well I'm in AP-land atm [13:27] mzanetti, I just checked with android, they do it power+home (at least old, + samsung), new one is supposed to be power + voldown, [13:27] mzanetti, and holding for 2s [13:28] Saviq: I personally really don't mind how to take screenshots, as long as it doesn't impact other behavior (e.g. no delay of other actions - be it a timer or onReleased) [13:28] in any case, I really don't think the onReleased change is so detrimental to the UX, you know because you know, 99% of other people won't even notice [13:33] it just adds a feeling of slowness [13:33] so they won't understand what's happening, but they'll say "the iphone reacts faster to input" [13:33] as an example [13:35] anyways [13:35] it works [13:35] gonna run AP now on the silo [13:35] looks ok otherwise [13:39] greyback_: are you already on it to fix qtubuntu? [13:39] mzanetti: yeah [13:39] ack [13:39] thanks [13:40] mzanetti, I have a working fix for LP: #1422711, just need to figure out a good way to test it... qml-wise [13:40] Launchpad bug 1422711 in unity8 (Ubuntu) "SIM pin notification breaks notification positioning" [Medium,In progress] https://launchpad.net/bugs/1422711 [13:43] Saviq: https://code.launchpad.net/~paulliu/unity8/lp1350891_ScrollBackground/+merge/249989 [13:43] Saviq: I'm wondering if this changes are ok. [13:44] paulliu, from a quick look - where's the DashBackground we had previously? [13:44] paulliu, but you'll need someone else to review [13:46] Saviq: previously I put the DashBackground alongside the ScopeListView. And scroll it by contentY. [13:46] Saviq: But the things is tryGenericScopeView works. The background scrolls. [13:47] paulliu, yeah, but what I mean is don't we already have a DashBackground, and if we do, why can't I see it being removed? [13:47] in the diff I mean [13:47] Saviq: got it. [13:47] Saviq: Let me remove it. === marcusto_ is now known as marcustomlinson === alan_g|lunch is now known as alan_g === marcusto_ is now known as marcustomlinson === sil2100_ is now known as sil2100 [15:34] tsdgeos, why don't we run the qmluitests during build anyway? [15:53] mterry: saviq was working on it [15:53] mterry: don't remember the exact problems he had [15:53] let me try to dig the review [15:53] suspect an xserver of some kind needed [15:53] plus maybe a mock dbus server [15:54] sure, both are needed [15:54] but both are things we have already [15:55] mterry: can't find the MR now, maybe he discarded it [15:55] fair [16:10] tsdgeos, mterry, yeah, I still have it in the pipeline, not sure they will run during build (waste of builders time), but definitely in DEP8 [16:11] Saviq, DEP8 would be great [16:12] mterry, yeah, that's prio [16:32] ChrisTownsend, hey, is your unity8 desktop session on vivid still working? mine freezes on the greeter [16:32] willcooke, ^ [16:33] seb128: Mine is working and I just updated the machine a couple of hours ago. [16:33] :-( [16:34] seb128: Which greeter is it freezing on? (Unity 8 or the one w/ LightDM) [16:36] unity8 in session [16:36] seb128: Hmm, any clues in ~/.cache/upstart/unity8.log? [16:36] let me see [16:43] ChrisTownsend, nothing useful I can see, complaining about loading the battery 100 charged icon, but I doubt it's it [16:44] seb128: Hmm, strange. Since you're in U8, I doubt anything useful will be in /var/log/lightdm/unity-system-compositor.log either, but it may be worth checking. [16:48] ChrisTownsend, tried with another user, my log is spammed with "Failed to get image from provider: image://theme/battery-100", I wonder if that's the issue [16:48] seb128: Hmm, it might be. I'll look to see if I get that error in my log as well. [16:49] ChrisTownsend, of course my batter is fully charged to get that error I guess [16:49] on thread is blick in QPixmap code [16:50] seb128: I have a fully charged battery as well, but I don't see that error. [16:50] :-( [16:53] seb128: Maybe find the package that the battery-100 image is in a force the package to be reinstalled? Just a thought... [16:53] mardy: was told to ask you about ubuntu-system-settings-online-accounts. It builds/installs, but trunk can't get a trusted prompt session (returns null). However when I build rtm branch, it works fine... [16:53] ChrisTownsend, I prefer not, the shell shouldn't freeze on a missing icon [16:54] ChrisTownsend, so that's still worth a bug report imho [16:54] seb128: True and I agree, it sounds like a bug. It will definitely prove it's a bug if you replace the missing icon and it all works. [16:56] ChrisTownsend, getting a bt of the hand first :p [16:56] seb128: Yep [16:57] ChrisTownsend, http://paste.ubuntu.com/10276002/ [16:57] getting debug symbols [17:00] http://paste.ubuntu.com/10276060/ [17:00] Saviq, others, ^ is that a known issue? currently untiy8 desktop hangs on start for me [17:00] on vivid [17:01] log has "Failed to get image from provider: image://theme/battery-100" warnings and the bt suggests it's stucked in some image loader [17:01] could be a qt5.4 fallout [17:07] not sure that'd even be unity8 necessarily [17:09] but i see what you mean... you wouldn't want unity8 to hang b/c of it [17:09] seb128: is it hanging, or using 100% cpu? [17:09] that bt is familiar to me [17:09] greyback_, cpu [17:09] yeah [17:09] which leads to hang [17:10] well, the UI isn't redrawing [17:10] it's stucked [17:10] seb128: is dbus very busy? [17:10] my theory was that the battery status dbus message is firing incredibly fast, causing qt to redraw the battery icon continually [17:11] completely blocking everything else [17:11] greyback_, did you open a bug about that? [17:12] seb128: no as I only occasionally discovered this happening [17:12] seb128: there is this one [17:12] https://bugs.launchpad.net/bugs/1417773 [17:12] Launchpad bug 1417773 in unity8 (Ubuntu RTM) "Unity8 completely frozen (unable to unlock, receive calls, etc)" [Undecided,New] [17:13] which got debugged down to 3 distinct probs....1 of which was dbus [17:13] kgunn, doesn't seem the same one [17:14] seb128: right...i would suppose gerry's theory would eventually recover ? [17:14] I never saw it recover [17:14] it doesn't recover [17:14] so busy forever ? [17:14] greyback_, dbus doesn't seem especially busy [17:14] kgunn, seems so [17:14] somehow yeah [17:14] and at every login now on that box [17:14] seb128: ok, there goes that theory [17:14] I can't logging anymore [17:14] tried 10 times with different users as well [17:15] seb128: when did this occur ? like do you know if it worked on Friday...but broke between then and now ? [17:15] greyback_, the unity8 log suggests it fails to find the icon [17:15] seb128: if you're in gdb, could you get the icon string or filename from around frame 4? Might give us a clue [17:16] greyback_, log has "Failed to get image from provider: image://theme/battery-100" [17:16] seb128: sure, but the BT indicates it's doing a pixmap fill [17:16] as if it's reading an image [17:16] * greyback_ not sure [17:17] greyback_, new bt seems a bit different, see bottom of http://paste.ubuntu.com/10276267/ [17:19] seb128: that's blocked on a dbus call [17:19] greyback_, how do I print the content of a QString in gdb? [17:21] seb128: you need a pretty printer:http://pastebin.ubuntu.com/10276312/ [17:21] will add "printqstring" should shoud do the job [17:22] I never got this working: http://stackoverflow.com/questions/10492290/gdb-pretty-printers-for-qt5 [17:27] greyback_, that gives me a "There is no member or method named len" :-/ [17:28] when trying to printqstring filename [17:29] pah [17:30] dumb script [17:30] :-/ [17:30] if filename hasn't been optimized out, it should be readable [17:31] my apologies but I have to go for a couple of hours [17:32] greyback_, no worry, if you/somebody else is interested in debug more later, let me know [17:32] meanwhile that box is a brick to me... [17:32] will be in touch when I get back [17:32] I'm probably going to be off for the day when you go back [17:32] but feel free to ping me tomorrow [17:33] ok [17:50] greyback_, filename = /usr/share/icons/sure/status/scalable/battery-100.svg [17:51] greyback_, /usr/share/icons/suru/status/scalable/battery-100.svg [17:51] rather [17:59] greyback_, opened https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1422835 [17:59] Launchpad bug 1422835 in unity8 (Ubuntu) "Unity8 stucks in icon loading code" [Undecided,New] === alan_g is now known as alan_g|EOD