=== jono is now known as Guest91841 === duflu_ is now known as duflu [04:34] greyback's qt5 prob (unity gets deinstalled) is indeed because of the dependency chain, but just more complex ie. bugs need to be fixed in other packages first === OutOfControl is now known as benonsoftware === It is now known as Wug === tjaalton_ is now known as tjaalton [08:26] mzanetti: you back' [08:26] ? [08:29] tsdgeos: hi [08:30] yes, I am [08:36] mzanetti: any chance we can do that "detach CI vm thing" so i can run the qmltests faster? Trying to get more info to make them work more reliable [08:38] tsdgeos: sure [08:43] tsdgeos: which one do you need? [08:43] mzanetti: qmluitests [08:47] tsdgeos: ps-trusty-server-amd64-3 [08:50] hi guys [09:05] Cimi: ho [09:14] so [09:14] this is interesting [09:14] any idea why this may happen [09:14] https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-trusty/797/consoleFull [09:14] search for "qmltestrunner::GenericScopeView::test_filter_expand_expand() A 0 [object Object]" [09:14] and see how a few lines up [09:14] it was [09:14] "qmltestrunner::GenericScopeView::test_filter_expand_expand() A Header_QMLTYPE_11_QML_52(0x1e3f020, "dashSectionHeader0")" [09:14] but the variable does not get reassigned [09:14] so it's just that the thing is disappearing? [09:16] yep it is [09:22] i think i have a fix for the expand_expand test! [09:23] which is not the one i wanted to fix originally [09:23] but oh well :D [09:30] Saviq: mzanetti: https://code.launchpad.net/~aacid/unity8/expand_expand_fix/+merge/200642 [09:31] tsdgeos: what exactly does this fix? [09:31] un unstable test? [09:31] ah right [09:31] not the one i wanted to originally fix [09:31] but i can't get the other to fail ^_^ [09:32] :D [09:32] busy looping now to see how much time the thing runs without failing [09:32] oh [09:32] i'm running the wrong unittest [09:32] tsdgeos: do we really need the tryCompareFunction? Looks a bit like a regular tryCompare would be enough [09:33] that'd make sense [09:33] mzanetti: the thing is that the header0 object goes away [09:33] ah ok [09:33] at some point [09:33] i need to requery it [09:33] no idea why tbh [09:33] otherwise you end up asking for an x of an object that "is not there" anymore [09:34] * tsdgeos runs the correct unittest [09:40] Saviq, ok to talk about previews and shell tomorrow? === greyback_ is now known as greyback [10:15] greyback: hi! regarding your unity/qt5.2 problem, I wrote about it before Christmas, when I identified at least hud needing a recompile, and hud requires gsettings-qt to be fixed (bug #1257322) - libqtdbustest was already fixed, another pre-requirement [10:15] bug 1257322 in qtdeclarative-opensource-src (Ubuntu) "gsettings-qt doesn't work with Qt 5.2" [Undecided,Fix committed] https://launchpad.net/bugs/1257322 [10:16] greyback: so there's a rather long chain of various dependencies that need to be fixed, about which I also wrote [10:16] Mirv: oh I missed that message, apologies. Ok, I will be patient, thanks [10:17] greyback: webbrowser-app needs UI Toolkit, which has some segfault in turn [10:17] but it's good to know that's also required for unity to not be removed when upgrading [10:18] I'll update the pad accordingly [10:20] mhr3, yup [10:21] Saviq, time ok, can move if you'd prefer [10:21] ? [10:21] mhr3, fine [10:21] k [10:56] Saviq: the Dash::test_show_scope_on_load failure is very weird [10:56] it is the listview that is actually not creating the delegate in some cases [10:57] tsdgeos, hmm, IIRC we have cacheBuffer MAX_INT there effectively? [10:57] yep [10:57] i mean [10:57] it works [10:57] most of the time [10:57] but sometimes not [10:57] which is weird [10:57] because i have never seen the phone just loading 2 of the 4 screens [10:59] i'm wondering if again we're just dealing with the evil scene graph loop from 5.0 failing to "do stuff" [11:03] tsdgeos, since you have the VM [11:03] tsdgeos, can you try and upgrade it to 5.2 and see if you can reproduce the issue? [11:05] makes sense [11:05] let me see [11:11] mzanetti, top-approve expand_expand fix? [11:12] Saviq: done. was waiting for jenkins to run through and forgot [11:12] mzanetti, cheers [11:23] Saviq: i've done something else before the 5.2.0 update [11:23] i've exported QML_BAD_GUI_RENDER_LOOP [11:23] tsdgeos, righ [11:23] t [11:23] that runs the non threaded render loop [11:23] and i'm running the tests in a loop [11:24] let's give it some time to see if it fails or not [11:24] and then [11:24] do the 5.2.0 anyway [11:24] since we don't want the non threaded thing === alan_g is now known as alan_g|tea [11:26] tsdgeos, yup === _salem is now known as salem_ [11:39] well, been running for 14 minutes [11:39] so the non theraded loop makes it work [11:39] let's see what 5.2.0 says === alan_g|tea is now known as alan_g [12:31] Mirv: i can't install qtdeclarative5-ubuntu-ui-toolkit-plugin when using the ppa for 5.2 [12:31] it complains it wants libqt5core5 but the ppa provides libqt5core5a [12:31] any idea? [12:33] Saviq: ↑ you? [12:33] Mirv: I've a branch attached to https://bugs.launchpad.net/platform-api/+bug/1266674 to fix the ftbfs, works for me now. Just need a reviewer [12:33] Ubuntu bug 1266674 in platform-api "platform-api fails build against libmirserver12" [Critical,In progress] [12:34] tsdgeos: I think UITK has a crash bug with 5.2, so wasn't pushed to the PPA yet [12:35] greyback: it is in the ppa [12:35] or so apt-cache says [12:35] 1:0.1.46+14.04.20131129-0~873~ubuntu14.04.1+disabletests2 0 [12:35] 500 http://ppa.launchpad.net/canonical-qt5-edgers/qt5-beta2/ubuntu/ trusty/main amd64 Packages [12:35] but it's pulling the wrong dependency :/ [12:36] tsdgeos: ah I see. Ignore me :) [12:37] but based on the date, I might guess that package was made before the lib name changed to libqt5core5a [12:38] seems [12:52] tsdgeos: UI toolkit has a failing test when compiled, so it hasn't rebuilt yet. it's blocking continuing from there currently together with gsettings-qt. [12:52] ok :( [12:52] tsdgeos: see http://pad.ubuntu.com/qt52-dependencies for the approximate dependency chain [12:52] plus bug links [12:53] edits welcome to the pad [13:13] greyback, tried running my carousel shader branch again locally on the nexus 10, it ends with "terminate called after throwing an instance of 'boost::exception_detail::clone_impl >' [13:13] what(): could not activate surface with eglMakeCurrent" [13:14] Cimi: often happens when backlight turned off. Hit the power button and try again === alan_g is now known as alan_g|lunch [13:19] tsdgeos, jenkins *really* doesn't want to merge the horizontal journal, it doesn't look like it's random any more... [13:20] tsdgeos, QFATAL : qmltestrunner::Dash::test_dash_bar_set_index_connection(applications.scope) ASSERT: "modelIndex == m_asyncRequestedIndex" in file /tmp/buildd/unity8-7.84+14.04.20131220/plugins/DashViews/listviewwithpageheader.cpp, line 874 [13:25] greyback: I'm testing that platform-api branch building and approving after that [13:25] Mirv: ack [13:28] greyback, ok, tested [13:28] greyback, it's slower [13:28] but it's already slow in the normal use case [13:29] and this looks much better [13:29] Cimi: I would argue that smoother is more important than looks [13:30] Cimi: but let's have a third party make that call [13:32] greyback, it's already bloody slow [13:32] greyback, we might though want to keep this to test again when mir will be faster [13:32] Cimi: making it slower isn't improving the situation then :) [13:32] greyback, so postpone the approve [13:33] greyback, I can try again using layers in qml but I ended up doing shader because layers wasn't working [13:33] there has been a regression though [13:34] when you open the preview, you see the carousel zoom popping in and out [13:34] I'll have a look later this afternoon, now off to the office + lunch [13:35] Cimi: I'm not up on the specifics, I just got worried when I saw you using a shader on a visually large ShaderEffectItem as that's expensive. Hopefully there's an alternative approach === dandrader is now known as dandrader|afk [13:44] Cimi, layers are basically equal to using a shader in terms of GPU usage === dandrader|afk is now known as dandrader [14:00] Saviq: right, somehow hadn't see it [14:01] Saviq: thanks for the pointer [14:01] dandrader: i should not have listened to you asking for the assert :D [14:01] now i need to find out why it's asserting :D [14:01] tsdgeos, :D [14:02] Saviq: ok, so i'll leave the qml error now, knowing it does not happen with the non threaded scenegraph and we can't test 5.2 atm :/ [14:03] tsdgeos, +1 === alan_g|lunch is now known as alan_g [14:11] mzanetti: ok, put the VM back into the pool i guess, we'll be back to it some other day [14:11] mzanetti: tx a log! [14:11] and a lot too! [14:11] :D [14:12] :) [14:12] tsdgeos: done. thanks [14:12] Mirv: fwiw i reported the gsettings-qt thing and is now fixed upstream, but i think it needs a string of fixes, so we're basically stuck waiting for 5.2.1 i'd say [14:13] Mirv: if you have spare cycles you can try adding https://codereview.qt-project.org/#change,74547 [14:16] ricmm|sick: are we good for Cimi to put together the sidestage preview for design to play with ? [14:16] * kgunn think |sick can't be good [14:17] Cimi, re qmlscene, didn't work for me and qtcreator seems to heavy [14:18] kgunn: we have this list of bugs: http://pad.ubuntu.com/qP0HD1BUn4 that we're working through [14:18] kgunn, Saviq is qmlscene --desktop_file_hint=foo.desktop foo.qml how I should be able to launch an app from an adb shell, su - phablet? [14:18] assuming that the files are in the same dir [14:19] olli_, you need a full path to the .desktop file [14:19] ah [14:19] that might explain the core dump [14:19] olli_, yeah, it aborts 'cause it's rejected by unity-mir, due to not finding the .desktop file [14:19] greyback: yeah, i know its not "finished"...but design needed to do some testing of some sort specifically for sidestage [14:19] so plan was to just provide them a snapshot of our branche [14:19] s [14:20] kgunn: then it's basically functional === ricmm|sick is now known as ricmm [14:20] sorry, hadnt changed the name on freenode [14:20] * kgunn relieved ricmm didn't relapse [14:21] kgunn: as it stands right now there arent any big functional bugs, except Mir screenshoting garbage sometimes [14:21] mostly just flickers [14:21] Saviq, thx! === bregma_ is now known as bregma [14:21] ricmm: cool...we can live with that (known bug & someones working) [14:21] olli_, cheers [14:21] the flickers and related ones come from feb/march [14:22] not necessarily new [14:24] Cimi: so its just the three branches, onto a fresh N10 [14:24] prereqs have hit the archive [14:24] greyback: lets land platform-api and qtubuntu branches asap, agree? [14:24] greyback: set-dimensions ones [14:24] ricmm: +1 [14:25] can you do a review of each? [14:25] although we've built and rebuilt a billion times [14:26] greyback: ricmm ....so is what you're landing equivalent to "v2" or "v3" referred to in the sketchpad ? [14:26] ricmm: in progress... [14:26] not landing anything unity-mir yet, will try to fix some of the flickers right now [14:26] just need to pop to post office, back in 5 [14:26] k np [14:26] kgunn: but cimi can flash latest of the three branches and it will work [14:27] thanks... ricmm can he just use your zip at people.canonical ? [14:27] or he needs to rebuild ? [14:28] its better if he builds his own [14:28] with the latest image [14:28] and creates his own deployable set of debs [14:28] takes about 15 min on a n10 [14:28] Cimi: can you give sidestage a shot ? on tingting's manta [14:28] kgunn, sure [14:29] Saviq, if they are equal, I'd keep the shader then [14:42] Saviq: oh idiot, bad timing for me to run to shop, sorry I missed standup [14:42] greyback, nw [14:42] greyback, just add your things to the doc if you have anything new to report [14:44] dandrader: ok, now i remember why the if is needed and why the assert has to go [14:44] dandrader: or maybe you prefer me to reflow the code [14:44] dandrader: let me explain [14:45] Saviq: i just flashed device, but cmake isnt installed with run_on_device anymore? :/ [14:45] tsdgeos, explain in a comment in the code, replacing that assert [14:45] ricmm, to build my own debs, I simply run dpkg-buildpackage on the device? [14:45] dandrader: ok [14:46] Cimi: build them as you usually build debs [14:46] first platform-api [14:46] then qtubuntu and unity-mir === dandrader is now known as dandrader|lunch [14:46] whatever your preferred way is :) [14:46] ricmm, don't know any other way :) but there might be 100 different in debian world :) [14:48] dpkg-buildpackage it then [14:48] install deps first, a simple build-dep should do it [14:52] dednick, you probably have an old build-deps .deb on the device [14:52] dednick, drop unity8-build-deps*deb from shell/builddir [14:53] dandrader|lunch: comment added [14:58] ricmm: happy to top-approve your papi and qtubuntu [15:00] go for it [15:01] mzanetti: Saviq: you guys going to do https://code.launchpad.net/~aacid/unity8/tabbar_dash ? Don't want it to get stale again, it's a bit of a pain to merge [15:01] mzanetti, will you? [15:01] greyback: you can probably approve both, qtubuntu will block on dep wait for 0.20 platform-api [15:02] I think, I no longer know how these things work [15:02] who pushes it and what not [15:05] tsdgeos, http://ubuntuone.com/7dAW1FaKYID6PTx0DnealB - the chevron got cut off [15:05] hmmm [15:05] let me see [15:06] Saviq, chevron == > ? [15:06] Cimi, yeah [15:06] Saviq, it's on the theme [15:06] Cimi, probably the wrong word to use [15:06] Cimi, that's tsdgeos's branch [15:06] Saviq, tsdgeos I can point you to the right piece in the sdk [15:06] Cimi, and it got cut off by the search input [15:06] Cimi, it's there, overlaid [15:07] Saviq, obviusly [15:07] Saviq: having dependency problems with unity8-build-deps [15:07] Saviq, it's external to the button [15:07] Cimi: i know where it is, i remade the whole code and then the sdk guys remade it all again [15:07] Cimi, I think you misunderstand [15:07] button == the title of the current lens [15:07] ah ok [15:07] Cimi, it uses a tab-like header in there [15:07] ah ok, custom one [15:07] ricmm: probably, was just worried that jenkins ci might override the approval before papi landed. Have approved, let's see what happens [15:08] Saviq: tsdgeosack [15:09] Saviq: actually i don't think it's overlaid, at most it's "underlaid" [15:09] tsdgeos, it's overlaid by the text entry ;) [15:09] Saviq: yep [15:10] Saviq: i don't think there's a way to fix that [15:10] Saviq: well we could do something like what the old code did, but someone told me it was a bad idea :D [15:10] * Saviq wonders who [15:11] tsdgeos, aah wait [15:11] Saviq: i.e. the old code took into account the width of the current label to decide if put the search bar on the right or not, which meant that depending on the title length you'd have it there or not [15:11] which is pretty weird if you ask me [15:11] now we're back to the simple "if it's wide enough it'll be there or not" [15:11] don't remember who i talked to regarding this tbh [15:12] may not be even in the company anymore :D [15:12] tsdgeos, think we could find out the longest label and base on that? [15:13] tsdgeos, or well... let's just say there needs to be at least 80 GU width? [15:13] tsdgeos, all labels need to fit in 40GUs to be visible on the phone [15:13] tsdgeos, and the search entry is 40GU wide [15:13] tsdgeos, so if we say it only goes to the side when > 80GU [15:14] that works [15:14] probably 90 [15:14] tsdgeos, 80 is fine [15:14] ok [15:14] sure i think that's a saner decision [15:14] tsdgeos, only other thing I thought of (which is not sane) would be to find out the longest label [15:15] tsdgeos, so no, let's not [15:15] updated to 80 [15:30] tsdgeos: should the >80GU thing fix this? http://i.imgur.com/qxsMKRY.png [15:31] mzanetti: no [15:31] mzanetti: that's just unfortunate widths [15:32] ah no [15:32] tsdgeos: could we fade it out so it doesn't look so cut off? [15:32] or yes [15:33] mzanetti: we could, but do we do that kind of fading anywhere? [15:33] fair point [15:34] tsdgeos: we do in the launcher [15:34] mzanetti: with text? [15:34] sort of [15:34] the icons [15:35] if they don't fit they are are folded and faded out [15:36] right [15:36] but it's the whole icon [15:37] in here we'd need some effect that applies to say, the last gu of the right hand side of the tabbar [15:37] yep [15:37] something like the BB10 phone does and Saviq said it sucks :D [15:38] ;D [15:38] that's just Saviq and BBX [15:38] anyways... I think we should ask jouni on this [15:38] * mzanetti continues with the review [15:38] mzanetti: we can, it'd be something for the SDK probably i'd say [15:39] I tend to agree [15:39] yup [15:48] tsdgeos, FTR, I said it sucked because it was overused, IMO [15:48] ok [15:48] i still like it [15:51] tsdgeos: do you really need the dashContentListHolder ? [15:52] yes? [15:52] why wouldn't i? [15:54] tsdgeos: ping [15:54] re https://code.launchpad.net/~aacid/platform-api/papi.rules.typo/+merge/182354 that saviq re-pinged me about the other day [15:54] libplatform-api-headers is the right package to ship that on [15:55] mzanetti: i may be able to refactor the code not to, but as it stands now, yes i need it, what's your problem with it? [15:55] ricmm: ok, i'll put it in there [15:55] tsdgeos: sorry... got confused [15:55] all right [16:16] I'm must be going crazy; neither of my devices are turning on [16:19] ricmm: https://code.launchpad.net/~aacid/platform-api/papi.rules.typo/+merge/182354 passed, you can now approve [16:23] mterry, no, they're just completely discharged [16:24] mterry, connect to a charger, leave them be for a half hour [16:24] Saviq, I figured that, but it's been a while [16:24] mterry, sometimes they really take their time [16:24] mterry, IIRC I've actually had more luck connecting to a PC [16:24] Saviq, maybe I just haven't waited long enough, yeah [16:24] mterry, /laptop [16:24] mterry: use the n10 charger for all [16:25] 2A juice makes the world a better place [16:26] shit [16:28] how do I keep a column from rearanging it's cells when an image in a top cell (in a 2line column) goes to visible:false [16:30] tsdgeos: made a few more small comments [16:30] olli_: put the Image inside an Item [16:30] olli_: or just change it's opacity instead of visible: false [16:44] mzanetti: answered your comments, tell me if i'm making any sense [16:45] tsdgeos: but isn't the pageHeader property in GenericScopeView always pointing to the one instance we have? [16:47] mzanetti: ah yes [16:48] the pageHeader property points to the single one [16:48] not the pageHeader of ScopeListView inside GenericScopeView but the pageHeader of GenericScopeView [16:48] what you would prefer me doing? [16:49] so I still think the two signals for headerHeight/SizeChanged are not needed [16:49] and it could be just set from within GenericScopeView just like all the rest [16:50] ah [16:50] i see what you mean [16:50] sure [16:50] can do that [16:50] please. the rest looks good to me === dandrader_ is now known as dandrader [16:54] tsdgeos: also, please do another check for warnings. for example: "GenericScopeView.qml:284: TypeError: Cannot read property 'category' of null" [16:55] just noticed right now [16:55] tsdgeos: the file needs to install to /usr/include/ubuntu/application/ [16:55] where it currently installs bsically, check on your local system where it lives [16:55] mzanetti: is this introduced by this? [16:55] if you just set the file it will install to the wrong loc [16:55] ricmm: ok [16:56] ricmm: honestly i'm going to abandon it and will let you guys fix it [16:56] makes no sense you telling me what to do [16:56] me doing it wrong [16:56] and then you telling me what to do agai [16:56] tsdgeos: change branch ownership to phablet-team [16:56] :) [16:57] seems it isnt [16:58] ricmm: done [16:58] ricmm: https://code.launchpad.net/~phablet-team/platform-api/papi.rules.typo [17:00] mzanetti: pushed [17:00] thanks [17:01] whaaat [17:01] the expand_expand test failed again ^_ [17:02] meh [17:02] tsdgeos: :D wanna know why I told you the thing about the warnings? [17:02] file:///home/mzanetti/Development/reviews/tabbar_dash/qml/Dash/GenericScopeView.qml:306: TypeError: Cannot set property 'y' of null [17:02] file:///home/mzanetti/Development/reviews/tabbar_dash/qml/Dash/GenericScopeView.qml:302: TypeError: Cannot set property 'height' of null [17:03] because I did exactly the same change here and got those two [17:03] that tricked me into believing the others were your's too [17:03] well i changed that [17:04] maybe the tests needs adjusting? [17:04] no... its not the tests. happens if you run it [17:05] tsdgeos: your latest change just misses a if (scopeView.pageHeader) [17:05] mzanetti, will try, thx [17:05] we don't really have one of those, but ok [17:06] tsdgeos: ? [17:06] yeah... actually a good question why this happens at all [17:08] ahhhhh [17:08] booo [17:09] not sure we're talking the same language today :D [17:09] ok, test fixed [17:09] thanks [17:09] the pageHeader has changed [17:09] so the thing that was looking for the pageheader in the fix i made this morning was wrong for this one [17:10] makes sense? [17:12] and that said, i have taxes stuff to do [17:12] so see you tomorrow [17:27] eh...wtf is going on with mir and unity8? installing mir 0.1.3 removes unity8... ? [17:28] all sorts of conflicts going on. === davmor2_ is now known as davmor2 === balloons_ is now known as balloons [17:36] Saviq: ping [17:36] dednick, pong [17:37] Saviq: hey, there seems to be a problem with latest mir and unity8. when doing build deps for unity8, it's trying to get mir 0.1.3, but that conflicts with current unity8 so it cant... [17:38] current installed ver of mir is 0.1.2, but it's trying to get 0.1.3 dev packages. [17:40] this is presumably only a problem on a clean device... [17:40] or if you do an apt upgrade. [17:40] s/upgrade/update [17:44] dednick: you may have to wait until this lands: https://code.launchpad.net/~ricmm/qtubuntu/papi-setdimensions/+merge/198498 [17:44] dednick: oh wait no, please ignore [17:44] different issue, crossed brain wires [17:45] dednick, hmm, there is no mir 0.1.3 in distro yet https://launchpad.net/ubuntu/+source/mir [17:45] dednick, maybe you have daily-build PPA enabled? [17:45] dednick, and unity-mir still only requires 0.1.2 http://bazaar.launchpad.net/~mir-team/unity-mir/trunk/view/head:/debian/control [17:46] dednick, the bump is only under review now https://code.launchpad.net/~kgunn72/unity-mir/mir-deb-bump-0.1.3/+merge/200349 [17:46] ah. damn. it's that stupid writable image phablet-config command [17:46] dednick, what's it do? [17:46] Saviq: phablet-config writable-image --ppa ppa:ubuntu-unity/daily-build [17:47] dednick, ;) [17:48] dednick, apt-cache policy is always helpful to see where stuff comes from [17:51] Saviq: thanks [17:51] wasted half a day... === balloons is now known as Guest99854 [17:59] * greyback eod === alan_g is now known as alan_g|EOD === balloons_ is now known as balloons === ssweeny` is now known as ssweeny === salem_ is now known as _salem [21:33] hi all i was hoping somebody could help me figure out how to remove app labels from desktop icons? [22:54] So the edge demo got disabled somewhere down the line [22:55] Saviq, was that intentional? ^ [23:00]