=== shiznix_ is now known as shiznix [05:54] where is the documentation for python bindings to appindicator? all the links here are broken: https://unity.ubuntu.com/projects/appindicators/ [08:09] morning [08:09] i think i either caught the ubuflu or the heathrow flu [08:10] feeling quite sick [08:36] tsdgeos, get a swap day, take it easy [08:36] tsdgeos, *in* Heathrow meaning you slept at the airport? [08:37] meanig i didn't sleep [08:37] but yeah :D [08:37] jeez [08:37] got a plane for sunday at 6am [08:37] so i had to be there at 4am anyway [08:37] so didn't make much sense trying to get somewhere to sleep [08:38] bought last Dan Brown's book and read it all :D [08:38] tsdgeos, :D how was it? [08:38] i actually liked it [08:39] tsdgeos, either way - get a swap day for that - especially if you're feeling sick [08:40] tsdgeos, and also, there's a flight delay compensation to be fought for within the EU [08:40] actually it was no that my flight got delayed [08:40] is that awesome piccadilly line broke [08:40] and i had to take an 1:30h detour [08:40] so i got there late [08:40] my plane actually flew as scheduled [08:45] tsdgeos, ah [08:45] that's "it's easier to get to Heathrow" for you... [09:14] MacSlow, hey, sorry to hear about your N10 [09:14] MacSlow, there's one thing, though - company hardware is somewhat insured - would be good to verify what's covered [09:23] Saviq, yeah... thanks... Murphy struck hard this weekend for everybody... in some way or another [09:23] MacSlow, yeah [09:23] Saviq, hm... who would know details about this? [09:24] MacSlow, start with msm [09:24] Saviq, ok [09:43] Saviq, hey boss, I'm looking for week's tasks [09:44] Cimi, "week's"? "this week's"? [09:44] this week [09:50] Cimi, how 'bout bug #1259005 [09:50] bug 1259005 in unity8 (Ubuntu) "Swiping over SEARCH header in top left of phone Dash does not show search" [Undecided,Confirmed] https://launchpad.net/bugs/1259005 [09:51] Saviq, ok [10:21] Saviq: did we switch to use the fake applications intentionally with ./run on the desktop or is there some import path breakage going on too? [10:21] mzanetti, we never had non-fake ones ;) [10:21] err... not the fake apps, but the fake scopes content [10:22] mzanetti, do you have unity-plugin-scopes installed? [10:22] ah... right... going back to qt 5.0 seems to have removed all that stuff [10:22] thanks [10:23] mzanetti, it now picks stuff up from ./plugins, /usr/lib/*/unity8/qml, ./tests/mocks - in that order, more or less [10:23] mzanetti, so we override system-wide plugins and fall back to mocks [10:23] s/system-wide/installed/ [10:23] ah ok. makes sense [10:25] biab, post office === iahmad is now known as iahmad|afk [11:28] greyback, ping [11:29] dednick, welcome back! [11:29] Saviq: :) thanks [11:29] dednick, were you flying during UK's air traffic shutdown? === _salem is now known as salem_ [11:32] Saviq: pong [11:42] Saviq: no, missed it by a day luckily [11:42] i arrived a few hours ago :) [11:42] dednick, straight into the office, eh? ;) [11:42] *home office* ;) [11:54] Saviq, does qmltestrunner run with touch emulation events? [11:54] Cimi, we have a custom touchtestrunner that does [11:55] Cimi, used by the indicator tests, for example [11:55] ok [11:55] maybe I can use that [11:55] Cimi, I *think* [11:55] Saviq, cause was failing [11:55] Cimi, either way - panel tests are fine with edge drags [11:55] Cimi, so that must work some way ;) [11:55] Saviq, I added this edgedragarea [11:56] Saviq, but my tests fail [11:56] Cimi, added? there is an EdgeDragArea in the panel already - would be best to use that one probably [11:57] Saviq, it's only for the indicators row iirc [11:57] Saviq, I'll have a look [11:57] Cimi, I'd rather not, if possible, have two separate ones - maybe some refactoring is in order [11:59] Saviq, in the meanwhile I'd like to have tests working :) [11:59] Cimi, as in trunk tests? [11:59] Saviq, nope referring to this search swipe [12:00] Saviq, took me 5 mins to get the feature in [12:00] Saviq, still debugging the tests failing [12:00] Cimi, ok sure [12:01] Saviq, ok was this mouseClick -> tap [12:01] Saviq, mouseFlick -> touchFlick [12:01] Saviq, when you use touch events [12:01] Cimi, yeah, thought we have those :) === MacSlow is now known as MacSlow|lunch [12:17] Saviq, having one big edgedragarea common won't simplify things [12:18] on the tablet it will be huge [12:18] while currently it's only 40gu [12:18] Cimi, right [12:19] also you have to abstract properties from the indicators to the panel, making them 'public' [12:19] Cimi, ok, fine with split areas [12:19] Cimi, makes sense when they're spread out on the tablet === boiko_ is now known as boiko === tvoss is now known as tvoss|lunch [12:49] Saviq, you remember the tolerance area of clicks? [12:49] how big is it [13:00] Cimi, you mean tap vs. long-press? [13:00] Saviq, I meant [13:00] Saviq, I'm faking mouse clicks in the edgedragarea [13:00] Cimi, that depends [13:00] Saviq, so I don't have to use two input areas [13:01] Cimi, EdgeDragArea.qml has them all set up [13:01] Saviq, I have mouseAbs (touchx - initialtouchx) < tolerance [13:01] Saviq, does not have onClicked IIRC [13:01] Cimi, clicks just go through the EdgeDragArea [13:02] Saviq, so you want me to use a edgedragarea AND a mousearea? [13:02] thought faking was better === alan_g is now known as alan_g|lunch [13:03] Cimi, yeah, just have a MouseArea below - that's how the EDA is supposed to be used [13:03] Cimi, or well, see how the IndicatorRow has those [13:03] Cimi, it does react to press [13:04] Cimi, and it seems it's the EDA that handles it [13:04] Saviq, they use tolerance 0 for swipes [13:04] Saviq, so you can swipe right to reveal [13:05] or swipe up [13:05] it seems wrong [13:05] to me [13:05] for reference, my current code http://bazaar.launchpad.net/~cimi/unity8/searchIndicator-swipe/revision/582 [13:06] two ways to show search indicator [13:07] 1. click within units.dp(2) from the original click [13:07] 2. swipe down of at least units.gu(1.5) === iahmad|afk is now known as iahmad [13:14] Saviq, edgedragrea eats mouse clicks [13:14] gnam gnam, it's hungry area [13:18] Saviq, forgot to bring up one topic last week - running apps [13:18] Saviq, i still don't want those in a scope [13:18] it's just silly [13:19] mhr3, they'll go away [13:19] mhr3, no worries, we might end up without them there completely [13:19] wooo :) [13:24] greyback: hmm... should ApplicationManager.startApplication() cause ApplicationManager.focusedApplicationId to change eventually? [13:28] Saviq, and one more, tests in the scopes plugin - gtest or qtest? [13:29] or rather do you see a reason why qtest would be a better choice? [13:29] mhr3, your call - we use qtest mostly, as it's just better integrated [13:29] mhr3, but I'm completely fine on gtest [13:29] Saviq, tell me more about the integration [13:30] mhr3, stuff is named Q* [13:30] ;) [13:30] mhr3, no really, don't see a reason [13:31] mhr3, maybe the fact that it launches QApplication automagically [13:31] mhr3, but it's not like you can't do it yourself if necessary [13:32] i'm all for simpler life :) [13:37] hmm, think i'll stay with qtest just to have QTRY_* [13:37] will be handy for the highly-async scope tests [13:39] mhr3, maybe the SignalSpy - useful, too [13:43] test rows/column might be useful too [13:43] and the nice output formatting :) === alan_g|lunch is now known as alan_g === Pici` is now known as Pici [14:04] Saviq: is there something special in regard to the order of the recent applications grid? as in, do mainstage apps need to be before sidestage ones? [14:05] or could I get rid of firstModel and secondModel and just show all of ApplicationManager's recent apps in whatever order the appmanager holds them [14:07] Saviq: snowboarding...ok...now i hate you [14:08] kgunn: finally back home? [14:08] mzanetti: sorry was at lunch. ApplicationManager.focusedApplicationId should update yes [14:08] mzanetti: yep, yesterday afternoon [14:08] mzanetti: heard you made with no issues...lucky === MacSlow|lunch is now known as MacSlow [14:09] kgunn: mostly, yes. 45 mins late [14:09] greyback: no problem. ok. seems our various ApplicationManager implementations start to drift apart [14:10] will fix [14:17] mzanetti, TBH it should be recency [14:18] mzanetti, but I don't think we have that proper yet [14:18] Saviq: I agree. but I think it'd be better if the applicationmanager is sorted itself according to that [14:18] but anyways, dropping that firstModel, secondModel hack [14:18] mzanetti, I'd be worried about reshuffling the launcher all the time [14:19] mzanetti, not sure what's the plan there [14:19] Saviq: no. the launcher doesn't follow that order [14:19] Saviq: as when an app is pinned it needs to be there, not where it is in the app stack etc [14:20] mzanetti, yeah exactly, but what about recent apps in launcher [14:20] mzanetti, not the pinned ones [14:20] Saviq: well, they are appended as they show up [14:21] Saviq: and stay there until either moved or removed from the recent list [14:21] mzanetti, that'd be good I think [14:21] I've dropped shell's knowledge about multiple stages... that's why I asked if we need to have that distinction in the recent apps grid [14:21] mzanetti, still, if we go for split right edge [14:21] mzanetti, we need separate models for main and side stage [14:21] Saviq: then we still can put a sortfilterproxymodel on top [14:22] mzanetti, or maybe the other way 'round - let app manager give up three models? main, side, merged? [14:22] but I'd say where we need it.. instead of keeping two separate lists in shell.qml and aggregate them again all over [14:22] Saviq: so when we designed the appmanager api we went for this: [14:23] appmanager itself is a model of all apps, having a role "stage" [14:23] kgunn, sorry... [14:23] so we could easily put a sortfilter on top [14:23] kgunn, it's not like you don't have places to go to do that [14:23] mzanetti, yeah, I'd rather it be provided by the app mgr itself [14:24] mzanetti, or well maybe it really doesn't matter in that case [14:24] ok well, sure... we can always create instances of filtermodel's inside the appmanager if really wanted [14:24] but yeah... the question was more the other way round === tvoss|lunch is now known as tvoss [14:25] mzanetti, yeah, we need a merged model from app manager indeed [14:25] Saviq: yeah... that's the thing... we do have that [14:25] Saviq: then ApplicationManagerWrapper was splitting it up [14:25] the split thing was used everywhere and merged again everywhere with fancy javascript [14:27] mzanetti, yeah, that's why I'm thinking we should simply have three - main, side, merged [14:31] Saviq: actually I was thinking to not have that distinction anywhere except inside the tablet stages code. Given that not all devices will have that [14:33] mzanetti, yeah maybe [14:33] greyback, dednick_ standup [14:53] tedg: hello! [14:53] tedg: not sure if it's still super valid, but we've been discussing and wondering if the ust requirement for upstart-app-launch could be somehow made optional? [14:55] tedg: by ust I mean liblttng-ust of course [15:09] mzanetti, hey, I can has you to look at https://code.launchpad.net/~saviq/unity8/new-scopes-integrate-card/+merge/197930/comments/459061 ? [15:13] heh, was just about to ping you about that [15:15] Saviq, how is the visibility of items going to work? [15:15] for example right now the height is too much, cause there's no subtitle, summary etc [15:15] mhr3, items meaning price etc.? [15:15] mhr3, that depends [15:16] mhr3, in a static-height scenario (grid, carousel), we'll to lay out a maximum card size and stick to it [15:16] Saviq, right, so for those it's about the components [15:16] but for the other ones it's about both components and the data [15:16] Saviq: ack [15:17] mhr3, yes, prices and attributes will be just stacked next to each other - based on contents [15:17] mhr3, but still we'll have to cater for the biggest case [15:17] mhr3, for variable-height/width (journals) it will be what it will be in reality [15:18] Saviq: line 117, 118 is some more commented stuff. reason? [15:18] mzanetti, because the card does not yet have those signals [15:18] mzanetti, but it will need them - can add FIXME [15:18] ok [15:20] mzanetti, shall I add FIXME? [15:21] Saviq: nah... given that you will change this soon enough in any case, no need to nitpick on it [15:21] mzanetti, yup [15:21] Saviq: I'd prefer to investigate the crash in the other fixme as I have a feeling otherwise that fixme will be in there for all eternity [15:21] mzanetti, oh no it won't, but yeah [15:22] mzanetti, we need the sourceSize there, and need it working [15:22] yeah... [15:22] mzanetti, otherwise we'll load huge images [15:22] mzanetti, so I basically need to try it out with 5.2 [15:22] seeing what sort of fixme's I came by in the application stuff today :D [15:23] Saviq: ah damn... I just went back to 5.0 as too much other stuff was crashing with 5.2 [15:23] otherwise I could have tested for you [15:25] mzanetti, will manage [15:25] mzanetti, we shall have it resolved before that whole thing gets into lp:unity8 === greyback is now known as greyback|london === alan_g is now known as alan_g|tea [15:39] although the sooner it would get there, the better :) [15:40] would lower the barrier for testing it [15:45] sil2100, We'd prefer to leave it in so that we could get timing results on images. Is it causing a problem? === alan_g|tea is now known as alan_g [16:08] mzanetti, I'm worried the "Qt.test.qtestroot" missing thing might be that we're trying to run the testcase under qmlscene [16:08] Saviq: not following [16:09] mzanetti, remember when you tried to do make tryCard under 5.2? [16:09] ah, ok [16:09] hmm... that'd be bad [16:09] yeah, we'd need to find a way to run it under qmltestrunner [16:09] which, btw, could be doable [16:10] yeah, quite sure. It's just the input thing right? [16:10] mzanetti, well, not even [16:10] mzanetti, for tryFoo we're not testing at all [16:10] mzanetti, just loading the QML and let you play with it [16:10] mzanetti, so we'd need to be able to the same under qmltestrunner [16:10] mzanetti, maybe our custom runner with a command line option to not actually start the tests [16:11] mzanetti, or set a context prop, for that matter [16:11] that we'll bind to TestCase.when: [16:11] which would mean we'll load everything under the test runner, just we won't start the tests [16:11] Saviq: but wait... iirc I had the crash when doing make testCard [16:11] mzanetti, yeah, unrelated [16:12] mzanetti, crash is there indeed [16:12] ppa:canonical-qt5-edgers/qt5-daily is for qt 5.2? [16:12] or we have others? [16:12] Cimi, qt5-beta2 [16:12] mzanetti, but it doesn't even start with qmlscene, complaining about Qt.test.qtestroot not being there - probably something qmltestrunner imports / adds an import path / something [16:13] mhm [16:36] mzanetti, the only symbols I got out of the qmltestrunner crash: [16:36] http://pastebin.ubuntu.com/6546538/ [16:36] mzanetti, look familiar? [16:36] yeah. the JIT was involved here too [16:37] not sure if its the exact same thing [16:37] qv4functionobject.cpp... yeah, I think it is [16:39] mzanetti, meaning https://bugs.launchpad.net/unity-api/+bug/1258057 ? [16:39] Ubuntu bug 1258057 in Unity API "unity-api fails to build against Qt 5.2" [Critical,New] [16:40] Saviq: right... I didn't try to debug testCard... so that must be the one [16:44] mzanetti, thanks, will try to add more info [16:46] mzanetti, https://bugreports.qt-project.org/browse/QTBUG-33658?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel [16:47] looks legit [16:49] Saviq: hmm... looks similar. but I wouldn't be able to say it's this one [16:51] mzanetti, yeah of course - I'll try to build it locally with the patch applied [16:52] ls [16:57] Saviq: do you have time for a quick chat about Unity 8 desktop mode? [17:02] mhall119, can you please put something in my calendar for tomorrow? [17:02] mhall119, I'm almost at the door today [17:03] mhall119, also, depending on the topic I might have more questions than answers for you... [17:32] Saviq: ok, I'll find us a time tomorrow [17:32] mhall119, thanks [17:34] Saviq: done [17:45] Saviq: hi. I am running run_on_devce and the build-deps for unity8 is missing deps lcov and gcov. are these imporant ? [17:46] untitled1, that's just a warning [17:46] camke warrning yeah but what are they used for ? [17:46] untitled1, unless you want to see test coverage, no - they're not important [17:46] thanks [17:46] untitled1, http://en.wikipedia.org/wiki/Gcov [17:47] I also had to mount my devce as dpkg would not work with out it === bschaefer_ is now known as bschaefer [17:47] maybe could add to bash script ? [17:47] untitled1, you mean make writable? [17:47] yeah [17:47] untitled1, that's dangerous, we don't want to do it by default [17:47] what is up with that ? [17:47] untitled1, we could bail out if not writable, though [17:48] I can not run_on_device with out that ? [17:48] no [17:48] because of the build deps and what not [17:48] yes [17:48] cmake and many other thins are not installed by default [17:48] untitled1, having a writable rootfs on a phone is not feasible - battery outage could result in unbootable phone [17:48] untitled1, exactly [17:48] untitled1, it's really like making the phone go into developer mode [17:48] I see [17:48] untitled1, we have plans to make it better / less invasive on your installation [17:48] thanks Saviq [17:49] untitled1, ideally you should be able to revert what you did during the development session [17:49] untitled1, at least to the rootfs [17:49] maybe I should make a qprocess app for this like a gui that uses them scripts [17:49] or add to qtcreator not sure what are your thoughts ? [17:50] at that what do you think about qprocess ? [17:50] untitled1, a qtcreator target could be used, sure [17:50] untitled1, although for that we'd rather cross-build and transfer pre-built onto the device [17:51] untitled1, this should happen relatively soon [17:51] Saviq: are the plugins for the creator still in source or are they outside branchs now ? [17:51] untitled1, the ubuntu plugins? [17:51] Saviq: yeah [17:51] untitled1, they are separate - and will mostly remain so [17:52] untitled1, until we get upstream support for Ubuntu in QtCreator at least [17:52] I built ubuntu-ui-toolkit to qt5.2 and might go after the creator today [17:52] qtcreator takes forever to build lol [17:52] untitled1, naah, qtcreator is nothing - try qt itself ;) [17:53] lol [17:53] esp on arm [17:53] done that many times when android was getting "wheels" on Qt [17:55] Saviq: one of the things that is hard for me to deal with is when I press the qt home screen and then press develop to go to last session and it opens like 5 windows that I can not close :( know any work a rounds ? [17:55] untitled1, here's a branch I did off of qtcreator 3.0-beta [17:55] https://code.launchpad.net/~unity-team/kubuntu-packaging/qtcreator-30 [17:55] thanks [17:55] untitled1, it should build now - importing rc could be useful, though [17:56] untitled1, ough [17:56] Oo [17:56] untitled1, if that's still the same in QtC 3.0 [17:56] untitled1, that's a bug that needs fixing [17:56] untitled1, in the mean time you can use File > Sessions or so [17:57] Thanks I think that I will build to rc and see what is going on in the welcome plugin [17:58] untitled1, actually, let me quickly update that branch to rc === mhall119 is now known as mhall119|afk === mhall119|afk is now known as mhall119 === salem_ is now known as _salem