[00:01] racarr: https://code.launchpad.net/~ricmm/unity-mir/shell-surface-monitor/+merge/189743 [00:01] kdub: what is interesting, is that i can get 80-100% cpu pretty consistently on top when i scroll the application lens on the dash with suggested apps expanded [00:01] that kinda fixes all (plus your shell keyboard focus for keys) [00:02] ricmm: *deep bow* [00:02] ricmm, racarr and why did we not have that by default? [00:02] s/by default/to start with/ [00:02] good question [00:03] we do in SF [00:03] because through the whole twisted chain of input [00:03] it was at one point decided we wouldn't use it in mir [00:03] then I forgot it was an option because we've never used it in mir [00:03] * Saviq was under the impression we couldn't do it [00:03] ugh [00:03] its still not... [00:03] well [00:03] I mean with the monitor stuff [00:03] you end up with doubled events [00:03] i.e. your gestures end up going throuhg to the app as well [00:03] yup [00:04] well not to the app, they go to the shell [00:04] sure [00:04] but the shell can still reorganize its input regions [00:04] racarr: ricmm ...so net-net...forget the branchs "default-input-focus" and "mir-map-audio-keys"....and work on landing "shell-surface-monitor" [00:04] and it wont be delivered events outside of those regions, no? [00:04] kgunn: Ok well either way we get mir-map-audio-keys [00:04] no no, this is just an idea [00:04] and we still need the keyboard focus [00:04] ricmm: No, it will get all events if its a monitor (irregardless of input region) [00:04] as its a different path [00:05] ricmm, racarr, I think we'll get dupe events for InputAreas, though? [00:05] yes [00:05] well not really dupe, as those events will be eaten and wont make it to the application surface [00:05] ricmm, dupe for shell [00:05] in the non-monitor case [00:06] in the monitor case, the shell should only get one event (the monitor event) [00:06] ricmm, once we get them due to monitoring, second time due to InputAreas set up [00:06] but the shell will never eat events... [00:06] I think the input areas wont get input if monitoring is on [00:06] because the input areas are just a filter on the surface input [00:06] right, so won't be able to block input - no go [00:06] but would have to do some investigation [00:07] mm [00:07] thats the problem [00:07] I really need to stretch my legs... [00:08] be back soon [00:08] Saviq: right, thats the thing... in the android app manager InputArea's also injected a null input area in the regions specified [00:08] so that the event wouldnt hit the application [00:08] although I think that is happening right now [00:09] ricmm, except the shell *only* gets input in InputAreas [00:09] whats a good way to test it? [00:09] ricmm, it's a flipped model [00:09] in the SF model the shell gets input everywhere [00:09] ricmm, under SF, we got all input, and were able to tell SF to not deliver events to apps in certain areas [00:09] ricmm, under Mir, we add areas we want input for, and those don't get to the app [00:10] ricmm, so currently it's either-or, not both, as in SF [00:10] ricmm, either the app gets the events or the shell takes them over [00:10] I think we are still blcokign where theres an InputArea [00:10] ricmm, yes, of course [00:11] ricmm, that's what I mean [00:11] so whats the issue? [00:11] ricmm, we need places where we listen, but *don't* block [00:11] ricmm, i.e. monitor - but it can't be a blanket like your branch [00:11] we listen in all places whrees theres no InputArea [00:11] ricmm, no can do, according to what racarr said [00:12] ricmm, once you're a monitor, you're a monitor and that's it, you don't have control over the events [00:12] ricmm, i.e. InputAreas don't matter anymore [00:13] ricmm, and don't receive any input - so can't block it, either [00:13] anyway... [00:13] o/ [00:13] more tomorra [00:13] one sec [00:14] well it doesnt seem to be true [00:14] Saviq: I just tested and I'm not receiving input events in the InputArea's [00:14] in client apps [00:14] where an input area is set, the shell sees the monitor event but the app is not delivered anything [00:14] so it looks like it does block the region [00:15] Saviq_: two saviqs ! [00:17] Saviq_: yea, input is blocked fine by the InputAreas [00:17] racarr: ^ [00:22] ricmm: hey...do you know if we're on a motley, trinity or M-kernel ?...or none of the above ? (or do any of those even mean anything to you?) [00:23] they mean nothing to me [00:23] what are they? [00:28] ricmm: so...i was reading about gpu clocking on xda...http://forum.xda-developers.com/showthread.php?p=36373811 [00:29] ricmm: i know we use CM kernels...wondered if they were related in any way...seems not.... [00:29] certainly not a random kernel [00:29] only CM-standard ones [00:30] racarr: looks like being a monitor helps to get key events... but only when the application receives them [00:30] for some reason the webbrowser is not getting such events [00:30] while calculator is [00:30] : [00:30] ooohhhhh [00:32] I get it now, the webbrowser spawns a secondary client which is the one in focus, but it is a transparen one [00:32] thats why events arent being delivered to the main client [00:32] but all other apps work fine and you can even see the volume keys [00:32] although it needs the qtubuntu branch to fix the keycode mappings [00:33] racarr: kgunn so I think we can agree that a one line fix is less risky than the injector interface MR + associated injection code in unity-mir [00:33] for now, post 13.10 is another story [00:39] ricmm: yes...one line sounds better than 2 related code changes in 2 diff components :) racarr ? [00:45] well its not just that [00:45] it needs the focus-setter branch as well [00:45] otherwise there will never be key evens unless a window is focused [00:45] but at least it saves us the injector stuff, that one sounds scary [00:45] for pre-freeze [00:52] kgunn: sending this out in an email [00:52] as I need to bounce [00:53] racarr: ill be back later for further discussion [02:02] Back for potential further discussion === _salem is now known as salem_ === salem_ is now known as _salem [03:50] tsdgeos: pong === _morphis is now known as morphis [07:48] Saviq: https://code.launchpad.net/~aacid/unity8/start_ssh [07:48] tsdgeos, yup, saw that [07:48] ok [07:52] pete-woods, ping [07:52] Saviq: hi [07:52] pete-woods, hey, think we should get an ABI bump on upstart-app-launch? [07:53] pete-woods, was able to build unity-mir against u-a-l trunk yesterday and it didn't pull the new library [07:53] and was missing symbols for the observers [07:53] Saviq: er, I think you have me confused with someone else. I have nothing to do with that lib [07:53] pete-woods, orly? [07:53] pete-woods, sorry, then :) [07:54] Saviq: I can certainly do an MR for you [07:54] pete-woods, right, it's all Ted's [07:55] okay [07:55] pete-woods, we should probably wait for him, not sure what his policy is on the -dev package version [07:55] as it's slightly different than the default [08:04] Saviq: on the "locking" thing, i think i know how to fix it (in Qt), not sure it's the proper fix though (and not sure it'll work, it's still all on my head), but at least it's making sense in my head :D [08:07] tsdgeos, ROCK ON [08:16] Mirv_, ping [08:17] Saviq: pong [08:17] Mirv_, hey, do you know if unity-mir is going to be published? [08:18] Mirv_, or are we waiting for more? [08:18] Saviq: I don't know, I'd guess landing of it would be talked in 10 mins in our call [08:18] Mirv_, ok thanks [08:19] "trying to land first thing in the morning", the unity8 part was to me and my morning is earlier, thus it's in [08:20] lool may land the unity-mir part then [08:28] Mirv_, yeah, we might actually wait for the two more fixes down below [08:31] Saviq: just let lool know [08:33] Mirv_, k, will do [08:35] aaaaaaaaaaaaaandddddddd it's fixed [08:35] * tsdgeos does the "i'm fixing too much of Qt" dance [08:41] Saviq, mind taking a look at this? https://code.launchpad.net/~nicolas-doffay/unity8/dismiss-keyboard/+merge/188599 [08:42] tsdgeos, \o/ [08:42] nic-doffay, you'll need to find someone else, I'm swamped [08:42] Saviq: http://pastebin.kde.org/pgn5xeua9 [08:42] creating the codereview entry now [08:42] let's see what the guys think [08:43] tsdgeos, awesome [08:45] Saviq, cool. [08:45] Cimi, ↑↑ [08:53] Cimi, can you do a small review for me dude? [08:53] nic-doffay, y [08:59] mzanetti: greyback: photo of the booth? [09:03] Cimi, https://code.launchpad.net/~nicolas-doffay/unity8/dismiss-keyboard/+merge/188599 [09:03] ta [09:03] mzanetti, ping [09:24] Saviq: done at https://codereview.qt-project.org/#change,67523 [09:25] tsdgeos, distro-patch straight away? [09:25] hmmm [09:26] Saviq: tbh i think it should be fine, but then it can always be an awful patch [09:27] otoh with devdays around i'm not hoping for people to review the patch "soon" [09:27] if this is something we want to have [09:27] i'd go for the distro patch i guess [09:28] distro patch! :) [09:32] tsdgeos, do [09:32] * tsdgeos does [09:42] Mirv_: https://bugs.launchpad.net/ubuntu/+source/qtdeclarative-opensource-src/+bug/1236765 [09:42] Ubuntu bug 1236765 in qtdeclarative-opensource-src (Ubuntu) "[PATCH] Fix infinite loop in QQmlIncubator::forceCompletion" [Undecided,New] [09:42] Saviq: ↑↑ === iahmad is now known as iahmad|afk [09:50] Saviq, quick question, any idea what file the category expansion is handled for the scopes? [09:54] tsdgeos, tx! [09:54] nic-doffay, E_SYNTAXERROR [09:55] dandrader, want another ~simple task re: input? [09:55] dandrader, with packages from http://people.canonical.com/~msawicz/mir-input/ [09:55] Saviq, looking into this bug: https://bugs.launchpad.net/unity8/+bug/1224552 [09:55] Ubuntu bug 1224552 in Unity 8 "[Dash] Category expansion transition has varaible speeds" [High,Triaged] [09:55] Saviq, yeah, I was actually looking at what to do next, considering the scramble to get unity in mir as default... [09:56] dandrader, input goes through OSK when in shell [09:56] dandrader, we have a MouseArea to catch that in Shell.qml [09:56] dandrader, but unity-mir does not have ApplicationManager.keyboardVisible .keyboardHeight implemented [09:57] dandrader, it does, though, have OSKController [09:58] dandrader, we need to get the visible and height out of the OSKController [09:59] Saviq, didn't get the "input goes through OSK when in shell" part [09:59] dandrader, search in dash, type - you'll launch stuff from the dash [09:59] dandrader, both OSK and shell gets the input, so we need to eat it behind the OSK in Shell.qml [10:00] ah, ok. [10:00] we are starting to have a scary amount of workarounds :) [10:00] dandrader, yeah... [10:00] eh, flashing failed for me, can get into recovery, but phablet-flash doesn't work from there, ideas? [10:01] ehm, not here === davmor2_ is now known as davmor2 [10:02] dandrader, bug #1236773 [10:02] bug 1236773 in Unity 8 "[mir] Need an input trap in shell behind the OSK" [Critical,In progress] https://launchpad.net/bugs/1236773 [10:02] mhr3, pass -d codename [10:02] mhr3, i.e. -d mako or something [10:03] thx, got help in -touch already [10:13] mhr3, how do I test scope-isactive easiest? [10:14] mhr3, start with no network, connect? [10:14] Saviq, no, it doesn't have the networking part yet [10:14] Saviq, installing a click should trigger it [10:14] Saviq, or adding media over mtp... theoretically [10:14] mhr3, k, checking [10:15] mhr3, the qml test got fixed somehow, btw? ;) [10:15] Saviq, oh??? [10:15] Saviq, i thought you said you *really* really need to fix that [10:15] mhr3, == make it pass ;) [10:15] mhr3, the test was correct - passing locally [10:15] mhr3, just not on jenkins [10:15] tsdgeos: ok. is it currently under control without the patch or are there infinite loops happening elsewhere? I'm just wondering about the (proposing) scheduling of the fix, with the mir switch etc upon us [10:15] hmm, magic fixes.. why not === dandrader is now known as dandrader|afk [10:15] if we have magic breaks, this just balances it [10:16] mhr3, we'll see in a few - have restarted the qmluitests [10:16] I can anyhow fire up a PPA with that it can be tested from [10:16] Mirv_, we had to revert something in unity8 that caused a lock up with this [10:16] Mirv_, that's not to say we won't get them with what we already have in unity8... [10:16] Mirv_, it's sheer luck we didn't hit it yet [10:17] Mirv_, but yeah, a PPA to test would be nice [10:17] +1 for the ppa [10:19] adding DEP-5 headers etc and uploading then [10:25] MacSlow, re: Snap decisions. we don't have a button for 'show password' while connecting to a wireless network ? [10:25] atleast its not in the implementation. I was wondering what the design says [10:27] dandrader|afk, got everything you need? [10:27] om26er, it's not there in the spec [10:28] Saviq, uh! regression for some (not me) [10:28] my wif is open :) [10:28] om26er, https://docs.google.com/a/canonical.com/document/d/1puQ9Z0yKqzsQ1VQ1OOBkxgp78iWGnAhAkFXWJFTWIrE/edit#heading=h.30kqxmjkvr4b [10:28] *wifi [10:28] om26er, scary! [10:28] yeah I missed it while testing, otherwise it was neat [10:29] Saviq, right, will report a bug for that. I think for people with complex passwords that can be a pain [10:29] om26er, I completely agree [10:29] om26er, I remember in Windoze when you had to type a 26-character DSL router password... [10:29] blind... [10:30] TWICE... [10:30] Saviq, btw re: Mir slowness. Its seem kdub is getting closer. The performance is satisfactory with the command he gave :) [10:30] om26er, yup, saw that [10:30] ouch! [10:31] om26er, that's what the design is like atm... no option for that [10:31] Saviq, do you know where I can find the code to fix this bug report? https://bugs.launchpad.net/unity8/+bug/1224552 [10:31] Ubuntu bug 1224552 in Unity 8 "[Dash] Category expansion transition has varaible speeds" [High,Triaged] === dandrader|afk is now known as dandrader [10:32] Saviq, would you know what dimensions of a preview image do we need for apps? [10:32] tsdgeos, can you point nic-doffay somewhere ↑? [10:33] mhr3, 22x13 GU by default [10:33] Saviq, I think so [10:33] Saviq, and 1gu on mako is...? [10:33] mhr3, 18 I think [10:33] thx [10:33] Saviq: nic-doffay: give me a sec [10:33] mhr3, yes [10:35] nic-doffay: filterAnimation in FilterGrid.qml [10:36] nic-doffay, you need SmoothedAnimation with velocity set instead of Duration [10:36] tsdgeos, ta [10:36] nic-doffay, http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-smoothedanimation.html#velocity-prop [10:37] Saviq, cool reading up on that SmoothedAnimation. [10:39] * om26er thinks there is room for improvement re: scrolling in the Dash. [10:51] om26er, sure there is [10:51] om26er, UShape breaks asynchronicity [10:51] lol someone created in Qt a bytearray in my honour :D [10:51] ./tests/auto/corelib/io/qurl/tst_qurl.cpp:2251: QByteArray tsdgeos("http://google.com/c?c=Translation+%C2%BB+trunk|"); [10:52] tsdgeos, !! congratz! [10:52] Saviq, so the UIToolkit needs to fix UbuntuShape ? [10:52] om26er, it's coming [10:52] om26er, there's a rewrite of UShape in the works [10:52] om26er, and we'll revisit when it's there [10:53] Saviq, good to know. Thanks! [10:53] mhr3, icon theme ubuntu-mono-dark? in -isactive? [10:53] really? did i do it again? [10:53] OOPS [10:54] can we have staging in bzr plz? pretty plz? kthxbai === MacSlow is now known as MacSlow|lunch [10:54] Saviq, pushed [10:55] mhr3, tx [10:55] maybe i should make a real fix that will set it to mono-dark on desktop :) [10:56] mhr3, damn, we broke autopilot tests there ;) [10:56] mhr3, they depended on isCurrent ;) [10:56] nooooooooooooooooooooooooooo [10:56] which doesn't happen now until you unlock [10:56] Saviq, one last question :) If you open an app from inside another app (gallery -> pull toolbar -> tap 'camera') there is no indication that the app is being opened unless it fully appears. I was told its something to be fixed in unity8 [10:57] its more of an issue when you open a video/song from dash [10:57] om26er, yes, and unity-mir [10:58] mhr3, ugh, click scope refreshes More suggestions onIsActiveChanged... [10:58] Saviq, I'm getting that when runnning unity8 manually -> "what(): display factory cannot create fb display" [10:58] Saviq, what am I missing? [10:58] dandrader, stale app running? [10:59] Saviq, no [10:59] dandrader, if you didn't close the apps before killing unity8 [10:59] Saviq, well... not a shell problem [10:59] dandrader, try power button then [10:59] mhr3, of course it's not, but... [10:59] but but... ; [10:59] ;( [11:00] doesn't help... [11:02] dandrader, hrmpf - you running as phablet? [11:02] dandrader, ssh or sudo -u phablet -i ? [11:02] * dandrader tries "adb reboot" [11:02] Saviq, click scope is being super stupid, emits invalidation inside search... looks easy to fix [11:02] Saviq, I logged with "ssh phablet@..." [11:03] dandrader, that's a mako? [11:03] Saviq, maguro [11:03] dandrader, try in #ubuntu-mir, I'm empty [11:04] mhr3, ok, /me looks into fixing ap tests, then - the feature is working [11:04] Saviq, thanks [11:05] dandrader, I assume it is working from upstart? [11:05] dandrader, you sure unity8 is stopped? [11:05] Saviq, yes [11:05] dandrader, yeah, I'm empty... [11:05] Saviq, but at least rebooting solve it (as in I can see a unity8 again) [11:06] Saviq, btw, I do have to "phablet@ubuntu-phablet:~$ export QT_QPA_PLATFORM=ubuntumirserver" right? [11:07] dandrader, no [11:07] ubuntumirclient is fine - unity8 "converts" that internally [11:07] dandrader, but should work nevertheless [11:07] Saviq, hmm.... gonna try with QT_QPA_PLATFORM=ubuntumirclient then [11:07] nah, same problem [11:11] dandrader, that ↑ is exported for you already [11:12] mhr3, at least it's easy to see that isCurrent is working ;) [11:12] Who is the QE with unity8 team ? [11:12] om26er, veebers [11:12] Saviq, that reminds me, http://paste.ubuntu.com/6208875/ [11:13] mhr3, pushed fix for autopilot [11:13] Saviq, wanna add that to the branch ^? [11:13] mhr3, doing [11:13] Saviq, ack, we need a test: bug 1236796 -- I might work on it in the weekends if its not written by someone else. [11:13] bug 1236796 in unity8 (Ubuntu) "[Autopilot test needed] Make sure closing an app from dash keeps dash in focus" [Medium,Triaged] https://launchpad.net/bugs/1236796 [11:14] mhr3, lowercase B [11:15] hmm, i could swear i tested it on the device [11:15] then again, oneliner, i guess i just rewrote it [11:15] mhr3, pushed [11:15] mhr3, let's see what jenkins says [11:16] mhr3, interrupting the current run to incorporate the latest two commit [11:16] s [11:16] I love jenkins... "Stop this build forever" [11:16] FOREVER [11:16] (alone) [11:17] makes you feel like you have superpowers, doesn't it? :P [11:17] ;) [11:17] Saviq, with do qt fix do we get weather and scopes renderer back? [11:18] s/do/the/ [11:19] mhr3, yes! [11:19] and we really, really need video renderers fixed, they look horrible [11:19] i'm going to bother Cimi with that :) [11:20] mhr3, yup, do that ;) [11:21] anyone here uses a maguro for development? [11:21] i do [11:21] dandrader, I just flashed, will try [11:22] Saviq, seems I have a fix for you bug. gonna push the proposals [11:22] it was simple indeed [11:25] dandrader, yup, cool [11:26] dandrader, unity8 by hand worked fine here [11:27] dandrader, just stop unity8; unity8 [11:27] dandrader, making sure that display is lit [11:27] bugger [11:27] btw, is there a powerd command or something to ensure the display doesn't turn off? === alan_g is now known as alan_g|lunch [11:32] dandrader, yeah, sudo stop powerd ;) [11:32] dandrader, there is powerd-cli that should work, in theory [11:32] guys use it in automation === hikiko is now known as hikiko|lunch [11:38] Saviq, https://code.launchpad.net/~dandrader/unity8/fix_lp1236773/+merge/189821 and https://code.launchpad.net/~dandrader/unity-mir/fix_lp1236773/+merge/189822 [11:39] dandrader, yup, looking good [11:40] dandrader, we *could* do OSKController for SurfaceFlinger, but it probably doesn't make sense now ;) [11:40] Do we plan to merge this anytime soon? https://code.launchpad.net/~saviq/unity8/fix-frequent-app-ids/+merge/188532 [11:41] if not how about removing Facebook and gmail icons from the Home scope ? [11:41] om26er, after https://code.launchpad.net/~ted/url-dispatcher/app-id-url/+merge/187913 most probably [11:44] Saviq, not so fast. spotted an issue [11:45] dandrader, ;P [11:47] mhr3, isactive lookin' good, passed ap [11:47] mhr3, passed qmluitests [11:47] Saviq, yey! [11:47] -touch is to follow [11:47] but I'm happroving already [11:50] Saviq, yeah, running unity8 from console works after stopping powerd. \o/ [11:55] dandrader, cool === MacSlow|lunch is now known as MacSlow [12:01] Saviq, ok, now the fixes seem to be really ready to go [12:01] dandrader, coolio === dandrader is now known as dandrader|lunch [12:02] dandrader|lunch, hum? you disable the InputArea? isn't that how the OSK *ever* got input/ === Mirv_ is now known as Mirv === alan_g|lunch is now known as alan_g === _salem is now known as salem_ === hikiko|lunch is now known as hikiko === iahmad|afk is now known as iahmad === jhodapp|afk is now known as jhodapp [12:48] nic-doffay, do we have a bgreport or a design guideline for that? [12:48] nic-doffay, I tested it, works but I don't like the interaction [12:48] I think we can do better [12:52] mhr3, it's our lucky day - merged on first try ;D [12:54] Saviq, feels like sweeping an issue under the rug and waiting for another branch to fail on it :/ [12:54] mhr3, no, really - the test was correct [12:54] mhr3, I didn't see what could be happening there [12:55] Saviq, then we're sweeping under the rug how much the ci infra is broken :) [12:55] mhr3, btw, YOU SHOULD SEE MY RUG! [12:55] lol [12:55] i can imagine those dead bodies there... you probably have a big rug, right? === dandrader|lunch is now known as dandrader [13:04] Saviq, that's just the initialization of that value [13:05] Saviq, who is responsible of the dash design? [13:05] Saviq, was reviewing mic branch but I don't like the UX [13:05] Saviq, it's overwritten later by a manual connection at the end of OSKController.qml [13:05] nic [13:05] dandrader, right! [13:06] Cimi, visual? Rosie [13:06] Saviq, UX [13:06] Cimi, Patricia [13:06] Saviq, ok I'll come up with my proposal of UX and wait her feedback later [13:07] mhr3, the carousel in the music scope does not make sense. That should only be used for Albums not individual songs. [13:08] om26er, design wanted that, we raised the same question, they had their reasons [13:08] right now I am starting songs blindly :/ [13:08] might be better to just open their preview first [13:09] om26er, you'd want https://code.launchpad.net/~unity-team/unity8/switching-previews/+merge/189556 then :) [13:09] Saviq, is that one going in btw ^? [13:09] mhr3, if mzanetti fixes [13:09] Saviq, that branch opens the preview instead of directly opening the song ? [13:09] om26er, yes [13:09] mhr3, I'd like it to, but there's an issue [13:09] mzanetti, it's not just pretty, but useful as well, pls fix ;) [13:10] mhr3, heh, I meant to ask you :) but Saviq already answered. Thanks [13:19] ping kgunn [13:22] dandrader: pong [13:22] Cimi: can you test the packages at https://bugs.launchpad.net/bugs/1236765 with https://code.launchpad.net/~aacid/unity8/unrevertr388/+merge/189839 [13:22] kgunn, is that a blocker for Oct 10th? https://bugs.launchpad.net/unity-mir/+bug/1234600 [13:22] Ubuntu bug 1234600 in unity-mir "Contacts Cancel/Save buttons don't work with OSK showing when running Mir" [High,In progress] [13:22] Ubuntu bug 1236765 in qtdeclarative-opensource-src (Ubuntu) "[PATCH] Fix infinite loop in QQmlIncubator::forceCompletion" [Undecided,In progress] [13:22] Cimi: i'll do the same and thus we have two people testing [13:23] tsdgeos, yup [13:23] tsdgeos, packages are already built or I build on phone? [13:24] Cimi: read the bug [13:24] dandrader: its not a blocker - but will be on the high visbility list when mir gets turned on... [13:24] where it says "packages are at foobar" [13:24] tsdgeos, ok [13:25] tsdgeos, https://launchpad.net/~canonical-qt5-edgers/+archive/qt5-daily/+build/5084928 [13:25] tsdgeos, it's easier to add the ppa [13:25] nevermind [13:25] yes, it's easier to add the ppa :-) [13:32] mhr3: what was wrong with the datetime indicators? [13:33] dednick, not showing up for me in --channel=saucy-proposed [13:33] mhr3: um, is that the same as --pending? [13:33] dednick, i was screamed at to not use cdimage [13:34] mhr3: er. ok [13:34] but theoretically, yes :) [13:35] mhr3: can you see if indicator-datetime is installed? [13:36] dednick, running upgrade right, so apt-cache doesn't talk to me, a file i can check to know? [13:37] mhr3: /usr/share/unity/indicators/com.canonical.indicator.datetime [13:37] yep, it's there [13:39] Cimi, I just tried out my branch, it def dismisses here when you press on anywhere aside from the keyboard/text area. [13:40] kgunn, "feeling froggy" [13:40] never heard that [13:42] mterry: you're not american :) [13:48] nic-doffay, :-\ [13:52] is there a known bug that once unity8 is in "close running software" mode (e.g long press in the app lens leading the red cross on the images to close what is running) it's stucked in there until you are done closing running softwares? [13:55] seb128: you jut long press again and the [x] should go away, doesn't that work? [13:56] dednick, ping [13:56] Saviq: pong [13:57] dednick, re: indicator redrag - using a single EdgeDragArea didn't work well? [13:57] dednick, it's probably too intertwined with the Revealer? [13:57] tsdgeos, oh, you need to long press again on one of the program ... I close one, I would expect the mode to end or that starting a new software would stop it [13:58] tsdgeos, I didn't expect I should try to click on one of the icons of the softwares I don't want to close and keep pressing [13:58] tsdgeos, e.g I was unsure that wouldn't close the thing I was clicking on [13:58] seb128: the UX wasn't really well defined when we implemented the feature afair [13:58] so yeah, it can for sure be better [13:58] Saviq: the Areas fo the initial drags are anchored to the bottom of the page. when you drag down, they go to the bottom of the page. [13:59] seb128: needs people that know about that to think about it [13:59] seb128, if you go away from the apps scope it will reset, too [13:59] seb128, at least in trunk - we just merged a fix [13:59] dednick, orly? ok then [13:59] Saviq: hang on, i might take another look [13:59] Saviq: possibly might be able to reuse the drag down. [13:59] Saviq, oh ok, is it worth opening a bug "starting a new software should exit that mode as well"? [14:00] seb128, no, that will happen [14:00] seb128, with trunk [14:00] ok, great [14:00] you guys are one step ahead as usual ;-) [14:00] Saviq, tsdgeos: thanks [14:01] Saviq, (if I get what you said, "exit the app scope", includes going to an app, not only changing lens in the dash) [14:01] seb128, yup [14:01] seb128, going to indicators, too [14:01] seb128, or locking the phone [14:01] excellent [14:01] * seb128 is happy [14:01] no bug from me for this one then ;-) [14:10] Saviq: hm. doesnt work. EdgeDragArea doesnt have the autocomplete/reject. That's why we need the DragHandles. [14:10] also for the hinting animation. [14:11] dednick, right, another question then - did you talk about the "commit" phase with design guys? [14:11] dednick, it feels too easy to change maybe? [14:12] Cimi, you'll do bug #1231937 ? [14:12] bug 1231937 in Unity 8 "Video & music renderers are not to spec" [High,Triaged] https://launchpad.net/bugs/1231937 [14:12] Saviq: not specifically. They said just drag your finger left to right and it should change... i had originally done it so that you had to drag down first. [14:13] as in left to right across the indicator bar [14:13] dednick, ah [14:13] dednick, ok, let's land this [14:13] Cimi, bug #1226288 is related [14:13] bug 1226288 in Unity 8 "Carousel should only be used when there's enough items" [Medium,Triaged] https://launchpad.net/bugs/1226288 [14:15] tsdgeos, can you review https://code.launchpad.net/~stolowski/unity8/category-reordering/+merge/189263 - you've been involved with that (can be *after* you do delegates ;)) [14:16] Saviq: sure, will do [14:16] Saviq: let me just double check with oren [14:16] dednick, already happroved ;P [14:20] Saviq: not anymore ;) [14:20] dednick, you! [14:21] Saviq: apparently the design has not gone through... JohnLea told me to do it, but now oren says no... [14:21] dednick, eh... [14:21] yup [14:21] dednick, k [14:25] mhr3: datetime is present for me. Can you check if the indicator-datetime process is running? [14:26] dednick, i updated to daily, it's fine now [14:26] mhr3: ok :/ [14:26] Saviq, yep both [14:26] Cimi, cool [14:31] greyback: are you actually out? https://plus.google.com/hangouts/_/6f7b22053ff6d243c9ba67afbf759cbf9d0c5e6a [14:32] kdub: ^ [14:33] dandrader, were you ever able to look at / reproduce https://bugs.launchpad.net/unity8/+bug/1228336 ? [14:33] Ubuntu bug 1228336 in Unity 8 "[DDA] Crash when revealing indicators" [Critical,Triaged] [14:33] sil2100, ping? [14:33] Saviq, no, is it showing up its ugly face? [14:34] dandrader, no [14:34] dandrader, does the trace make any sense, though? [14:34] dandrader, or shall we just mark incomplete? [14:34] Free Karma: https://code.launchpad.net/~nick-dedekind/unity8/indicators-client.item-title/+merge/189868 [14:35] Saviq, it does [14:35] make sense === charles_ is now known as charles [14:37] Saviq, although it's a mystery how is it possible for it to get into such state [14:37] dandrader, mhm, mark Incomplete then? [14:37] mhr3: pong [14:37] mhr3: what's up? [14:38] dednick, MINE [14:38] Saviq: :) [14:38] sil2100, hey, we're seeing a crasher in mediascanner tests, identified it to be grilo bug, mind distropatching it? [14:38] one of my easier ones [14:38] Saviq, yeah, incomplete as there are no steps to reproduce it [14:38] in a reasonably reliable way [14:38] sil2100, it's in lp:~mhr3/+junk/grilo-fix-1236637 [14:39] mhr3: sure thing! Where's the bug/patch/branch? [14:39] ^^ [14:39] Doing [14:39] nic-doffay, not everywhere [14:39] dednick, got a moment to test out a branch for me? [14:39] nic-doffay, searching from the home screen [14:39] I still can click elements on the application section [14:39] nic-doffay: funny man. [14:39] nic-doffay: but yeah, sure [14:40] Cimi, I just tried it again on the home screen. [14:40] It dismissed when I select anywhere? [14:41] nic-doffay, I think your improves the situation [14:41] nic-doffay, with the keyboard [14:41] nic-doffay, but it's still unclear to the user when it will dismiss or not [14:41] dednick, i'm a comedian, what can I say? https://code.launchpad.net/~nicolas-doffay/unity8/expansion-transition-fix/+merge/189872 [14:41] Cimi, that's not part of the bug though? [14:42] nic-doffay, which bug? [14:42] When it lands design can go over it. [14:43] Cimi, I thought I'd linked to the branch, apparently not. [14:43] Cimi, here's the bug: https://bugs.launchpad.net/unity8/+bug/1213034 [14:43] Ubuntu bug 1213034 in Unity 8 "Can't dismiss keyboard by tapping outside of search entry" [High,In progress] [14:43] I've lined it to the MP now too. [14:43] *linked [14:46] nic-doffay, so, it's not fully working for me [14:46] nic-doffay, I invoke the search entry in the home screen [14:46] nic-doffay, keyboard and recent searches menu appear [14:47] nic-doffay, I touch on "applications" category header [14:47] mhr3: do you have a bug for the crash btw.? [14:47] nic-doffay, it expands the applications category but does not dismiss the search box, even if it's empty [14:47] sil2100, don't you wonder what is that number in the branch name? :P [14:48] mhr3: pfff, some random numbers [14:48] ;) [14:48] sil2100, or do you want upstream bug? [14:48] i'll open that [14:49] kgunn, ping [14:50] nic-doffay: reviewed. [14:50] dednick, cool ta [14:51] Cimi, that's weird man it def works for me... [14:52] nic-doffay: both of you on surfaceflinger or mir? [14:52] tsdgeos, no mir here [14:54] dednick, can you try https://code.launchpad.net/~nicolas-doffay/unity8/dismiss-keyboard/+merge/188599 ? [14:54] kdub: pong [14:55] dednick, switched to equivalent grid units dude. [14:55] Totally slipped my mind. [14:58] nic-doffay: i'm not sure about this. we should probably have a max time. it can take quite a while to open if there are lots of entries... have you checked with design? === dandrader is now known as dandrader|afk [15:02] dednick, not yet Oren wasn't online. [15:02] Let me check if he's around. [15:02] nic-doffay: he's not at his desk [15:02] I assumed there shouldn't be since it wasn't mentioned in his bug report. [15:02] But lets hold off until I double check. [15:03] nic-doffay: my file scope has a hundred entries or so, and it takes quite some time to close... [15:04] Cimi: testing now [15:05] nic-doffay: lp:~nicolas-doffay/unity8/dismiss-keyboard has conflicts. [15:09] dednick, https://wiki.ubuntu.com/Touch/Testing [15:09] mhr3: thanks [15:13] sil2100, once you have the patch, can you link it to lp:1236637 pls? [15:13] Saviq: need this one https://code.launchpad.net/~dandrader/unity8/fix_lp1236773/+merge/189821 [15:13] as well as your mp that's waiting to merge right? [15:13] kgunn, yeah [15:13] mhr3: suar [15:13] anymore other than those 2 ? [15:14] kgunn, no [15:14] cool [15:14] sil2100, txh [15:14] kgunn, refocus-shell should land within minutes (and if it fails on mako I'll just push it through) [15:14] ah, SUCCESS [15:14] scared, huh! === dandrader|afk is now known as dandrader [15:16] kgunn, the other one just started [15:16] aaand I'm off for now... did not get up from this chair yet today [15:16] Saviq: please don't injure yourself [15:17] kgunn, on the other hand that would ensure he doesn't move from there :) [15:18] dednick: hi! Is ubuntu-settings-components ready for release? And daily-release? [15:18] :) [15:18] sil2100: yup [15:18] dednick: awesome! [15:19] mhr3, ;D [15:20] Saviq, j/k, go enjoy the sun, while there's still some left [15:24] mhr3, I *did* merge scope-isactive today did I not! [15:24] mhr3, you can try and fix switching-previews in your spare time ;) [15:24] Saviq, yes, you restored my faith in humanity, thank you [15:31] mhr3: package uploaded! Sponsored by kenvandine, it's building now ;) [15:32] sil2100, awesome, thanks [15:35] Cimi, here's the spec https://docs.google.com/a/canonical.com/presentation/d/1X3_G6reCpcFJxzxn1Xtv9wK5n3-wOW-6z2l0qSZ8-sg/edit#slide=id.g103714763_47 [16:07] dandrader: since greyback is out...do you know if unity-mir launching mir is running mir with user permissions or as root ? [16:07] racarr: ^ ? [16:08] kgunn, afaik it's running as phablet user [16:09] kgunn, why? [16:09] dandrader: :) [16:09] dandrader: we're talking about moving the mir_socket to a diff dir associated with a running user session [16:11] kgunn, /run/user/32011/mir_socket or something [16:12] right talking about how to do that w/o breaking xmir as well [16:12] right [16:14] fginther, http://10.97.0.26:8080/job/generic-mediumtests-runner-maguro/2161/console stuck? [16:14] Saviq, look like [16:15] Saviq, I took the node offline [16:17] Saviq: yeah...i just reapproved dandrader's mp to get some jenkins love.... [16:25] EOD === jhodapp is now known as jhodapp|vet [16:41] Saviq, albert is OED [16:41] Saviq, I tested the qt fix and is working [16:52] charles: ping [16:55] kgunn: it would appear that it's the location service backend that is misbehaving. [17:04] mterry, hey :) === alan_g is now known as alan_g|EOD [17:08] charles: location indicator doesnt seem to be working correctly. action activations arent doing anything on backend. Seem not to be getting updates from platform api. [17:08] Cimi, comment on bug #1236765 [17:08] bug 1236765 in qtdeclarative-opensource-src (Ubuntu) "[PATCH] Fix infinite loop in QQmlIncubator::forceCompletion" [Undecided,In progress] https://launchpad.net/bugs/1236765 [17:08] Cimi, ah you did [17:08] Cimi, ok then, Timo will put it in our qt tomorrow and we'll unrevert then [17:10] om26er, what's up? [17:10] mterry, bug 1236960 [17:10] bug 1236960 in Unity 8 "[Mir] greeter matrices are not updating, requires unity8 restart" [Medium,New] https://launchpad.net/bugs/1236960 [17:11] om26er, marked as dup [17:11] om26er, of bug 1234904 [17:11] bug 1234904 in unity-mir (Ubuntu) "[mir-only] infometric values are not updated" [Critical,Confirmed] https://launchpad.net/bugs/1234904 [17:11] haah [17:11] om26er, a fix on the way, just needs to be reviewed [17:11] mterry, cool. thanks [17:22] Saviq: I'll need to run all AP:s to convince the rest of the integration team, but yes I'll look at that in the morning and find a sponsorer then [17:22] g'night. === jhodapp|vet is now known as jhodapp [17:53] Saviq, if the greeter is visible, should the focused app still be focused? [17:56] * mterry thinks not for the same reason that we unfocus on display-off [18:21] mterry, no [18:21] Saviq, cool, fixing that optimization makes it easier to detect when a snap decision launches an app [18:30] Saviq: fginther ...i notice we're still waiting on https://code.launchpad.net/~dandrader/unity8/fix_lp1236773/+merge/189821 [18:30] however, we're hoping to get that into qa test... [18:30] are we confident ci will run ok ? [18:31] or should we consider manually landing onto trunk (ci run be damned) [18:31] kgunn, let me push that through [18:31] kgunn, it's building right now, let me see how far it is [18:32] kgunn, right, it's building atm [18:32] what's the normal turn time....like 30 minutes? [18:32] kgunn, more, currently, I'm afraid [18:33] kgunn, 48 mins on a good run [18:33] kgunn, 1:48 on a bad one (when queued up) [18:34] kgunn, Saviq, the mako/maguro tests are queued waiting for the current build to finish. [18:34] kgunn, Saviq, it should start within 10 and another 20 to run [18:34] fginther, kgunn yup [18:35] kgunn, it's green up 'til now - if it fails on touch, I'll push through [18:35] ack, thanks Saviq [18:35] kgunn, Saviq, the desktop tests passed [18:37] fginther, yup [18:38] kgunn, for future reference, you got access to rebuild jobs on s-jenkins? you want them? [18:39] ooo...i like that [18:39] how do i get that [18:41] kgunn, register at http://10.97.0.26:8080/signup [18:42] kgunn, and fginther will set you up for access to triggering rebuilds [18:43] kgunn, looks like you won't get the chance this time - things look green for that MP [18:47] Saviq: never been so happy to be disappointed :) [19:04] kgunn, Saviq is this stuck? http://10.97.0.26:8080/job/unity-phablet-qmluitests-saucy/2178/console [19:05] kgunn, Saviq, it's a ci run for https://code.launchpad.net/~unity-team/unity8/albumart-provider/+merge/189923 [19:05] but it looks to be blocking the autolanding you're interested in [19:05] well that explains why the merge wasn't showing up [19:06] kgunn, should I kill and retrigger the -ci jo? [19:06] fginther: yes, please makes sense [19:07] fginther, yeah, looks like it [19:10] Saviq, kgunn, it failed a qmluitest, do you want it pushed as is [19:10] ? [19:10] fginther, already pushed [19:10] Saviq, :-) [19:11] fginther, we should probably skip that test - it's unstable under Qt 5.0 [19:11] kgunn, it's merged [19:12] Saviq: ack, thanks [19:22] Saviq, ping, you want to bump ABI? [19:25] tedg, or even API - I managed to build unity-mir packages against new u-a-l and install them, just to get missing symbols [19:26] Oh, this is because we don't have a symbols file... [19:26] No one tell didrocks [19:26] tedg, ;) [19:26] Saviq, So I think if I just add a symbols file we should be good there, because then dpkg will make the version you depend on based on the symbols you use. [19:28] Saviq: still awake? === thomi_ is now known as thomi [19:32] thomi, yup, here [19:32] Saviq: do you know what would be required to fix this? https://bugs.launchpad.net/unity8/+bug/1237022 [19:32] Ubuntu bug 1237022 in Unity 8 "Silence unknown option messages" [Undecided,New] [19:32] Saviq: is it as easy as not passing those merssages to Qt itself? [19:33] if so, I'll propose a branch myself that fixes it [19:33] thomi, sure, but how would you know? [19:33] how would you know what? [19:33] thomi, which options are "known" and which are not? [19:33] well, I assume that unity8 knows "I handle options X, Y, and Z", everything else pass to Qt itself [19:34] thomi, and then there's Mir [19:34] thomi, and the QPA [19:34] thomi, but anyway, it's not unity8 that spits out those unknown options warnings [19:34] thomi, it's Qt, AFAICT [19:35] Saviq: I thought it was Qt [19:35] yeah [19:35] thomi, yeah, so you'd have to know all options Qt and Mir and QPA handle to filter out those unknown [19:35] so if I make sure -testability and -fullscreen have been removed before passing argv to Qt, that should be OK? [19:35] thomi, sure, we can pop [19:35] Saviq: ideally yes, but in reality it's only ever these two options that get printed [19:35] thomi, but that won't mean we'll actually silence it ;) [19:36] thomi, granted, false negatives will be silenced [19:36] Saviq: it'll still print if someone uses some mir / QPA option, but that seems rare [19:36] thomi, not a huge gain if you ask me ;) [19:36] thomi, we generally need to silence Qt's and QML warnings [19:37] hmmm [19:37] thomi, only problem is... that can only be done build-time with current Qt's logging facilities [19:37] ugh, really? bum [19:37] Saviq: do you know whwre I should start looking inside Qt for that stuff? [19:37] roughly which class/file it's in? [19:37] thomi, yeah, 5.3 is supposed to have a more sophisticated logger [19:37] thomi, http://qt-project.org/doc/qt-5.0/qtcore/qdebug.html [19:38] Saviq: ahh, it uses QDebug - where is that called though? [19:38] tell you what, I'll stop being lazy and look for it myself :) [19:38] thomi, good boy! [19:38] :) [19:38] need more coffee first though [19:40] thomi, or well - http://qt-project.org/doc/qt-5.0/qtcore/qtglobal.html#qInstallMessageHandler [19:41] thomi, we could make and install one that would be quiet by default [19:41] and only enable debug on some env var and/or argument [19:41] Saviq: I was hoping to find where in QApplication (or wherever) those log messages are generated, and see if I couldn't override that code to be silent [19:41] thomi, that does *not* sound like the correct solution ;) [19:42] why's that? We want the other Qt debug output, just not those messages [19:42] thomi, what's so bad about those two> [19:42] thomi, and anyway - just pop them from argv when you handle them, is all [19:42] Saviq: OK, that makes more sense - I'll do that [19:43] Saviq: they really clutter up the AP test logs [19:43] making the test logs much less useful [19:44] thomi, I usually use QStringList::removeAll(const QString&) [19:44] thomi, it spits out the number of items removed [19:45] thomi, so if(argv.removeAll('-fullscreen') > 0) { foo; } [19:46] Saviq: gotchya [19:46] Saviq, https://code.launchpad.net/~ted/upstart-app-launch/symbols-file/+merge/189951 [19:48] tedg, should you do REPLACEME there? [19:48] Saviq, Since I bumped the version I don't need the merge to do it. [19:48] tedg, right, it will always be >= 0.2 [19:50] Saviq: lool> kgunn: so I wanted to run the webbrowser-app-autopilot, these stop unity8 and start it again immediately, but it doesn't seem to come up [19:50] kgunn: like can't unblank display thing [19:51] did we test the webbrowser-app [19:51] ap suite? [19:51] kgunn: presumably the reason it does that is because we don't run in testability mode by default. [19:52] ...and it needs the shell to be introspectable for some reason [19:52] thomi, no [19:52] thomi, it was lool running with -n [19:52] phablet-test-run I mean [19:52] ahhh [19:52] well, you can't do that with the mir stack any more I guess [19:53] thomi, well, you *should* be able to [19:53] well, sure, if you start the shell yourself [19:53] thomi, but the powerd issue + the leftover socket makes it difficult enough [19:53] Saviq: exactly [19:53] and it shouldn't really be the concern of the test case [19:54] unless it's testing the shell itself [19:54] thomi, of course [20:06] Saviq: i wanna top approve this one...as i think it will solve _some_ crashes... https://code.launchpad.net/~afrantzis/unity-mir/fix-1236898/+merge/189894 [20:06] Saviq: shold i wait for gerry...seems sensible and safe [20:06] tho [20:07] kgunn, no need to wait, just top-approve if you think is good - if in doubt - ask someone else to look (like racarr) [20:08] Saviq: he already approved :) [20:08] kgunn, right, so yeah - go for it [20:19] Saviq: is the recommended way to build unity8 still './build -s && ./build ' ? [20:20] I'll assume that 'CODING' is accurate... [20:20] thomi, yeah [20:20] thomi, bzr bd does fine, too [20:20] thomi, no scary PPAs added this time ;) [20:21] Saviq: sweet - it still asks me for my password though, which is kind of scary :) [20:28] Saviq: ping, hey is it possible to launch unity8 without the greeter? [20:30] Saviq: hey, so when I run unity8 locally with -testability I don't see that output... is there some kind of race condition, do you think? [20:32] veebers, nope [20:33] veebers, there is something mterry added - it listens to "ShowHome" on com.canonical.Unity.Greeter.Broadcast on the system bus [20:34] Saviq: ah ok, so that's a potential solution if people are having 'greeter swipe' issues? [20:34] elopio: ^^ I don't know if that helps you at this point? [20:35] veebers, elopio, we *could* add a env var we'd look at, but hopefully we won't need to? [20:35] mterry, ideas about unlocking unity8 for testing ↑↑ ? [20:35] veebers, Saviq: at this point, I think we have to live with the fact that our app tests interact a lot with unity. [20:36] but for the future, I would like to split the suites. If we are not testing any integration with unity on a test suite, I would like unity to be as invisible as possible. [20:36] elopio, actually, if you start them with upstart after https://code.launchpad.net/~mterry/unity8/hide-greeter-on-app-launch/+merge/189948 is merged [20:36] elopio, it should just unlock and focus the app [20:37] mterry, correct ↑ ? [20:37] oh, that's handy. [20:39] Saviq: so that "unknown option" string is not being printed by Qt it seems [20:39] Saviq: I wonder if it's unity-mir maybe? [20:40] thomi, no, not there for sure [20:40] hmmm [20:41] I grepped for "unknown option" through the unity source code, and couldnt' see anything that was related to what we're doing [20:41] err, I mean "Qt source code" [20:42] thomi, the interesting thing... if I go ./builddir/unity8 -blahblah [20:43] thomi, I don't get it printed out? [20:43] Saviq: yeah, me neither [20:43] Saviq: which made me think maybe it only happened when you ran on top of mir? [20:43] thomi, https://bugs.launchpad.net/mir/+bug/1226227 then? [20:43] Ubuntu bug 1226227 in unity8 (Ubuntu) "libmirserver parses arguments and fails if it's not something it understands" [Medium,Fix released] [20:44] Saviq: no, that's something separate [20:44] Saviq: what's more, I'm *sure* this happened with unity8 on SF as well [20:44] thomi, yeah, so QPA-specific maybe? [20:45] Saviq: which project is that? [20:45] Saviq, yeah, right. That should work [20:46] thomi, that'd be qtubuntu [20:46] I looked there already :-/ [20:47] thomi, it's "invalid option" - grepped for that? [20:47] yeah [20:47] thomi, so I'm getting it on the device, yes [20:47] thomi, sure it's not QApplication? it shows up just after "WARNING: QApplication was not created in the main() thread." [20:48] Saviq: It might be constructing that string in a way that makes it un-greppable [20:49] thomi, yeah, looks like it [20:50] thomi, so - both on SF *and* Mir indeed [20:51] Saviq: hmmmm [20:51] thomi, but not on desktop [20:51] thomi, so arguably not QApplication [20:51] anyway, I'm afk [20:52] Saviq: thanks for your help === salem_ is now known as _salem [22:14] Is the ListViewWithPageHeaderTest.testMaximizeVisibleAreaMoveUpAndShowHeader qmltest causing problems for others? [22:56] mterry, we have one that's flaky in Qt 5.0 and it's failing in jenkins sometimes - not sure that's the one, though