=== duflu_ is now known as duflu [08:37] tsdgeos: the AP is in the PPA. the test case seemed fixed for me too, but it'd look like running full UITK test suite just hangs at some point - in the sense the test app is loaded but it stays there. [08:50] I killed now one process and will see if the current run finishes eventually [08:55] ok [08:56] i'm also a bit concerned about richard's report of the addressboox app hanging [08:57] trying to reproduce [09:20] i can't reproduce richard's issue [09:22] Cimi: can you do https://code.launchpad.net/~aacid/unity8/urldispatcher_hideeverything/+merge/253527 ? [09:22] and someone https://code.launchpad.net/~aacid/unity8/passphrase_kewboard/+merge/253657 ? === davidcalle_ is now known as davidcalle [09:51] pstolowski: commented on https://code.launchpad.net/~stolowski/unity-api/filters-iface/+merge/252890 [09:52] tsdgeos, thanks, looking [10:01] tsdgeos, is this also a problem with roles, do I need to avoid "id" in role names? [10:02] pstolowski: roles are fine since they're not objects per se [10:02] afaik we have have some other id roles [10:02] in there [10:02] yeah [10:03] ./include/unity/shell/scopes/ScopesInterface.h:109: roles[RoleId] = "id"; [10:03] yeah, you're right [10:03] thanks [10:14] tsdgeos, about your second comment, i can added optionsChanged signal, but it will never be triggered since any changes are changes to options rows (model changes) [10:14] tsdgeos, makes sense? [10:14] pstolowski: is the options varaible set correctly on creation alrady? [10:15] tsdgeos, yes, it's going to be set on creation of the filter only [10:15] then it's fine [10:15] tsdgeos, fine to add but never trigger? [10:15] yes [10:16] ok === greyback_ is now known as greyback [10:54] hmm. so slow. [11:03] dednick: i think we should do the same thing i did for all the drags and see fi that helps [11:03] since it seems it may very well be that [11:03] looking at the logs the last movement is very few pixels === JMulholland_ is now known as JMulholland [11:47] greyback: does MirSurfaceManager have any reason to hold m_mirServer? I don't see it being used. [11:52] alan_g: quite right, it is unused [11:52] * greyback expcted compiler to notice that [11:52] no actually, that would be presumptuous of it [11:53] The compiler knows there's a non-trivial dtor === MacSlow is now known as MacSlow|lunch [11:54] I just wanted to double check there isn't a weird lifetime issue. [11:56] none I'm aware of, this it an oversight [12:15] Hello. I am trying to boot a 15.04 next live cd. Whenever i enter the username ubuntu-desktop-next and password, its stuck at login. i can move the cursor around but it wont let me in the gui === MacSlow|lunch is now known as MacSlow === Pici` is now known as Pici [12:54] greyback: not urgent, but would something like this make sense to you? https://code.launchpad.net/~alan-griffiths/qtmir/spike-using-WindowManager/+merge/254235 [12:58] alan_g: looks reasonable at first glance. You've got work to do to replace our custom input dispatcher with the handle_*_event bits, but nothing huge. Direction looks good [12:59] greyback: I don't *have* to change that. (At least not for a stable form.) [13:00] alan_g: true [13:01] alan_g: question, would "handle_*_event" supposed to communicate things like close event or focus/unfocus? A nested compositor would probable need to know such things [13:02] BTW how do you go about building the stack and deploying on a phone? Every time I try I can't help feeling there's a better way... [13:03] alan_g: either building on device, or an sbuild chroot [13:03] if you disable tests building, dev on device isn't so bad [13:04] How do you get space on the device to install the build tools? [13:04] I don't hit that issue usually [13:05] if you're changing mir too, then yeah, that's a problem [13:05] Well, it is up to the WM to set focus etc, so yes [13:06] I can probably work around that. [13:06] Gotta go === alan_g is now known as alan_g|lunch [13:06] enjoy [13:30] *giggle* " [13:30] Please note that Antti's "Disapprove" review is a joke. [13:50] @unity: guys my unity8 is not accepting any touch at all [13:50] http://paste.ubuntu.com/10683845/ [13:52] any clue? [13:53] greyback: thread 22 suspicious? [13:53] tsdgeos, no [13:53] * tsdgeos gets more debug symbols [13:53] tsdgeos: thread 22 in that state means the main loop is blocked [13:54] greyback: that sounds bad [13:54] :D [13:54] yeah [13:54] it is mir asking qt to authorize an application connection [13:54] and is waiting for a blockingqueuedconnection to return [13:54] yet [13:54] on screen [13:55] i have soemthing that used to be a sdk test application [13:55] that is not even running anymore [13:55] renderer is blocked too then [13:57] greyback: anything else more i can do than filing a bug? [13:58] tsdgeos: is anything using lots of cpu? is dbus busy? [13:58] I don't see a good clue in that BT sadly [13:58] cpu is idle [13:58] dbus-monitor says nothing [13:58] ok [13:59] can you get the thread names [13:59] sure [14:00] greyback: http://paste.ubuntu.com/10683891/ [14:00] tsdgeos: any ide what those pool threads are? [14:00] idea [14:00] greyback: gio [14:01] which we don't use but somehow we're stuck with afaik [14:01] hmm ok [14:02] tsdgeos: oh hang on, thread 9 is weird [14:02] that's the render thread [14:03] it's blocked waiting for a new buffer from mir [14:03] mir is still unity8 here, no? [14:03] or is it usc? [14:03] yeah, mir is a lib unity8 uses [14:03] any clue which thread shuould supply that buffer? [14:05] not sure, I suspect thread 23 but am not sure [14:05] asking mir folk... === alan_g|lunch is now known as alan_g [14:05] tsdgeos: please log a bug anyway, including unity8.log [14:06] k [14:06] i'll get some more debug symbols intsalled [14:06] greyback: this is not "hard" to get it seems [14:06] happens if you run the sdk suite [14:07] hmm [14:07] not what I like to hear :( [14:07] seems more easily if you have the silo we have that makes qt dbus locking less [14:08] but there's nothing that would seem to blame qtbus in that bt [14:08] i'll note it in the bug anyway [14:08] just in case [14:11] tsdgeos: what mir version? === dandrader is now known as dandrader|afk [14:29] QQuickAsyncImageProvider merged \o/ === davidcalle_ is now known as davidcalle [14:38] tsdgeos: nice one [14:38] tsdgeos: link? We backported to 5.4? [14:40] greyback: not yet [14:40] greyback: https://codereview.qt-project.org/#/c/108540/ [14:40] only needed 17 patchsets :D [14:41] and https://code.launchpad.net/~aacid/thumbnailer/asyncprovider is our counterpart [14:41] which i think doesn't compile now [14:41] since i had to change stuff upstream to make them happy [14:42] tsdgeos: no worries. just wanna be familiar with the api [14:42] there's not much [14:42] basically you get a job [14:42] and then you signal done() [14:42] it's up to you to spawn threads and stuff [14:42] i think it's not done() but finished() [14:42] :D [14:42] but you know what i mean [14:48] gotcha === dandrader|afk is now known as dandrader === lool- is now known as lool [15:19] lol [15:19] make testDialogs was doing nothing [15:19] :D [15:20] Mirv: \o/ [15:20] Loading tests from: /usr/lib/python3/dist-packages [15:20] Tests running... [15:20] Ran 365 tests in 2579.659s [15:20] OK [15:20] greyback: that silo seems ot really help [15:20] * tsdgeos launches again [15:21] tsdgeos: sweet [15:21] tsdgeos: good that you caught me still, I can still launch a 018+013 full suite then now [15:23] and, \o/ if a working combo of Mir + Qt + Autopilot is found [15:23] he he [15:24] mzanetti: test updated https://code.launchpad.net/~aacid/unity8/passphrase_kewboard/+merge/253657 [15:26] Mirv: you probably need my autopilot change too, do you have that one? [15:27] tsdgeos, ack [15:27] tsdgeos: hmm? you mean the one I submitted to ap1.5 and added to the PPA, or something new? [15:28] Mirv: that one [15:28] ah didn't know you had it in the ppa already [15:28] good then [15:28] tsdgeos: right, the MP is there in the PPA. [15:28] tsdgeos, yeah, it's there only for its "try" version [15:29] dandrader: lol right i broke the try now [15:57] mzanetti: there was no test for this functionality https://code.launchpad.net/~aacid/unity8/urldispatcher_hideeverything/+merge/253527 originally [15:58] but i guess i can write one for my fix [15:58] and for the feature itself [15:58] will take a bit but it's good to have another test [16:22] tsdgeos: can any of the properties of a Result be used as a unique id? [16:22] to look for an item and similar tasks [16:22] pstolowski: ↑↑ [16:22] uri is probably not unique enough [16:24] faenil: result is [16:24] since it's basically the pointer [16:24] case RoleResult: [16:24] return QVariant::fromValue(std::static_pointer_cast(m_results.at(index.row()))); [16:24] not sure the mock implements it that way [16:24] if it doesn't it should :D [16:25] or at least return something that is unique [16:26] tsdgeos: in the mocks it's a string [16:26] but ok, thanks! [16:26] faenil: but is it unique? [16:26] yeah, [16:27] then it's "ok2 [16:27] " [16:31] faenil, none of the results' attributes should be treated as a key, we do not impose any constraints on any values to be unique.. as tsdgeos says you may relay on the pointers, but this is of course a hack/workaround. note, when you do a new search or refresh, you get new result instances even though the data looks the same [16:33] pstolowski: mmm ok... [16:33] * tsdgeos missed something [16:33] what's the ok for? [16:34] faenil, on new search we basically wipe the model and create new result instances [16:34] pstolowski: I see [16:34] tsdgeos: in reply to pstolowski [16:34] oki [16:36] tsdgeos, fyi "faenil, none of the results' attributes should be treated as a key, we do not impose any constraints on any values to be unique.. as tsdgeos says you may relay on the pointers, but this is of course a hack/workaround. note, when you do a new search or refresh, you get new result instances even though the data looks the same" [16:37] sure [16:37] i meant that for a given set of results [16:37] you can use the pointer as key [16:37] but won't be stable or regresh or something [16:37] tsdgeos, ok.. i didn't know what is it for [16:38] pstolowski: me neither :D === dandrader is now known as dandrader|lunch [17:58] mzanetti, Hi! who is the best person for bug 1436982 ? [17:58] bug 1436982 in unity8 (Ubuntu) "Unity freezes for a few seconds when 'clear messages' is tapped in Notifications" [High,New] https://launchpad.net/bugs/1436982 [17:59] actually it should be dednick ^ [17:59] probably [17:59] let me try to reproduce [18:02] 'Clear all' button. [18:02] yeah... so far not happened. but only had telegram notifications in it. let me try with a missed call === alan_g is now known as alan_g|EOD [18:05] mzanetti, same happens when you change volume from sound Menu [18:06] when notification appears, try to flick close the menu up [18:06] hmm... no problem here... [18:06] om26er, maybe try checking if unity8.log prints something suspicious [18:09] mzanetti, on rtm right now, will flash and update you/bug_report. === dandrader|lunch is now known as dandrader [18:16] mzanetti, om26er: yeah, that's probably me. [18:16] ack [18:22] dednick, mzanetti try to clear while the notification is still on screen. === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === greyback__ is now known as greyback === tmpRAOF is now known as RAOF