/srv/irclogs.ubuntu.com/2015/02/26/#ubuntu-unity.txt

* tsdgeos looks at dbus and asks it why so much locking09:19
tsdgeoshttp://paste.ubuntu.com/10425178/09:19
Saviqtsdgeos, multiple locks get you a discount on home insurance, maybe that?09:39
tsdgeosCE 0x7f506c00fa10 009:40
tsdgeosQDBusConnection: warning: blocking call took a long time (803 ms, max for this thread is 500 ms) to service "com.canonical.Thumbnailer" path "/com/canonical/Thumbnailer" interface "com.canonical.Thumbnailer" member "GetArtistArt"09:40
tsdgeosCE2 0x7f506c00fa10 0 80109:40
tsdgeosSaviq: there's a hidden joke in there i didn't get i'd say :D09:40
Saviqtsdgeos, nvm ;)09:41
tsdgeospreliminary analisis, it's hard to fix :D10:03
tsdgeosqt likes doing it's dbus on the main thread it seems10:03
tsdgeosso if you do a sync call from a thread it still blocks the main thread10:03
tsdgeosneed to verify10:03
tsdgeoss/it's/its10:04
Saviqtsdgeos, yay :/10:41
Saviqmzanetti, tsdgeos, about the one failing autopilot test, if you have enough items in the launcher for it to start collapsing, one of the tests will fail, could that have been your issue?10:49
tsdgeoswasn't launcher related as far as i remember10:50
Saviqwell, it's not a launcher-related test that fails10:50
Saviqit was test_click_app_icon_on_dash_must_focus_it IIRC10:50
tsdgeosthen this is the one that mzanetti had failing afair10:58
mzanettire (was in a meeting)10:58
mzanettithat was my failing one: test_get_applications_should_return_correct_applications10:59
Saviqok, different one, then10:59
tsdgeosah11:00
tsdgeosbad memory :D11:00
Saviqgreyback, shall I land the two top-approved branches for qtmir?11:27
greybackSaviq:  lp:~gerboland/qtmir/fix-lifecycle-exempt-keeps-wakelock is safe to land. The other one I think requires some unity8 changes on desktop (mouse != touch any more)11:28
Saviqgreyback, ack11:29
=== MacSlow is now known as MacSlow|lunch
=== alan_g is now known as alan_g|lunch
=== MacSlow|lunch is now known as MacSlow
Saviqmterry, hey, not sure if you saw - I can confirm jibel's wizard-interrupts bug13:41
mterrySaviq, I did see!  Bummer.  I couldn't reproduce when I tried yesterday, but will try again, after some MIR stuff13:42
mterrySaviq, I don't *think* it's the dash stealing focus, because I think we ignore the dash when killing the wizard on a focus13:42
mterrySaviq, but will double check13:43
Saviqkk13:43
om26ermzanetti, yo. whats the best place to download unity8 with the inprogress window management features ?13:58
Saviqom26er, vivid silo 0 has a bit more than vivid itself13:58
om26erSaviq, so vivid is a must ? I can't try on 14.10 ?13:59
om26erSaviq, also just apt-get install unity8 ?13:59
Saviqom26er, we can't afford to backport everything to stable13:59
Saviqom26er, `citrain host-upgrade 0` more like14:00
Saviqom26er, but yeah, vivid is the only thing we've tested with14:00
=== alan_g|lunch is now known as alan_g
Saviqhmm hmm14:17
Saviqdandrader, "/usr/bin/ld: cannot find -llightdm-qt5-2", rings a bell?14:18
dandraderSaviq, no, why you ask?14:18
Saviqdandrader, that's me trying to build http://people.canonical.com/~platform/citrain_dashboard/#?distro=ubuntu&q=landing-019 locally14:18
Saviqdandrader, and I remember something similar initially after your lightdm mock refactor14:19
* Saviq tries trunk14:25
Saviqok yeah, trunk doesn't build either14:25
Saviqit's probably ninja or ccache :/14:26
Saviqok, so a missing depends between targets it must be14:33
=== dandrader is now known as dandrader|afk
mterrySaviq, and in that wizard-disappears bug, you don't see the greeter either?  It goes right to the dash?14:40
Saviqmterry, correct14:40
mterrySaviq, that's super werid14:40
Saviqmterry, reproduced on both mako and arale, too14:41
mterrySaviq, is the factory reset important, or can it happen after a normal flash?14:41
Saviqmterry, it happened after wipe for me14:41
mterryoh you say bootstrapping/wiping works14:41
mterrySaviq, is picking "French" part of the repro steps?14:42
Saviqmterry, "Polish" in my case, but might be the thing14:42
Saviqmterry, like that might be the thing that triggers a restart of services14:42
mterrySaviq, yeah14:42
mterrySaviq, and out of curiousity, what is your SIM setup at the time?14:43
mterry(like slots, locked, etc)14:43
Saviqmterry, none14:45
mterrySaviq, OK (SIM page is next, thought maybe that was involved14:45
=== dandrader|afk is now known as dandrader
mterrySaviq, just got it!  Yeah, it does feel a lot like the dash is getting focused, killing the wizard14:55
Saviqmterry, looks like it from the console log as well14:55
mterrySaviq, you mean you see the dash being launched a second time?14:57
mterryYeah, I see that too14:57
Saviqgreyback, how do I verify your qtmir wakelock fix?15:45
greybackSaviq: start music app, play a song, blank the screen, run "sudo powerd-cli list" and verify the only wakelock held is pulseaudio's15:47
Saviqgreyback, btw, we could use a better name than "active" there ;)15:49
greybackthe concept will migrate out of qtmir eventually15:51
Saviqgreyback, in any case, doesn't seem to work15:51
greybackwhat doesn't?15:52
Saviqgreyback, permanent "active" lock held when music is running15:52
greybackby pulseaudio?15:52
greybackwithout my patch, there would be 2 locks, one by shell, one by PA15:52
Saviqgreyback, not playing yet15:52
Saviqgreyback, just launched music, locked screen15:53
greybackthat should not be the case15:53
Saviqgreyback, goes away when I put music in background15:53
Saviqgreyback, that's the silo http://people.canonical.com/~platform/citrain_dashboard/#?distro=ubuntu&q=landing-01915:54
Saviqbrb, testing desktop fixes15:57
greybackSaviq: hmm you're right, wtf16:01
greybackmy patch releases it ok, but something re-acquires another one16:03
Saviqgreyback, FWIW it seems to be qtmir still, as the name is the same16:15
greybackSaviq: it is, was my fail, I'm just testing my fix now16:15
Saviqkk16:15
Saviqgonna test first thing tomorrow then16:15
dandradergreyback, do you know what's holding it from getting merged? https://code.launchpad.net/~mir-team/qtubuntu/port-to-mirclient/+merge/24516416:27
greybackdandrader: on desktop, indicators are not reponding to mouse input events any more16:27
greybackwe may need to add MouseAreas where DirectionalDragAreas are, to do something sensible.16:28
dandradergreyback, ermm... what does it have to do with this branch?16:36
greybackdandrader: it breaks indicators16:37
greybacktry it and see16:37
dandradergreyback, you know why? it's not supposed to cause any behavioral changes...16:37
greybackdandrader: I didn't look closely into it, but my guess was that DDA only listens for touch events, not mouse events16:38
dandradergreyback, but this qtubuntu branch is meant to make qtubuntu bypass papi and use mirclient directly....16:40
dandraderno change how input events are handled... or maybe I'm missing something16:40
greybackdandrader: oh sorry, I mis-read the name. I thought it was the qtmir branch you were mentioning16:40
greybackno there's nothing stopping the qtubuntu branch from landing, other than me liking to land it with the qtmir one16:41
dandradergreyback, it can't be landed by itself?16:41
dandradergreyback, along with the papi one, that is16:41
greybackit could, but testing it in the silo will not be that useful, as the new code-paths won't be exercised until the qtmir bits are in there too, no?16:42
mterryTrevinho, for bug 1425362, is there an easy way to get the same env as you?  Make a VM, install unity8-desktop-session-mir, set lightdm.conf.d to use unity sessions and then log into unity8?16:42
ubot5bug 1425362 in unity8 (Ubuntu) "unity8 crashed with SIGSEGV in QLightDM::GreeterImpl::authenticateWithPam() when logging in" [Medium,New] https://launchpad.net/bugs/142536216:42
dandradergreyback, ah, you mean the mouse input handling part will be left unused until the qtmir branch lands16:43
greybackdandrader: yeah16:44
dandradermake sense16:44
dandradermakes16:44
dandradergreyback, I think mzanetti made a branch that sprinkles some MouseAreas around to make controls usable with mouse input16:45
greybackdandrader: indeed, think we'll grab that16:45
dandradergreyback, so we could get that branch + qtmir port-to-event-2.0 + qtubuntu's use-mirclient + the papi one and land them now?16:47
greybackdandrader: I think so. Want to look after that?16:47
dandradergreyback, I would. but I don't have landing super powers. What can I do?16:47
greybackI'm still stuck in usc/mir16:47
greybackdandrader: get the branches ready, then ping me and I can request a silo16:48
dandradergreyback, ok16:48
mzanettigreyback, dandrader: yes. (was in a meeting)16:49
dandradermzanetti, so is https://code.launchpad.net/~mzanetti/unity8/indicators-mouse/+merge/250429 still WIP?16:49
mzanettidandrader, yeah, that's really just a hack for silo016:49
mzanettidandrader, I think this requires design input16:50
dandradermzanetti, well, it serves to avoid regressing when landing qtmir's port-to-event-2.016:50
mzanettidandrader, ah, I see...16:51
greybackwill also need to see if we need a FFe or not16:51
mzanettiok well, I guess we can put it in then... it's quite simple anyways, just adds a MouseArea16:51
mzanettibut the whole topic requires design input and more work16:52
=== dandrader is now known as dandrader|lunch
=== alan_g is now known as alan_g|EOD
=== dandrader|lunch is now known as dandrader
=== dandrader is now known as dandrader|afk
veeberskgunn, et. al regarding this bug: https://bugs.launchpad.net/autopilot/+bug/1422797, I have a better idea what's happening with it, I'm just not sure what else autopilot can do in this situation20:13
ubot5Launchpad bug 1422797 in Autopilot "Mir refuses the app to connect" [High,Incomplete]20:13
veebersThe camera app is crashing and when autopilot attempts to stop the app it times out as ual-stop doesn't stop it within 10sec (due to the crashing/apport).20:14
veebersthen the next couple of tests fail to launch the app again as ual fails to launch it (again, due to it still in the process of crashing(20:15
veebersAs far as I can see autopilot is stopping/launching things as expected, unless I'm missing something?20:15
kgunnveebers: but sounds like there is no time accounted for a crashing app....20:16
kgunnveebers: question...20:16
kgunnif the camera app does not crash...is there a problem ?20:16
veeberskgunn: not that I'm aware of, I've been unable to reproduce, so relying on the logs from ci.20:16
veeberskgunn: I should have time later this morning to check that20:16
kgunnveebers: i guess i'm asking, if the AP test is meant to catch problems, and the camera app crashes....that is a problem20:17
veeberskgunn: the suggestion is that ual should be doing the right thing too20:17
kgunnbut...if apport is running the whole system freezes (i've witenessed this as well)20:17
veeberskgunn: that's a good point, and perhaps a possible feature.20:18
kgunnveebers: i guess i'm saying, while it'd be nice to run the remaining tests....technically, the AP caught the crash20:18
kgunnand you can always find out what the remaining test results are, after you stop and fix the crash you found20:18
kgunnyou=proverbial you20:19
veeberskgunn: sure, so what makes sense in the event of a crash, should ap wait for apport to finish (if it's running), how long should it wait before continuing or doing something more heavy handed20:20
kgunnveebers: yeah, i would argue wait...b/c engineers are gonna ask "did you get a crash file" :)20:20
kgunnto try and debug th eproblem20:21
veeberskgunn: isn't a crash file generated anyway?20:21
kgunnas for a time to wait...i've seen it run for what felt like a solid 10 seconds before....20:21
kgunnveebers: true...i guess it doesn't have to process it20:22
veeberskgunn: well, at the end of the test autopilot fires of an attempt to ual-stop the application and times out after 10 sec20:22
veebersalso, when attempting to launch the app it gives it 10 to start successfully20:22
veebersso in the test log I reviewed, it appears there are 30+ seconds from failure to stop and the next test being able to start it again20:23
kgunnveebers: i suppose there's always polling? is apport running ?....or diable apport processing like you say20:23
kgunn?20:23
veebersI don't think it's autopilots place to disable apport20:24
kgunnveebers: yeah, then you gotta go with either longer waits or polling to see what's running20:24
veebersas to polling, how long makes sense to wait for apport, what if that goes bung too and then we have infinite wait20:24
kgunnman...30+seconds...wow20:24
kgunnveebers: does apport go bung ?20:24
veeberskgunn: Not sure, I haven't seen it personally , just thinking of things to consider :0)20:25
kgunnimho, solve the first problem, worry about finding an infinite wait some other day :)20:25
kgunnveebers: hopefully....we shou;dn't be having too many crashes...and when we do if apport hangs as well, i would hope that would be a very very very small amount of instances :)20:26
veebersyes, true20:26
veeberskgunn: so at this point, I don't think the linked bug is related to autopilot (ap just triggers it perhaps) and there should be a new bug assigned to autopilot related to doing <something> when it detects the app under test has crashed20:27
veebersThat way we can backlog it and get something done on it20:27
kgunni concur20:27
kgunnveebers: if my concurrence helps in any way :)20:28
veeberskgunn: can I ask you to file that bug with your thoughts and expectations for it please?20:28
veebersit sure does20:28
kgunnveebers: isn't this bill's bug ?20:29
kgunni don't own ual either20:29
veeberskgunn: oh perhaps, sorry we were talking about this the other day hence me bothering you, sorry ;-P20:29
kgunnnp20:29
redlama42Is this a good place for questions about Unity8 or is that elsewhere?21:27
spaceindaverHi all, I am having getting unity to run after logging in. I just get my desktop background without any launcher or status bar21:46
spaceindaverAfter checking the compiz logs I see that it failed to load opengl21:46
spaceindaverany idea how I could fix this?21:47
=== dandrader|afk is now known as dandrader
dandraderredlama42, yes, this is the place22:01
redlama42dandrader: Whenever I try and log into a unity8-mir session lightdm just makes logging in anywhere impossible and I have to restart it.22:03
redlama42dandrader: You know what's up?22:03
dandraderredlama42, hmm, no. I very seldomly try out unity8 in a desktop/laptop machine. I'm almost always working with it in phones and tablets only22:05
redlama42dandrader: Damn, thanks anyways.22:08

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!