=== kdub is now known as kdub^holiday === duflu_ is now known as duflu [02:11] duflu, re: bug #1255045 - it's very difficult to see on maguro the difference between black and turned off [02:11] bug 1255045 in mir (Ubuntu) "screen does not turn on on maguro when pressing the power button" [Critical,Confirmed] https://launchpad.net/bugs/1255045 [02:11] duflu, but I can see completely no reaction on the screen when pressing the power button [02:15] duflu, and indeed the patch didn't help [02:19] Saviq: Can you look from a wide angle? A backlight usually bleeds through even a black screen from a sufficient angle [02:19] duflu, yeah, problem is maguro is AMOLED - no backlight... [02:19] I suspect I've seen a similar bug on either Nexus 4 or Nexus 7. Can't remember which [02:19] Saviq: Oh, AMOLED FTW :/ [02:20] duflu, https://bugs.launchpad.net/ubuntu/+source/mir/+bug/1255045/comments/13 for some syslog activity [02:20] Ubuntu bug 1255045 in mir (Ubuntu) "screen does not turn on on maguro when pressing the power button" [Critical,Confirmed] [02:21] duflu, so yeah, the lack of activity between power button and touch does not... prove anything... but also does not rule out your "it's black, not blank" suspicion [02:21] * Saviq tries surfaceflinger [02:21] Saviq: We might have to start with one of the Nexus's where it's more obvious on the traditional LCD [02:22] duflu, yeah, except that bug is only happening on maguro isn't it? [02:23] Saviq: Maybe... I don't have maguro but am sure I saw a similar wake-on-touch problem on Nexus 4/7 some time recently [02:24] Hmm... and then there's the question; if the screen is truly asleep why power the touch sensor? [02:24] duflu, ok, under surfaceflinger I managed to see the difference between blank and black... shite it's subtle, trying again under mir [02:25] doesn't help it's past 3am here.. [02:30] so yeah, one step closer... I'm gonna go now... o/ === xubincs is now known as xubincq === xubincq is now known as xubincs === chihchun_afk is now known as chihchun === xubincq is now known as xubincs === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [09:45] didrocks, https://code.launchpad.net/~vanvugt/unity-mir/fix-1255045/+merge/197016 seems to work! [09:45] oh [09:45] really? [09:46] didrocks, just doing some more tests, but yeah [09:46] as in... waiting 1 minute ;) [09:46] Mirv: ^ [09:46] Mirv: so, if all planet aligns, can you publish Mir? [09:47] didrocks, Mirv, I've unity-mir packages, but building unity8, too, as they need to be released together this time [09:47] Saviq, happy to test, too [09:49] didrocks, Mirv, tvoss_ http://people.canonical.com/~msawicz/mir-maguro-blank/ [09:49] echo "expect stop" > ~phablet/.config/upstart/unity8.conf [09:51] didrocks, Mirv, tvoss_, oh and yeah, that with ppa:ubuntu-unity/daily-build [09:51] didrocks: Saviq: so the Mir landing actually becomes Mir + unity-mir + platform-api + u-s-c + unity8 landing? [09:52] but yeah, preparing at least for such a beast [09:52] Mirv, ultimately yes, but we still have a commit missing in unity8 before that can happen [09:53] Saviq: ok [09:53] ah crap [09:53] Saviq, ? [09:53] didrocks, Mirv, tvoss_ the above ↑↑ not to .conf, but to .override [09:54] Saviq: and that unity-mir fix too [09:54] Mirv, right, that too [09:54] ok, just updated the landing plan to reflect [09:55] Mirv, could we publish Mir first, then land stuff in unity and unity-mir (as we're blocked on unity8 due to bug #1255578) [09:55] bug 1255578 in Ubuntu CI Services "dependency issues on libunity-mir1 in testrunner-otto" [Undecided,Confirmed] https://launchpad.net/bugs/1255578 [09:55] Mirv, and then land unity8, unity-mir, unity-scopes-shell? [09:55] Saviq: land mir, platform-api, unity-mir, unity-system-compositor, unity8 is the current list [09:56] Saviq: only what is absolutely required, is unity-scopes-shell part of it too? [09:56] Mirv, no, not absolutely required [09:56] Mirv, https://code.launchpad.net/~saviq/unity8/raise-sigstop/+merge/191212 is though [09:56] ok, a separate landing item for that then please [09:56] right, marking that to the sheet too [09:56] Mirv, was already there [09:57] Saviq: so is that MP "a commit missing"? [09:57] Mirv, but "waiting for code" [09:57] Mirv, not sure what you mean? [09:58] Saviq: you said that unity8 is missing a commit, so is thar raise-sigstop the only commit needed still for unity8, similar to maguro fix being the only commit needed for unity-mir? [09:58] Mirv, yes [09:58] ok :) [09:58] and yes the scopes is there in the landing asks [09:59] Mirv, but we can't merge into unity8 until Mir is published, due to the above bug [09:59] Mirv, or well, we can force it in [09:59] Saviq: aha, ok, then a manual merge would be needed [09:59] Mirv, ok, will do once there's more people +1'd [10:00] yep, that's useful. bzr branch lp:unity8, bzr merge, bzr commit --author , bzr push. but yeah, only in very special cases, after enough +1:s and with landing team's approval I'd say :) [10:00] or well, we can add daily build to it temporarily [10:01] will run -ci with daily build on them, then [10:01] oh right that technique too [10:13] Mirv, hrm, any idea why that'd have failed http://s-jenkins.ubuntu-ci:8080/job/unity-mir-trusty-armhf-ci/29/console ? [10:13] libmirclient-dev (>= 0.1.2) but 0.1.1+14.04.20131120-0ubuntu1 is to be installed. :/ [10:14] Mirv, ah, Mir building in daily-build? [10:15] seems we'll have to wait for that [10:25] alan_g, re: https://code.launchpad.net/~vanvugt/unity-mir/fix-1255045/+merge/197016/comments/456116 [10:25] alan_g, http://bazaar.launchpad.net/~mir-team/unity-mir/trunk/view/head:/src/unity-mir/dbusscreen.cpp#L45 [10:26] wouldn't configure_output take care of pause/resume? [10:26] or well, display->configure(), for that matter? [10:28] Saviq: it just seems strange to pick out compositing [10:29] alan_g, sure, I agree (bear in mind, I dunno Mir enough), but since we're already configuring the display, couldn't it do pause/resume internally, if there's no screens enabled? [10:30] Saviq: I know the infrastructure isn't in place, but there should be a change notification for the display area covered by the screen [10:30] At present it is a pretty brutal "there are changes somewhere" [10:32] Saviq: seems not using daily-build PPA the 0.1.2 was already there [10:32] Mirv, yeah, it's weird, but I think it got superseded by the new build already [10:33] Mirv, and it's only libmirclient-dev that complained about the version, although that, by apt dep error standards, might not mean anything... [10:33] Mirv, will restart once the new build gets published [10:34] alan_g, ok understood [10:35] * alan_g disgs through the code... [10:41] Saviq: OK, I suspect this is a hack around the absences of a schedule_compositing() call in compositor - and not really intended as a pause()/resume() cycle [10:42] alan_g, sounds right, compositor should know when display comes on and schedule then, right? [10:45] Saviq: well, I think it should be notified (and, in theory, told what area needs display) - unless the idea really is to pause the server [10:46] Mirv, grrr, dep fail again :/ [10:47] Mirv, the hook name is definitely added http://s-jenkins.ubuntu-ci:8080/job/unity-mir-trusty-armhf-ci/30/console [10:47] Mirv, but indeed it doesn't seem to be executed?? [10:48] * Saviq goes to -ci-eng [10:49] Saviq: yeah, probably a job for CI team to check [10:57] alan_g, FWIW, we would actually need a frame composited before turning the display back on, as now you can see the previous frame for a split second when turning the device on [10:57] Saviq: ack [11:00] Saviq: is the code we're looking at executing after the device is turned on or before? [11:02] alan_g, I'd say after [11:02] alan_g, i.e. we get notified that the screen was turned on, not that it's going to be turned on [11:02] alan_g, that's something that will be better when we move the responsibility on that from powerd to unity-mir/unity8 [11:03] Saviq: that makes it hard to composite a frame beforehand [11:03] alan_g, i.e. it will be told "you need to turn on now", and it will decide what to do and when to actually turn on [11:03] alan_g, yeah of course [11:03] Saviq, do we have updated packages, yet? [11:03] Although, I gather there wasn't a flash and then black [11:04] tvoss_, still the same http://people.canonical.com/~msawicz/mir-maguro-blank/ [11:04] tvoss_, + daily-build ppa + unity8.override [11:04] alf_: thoughts? https://code.launchpad.net/~vanvugt/unity-mir/fix-1255045/+merge/197016 (and ^^) [11:04] tvoss_, although I have a unity8 package now [11:04] Saviq, okay, those work for me [11:04] * Saviq adds [11:04] Saviq, will give that package a spin, too [11:05] tvoss_, pushed [11:05] tvoss_, drop the .override for good measure [11:05] Saviq, ack [11:06] /food [11:10] Saviq: alan_g: the DisplayChanger class holds the logic for properly reconfiguring the display (pausing/resuming), so that's what we should expose (or a similar interface), i.e., I don't think display->configure() should pause resume internally [11:14] Saviq: alan_g: about whether we should pause/resume or schedule_compositing() instead, it depends on how well our display classes deal with posting buffers if the display is off (right now not very well it seems) [11:16] Saviq: alan_g: and whether it actually makes sense to disable communications and input [11:16] alf_: I also wonder if we should be processing input when the display is off [11:17] alan_g: don't we need to handle some kind of input event to wake up (e.g. power button press)? [11:18] Hmm. but touch events going to apps? [11:21] alan_g: Turning off communications would solve that. [11:22] alf_: maybe not - there was a time input bypassed the thread pool. I don't remember that changing. [11:24] alan_g: yes, so turning off communications _properly_ would solve that :) [11:26] alan_g: and then there is also the issue of showing previous state when we wake up (because the buffer contents are obsolete)... [11:26] Anyway, it does look like code around patch is a cut down version of MediatingDisplayChanger [11:27] Which, like you say is what we should expose === chihchun is now known as chihchun_afk [11:39] any reason why all shell::Surface methods are virtual whereas surfaces:Surface are not. Both are concrete classes. [11:41] dandrader: insanity [11:41] lol [11:41] dandrader: you're also looking at an old version [11:42] alan_g, hmm, true! [11:43] I'm playing with gerry's "qt scene graph" mir branch [11:44] The newer version doesn't fix all the insanity, but is a step in a right direction. [11:44] * alan_g remembers he wants to look at that before next week [12:25] Mirv, tvoss_, didrocks, ok so bug #1255948 will not let us through upstream merger, OK to force land those two commits before release? [12:25] bug 1255948 in Ubuntu CI Services "upstream merger hooks do not propagate to downstream jobs' builder_hooks" [Undecided,New] https://launchpad.net/bugs/1255948 [12:25] let me look, one sec [12:25] greyback, can we have your ACK on https://code.launchpad.net/~vanvugt/unity-mir/fix-1255045/+merge/197016 ? [12:26] didrocks, it's: [12:26] https://code.launchpad.net/~saviq/unity8/raise-sigstop/+merge/191212 [12:26] and https://code.launchpad.net/~vanvugt/unity-mir/fix-1255045/+merge/197016 [12:26] Saviq: if it's tested and really ready to land (with the whole Mir stuff), +1 [12:26] Saviq: looking === chihchun_afk is now known as chihchun [12:26] * greyback confused why he doesn't get mail when a unity-mir MR is made [12:27] greyback, there's a lot of confusion about LP mails if you ask me... [12:28] * Saviq installs the packages on mako and maguro, too [12:29] s/maguro/manta/ [12:49] didrocks, pushed a fix to debian/control in https://code.launchpad.net/~saviq/unity8/raise-sigstop/+merge/191212 === dandrader is now known as dandrader|afk [12:50] Reading the meaningless (to most people) Android codenames, I think we should use mythical monsters for Ubuntu Touch codenames: Ubuntu Touch 2.0 Kraken, 3.0 Hydra etc :P [12:50] tvoss_, Mirv, I can has more +1s? [12:50] alf_, who would you get the billions of $$ from then? [12:50] Saviq: looks good for packaging change [12:50] alf_, like android did with KitKat? ;) [12:52] Saviq, installing too === chihchun is now known as chihchun_afk === dandrader|afk is now known as dandrader [13:04] hmmmm, what did I do to my nexus? Saviq, have you seen INFO:phablet-flash:Device detected as [13:04] ERROR:phablet-flash:Unsupported device, autodetect fails device [13:04] before? [13:05] tvoss_, that happens when your device isn't booted [13:05] tvoss_, -d [13:05] tvoss_, to skip autodetection [13:06] Saviq, thx, works === alan_g is now known as alan_g|lunch [14:04] tvoss_, any updates? [14:04] * Saviq is tempted to just push the button [14:05] Saviq, works here [14:05] ok, /me pokes the buttons [14:09] Mirv, ready [14:10] Mirv, hope it's not too late for you :/ [14:10] Mirv, both MPs are in trunks [14:15] didrocks, ↑↑ [14:21] \o/ [14:21] I guess Mirv left [14:21] Mirv: ? === alan_g|lunch is now known as alan_g [14:43] ok, no more Mirv I guess :) [14:43] Saviq: so, mir, platform-api, unity-mir, unity-system-compositor + unity8, right? [14:44] Saviq: do you know if the u-s-c FTBFS fix was merged? [14:44] didrocks, no [14:44] don't know [14:44] seems, is was [14:44] it* [14:45] didrocks, looks like it, yeah [14:45] https://code.launchpad.net/~unity-system-compositor-team/unity-system-compositor/trunk [14:45] Saviq: so, I'm adding that to robru's list [14:45] didrocks, thanks [14:45] Saviq: any particular thing to look at in unity8 trunk? [14:45] didrocks, that it starts ;) [14:46] well, that's a given :p [14:46] didrocks, other than that there's a lot happened... 3 weeks no release almost === dandrader is now known as dandrader|lunch [14:50] ok [16:41] alf_: are you fixing the the merge conflicts? https://code.launchpad.net/~afrantzis/mir/mir-client-ensure-global-symbol-resolution/+merge/196110 [16:42] alan_g: I think I will wait for lp:~afrantzis/mir/build-options-for-tests to be merged, so I can follow the same scheme for special linkage tests build options [16:43] alan_g: do you want me to mark as WIP? [16:43] alf_: fair enough [16:43] alf_: No. Nobody's around anyway [16:43] alan_g: ok === dandrader|lunch is now known as dandrader === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === seb128_ is now known as seb128 === thomi_ is now known as thomi