[00:32] back in a few hours [01:25] Bah. We have two nearly identical places where we probe the graphics platform. [01:25] I always forget to update one :* === chihchun_afk is now known as chihchun [07:13] Bah, cmake. [07:15] Why can't you pkgconfig correctly? [07:33] racarr_: yes, at some point when heavy rendering is going on the whole system freezes for some time and then continues normally, I "solved" it by going back to the latest 3.15 kernel. Perhaps an early 3.16 kernel will do but I haven't had the time to bisect. [07:46] RAOF: duflu: Hi! Any idea about http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-mako/2347/console ? [07:46] (see failure at end of log) [07:47] alf_: Yup: Package libboost-program-options-dev is not installed. [07:48] Not using apt + a repository strikes again! [07:48] alf_: It seems build-deps are prepared based on the system installation. And those are wrong for the new tree being built. Someone needs to kick in the new package ... libmirplatform-dev:armhf depends on libboost-program-options-dev; however: [07:48] Package libboost-program-options-dev is not installed. [07:48] So.. it's only CI that's broken :/ [07:49] ... or our x-compile scripts. I thought Alberto made them smart and this should happen automatically [07:49] ... assuming CI uses our scripts [07:52] duflu: Incorrect! [07:52] :) [07:52] RAOF: OK *shrug*. I care not for new problems this week :/ [07:52] There's plenty of old problems still [07:54] alf_: Want to try with /~raof/+junk/testrunner-testing/? :) [08:01] alf_: I stole your bug BTW ... bug 1340510 [08:01] bug 1340510 in Mir "session screen seen upon quick power key strike" [Medium,In progress] https://launchpad.net/bugs/1340510 [08:02] duflu: Ah, good, thanks. I wanted to unassign myself since I got caught up with other things, but forgot. [08:05] RAOF: sure, how do I try that? [08:05] alf_: I've already started - http://s-jenkins.ubuntu-ci:8080/view/mediumtests/job/mir-mediumtests-runner-mako/2348/console [08:05] RAOF: great, thanks [08:06] Faster, silly device! [08:09] Bah. Stupid thing. [08:12] alf_: Died with a transient unrelated error, so I've started it again. [08:24] Sweet. [08:24] Looks like that's working. [08:30] alf_: That looks like it's good; feel free to push that branch to the main testrunner bit. [08:30] EOW for me! [08:31] RAOF: Thanks! Enjoy you weekend! [08:38] alf_: is "libmirplatform:armhf to be configured is 0.5.1+14.10.20140728-0ubuntu1" being addressed or should I jump on it? [08:42] alan_g: RAOF's update to the test runner fixes this /~raof/+junk/testrunner-testing/, I will merge into our "upstream" testrunner branch [08:44] Excellent [08:45] * alan_g goes off to play with parsing our code with libclang === tvoss is now known as tvoss|errands === chihchun is now known as chihchun_afk === dandrader is now known as dandrader|afk === tvoss|errands is now known as tvoss === lool- is now known as lool [11:59] alf_: hey, I've not dug yet, but I fear our "last frame not drawn" bug has appeared again https://bugs.launchpad.net/unity8/+bug/1354407 - it may be qtmir, and it may be mir. Just an FYI as yet [11:59] Ubuntu bug 1354407 in Unity 8 "last frame not being drawn - left edge swipe reveals old frame of dash, need to interact with it to refresh it" [Undecided,New] [12:00] greyback: ack === alan_g is now known as alan_g|lunch === greyback is now known as greyback|lunch === dandrader|afk is now known as dandrader === alan_g|lunch is now known as alan_g === greyback|lunch is now known as greyback === dandrader is now known as dandrader|lunch [15:25] Morning. [15:26] late psuedo standup: Still working on emulator bug...have something to test now but not 100% sure how to build an appropriate image (modified the emulator target egl impl) [15:27] proposed a qtmir test skeleton === om26er is now known as om26er|dinner [15:33] sigh touchspot renderables failed [15:33] the sha1sum check again due to other stuff landing [15:34] groan [15:34] I'm considering making an mp that deletes it and adds a file like [15:35] README-release.txt [15:35] that has a checklist which includes [15:35] 1. Have any headers in include/ changed [15:35] which is just as easy to tell from the diff [15:35] as looking at the sha1sum file [15:36] yeah, I think part of the problem too is we need to firm up the interfaces [15:37] like look at how the downstreams are using the interfaces [15:37] and then bury the stuff they arent using [15:37] mm [15:37] and make sure those that are used are sensible [15:38] thats not quite the immediate problem, but along the same tangent and important [15:41] my guess is there is a bunch of stuff that could be hidden at this point but not sure [15:42] yeah, and even more if you take the set of stuff that is only used by the demo shell [15:43] We don't want to be quite that strict - there's likely stuff we should expose that hasn't been used for the phone [15:45] yeah, agreed [15:45] I guess my main point is that an audit of the interfaces would help [15:54] Can I turn on the Mir LTTng tracepoints with an env var, or is it something I need to compile in? [15:57] env variable! [15:58] tedg: see doc/component_reports.md [15:59] http://unity.ubuntu.com/mir/component_reports.html [15:59] so e.g. MIR_SERVER_INPUT_REPORT=lttng [15:59] I doubt that list of reports is up to date [15:59] lol [16:00] Ah, cool. [16:00] So I guess that needs to go in the session startup script [16:01] I'd like to get a report on when surfaces are allocated and displayed [16:01] Looking at app startup [16:01] Which would be helpful? :-) [16:03] tedg: You could get that from the session mediator report which logs IPC calls I guess [16:03] i.e. create_surface [16:03] racarr, Cool, thanks! === dandrader|lunch is now known as dandrader [16:05] :) np === chihchun_afk is now known as chihchun === om26er|dinner is now known as om26er [16:38] rsalveti: Ping? [16:39] I am trying to test changes to the emulator libEGL...and having some difficulty understanding exactly how the emulator images are assembled === alan_g is now known as alan_g|EOW [17:19] racarr: https://wiki.ubuntu.com/Touch/Emulator === om26er_ is now known as om26er [18:32] rsalveti: So I see on one hand there is the -system parameter to the emulator binary and then also [18:32] the system.img is copied in to the [18:32] /var/lib/lxc...etc and mounted at boot [18:33] what is the deal with that? [18:33] Having to deviate from the instructions slightly for various reasons so trying to understnad what is happening [18:34] rsalveti: also having terouble understanding the raw disk vs ...non raw distinction [18:39] also what is the ubuntu-*system.img vs just [18:39] system.img [18:40] sorry...im really unknowledgeable about the whole image business I guess :) [18:47] haha...ok now I tried using raw disks, so now I can mount my new ubuntu-system.img and found [18:47] a libEGL with a new md5sum so I guess that worked [18:48] well [18:48] nvm lets see what happens when I replace system.img with ubuntu-*system.img [18:48] actually lets have lunch [18:48] brb [19:02] ohhhh [19:02] theres a system.img [19:02] inside the system.img [19:02] ... [19:05] grr [19:33] Yay it works [19:35] it turns out the system.img that building the android source package produces is the system.img inside the system.img inside the sdcard.img that creating an emulator instance [19:35] creates [19:35] and the third system.img is just a decoy [19:35] (the one outside the sdcard.img) [19:35] lol [19:35] how does something like that happen [20:07] racarr: yeah, it's horrible [20:07] we're fixing that soon [20:07] really confusing atm [20:14] rsalveti: :D [21:15] racarr: Hi hi. I tried to convert your Chromium ozone-mir branch into a distro patch for current stable Chromium -- really the only way I can use it -- and I had some trouble. Do you have some time to help me debug it? [21:17] As I understand it, we would like to avoid code duplication with ozone-wayland, so we try to extend it, overriding where we need to.