[09:46] does anyone know why the "mediumtest" jobs are not running in CI since the switch to vivid? [09:48] alan_g: no [10:19] alf_: " alan_g: looks like the mir-mediumtests-vivid-touch doesn't exist, will ping fginther" [10:20] alan_g: ack, thanks [10:21] alan_g: the bright side is that our MPs pass through CI a lot faster now :) [10:21] LOL [10:23] https://code.launchpad.net/~andreas-pokorny/unity-system-compositor/expose-powerkey-state-and-screen-toggle-controller-interface would be nice if someone reviews this change.. but not top approve.. it is not meant to be released next week [10:24] i found a lot of low hanging fruits while doing this.. [10:24] anpok: ok [10:25] the actual change is kind of trivial... but i needed the code separation.. [10:25] i mean.. i struggled with relating the code to the actual system behavior [10:25] especially those timeouts.. [10:26] I wonder why/when the tests where commented out... [10:27] alf_: it was an accident [10:27] because they werent working with the cross compile setup [10:27] usc is also one of the pacakges that do not build with a x-sbuild [10:27] alan_g: anpok: thanks [10:36] alan_g: anpok: I was thinking about whether it's worth for time::Timer to have the convenience functions notify_in() and notify_at(), since we can just say "auto alarm = timer.create_alarm(); alarm->notify_in(..);" instead. [10:38] That looks like a repeating pattern that could be simplified. So yes. [10:38] i would have no objection to reduce it to just that [10:38] ...and the convenience functions return an Alarm object anyway, which we still have to keep alive [10:39] so from looking at usc .. it is also just .. create_alarm + notify_in that is in use [10:42] alan_g: anpok: Even if we want them as part of the interface, I am not sure they are worth it as pure virtual functions. Perhaps they could have a default implementation. [10:43] anpok: yes, the same in Mir proper [10:44] I agree they're not appropriate for overriding. Technically they could even be free functions. [10:45] But I'd go for non-virtual TemplateFunctions (not C++ templates, just the design pattern) === dandrader is now known as dandrader|afk === alan_g is now known as alan_g|lunch === dandrader|afk is now known as dandrader === chihchun_afk is now known as chihchun === alan_g|lunch is now known as alan_g === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === alan_g is now known as alan_g|EOD === chihchun is now known as chihchun_afk [20:02] hi hi miristoj [20:02] what's the word on the prerolled nature of the cursor-names available for use with mir? [20:51] hi, when i try to run an app from a vt, i keep getting an error like Failed to connect to Mir: connect not called or Failed to send message to server: Broken pipe [20:54] attente_, make sure you've set the socket permissions [20:54] as well as [20:55] export MIR_SOCKET=/tmp/mir_socket (or /path/to/socket) [20:55] bschaefer: the socket is in the xdg runtime dir with the correct perms [20:55] desrt, when i run mir_demo_server_* it puts it in /tmp for me :( [20:55] this is a unity8 graphical session [20:55] oo i see [20:56] usually if you've mir in a vt you've got a mir_demo running [20:58] bschaefer: i'm just trying to figure out how i can run something under unity 8 in a terminal session [20:59] attente_, oo hmm i think ubuntu-app-launch? [20:59] will let you do that, or --desktop_file_hint=/path/to/desktop/file [21:01] bschaefer: ubuntu-app-launch is giving me Unable to connect to Upstart bus: Cannot launch D-Bus without X11 $DISPLAY [21:02] attente_, i've had that before ... but i didn't get around it :( [21:02] thats a ted question [21:02] tedg, ^ [21:03] attente_, usually means for some reason you didn't get the DBUS_SESSION_BUS_ADDRESS variable set in your environment. [21:04] attente_, How is your shell session being created? [21:04] tedg: i'm logging into unity8 from the greeter [21:04] attente_, And then opening a terminal? [21:05] Command line shell, not graphical shell. [21:05] * tedg might be partially to blame for that confusion [21:06] tedg: the command line shell is from a vt [21:07] attente_, Okay, so you'll need to get the dbus session bus address. It should be /run/user/$(user number)/dbus-session [21:07] bschaefer: is MIR_SOCKET required to be set? i thought it derived it from the XDG_RUNTIME_DIR [21:07] attente_, no i was wrong in what i thought you were doing [21:07] attente_, No [21:07] i thought you were using a mir demo, not unity8 haha [21:07] It'll get MIR_SOCKET from Upstart though [21:07] yeah [21:32] attente_, This might also be helpful depending on what you're doing: https://lists.launchpad.net/ubuntu-phone/msg09839.html [21:35] tedg: thanks, i'll try it! [22:01] camako, is there someone who can take a look at http://s-jenkins.ubuntu-ci:8080/job/mir-mediumtests-runner-mako/3297/ ? I've updated this to run with vivid bits and it's failing. [22:02] camako, It's currently not enabled as part of the mir ci jobs, but I can if that's the right thing to do [22:16] fginther, I'll take a look [22:17] camako, it looks like for this run, mir_performance_tests, took too much time and the job timed out. The timeout is 60 minutes and the whole thing usually takes < 20 minutes [22:18] fginther, I see [22:22] fginther, we were having problems with this test and disabled it in CI... Can't remember if we reenabled it later [22:22] camako, it was re-enabled right after the device sprint [22:24] camako, I ran it again without mir_performance_tests, http://s-jenkins.ubuntu-ci:8080/job/mir-mediumtests-runner-mako/3298/console [22:25] * camako looks [22:27] camako, looks like this one failed: mStaleFrames.are_dropped_when_restarting_compositor [22:27] fginther, ok let me investigate and I'll get back to you [22:28] camako, thanks. Shall I enable the test for -ci jobs or just hold off until you've investigated? [22:30] fginther, aren't -ci jobs currently enabled against our MPs? Not sure if I understand correctly what is being enabled. [22:31] desrt: What do you mean? [22:31] camako, sorry, I need to give you more context... When the jobs were converted to vivid, the mir mako test was missing and as a result wasn't running on MPs within the last day [22:32] camako, I would prefer to enable it so that it can be fixed with just an MP [22:32] fginther, Ah ok... I think we should enable them... Just to increase urgency on our end [22:32] camako, thanks [22:40] fginther, have you committed your vivid changes? Can I get a link to the tree? [22:41] camako, there were no changes to how the tests are executed. Only that the packages are now built for vivid (and the image is devel-proposed which is now vivid) [22:42] fginther, I see... [22:44] fginther, so still using lp:~mir-team/+junk/mir-medium-test-runner-for-jenkins, right? And if I download the archive.zip mentioned on the link you sent above, then I should be able to repro, correct? [22:44] repro locally , I mean [22:45] camako, correct