[00:41] desrt: Hey, how would you like your client logging integration? [00:42] (As in: how would you like messages from libmirclient to be presented to you) === chihchun_afk is now known as chihchun === 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 === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [10:29] alf__: it isn't urgent, but as no-one "owns" USC could you please review this at a convenient point? https://code.launchpad.net/~alan-griffiths/unity-system-compositor/migrate-to-mir-Server-API/+merge/240566 [10:33] alan_g: sure === dandrader is now known as dandrader|afk === greyback_ is now known as greyback [11:13] what do I need for setting up lightdm correctly to run unity8 on a normal ubuntu install? [11:17] mlankhorst: http://unity.ubuntu.com/mir/using_mir_on_pc.html (I hope it is up to date - I've not used if for ages) [11:18] i don't think it is [11:21] What happens? [11:22] mzanetti: IIRC you've used unity8 desktop recently. ^^ [11:23] I've installed unity8-desktop-session-mir, but it fails to spawn a native mir session [11:23] * mzanetti reads [11:24] mlankhorst: hmm... that should work. I didn't do anything special. [11:24] mlankhorst: try to first log in into a unity7 session, then logout again and relogin to the unity8 session [11:24] that seems to improve chances of a successful start [11:24] it is still a bit flaky [11:25] mlankhorst: I've heard of people with nvidia cards having issues. I'm using an Intel one === dandrader|afk is now known as dandrader === dandrader is now known as dandrader|afk [14:22] RAOF: 'just use the journal'? [14:22] RAOF: otherwise i guess a callback is fine [14:22] btw does the unity8 session run fullscreen on mir? [14:25] mlankhorst: yes it does [14:25] so if I create a Xmir window there, it will be hidden by the u8 session? [14:30] yes, if that mir client is connecting to unity-system-compositor. But if you connect it to unity8 (MIR_SOCKET=$SDG_RUNTIME_DIR/mir_socket) it should appear in unity8 [14:30] you'll need to append "--desktop_file_hint=/usr/share/applications/some-desktop-file.desktop" to the process, else it won't be accepted [14:31] as unity8 is strict, and requires either upstart to launch the app, or else that command line arg appended to the command [14:35] ah [14:35] would that explain the broken pipe? :P [14:41] where can I find that in the source :p [14:46] that policy is defined in lp:qtmir:/src/modules/Unity/Application/application_manager.cpp:ApplicationManager::authorizeSession [14:49] ok [14:49] thanks [14:53] hey [14:53] is it possible to start a demo mir server somewhere while keeping the an xorg/unity7 session running? [14:54] I tried to "sudo mir_demo_server_shell --vt 1" and switch to vt1 but I couldn't come back (or there is a bug with my i965 which bugged the machine) [14:55] seb128: that used to work, lemme try [14:56] greyback, thanks === chihchun is now known as chihchun_afk [15:00] seb128: yep worked here [15:01] seb128: ctrl+alt+backspace quits the mir server [15:01] as it tends to ignore Crtl+C in the console that started it [15:02] greyback, thanks [15:02] greyback, can you ctrl-alt-f1 ctrl-alt-f7 to switch between mir and unity7? [15:02] seb128: yes [15:06] greyback, k, thanks for that and for the email ;-) [15:06] seb128: thought it might be handy [15:07] it is indeed :-) [15:07] seb128: while I've got you, somehow unity7 isn't showing up as an option in my lightdm Greeter. Only unity8 is there. ubuntu-desktop is installed. Any ideas? [15:07] greyback, install ubuntu-session [15:07] seb128: getting the server working is another issue... [15:08] seb128: you need to change ownership of the socket and symlink it to the correct location in your user's xdg_runtime_dir (or set env variables to work around) [15:08] seb128: it is installed [15:09] greyback, weird then, is "unity" installed? [15:09] yep, it is [15:09] hum [15:09] I'm a tad confused [15:09] but it's not end of world, it's my unity8 desktop machine anyway [15:09] do you have anything useful in /var/log/lightdm/ ? === chihchun_afk is now known as chihchun [15:09] x-0-greeter.log [15:10] desrt, do you know if there is a wikipage or something about the socket thing? [15:10] that the greeter doesn't use mir makes things hard for me. :P [15:10] seb128: no. i don't. [15:10] seb128: unfortunately this is sort of something that everyone has to discover for themseelves.... [15:10] seb128: also: you need to run the server as root [15:11] desrt, the email I crossed had those instructions [15:11] * desrt has burned half an hour or so figuring these things out, twice (after having forgot the first round) [15:11] sudo mir_demo_server_shell --vt 1 [15:11] sudo env XDG_RUNTIME_DIR=/run/user/1000 MIR_SOCKET=/tmp/mir_socket [15:11] oh. nice. [15:11] the second command is missing the [15:11] that's a strange combination of environment variables :) [15:11] but you are saying that's not going to be enough? [15:11] you need to change ownership of the socket if you wnat to connect as a normal user [15:12] k [15:12] I'm fine running the command under sudo [15:12] but thanks for the hint [15:12] * desrt has had a couple of outstanding requests to improve the situation here... [15:12] seb128: 'don't gtk as root' ;) [15:12] desrt, on a test laptop EDONTCARE ;-) [15:13] * greyback has todo item to write up this stuff, sorry hasn't got it done [15:13] that would be useful [15:13] http://unity.ubuntu.com/mir/ doesn't have a lot of useful user informations [15:14] it seems to be hard to run your own unity8 currently.. [15:14] at least when not going through lightdm [15:18] * desrt is reminded that he has been meaning to file some bugs, and does so: https://bugs.launchpad.net/mir/+bug/1398038 and https://bugs.launchpad.net/mir/+bug/1398039 [15:18] Launchpad bug 1398038 in Mir "need nested mir (in X) server" [Undecided,New] [15:18] Launchpad bug 1398039 in Mir "test servers should have commandline options for socket owner, location" [Undecided,New] [15:22] Morning [15:22] seb128: extremely messy, but here's the biggest of my notes on unity8/mir http://pad.ubuntu.com/using-mir [15:28] greyback, thanks [15:29] * alan_g finds mir doesn't build today === dandrader|afk is now known as dandrader === alan_g is now known as alan_g|tea [15:34] alan_g|tea, what's not building === alan_g|tea is now known as alan_g [16:22] https://errors.ubuntu.com/problem/318a0132ff5aca50838dfffb9bdbe88d1edebf01 is showing on e.u.c for vivid [16:23] hitting abrt in mir_connection_create_surface() [16:23] kdub: it turns out to be vivid migration detritus [16:23] (just pointing it, dunno if you guys watch e.u.c) [16:47] seb128: not sure where that report originates, but a test called "createing_surface_on_garbage_connection_is_fatal_Test" forking a process that aborts doesn't seem like anything to worry about [17:04] New rev for mir-event-2.0 :) [17:09] alan_g, right, it's still generating some noise on the error tracker so would be good to fix/disable [17:12] seb128: I've not looked closely, but we have a number of tests whose purpose is to ensure an abort occurs when required. Are you suggesting that we shouldn't test those scenarios? [17:12] alan_g, hum, no, I just wonder if we could do something to avoid getting e.u.c reports for them [17:13] that's kind of low importance though, so feel free to ignore me ;-) [17:13] If you can suggest a way to do that in Mir I can follow it up. Otherwise... [17:14] good point [17:15] I'm going to investigate if that's possible, thanks ;-) === dandrader is now known as dandrader|afk === greyback is now known as greyback|afk === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === dandrader|afk is now known as dandrader === alan_g is now known as alan_g|EOD === ValicekB_ is now known as ValicekB [18:28] camako, was thinking about how you mentioned that you were looking through the platform code in trying to not tear down the display system on compositor start/stop [18:29] I can point to what android is doing there if that helps [18:29] but it seems to mostly be something the server is doing [18:29] and perhaps some improvement has to be done to the mesa platform too [18:59] kdub, thanks. I looked at Android side, and am now looking at the Mesa side which is more involved. [18:59] camako, yeah, the android side its easy to keep the display around... iirc, the mesa side has some complication with VT's === chihchun is now known as chihchun_afk [19:03] camako: Thanks for your comments on event-2.0 will get back toyou right after lunch === chihchun_afk is now known as chihchun === Trevinho_ is now known as Trevinho [22:32] desrt: Oh, I was wrong about how much done eventloop fds is; it's much closer to finished than I thought :)