[00:13] RAOF, Yeah, I tried xeyes and xterm [00:13] Hm. They should trigger stuff. [00:14] Setting the MIR_CLIENT_RPC_REPORT=log environment variable might give a clue. [00:14] Oh, wait. It rendered circles [00:16] RAOF, http://paste.ubuntu.com/11802164/ [00:16] Hm, that's not with MIR_CLIENT_RPC_REPORT set, is it? [00:17] No, trying to set that now [00:20] tedg: I don't suppose you could try it *outside* a VM? You might simply be hitting a qxl bug. [00:23] RAOF, Seems the same with that env var set? [00:23] RAOF, Uhm, I don't have anything running wily on real HW. [00:24] Hm, it should spew at least some RPC log... [00:24] Softcock! [00:24] Where would it? To stdout? [00:26] Here is unity8 log: http://paste.ubuntu.com/11802196/ [00:27] It should print to stdout, yeah. [00:36] tedg: You should get something like: http://paste.ubuntu.com/11802221/ [00:36] Oh! Maybe you need to pass -verbose 7 to XMir; it might be doing stupid things with stdout. [00:39] Let me try that, rebuilding. [00:47] RAOF, (EE) Unrecognized option: -verbose [00:47] Gah, stupid bloody DDX split. [00:49] Why is -verbose not a core X option? Bah. [01:21] RAOF, Got a new error in compwindow.c:488 http://paste.ubuntu.com/11802347/ [01:21] Turned on -rootless [01:22] Same error with xterm as well. [01:22] (first was xeyes) [01:40] Harumph. [01:41] tedg: Ok, so Xmir seems to work fine in mir_demo_server locally. [01:44] So, then could it be Unity8 playing with us? [01:45] It seems that I get a updateMirSurfaceSize and then a stopping [01:46] old (1024,768) new (480, 376) [01:46] Seems to be from the screen size 1024x768 to a window size [01:47] Hm, that's odd. [01:47] With rootless you shouldn't see the first; without rootless you shouldn't see the second. [02:55] RAOF, So any recommendation on where to go with this? I feel like I'm starting things correctly, but it doesn't work. [02:55] Not sure how to keep this moving forward. [02:56] I guess first check whether you can get Xmir working not-rootless in mir_demo_server. If that works, then it's something unity8 ish. [02:56] If it doesn't, that suggests a problem in qxl-glamor land. [02:57] Well, I'm kinda at the integration point. I'm not sure anything I have would work really without Unity8 [02:57] Urgh. [02:58] Though, I mean, if you think it's likely to be an QXL thing, I'm happy to ship it. [02:58] Just worry when the libertine container folks start to get to this point I'll have lied about its completedness :-) [02:58] It *could* be a qxl thing. [03:01] I guess you could also try deleting /usr/lib/x86_64-linux-gnu/dri; that should make Xmir drop into software-rendering-only mode. [03:01] Which would bypass any qxl bug. [03:06] RAOF, That still got the compwindow.c:488 error [03:07] You're running not rootless? [03:07] I am running with -rootless [03:32] I think -rootless will need a compositing window manager running in order to work. [03:32] Maybe? === chihchun_afk is now known as chihchun === duflu_ is now known as duflu [07:46] oh we forgot to publish the mesa platform operations [07:46] hm [07:46] why does xmir require those [07:59] * conyoo error parsing serv.dat [08:07] alf_: xmir used to do mir_connection_drm_auth_magic.. [08:10] i changed that to a mir_connection_platform_operation [08:11] i think we have to publish mir_toolkit/mesa/platform_operation.h [08:20] anpok_: it is in mesa-client-platform-dev [08:21] anpok_: or rather, mir-client-platform-mesa-dev === chihchun is now known as chihchun_afk [08:30] oh, thx [09:33] https://plus.google.com/u/1/+PopescuSorin/posts/GNRw5yjHEgF shows qtcreator on mir ... does it mean we got multisurface handling in mir now? ;-) [09:36] anpok_: Is there a way to turn off input resampling from the command line (env. variable or cmdline option) for a demo server? [09:36] yes [09:37] MIR_CLIENT_INPUT_RATE=0 [09:38] ah, so resampling is a client side operation? [09:38] anpok_: ^^ ? [09:38] yes [09:39] anpok_: Does it apply to both nested server and normal client? That is, should I set MIR_CLIENT_INPUT_RATE=0 in both? [09:39] and you can disable it for the nested server too [09:39] yes [09:39] anpok_: thanks === 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 [13:32] anpok_: ping [14:01] greyback: yip? [14:01] oh what a delay [14:01] anpok_: hey I need a hand with something: lp:~unity-team/unity-system-compositor/toggle-cursor2 [14:02] anpok_: that should build against mir 0.13 [14:02] anpok_: I'm unable to make the TestCursorEnabler test pass, and I simply don't understand why [14:03] I'm using mir::event::make_event to assemble a mouse event, but in the event handler, it's detected as a touch [14:04] can you see my error? [14:09] you should store mir::EventUptr instead [14:10] because those events you create inside the test only live till the end of the expression [14:10] make_event returns unique_ptrs [14:11] hum how do I push a new branch with git to launchpad? [14:13] ah ok .. just git push rep branch.. [14:14] greyback: and because of the way the allocator works in that case the memory pointed to by the event contained the data of the touch event [14:15] anpok_: unique pointer deletes the MirEvent once it goes out of scope? [14:15] unique ptr has exclusive ownership [14:16] and when it dies .. it takes the object pointed to with it [14:16] so to keep it alive, you have to store it.. [14:17] and unique_ptr is move only.. so you can never have a situation of shared ownership of two unique ptrs.. [14:29] anpok_: yep that was it, ta [14:29] I got get() and release() confused ;) [14:34] hi, i'm trying to compile SDL2-2.0.2 with --enable-video-mir but i get this error: recive for target 'build/SDL_mirdyn.lo' failed [14:37] ../configure --disable-mir-shared --disable-video-opengl --disable-video-wayland --disable-video-x11 --enable-video-mir [14:37] am i doing something wrong? [14:38] alber_m: can you shar ethe whole build log please, that's not much info to go on [14:42] :/ i can't copy on Xmir [14:42] i'm running xmir/firefox/irc [14:43] pff [14:44] i;m on wily [14:45] hm why did you disable-mir-shared? === ogra_` is now known as ogra_ [14:47] alber_m: hm bschaefer would know better.. but he is not yet around [14:47] anpok_: any idea where the mir branch for silo0 has gone? [14:47] nope [14:48] hmm === andyrock_ is now known as andyrock [14:48] you want the up to date android multi monitor handling, and the override-orientation stuff? [14:48] anpok_: mostly the override orientation stuff [14:49] i mean without the old android mm stuff.. [14:49] right [14:53] anpok_: i get the same error with or without --disable-mir-shared [14:58] alber_m: here's my instructions to build SDL on the phone platform. Might help in your quest: https://github.com/mcphail/ubuntu-touch-sdl-template/blob/master/lib/src/how_to_build_sdl_for_ubuntu_phone.txt [14:59] guest1244: which libmirclient version are you building with? [15:01] anpok_: if you don't disable-mir-shared you get segfaults like this: https://bugs.launchpad.net/ubuntu/+source/libsdl2/+bug/1448544 [15:01] Launchpad bug 1448544 in libsdl2 (Ubuntu) "[MIR] SDL_Init() crashes on the bq Aquaris E4.5 phone" [Undecided,New] [15:10] mcphail: oh looks broken.. maybe there was an abi break that was not detected.. because libsdl2 dynamically loads libmirclient [15:11] anpok_: I don't really understand what layer is broken, but know SDL plays nicely with libmirclient if the dynamic loading is disabled. bschaefer gave the tip [15:12] anpok_: don't know if it has been a new abi break - afaik it has )never_ worked [15:12] ah [15:13] s/ah/no clue/g .. that failue would make sense on top of mesa.. [15:14] anpok_: note that I have not tried sdl with mir on anything but the phone === dandrader is now known as dandrader|afk [15:37] greyback: hmm do you really need the override-orientation branch? [15:37] greyback: oh for silo-0 we will still let usc display the cursor? [15:38] anpok_: well if I can get it working quickly, it'll have to do [15:39] mir is missing ability for unity8 to receive relative pointer coordinates [15:39] which we need to have unity8 draw the cursor === dandrader|afk is now known as dandrader === yofel_ is now known as yofel === dandrader is now known as dandrader|lunch === dandrader|lunch is now known as dandrader === \b is now known as benonsoftware === mibofra is now known as Guest47179 === dandrader is now known as dandrader|afk === Guest47179 is now known as mibofra === dandrader|afk is now known as dandrader [23:40] Hey, vogons! Has anyone seen this failure before: https://jenkins.qa.ubuntu.com/job/mir-wily-amd64-ci/394/consoleFull [23:41] It seems that something in glib is leaking an fd created in GSourceHandle's destructor?