=== chihchun_afk is now known as chihchun [10:28] racarr: what was the little handy trick to get jenkins to pick up an mp ?..."request review" pick jenkins bot user...but what do you put in "type of review" ? "continuous integration"? === chihchun is now known as chihchun_afk [11:04] kgunn: Yes [11:04] continuous-integration === chihchun_afk is now known as chihchun [11:42] RAOF: Does this match the ugliness of the surrounding code? https://code.launchpad.net/~alan-griffiths/mir/hack-the-hack/+merge/197501 === chihchun is now known as chihchun_afk [13:04] anpok: http://unity.ubuntu.com/mir/index.html === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [14:27] mterry, so the nested was working with the simple client [14:27] kdub, awesome, as I'd expect [14:31] kdub, so do you want help setting up greeter mode? [14:32] mterry, probably not enough time, have to jump into meetings [14:32] could you point me at the code? maybe a visual inspection will give clues [14:32] kdub, OK [14:32] kdub, heh, it's not a small branch. lp:~mterry/unity8/split === chihchun_afk is now known as chihchun [14:45] anyone wants to test mesa10 before I upload it to the archive? I've copied it to https://launchpad.net/~ubuntu-x-swat/+archive/ppa/+packages -- should be published shortly [14:46] kdub, hmm, my other crash I was looking at (which might be same issue?) is crashing in libhybris. Will try to get more info for ya, to confirm it's a Mir problem [14:51] ok it's up now [16:32] RAOF: can you please please accept pixman sru to saucy? :P [16:33] and probably raring, quantal, precise too but less priority there [16:40] mlankhorst: Sure, in a bit. === chihchun is now known as chihchun_afk === olli_ is now known as olli [17:31] kdub, heyo! [17:32] kdub, so I think this other crash I'm seeing on the N7 is related to the nested-mode-does-too-much issue. It's also crashing in the_graphics_platform(), possibly when trying to load the graphics library [17:32] kdub, this N7 crash is easier to setup than the greeter crash [17:32] kdub, and in fact, is actually blocking nested mode from landing, whereas the greeter one isn't [17:33] kdub, so I'd love some of your time tomorrow for debugging [17:33] mterry, is anyone beyond me looking into maguro at the sprint ? [17:34] i'm at the point where i'm stuck in a loop of: [17:34] ioctl(14, 0xc01c6730, 0xbe99708c) = 0 [17:34] writev(15, [{"\6", 1}, {"IMGSRV\0", 7}, {":0: PVRSRVCreateDCSwapChain: Err"..., 50}], 3) = 58 [17:34] writev(15, [{"\6", 1}, {"IMGSRV\0", 7}, {":0: framebuffer_device_open: Fai"..., 68}], 3) = 76 [17:34] nanosleep({1, 0}, 0xbe997158) [17:34] which points to the driver ... which i have not much clue about [17:34] ogra_, last time I checked in, it was using SF mode. Did you force it to use mir mode? [17:34] mterry, maguro uses mir since saucy [17:34] we released it with mir [17:34] ogra_, kdub thinks nested mode is doing too much hw configuration [17:34] ogra_, no no [17:35] yes yes :) [17:35] ogra_, I mean your stracing was showing it trying to use SF mode [17:35] oh, right [17:35] ogra_, despite the fact that it should use Mir [17:35] well, i forced it to Mir by exporting the var before stracing [17:35] yeah [17:35] ogra_, so if nested mode is doing bad hw config, that would explain why we see such weird results across hardware [17:35] there are other issues too for sure [17:35] ah, k [17:35] ogra_, so I'm hoping for some sort of silver bullte [17:36] ogra_, though having to force Mir is bad... [17:36] hehe, well, you are surely at the right place to catch one [17:36] yes, there are issues in lightdm or logind [17:36] * mterry will get kdub drunk enough to agree to help [17:36] but these are separate from unity8 falling over [17:36] ++ [17:36] ogra_, I'll have to find a maguro to see why it's unique wrt to lightdm [17:36] i'll happily pay the beer, just send me the bill [17:37] lightdm at least I can navigate, unlike this Mir stuff [17:37] ogra_, so you forced MIR_SOCKET. What did you do to force /run/user to behave? [17:37] well, i suspect lightdm doesnt do its job because Mir doesnt do right .... but we'll see [17:38] i added some mkdir, chown, chmod stuff the the wrapper to get the dir [17:38] that gets me a proper upstart session in which i can play with unity then [17:39] and stracing unity8 gets me the above framebuffer error in a loop until it gets frustrated any exist [17:39] *exits [17:40] ogra_, I'm very confused by the runtimedir failure. That really shouldn't be affected by Mir. But it seems to work in non-nested mode... /me hopes for large silver bullet [17:41] well, lets see, i'll hunt down more people tomorrow ... i'm sure we'll get somewhere before end of the week [17:41] as long as we can get Mir to start at all :) [17:51] alan_g: https://code.launchpad.net/~raof/mir/hack-hack-hack-remove-a-hack/+merge/197581 [17:51] RAOF: looking... [17:53] * alan_g thinks that looks much better! (but does it work? ...) [17:56] alan_g: Works for me :) [17:58] RAOF: and me. APPROVED!!!! [18:13] duflu: so [18:13] for platform-api clients, xbgr or bgr for occlusion support? [18:13] we can just select that one directly in src/ubuntu/mirclient/ [18:13] no need for more smart opperation [18:13] ricmm: Either should be fine [18:14] ricmm: Don't stick to only one pixel format though. Expect the next device you try to not support it [18:15] duflu: everything else is in development-branch? [18:15] ricmm: Even better. Everything else is released in archive Mir 0.1.2 already [18:15] ah perfect === mhall119 is now known as mhall119|afk