/srv/irclogs.ubuntu.com/2013/09/25/#ubuntu-mir.txt

robotfuelx is locking up when using xmir tonight, I was able to grab this right before the machine started a new test http://pastebin.ubuntu.com/6152767/02:21
robotfuelI'll have a better bug in about an hour02:21
robotfuelrobert_ancell: ping02:44
robert_ancellrobotfuel, hey02:44
robotfuelrobert_ancell: xrandr doesn't say xmir is running when but I see /tmp/mir_socket did that change?02:45
robert_ancellrobotfuel, can you confirm /tmp/mir_socket is owned by u-s-c? Or has this test completed02:45
robert_ancellrobotfuel, /tmp/mir_socket indicates some Mir server is running, which is probably u-s-c02:46
robotfuelsrwxr-xr-x  1 root    root       0 Sep 25 02:21 mir_socket02:46
robert_ancellrobotfuel, run fuser on it02:46
robotfuelrobert_ancell: fuser /tmp/mir_socket returns nothing02:49
robert_ancellrobotfuel, that implies that file is left over from something02:50
robert_ancellrobotfuel, you are checking for that existence of that file to confirm u-s-c is running?02:50
robotfuelroot      1035  1.0  1.0 906132 41188 tty7     Ssl+ 02:21   0:18 /usr/sbin/unity-system-compositor --from-dm-fd 10 --to-dm-fd 13 --vt 702:51
robotfuelrobert_ancell: shows it's running02:51
robotfuelrobert_ancell: /var/log/Xorg.0.log shows mir stuff in it.02:51
robotfuelrobert_ancell: it's at 10.97.2.151 if you want to log in02:52
robert_ancellrobotfuel, hang on, I'll just restart to confirm the /tmp/mir_socket behaviour02:52
robert_ancellrobotfuel, aha, you have to sudo fuser /tmp/mir_socket since it's root owned02:56
robert_ancellkind of expected fuser to consider that an error02:56
robert_ancellrobotfuel, right, xrandr outputs now match their old names, so you can't tell if you're running Mir using xrandr02:57
robotfuelrobert_ancell: doh! yes it's being used by u-s-c02:57
robert_ancellRAOF, there's no X call you know of that can give you information about the driver in use is there?02:57
robotfuelrobert_ancell: ok I'll use mir_socket and fuser to make sure the process id's match instead of xrandr02:58
RAOFrobert_ancell: You can call into DRI2 to get the name of the DRI module to load?02:58
RAOFrobert_ancell: But that's likely not what you want ;)02:58
robert_ancellRAOF, I was kind of hoping there was some sort of convenient vendor string that an X call returned, but I don't remember anything like that02:59
RAOFNope, no dice.02:59
robotfuelI need to get some sleep thanks robert_ancell and RAOF03:00
=== chihchun_afk is now known as chihchun
robert_ancellracarr, kdub, hikiko, alf_, kgunn_, meeting06:00
robert_ancelllooks like just me and alf_ here, and his microphone/camera is broken :)06:05
hikikostill trying to join...06:06
robert_ancellbye all06:10
hikikobye robert_ancell06:10
=== chihchun is now known as chihchun_afk
=== chihchun_afk is now known as chihchun
robert_ancelldidrocks, hey, I noticed the daily releaser did a nice clean up of the changelog. It's nice having guaranteed sane changelogs now :)06:28
didrocksrobert_ancell: oh, daily-release does that from day one FYI ;)06:28
didrocksrobert_ancell: I had to iterate 3 times with Kevin to have the diff you saw already06:29
robert_ancelldidrocks, I wasn't sure if the project version would confuse anything - but if you also can't think of anything I'll update it for consistency06:29
didrocksso knowing the rest would be cleaned up automatically, I didn't fight more ;)06:29
didrocksrobert_ancell: yeah, just update for consistency, from what I know, it's not picked up by anything06:29
didrocksso in case you didn't notice, Mir is in the release pocket ;)06:30
didrocks(had to bribe some release team member to accept the new xorg-server despite the beta freeze)06:30
robert_ancelldidrocks, nice :)06:31
robert_ancelldidrocks, the only place the version is used appears to be the .pc files06:31
didrocksyeah, and has the deps weren't forced, we should be fine06:33
dandraderalf_, so commits flow from development_branch (where the real action happens) to trunk (stable branch)?08:23
alf_dandrader: That's the plan for now. It's not a blocker for branches that don't break the ABI, but keeping merges going in one direction only simplifies the process.08:25
alf_dandrader: (i.e. no need to merge from trunk back into devel)08:26
dandraderalf_, kdub, racarr  ok I have resubmitted to the development branch. But now I need you guys to reapprove it, please :)08:29
alf_dandrader: np, I will reapprove and top-approve when CI is done.08:30
dandraderalf_, thanks!08:30
=== chihchun is now known as chihchun_afk
=== chihchun_afk is now known as chihchun
alan_galf_: I suspect there's no auto-landing job for development-branch (i.e. we have to do the merges ourselves)10:07
=== smartboyhw_ is now known as smartboyhw
=== greyback is now known as greyback|lunch
alf_alan_g: hmm, I remember Thomi saying that he could arrange that?11:44
alan_galf_: looks as though I was wrong.11:45
alan_gHow do I suppress this: /tmp/buildd/mir-0.0.11+13.10.20130904/examples/basic_server.cpp:49:83: error: ignoring return value of 'int system(const char*)', declared with attribute warn_unused_result [-Werror=unused-result]11:46
alan_g             (void)system((the_options()->get(launch_child_opt, "") + "&").c_str());11:46
alan_gSurely casting the result to void isn't ignoring it11:47
=== greyback|lunch is now known as greyback
=== hikiko is now known as hikiko|lunch
alf_alan_g: In the past I have used an "if (bla()) {}" construct to work around this (in test code), but I guess in this case you might as well do something more useful with the return value.12:02
=== alan_g is now known as alan_g|lunch
=== chihchun is now known as chihchun_afk
=== alan_g|lunch is now known as alan_g
alan_galf_: I've tidied up https://code.launchpad.net/~alan-griffiths/mir/socket-connection/+merge/187326 if you want to take another look12:55
alf_alan_g: will do12:56
=== hikiko|lunch is now known as hikiko
=== chihchun_afk is now known as chihchun
=== dandrader is now known as dandrader|lunch
kgunnoh bravo alan_g..... Thumping Termite14:09
kgunnthomi: ^14:09
davmor2kgunn: Teenage-mutant-ninja Turtles is the obvious choice14:11
kgunndavmor2: dang it...how did we miss that one14:12
alan_gkgunn: that wasn't intended for publication14:12
kgunn:))14:12
=== mzanetti is now known as mzanetti|food
=== alan_g is now known as alan_g|tea
=== alan_g|tea is now known as alan_g
alan_gkgunn: are you trying to change the Mir culture of only top-approving when discussion is over? (c.f. https://code.launchpad.net/~alan-griffiths/mir/socket-connection/+merge/187326/comments/427682)14:40
kgunnalan_g: sorry...i missed that14:42
kgunnthanks14:43
kgunnback to needs review14:43
alan_gkgunn: yw. Just wanted to be sure it was intentional14:44
alan_gkgunn: and thanks for getting the ci set up for development-branch. Nice14:45
kgunnalan_g: i know....much better...thank francis :)14:45
=== dandrader|lunch is now known as dandrader
kdubgood morning15:04
alan_gkdub: kgunn hangout?15:04
kdubsure15:04
kgunncoming15:04
=== mzanetti|food is now known as mzanetti
hikikobye15:31
=== chihchun is now known as chihchun_afk
=== alan_g is now known as alan_g|EOD
alan_g|EODkdub: happy reviewing! (https://code.launchpad.net/~alan-griffiths/mir/socket-connection/+merge/187326/comments/427682)17:01
kdubalan_g|EOD, could you switch lp:~kdub/mir/fix-1215979 to approve/abstain before you go?17:02
kdubalan_g|EOD, will review that today17:02
kgunnkdub: so, do we need to wait for alf_ ? on https://code.launchpad.net/~kdub/mir/fix-1215979/+merge/18734618:58
kgunnor is ready for ta18:58
kdubkgunn, comment should be addressed, so it should be okay19:00
kgunnkdub: ok...i'd like to get it landed tomorrow...only way that happens is to TA right now19:01
kgunni'll be doing that now19:01
kdubkgunn, okay19:02
kdubkgunn, we'll might get some complaints from the n4 users about worse stuttering if the users have mixed and matched the phablet build and the mir build19:03
kdubperhaps i should send out a wider email to mir/unity8 folks working on the n419:04
kgunnkdub: understand....this just gets it merged onto our dev19:04
kgunnthat's probably a good idea...19:04
kgunnkdub: does rsalveti already have matching mp's on FB/hwc side ?19:04
kgunnor ?19:04
kgunnmaybe you put some mp's up?19:05
kdubaccording to email, should be in the next image19:05
rsalvetikdub: kgunn: it's already in today's image19:08
kgunnhey hey...even better19:08
* kgunn wonders how rsalveti lands stuff so fast19:09
rsalveti:-)19:09
kgunnkdub: do you still have a GalaxyNexus?19:17
kdubyep19:17
kgunncurious if you think with mir if its usable as a dev device (i know the perf aint great)19:17
kgunnkdub: one other thing to check my thinking....i should just bump our so name on our dev branch, then push the whole damn thing to trunk....e.g. we want it all w/ history right ?19:19
kgunne.g. push --overwrite lp:mir19:19
kdubnot sure what --overwrite does :/19:19
kdubmerging is probably better19:20
kgunnmmmm...it ignores any conflict, but that would only matter if an mp went in....19:20
kgunnso i guess thats not right19:20
kgunnit would be better to merge....but somehow perserve history19:20
kdubkgunn, what we should do is bump the so in the dev branch, get that landed, and then instruct CI to autoland the revision that has the bump19:21
kdubwell, jenkin's autolander for lp:mir19:21
kgunnkdub: "get that landed" means onto lp:mir trunk right19:21
kdubno, land onto the dev branch19:22
kduband then merge that into  lp:mir19:22
kgunnkdub: oh...ok...yeah...we're thinking the same19:22
kgunntoo many "lands"19:22
kgunnkdub: i'm just wanting to make sure we pull over history per commit with it19:23
kdubkgunn, after a merge, bzr log -n0 shows the individual revisions in the branch that was merged19:24
robotfuelrobert_ancell: ping22:37
robotfuelRAOF: ping22:37
robert_ancellrobotfuel, hello22:37
RAOFrobotfuel: Pong22:37
robotfuelwe have this new bug from the xmir stuff that landed last night https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/123108422:38
ubot5Ubuntu bug 1231084 in xserver-xorg-video-intel (Ubuntu) "xorg freezes with xmir after changing resolution with xrandr " [Undecided,Confirmed]22:38
robotfuelit's only on intel and only 2 of the 3 intels in the lab22:38
RAOFHrm.22:40
RAOFDoes running with MIR_BYPASS=0 fix it?22:40
* RAOF is currently in a moderately interesting XDC talk, so mental bandwidth is low.22:41
robotfuelRAOF: where do I add MIR_BYPASS=0?22:45
RAOFrobotfuel: Export it from /usr/bin/unity-system-compositor.sleep22:46
RAOFrobotfuel: Ahem. /usr/sbin/unity-system-compositor.sleep22:49
robotfuelRAOF: MIR_BYPASS=0 fixed the issue. it's actually not freezing when the resolution changes it's freezing when using xrandr --output VGA-0 --off23:00
robotfuelRAOF: the next step is the resolution change which hangs, but the ui is frozrn after the xrandr --output VGA-0 --off I'll update the bug23:01
RAOFOk.23:01
RAOFI think I've seen that before. We're probably blocking the (Mir) frontend, so messages are no longer being handled.23:03
mlankhorstRAOF: hm considering you're at XDC (part of me is jealous, but I'm also glad), want me to look at any xmir things?23:19
mlankhorstany specific23:19
RAOFmlankhorst: Not that I'm aware of :)23:20
RAOFmlankhorst: Oh, you could push my xf86-video-ati patch if you felt like it :P23:20
mlankhorstah forgot :P23:20
mlankhorst*writes it down*23:20
mlankhorstok will probably do it tomorrow, bedtime for me :P23:21
RAOFYeah, no rush.23:21
RAOFI've applied the most interesting bit in Ubuntu already.23:21
* mlankhorst is in the mood for some triaging tomorrow23:21
RAOFmlankhorst: Oh, and there's always https://bugs.launchpad.net/mir/+bug/1195811 :)23:31
ubot5Ubuntu bug 1195811 in linux (Ubuntu) "nouveau: Abnormally high FPS and tearing (no vsync) with nouveau" [High,In progress]23:31
mlankhorstRAOF: Yeah no luck there, requires upstream cooperation :/23:37
mlankhorstI did write a fix that I keep in my tree23:37
mlankhorstRAOF: oh bug darktama for me please23:38
mlankhorstpester him as much as you can about high resolution vblank timing23:38
mlankhorstthat patch fixes that issue :P23:38
robert_ancellracarr, around?23:44

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!