RAOFNote to self: don't run xorg integration tests while performing tasks you care about.01:28
dufluThat's odd. My main machine can't run GL clients any more. And I upgraded _nothing_ since yesterday01:40
dufluBack in a minute01:44
dufluI hope01:44
dufluWoo, it's not PPA problems this time. We landed something in multi-monitor which makes half the example clients crash01:51
dufluping kduv01:56
dufluping kdub01:56
dufluYeah, right, hurray02:01
* RAOF is off for lunch03:15
duflualf__: out->used implies out->connected right?07:03
dufluActually, it doesn't have to... but used is still more important I guess07:03
alf__duflu: As things are now (and I don't really expect this to change) out->used implies all of out->connected, has valid mode etc07:04
duflualf__: What about when an output is in use, but suddenly disconnected? Does the server reconfigure or would that make used && !connected ?07:05
dufluI'm guessing an output could be "used" but possibly not "connected". That's an error to log, but theoretically possible07:06
alf__duflu: the new configuration that reaches the client should have an updated used field (i.e. not used) if an output is not connected07:08
duflualf__: Sounds reasonable. But I will check all conditions for now...07:08
alf__duflu: but I guess it doesn't hurt to sanity check all the relevant fields07:08
alf__duflu: ok07:08
dholbachgood morning07:18
didrocksRAOF: good evening!07:37
RAOFdidrocks: Aloha!07:37
didrockshow's life in the winterish side of the world? :)07:37
RAOFDark, now :)07:38
didrocksbut but… you're not in the IoM!07:38
didrocksdid you see my emails about xmir and so on?07:38
RAOFdidrocks: I did indeed. The XMir stack is preparing for an upload.07:39
didrocksexcellent! so still on track for today?07:40
RAOFIt might even include an autopkgtest for X, depending on how ambitious I am.07:40
didrocksmaybe let's try to have a first upload, and then adding it :p07:40
didrocksnot adding a potential last minute "zomg"07:40
RAOFI should be able to finish it this evening.07:40
didrocksexcellent, do you mind poking/sending me a quick shotout?07:41
didrocksso that I promote stuff in main in proposed07:41
RAOFdidrocks: Sure.07:42
dufluUgg. bzr tells me there's nothing to merge, but also my diff is 10kloc. It's not.07:42
didrocksRAOF: excellent! thanks a lot. Let's cross fingers07:43
didrocksnothing to merge, 10kloc? hum…07:43
didrocksare you in a case with \r\n eventually?07:43
didrocksI've already been trapped by that07:43
dufludidrocks: bzr was confused by me merging branches, reverting, and then trying to merge again. The diff remained but there was "nothing to merge"07:47
dufluFixed now07:47
didrocksah, but the revert was07:47
didrocksbzr merge . -r -1..-2, right?07:47
didrocksso you have a commit07:47
didrocksthen, merging a revert07:47
didrocksand committing, right?07:48
dufludidrocks: Yes I think I did that. Because I'd already backed it up offsite. So didn't want to uncommit07:48
didrocksok, so that makes sense, not bzr being confused :)07:48
didrocksbasically the revert is "new code" more advanced than your merge itself07:48
didrocks(there is no marker that this commit is a revert of this one, and should be considered again then)07:49
dufludidrocks: Yes it looks like bzr is trying to mark branches as merged, even when they differ07:49
didrocksright, because the commit id is already in07:51
didrocksbasically, you have to bzr branch trunk, then, creating a manual diff (from your commit change) and reapplying it :(07:51
kgunnduflu: hey...tvoss keen to push for your switch branch because it fixes some issues....but07:51
kgunni remembered kdub had found a perf hit on android07:51
duflukgunn: Fixed AFAIK07:51
kgunnjust curious if you had a chance to investigate07:52
kgunnoh swet07:52
dufluHe was testing it too early07:52
kgunn...or sweet even07:52
duflukgunn: Nexus4 gives 60FPS on everything07:52
kgunnkdub: ^ wanna give another go ?07:52
kgunnduflu: is it switching or bypass branch ?07:52
duflukgunn: switch07:52
kgunnduflu: awesome07:53
duflukgunn: I will still ask kdub to re-test though07:53
kgunnduflu: think you'll be able to mp next week ?07:55
duflukgunn: I was aiming for today. Fingers crossed07:55
duflukgunn: That's "switch", not "bypass"07:55
kgunnduflu: right...still lot's of effort...great07:56
duflukgunn: Incidentally, I suspect the fix I did for the Android frame rate might resolve some of greyback's concerns about Mir's swapbuffers being slow for Qt08:02
dufluDoes anyone have operator permission to ban morphis?08:16
ubot5Help! Channel emergency! (ONLY use this trigger in emergencies) - elky, Madpilot, tritium, Nalioth, tonyyarusso, PriceChild, Amaranth, jrib, Myrtti, mneptok, Pici,  jpds,  gnomefreak, bazhang,  Flannel, ikonia, maco, h00k, IdleOne, bkerensa, nhandler, Jordan_U, DJones or k1l!08:17
dufluNot big deal. Wait and see if the noise stops I guess08:18
seb128duflu, I'm asking on #ubuntu-devel, that's actually happening on -devel and -touch as well and is quite annoying08:20
duflukgunn: Regarding 30FPS on Android... I just discovered that can happen if you accidentally have multiple Mir servers running. It actually works, at half framerate08:29
kgunnmakes sense08:29
duflualf__: Would this be multimonitor? https://bugs.launchpad.net/mir/+bug/120722608:39
ubot5Launchpad bug 1207226 in Mir "[regression] Synchronous clients become asynchronous framedropping clients after a VT switch" [Undecided,New]08:39
alf__duflu: I can't reproduce this. Is this with lp:mir? What's your monitor setup exactly, and what commands are you running running?09:01
duflualf__: Yes plain old lp:mir. One monitor connected (but three supported). Just run any client and it happens09:02
alf__duflu: which server?09:02
duflualf__: demo_server_shell at least. Let me test demo_server09:02
duflualf__: If you can't reproduce it, I will debug tomorrow09:03
alf__duflu: could it be that you have left-over processes?09:03
duflualf__: No, I've rebuilt and retested several times09:03
dufluActually it's probably just the compositor losing vsync on VT switch09:04
alan_gduflu: alf__ - anyone need to review https://code.launchpad.net/~robertcarr/mir/client-focus-notifications/+merge/166440 again - or shall I top-approve to get it off the review queue?09:15
duflualan_g: Yes I need to09:16
dufluOr at least verify my previous concerns were addressed09:16
duflualf__: Apologies. I thought I had tested with trunk, twice. Now only happens in my own branch :)09:21
alf__duflu: no worries09:24
dufluToo many machines/branches going at once09:24
dufluI did have the forethought to test trunk. And yet I messed that up somehow09:25
alan_gduflu: time for rest09:25
alan_gRAOF: got time to fix https://code.launchpad.net/~mir-team/mir/expose-debug-buffer-id-to-client/+merge/176843 so that we can land it?09:26
duflualan_g: I won't get to that review today. But I would like to tomorrow....09:44
alan_gduflu: that's OK, I'm just keen to reduce the review backlog (but only where possible)09:45
RAOFalan_g: Not today; barring firefighting, tomorrow.09:48
alf__RAOF: alan_g: could we change it to WIP then?09:54
RAOFalf__: Done.09:55
alf__RAOF: great thanks09:55
alf__RAOF: let's see how light we can make our MP list :)09:55
didrocksRAOF: can I help you in any way?09:58
RAOFdidrocks: Not really; I'm just preparing the packages, testing them, and then I'll upload them. In a bunch.09:59
dufluIt's a bad sign that I read "barring firefighting, tomorrow" as "barfighting tomorrow"10:00
alf__duflu: :D10:00
didrocksRAOF: good luck! :)10:02
RAOFdidrocks: Hm. This might not all be ready and sufficiently tested for me to upload tonight. Is it urgent that I push through, or is your late today/early morning your time ok?10:37
didrocksRAOF: I prefer we push safe things, just be aware that as I have then to promote build-deps in main, it will take time10:37
didrocksRAOF: I wanted to ensure that today's Mir + u-s-c worked TBH (and be done with this)10:38
didrocksbut again, better to be on the safe side of things :)10:38
kgunnrobert_ancell: i was gonna test lankhorst's mp....unless, you tell me you're already doing it10:39
RAOFdidrocks: I probably _could_ get it done today, but it'd be late.10:39
didrocksRAOF: that would be better, but I don't want you having to overwork for it :)10:40
didrocksRAOF: just keep me posted, we'll see10:40
alan_ghikiko: BTW your work got a mention: http://www.phoronix.com/scan.php?page=news_item&px=MTQyNTI10:46
hikikohaha :)10:49
hikikothat's only an option10:50
hikikoanyway, I wope we have the full mir on mir soon10:51
robert_ancellkgunn, testing it now10:56
hikikoalan_g, I have a question11:09
hikikoI added a check that the host socket and the socket that is used by the nested mir to accept connections is not the same11:10
RAOFGrah. Stupid alioth!11:10
hikikoand if it11:10
hikikos the same I just throw an exception11:10
hikikobut I wonder if it's better to rename the socket myself px add a 0 at the end11:11
hikikoeg /tmp/mir_socket for host /tmp/mir_socket0 for nest11:11
alan_ghikiko: no. it is better to report the problem than hide it11:11
hikikocool :)11:12
alan_ghikiko: Use AbnormalExit to give a nice explanation11:12
RAOFdidrocks: Grr. Alioth seems to be down again; I won't have the full stack ready tonight.11:19
=== hikiko is now known as hikiko|lunch
didrocksRAOF: don't worry then, thanks for trying :)11:40
=== alan_g is now known as alan_g|lunch
xnoxhikiko|lunch: so can I run and test it yet? =)11:47
hikiko|lunchxnox, not yet I am afraid :) sorry eating! brb in a while11:52
=== dholbach_ is now known as dholbach
=== hikiko|lunch is now known as hikiko
thomirobert_ancell: https://bugs.launchpad.net/mir/+bug/120731212:31
ubot5Launchpad bug 1207312 in Mir "Need the ability to record video of the screen " [Undecided,New]12:31
=== alan_g|lunch is now known as alan_g
robert_ancellalf__, regarding https://code.launchpad.net/~robert-ancell/mir/alt-ctrl-backspace/+merge/178036 - do you want to merge your common code in now and I'll rebase on it, or do the rebase yourself when merging it in?13:29
alf__robert_ancell: Let's get the common code in first, and I will give rebasing a try. I will let you know if there are any issues. BTW, we also need VT switching code.13:31
alan_ghikiko: I've made a second review comment - after more thought than the first.13:31
robert_ancellalf__, I have the VT switching code done13:32
robert_ancellin lp:~robert-ancell/mir/setsid-alt-ctrl-Fn13:32
robert_ancelltvoss_, what was the problem we had in the system compositor that we turned the input off for? We need to re-enable the input for mlankhorst patch to go through13:33
alf__robert_ancell: ok, if the common code (multimonitor-examples) lands early enough today I will take your branches, rebase them and MP them. Otherwise whoever gets to it first can do it.13:34
tvoss_robert_ancell, it makes the ati tests fail when trying to allocate a bo for the hw cursor13:34
robert_ancelltvoss_, right, so we just need to disable the cursor then13:34
tvoss_robert_ancell, right13:34
hikikoalan_g, I saw it thanks, just a question: it's not possible at all that we release the connection before deleting the platform? I mean13:35
hikikowe are sure that nested mir will never be disconnected before the exit?13:35
alan_ghikiko: we don't need it (yet)13:36
alan_gso don't make anything more complicated to allow it.13:37
robert_ancellracarr, are you awake?13:37
alan_ghikiko: http://en.wikipedia.org/wiki/You_aren't_gonna_need_it13:38
hikikook alan_g one more question:13:39
hikikoif I make mir connection const13:39
hikikoi wont be able to call is valid and get error and release13:40
hikikowhithout cast13:40
hikiko:s/without cast//13:41
alan_ghikiko: why do you think that?13:41
hikikobecause they get char* not const char*13:43
hikikoconst MirConnection* :p13:44
hikikoand MirConnection*13:44
alan_gMirConnection* const connection;13:44
alan_gMake connection const, not what it points to13:45
hikikothis doesn't make it read-only?13:46
hikikoshall i make it mutable?13:46
alan_gWhat for?13:50
hikiko I call mir_connect_sync in the constructor13:50
hikikoconnection = mir_connect_sync13:50
alan_gBut you shouldn't - call it in the initializer list13:51
hikikothen makes sense :D13:51
robert_ancellCan I get a second opinion on https://code.launchpad.net/~robertcarr/unity-system-compositor/disable-cursor/+merge/174006?14:03
alan_grobert_ancell: done14:07
robert_ancellalan_g, ta14:07
robert_ancellalan_g, feel like https://code.launchpad.net/~robert-ancell/unity-system-compositor/vt-switch/+merge/178072 too?14:08
alan_gwhat look at usc twice in one day?14:09
robert_ancella wild Thursday for all involved14:09
alan_grobert_ancell: is there a way to test it without forcing my system to reboot into usc world? (I never do that.)14:13
robert_ancellalan_g, not easily14:17
robert_ancellrather, no, you would have to go into u-s-c world14:17
alan_gI'll have to pass then14:17
=== alan_g is now known as alan_g|tea
=== alan_g|tea is now known as alan_g
kgunn_alf__ hey so that was it....14:52
kgunn_thnaks for the help14:52
alf__kgunn_: no problem, let me know how things go when you connect a second monitor14:53
kgunn_i'll probably demo tomorrow at wrap up....i still need to find an extra mon14:53
kgunn_you bet14:53
kgunn_exciting stuff14:53
alf__kgunn_: It would be great if you could find a monitor today and try it out, so we have more time in case of problems14:54
kgunn_yep...i think i may creep out of the meeting i'm in14:56
=== sfeole` is now known as sfeole
robert_ancellmlankhorst, around?15:47
seb128robert_ancell, he's off today15:52
robert_ancellseb128, ah, thanks15:52
jonorobert_ancell, would you mind responding to the post on mir-devel with the guy trying to get Mir running?16:00
robert_ancelljono, looking...16:00
jonothanks, pal16:01
kgunnracarr: ping16:10
kgunnrobert_ancell: your answer on mir-devel....didn't i just prove you need the mesa bit16:43
robert_ancellkgunn, only if you're running a client right?16:44
robert_ancellwhich is the obvious next step16:44
robert_ancellI'll clarify16:44
kgunncool...was just double checking, it sort of hit me since i just stumbled today16:46
robert_ancellkgunn, this problem will all disappear soon when the patch is in the archive16:46
kgunnrobert_ancell: that's what you say to all the girls16:47
robert_ancellkgunn, that's how you end up with two children16:48
alan_galf__: happy with this version? https://code.launchpad.net/~alan-griffiths/mir/negate-negative-error-codes/+merge/17804116:58
alf__alan_g: yes, approved17:05
alan_galf__: thanks17:06
alan_gHave a great evening17:06
=== alan_g is now known as alan_g|EOD
robert_ancellracarr, around?17:14
om26erhow do I stop surfaceflinger ?17:35
kdub_om26er, 'stop'17:38
om26erkdub_, figured that, though I had to first 'android-chroot'17:39
om26erthe doc probably needs to be updated17:39
kdub_om26er, would be an easy patch to make :) (docs are generated from mir source)17:42
om26erkdub_, sure, branching lp:mir :)17:42
racarrrobotfuel: kgunn: Oi. sorry. family stuff17:47
racarrshould have sent an email17:47
racarrerr robotfuel sorry, tab completion error17:47
racarrcan I top approve?18:16
racarrLet's land this too https://code.launchpad.net/~mterry/mir/session-for-surface/+merge/174406 as per the discussion at end?18:22
kdub_racarr, i suppose, i think we'll have to unwind it though eventually18:24
racarrkdub_: Yes definitely18:25
racarrI feel sort of like it doesn't make our unwinding much harder than it already is though18:25
racarrespescially as long as we don't use it internally and just let it be to unblock mterry18:25
racarrand I feel like we have blocked mterry for too long on something that should be trivial18:26
kdub_racarr, well, switched to abstain18:34
racarrkdub_: Ok. I think it conflicts with some stuff so I will merge it later once I have an idea of what all that is up conflicts18:35
smintheuhi all18:48
smintheuI just saw that a bunch of mir-related packages are in my dist-upgrade18:48
smintheuis there any way to check if i'm running mir now?18:48
kdub_smintheu, look for a unity-system-compositor executable with 'ps ax' is a good way18:50
smintheufrom the documentation i read earlier a 'ps aux | grep unity' should show me that if unity-system-compositor is running, i'm running mir, is that right?18:50
smintheukdub, it isn't18:50
smintheuafter the update, is there anything I should do to 'activate' mir?18:50
racarrCan we try and land death-to-single-visibility today20:23
racarrwould like to land that and no-input-for-hiddne-surfaces20:24
racarrso I can unjam test_client_input.cpp20:24
racarrand begin improving that20:24
racarrTaking off for a while. my entire family is in town for just the day so going to spend some time with them . I will come back for a while in the evening to iterate if anyone has time to do some reviews.22:53

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