/srv/irclogs.ubuntu.com/2013/11/06/#ubuntu-desktop.txt

=== duflu_ is now known as duflu
=== thumper is now known as thumper-afk
=== lifeless_ is now known as lifeless
Mirvmorn04:58
pittiGood morning05:42
seb128good morning desktopers06:43
=== ohama_ is now known as ohama
mlankhorstmorning!07:26
=== thumper-afk is now known as thumper
pittisince today's dist-upgrade unity panel crashes every few mins, bug 124844608:32
ubot2Launchpad bug 1248446 in unity (Ubuntu) "unity-panel-service crashed with SIGSEGV in g_object_unref()" [Undecided,New] https://launchpad.net/bugs/124844608:32
pittithis keeps rearranging windows and screen flicker, so it's quite noticeable08:33
seb128pitti, hey08:34
pittibonjour seb128, ça va ?08:34
seb128pitti, that's an ido-trusty issue, I'm debugging it with tsdgeos on #ubuntu-unity, downgrade libido-3-0 if you need a workaround08:34
seb128pitti, oui, et toi ?08:34
pittiseb128: ah, so known? thanks08:40
pittiseb128: I'm good, thanks!08:40
seb128pitti, "known", you are the second one to ping me about it, tsdgeos is getting a valgrind log08:40
pittiapport traces arrived, too08:41
tsdgeosseb128: i'll add the valgrind log to the bug08:41
seb128tsdgeos, thanks08:41
seb128pitti, retraces matches http://paste.ubuntu.com/6369241/ from tsdgeos08:42
tsdgeosthe valgrind log is not very interesting though08:42
seb128tsdgeos, doesn't tell where the other unref is?08:44
tsdgeosaccording to the trace it'd seem it's not a double delete, just a bad delete :D08:44
tsdgeosbut i haven't done much/any glib "bad deletion" code on valgrind so no idea if this is what one would expect from a double delete or not08:45
pittiyou usually get a rather clear warning if you try to unref an already free'd object08:50
pitti(on stderr)08:50
pittis/warning/critical/08:50
Laneymorning desktopperinos09:05
seb128Laney, hey09:06
seb128larsu, there?09:06
* seb128 wonders wth about this code09:06
larsuseb128: now I am09:20
* larsu was up long last night09:20
seb128larsu, good morning!09:20
seb128larsu, the issue is the g_object_unref (file); in l30809:21
larsuseb128: wait, are we talking about the same thing here as in #ubuntu-unity?09:21
seb128larsu, I knew I should have waited for you to be around, that you fix it easily :p09:21
* seb128 just spent 30min looking to the changes to come to the conclusion you had directly :-(09:21
seb128larsu, yes, we moved there because pitti reported the issue as well09:22
larsuseb128: I wonder how this wasn't a problem during my testing09:23
* seb128 gets the bug as well since today's update09:23
seb128larsu, yeah, me too :/09:23
seb128larsu, it segfaults every time there is a as refresh it seems09:23
seb128I can trigger by calling SetIconFile in dfeet09:23
seb128or just by waiting a bit09:23
larsuah, doesn't happen in the loader…09:26
larsuoh, it does when closing it :)09:26
seb128larsu, you didn't test the changes on your real session? no cookie :p09:26
larsu:'(09:27
seb128larsu, you can still get cookies by get a fix up for review ;-)09:27
* seb128 hugs larsu09:27
larsu\o/09:27
seb128larsu, bug #1248446 if you want a bug reference09:28
ubot2Launchpad bug 1248446 in ido (Ubuntu) "unity-panel-service crashed with SIGSEGV in g_object_unref()" [High,Confirmed] https://launchpad.net/bugs/124844609:28
larsuseb128: thanks. https://code.launchpad.net/~larsu/ido/remove-superfluous-unref/+merge/19408809:31
larsutsdgeos, pitti: sorry for that crash :/ (and thanks for tracking it down)09:32
tsdgeoslarsu: no worries09:32
tsdgeosit took me a while to see it was crashing for how fast it restarts, but the konsole area was shirking/enlarging when switching tabs, and there's been no konsole update so i thought "WHAT?¿?"09:33
seb128larsu, I wish the glib documentation was better about those things, I looked at it earlier but "transfer-none" doesn't speak to me09:33
larsuseb128: you should learn those, they're all over the docs. I agree its not very intuitive, but it's _very_ consistent09:34
larsutransfer-none: ownership is not transferred in any way (no freeing/unreffing)09:34
seb128larsu, I though they were tags for gir, not human documentation :p09:34
larsutransfer-full: free all the things you get back09:34
larsutransfer-container: free the passed back array/list, but not its contents09:35
larsuseb128: they're used for both :)09:35
seb128larsu, thanks for the explanations ;-)09:35
seb128larsu, btw I set the commit message on the mr (and fixed the "trans*f*er-none" typo)09:35
larsutsdgeos: I've long been a proponent of _not_ restarting things that crash so that it's more noticable. I'm very alone with that opinion though09:36
larsuand I can understand the other side of it09:36
larsuseb128: oh, thanks!09:36
tsdgeosyeah09:36
tsdgeoslarsu: for devels it's better the permanent crash, for Joe User it's better the restart09:37
tsdgeosi guess09:37
larsumaybe we should only enable restarting on releases?09:37
seb128larsu, what would we win of not restarting?09:38
seb128that wouldn't give us more infos on the segfault09:38
seb128but it would let those of us who run $unstable without a working desktop09:39
seb128it seems getting the annoyance for no benefit09:39
larsupeople would complain louder09:39
larsuseems like many just click away the crash reporter09:39
larsubut yeah, I totally see your point about a working system09:39
seb128you can't really "click away the reporter", since closing it submits by default (you have to uncheck before closing to don't report)09:40
seb128larsu, the current system didn't stop you to have several users reporting the issue the day the bug landed as well...09:40
larsuha, fair enough09:41
larsubut they still had to revert to a lower version09:41
seb128sil2100, can we get a new ido landing once https://code.launchpad.net/~larsu/ido/remove-superfluous-unref/+merge/194088 is in trunk?09:41
sil2100seb128: I think we can do that, let me keep that on the radar09:42
seb128sil2100, ok, earlier is better, unity-panel-service keeps segfaulting with yesterday's update which is going to be annoying for trusty users09:43
larsulol, "trusty users"09:43
sil2100;)09:44
seb128;-)09:44
sil2100seb128: if I would miss the moment when it's merged in, please poke me if you can09:44
seb128larsu, best release codename ever :p09:44
sil2100And I'll spin stuff09:44
seb128sil2100, ok09:44
seb128sil2100, it's merged09:53
vilamlankhorst: ping regarding bug #124432410:02
ubot2Launchpad bug 1244324 in glamor-egl (Ubuntu) "glamor-egl crashes when running autopilot tests" [High,In progress] https://launchpad.net/bugs/124432410:02
vilamlankhorst: can you get the fix in ubuntu while it's dealt it upstream so we can restore the ci service ?10:03
vila*dealt with10:03
mlankhorstpatience ;)10:04
mlankhorsthttps://launchpad.net/ubuntu/+source/glamor-egl/0.5.1-0ubuntu610:04
vilamlankhorst: wow, magic !10:05
mlankhorstshould autoclose the bug in an hour or so10:05
vilaohhhhh10:05
sil2100seb128: ok, spinning!10:07
seb128sil2100, thanks10:07
vilamlankhorst: hmm, I'm not sure I understand, none of xserver-xorg-glamoregl libglamor0 libglamor-dev in installed on qa-radeon-775010:58
vilamlankhorst: did your fix propagates somewhere else I should look for ?11:00
mlankhorstvila: 7750 is radeonsi right?11:01
mlankhorstxserver-xorg-video-ati pulls it in11:01
vilanot installed either11:02
vilaerr wait11:02
vilawait wait wait11:02
mlankhorst;)11:02
vilaI'm looking at the host not at the container ;) scratch everything I said in the last minutes ;)11:03
mlankhorstI had a feeling11:04
mlankhorstconsidering 7750 wouldn't work otherwise11:05
seb128bah, hate tests11:06
seb128gdk-pixbuf fails to build due to tests, of course it works locally11:07
mlankhorst;)11:07
seb128https://launchpadlibrarian.net/155950592/buildlog_ubuntu-trusty-i386.gdk-pixbuf_2.30.0-0ubuntu1_FAILEDTOBUILD.txt.gz11:07
mlankhorsttried on a i386 chroot?11:07
seb128I should11:07
seb128it looks like they might be running installed-tests are build-time11:08
seb128or at least relying on the package to be installed11:08
mlankhorstseems that way11:09
vilamlankhorst: yeah, xserver-xorg-glamoregl installed in the container11:10
mlankhorstseb128: yeah same failure here in a pbuilder11:10
vilamlankhorst: but I'm hitting a different bug so will need to tweak before confirming your fix ;)11:10
mlankhorstERROR:animation.c:11:test_animation: assertion failed (error == NULL): Couldn't recognize the image file format for file '/tmp/buildd/gdk-pixbuf-2.30.0/tests/test-animation.gif' (gdk-pixbuf-error-quark, 3)11:12
mlankhorstafter i created the file it asked for, but empty11:12
seb128mlankhorst, it seems to be because it looks for the loader in the system location11:12
mlankhorstyeah11:13
seb128I'm pondering just disabling the tests :p11:13
seb128but Laney and pitti are going to look angry at me if I do that11:13
mlankhorstseb128: convert to autopkgtest or figure out how to override it?11:13
mlankhorst:>11:13
* Laney stares11:14
mlankhorstor define failure as another way to succeed. should be laney approved11:14
seb128that's going to teach me to update to new versions when they seem fine, it was mostly new tests there11:15
Laneygo ask mclasen, he fiddled with that stuff11:15
seb128https://bugzilla.gnome.org/show_bug.cgi?id=70301211:16
ubot2Gnome bug 703012 in general "make check fails: Cannot open pixbuf loader module file '.../gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory" [Normal,Unconfirmed]11:16
pittiseb128: that doesn't fail on a local build?11:16
pittiseems strange11:16
pittiit doesn't sound very sensitive to a buildd environment11:16
mlankhorstpitti: not really, it can use the system version to test11:16
seb128pitti, it doesn't fail on my system which has gdk-pixbuf installed11:16
Laneyit fails in a clean chroot11:16
pittioh, you didn't build in sbuild/chroot/pbuilder, I figure11:16
seb128no I didn't11:16
seb128I can't be bother to test build everything in pbuilder :p11:17
seb128(that's what I get in return :/)11:17
pittisbuild == ♥ :)11:17
* seb128 needs a bigger disk than 80G11:17
pittiwell, sbuild + apt-cacher-ng + lots of RAM and building everything in /tmp, that is (/tmp on tmpfs)11:17
=== alan_g is now known as alan_g|afk
mlankhorsthehehhehe11:18
mlankhorstpbuilder-dist + 16 gb ram to build in works for me11:19
LaneyI think you need to set GDK_PIXBUF_MODULEDIR11:20
mlankhorstafter I modded pbuilder to kill the archives after installing deps, before building, using 8 gb for copies of archive was a recipe for disaster11:20
Laneyfor the testsuite11:20
mlankhorstLaney: maybe, but the testsuite should do that upstream11:21
Laney?!?!?!?!11:21
Laneyof course, any fix is going to be an upstream fix11:22
seb128Laney, do you want to have a look at it, since it seems you already have an idea what to do? ;-)11:25
LaneyI don't :P11:25
LaneyI just looked at the manpage for gdk-pixbuf-query-loaders11:25
=== alan_g|afk is now known as alan_g
seb128sil2100, what happened to the ido landing?11:38
sil2100seb128: the indicator stack is running, when it is done we can release11:39
seb128sil2100, ok, that's taking ages it seems11:40
seb128on that note, lunch11:40
seb128bbiab11:40
sil2100seb128: yes, it was like that since all the stacks got ran, and this meant that indicators wasn't allowed to run by itself and was halted by waiting on other dependent stacks to finish11:40
sil2100seb128: so platform had to be built first etc...11:41
sil2100seb128: there was no way of dealing with this without breaking the stack builds that others prepared11:41
seb128sil2100, ok, well as long as it lands soon that's ok11:57
seb128sil2100, it would be nice if the fix was in by the time the US wakes up and start upgrading ... which is soon ... do you know when the stack is going to be ready?11:58
sil2100seb128: sorry about that in overall, hope we'll release it in like 15 minutes or such11:58
seb128ok11:58
=== MacSlow is now known as MacSlow|lunch
Laneyoh12:08
LaneyI got the tests to pass12:08
Laneynow to see where to insert that into the build system12:09
seb128Laney, gdk-pixbuf?12:20
Laneyye12:20
seb128oh, great ;-)12:20
Laneyit was what I said12:20
seb128I was just about to start looking at it, I though said "no" earlier12:20
seb128thanks ;-)12:20
Laneyyeah, I did12:20
Laneybut then I thought "oh, I'll just try it quickly"12:20
seb128;-)12:21
didrockssil2100: please don't land anything that is installed on touch12:21
didrocksseb128: what about the ido landing?12:21
didrocksis there a regression?12:22
seb128didrocks, the landing from yesterday makes unity-panel-service segfaults every few minutes12:22
* sil2100 checks for ido in the touch12:22
didrockssil2100: it is in touch12:22
seb128didrocks, that's fixed in trunk, I really want that to land before U.S wake up/upgrade12:22
didrocksseb128: well, we can as well backout12:22
didrocksI first want to know if ido is what is getting one indicator crazy12:22
seb128didrocks, "backout"?12:22
didrocks(the name is cut)12:22
didrocksrevert to previous ido12:22
seb128shrug12:22
seb128the fix is a one liner12:23
didrocksit seems it wasn't network-manager12:23
seb128dropping an incorrect g_free12:23
sil2100hm, I was really sure that ido had nothing to do with touch12:23
didrocksseb128: yeah, but we have no idea if ido is what making all touch testss failing12:23
seb128didrocks, I fail to see how ido could be used on the phone12:23
didrockssil2100: ok, do you have any other idea?12:23
didrocksseb128: I just know it is in the phone by a dependency, I'm unsure if it's loaded or not12:23
seb128didrocks, that doesn't make any sense to me, I checked the rdepends, no phone component uses it12:23
seb128didrocks, do you have a log/error from a failing test?12:24
seb128didrocks, what components have failing tests?12:24
didrocksseb128: it's not a failing tests, it's one indicator going crazy12:24
didrockstaking all CPU12:24
seb128which one?12:24
didrocksand so a lot of tests failing12:24
didrocksindicator-*12:24
didrocksit's truncated, as said12:25
seb128oh, "fun"12:25
didrocksindeed12:25
didrocksI was hoping for indicator-network12:25
seb128is there any way to untruncate?12:25
didrocksas we got a network-manager update12:25
didrocksseb128: not that I know of12:25
seb128didrocks, what's the diff between not-buggy and buggy?12:25
didrocksseb128: http://people.canonical.com/~ogra/touch-image-stats/20131105.changes12:25
sil2100hmm12:26
didrockspsivaa downgraded all nm component12:26
didrockswe still have this indicator12:26
seb128hum12:26
seb128didrocks, can somebody reproduce the issue on an actual device?12:26
sil2100psivaa: can you try downgrading ido?12:26
didrocksseb128: https://jenkins.qa.ubuntu.com/job/trusty-touch-maguro-smoke-ubuntu-rssreader-app-autopilot/9/artifact/clientlogs/top_before.log/*view*/12:27
didrockssil2100: already what I asked 10 minutes ago12:27
didrocksseb128: see different top run (for 2 minutes)12:27
didrocksseb128: we are trying on the test env directly now12:27
didrockseasier to rerun the tests12:27
seb128didrocks, if an actual human could reproduce on a device where we can adb to that would be useful12:27
psivaasil2100: downgraded to 13.10.0+13.10.20131031-0ubuntu1 and running a test now12:27
sil2100psivaa: thanks!12:28
seb128didrocks, psivaa, sil2100: do you have an uptodate device? what is apt wanting to do if you apt-get remove libido3-0.1-012:28
seb128?12:28
didrockssystem settled. SUCCESS12:28
* seb128 has saucy still there, needs to upgrade12:28
didrocksseb128: no, I don't have it handy, trying to get back production working first12:29
seb128didrocks, what does "system settled. SUCCESS" means?12:29
didrocksseb128: that the system was "quiet enough"12:29
seb128didrocks, e.g that the ido downgrade fixes the issue?12:30
didrockswhich should indicate that the indicator wasn't at that CPU % usage12:30
didrocksmaybe, I want multiple runs before affirming it's the case12:30
didrocksseb128: seems to indicate ido is guilty: http://10.97.0.1:8080/job/trusty-touch-maguro-smoke-ubuntu-calculator-app-autopilot/12/artifact/clientlogs/top_before.log12:33
didrockssee the last top run12:33
didrocksno indicator*12:33
didrockspsivaa: running others?12:34
seb128didrocks, so it's unity-services which brings ido on the image12:35
didrocksseb128: but unity-services isn't started, right?12:35
didrockson the device12:35
seb128didrocks, well, if it's not a fail to see how ido can impact on indicators :/12:35
didrocksseb128: let's see with other runs, but at least, if we identify that one as being the cause… there is clearly something to look into12:36
seb128didrocks, can you see if unity-panel-service is running on that test machine?12:36
didrockspsivaa: you are the one connected ^12:36
seb128sil2100, meanwhile please publish the new ido12:36
seb128that can't hurt12:37
seb128if we need to do a revert we can still do it12:37
seb128but at least we don't stay on a buggy version while deciding on the revert12:37
didrockssil2100: no, please don't publish it12:37
seb128grrrr12:37
didrocksI would rather go with a backout versoin12:37
* seb128 manual upload12:37
didrocksseb128: come on, you screw us12:37
seb128didrocks, and you screw us12:37
didrocksso please try to be cooperative12:37
seb128difference is that desktop is in production12:37
didrocksseb128: I wasn't the one pushing a broken version to the archive12:37
seb128where your image doesn't reach users12:38
didrocksseb128: yeah, so let's rollback12:38
didrocksI'm happy to handle that rollback12:38
seb128didrocks, that version is not broken afaik, but it seems unity8 is doing something weird trying to load gtk components12:38
didrocksseb128: well, we don't know, you thought yesterday's version wasn't broken12:38
didrocksand it segfaults12:39
seb128you better fix whatever tries to load a gtk library12:39
didrocksso I'm trying the safest path for now12:39
seb128didrocks, still, something on the device is doing something stupid12:39
seb128didrocks, and you better get at the bottom of it12:39
didrocksseb128: I don't say the contrary12:39
didrocksseb128: *but* we are doing a complex 1.4 AP transition12:39
didrocksI still don't have a status on the 1.3 last release because of this12:39
didrocksso let's get some other test runs12:40
didrocks(and no, no unity-panel-service is running now on my device with a recent image)12:40
seb128didrocks, I guess no indicator is using 100% cpu either?12:40
didrocksseb128: I don't have latest12:41
didrocksseb128: sorry, but keep being interrupted, even no coffee from this morning12:41
didrocksmostly because of that issue12:41
seb128didrocks, stop starting IRC before coffee!12:41
didrocksthat's not related to it12:42
* seb128 installs new ido on his saucy12:42
didrockspsivaa: around at all?12:43
psivaaseb128: didrocks: so with nm downgraded (maguro) i still see some systemsettle failures. but with ido downgraded (mako) i see some improvements12:43
psivaadidrocks: yes, the tests take time to run :)12:43
didrockspsivaa: how many tries with ido downgraded?12:43
didrocksjust one or more?12:43
seb128psivaa, "some improvements" or "it's fixed"?12:43
psivaadidrocks: seb128: one test, still running12:44
seb128psivaa, can you "grep ido /proc/*/maps"12:44
seb128and see if that returns something?12:44
didrocks(flashed latest image and of course, can't reproduce…)12:45
* sil2100 is confused12:45
seb128didrocks, ok, so you want to revert ido in trusty now?12:45
didrocksseb128: I wish we can wait for 30 minutes to get more results12:45
sil2100I'll just get back to fixing AP issues maybe... ;)12:45
* sil2100 walks away silently12:45
seb128didrocks, well, if we wait for 30 min we can land the segfault fix, so it's in, then we can revert12:45
seb128if needed12:45
* seb128 doesn't get what blocking the 1 liner segfault fix brings us12:46
didrocksseb128: as you wish, but if we see that ido is really the faulty one, I don't want that we wait for testing the ido fix, agreed?12:46
psivaaseb128: i ran that command on both devices and both return nothing12:46
seb128didrocks, sure12:46
didrockssil2100: ok, please release this ido ^ know that we maybe we'll revert all ido soon though12:47
seb128didrocks, I've installed the new ido on my device (saucy) and it creates no issue12:48
didrockspsivaa: what do we have on mako?12:48
psivaadidrocks: downgraded ido12:49
didrocksdowngraded network-manager as well?12:49
psivaadidrocks: no12:49
didrocksok, so calendar-app seems to show that ido is really guilty12:49
didrocksreally really weird…12:49
seb128didrocks, calendar-app?! wth...12:50
didrocksseb128: the test of calendar-app (systemsettle one)12:50
didrocksseb128: basically ALL tests were screwed because of this indicator12:50
didrocksseb128: http://reports.qa.ubuntu.com/smokeng/trusty/touch/12:51
seb128didrocks, calendar-app is not an indicator though12:51
didrockshttp://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/12:20131105:20131031.1/4903/12:51
didrocksseb128: yeah, but if an indicator is using all the CPU…12:51
didrocksthe tests will fail12:51
seb128ok12:51
didrocks67% of mem for an indicaotr12:51
seb128how do we go about debugging the issue next?12:52
didrocksseb128: so, if it's not ido as it seems it can't, if it's not netowkr-manager, what can it be?12:52
seb128can we figure out what loads ido12:52
seb128and what indicator uses the cpu?12:52
didrocksseb128: I guess that's more a question for upstream to figure it out12:53
seb128didrocks, we had an unity8 landing, did we rul that one to be buggy?12:53
didrocksright now, I'm trying to ensure we find the right guilty12:53
seb128didrocks, or are we sure than downgrading ido is workaround it?12:53
didrocksseb128: it can be unity8, indeed12:53
didrocksseb128: what I'm telling, I'm not ensured it's ido right now12:53
seb128didrocks, also, whoever "downgraded nm", did they include the libs or just the network-manager binary?12:53
didrocksseb128: psivaa told me he dowgraded all binaries from nm source12:54
seb128ok12:54
didrocksif only we knew which indicator was at fault…12:54
seb128where is the script doing the top? can we tweak it?12:55
didrocksseb128: we can tweak it, do you know what's the parameter for top to not truncate?12:55
didrocksit's in utah test suite12:55
sil2100seb128, didrocks: ido publishing12:55
psivaadidrocks: seb128: http://pastebin.ubuntu.com/6370333/ is the nm related packages in the nm downgraded device12:56
didrockssounds like the whole selection12:56
didrockssystem settled. SUCCESS12:56
didrockswith friends tests12:56
didrockson mako12:56
didrocks(with thus only ido downgraded)12:56
psivaadidrocks: yes12:57
didrocksso 3 on 3 runs?12:57
=== alan_g is now known as alan_g|lunch
sil2100seb128: ido should be in soon12:59
didrockslocal run on a really fresh install -> nothing13:00
* didrocks wonders why he's the only one caring about this13:02
didrocksok, giving up…13:02
seb128didrocks, top -b -n1 -c  > log13:02
seb128didrocks, that gives non truncated commands13:02
seb128didrocks, sorry was googling for the top truncating13:03
=== Rodge is now known as Rodge_
seb128Laney, thanks for the gdk-pixbuf fix!13:05
psivaadidrocks: terminal app tests, systemsettle success: http://10.97.0.1:8080/job/trusty-touch-mako-smoke-ubuntu-terminal-app-autopilot/15/console13:05
psivaawhich earlier failed with the latest ido13:05
Laneyseb128: np!13:05
Laneydon't really understand these new test harnesses13:06
Laneyso I have trusty/13 on my mako13:06
didrockspsivaa: ok, continue running mako with this ido downgraded, let's try to reflash maguro13:06
Laneyhow do I get this bug?13:06
psivaadidrocks: do you want me to get maguro with the latest image and downgrade ido as well?13:07
didrocksLaney: I don't get it as well, locally13:07
didrockspsivaa: no, just get latest image on maguro, then, let me give you a branch so that we know which indicator is failing13:07
psivaadidrocks: ack13:08
didrockspsivaa: can you apply that to the systemsettle script branch? http://paste.ubuntu.com/6370377/13:08
didrockshopefully, the command name won't be truncated anymore13:08
didrocksso that we see what indicator is guilty13:09
didrocks(so just run one test)13:09
didrocksthen, let's downgrade ido13:09
psivaadidrocks: ack13:09
didrocksand retry all tests on maguro as well13:09
Laneyare there any crash files?13:11
didrocksLaney: none from the dashboard. For instance: http://reports.qa.ubuntu.com/smokeng/trusty/touch/maguro/12:20131105:20131031.1/4905/ubuntu-terminal-app-autopilot/13:12
Laneymmm13:12
didrocksseb128: I don't htink that will work through adb though13:15
seb128didrocks, the top?13:15
didrocksyep13:16
didrocksbecause of adb TERMCAP13:16
didrocksjust try adb shell13:16
didrocksand your command13:16
didrocks-> see top being truncated13:16
seb128didrocks, COLUMNS=900 top  -b -n1 -c  > log13:17
seb128that works13:17
didrockspsivaa: mind prepending this COLUMNS=900 as well?13:18
didrocksseb128: nice catch13:18
psivaadidrocks: ack will do. the flashing ongoing btw for maguro13:18
Laneybtw:13:18
Laneyphablet@ubuntu-phablet:~$ aptitude why libido3-0.1-013:19
Laneyi   ubuntu-touch        Depends unity813:19
Laneyi A unity8              Depends unity8-private (= 7.83+14.04.20131105.1-0ubuntu1)13:19
Laneyi A unity8-private      Depends libunity-core-6.0-8 (>= 7.0.0)13:19
Laneyi A libunity-core-6.0-8 Depends unity-services (= 7.1.2+13.10.20131014.1-0ubuntu1)13:19
Laneyi A unity-services      Depends libido3-0.1-0 (>= 13.10.0daily13.06.19)13:19
didrocksahah, so unity8 is using it13:19
didrocksyeah but13:19
didrockslibunity-core-6.0-8 doesn't start unity-services13:19
Laneyit's probably unused13:19
seb128it's not13:19
didrocks(u-p-s)13:19
Laneybut that's why it is there13:19
seb128right13:19
seb128libunity-core-6.0-8 depends on it (is that buggy?)13:19
didrocksif it's optional now, yeah, it is13:20
didrocksit was required before13:20
seb128because of a schemas or something?13:20
didrocksbut still, that doesn't explain if downgrading ido reliably "fix"13:20
Laneyit has the upstart job13:20
Laneyto start ups13:20
didrocksLaney: but not on touch, right?13:20
Laneyoh and ups itself13:20
=== Rodge_ is now known as recoveryRodge
seb128but yeah, that ido stuff doesn't make sense to me13:20
Laneyno13:20
Laneydon't know if it needs to be a depends of the library or could be from unity itself or so13:21
didrocksI still don't know why I don't see that crazy indicators here13:21
didrocksthe only thing I don't have is messages and sim card13:21
didrocksmaybe popey can see the issue13:22
popeyhm?13:22
didrockspopey: with latest image, if you adb shell just after you boot your phone13:22
didrocksand run COLUMNS=900 top  -b -n1 -c13:22
popeyimmediately after boot?13:22
didrocksdo you see some indicators being at the top, like taking most of the CPU (for the first 2 minutes13:23
didrocks)13:23
didrocksyeah13:23
popeylets see13:23
didrocksand then 2 minutes after boot13:23
didrocksthanks!13:23
popeyno13:24
popeyi see hud-service initially13:24
didrocksand then everything's idled?13:24
popeynothing especially eating cpu13:24
popeyaside from top13:24
didrockshum…13:24
didrockswhy? it's so reliable on the DC…13:25
=== MacSlow|lunch is now known as MacSlow
popeycurrent build number: 1313:25
popeydevice name: mako13:25
popeythat the right one?13:25
ogra_we probably just need to send popey over and it will work magically13:25
didrocksyeah13:25
popeyooh13:25
didrocksogra_: well, can't get it here as well :/13:25
didrocksLaney neither it seems :/13:25
didrocksI was hoping it was sim-card influenced13:26
ogra_weird13:26
popeytop - 13:26:23 up 3 min,  1 user,  load average: 1.70, 1.34, 0.5813:26
didrocksyeah, sounds reasonable13:26
didrockspsivaa: do you have other mako results while maguro is flashing?13:26
didrockspsivaa: like, we do have 0 system settle idling issue?13:27
popeylots of wifi networks in the DC?13:27
* ogra_ doesnt fine a load above 1 reasonable ... 13:27
popeyor indeed, none?13:27
ogra_but thats another topic :)13:27
popeythats mostly "top" ogra_13:27
popey 2969 root      20   0    2540   1096    708 R  2.3  0.1   0:01.26 top13:27
psivaadidrocks: yes terminal app is finished with success13:27
ogra_popey, top cant produce a load above 1 ... thats something else13:27
didrockspopey: yeah, maybe…13:27
popeywell, top is more than everything else13:28
ogra_popey, try on a decent PC ...13:28
popeythere is other stuff, but top is the largest proprtion13:28
didrocksogra_: TBH, I would appreciate that we all focus the effort on fixing the regression first13:28
ogra_load should always be under 1 unless there is something actally demanding going on13:28
popeyindicator-power wakes up now and then13:28
Laneyqengho: looks like the chromium tests fail newly now: https://jenkins.qa.ubuntu.com/view/Trusty/view/AutoPkgTest/job/trusty-adt-chromium-browser/11/ARCH=amd64,label=adt/console13:28
popeyhttp://paste.ubuntu.com/6370456/13:29
popeythose are all eating a little bit all the time13:29
didrockspsivaa: so you continue on running everything manually on mako?13:30
didrocks(all the one which had system settle idling issue)13:30
ogra_didrocks, agreed, but what if my assumption is that top gets us wrong data ;)13:30
psivaadidrocks: yes, ubuntuuitoolkit is running and addressbook is the next13:30
didrockspsivaa: I'm so puzzled…13:30
psivaadidrocks: ?13:31
didrockspsivaa: like, we have really nothing using ido13:31
didrocksand you just downgraded ido on mako13:31
didrocksand we see no failure…13:31
psivaadidrocks: right, these are the nm related packages in mako: http://pastebin.ubuntu.com/6370341/13:31
didrocksyeah, so latest of latest…13:32
psivaadidrocks: yes13:32
psivaaso afaik only change is downgraded ido13:32
didrocksLaney: seb128: any idea? ^13:33
seb128can you upgrade it again and get the non truncated top output?13:33
didrocksseb128: that's why we reprovision maguro13:33
didrocksthen I guess psivaa will run one13:34
didrocksthe first test should get the failure13:34
seb128didrocks, not really, as said I can't make any sense of the situation, I fail to see why ido would be anything else than a file on disk13:34
didrockswith the non truncated top output13:34
seb128e.g the only thing linking to it is unity-panel-service13:34
didrockssystem settled. SUCCESS13:34
didrocksfor ubuntuuitoolkit13:34
seb128that's not used on the device13:34
didrocksseb128: before, most of the tests (like > 60% of them) were failing13:35
didrockson image 12 and 1313:35
didrockson both devices13:35
seb128didrocks, image 12 had the new ido already?13:35
didrocksyeah13:35
seb128k13:35
ogra_indicator-sound and  indicator-datetime are in libdio3's rdeps13:35
didrocksimage 12 is the diff I pasted13:35
seb128ogra_, ?13:35
ogra_ogra@anubis:~$ apt-cache rdepends libido3-0.1-0|grep indicator13:35
ogra_  indicator-sound13:35
ogra_  indicator-sound13:35
ogra_  indicator-datetime13:35
seb128ogra_, is that in saucy?13:36
seb128ogra_, or do you run that on a box with a raring source?13:36
ogra_oh, sorry, thats precise even13:36
jpdslarsu: 'sup.13:36
seb128ogra_, that was before the convergence codebase13:36
larsuseb128: do you know the state of desrt's app info stuff?13:36
seb128larsu, no13:36
larsuthe branch looks pretty complete to me...13:36
larsuokay, I'll just wait for him to wake up then13:36
larsuhi jpds!13:36
=== alan_g|lunch is now known as alan_g
didrockspsivaa: so trusty-touch-maguro-smoke-ubuntu-clock-app-autopilot failed to settle right?13:57
psivaadidrocks: yes13:57
didrocksbut I don't see anything in top_before…13:57
didrockspsivaa: do you see anything here? ^14:00
psivaayea, i see indicator-datetime-service once no other related to indicator14:01
didrocksbut nothing reliably on top14:01
didrockson top of top14:01
didrocksas we had14:01
psivaayea, i'll run once more14:02
didrocksasac: I'm completely out of ideas ^14:02
didrocks4 non stop hours I'm struggling with this…14:03
didrockssystem is settled now14:05
didrocksok psivaa: let's try to increase the timeout14:05
didrocksto 5 minutes14:05
didrockstrying to get one run14:05
didrocksthen, we'll look later on the regression14:05
didrocksI'm completely out of ideas14:05
didrockswe have way more hectic transition to handle with14:06
didrocksso, let's hope that we "only" take longer to settle14:06
didrocksand that's not reliable14:06
didrockspsivaa: making sense?14:06
psivaadidrocks: yea, we could try that14:06
didrockspsivaa: so, you do continue with a longer timeout to run all mako + all maguro which system settled fail?14:08
didrocksso that we can have run 13 with real results14:08
didrockson both14:09
didrocksbefore AP 1.4 transition14:09
psivaadidrocks: ack14:09
psivaadidrocks: im changing the timeout from 120 -> 30014:09
didrockspsivaa: thanks, keep me posted in your rerun if it's not enough and we'll see14:09
psivaadidrocks: one more thing, do you want me to upgrade ido to the latest in mako when running with increased timeout?14:10
didrockspsivaa: for the remaining one? yeah, why not14:11
didrocks(don't rerun those for which system settle passed)14:11
psivaadidrocks: ok, i've run most of the tests in mako that failed systemsettle_before with older ido.. ill run them if there is anything left14:12
didrocksok14:13
didrockspsivaa: so, friends couldn't system settle on mako, this is just after the ido upgrade + timeout?14:27
psivaano, this was with downgraded ido and shorter timeout. the one that's running now is with the increased timeout,14:28
psivaaforgot to upgrade ido. sigh...sorry, will do that after this run14:28
didrockspsivaa: ok, you kept the "longer top command right"?14:28
psivaadidrocks: yes14:29
didrocksgreat, let's hope :)14:29
didrocksseb128: so, that seems to confirm it's really an annoying-unknown-issue14:30
didrocksand not ido related14:30
didrocksnot sure why we had ido downgraded passing test reliably for a while, like multiple times in a row14:30
seb128didrocks, yeah, that's quite weird indeed14:30
didrocksSaviq: did you make any change in unity8 that can make an indicator taking > 60% of CPU at load time?14:31
didrocksboot*14:31
Saviqdednick, ↑?14:33
dednickSaviq: :/ not that I know of14:36
dednickdidrocks: an indicator, you mean the backend14:37
didrocksseb128: so, things we have better top log, I can't find a case where there is an indicator at the top14:38
didrockssince*14:38
seb128:-(14:38
seb128red herring?14:38
didrockswell, depressing, annoying, bothering… whatever word you want to put…14:39
psivaadidrocks: http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/13:20131105.1:20131031.1/4910/friends-app-autopilot/ is with increased timeout and old ido14:44
* didrocks sigh14:44
didrocksso at least: 2151 phablet 20 0 212204 177852 3848 R 69.5 9.3 2:45.98 /usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service14:44
didrockspsivaa: but with downgraded nm, we got it, right?14:45
dednickdidrocks: ubuntu-settings-components landing. Says needs upstream feedback. ?14:45
didrocksdednick: yeah, about why is it needed, what will use it14:46
psivaadidrocks: at the moment we dont have anything running with downgraded nm, but we still got systemsettle issues with downgraded nm14:46
didrockspsivaa: and on those systemsettle issues, it was an indicator at the top?14:46
dednickdidrocks: sharing qml components between ubuntu-system-settings and unity814:47
didrocksdednick: can you write that on the spreadsheet please? (or someone to add that)14:47
dednickdidrocks: ok14:47
didrocksthanks!14:47
didrockspsivaa: I wonder if we are not mixing 2 issues14:47
didrocksone which can be the nm one14:48
didrocksand another one which happens way less often14:48
psivaadidrocks: quite possible, let me look at some of the runs and see if we could see a pattern14:48
didrockspsivaa: meanwhile, can you retry downgrading nm14:49
didrocksand rerun the one with system settle idle failing?14:49
didrockscyphermox: hey, around?14:52
psivaadidrocks: so now, on maguro with all the latest nm and ido but with increased systemsettle timeout, we see the systemsettle_before pass now.14:53
psivaahttp://reports.qa.ubuntu.com/smokeng/trusty/touch/maguro/13:20131105.1:20131031.1/4909/14:53
psivaadidrocks: which device you want me to downgrade nm?14:53
didrockspsivaa: the mako one (the one which failed friends)14:53
cyphermoxdidrocks: what's up14:53
didrockspsivaa: on the maguro, this is the status with no systemsettle issue?14:53
didrockscyphermox: we have indicator-network going crazy sometimes in the lab just after boot14:54
didrockscyphermox: see those top (taking at a some intervals): https://jenkins.qa.ubuntu.com/job/trusty-touch-mako-smoke-friends-app-autopilot/24/artifact/clientlogs/top_before.log/*view*/14:54
psivaadidrocks: yea still need to run some tests on maguro14:54
didrockspsivaa: ok ;)14:54
didrockspsivaa: if you can try to downgrade nm here as well14:54
didrocksthe more info we have…14:54
didrockscyphermox: do you think you nm upload can be linked to that?14:55
cyphermoxno14:55
didrockscyphermox: do you have any idea of what can be the cause then?14:55
didrocksit's really happening like 50-60% of each boot14:56
cyphermoxno idea14:56
didrockscyphermox: can you help us on that?14:56
cyphermoxyou said indicator-network? I don't see it in that listing14:56
didrockscyphermox: here is the diff list: http://people.canonical.com/~ogra/touch-image-stats/20131105.changes14:56
didrockscyphermox: in top you mean?14:56
didrocksor in the diff?14:56
cyphermoxin top14:56
didrocks2151 phablet 20 0 212204 177852 3848 R 69.5 9.3 2:45.98 /usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service14:57
didrockslook at the last runs14:57
didrocks(5 minutes after the device booted)14:57
cyphermoxoh, I see, it's more than one top14:57
didrocksyeah, it's regularly top snapshot14:58
didrocksthe last one is 5 minutes after the first one14:58
maxiaojunoj, you are discussing touch in #ubuntu-desktop14:58
maxiaojunoh14:58
didrocksmaxiaojun: depends on where people are :)14:58
didrocksnm and indicators are also desktopish14:58
didrockscyphermox: indicator-network is rocking between 63 and 70%14:58
didrockscyphermox: from the diff between images, apart from nm, I don't see what can cause this…14:59
cyphermoxunity8\14:59
cyphermoxlxc-config-android could, depending on what changed there...14:59
didrockscyphermox: Saviq is telling the same thing than you for nm: they don't think they are impacting it14:59
cyphermoxand regardless, the bug is in indicator-network, not in NM15:00
=== tkamppeter_ is now known as tkamppeter
cyphermoxI'm not denying it's possible, I'll look at it15:00
didrockscyphermox: right, but right now, I'm trying first to get us to a state where we can test things15:00
didrockscyphermox: and we are totally blocked on that one15:00
cyphermoxhow is that?15:00
didrockscyphermox: because a lot of tests are failing because of this CPU usage15:01
didrockscyphermox: so we even don't know where we are test-wise15:01
didrockswhich will make regression from AP 1.4 harder to read15:01
didrockspsivaa: I found the log with downgraded nm: http://10.97.0.1:8080/job/trusty-touch-maguro-smoke-ubuntu-system-settings-online-accounts-autopilot/11/artifact/clientlogs/top_before.log/15:05
didrocksso if indicator-* is indicator-network15:05
psivaathat's more confusing because that was with the downgraded nm15:08
didrocksright15:09
psivaadidrocks: straight away a contradiction: on mako, with downgraded nm, systemsettle success :(((((15:10
psivaahttp://10.97.0.1:8080/job/trusty-touch-mako-smoke-friends-app-autopilot/25/console15:10
didrockspsivaa: ok, can you try to push more and more mako tests15:11
didrockseven those which succeed15:11
psivaadidrocks: ack will do.15:11
didrocksif we see no issue, we can say it's nm15:13
psivaadidrocks: rssreader on mako with downgraded nm has systemsettle_before failure15:23
psivaahttp://10.97.0.1:8080/user/psivaa/my-views/view/trusty_mako_smoke/job/trusty-touch-mako-smoke-ubuntu-rssreader-app-autopilot/14/artifact/clientlogs/top_before.log15:23
didrockspsivaa: ok, can you try unity8 itself?15:24
didrocks(there are multiple binaries)15:24
didrocksit's the last shot :/15:24
psivaadidrocks: you want me to downgrade unity8 or run the unity8 test without downgrading unity815:26
didrockspsivaa: downgrading the whole unity8 source15:27
didrocksand related binaries15:27
psivaadidrocks: ack, will do15:27
psivaadidrocks: ok, i'll replace the current unity8 with the one before, 7.83+14.04.20131031-0ubuntu1 but not sure if that had any issues15:30
didrockspsivaa: it didn't in the previous run15:33
psivaadidrocks: ack, running tests with older unity815:34
=== alan_g is now known as alan_g|tea
seb128cyphermox, do you still plan to look at those wpasupplicant issues that are ranked high on the saucy e.u.c list?15:44
seb128happyaron, that ibus abrt, can you at least upstream it with the stacktrace? that might be enough for them to have a clue?15:44
seb128happyaron, https://errors.ubuntu.com/problem/7d83460403cc3ff966c466f15d32aecec048c0dc is ranking high on the saucy issues15:45
happyaronok15:46
seb128happyaron, thanks15:46
psivaadidrocks: clock_app and rss reader tests have failed systemsettle before even with older unity8.15:50
didrocks*shrugh*15:50
=== alan_g|tea is now known as alan_g
didrockspsivaa: still indicator-network on top?15:52
psivaadidrocks: yes, for both app tests15:54
psivaadidrocks: http://pastebin.ubuntu.com/6371157/ are the relevant packages just to confirm15:54
didrockspsivaa: you reverted unity8-private as well, right?15:54
psivaadidrocks: yes, http://pastebin.ubuntu.com/6371172/15:55
psivaaand unity8-fake-env:armhf  as well15:55
didrocksok… all libnm* as well15:56
psivaadidrocks: yep, http://pastebin.ubuntu.com/6371182/15:57
didrockspsivaa: can you rm /etc/init/android-usb-state.conf ?15:57
didrocksand retry?15:57
psivaadidrocks: running clock with that removed15:58
didrocksogra_: you told that the job is not wired, but do you think that we can have indicator-network reacting to it? ^15:59
didrocks(it's started by mtp)15:59
ogra_nah16:01
ogra_it would only even do something on cable disconnect/reconnect16:01
ogra_didrocks, i think the probalm is on a different levelö (as i told you a few hours ago)16:02
ogra_*problem16:02
didrocksogra_: what level?16:06
ogra_kernel ...16:06
ogra_or rather cpufreq in the kernel16:06
psivaadidrocks: so the clock app test failed on syssettle_before again with /etc/init/android-usb-state.conf removed16:08
psivaaindicator-network-service is at the top16:08
psivaarss app test is running to confirm16:09
dednickdidrocks: comment has been added to landing pipeline doc.16:11
didrocksogra_: did we got kernel change?16:17
ogra_didrocks, no, but we have a cpufreq governor (ondemand) set that makes the device always operate slightly on the edge (see the load values in the top output) ... so it can easily be that the scaling happens to slow and we just hit a threshold ...16:18
Laney"will be used by […] ubuntu-system-settings" - what is this?16:18
ogra_moving to the interactive governor (which i'm just testing here) actually makes me see the load way below 1 (0.2 or some such on idle) and processes seemigly dropping down faster in CPU usage16:19
seb128Laney, shared UI components for indicators/unity/settings16:20
seb128Laney, e.g calendar widget16:20
ogra_didrocks, if your timeout change improved the test, thats definitely an indicator for the slow app reaction i'm seeing here too16:20
psivaadidrocks: rss also fails with /etc/init/android-usb-state.conf removed. let me know if/when i could revert the changes to the tests to increase the timeout16:22
didrockshum, I've no idea16:22
didrocksogra_: no, it didn't improve16:22
didrocksogra_: the indicator is still crazy after 5 minutes16:22
didrockspsivaa: are you connected on the device?16:22
psivaadidrocks: yes16:22
didrockscan you try get a backtrace to see what the indicator is doing?16:23
ogra_didrocks, ok, then my issue is separate16:23
didrockspsivaa: like, right now, the indicator is still skyrocketing, right?16:23
Laneyseb128: seems like they have a left checkbox and a rewritten timezone displayer thingy in there16:23
Laneyis this coming from design?16:23
ogra_(teh device feels a *lot* more responsive with interactive though)16:23
ogra_didrocks, do the devices in the lab have SIMs ?16:24
didrocksogra_: I guess they do have some, psivaa?16:24
ogra_didrocks, i wonder if there are issues with them since you say you can reproduce it anywhere else16:25
psivaadidrocks: i think they have. let me confirm16:25
seb128Laney, check with dednick16:25
ogra_would be intresting to see what happens if someone removes the SIM and we test again16:25
psivaadidrocks: indicator is still coming on the top16:25
didrockspsivaa: ok, so installing debug symbols and gdb into the running process?16:25
seb128Laney, I think they had the checkbox on the other side like us and they moved back to the standard UI, maybe outdated description/code somewhere?16:25
Laneylooking in the example file16:26
psivaadidrocks: ok, this is new to me. let me try that16:26
didrockspsivaa: do you need help?16:26
seb128Laney, I didn't follow the details, I just know what's the project is about16:26
didrocksseb128: would is the right person to help debugging indicator-network?16:27
seb128Laney, we discussed it previous cycle when we were talking about custom widgets in indicators/settings16:27
seb128didrocks, tedg16:27
Laneyokay, weird, I'd have contributed / tried to use it earlier if I knew about it16:27
seb128didrocks, or Wellmark16:27
seb128didrocks, or cyphermox16:27
seb128Laney, well, it was not really a thing yet, it just starts shaping I think16:28
psivaadidrocks: yea i need help in that16:28
seb128Laney, e.g everybody start with "get stuff somewhat working for their component"16:28
seb128started*16:28
didrockstedg: cyphermox: one of you around to help?16:28
cyphermoxyes16:28
tedgdidrocks, ^16:28
tedg;-)16:28
didrockspsivaa: can you give cyphermox ssh access?16:28
didrockshe has access to the lab16:29
cyphermoxtedg: I think you really should look at it though16:29
didrocksI just want someone who can install the dbgsym with psivaa16:29
didrocksand gdb attach16:29
didrocksbt full16:29
tedgI'm confused on what's happening.16:30
cyphermoxrocket science.16:30
* tedg looks for bryceh16:30
ogra_lol16:30
ogra_pass the bucket ...16:30
didrockstedg: so, we have indicator-network taking > 60% of CPU16:30
tedgHmm, that's odd.  Look at dbus and see if there's a bunch of signals coming to it?16:31
tedgIt doesn't really do any processing.16:31
didrockstedg: can you ssh to the phone with psivaa's help and debug live?16:31
didrockstedg: FYI, the only diff we have is: http://people.canonical.com/~ogra/touch-image-stats/20131105.changes16:32
didrockstedg: we tried to backout ido, network-manager, lxc-android-config and unity816:32
* ogra_ is wondering if the SIM cards are run out of credit or some such and spam NM with error messages16:32
didrocksogra_: ohhhh16:32
didrocksrunning out of credit16:32
didrocksmaybe…16:32
ogra_didrocks, thats what i wanted to say before :)16:32
tedgHmm, network-manager from 0.9.8.0-0ubuntu22 to 0.9.8.4-0ubuntu3, cyphermox might not be out of the woods yet ;-)16:32
didrocksogra_: well, you told "with SIM card", that's why I asked you and popey :)16:32
* ogra_ is just dsitracted being in a meeting16:32
cyphermoxmeh16:32
didrockstedg: we tried to downgrade16:33
cyphermoxdidrocks: did libnm* also get downgraded?16:33
didrockswe still have this indicator-network going crazy after boot more than 30% of the times16:33
didrockscyphermox: yep16:33
cyphermoxalright16:33
cyphermoxthen yeah, have fun tedg ;)16:33
cyphermoxj/k :)16:33
cyphermoxpsivaa: how can I get there?16:33
cyphermoxin the meantime my phone was just reflashed with trusty-proposed, I'm hoping to reproduce here16:34
didrockscyphermox: we were 5 trying to reproduce…16:34
didrocksnobody was able to16:34
didrocks(tried to start 10 times)16:34
cyphermoxah16:34
didrocksso it's just, after boot indicator-network is using > 60% CPU16:34
cyphermoxand what's the difference between our own phones and what's in the lab then?16:34
didrocksand never calm down16:34
didrocks(we get that I would say 30% of the time in the lab)16:35
tedgIs there anything in the log file?16:35
didrocksapart from SIM card…16:35
cyphermoxotherwise we're testing something different, and we can't have that16:35
cyphermoxSIM card shouldn't affect the indicator directly in any way16:35
didrockspsivaa: can you pastebin the indicator-network logs16:35
tedgcyphermox, It effects us when we read the oFono dbus interface to show status, but yes.16:35
didrockspsivaa: should be ~/.cache/upstart/indicator-network.log16:36
cyphermoxtedg: well, you need to do that regardless -- that's what I mean by it shouldn't make a difference what kind of brand the SIM is for :D16:36
didrockstedg: the question is "what changed?" from what we tried to downgrade, I don't see any other potential culpurit…16:36
cyphermoxI could understand if it was breaking here for me, with my weird provider that uses UTF8 in some names16:36
didrockscyphermox: let's hope you can reproduce :)16:36
cyphermoxseems to me like it's going to be hard to reproduce...16:37
tedgdidrocks, I think the only other possibility would be unity8 itself, it does talk to the indicator, but it'd be weird to cause that kind of traffic.16:37
cyphermoxlxc-android-config seems to have changed16:37
didrockstedg: as told, we reverted it16:38
psivaadidrocks: http://pastebin.ubuntu.com/6371408/ is the log16:38
cyphermox(ie. I don't have access to /var/lib/dpkg even though I rebooted to a writable image)16:38
psivaawell the last bit, let me retrieve the whole log16:38
didrockstedg: seems spammy? ^16:38
cyphermoxit kind of is, but that shouldn't break things16:38
didrockscyphermox: lxc-android-config reverted and tested as well16:38
cyphermoxcould the fact that phonesim is there extra could be breaking things?16:39
tedgdidrocks, blame cyphermox, we get lots of messages from NM :-)16:39
didrockstedg: can be, but not the cause of the regression ;)16:39
tedgThis is an N4?16:39
didrockswe both have the issue on mako and maguro16:39
didrockssince we had that diff: http://people.canonical.com/~ogra/touch-image-stats/20131105.changes16:40
tedgJust to be curious, why is IDO in the phone image?16:40
didrockstedg: seems to be a packaging error/change16:41
didrockstedg: but it's not loaded (u-p-s isn't started)16:42
tedgYeah, it just pulls GTK along with it.16:42
didrockstedg: right, but not the cause of the current issue ;)16:42
didrockswe even tried, as not loaded, to revert it16:42
didrocksno effect16:42
didrockstedg: you do have access to the QA lab?16:43
tedgVPN?  Yes.16:43
tedgBut I don't know what to do from there.16:43
psivaadidrocks: tedg: http://pastebin.ubuntu.com/6371436/ is the whole log16:43
didrockspsivaa: I think we really need to give the creds for tedg to connect and gdb16:43
dednickLaney, seb128: at the moment all the ubuntu-settings-components are using components from the indicators in unity8, for which we didn't really have a clearcut design for each component. I guess once we have it in archive we'll do the transision for unity8 first, update for correct design and switch then ubuntu-system-settings.16:43
didrocksyeah, it seems a lot of spam16:43
didrocksnot sure why16:43
dednickLaney, seb128: just need to get it into archive first!16:43
didrockswill be better for tedg to see where the indicator is looping16:44
LaneyI don't know why you need to do the spreadsheet stuff when it's not going to be on any image16:44
Laneyshould just be able to release it16:44
dednickLaney: to where?16:46
Laneyuniverse16:46
dednickLaney: well i have no idea about packaging. But it needs to go in trusty archive surely no?16:48
bregmaseb128, do you think the g-s-d/unity keygrab issue is worthy of a vUDS session or is it more of a "just do it" kind of thing?16:49
Laneydednick: yeah, that's what I mean, but you shouldn't have to go through the spreadsheet stuff if there is no way it'll impact the touch images16:49
Laneydednick: When stuff is moved to use it then that'll probably need to do it16:49
didrocksLaney: before, it was linked to the unity8 move to it16:49
dednickLaney: tis what I was told do to.16:49
didrocksI guess that's why it ended up here16:49
didrocksright now, it's for us to track the release to distro then16:50
didrocksbut right now, as you noticed, we are doing a complex transition and have issues :)16:50
didrocksso rather than discussing process, I would appreciate all available help16:50
Laneyhmm?16:50
LaneyI'm trying to get it off your plate, not attacking :/16:50
didrocksLaney: if you can release it from cu2d (I think it built) and get an AA review, that's fine16:50
didrocksIIRC, everything is wired up16:51
cyphermoxdidrocks: got something suspicious16:51
didrockscyphermox: oh?16:51
didrockstell me you can reproduce it :)16:51
Laneyok cool16:51
tedgdidrocks, I'm logged in, but indicator-network doesn't seem to be using excess CPU?16:52
didrockspsivaa: didn't you tell me it was still going crazy?16:52
psivaadidrocks: it is at the top even now.16:52
cyphermoxdidrocks: reproducing on the lab device16:52
didrockstedg: connected on the same device? ;)16:53
tedgAFAIK16:53
* didrocks wonders why he needs to act as a proxy16:53
didrockscyphermox: ah? what's up then?16:53
tedgAh, interesting.16:53
didrocks(please nobody rebooting anything)16:53
tedgIt wasn't before, but is now.16:53
cyphermoxit gets borked by phonesim, AFAICT16:54
cyphermoxwhen phonesim starts, there's a small period of time where the device isn't quite ready16:54
tedgYeah, it seems to be the system bus.16:54
cyphermoxI'd look at that piece of code for when devices appear, first16:54
tedgThe system dbus-daemon is at 20% CPU16:55
tedgofonod is at 10%16:55
tedgSeems like a game of crack the whip16:55
cyphermoxindicator-network all threads are stuck in poll()16:56
didrocksbut phonesim wasn't changed?16:56
cyphermoxnot that I know16:57
seb128bregma, it seems a technical details, not really a session topic16:59
seb128bregma, we might have one on input methods/ibus/fcitx though17:00
bregmaseb128, https://blueprints.launchpad.net/ubuntu/+spec/desktop-1311-default-imf-review17:00
cyphermoxahahaha17:00
seb128bregma, indeed17:01
Laneybah17:01
Laneyseems the sdk changed ItemSelector.expanded to read only17:01
Laneybreaks half of the u-s-s panels17:01
seb128Laney, did that change land?17:02
LaneyI guess so17:02
seb128Laney, shrug17:02
seb128Laney, https://code.launchpad.net/~nicolas-doffay/ubuntu-ui-toolkit/selectors-multi-selection/+merge/190605 ?17:03
Laneylooks like it17:04
Laneyis that a buggy search and replace?!17:04
tedgcyphermox, It seems the modem state keeps changing.17:04
cyphermoxI'm changing it17:05
tedgcyphermox, http://paste.ubuntu.com/6371532/17:05
tedgOh17:05
seb128Laney, not sure, pinging the sdk guys17:05
Laneyok17:05
Laneyit was prelimenary, so maybe that means that we lose17:05
seb128Laney, that's going to teach us to not have tests :p17:06
Laneyheh17:06
seb128that would have stopped their merge17:06
seb128didrocks, ^ btw, ubuntu-ui-toolkit's update has incompatible changes that makes settings really unhappy17:06
didrocksseb128: I guess you need to talk with the sdk guys… :/17:07
seb128didrocks, doing that right now17:07
Laneythe change was actually noted17:07
Laneyin some comments from timp17:08
seb128Laney, SHRUG17:24
Laneyhaha17:24
seb128nic-doffay is giving me the "you guys should just update your code"17:24
Laneyso much for APIs17:25
seb128yeah, that's mind boggling17:25
seb128I'm not porting system settings, I told them I refuse to ack an incompatible API change in our toolkit like that17:26
=== recoveryRodge is now known as Rodge
* Laney nods17:28
Laneyhave a nice fight then :-)17:28
=== Rodge_ is now known as Rodge
cyphermoxpsivaa: I'll reboot the device now, need to check what's up17:39
psivaacyphermox: ack, hope tedg is ok with it17:39
tedgcyphermox, That's fine.17:41
seb128Laney, can you open a bug with the actual error? I'm still arguing but they want a bug report...17:42
Laneyokay17:42
seb128Laney, seems like they are going to revert the incompatibility though (thanks Kaleo for bringing some sanity)17:43
Laneyit's just what you'd expect though: that the property is read only17:43
Laneyoh nice17:43
Laneyseb128: bug #124864617:46
ubot2Launchpad bug 1248646 in ubuntu-system-settings (Ubuntu) "API break: ItemSelector.enabled changed to read-only" [Undecided,New] https://launchpad.net/bugs/124864617:46
seb128Laney, thanks, they are going to fix the incompatibility, we should wait on their changes to land to port our code, I'm not sure what is going to be renamed and how17:50
Laneyif they fix it then nothing ;-)17:50
seb128Laney, well, the intend is still to deprecate the property so we should use the new one17:51
attentebschaefer, hi19:07
kenvandine:q19:18
bschaeferattente, pong19:21
bschaeferattente, hows the dbus stuff going?19:22
attentebschaefer, i think it's make the grabs now19:24
attente*making19:24
attenteso it's some progress at least19:24
attentebut i'm wondering why the action callbacks aren't getting called19:24
attentecan you explain a bit when the initiate and terminate callbacks are supposed to be called on an action?19:25
bschaeferhmm from what i gather, and event comes in19:25
bschaeferthen it sends a ::handleEvent out to all the plugins first19:26
bschaeferthen it checks if any grabs are pressed19:26
bschaeferIIRC19:26
attenteok19:27
bschaeferit should eventually get: PrivateScreen::triggerKeyPressBindings19:27
bschaeferattente, we use to do this in umm19:27
bschaeferthe switcher, which i removed a bit ago, one problem to keep in mind19:28
attenteand for a KeyBinding action, do initiate and terminate represent like the key press and release? or they have some alternate meaning?19:28
bschaeferif you even remove the key grab, it'll remove it form compiz entirely19:28
bschaeferyeah19:28
bschaeferso like Super+A when thats pressed, the state is Init19:28
bschaeferthen when you release it, it goes to Term19:28
attenteah, ok19:29
bschaeferattente, IIRC, once you get an init you have to remove it from the state19:29
* bschaefer needs to check unity19:29
attentebschaefer, i think the state passed into the callback is an in parameter only, no?19:30
bschaefersomething like:19:30
bschaefer  if (state & CompAction::StateInitKey)19:30
bschaefer    action->setState(action->state() | CompAction::StateTermKey);19:30
bschaeferits in the state19:30
bschaeferparam19:30
bschaeferlooking at: bool UnityScreen::ShowHudInitiate19:31
attenteoh. i see, you meant on the action itself19:31
bschaeferwell this is once you get the press :)19:31
bschaeferso im jumping a bit to far ahead19:31
attenteright, no, this is useful info :)19:31
bschaeferyeah, compiz is pretty ... confusing and i still don't understand most of it :)19:32
bschaeferfrom what i see you have to do:19:32
bschaefer    screen->addAction(action.get());19:32
bschaeferadd an action to the screen19:32
bschaeferwhich the action will be a key binding19:32
bschaeferattente, take a look at: void UnityScreen::CreateSuperNewAction19:32
bschaeferthere it sets up a keybinding, and adds that action to the screen19:33
attentebschaefer, but it doesn't seem to ever set those callbacks19:34
bschaeferhmm right19:35
bschaeferattente, the only way i see the callbacks getting set up is through all those xml stuff we have19:36
bschaeferlike19:36
bschaefer     optionSetAltTabForwardAllInitiate(boost::bind(&UnityScreen::altTabForwardAllInitiate, this, _1, _2, _3));19:36
bschaeferwhich that is generated code in umm19:37
bschaeferbuild/generated19:37
bschaeferunityshell_options.h19:37
bschaeferattente, so, you could possibly look in there to see how those hotkeys are getting set up19:38
bschaeferand instead of using the xml file to generate it you can just code it in yourself somewhere else19:38
attentebschaefer, yeah, i've tried setting callbacks manually on the actions in the key grabber19:39
bschaeferattente, looking at: void UnityScreen::initAltTabNextWindow()19:39
attenteit's just a bit of a mystery why they're not being triggered..19:39
bschaeferyeah...19:39
bschaeferare you doing something like:19:40
bschaefer      optionSetAltTabNextWindowTerminate(boost::bind(&UnityScreen::altTabTerminateCommon, this, _1, _2, _3));19:40
bschaeferwhere you bind the callback?19:40
attentei'm doing something like:19:41
attenteCompAction* action = new CompAction;19:41
attenteaction->keyFromString(accelerator);19:41
attenteaction->setInitiate(boost::bind(&GnomeKeyGrabber::Impl::InitiateAction, this, _1, _2, _3));19:41
attentewhere InitiateAction is the method on the GnomeKeyGrabber::Impl class19:41
bschaeferhmm I dont see that setInit used in unityshell else where19:42
bschaeferattente, do you add that action tot he screen?19:42
bschaeferscreen->addAction()19:42
attentebschaefer, yes19:42
attentemight be better if i do a paste instead :)19:43
attentehttp://paste.ubuntu.com/6372383/19:43
bschaeferhmm19:44
attentewhen i log the addAction method in compiz, it seems to be working and returning true19:44
bschaeferhmm possibly the hotkey is setup different from gnome -> compiz?19:45
attentebschaefer, you mean in the parsing of the accelerator?19:46
bschaeferyeah, im nost sure what format compiz takes in19:46
attentei was just testing it with '<Shift>F11' and '<Shift>F12'19:46
bschaeferhmm i think compiz needs +19:46
bschaeferor x11?19:46
bschaeferhmm I had to write a converter for ibus hotkeys19:46
bschaefertry19:46
bschaeferShift+F1119:46
bschaeferthe conversion is pretty simple if that ends up working...19:48
attenteno luck with that either19:48
bschaeferwell shoot19:48
bschaeferi do wish I understood this part of compiz a bit more :)19:48
attenteheh, don't we all :)19:48
bschaeferattente, oo possibly:19:50
bschaefer      action.setState (CompAction::StateInitKey);19:50
bschaeferas if you don't set the state, it might just ignore that action19:50
attenteooo19:50
bschaefer(as it'll have an empty state)19:50
attenteok, i'll try it19:50
bschaeferim looking at the trigger code in compiz19:50
attentehmm. no luck there either19:55
bschaeferdammit19:55
attentebut something is happening here19:55
attentewhen i press Shift+F8, more calls to addAction are happening19:55
attentethey happen only after Shift+F8 are grabbed19:56
attente*is19:56
bschaeferhmm19:56
attentesame for Shift+F11,F1219:56
bschaefermore calls to addAction from your code?19:56
attentei hope not..19:57
* attente checks19:57
bschaefertry:  action.setState (CompAction::StateInitKey | CompAction::StateAutoGrab);19:57
bschaefertheres a lot of addActions19:57
bschaeferit seems...19:57
bschaeferin unityshell.cpp19:57
attentei wonder why pressing that keybinding would trigger more addAction calls19:57
bschaeferwell the workhorse in compiz is: PrivateScreen::triggerKeyPressBindings19:58
bschaeferso something is happening there that either, could be finding your hotkeys19:58
bschaeferand the callback is failing, or it can't find the hotkeys19:58
attenteok, thanks for the guidance bschaefer19:59
attentei should be able to manage from here19:59
bschaeferattente, good luck, and sorry for the lack of hlep19:59
bschaeferhelp*19:59
attenteno, you were very helpful, thanks!19:59
bschaeferyou could also look in: PrivateScreen::triggerStateNotifyBindings20:00
bschaeferpretty much in compiz/src/event.cpp20:00
bschaefertheres a bunch of stuff going through and checking all the bindings20:00
=== seb128 is now known as 3JTAAAD4U
=== Sweetsha1k is now known as Sweetshark
robert_ancellmterry, still around?22:58
mterryrobert_ancell, yeah23:27
robert_ancellmterry, was wondering if you saw https://code.launchpad.net/~robert-ancell/unity-greeter/end-session-dialog/+merge/19388023:29
mterryrobert_ancell, sure, I put it on my TODO then forgot about it.  Soryr23:30
mterryrobert_ancell, will look now23:30
robert_ancellta23:31
mterryrobert_ancell, looks good23:53
mterryapproved23:53
robert_ancell:)23:53
robert_ancellthanks!23:53

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