/srv/irclogs.ubuntu.com/2014/03/26/#ubuntu-quality.txt

=== _salem is now known as salem_
=== maclin__ is now known as maclin
=== salem_ is now known as _salem
pittiGood morning05:51
DanChapmangood morning06:49
dkesselGood morning07:24
elfymorning dkessel07:25
dkesselmorning elfy07:32
jibelGood morning07:46
elfymorning jibel07:47
jibelmorning elfy07:52
=== maclin_ is now known as maclin
=== _salem is now known as salem_
davmor2jibel: uefi+sb first issue discovered I'm hitting ENTER and RETURN on the keyboard after ejecting the cd and nothing is happening :(12:56
davmor2jibel: looks like the keyboard is completely disabled ctrl+alt+Fx isn't working either12:57
davmor2jibel: and now the second clicking on the cross on the keyboard shortcuts doesn't actually close the keyboard shortcuts other than that perfect install13:02
=== roadmr_afk is now known as roadmr
davmor2jibel: if you open say launchpad and a webapp you get 2 firefox windows and 2 markers to the left of the FF icon in the Launcher correct?  If you then close the webapp but keep the launchpad window open do you still see the 2 markers to the left of the FF icon?13:43
jibeldavmor2, I do.13:47
jibeldavmor2, this is a bug in bamf13:47
jibeldavmor2, actually it happens only on first try, if you try again is works properly13:48
davmor2jibel: confirmed14:16
jibeldavmor2, file a bug but I think it's minor14:16
davmor2oh indeed just something I noticed :)14:17
pittijibel: darn; so after hours of debugging I finally found the cause of bug 1283966 (testing fix now), and it's not the same as bug 1267116 :(14:28
ubot5bug 1283966 in autopilot-gtk (Ubuntu) "Looping while selecting the same object is causing the application to die raising a dbus no reply exception" [High,Triaged] https://launchpad.net/bugs/128396614:28
ubot5bug 1267116 in autopilot-gtk (Ubuntu) "ubiquity crashed with SIGSEGV in GtkNode::MatchStringProperty()" [High,Confirmed] https://launchpad.net/bugs/126711614:28
pittiso back to the drawing board14:28
* davmor2 hugs pitti it sounds like he needs it14:29
pittidavmor2: thanks14:29
* davmor2 also passes pitti a new marker for the drawing board, also sounds like he'll need it :)14:30
pittidavmor2: I also need a drawing board in the first place :)14:30
jibelpitti, great! do you think 1267116 and 1284671 are really duplicates? traces are slightly different and the bucket in errors.u.c corresponds to 128467114:30
davmor2pitti: sorry I don't keep spare drawing boards request one from msm I'm sure the office have extra ones :)14:31
* jibel hands pitti white paint to repaint his office14:31
davmor2hahaha14:32
davmor2yeah use the walls and repaint once a month :D14:32
pittijibel: they look similar enough to me; as long as we don't have a practical reproducer (i. e. environment where one can debug this, and where turnaround time is only a minute or less),  I think we can treat them the same14:32
pittijibel: I doubt that for a memory corruption like that the traces will always be identical14:32
pittijibel: haha14:32
pittijibel: ah, but indeed http://paste.ubuntu.com/6725605/ also has a critical14:39
pittijibel, DanChapman: how reliably can bug 1284671 be triggered?14:40
ubot5bug 1267116 in autopilot-gtk (Ubuntu) "duplicate for #1284671 ubiquity crashed with SIGSEGV in GtkNode::MatchStringProperty()" [High,Confirmed] https://launchpad.net/bugs/126711614:40
pittiany chance this could be run with G_DEBUG=fatal-criticals and "ulimit -c unlimited" to get a proper core file for this atk_object_set_parent: assertion 'ATK_IS_OBJECT (accessible)' failed14:40
jibelpitti, easy, I can remove the workaround (uninstall ubiquity slideshow) and run the tests14:40
pittijibel: how much effort is it to set up this testing rig locally?14:41
DanChapmanpitti there is the test runner in ubiquity trunk, (just so you know i have only managed to reproduce locally 3 or 4 times out of 100's of attempts)14:42
pittijibel: but yeah, if it's not too much effort for you to run the test under that environment, a core file would be great14:42
jibelpitti, I forwarded you what I sent to veebers with the instructions to run locally14:43
jibelpitti, but locally I can reproduce it like 10% or less of the time14:43
pittijibel: well, I can of course just poke around on albali or wherever that's running14:43
pittijibel: ah, thanks for the mail14:45
jibelpitti, where will the core file be generated?14:49
jibelpitti, also do I need to reset core_pattern to disable apport?14:51
jibelor I just leave everything as it is and let apport do its work14:52
pittijibel: apport checks the crashing processes' ulimit and create core files just like the kernel would, so no need to touch14:53
jibelok14:53
pittijibel: core file will be in the cwd of the crashed processs (ubiquity), so probably /root or /home/ubuntu14:53
davmor2jibel: https://bugs.launchpad.net/ubuntu/+source/bamf/+bug/1297886 can you do the honours please15:01
ubot5Ubuntu bug 1297886 in bamf (Ubuntu) "two marker on FF after closing a webapp" [Undecided,New]15:01
jibelpitti, tests running, without the workaround + G_DEBUG + ulimit -c unlimited15:06
jibelpitti, not sure what it means but with G_DEBUG ubiquity crashed almost immediately https://jenkins.qa.ubuntu.com/job/ubiquity_ap-ubuntu_devel_daily-test_english_default/ARCH=amd64,label=rabisu/128/15:19
jibelwell, stopped with SIGTRAP15:20
jibel/var/local/autopilot//logs//autopilot.log: [Gtk] UNKNOWN: gtk_radio_button_set_group: assertion '!g_slist_find (group, radio_button)' failed15:21
jibel/var/log/syslog: Mar 26 15:07:46 ubuntu kernel: [   77.614241] traps: ubiquity[4827] trap int3 ip:7fd8aac19c13 sp:7fff30156970 error:015:21
pittijibel: ah, crap; yeah, I guess if there's tons of "expected" criticals that G_DEBUG approach won't get us very far :(15:36
=== chihchun is now known as chihchun_afk
=== chihchun_afk is now known as chihchun
=== bfiller is now known as bfiller_afk
=== roadmr is now known as roadmr_afk
=== roadmr_afk is now known as roadmr
jibeljamespage, hi, some server tests are not covered for this milestone, particularly MaaS and iSCSI. Could the server team test those?18:11
jamespagejibel, sure18:17
davmor2jibel: right back from tea images downloading I'll hit uefi and mac first18:25
=== bfiller_afk is now known as bfiller
jibeldavmor2, did you download them from pending/ ?18:33
jibeldavmor2, because only latest mac is available from current/18:33
davmor2meh18:33
jibelstgraber, the script updates the tracker when images are not yet in current/18:34
jibelfor example in http://cdimage.ubuntu.com/daily-live/current/ there are both latest and previous builds depending on the arch18:34
stgraberjibel: nusakan updates the tracker once it has asked all mirrors to sync from it18:39
stgraberjibel: we have no way to know when the rsync is actually done so that's as good as it gets...18:39
davmor2jibel: oh so hang on queuebot said Builds for ubuntu had been updated does that not mean released then18:57
davmor2jibel: does it take ages to get from the requirements check to the partitioning step20:42
jibeldavmor2, no, it should be almost immediate20:45
davmor2jibel: it's taking like a minute or so here.  I wonder if the TB drive has anything to do with it :)20:46
jibeldavmor2, you can open a terminal and check what it is doing. also /var/log/syslog and /var/log/installer/debug might give you an indication if something went wrong20:47
davmor2jibel: I'll have a look next time20:47
jibelso ubiquity-dm and encrpyted home are working now.20:48
jibelI'm now verifying oem installationds20:48
jibel-d20:48
davmor2jibel: yeah setup of the tools and then scan of the drive by the look of it21:17
davmor2jibel: that would slow it down a on 1TB drive I guess :)21:17
jibeldavmor2, who needs a 1TB drive now that everything's on the cloud :)21:18
davmor2jibel: well when you need a machine that can install the entire myapps for 6 versions of ubuntu you need a bit of space21:19
davmor2jibel: the mac with it's smaller drive is much quicker there, but I'm assuming you in vm have like 23 GB driver or something21:20
jibeldavmor2, exactly21:20
davmor2once that is loaded the installer is rapid it's just that one step that is slow21:21
=== roadmr is now known as roadmr_afk
davmor2jibel: is it me or is there no option for encrypted home on 64bit?  I see encrypted lvm disc but not just home unless I'm reading them wrong21:43
jibeldavmor2, it is in the user setup step. I tried it earlier tonight21:44
jibellet me verify again21:44
davmor2jibel: no I mean a test listed for it21:45
davmor2jibel: I can confirm that there is a encrypted home option I just don't see a test to mark it against21:48
jibeldavmor2, ah on the tracker, there is a gap in the test plan. It would be worth adding a mandatory test for this21:48
jibelI added an automated test this afternoon, just not deployed yet.21:48
davmor2jibel: I think so being as it is available as a default option.21:49
davmor2jibel: I'll add it to other features I think21:49
jibelsounds good21:52
=== salem_ is now known as _salem

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