[05:29] <pitti> Good morning
[05:29] <pitti> thomi: what's up?
[05:29] <thomi> hi pitti
[05:30] <thomi> pitti: ummm, I think I got someone else to answer my question
[05:31] <thomi> what's up with you pitti anyway?
[05:32] <pitti> thomi: how do you mean?
[05:32] <DanChapman> morning all
[05:32] <thomi> dunno, was hoping you'd distract me with exciting takes from mainland europe :)
[05:32] <pitti> thomi: me being back in the one and only timezone? :-)
[05:32] <thomi> hi DanChapman
[05:32] <jibel> Good morning
[05:33] <thomi> pitti: exactly
[05:33] <DanChapman> hey thomi
[05:33] <pitti> bonjour jibel, comment vas-tu ?
[05:33] <thomi> hey jibel, do you know how to un-stick the daily-release jobs?
[05:34] <thomi> last time I looked, some job had been going for like 6 hours or so, blocking everything else
[05:35] <jibel> pitti, ça va bien et toi?
[05:35] <jibel> a bit less fine since thomi asked a question about daily release
[05:35] <jibel> :)
[05:35] <jibel> thomi, which job ?
[05:35] <jibel> thomi, good evening
[05:35] <pitti> jibel: hah
[05:36] <thomi> jibel: ummm, will have to find it again - actually, don't worry about it. I'm about to EOD, and I'm sure the european CI crew will find it
[05:37] <pitti> jibel: je vais bien, mais après le premier Taekwondo hier j'ai les muscles endoloris :)
[05:44] <jibel> thomi, okay, it was just to make sure I'm in a bad mood to start my day ;)
[05:45] <thomi> jibel: sorry
[07:09] <DanChapman> thomi, if I want to patch an environment var in an autopilot test. Do I just set the LC_ALL environment var?  something like .patch_environment("LC_ALL", "zh_CN.utf8") and would this cause the x11 keyboard input to change?
[07:09] <DanChapman> sorry set locale environment var
[07:10] <thomi> DanChapman: yeah, using 'self.patch_environment' will affect just that one test, and only processes that are spawned after you make that call
[07:11] <DanChapman> thomi, awesome cheers. :-)
[07:11] <thomi> yw
[07:52] <jibel> pitti, autopkgtest 2.3.5 didn't fix the timeout issue with firefox and bluez. I think it must kill process tree not only 'tee's. I'm testing something but FF takes a while to run.
[08:00] <pitti> jibel: you can specify --timeout-test=20 or so to make it fail quickly
[08:00] <pitti> jibel: but weird, in the ps tree that you had yesterday it was the stdout/err tee child which hung
[08:00] <pitti> jibel: is that VM still running somewhere where I could have a look?
[08:02] <jibel> pitti, it is running on wazn
[08:02] <jibel> pitti, $(ps a -ocmd|grep ^ssh) should connect you to the vm
[08:03] <jibel> pitti, you can restart adt-run in the vm if you wish
[08:03] <pitti> jibel: I'm in, looking
[08:04] <pitti> jibel: there's also the "tee" program from this Errplumb thing (which I don't really understand yet)
[08:04] <pitti> so it seems right now all processes are still running, the topmost adt-run doesn't want to die yet
[08:05] <pitti> jibel: but that only started about an hour ago, so it shouldn't time out yet?
[08:05] <pitti> jibel: what's confusing is that there are two adt-runs, one from 06:59, the other from 07:39
[08:10] <pitti> jibel: looks like it still has some trouble with the session manager?
[08:11] <pitti> http://10.98.0.1:8080/job/saucy-adt-firefox/ARCH=amd64,label=adt/164/console
[08:11] <pitti> that still seems to be right in the middle of tests
[08:19] <pitti> jibel: I'm tailing the log files, there's still some actual progress there
[08:21] <jibel> pitti, it should hang at the end of the browser test
[08:21] <pitti> jibel: ok, so let's check in about 3 hours when the timeout is due?
[08:21] <jibel> pitti, yes
[08:59] <slickymaster> morning all
[10:11] <jibel> pitti, FYI firefox test is stuck on wazn
[10:12] <pitti> jibel: hm, did you kill it?
[10:12] <pitti> jibel: just logged into wazn, there's no kvm running
[10:13] <pitti> ah, jenkins still thinks it's running on http://10.98.0.1:8080/job/saucy-adt-firefox/ARCH=amd64,label=adt/164/
[10:13] <jibel> pitti, no I didn't touched it
[10:13] <jibel> pitti, it is a manual run not visible on jenkins
[10:13] <pitti> jibel: nevermind, silly me
[10:13] <jibel> ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -o CheckHostIP=no -i /home/auto-package-testing/cache/disks/adtkey -p 54326 -l ubuntu localhost
[10:13] <pitti> jibel: yes, I see it; forgot the "a" in ps, sorry
[10:13] <jibel> np :)
[10:17] <pitti> adt-run1: testbed executing test finished with exit status 0
[10:17] <pitti> hm, that's the last line
[10:18] <pitti> jibel: did you start this from a shell? I guess it didn't show any kind of exception or something?
[10:19] <jibel> pitti, this is the output on the terminal http://paste.ubuntu.com/6158145/
[10:19] <jibel> there is an exception but for firefox's runner
[10:20] <jibel> it shouldn't interfere with adt-run itself, just fail the test IMO
[10:20] <jibel> the test ended actually INFO | runtests.py | Running tests: end.
[10:22] <pitti> ok, so that seems to reflect the log file
[10:22] <pitti> so it seems it didn't actually bomb out
[10:36] <pitti> jibel: does your command continue to run now?
[10:37] <pitti> jibel: I killed the hanging tee for stdout which was stuck on a read() call
[10:37] <pitti> jibel: I have an idea what happened there, will try to reproduce in a test
[10:37] <pitti> jibel: but either way, I know how to make this more robust so that it never hangs
[10:50] <pitti> jibel: ok, should be fixed now; fixing a build failure on lucid/precise/wheezy, then uploading .6
[11:16] <pitti> jibel: 2.3.6 uploaded
[11:20] <davmor2> Morning all
[11:20] <jibel> pitti, sorry, lunch time. The command continues
[11:20] <jibel> pitti, thanks, I'll retry with 2.3.6 when it is in Ubuntu
[11:21] <pitti> jibel: dejeuner> ici aussi maintenant :)
[12:04] <infinity> Oh hey, I'm in this channel.  Look at that.
[12:04] <infinity> So, is anyone testing the current lubuntu images?
 I've noticed.... and tests are being run.
[12:09] <smartboyhw> inara, ^
[12:09] <smartboyhw> Oops
[12:09] <smartboyhw> infinity, ^
[14:02] <balloons> jibel, coming mate?
[14:05] <elopio> hello!
[14:09] <balloons> howdy elopio :-)
[15:14] <balloons> jibel, DanChapman http://pad.ubuntu.com/k6nvveLDHX
[15:14] <balloons> please add anything I missed, etc
[15:14] <jibel> pitti, I get http://paste.ubuntu.com/6159102/ after running apport-cli on a phone and asking it to retrace any idea what I did wrong?
[15:18] <balloons> elopio, how's https://code.launchpad.net/~elopio/qtcreator-plugin-ubuntu/uitk_base_class/+merge/187313 coming?
[15:21] <jibel> balloons, thanks
[15:28] <balloons> can someone who is running saucy cofirm something for me? if you ctl+alt+numpad 7 does your window snap to the upper left? And if so, does pressing it again scale it smaller in the upper left?
[15:30] <balloons> weird, the grid cycle settings got reset..
[15:43] <davmor2> elopio: Morning good sir
[15:54] <pitti> jibel: not immediately; what's the command line? Do you use -S system ?
[16:50] <cgoldberg> mhall119, hey... for getting documentation into developer.ubuntu.com/api  .... do you have a generic html importer?  or do I need to write a special one for autopilot like import-autopilot.py ?  I'm tweaking the Autopilot docs and want to try importing them "as-is"... importer just needs to strip the html <head>
[16:51] <mhall119> cgoldberg: you'd need to write an import-autopilot.py, but if there's not much work to do on the content it can probably follow import-juju.py pretty closely
[16:52] <cgoldberg> mhall119, ok.. will do.
[16:52] <cgoldberg> mhall119, so how does the deploy process work once the command is implemented?  basically how does the import-autopilot get called, and where would docs need to reside?
[16:53] <mhall119> cgoldberg: we haven't finalized that yet
[16:54] <mhall119> currently I branch and build on my laptop, then scp the generated docs to our staging canonistack server, then run the django manage.py commands to import them
[17:00] <cgoldberg> mhall119, so I should run a local version of ubuntu-api-website to view the generated docs?
[17:00] <cgoldberg> for testing/tweaking
[17:01] <mhall119> cgoldberg: that's the easiest yeah
[17:01] <mhall119> setting up a dev instance is simple, just plain virtualenv and django runserver
[17:01] <mhall119> everything you need should be in the README
[17:02] <cgoldberg> mhall119, great.. thanks
[17:39] <DanChapman> balloons, jibel i'm just trying to test all the languages are unicode but the type being returned is  <class 'autopilot.introspection.dbus.String'>.  Not sure what to do with it
[17:43] <balloons> DanChapman, what properties does it have?
[17:49] <DanChapman> balloons: i'll just check
[17:53] <DanChapman> balloons: it is already a property its the GtkTextCellAccessible.accessible_name.
[17:55] <DanChapman> which in effect a GtkLabel.label value, as ATK uses the label value if a parent widget has a GtkLabel as a child
[18:54] <balloons> DanChapman, so heh, sorry I was kneedeep in my own autopilot fun. So are you ok, or stuck?
[18:58] <DanChapman> balloons: more confused at the moment. See if I get a GtkTextCellAccessible.accessible_name property and then do .encode('UTF-8') for some weird reason i get <type str>, so then i suppose i can just stick u'' in front of them all. But this feels a bit wrong
[18:58] <balloons> I really dislike unicode and python
[18:59] <balloons> I've given up trying to get it right
[18:59] <balloons> DanChapman, I've got stuck on similar things.. autopilot is moving to python 3 which should be the magic pixie dust we want ;-)
[19:01] <DanChapman> so if I stick 'u' in front of each str and then just do an IsInstance on it to test its unicode would that be enough for now?
[19:02] <DanChapman> i've read unicode is supposidly alot easier with py3
[19:02] <DanChapman> :-)
[19:05] <DanChapman> scrap that it don't work lol
[19:06] <balloons> DanChapman, sadly I'm not pythonic enough to help.. thomi will be around shortly and he is a king of python.. I'm sure he'll have a proper answer
[19:11] <cgoldberg> balloons, if you haven't seen this presentation, it's very helpful:  http://nedbatchelder.com/text/unipain.html
[19:12] <DanChapman> balloons: I've got to shoot now, i will try and grab him in the morning.
[19:12] <balloons> DanChapman, bookmark cgoldberg 's link too.. looks useful
[19:12]  * DanChapman has bookmarked
[19:13] <balloons> DanChapman, umm as far as timing, hmm..
[19:13] <balloons> thomi is in NZ, so :-)
[19:13] <balloons> that's a hard one for UK'ers
[19:13] <DanChapman> its not so difficult with a two year old ;-) 5 am is the norm for me
[19:14] <balloons> haha! alrighty then
[19:14] <DanChapman> anyway catch you all tomorrow. Night folks
[19:14] <balloons> night DanChapman !
[19:53] <balloons> letozaf, how are you?
[19:54] <letozaf> balloons, hello
[19:55] <letozaf> balloons, I'm quite fine and you ?
[19:55] <balloons> hungry.. :-) hah! Just looking through some of the unconverted autopilot apps
[19:55] <balloons> You up to try another conversion and test writing?
[19:58] <letozaf> balloons, I got stuck a couple of times, once  with sudoku app integrating emulator I'm doing something wrong, can't figure out what :( , then the rssreader app, the next button adding a feed has dissapeared
[19:59] <balloons> letozaf, yes, I'm trying the sudoku app now because I've been trying to add click support for it, so, I'll cover that one :-)
[19:59] <letozaf> balloons, ok
[19:59] <balloons> the rssreader has a known bug, as you said, so it's good for now. However, the doc viewer devs just said 2 tests are unblocked
[20:00] <balloons> https://bugs.launchpad.net/ubuntu-docviewer-app/+bug/1188373
[20:00] <balloons> https://bugs.launchpad.net/ubuntu-docviewer-app/+bug/1188762
[20:00] <letozaf> balloons, ok I will try this one then
[20:00] <balloons> I noticed the tests also need converted to the sdk.. I've not looked further, but ;-)
[20:00] <balloons> it can only be exciting right?
[20:02] <letozaf> balloons, sure :p
[20:12] <thomi> morning
[20:25] <balloons> morning thomi
[20:25] <thomi> hey balloons, what's up?
[20:26] <balloons> not much any longer.. updating some old autopilot tests and playing with click.. poor DanChapman was being roughed up by unicode.. I told him I would have to defer as unicode and python is lost on me
[20:47] <thomi> balloons: unicode and python is a PITA in python2, but much much easier in python3
[20:49] <balloons> thomi, yea, that's pretty much what I shared. python3 is magic pixie dust and autopilot is migrating to it
[20:49] <thomi> balloons: s/is migrating/has migrated/
[20:50] <balloons> well, we need to just use 1.4
[20:50] <thomi> balloons: I mean, you still need to think about what you're doing, it doesn't remove that obstacle
[20:50] <balloons> new answer right :-)
[20:50] <thomi> ooh, which reminds me....
[20:50]  * thomi files a bug
[20:50] <thomi> before I forget
[21:37] <phillw> Hi, did beta-2 go out yet?
[21:38] <knome> no
[21:39] <phillw> okies, thanks, was worried I'd un-subscribed from too many mailing lists! Catch you all later :)
[22:13] <thomi> veebers: could you please review this for me? https://code.launchpad.net/~thomir/autopilot/trunk-add-wait_select_single/+merge/187937
[22:14] <veebers> thomi: sure, I'll do that now
[22:15] <veebers> thomi: all the 1.4 autopilot bits live in ppa:autopilot/experimental right?
[22:15] <thomi> veebers: for now, yes
[22:15] <veebers> cool, thanks
[22:17] <veebers> thomi: Initial thought is the timeout should be mentioned in the intial docstring sentence lines: 48-51. I.e. . . . until a valid object is found or the timeout is reached
[22:18] <thomi> veebers: ok, will fix that now
[22:19] <veebers> thomi: also in the docstring of that method, it states the param type_name, but not what's expected in **kwargs
[22:20] <thomi> veebers: that's probably because **kwargs is free-form. I wanted to avoid copying the entire docstring from select_single
[22:20] <thomi> but I'll copy that part as well
[22:20] <thomi> ugh, fuck it, I'll rewrite the whole thing
[22:20] <veebers> fair enough, just found it odd that one param is documented but not the second. Perhaps just a one-liner stating "**kwargs is the same as select_single or something
[22:27] <thomi> veebers: docstring updated
[22:27] <thomi> and pushed
[22:27] <veebers> thomi: awesome
[22:32] <thomi> veebers: just saw another issue, don't approve yet
[22:32] <veebers> thomi: ack
[22:33] <thomi> veebers: ok, pushed revision 341
[22:36] <thomi> veebers: erp, spotted another issue :-/
[22:36] <veebers> thomi: heh, I'm just waiting for dist-upgrade so I can run tests
[22:37] <veebers> thomi: I thought it was :py:meth: not :py:met:?
[22:37] <thomi> veebers: exactly :)
[22:37] <veebers> ah, I see :-)
[22:38] <thomi> veebers: ok, fix pushed
[22:38] <thomi> veebers: in this MP, those docstrings aren't actually parsed, but I'm preparing a second MP that does parse those...
[22:39] <veebers> ok sweet, I've gotten into the habit of writing docstring in that method regardless if they're parsed or not :-\
[22:39] <thomi> well, the autopilot docs need some attention
[22:40] <thomi> ideally we'd parse everything, and make the navigation a bit better
[22:57] <thomi> veebers: second MP is here: https://code.launchpad.net/~thomir/autopilot/trunk-fix-docs/+merge/187951
[23:17] <thomi> veebers: any progress with those MPs?
[23:18] <veebers> thomi: Just finished running the tests for the first one
[23:18] <veebers> approved
[23:18] <veebers> looking at the 2nd now
[23:19] <thomi> heh, I wonder if I can produce MPs faster than you can review them?
[23:19] <thomi> maybe if I stopped reading the 'nets :)
[23:21] <veebers> thomi: line 116, why is it ``QPushButton`` and not "QPushButton"?
[23:22] <thomi> veebers: because... that's how sphinx works?
[23:22] <thomi> ``foo`` makes it a keyword
[23:22] <thomi> and is rendered as monospace font etc
[23:22] <veebers> thomi: ah ok I see
[23:22] <thomi> I tend to use it for class names etc
[23:23] <veebers> thomi: sweet. Approved
[23:23] <thomi> thanks
[23:23] <thomi> will have another one for you soon
[23:25] <veebers> nw, sweet
[23:36] <thomi> veebers: another MP for you: https://code.launchpad.net/~thomir/autopilot/trunk-support-point3d/+merge/187954
[23:39]  * veebers looks