[05:12] <pitti> Good morning
[07:50] <jibel> Good morning
[07:55] <jibel> pitti, what is the result of autopkgtest on ARM, is it stable?
[07:57] <pitti> jibel: two nodes finished their quarter of tests this morning, two more are still running
[07:58] <pitti> libnih hung and one node got firefox etc.
[07:58] <pitti> jibel: but so far the tests are progressing well indeed, no crashes or anything \o/
[07:58] <pitti> jibel: http://people.canonical.com/~pitti/tmp/fails-armhf.txt
[07:58] <pitti> jibel: these are my notes so far for tests "pandas" to "z.."
[07:58] <pitti> jibel: I'm re-running the failed ones on my amd64 LXC to see which ones fail due to LXC, and which ones are ARM specific
[07:59] <pitti> jibel: would it be possible to use your superpowers to install byobu or at least screen on wazn?
[07:59] <pitti> I already had to interrupt a test run due to my reboot
[08:00] <jibel> pitti, it is much more reliable than Pandaboards
[08:01] <pitti> indeed
[08:01] <pitti> I'm quite happy about them
[08:01] <jibel> pitti, how much slower than x86 is it?
[08:01] <pitti> jibel: hm, I didn't make precise measurements yet, and I ran these tests without apt-get update/dist-upgrade
[08:02] <pitti> but compared to the ones I'm currently running on x86 (I'll brief you in a minute), my gut feeling is factor 2 to 4
[08:02] <pitti> most tests are really IO/network bound due to having to install lots of dependencies
[08:03] <jibel> pitti, screen is installed
[08:03] <pitti> jibel: thank you
[08:04] <pitti> jibel: so, doko asked me to run all our autopkgtests against https://launchpad.net/~ubuntu-toolchain-r/+archive/python3 which flips the default python3 to 3.4
[08:05] <pitti> jibel: so I took our list of autopkgtests, removed the ones which are obviously not related to python and are big (linux, libo, and the like), divided them in four, and started running a quarter each on alderamin, aldebaran, and wazn
[08:05] <pitti> jibel: their results go into ~auto-package-testing/py34-$pkgname
[08:05] <pitti> jibel: I run them serially, so that they don't take too many resources away from the "real" autopkgtests
[08:06] <pitti> jibel: I now want to do the same on albali, but it's configured differently; is that still the user ~usit, or ~auto-package-testing? both exist, but I don't know how to log in as usit
[08:08] <jibel> pitti, for historical reason jenkins is running under usit. I can sudo to it but if you don't retoaded can give you access
[08:11] <pitti> jibel: or could I just ask you to start running the tests there?
[08:11] <pitti> jibel: http://paste.ubuntu.com/6807223/ is the command (run-run-adt-test path might need adjusting)
[08:12] <pitti> jibel: otherwise I'll distribute them evenly to the other three nodes; they ought to have enough horsepower anywa
[08:12] <pitti> y
[08:17] <jibel> pitti, it is running on albali
[08:18] <pitti> jibel: merci
[08:19] <pitti> jibel: I see py34-python-tornado/ already, so it's working
[08:21] <pitti> jibel: btw, the failures are not that specific to a clear subset of packages; I think it might be best to run them for all packages, and only consider failures in britney if the test ever worked
[08:21] <jibel> pitti, I changed the workspace to ~usit/py34/
[08:21] <pitti> jibel: not only on arm, but on all platforms; that would also stop blocking packages with new autopkgtests which are broken
[08:22] <jibel> pitti, and there is a log in /tmp/run-adt-py34.log
[08:22] <pitti> jibel: ah, so you can remove /var/lib/ubuntu-iso-testing/py34-python-tornado ?
[08:22] <jibel> yes
[08:22] <pitti> nice, so  I can tail
[08:23] <jibel> pitti, agreed, in britney we should consider only regression
[08:23] <jibel> +s
[08:29] <pitti> jibel: did you get any autopkgtest fail/fixed notifications today? I didn't
[08:29] <pitti> and I got some yesterday, but some were missing, too
[08:29] <pitti> and we definitively got failures this morning due to some uninstallability and VM failures
[08:29] <pitti> I fixed the uninstallability and retried the packages, so it's all good now, but I didn't get mail
[08:29] <jibel> pitti, nothing, only notifications from ubiquity tests
[08:30] <pitti> yay. not.
[08:33] <jibel> pitti, I asked the CI team, now waiting of UK to wake up
[08:33] <pitti> thanks
[08:34] <pitti> jibel: btw, some failures were due to having en_US locale on the nodes; I fixed the nodes and their containers to use en_US.UTF-8 now, and re-ran a few tests which looked like they failed due to that
[09:02] <knome> balloons, plong, we need to dicsuss -qa logos ;)
[09:10] <jibel> pitti, configuration of precise amd64 with all of main finally finished. 8125 packages, 54GB. I didn't install extra kernels, and skipped cgroup-lite (fails in a container) and the following packages fail to install on Precise http://paste.ubuntu.com/6807376/
[09:11] <jibel> I'll do an upgrade to Trusty an see how it goes before automating it.
[09:29] <jibel> pitti, that was quick :) It cannot calculate the upgrade
[09:30] <jibel> pitti, python-sip is held back due to python-kde4
[09:40] <pitti> jibel: heavy trucking :)
[09:40] <pitti> jibel: sip> ah, does that need another rebuild due to an ABI change?
[09:42] <jibel> pitti, I dont think so, python-kde4 depends on sip-api-10.1 which is provided by python-sip
[09:44] <jibel> pitti, ah, kde-runtime has been demoted to universe
[09:44] <pitti> jibel: ah, I guess that'll hit us quite often for upgrades
[09:44] <pitti> packages being un-upgradeable due to universe demotion
[09:44] <jibel> so the whole kde stack is not upgradeable
[09:45] <jibel> pitti, how should it be handled?
[09:45] <pitti> jibel: I think for this case we should enable universe for the dist-upgrade
[09:45] <pitti> jibel: and limit the "main only" requirement for an upgrade of the default install
[09:45] <pitti> the latter really ought to work without universe
[09:46] <pitti> but as "all of main" includes the other flavours which used to live in main, it's not going to work with main only, I'm afraid
[09:46] <pitti> and "main only" is also interesting for servers
[09:46] <jibel> pitti, okay, I'll enable universe to see if it improves the situation. We already have profiles with main only for desktop and server
[09:50] <jibel> meh, 414 packages are going to be removed. 1147 new packages are going to
[09:50] <jibel> be installed. 6523 packages are going to be upgraded
[09:50] <jibel> :/
[09:50] <pitti> jibel: i. e. "takes effing long", but "upgrade calculated successfully"?
[09:50] <pitti> jibel: I take it that doesn't fit into RAM/ephemeral any more?
[09:51] <jibel> pitti, yup, pressing Y, and see the result on monday :)
[09:51] <pitti> there, data center, no weekend for you!
[09:51] <jibel> pitti, I'm trying with an ephemeral on disk
[09:52] <jibel> that won't fit in memory, the download only is 10GB
[09:53] <pitti> jibel: this is at least as much a test for overlayfs as for apt/dpkg :)
[09:53] <pitti> jibel: btw, the espeak-data workaround landed after the a2 freeze; /me checks current jenkins whether that shows as fixed now
[09:54] <pitti> jibel: ah, can we re-run http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-quantal-trusty-desktop-amd64/ ?
[09:54] <davmor2> Morning all
[09:54] <pitti> jibel: last run was two days ago, then we didn't have the fix yet
[09:54] <jibel> pitti, oh sure, I'll setup a periodic run
[09:54] <pitti> jibel: i. e. is it safe to just click on "play"?
[09:54] <pitti> poor wazn
[09:54] <jibel> pitti, it is
[09:55] <pitti> these upgrade tests, your "all of main", plus my "python3.4" ones
[09:55] <jibel> + image smoketests
[09:56] <pitti> oh, I thought these were on albali
[09:58] <pitti> jibel: ok, starting upgrade-ubuntu-quantal-trusty-desktop-amd64
[09:58] <jibel> pitti, scheduled at 1934UTC every days
[09:59] <pitti> jibel: btw, https://jenkins.qa.ubuntu.com/view/Trusty/view/Upgrade%20Testing/ is empty, but https://jenkins.qa.ubuntu.com/search/?q=upgrade has them; is that a problem for retoaded?
[09:59] <slickymaster> morning all
[10:22] <jibel> pitti, yeah, he'd like to wait for rick and create the view next week. it is tracked in RT67161
[10:23] <pitti> jibel: cheers
[10:34] <pitti> jibel: meh, http://d-jenkins.ubuntu-ci:8080/job/upgrade-ubuntu-quantal-trusty-desktop-amd64//6 failed again; it still got the previous espeak version without the workaround :/
[10:35] <pitti> the new one landed in trusty 17 hours ago, maybe squid is lagging behind or so
[10:35] <pitti> but these run automatically now, right? so I'll just check on Monday
[10:39] <jibel> pitti, I'll have a look, on another system the version is correct
[10:39] <jibel>   Installed: 1.46.02-0ubuntu1
[10:39] <jibel>   Candidate: 1.47.11-1ubuntu1
[10:39] <pitti> no hurry/worry for now
[10:40] <jibel> during this Quantal upgrade dpkg reloaded its database 132 times :)
[13:40] <jibel> pitti, albali is done with py34 autopkgtests
[13:40] <jibel> pitti, do you want the results somewhere?
[13:45] <pitti> jibel: yep, let me scp them (already done for two other boxes; aldebaran was still busy)
[14:00] <pitti> jibel: they are all on http://people.canonical.com/~pitti/tmp/autopkgtest-py34/, so you can clean up
[14:00] <pitti> jibel: I'll clean up on the other three boxes
[14:01] <jibel> pitti, done, thanks
[16:36] <pitti> jibel: *phew*, I'm done staring at test logs all day; I sent results to u-devel@
[16:36] <pitti> have a nice weekend everyone!
[16:37] <elfy> cya pitti
[16:40] <jibel> pitti, I've read your post, it is very good. Have a nice week end.
[17:21] <cgoldberg> balloons, just saw your post with the core-apps contributor stats... awesome stuff!  on a related note, I'm gonna update my gource video for core-apps.. did you ever see this one I did in november?  http://coreygoldberg.blogspot.com/2013/11/gource-visualization-of-ubuntu-touch.html
[17:38] <balloons> cgoldberg, yes I remember it well
[17:38] <balloons> good idea :-)
[17:39] <cgoldberg> balloons, cool.. I'll make one either today and this weekend
[17:39] <balloons> there's a new app as well, reminders :-)
[19:57] <balloons> hallo dkessel
[19:57] <balloons> buonasera Letozaf_
[19:57] <balloons> Happy Friday to you all.. Happy Saturday to you Noskcaj :-p
[19:57] <Letozaf_> balloons, buonasera
[19:58] <Noskcaj> hey balloons
[19:58] <Letozaf_> balloons, happy Friday to you too :)
[19:59] <dkessel_> hello balloons. hm. seems xchat still is not unminimized but launched a second time...
[20:00] <dkessel_> how are you? it's been a busy start of the year for me...
[20:00] <balloons> dkessel_, I'm doing well. Just had to run the dog to the vet, but he'll be ok, so :-)
[20:01] <dkessel_> good to hear that :)
[20:01] <Letozaf_> balloons, I found another issue with calendar-app and dates: https://docs.google.com/file/d/0BwSy2uwGals0cmFjcEtHbVVXVUE/edit
[20:02] <Letozaf_> balloons, :( sorry for your dog
[20:02] <balloons> awesome Letozaf_ .. looks like the current Year is instead the current date and time
[20:03] <balloons> as a unix timestamp :-)
[20:03] <Letozaf_> balloons, and it's not today 24 but 23 yesterday
[20:03] <balloons> Thanks for the thoughts.. The hard part now is getting him to take it easy so he recovers
[20:03] <Letozaf_> balloons, should I upload this to the bug I reported or open another one
[20:04] <balloons> Letozaf_, I think this is a bit different.. how did it come up?
[20:04] <Letozaf_> balloons, oh yes I think it will be hard, dogs don't like to stay calm
[20:05] <Letozaf_> balloons, I was trying to fix other locale issues and had a test on year tab that didn't work due to locale so I was looking for another way to run the test in autopilot vis
[20:06] <Letozaf_> balloons, looking for something to inspect in autopilot vis made me run into this
[20:06] <Letozaf_> balloons, should I report another bug ?
[20:06] <balloons> Letozaf_, which page is it on?
[20:06] <balloons> we can look into the qml file a bit
[20:06] <balloons> it might simply be misnamed
[20:06] <balloons> but yes, this would be a separate bug I think
[20:08] <Letozaf_> balloons, the page should be YearView.qml
[20:08] <Letozaf_> balloons,  property var currentYear: DateExt.today();
[20:10] <balloons> Letozaf_, ahh.. so indeed it's showing up as per the qml.. but property var currentYear: DateExt.today(); is not going to be equal to the currentYear :-)
[20:11] <balloons> so it's a bug perhaps in the naming
[20:12] <Letozaf_> balloons, ok but shouldn't it be 24 Jan instead of 23 Jan ?
[20:12] <Letozaf_> balloons, today = 24 not 23
[20:12] <balloons> Letozaf_, ohh, well right, hehe
[20:13] <balloons> so that might be a side effect of your bug from yesterday
[20:13] <balloons> I get it now :-)
[20:13] <Letozaf_> balloons, don't think it important though that property might not be usede
[20:13] <Letozaf_> used
[20:31] <Letozaf_> balloons, I'm not sure if it's better I wait for that bug to be fixed to continue with locale issues in test failures as I'm finding other properties with weired dates
[20:32] <balloons> Letozaf_, I would link what you found and say you see lots of weird dates
[20:32] <balloons> and you are right, they are probably connected
[20:33] <Letozaf_> balloons, ok
[20:35] <dkessel_> good night, bb
[20:37] <balloons> Letozaf_, calendar hack day is monday
[20:37] <balloons> we can poke the devs then :-)
[20:37] <Letozaf_> balloons, right! good idea