=== _salem is now known as salem_ === salem_ is now known as _salem === maclin__ is now known as maclin === bfiller is now known as bfiller_afk [07:06] Good morning [07:57] jibel: poor wazn, load of 9 and 6 QEMUs running in parallel.. [07:57] (DKMS at the moment, apparently) [07:59] jibel: someone disabled the http_proxy setting from wazn's /etc/environment, and it seems when calling prepare-testbed from jenkins it doesn't get read (through PAM) anyway, so on alderamin/aldebaran/wazn I now added them to ~/.adtrc [07:59] jibel: could you please do the same on albali? I. e. append [07:59] http_proxy="http://squid.internal:3128" [07:59] https_proxy="http://squid.internal:3128" [08:00] Bonjour pitti [08:01] pitti, done [08:17] jibel: merci [08:17] jibel: bonjour! [08:22] pitti, do you think autopkgtest of apt timed out because a proxy is set? [08:23] we should also add a no_proxy=localhost [08:26] pitti, also for libreoffice, it fails because a chown of the test tree takes more than 100s and times out [08:27] do we really need all these timeouts in autopkgtest, or only 1 for the build and 1 for the test? [08:43] jibel: re [08:43] jibel: haven't looked at apt yet, next on my list [08:43] jibel: ah, I thought localhost was kind of implied, but sure, let's add that too [08:44] jibel: I didn't introduce them, but I think having *some* time out is good as sometimes copying stuff around involves tar | scp | tar etc. which can easily hang indefinitely [08:55] pitti, I know you didn't, but there are timeouts for nearly every action, which is probably too much e.g for dir copies or chown, I hardly see how these commands can block [08:57] I'll increase the short timeout to see if it makes libreoffice test go further [08:57] *nod* [09:09] jibel: I added no_proxy=localhost to .adtrc on the usual three and pushed a commit to also transition $no_proxy to the VMs [09:10] pitti, added to .adtrc and pulled on albali [09:10] jibel: merci [09:12] jibel: and FTR, yay for fixing the precise-lts-backports upgrades [09:12] * pitti looks at the remaining failures [09:28] Morning all [09:35] jibel: output_file = '/tmp/obsolete_conffiles.log' [09:35] jibel: how does that become a jenkins artifact? [09:35] jibel: does it just attach /tmp/*.log or do I need to register new files somewhere? [09:36] I think it's hardcoded somewhere :/ [09:36] jibel: I want to make the conffile test more verbose, and add old/new files (perhaps a diff) as a new /tmp/conffile_prompts.log [09:36] in UpgradeTestBackendSSH.py [09:36] ./AutoUpgradeTester/UpgradeTestBackendSSH.py: self._copyFromImage("/tmp/obsolete_conffiles.log", self.resultdir) [09:36] I only see that [09:37] does that apply to LXC as well? [09:37] pitti, it does [09:37] because these tests do get obsolete_conffiles.log [09:37] jibel: ah, ok; thanks! [09:37] # find /etc -name '*.dpkg-dist' [09:37] /etc/default/bind9.dpkg-dist [09:37] I found it, but it's still nicer to see it right away in jenkins [09:37] (reproduced locally in a chroot, I meant) [09:38] but there are half a dozen files in /etc named "bind9", it's too obtuse ATM [09:39] pitti, I should copy to a /tmp/result in the guest and copy that instead of individual files and preserve the directory tree structure, at least in the name [09:42] jibel: right, that should be easy; queueing that [09:48] * pitti pushes another commit that fixes upgrade-ubuntu-quantal-saucy-desktop-{i386,amd64} [10:01] jibel: do you know why test_dpkgdist does all that dance with copying stuff to /tmp ? [10:02] jibel: it would seem much more useful to me to compare the whitelist against the original paths, and print out the list comparison with that [10:02] jibel: in other words, do you mind if I completely rewrite that? (I have a work item "fix conffile upgrade test" anyway) [10:31] jibel: I pushed r94 for that, it just puts the diffs into the main log now [10:31] * pitti looks into the actual bind issue now [10:45] pitti, sorry was otp, it is just to collect them after the test [10:45] pitti, but it cannot work because conffiles = glob.glob('/tmp/*.dpkg-dist') is executed on the host :/ [10:47] jibel: yeah, I think with r94 it's more useful [10:48] http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-precise-trusty-server-tasks-amd64/30/console [10:48] jibel: of course now the tests fail due to the libgcc mess, but it also has the conffile prompt [10:48] jibel: no diff in this case as .dpkg-dist and /etc/default/bind9 are indeed identical, but I tested it on my workstation where I do have some changed conffiles [10:49] http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-precise-trusty-server-tasks-amd64/30/artifact/results/bootstrap.log is better [10:55] https://jenkins.qa.ubuntu.com/job/trusty-adt-php5/106/ looks like an infrastructure related issue. [10:55] Is there anything I should be doing about this, or anybody I should be pinging? Or is this something that someone already keeps an eye out for? [10:55] rbasak: already handled [10:55] Ah great. Thanks! [10:55] rbasak: it's because of the botched libgcc1 which breaks *everything* in trusty now [10:56] Ah [10:56] we'll wait for the fixed gccgo to go in, then retry everything [10:56] see #u-devel backscroll [10:56] * rbasak looks [10:56] rbasak: it's not just php, there are a bazillion failures [10:56] jibel is currently looking why it wasn't held back in -proposed === _salem is now known as salem_ [11:01] Thanks. That's an interesting/enlightening read. [11:27] jibel: darn, and I was soo close on http://d-jenkins.ubuntu-ci:8080/view/Upgrade/ [11:27] the latter two failures should work as well now, after libgcc [11:29] pitti, nice. and from the logs I cannot find why it libgcc has been promoted :/ now reading the code of britney [11:29] I: [Fri Feb 14 09:42:55 2014] - Requested autopkgtest for glib2.0_2.39.4-0ubuntu1 (NEW gccgo-4.9 1:4.9-20140213-0ubuntu1) [11:29] I: [Fri Feb 14 09:44:03 2014] - Collected autopkgtest status for glib2.0_2.39.4-0ubuntu1: RUNNING [11:29] Copying: gccgo-4.9/4.9-20140213-0ubuntu1 [11:30] for example, glib triggered by gccgo, correctly identified as running but gccgo has been copied anyway [12:03] jibel: I moved precise-i386.qcow2 away (to *.old) on alderamin, as they are ancient (may last year), and dist-upgrading them takes very long [12:03] jibel: I was hoping the job would rebuild them, but http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-precise-trusty-desktop-i386_vm/29/console is just sitting there without any noticeable progress (or started processes) [12:04] jibel: does precise-{i386,amd64}.qcow2 need to be maintained/updated manually then? (I can do that) [12:10] jibel: ah, I have a ubuntu-kvm-i386-precise/tmpDdXOIt.qcow2, apparently it's rebuilding now (just curious that I don't have a kvm/qemu process), so nevermind [12:11] pitti, no maintenance required, it is all created by vm-builder [12:13] pitti, although the resulting VM may require an update from time to time otherwise the dist-upgrade at the beginning of the test can take a while [12:13] I have a script somewhere to automate this [12:13] jibel: right, that's why I blew it away and let it rebuild [12:13] (well, I kept the previous .qcow until they get rebuilt) [12:14] ah, and there it is, a fresh precise-i386.qcow2 \o/ [12:14] not 12.04.2 but .4 now [12:25] hi all === bfiller_afk is now known as bfiller === ara is now known as Guest55777 === WebbyIT is now known as rpadovani [15:07] test [16:05] * balloons waves happy friday to all [16:27] balloons: haha don't forget the <3 [16:29] roadmr, :-) May cupid's arrow find it's way to all the loves in your life [16:29] thanks <3 === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [17:27] evening folks [17:27] g'evening :) [19:57] balloons, hi [20:45] Letozaf_, hello [20:47] balloons, hi I have a quite strange problem with reminders-app I uninstalled it and now I am unable to install it again I get: http://paste.ubuntu.com/6933106/ the fact is that I cannot find this qtdeclarative5-evernote0.1 anywhere, have you idea how to fix it ? [20:47] Letozaf_, apt-get update? [20:48] balloons, I tried and it does not work, I have been trying a lot of stuff but nothing worked [20:49] balloons, I was thinking to try on a new Trusty install on my notebook to see if it gets installed, because on my PC there is not way to get it installed [20:49] Letozaf_, likely the dependencies are not correct [20:49] so it's a packaging issue [20:49] it's pulling from the ppa [20:50] balloons, yes I know but I cannot find the way to fix this :( [20:50] so, you can force install it anyway, odds are it will work. We have to look at what the package is called now [20:53] balloons, ok I will try to force install === salem_ is now known as _salem [21:22] Letozaf_, any luck? [21:23] balloons, no :( [21:23] Letozaf_, ok, let's look @ the packaging [21:41] Letozaf_, https://launchpad.net/~ubuntu-touch-coreapps-drivers/+archive/daily?field.series_filter=trusty. you can see the version in there for the plugin [21:41] * Letozaf_ is looking at the link [21:45] balloons, I do not see this qtdeclarative5-evernote0 [21:45] balloons, and I have already the evernote plugin installed [21:45] account-plugin-evernote [21:46] ok, I removed the package and got the same issue [21:46] balloons, argh! [21:46] Letozaf_, it's a good thing [21:47] it confirms it's a packaging issue [21:47] balloons, good ? [21:47] bug filing time [21:47] balloons, oh! now I understand what you ment with good [21:47] balloons, ok so I will report a bug on this issue, ok? [21:52] balloons, bug 1280459 [21:52] bug 1280459 in Ubuntu Reminders app " unmet dependency when installing reminders-app" [Undecided,New] https://launchpad.net/bugs/1280459 [21:52] Letozaf_, yes :-) [22:17] balloons: we're getting a failure on an integration test on the latest image. any ideas on this? was previously working fine http://pastebin.ubuntu.com/6933841/ [22:17] autopilot test that is [22:21] bfiller, does the test use the uitk emulator? [22:22] balloons: let me check, here is the full log. getting tons of introspection errors http://pastebin.ubuntu.com/6933871/ [22:26] bfiller, the toolkit, the toolkit emulator, and autopilot have all changed within the past 24 hours [22:26] balloons: yup, this broke something [22:26] I take it the codebase did not right? [22:27] balloons: it's failing on this line: get_proxy_object_for_existing_process(pid) which looks like it's imported from from autopilot.introspection import get_proxy_object_for_existing_process [22:28] balloons: the code did change in camera-app, but nothing at all that would affect this test [22:28] balloons: but I'm getting blamed :) [22:29] bfiller, well first things first. Let's pull the old code and run with the new stuff and see what happens [22:29] balloons: apparently that works, rsalveti tried it. where can I get the previous camera-app package to try? [22:31] bfiller: http://ports.ubuntu.com/pool/universe/c/camera-app/ [22:31] rsalveti, bfiller I was going to checkout the old code and run it. [22:31] previous package is from 06-Nov [22:31] going to grab the code now, reproduce locally and go [22:31] rsalveti, oO why so old? [22:32] do we know what revno the package was built from? [22:32] bfiller might know :-) [22:32] hm, let me check the changelog [22:32] rsalveti: let me check trunk [22:32] trunk here btw https://code.launchpad.net/~phablet-team/camera-app/trunk [22:33] yeah, a lot was done that is test related since then [22:33] mostly porting to python3 [22:33] balloons: rev 225 [22:33] ahah [22:33] I see the problem [22:33] ahh I see omer did a py3 port in between [22:34] rev 229 [22:34] omer added the test that is failing now [22:34] so it was never being run previous to this release [22:34] but I did build from trunk and test it a few days ago and it was working [22:35] unittest.loader.ModuleImportFailure.camera_app.tests.test_gallery_integration [22:35] I pulled trunk and hmm.. ohh, unity 8 :-) [22:36] bfiller: just fix it lol [22:37] let's land a new version [22:37] haha [22:37] rsalveti: skip()!!!! [22:37] what [22:37] haha [22:41] balloons: when do you see that ModuleImportFailure error? [22:41] bfiller, trying to list the tests.. [22:43] balloons: shows up fine for me, you need unity8-autopilot installed as well [22:43] yes exactly [22:43] I did install build deps [22:43] *didn't [22:44] is webbrowser tests still ok? [22:44] they use the same calls [22:49] phablet only.. argh [22:50] balloons: browser seems ok [23:17] balloons: I need to EOD, I filed this bug if you figure anything out: https://bugs.launchpad.net/camera-app/+bug/1280485 [23:17] Launchpad bug 1280485 in camera-app "[autopilot] integration test failing" [High,New] [23:17] bfiller, :-) I'm in the same boat [23:17] balloons: in the mean time, I'm going to skip the test to unblock the image promotion [23:18] bfiller, since you filed a bug, no worries [23:18] balloons: thanks, have a good weekend [23:18] rsalveti is correct.. it looks like this never ran [23:18] and it's not systematic, so it's just camera [23:18] enjoy bfiller [23:23] balloons: ping === bfiller is now known as bfiller_bbiab === bfiller_bbiab is now known as bfiller