=== chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [06:24] correct me if I'm wrong, but testdrive in ubuntu 14.04 is broken to the point of unusability and I should try to use bzr trunk, yes? [07:08] balloons: ftr - new image - still see bug 1370707 [07:08] bug 1370707 in plymouth (Ubuntu) "Plymouth does not display the graphical boot splash" [Undecided,Confirmed] https://launchpad.net/bugs/1370707 [07:08] not had time to check in hardware though [07:18] so I was wrong and testdrive can be wrestled to submission in trusty === bipul_ is now known as bipul === spineau is now known as spineau_afk === chihchun is now known as chihchun_afk === _salem is now known as salem_ [13:41] elfy, bug 1370707 is still in new image yes? [13:41] bug 1370707 in plymouth (Ubuntu) "Plymouth does not display the graphical boot splash" [Undecided,Confirmed] https://launchpad.net/bugs/1370707 === chihchun_afk is now known as chihchun === salem_ is now known as _salem [15:04] balloons: I'll double check - got to do a hardware test shortly [15:10] ack, ty ty [15:11] I think I might possibly have hit a key to early in vm this morning [15:30] balloons: I has plymouth \o/ === chihchun is now known as chihchun_afk [15:34] elfy, :-) [15:34] rushing this morning I suspect === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [16:22] plars: good news! bug 1383878 [16:22] bug 1383878 in autopkgtest (Ubuntu) "make temp unpack mode work for dpkg-source" [Medium,Fix committed] https://launchpad.net/bugs/1383878 [16:22] plars: i. e. dpkg-source now works in "temp unpack" mode, test deb available (see bug) === chihchun is now known as chihchun_afk === roadmr is now known as roadmr_afk [17:06] we just had a world respin? [17:06] khave a link handy phillip ? [17:07] argh wrong channel :/ [17:07] wxl: ? [17:34] wxl, yes world respin [17:35] doesn't the world spin continuosly without stopping anyway? [17:35] knome, it's been suggested that indeed this is in the case [17:35] * balloons wonders how is it we don't fall off anyways? [17:36] balloons: full disk encryption or what's the deal? [17:36] you just haven't been close enough to the edge [17:36] wxl, yes, that's been fixed. I'm mailing the list [17:37] balloons: ok, well release is tomorrow so get your rear in gear XD === roadmr_afk is now known as roadmr [17:39] wxl, lol, it's done my friend === chihchun_afk is now known as chihchun [18:41] we're rebuilding again balloons ??? === chihchun is now known as chihchun_afk [18:43] :) [18:43] that's what i meant rafaellaguna [18:43] i was just so freaking mad [18:43] wrong channel [18:43] wxl [18:44] ianorlin: no [18:44] ianorlin: yes [18:44] why am i so good at this???? [18:44] * wxl bashes head against keyboard === chihchun_afk is now known as chihchun [19:06] balloons, hi === chihchun is now known as chihchun_afk [19:07] balloons, I used pitti's autopkgtest to run also clock app click package and have the same error as yesterday, should I go on to report a bug or have you guys made other changes meantime ? [19:10] evening Letozaf_ :) [19:10] elfy, hello :-) [19:18] Letozaf_: which "same error"? [19:19] pitti, hello, well if I run : adt-run ubuntu-clock-app/ --click com.ubuntu.clock_3.2.158_armhf.click --- ssh -s adb [19:19] Letozaf_: oh -- it seems you have a password other than 0000, so your test doesn't have root privs [19:19] pitti, I get: Cannot install /tmp/adt-run.ixZeib/com.ubuntu.clock_3.2.158_armhf.click: Cannot acquire permission to write to /opt/click.ubuntu.com; either run as root with --user, or use "pkcon install-local" instead [19:19] adt-run [21:18:00]: ERROR: unexpected error: click install failed with status 1 [19:19] pitti, oh! so my password should be 0000 :P [19:19] pitti, ok I will change it :P [19:19] Letozaf_: so you need to specify it by appending "-- -p s3kr1t" [19:20] pitti, oh I see thanks a lot [19:20] pitti, I will try again [19:20] Letozaf_: but it's a good point, if we don't have root we shuldn't install the click for all users [19:20] Letozaf_: so bug report is appreciated for that case [19:20] Letozaf_: as for most other things we don't really need root privs [19:21] pitti, ok so I will also report the bug [19:21] pitti, thanks [19:21] Letozaf_: thanks! in the meantime -p should work [19:21] Letozaf_: TBH I'm mostly testing running tests for already installed clicks [19:22] pitti, I don't as I have to test the tests that I am writing [19:22] Letozaf_: so thanks for being the guinea pig for testing local clicks :) (I do have automatic tests for that, but not for a r/o env aparently) [19:22] Letozaf_: yes, but the tests are in the source, not in the .click [19:22] pitti, my pleasure, it's also kind of fun [19:23] Letozaf_: i. e. your command runs the tests from the local ./ubuntu-clock-app/ [19:23] pitti, oh! got it [19:24] Letozaf_: i. e. if you don't change the .click, just the test, you could do: adt-run ubuntu-clock-app/ --click com.ubuntu.clock --- ... [19:24] pitti, now I understood [19:24] pitti, thanks that's very useful [19:24] Letozaf_: I should still fix that bug, of course, as a developer usually wants to test a locally changed .click against the existing or locally changed tests [19:26] pitti, yes right [19:27] Letozaf_: also, --click clock has a nice ring to it! [19:28] pitti, :) [19:38] pitti, fyi bug 1384417 [19:38] bug 1384417 in autopkgtest (Ubuntu) "Running adt-run without root installs the click for all users" [Undecided,New] https://launchpad.net/bugs/1384417 [19:46] \o/ it's working :-P [19:46] thanks pitti [19:54] Letozaf_: btw, if you test on an actual phone, can you make sure you are using autopkgtest 3.6git1 (utopic today) or http://people.canonical.com/~pitti/tmp/autopkgtest_3.6pitti2_all.deb ? [19:54] Letozaf_: 3.6 from utopic takes painfully long [19:55] plars: speaking of which, did you get my ping about this? [19:55] plars: (fixed the dpkg-source thingy) [19:55] pitti: I saw, is there a new package I can grab? [19:55] oh [19:55] nm [19:55] :) [19:55] plars: yeah, there's also a link in the bug [19:55] pitti, balloons gave me that link yesterday and today I am using http://people.canonical.com/~pitti/tmp/autopkgtest_3.6pitti2_all.deb :P [19:55] pitti: I had seen the bug but not a link to the built package [19:56] pitti: I'll try it shortly here. I did a bit of experimenting with creating a very minimal dep8 test in an unbuilt directory and running it, and it worked at least [19:56] pitti: I need to sort out how to make it provide useful results is all [20:00] plars: i. e. test things like "my testbed is in a state that I expect"? [20:01] plars: the things that come to my mind is doing the d-bus call to check whether unity is unlocked, and some pidof system-settings-bla to check that the wizard isn't running, and that you have a default route? [20:02] plars: or better, drop the default route and add a test Depends: pmount or something; if that's available (in the tmp path), then network and package (fake) install obviously work [20:02] pitti: I think so, I need to reinstall and start from scratch - I briefly saw that ssh error again earlier today but on the second run it passed (new cable too) so I was wondering if it's similar to the thing you saw that had to run twice [20:03] plars: no, for me that wasn't an ssh error, it looked like a genuine testfailure [20:03] plars: i. e. I saw calculator starting, stopping, starting etc., but no events in calculator [20:03] and thus test failures [20:07] Letozaf_: FTR, as far as I remember you still can't install a click even for one user only without root privs, so fixing that might be difficult === chihchun_afk is now known as chihchun [20:08] pitti, ok [20:11] pitti: for tests that require autopilot, if autopilot is not in the image already, istr you saying that it would still work somehow? does it pull it in as a test dependency without actually having to install autopilot-touch and dependencies? [20:13] plars: yes, that's the "unpack into /tmp" mode [20:13] pitti: great! [20:13] plars: any test Depends: will be downloaded, unpacked into /tmp/, and then the test runs with $PATH, $LD_LIBRARY_PATH, $PYTHONPATH etc. set accordingly [20:14] plars: of course that's much less robust than actually installing them (for example I haven't yet covered udev rules) [20:14] right [20:14] plars: but for python modules and the like it should be good enough [20:14] plars: it obviously fails for packages which are looking for plugins in /usr/lib/ etc. [20:15] but oh well, don't run them on a r/o image then :) [20:19] pitti: well, libpng passed (build test is all it has of course), but unfortunately it looks like none of the packages have the dep8 tests in them (unity8, ubuntu-ui-toolkit, webbrowser-app, etc) [20:19] so I can't test it easily with something autopilot-y [20:19] plars: yeah :/ [20:19] pitti: but we know libpng wasn't working yesterday, and it is now [20:19] plars: well, you can cheat a bit [20:19] I should try a clean system too though, mine is installed rw at the moment [20:20] plars: oh? that's actually surprising; it should fail on gcc not finding libraries [20:20] plars: ah yes, then it actually ran apt-get install [20:20] yay! autopilot! [20:20] sorry [20:20] * pitti ^5s thomi [20:21] plars: so, you can create a local fake source package (with pretty much just debian/tests/) which depends: on e. g. unity8 and other test deps, and then put the autopilot call into Test-Command: === chihchun is now known as chihchun_afk [20:21] plars: while it's customary to put tests for binary package X into the source package of X this isn't a technical requirement :) [20:21] hey guys anyone testing Ubuntu Desktop amd65 ISO on Virtualbox ? on my PC Virtualbox crashes [20:22] oh, where can I buy an amd65? :-) [20:22] :( [20:22] amd64 [20:22] wrong typing [20:22] :P [20:22] Letozaf_: davmor2 says the same thing [20:23] pitti: right, but that seemed to work before even - running out of an ubuilt tree [20:23] plars: yes, sure (that has always worked) [20:23] pitti, oh fine, well not fine, but at least someone confirms my problem [20:24] works quite ok in kvm [20:25] so it's Virbualbox [20:26] I'll check our's quickly now [20:27] you want "-vga vmware" with kvm, BTW (otherwise the background image is quite broken; it's not that important, but looks weird) [20:31] my virtualbox crash has been uploaded automatically so I do not think thers is a bug to be reported, right ? === chihchun_afk is now known as chihchun === _salem is now known as salem_ [21:16] vila: des bonnes nouvelles ! http://anonscm.debian.org/cgit/autopkgtest/autopkgtest.git/commit/?id=0dc5fe [21:16] vila: should I build a .deb for you, or do you run out of git? [21:17] pitti: yeehaa ! [21:18] pitti: the tmux... funnily enough you *were* suspicious about it ... [21:18] vila: no, it's not tmux, it's ssh's connection sharing muxer [21:18] ha, another one then, ack [21:19] so, to get the fix, the process is to pull, recipe, etc, I'll do it right now [21:19] vila: ah right, your magic PPA [21:19] vila: I'm building a .deb, too [21:20] pitti: s/magic/gating/ ;) [21:27] vila: (in case you need: http://people.canonical.com/~pitti/tmp/autopkgtest_3.6pitti3_all.deb) [21:27] pitti: thanks [21:32] elfy: you around? [21:33] just about for a short while longer [21:33] can anyone not solely with ubuntu explain their reasoning not to use the desktop encryption testcase that k/ubuntu uses? http://iso.qa.ubuntu.com/qatracker/testcases/1451/info [21:33] elfy ^ [21:34] well ... [21:34] the reason is just that we don't - never have afaik [21:36] figured as such, thanks :) [21:36] afair there was a time when that option was dropped from the GUI [21:36] so it was only available from the alternate installer for a while [21:37] given we have enough trouble getting test on what we already have I'd not be happy about adding anything === salem_ is now known as _salem [21:39] practically any option that the ubiquity installer provides is tested if it's tested on one flavor === chihchun is now known as chihchun_afk [21:40] that's a very b&w way to look at it, but there are hardly any reasons why some option would be broken in a flavor but not the other === chihchun_afk is now known as chihchun [21:56] pitti: eeerk: https://launchpadlibrarian.net/187924725/buildlog_ubuntu-precise-i386.autopkgtest_3.7-0~966~ubuntu12.04.1_FAILEDTOBUILD.txt.gz [21:56] pitti: from https://code.launchpad.net/~canonical-ci-engineering/+recipe/autopkgtest-phase-0 [21:56] pitti: only the precise one really matter though [21:57] pitti: but I guess you don't want the trusty one to fail either ;) === chihchun is now known as chihchun_afk