=== chihchun_afk is now known as chihchun === _salem is now known as salem_ [03:15] Good morning [03:15] balloons: I think for apt-get update temporarily switching to r/w isn't too bad === chihchun is now known as chihchun_afk === salem_ is now known as _salem === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [06:19] vila: bonjour [06:19] vila: not sure whether you saw, but autopkgtest 3.2.2 is in utopic since yesterday [06:19] vila: so now, depending on >= 3.2 is officially correct :) [06:42] pitti: hehe, I saw 3.2 then 3.2.1 and was willing to upgrade indeed ;) [07:12] vila: don't use 3.2; that's not the version you are looking for :) [07:12] 3.2.2 is good [07:12] vila: 3.2 had a giant boo-boo which unfortunately was hidden by how I test; yay PYTHONPATH including '.' by default :) [07:12] pitti: yeah, just noticed the changelog for 3.2.2 and will aim for that [07:12] I now wrote a test to avoid this [07:12] I mean s/avoid/detect/ [07:13] urgh, I hate that one ( '.' in PYTHONPATH) but confess I sometimes rely on it as it's handy too ;-) [07:14] pitti: I was wondering (a tiny bit) if it would be worth for me to test 3.2.1 and see if my tests detect the issue with --apt-source fixed by 3.2.2 [07:17] vila: yeah, it's both incredibly useful and sometimes also incredibly dangerous :) [07:18] vila: you use libpng, right? that doesn't expose the apt-source bug, only packages with "build-needed" do [07:18] vila: but sure, covering that case is certainly useful [07:18] pitti: thanks, no need for the experiment then, good ;) === chihchun_afk is now known as chihchun [07:54] vila: I saw that your recent branch has landed, that means CI is using current autopkgtest now? [07:55] vila: with the cloud instances and ssh? [07:56] pitti: yes and more in the pipe [07:56] nice! [07:57] pitti: the current version has a design bug and works for the wrong reasons ;) [07:57] vila: how is the HP cloud stuff holding up so far? do you see some instability there with creating VMs, etc? [07:57] pitti: well, current as in trunk [07:57] pitti: still some transient errors in various places [07:58] pitti: but all diagnosed as being caused by HP [07:58] vila: right, but I suppose we still need to add workarounds/robustifications on the uci-vms end? [07:58] pitti: yup, but the core features work === pitti is now known as somedude === somedude is now known as pitti [13:44] elopio, hello [14:13] hey pitti.. yea, I see you saw the bug.. I was trying out the emulator last night with Jason who's trying to use it on trusty for testing with adt [14:14] balloons: hey; nice to know that the emulator works with the adb setup script in general! [14:15] pitti, yea, on my version I had trouble getting ssh to work. I couldn't ssh as phablet, and therefore in order to run adt, I had to run as root as well [14:15] Jason didn't have the issue.. [14:15] That's one for the sdk team I think [14:16] so it's just a packaging issue.. that's nice to know [14:29] brendand: hi! [14:38] balloons, elopio: so how long do I have for the autopkgtest lightning talk? i. e. into how much detail can I go? [14:39] I'll let elopio answer, but 10-15 mins I believe. it's 30 mins in total, 2 speakers [14:39] pitti: you have 15 minutes. And our host jose is really serious about that. [14:39] elopio: ack, thanks [14:39] should be exciting stuff [14:40] pitti: if you want you can talk for ~12 minutes and take a couple of questions at the end, or use all the 15 minutes and hang around after the talk to get questions on IRC. [14:41] elopio: sure; I'll try to keep it to 10 mins, to have room for some overflow [14:41] jose, elopio: you'll give me the hangout URL some minutes before, and we sync up off the record? [14:42] pitti: yes, jose takes care of all that. [14:47] elopio, we need to get those unity8 branches landing. anything blocking it? i see CI failures [14:48] brendand: yes, but unrelated. [14:48] elopio, also i was able to reproduce the CI failures in dialer-app [14:48] some caused by the qmlscene crash. Some are just weird. [14:48] elopio, i didn't investigate them much though [14:48] I'll try to get a green run, shouldn't be too hard. [14:49] brendand: take a look at how we are using the env vars here: http://bazaar.launchpad.net/~ubuntuone-control-tower/unity-scope-click/trunk/view/head:/autopilot/CMakeLists.txt [14:50] it might be different on your url-dispatcher branch, because these are not ctests. But it might help. [14:52] elopio, i got a solution which works [14:52] awesome. [14:59] brendand: oh, you meant the py3 dialer errors. [15:00] I didn't get back to it yesterday. [15:00] I'll grab barry's updates and see how it goes. Thanks for looking at it. [15:53] oO first builds coming up [16:11] elopio, can you re-review https://code.launchpad.net/~canonical-platform-qa/url-dispatcher/fake_dispatcher/+merge/227778 [16:11] balloons, did dpm get on to you about the calendar-app AP failures? [16:11] brendand, which ones? [16:12] balloons, from this bug: https://bugs.launchpad.net/ubuntu-calendar-app/+bug/1343916 [16:12] I assume dashboard [16:14] brendand, I hadn't seen the bug. the new event test is being tweaked a bit sitting in some mp's [16:14] balloons, yes dashboard failures [16:14] I suppose this could be simple enough to do a one-off mp for [16:18] brendand: sure. [16:23] brendand: I don't like the .in solution, because that means you can only run the tests from cmake. [16:24] but I approved it. If ted likes it, I'm happy. [16:25] elopio, i wish there was a better way - cmake is really not flexible though [16:26] brendand: yes. dobey is good at this, maybe he know about a better way. [16:26] but lets leave that for the future. [17:03] hmm can't get to the testing tracker for testcases [17:03] ianorlin, same.. hmm [17:05] I'll have someone check [17:11] ianorlin, back up :-) Thanks for pinging [18:00] balloons: zsync on http://iso.qa.ubuntu.com/qatracker/milestones/318/builds reports [18:00] phillw@piglet:~/Desktop/ISOs$ zsync http://cdimage.ubuntu.com/lubuntu/daily/20140722/trusty-alternate-amd64.iso.zsync [18:00] failed on url http://cdimage.ubuntu.com/lubuntu/daily/20140722/trusty-alternate-amd64.iso.zsync [18:00] could not read control file from URL http://cdimage.ubuntu.com/lubuntu/daily/20140722/trusty-alternate-amd64.iso.zsync [18:03] phillw, http://cdimage.ubuntu.com/lubuntu/trusty/daily/current/trusty-alternate-amd64.iso.zsync [18:03] balloons: also, F.Y.I., http://iso.qa.ubuntu.com/qatracker/milestones/318/builds only shows lubuntu at present.. no other flavours? Has the build had a hiccup ? [18:03] they are still working on the rest [18:03] balloons: iso tracker down? [18:03] okies.. thanks... [18:03] it was, it should be up now [18:04] zync for amd64 lubuntu now working.. thanks! We don't have a lot of time :D === roadmr is now known as roadmr_afk [18:24] hmm what package should a bug be reported agianst if virtualbox or something freezes when you get to the restart now prompt? [18:24] ianorlin, that's a known bug [18:25] ah which number? [18:27] ianorlin, ohh wait, that sounds different I'm sorry.. You mean the ubiquity prompt to restart now or continue working? [18:27] if so, ubiquity is your answer, and it's not filed afaik [18:28] ianorlin: press enter? [18:29] pressing enter didn't work and didn't get the prompt to press enter [18:29] ianorlin: how did you get access to test results? [18:30] the system boots fine after but would not respond to keyboard at all [18:30] phillw is iso.qa.ubuntu.com still down to you? [18:31] ballons said he got the site back up [18:31] and it came back up for me [18:31] it's hangiing on by finger tips... I cannot state that I'm doing tests [18:32] phillw are you logged in? [18:32] also, I see all the alternates have just been respun [18:32] ianorlin: ,,, yes :P [18:33] ianorlin: if tbings settle down, I'll grab the amd64 altrnate ones... feel free to pick any other :) [18:35] ianorlin: I've got them marked as being tested by me. [18:42] I am spinning up these images as fast as I can also. Why are people jumping right on this? Most days I consider running through the daily's but I see that noone else seems to care. [18:42] just trying to figure out when to test [18:43] average_guy, during milestones like this all the tests need to be run in order to feed into a thumbs up or thumbs down on releasing [18:43] ubuntu-qa: can I get a review here please? https://code.launchpad.net/~canonical-platform-qa/messaging-app/qmltests1/+merge/227661 [18:44] average_guy, this might help: https://wiki.ubuntu.com/QATeam/Overview/TheStages [18:44] ty balloons, I will check it out [18:45] average_guy, basically if you look at the release schedule or our calendar https://wiki.ubuntu.com/QATeam/Calendar, the milestones are marked [18:45] releases are on thursday's generally, and the days leading up to thursday the same week are reserved for testing [18:49] balloons: that text on TheStages has a familiar feeling :D [18:52] :-) [18:55] balloons: As lubuntu now has an LTS, do you want a mention adding for LTS point testing? [18:55] Or, would prefer to do task it to others? [18:55] Letozaf_, hello. I figured out my mistake in telling you the command last night [18:55] balloons, hello :) [18:56] Letozaf_, the click argument is --click :-) [18:56] Letozaf_, so like adt-run -d ubuntu-clock-app/ --click=com.ubuntu.clock --- ssh -s /usr/share/autopkgtest/ssh-setup/adb [18:56] balloons, let me try [18:56] phillw, for the wiki page? You could talk about point releases sure [18:56] not a bad idea [19:01] elopio: did you find a reviewer? [19:02] balloons, yes it works on the clock app, the one I got form lp:ubuntu-clock-app [19:03] balloons, I want to try this on the new clock-app do you think it will work ? [19:03] Letozaf_, yep should be golden [19:03] Letozaf_, now I assume the new app has a seperate click as well right? [19:03] you'll want to run the tests against the right package [19:04] so probably build it locally, and then tell autopkgtest to use it [19:04] balloons, yes when I tried the old clock app used to run and not the new oen [19:04] so instead of --click=com.ubuntu.clock, pass instead the path to the click you build locally [19:05] balloons, ah ok thanks I will try [19:24] balloons, is this stilla ok to use http://www.theorangenotebook.com/2014/05/building-cross-platform-click-packages.html ? I used pbuilder-dist on chroot got the click pagkage, but it does not work :( [19:25] still not stilla [19:25] Letozaf_, the original clock has no native code.. does the rewrite have some native code? [19:25] if it doesn't you can build locally ofc.. [19:26] but yes, building in a chroot works fine [19:26] you'll want a utopic chroot now [19:26] balloons, yes I used utopic chroot [19:27] well, I guess I need context on what is or isn't working ;-) But sounds like it should work fine [19:27] balloons, I will first try again, if it doesn't work I will tell you whats wrong [19:29] Letozaf_, feel free to paste your output [19:29] so I can see what's up [19:30] balloons, yes thanks, let me try it again, so I will be sure it's not a silly error I made [19:39] elopio, ping ;-) I meant to ask you yesterday, but reminders doesn't run on the devices again.. that same error with accounts not creating I believe === roadmr_afk is now known as roadmr [19:48] balloons, http://paste.ubuntu.com/7837995/ [19:49] Letozaf_, you have an extra space in there [19:49] balloons, :( let me check [19:49] adt-run -d reboot/ --click= com. [19:49] but also the syntax is wrong.. if you pass a click file, just pass the filename [19:49] adt-run -d reboot/ com.ubuntu.clock.devel_0.1_all.click --- ssh -s /usr/share/autopkgtest/ssh-setup/adb [19:50] balloons, thanks let me try again :P [19:50] :-) [19:52] balloons, :( I probably built a wrong click package: SKIP no tests in this package [19:53] balloons, I ran: adt-run -d reboot/ com.ubuntu.clock.devel_0.1_all.click --- ssh -s /usr/share/autopkgtest/ssh-setup/adb [19:54] Letozaf_, never fear [19:54] ok, so the tests are controlled by the manifests file [19:55] * Letozaf_ is looking at it [19:58] balloons, looks ok to me, but I am not "an expert" [19:58] Letozaf_, can you link it? [19:59] balloons, how ? [20:00] Letozaf_, is this it? [20:00] http://bazaar.launchpad.net/~ubuntu-clock-dev/ubuntu-clock-app/utopic-3.0/view/head:/manifest.json [20:00] balloons, yes it is [20:01] Letozaf_, yep so it needs to have the x-test key in it [20:01] notice http://bazaar.launchpad.net/~ubuntu-clock-dev/ubuntu-clock-app/trunk/view/head:/click/manifest.json.in [20:02] balloons, ok thanks, just the x-test ? [20:02] Letozaf_, yes and put the actually value of the folder in there as well [20:02] don't use the @AUTOPILOT_DIR@.. that's a variable for cmake and it won't resolve [20:03] balloons, ok let me try [20:03] so the folder name under tests/autopilot [20:03] put that in there instead [20:03] ubuntu-qa any takers https://code.launchpad.net/~canonical-platform-qa/address-book-service/test_dummy_service/+merge/225846 ? It implements a standard contact fixture for all your integration tests [20:04] balloons, thanks I will now and retry to build [20:04] Letozaf_, you don't have to rebuild [20:04] that's the nice part, you can just specify the new manifest with the click [20:04] balloons, you are righe :P [20:04] martin did that to save time ;-) [20:04] right [20:04] balloons, cool [20:05] Letozaf_, I just have to remember how to do it [20:05] ;p [20:06] balloons, :) [20:10] Letozaf_, ok, so it's --override-control [20:11] balloons, in the adt-run command ? [20:11] Letozaf_, yes, so add that on and point it at the new manifest file you made [20:12] Letozaf_, so adt-run -d reboot/ --override-control=reboot/manifest.json --click= com.ubuntu.clock.devel_0.1_all.click --- ssh -s /usr/share/autopkgtest/ssh-setup/adb [20:12] ? [20:16] balloons, http://paste.ubuntu.com/7838168/ [20:17] Letozaf_, no --click= [20:17] balloons, ok, thanks again, let's try [20:17] Letozaf_, sorry I realized I told you incorrectly :-) [20:17] copy paste fail [20:18] balloons, doesn't matter I should have read, I also copied and pasted :P [20:20] there's alot in there to read [20:21] balloons, http://paste.ubuntu.com/7838177/ :( [20:22] Letozaf_, can I see the manifest? [20:23] balloons, sure... http://paste.ubuntu.com/7838186/ [20:23] Letozaf_, "autopilot": "tests/autopilot" needs to reflect the module to test [20:23] what's inside tests/autopilot? [20:24] balloons, ubuntu_clock_app [20:24] for instance, for the old version of clock, the testsuite is ubuntu_clock_app [20:25] Letozaf_, ok then "autopilot": "ubuntu_clock_app" [20:25] one more try :-) [20:25] balloons, oh! I misunderstood... ok let's try again [20:28] balloons, it ran, but I get the old clock app running on device the one with the purple background instead of white :O [20:28] Letozaf_, output again ;-) [20:28] balloons, ok, and thanks for you patience [20:28] Letozaf_, may I suggest you name your testsuite something else [20:28] that will give sanity anyway [20:29] balloons, ok [20:29] rename the folder under tests/autopilot to clock_reboot and update the manifest as well [20:29] balloons, ok I will [20:32] balloons, I will have to rebuild now, as I changed the testsuite's name [20:32] Letozaf_, you shouldn't. change the manifest and run again [20:33] balloons, ok let me do it [20:35] balloons, the full output is long but these File "/tmp/adt-run.gZTjWx/tree/tests/autopilot/clock_reboot/tests/__init__.py", line 34, in [20:35] from ubuntu_clock_app import emulators [20:35] ImportError: No module named 'ubuntu_clock_app' [20:35] balloons, suggest that I have something in the code to check and rename too [20:36] Letozaf_, do the tests run locally? since I had you rename the testsuite, yea I made it rough for you ;-) [20:36] i guess you can undo that.. I'd still like to see the original paste [20:36] where you say the wrong version ran [20:37] balloons, ok, let me "roll back" [20:40] balloons, I couldn't get the whole output, this is what I could grab: http://paste.ubuntu.com/7838369/ [20:40] Letozaf_, try running with >& dump at the end to capture it [20:41] the top is the useful bit ;-) [20:41] balloons, ok :) [20:43] balloons, http://paste.ubuntu.com/7838392/ [20:46] Letozaf_, ok so the trouble is ubuntu_clock_app testsuite already exists.. bah, I think I might be correct you should name it something else [20:47] balloons, ok I will rename it and check the code and rebuild [20:47] Letozaf_, well [20:48] is the binary also uniquely named? [20:48] the tests do appear to launch from the tmp dir [20:48] Loading tests from: /tmp/adt-run.yWY2ti/tree/tests/autopilot [20:48] so inside the test, are you calling the new app or the old one? [20:49] balloons, let me check [20:49] Letozaf_, yea, check tests/__init__.py [20:49] that's probably the actual issue :-) The log doesn't lie, so I think the test name is fine [20:51] balloons, def launch_test_click(self): [20:51] self.app = self.launch_click_package( [20:51] "com.ubuntu.clock", [20:51] emulator_base=toolkit_emulators.UbuntuUIToolkitEmulatorBase) [20:52] balloons, must be this [20:53] balloons, it launches com.ubuntu.clock, must be the old one that's on the device [20:54] om26er: I'll review it. [20:54] please review mine. [20:54] aha trading [20:54] elopio, which one ? [20:55] Letozaf_, yep.. that's it. fix that to be the proper one [20:55] om26er: https://code.launchpad.net/~canonical-platform-qa/ubuntu-ui-toolkit/fix1342521-find_element_on_top/+merge/226961 [20:55] try again.. [20:55] again, no rebuild should be needed :-0 [20:57] balloons, but the new one has the same name com.ubuntu.clock or am I worng ? [20:57] wrong [20:57] balloons, unless I change the name [20:58] Letozaf_, it shouldn't have the same name [20:58] look at manifest [20:58] com.ubuntu.clock.devel looks like [20:58] balloons, yes I saw it now [20:59] balloons, com.ubuntu.clock.devel [21:00] Letozaf_, yep so just update that to launch the right one [21:01] balloons, ok I am fixing and I will rebuild [21:02] Letozaf_, shouldn't have to rebuild for test change [21:02] balloons, :-) ok so I will just re-launch [21:03] elopio, on it. [21:04] balloons, yay!!! \o/ [21:04] balloons, it worked [21:04] balloons, thanks a lot [21:04] Letozaf_, awesome! [21:04] you are an expert now [21:05] balloons, :) I must thank you for your patience in helping me out [21:05] no worries.. I trust it makes more sense now [21:05] balloons, yes I am sort of an expert now [21:05] balloons, yes it does [21:05] balloons, ok I better go off to bed now :P [21:06] balloons, buona notte ! [21:07] Letozaf_, buona notte! === salem_ is now known as _salem [21:42] balloons: just how much have things been brocken? I cannot even sign on with phillw@uk.vpolink.com to give results? [21:42] balloons: No account associated with phillw@uk.ubuntu.com [21:43] which happens to be my master account for iso.tracker [21:43] phillw, tracker uses ubuntu sso [21:44] balloons: did they change that this evening? [21:44] phillw, it's used it since as long as I can remember [21:44] several cycles [21:44] balloons: I've used SSO long before ubuntu sso arrived.... [21:45] But, there you go.. not allowed to sign on... ce'est la vie. [21:45] I would log into your sso account [21:45] odds are you can't.. fix it and you'll get int [21:46] not quite sure what else they want to break on iso tracker... [22:01] balloons: how did desktop amd64 build for lubuntu vanish?... While you look into that, i'll go grab another ISI [22:01] *ISO* [22:16] balloons: are you seeing the rest of the ISO's at http://iso.qa.ubuntu.com/qatracker/milestones/318/builds for lubuntu? [22:23] om26er: shouldn't you clean up the dummy backend fixture to return to the real backend? [22:25] elopio, I am killing the service in the end. Any app that starts after that spawns the service [22:26] om26er: ah, I see it. You do it on the other fixture. [22:28] om26er: on SetupEnvironmentVariable, it's not clear that you can also set up some demo data. [22:28] I have no idea how to make it clear with code, so a docstring for the class is probably a good idea. [22:30] om26er: I like your branch, I love these new testability packages in python. [22:30] om26er: who is the other one affected by this? brendan or richard? [22:31] I think it would be nice if you request a review from him, so you all agree on the strategy. [22:37] elopio, brendan depends on that