=== vrruiz_ is now known as rvr_ === rvr_ is now known as rvr [05:14] veebers: you are working on the nfss data file for app-startup, that's great. [05:14] I thought I would had to do it tomorrow. [05:14] you are kicking ass. [05:15] your email makes sense. I was hoping to upload one observation per test, but thomi doesn't seem to like that. [05:16] and as the project was already sending data, I think that for now we should stick with the same agregated format. [05:17] I hope tomorrow by the time you wake up I'll have the tests running and collecting the data, so we can just pass it through the export script. [05:21] elopio: awesome. How I understand it is that CI needs the results aggregated into a single file, so they will still be set per test, but it's just one file [05:22] elopio: if you could review the 2 MPs that I have for my cards that would be great, then I can move them to Done in prep. for Fri morn [05:22] veebers: yes, I get that part. What I mean is that thomi is suggesting on the document that on the dep8 tests we do a loop of 10 tests or so, and send to nfss the agregated results of those 10 tests. [05:22] oh, I don't know anything about that sorry :-P [05:24] veebers: I've just left my comments on two of your branches. Am I missing one? I [05:24] 'm not sure which are the 2 you are referring to. [05:24] elopio: no that should be it, let me check [05:25] elopio: hah, damn ipdb :-\ [05:27] veebers: that's why I commented about testing the scripts. A simple and quick test would have caught that. I just don't know where to put it. [05:28] elopio: yeah, I wrote tests for the memevent one, but after that felt it was out of scope [05:28] but you're right, tests for that would be nice [05:28] I'll see what I can come up with tomorrow [05:29] elopio: what's your reason for wanting single quotes over double? (I understand having it consistent, which I haven't here) [05:31] veebers: that's what most people on QA preferred, according to my quick poll some months ago. As long as it's consistent, I'm ok. But I would love consistency between all our projects. [05:32] ack fair enough [05:38] elopio: I have made your suggested changes on both those MPs, I'll add a readme tomorrow, but for now I'm off for the night [05:40] elopio: one request, with the app-startup tests, can you provide the results to me in an email or something, so I don't have to spend time running the tests etc. just to get results to test the nfss-generator script against? [05:50] veebers: sure, no problem. [05:51] let me see what I get tomorrow. [06:11] awesome, cheers leo [06:11] * veebers is out o/ [06:19] Good morning [10:02] good morning pitti [10:03] balloons: i just froze unity-next in preparation for the UOS session :p [10:04] hey dkessel [10:04] balloons is going to be "I hate the day after a day off" today ;) [10:04] morning pitti [10:04] heh [13:23] elfy, :-) [13:23] good morning! [13:23] hey balloons - welcome to day 1 of your 3 day nightmare :D [13:24] I do believe yesterday was day 1 :p [13:24] I'm going to try to get to the manual testing session - but the mention of phones fills me with sense of depreciating interest ;) [13:25] there's a few I'll try to get to this week [13:29] balloons: thanks for doing that will cooke thing - couldn't see any way to do that - at all, I looked for the manual but couldn't find it :p [13:29] Hi! where I can find something about old UDS ? i.e. 14.06. [13:31] stblack, was it online? [13:31] i think it is still there: http://summit.ubuntu.com/uos-1406/ [13:31] elfy, heh, phones are something we should consider. Most of the "real" discussion will be on the tracker [13:32] stblack, yep dkessel is correct. If it was an online summit, it's still there; the links should still work and the videos too [13:32] balloons: when I have a phone I'll consider it :) couldn't get involved in testcases for it - I'd have to guess :p [13:35] thanks balloons, I was wrongly looking for uds-1406, not uos. [15:52] balloons, hi [15:59] Letozaf_, hello [15:59] Letozaf_, joining fm session? [16:00] balloons, yes [16:01] balloons, sorry but I arrived late at yesterday's filemanager meeting, think I got the wrong time [16:01] Letozaf_, I was away all day national holiday for me, so I certainly missed :-) [16:02] you can ask in the hangout about times for meeting [16:02] balloons, yes [16:02] pitti may I have a few minutes with you before your EOD? I'm trying to launch an app on device (via adt) as root (needs to run as root), and getting a d-bus denial--attempting to su to a normal user isn't helping, any advice? http://pastebin.ubuntu.com/8929277 === chihchun is now known as chihchun_afk [16:03] alesage: oh, you put the device into r/w mode? [16:03] pitti I believe so, yes--should I put the device into r/w mode? [16:04] alesage: well, preferably not, but if your test wants to change things in /etc/init, you have to [16:04] alesage: anyway, 3.6git1 doesn't work with current touch images any more, can you please use 3.7.1 from vivid? [16:05] alesage: http://archive.ubuntu.com/ubuntu/pool/main/a/autopkgtest/autopkgtest_3.7.1_all.deb [16:05] pitti, I am cleaning up /var/crash and so need r/w [16:05] alesage: you can install that .deb from precise upwards [16:05] alesage: /var/crash/ shoudl be writable anyway, though [16:05] pitti ok will verify [16:05] pitti I'm hearing you say "should work", yes? :) [16:06] alesage: I don't know your particular test, and presumably su etc. will cause some trouble, but let's at least rule out that it's due to something I already fixed [16:07] pitti ok thanks, let me have another go [16:08] alesage: also, your second sed command is likely broken -- it won't expand the local $CRASH_ID from the script, but instead will add $CRASH_ID verbatim to /etc/init/whoopsie.conf [16:08] alesage: powerd-cli display on bright -> that bit is already taken care of by the adb setup script, you can drop it [16:09] pitti ok for mini-review, will amend [16:43] pitti ok this is on a fresh image, no r/w, 3.7git1--this is essentially what I need to do in my adt script: http://pastebin.ubuntu.com/8967208/ --is there a more correct way to launch an app from within autopilot? [16:43] pitti (i.e. more correct than using ubuntu-app-launch) [16:44] alesage: yeah, you can't run apps as root, they don't have an upstart session running [16:44] alesage: that needs to happen as phablet [16:45] pitti would su'ing over give me the right env to do so, do you think? [16:45] alesage: I'm not sure, I suppose through the sudo/su/su you are losing env variables, but it's worth a try [16:46] pitti confirmed, yes I'm able to do (with --login) [16:46] alesage: the other q is, why do you need root privs in the first place? [16:46] pitti well b/c of the whoopsie tinkering [16:47] pitti possibly we have a different schema for that, handing over to CI, will verify [16:47] pitti thanks for the coaching [16:47] alesage: the whoopsie bits could be done in --setup-commands, and then your actual test would jsut run as user [16:48] pitti hmm ok thx will look into [16:48] alesage: calling sudo out of your test makes it very specific to current touch, and you somehow need to put the passphrase into the test (and then assume that whoever runs the test uses that) [16:49] sorry, need to run out, time for sports [16:49] pitti all of this is true, yes--agreed it's better as user, will depend on how CI prefers to set up (will ask in standup today) [16:49] pitti enjoy! === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [18:11] knome, will you be about for http://summit.ubuntu.com/uos-1411/meeting/22340/improving-manual-testing/ which is in 50 mins? [18:11] knome, I'm thinking about showing the UI mockup you did again. thoughts? === chihchun is now known as chihchun_afk [18:43] balloons: I'm kind of going to be about for that session [18:43] elfy, awesome. I put the mockup in the notes [18:43] yea - read those ;) [18:44] got someone visiting at some point - and I've got a wicked cough so might be just on irc [19:00] balloons, i'm here [19:00] brendand, awesome, check pm [20:06] balloons, have you got time for filemanager app now ? [20:07] Letozaf_, I saw that it didn't work for you.. but it works locally? in both phone and tablet mode? [20:07] Letozaf_, btw, it would be cool to have launchers for both so the tests are duplicated and run during jenkins [20:07] balloons, I tested the app on the desktop and on my nexus4 [20:08] balloons, the tests paseed but on jenkins one test fials [20:08] fails [20:08] balloons, the error is quite weired to me [20:09] balloons, do you mean launch the tests with the -p and -b swtich ? [20:09] switches [20:11] balloons, by the way if I launch filemanager with the -p switch on desktop, the places side bar is displayed have to change the dimensioin of the window to the phone one to get the places side bar to hide and get the bottom edge for places [20:12] dimension [20:12] Letozaf_, are you saying the -p doesn't work properly? [20:12] Letozaf_, and yes I mean run the tests under each mode [20:12] another MP though, not this one :-) [20:12] balloons, ok [20:13] balloons, I would expect the filemanager app to open with the phone dimensions with -p [20:13] balloons, but it is a bit bigger and has the places side bar visible [20:14] Letozaf_, ok so that's probably something worth asking arto about. Did you talk about it today? [20:14] ideally it would be exactly the same. [20:14] balloons, nope as I wasn't sure if I was getting this wrogn [20:14] wrong [20:14] Anyways, to your error, if it passes locally that's confusing to me why it's not passing in jenkins [20:15] balloons, http://91.189.93.70:8080/job/generic-mediumtests-vivid/156/testReport/junit/filemanager.tests.test_places/PlacesTestCase/test_go_to_root_must_open_the_root_directory/ [20:15] the MismatchError: After 10.0 seconds test failed: '/' != '/tmp/tmp2tsfl4lb' [20:15] balloons, why is it in /tmp/tmp2tsfl4lb and not / [20:15] balloons, is that right ? [20:16] balloons, by the why with the -p switch you can read in the logs that the app is running in phone mode, but it's not what you seee [20:17] balloons, Patched home to fake home directory /tmp/tmp2tsfl4lb [20:17] balloons, shouldn't it be in / [20:18] Letozaf_, let me look at the test since you updated it [20:18] so I can be correct in my thoughts :-) [20:20] balloons, if I run the app on my device and go to device place I go to / shouldn't that be the same on Jenkins? [20:20] balloons, or maybe you use a temp directory for moking... [20:21] Letozaf_, right that's the issue [20:21] * balloons pulls code [20:21] Letozaf_, I think the only thing needing changed is what the assert looks for [20:21] it should look for the mocked / when it's mocked [20:22] balloons, ok, so I got it wrong :P I thought that only home was mocked, I will look at the code again and figure this out [20:24] Letozaf_, ok I have the code [20:25] the difference between your local version and jenkins is that jenkins uses the debian package version [20:25] otherwise they are the same [20:25] balloons, ah ok [20:29] Letozaf_, ahh right.. So look at test_go_home_must_open_the_home_directory [20:29] it asserts to self.fakehome [20:30] balloons, so in the root directory test I suppose that if the test type is 'deb' I will have to assert to self.fakehome [20:37] Letozaf_, it's confusing to me actually why that is [20:37] I wouldn't do anything special for .deb no [20:39] balloons, yea: Location: "/tmp/tmp2tsfl4lb" [20:39] Location: "" [20:39] Location: "" [20:39] Location: "" [20:39] Location: "" [20:39] Location: "" [20:39] Location: "/" [20:39] balloons, Location: "/" [20:39] balloons, it's / [20:39] * Letozaf_ is a bit confused [20:39] Letozaf_, if you read the log, home is set to Patched home to fake home directory /tmp/tmp2tsfl4lb [20:40] so '/' should not be the same as home [20:40] balloons, the test that's failing is the root directory one [20:40] Letozaf_, lol ohh my. should have watched the video [20:40] Letozaf_, it fails to wait for the animation [20:40] simple fix [20:41] balloons, how do get the video ? I tried to [20:41] * Letozaf_ blushes [20:41] http://91.189.93.70:8080/job/generic-mediumtests-vivid/156/artifact/filemanager.tests.test_places.PlacesTestCase.test_go_to_root_must_open_the_root_directory.ogv [20:41] :P [20:43] Letozaf_, so _drag_Bottomedge_to_open_places or open_places should wait for the places page to load [20:43] Letozaf_, also I would make those methods all lower case instead of mixed case, to follow the rest of the app [20:44] balloons, ok thanks [20:44] cool! Shold hopefully we a simple thing to fix [20:44] * balloons would put the wait_for in _drag_Bottomedge_to_open_places [20:44] hopefully there is an expanded property or ? [20:45] balloons, thanks I will check now [21:02] balloons, sorry, wasn't around... and feel free to show any of the stuff i've done :) [21:03] knome, no worries, it was a good session.. heh, but that I mean plenty o stuff for me [21:03] ;)= [21:43] balloons, I will fix the test tomorrow, I am tired and probably the two properties I tried to use on the places page aren't the right ones, I will find the "right" one tomorrow now I am KO :P [21:44] balloons, good night :-) [23:05] * ianorlin is wondering why all of the reporting bugs documenation doesn't have a link to help describe different ui elements so they could use precise langauge and know the different ui elements they are talking about when writing a bug report [23:06] because if someone basically taught themselves how to use computers they and use them a lot they might not know ui elements are called to use in bug reports to be specific in what they are saying. [23:07] It is a lot easier to describe what is going on if you know what to call a widget [23:20] defining UI elements.. sadly they change :-( [23:20] and defining them would mean somebody reading the bug report might need to double check [23:21] describing them clearly enough in human-readable format is much better [23:21] and doesn't take that much longer time [23:29] * balloons floats towards the light [23:29] don't hit the sun, it'll blow you [23:29] ..."blow" in the balloon-pop sense [23:30] reminds me, should play the incredible machine some day [23:47] elopio, ping [23:47] alesage: pong [23:47] elopio in your DEP 8 travels have you ever discovered which scripts to use for provisioning? [23:47] elopio, and/or have you ever discussed with CI folks? [23:48] elopio, i.e. if LRT has special needs, I assume I'll be discussing with them but would like to get a head start [23:48] alesage: this should be ok for simple cases: [23:48] $ adt-run -B --built-tree=tests/app-startup --output-dir=output --- ssh -s adb [23:48] i.e, just set up ssh through adb, the phone will already be provisioned [23:49] elopio, yes I know how to run 'em, wondering if you've been exposed to the scripts they use to provision (they = CI) [23:49] now I'm having problems because my test needs root access. I'm not yet sure how to fix that. [23:49] elopio, there's a switch for that in debian/tests/control [23:49] (one sec) [23:49] alesage: I don't know about those scripts because they told us the phone will be provisioned, so I didn't dig more. [23:49] add Restrictions: needs-root [23:50] elopio, if you find yourself still blocked as again :) (my pw was set specially so I needed a special invocation of adt-run) [23:50] s/as/ask/ [23:50] elopio, right, makes sense re: provisioning [23:50] alesage: I got SKIP Test needs root on testbed which is not available [23:51] elopio, one sec [23:51] I was missing -P 1234 [23:51] to pass the password. Looks better now :) [23:51] elopio riight ok :) [23:56] alesage: I found that I don't need to run the tests as root. I just need to be able to sudo for one command [23:56] elopio, o ok interesting, and that works? /me hadn't considered [23:57] no. I guess can sudo su phablet