=== DalekSec_ is now known as DalekSec === chihchun_afk is now known as chihchun [05:52] thomi: hm, what's the official way to run an autopilot test in-tree now? [05:52] $ PYTHONPATH=. python -m autopilot.run run autopilot.tests.functional.test_introspection_features.IntrospectionFeatureTests.test_window_geometry [05:52] $ python3 -m testtools.run discover autopilot.tests.functional.test_introspection_features [05:52] both of these fail [05:52] and there is no bin/autopilot any more which DTRT [05:52] pitti: yes, that test fails, veebers found that today, and he's looking at it :) [05:52] pitti: DTRT? [05:52] thomi: no, I mean it doesn't even start running the tests [05:53] thomi: I wanted to see what's wrong with it as I recently touched it, but I don't know how to run a test [05:53] thomi: "autopilot" uses the system-installed ap, that doesn't work as it's way too old in trusty [05:53] I need to run it from within the tree [05:53] pitti: well, you don't need to set PYTHONPATH, but aside from that, what you're doing is correct [05:54] pitti: also, the system autopilot in trusty should work fine [05:54] pitti: I think maybe there's something funny happening on your system? [05:54] thomi: http://paste.ubuntu.com/6719045/ [05:54] thomi: [05:54] $ autopilot run autopilot.tests.functional.test_introspection_features.IntrospectionFeatureTests.test_window_geometry [05:54] from autopilot.testresult import AutopilotVerboseResult [05:54] ImportError: cannot import name AutopilotVerboseResult [05:54] ^ that's a change in trunk which isn't in trusty yet [05:55] pitti: oh wait - sorry - I forgot I built & installed it from source :-/ [05:55] thomi: trusty has 20131125, that's fairly old [05:55] sorry :) [05:55] thomi: np; hence I wonder how to run it out of the tree now [05:55] thomi: i. e. the equivalent of bin/autopilot [...] [05:55] this works for me: [05:55] python3 -m autopilot.run run autopilot.tests.functional.test_introspection_features.IntrospectionFeatureTests.test_window_geometry [05:55] as does python 2 [05:56] I get above pastebin error (AttributeError: 'module' object has no attribute 'functional') [05:56] http://pastebin.ubuntu.com/6719052/ [05:56] with both py 2 and 3 [05:56] O.0 [05:56] and your CWD is inside the lp:autopilot checkout location? [05:56] thomi: yeah, that's the error I wanted to get and investigate [05:57] thomi: yes, I'm in the root dir of the lp:autopilot checkout [05:57] hmmmm [05:57] thomi: ok, but that's the official command? [05:57] then I'll investigate what's wrong with that [05:57] pitti: yeah [05:57] (we should add that to README) [05:58] thomi: even tox.ini says "autopilot run", which again uses the system installed ap, right? [05:58] pitti: no, because tox installs into a virtualenv [05:58] pitti: but the tox config is broken on trusty for some reaon [05:58] *reason [05:58] if you happen to find the problem.... you know what to do :) [05:59] thomi: yeah, I just wanted to make sure I'm actually using the right command this time; thanks [05:59] yeah. generally, 'tox ' should do what you want [06:00] aah! [06:00] $ sudo apt-get install python3-fixtures python-fixtures [06:01] of course that error message made it *absolutely clear* that the "fixtures" module was missing [06:01] isn't that a dependency on python-autopilot & python3-autopilot? [06:01] (not!) [06:01] thomi: it is apparently fairly recent (not in trusty yet) [06:02] thomi: anyway, the test_window_geometry WFM [06:02] pitti: the failures are specific to desktop configuration [06:02] * pitti looks at code to get some ideas what could go wrong [06:02] the test assumes that the desktop starts at 0,0 [06:03] and also assumes that there's only one screen configured [06:03] and further assumes that applicaiton geomentry will always be wholly within the desktop area. [06:03] all in all, not a great test :) [06:03] I enabled my internal monitor, still works [06:03] I moved it to the left of my primary monitor, still works [06:04] (I tested these cases back then) [06:04] hmmm [06:04] but I guess it's related to that, I'll have a closer look and follow up to the bug [06:04] if you can't reproduce, then veebers or I can take a look - at least we have a reproducer already :) [06:09] thomi: followed up [06:10] thomi: so apparently I don't understand what the window geometry says, or there is actually a bug somewhere that it sometimes ends up being negative [06:13] yeah, me too [06:13] there's something screwy going on, for sure [06:13] anyway, I gotta go eat dinner. [06:13] might be on later, but probably not [06:14] talk to you tomorrow! [06:17] Good Morning [06:28] hey DanChapman [06:28] pitti hey :-) how are you? [06:29] DanChapman: I think only ints and strings are matched on the server (libautopilot-gtk/qt, inside the application) side; any other properties (like float) are matched on the client (i. e. autopilot) side [06:29] or more complicated selects [06:29] in those cases, the whole widget (set) is transmitted over D-BUS, and ap filters the result [06:30] DanChapman: quite fine, thanks! and you? [06:34] pitti, ahh ok I see, that makes sense. I mentioned to jibel yesterday about how in the autopilot log of the failed tests that there is an "atk_object_set_parent: assertion 'ATK_IS_OBJECT (accessible)' failed" right before the segfault. COuld it be a Gtk/Atk bug causing our crashes? [06:35] pitti, awesome, Yes I'm really good thanks [06:35] DanChapman: in principle, yes; if something before that code corrupts the stack, then anything is possible [06:36] i. e. we need to find out what calls set_parent() and why it's calling that on an invalid object [06:37] DanChapman: oh, is that really an assertion failure, or a CRITICAL? [06:37] DanChapman: an assertion failure should terminate ubiquity, with SIGABRT [06:38] CRITICALs are merely printed out and then the program continues, unless you run ubiquity with G_DEBUG=fatal-criticals [06:38] pitti its a CRITICAL [06:38] that's useless for debugging where a critical happens, but of course inconvenient if there are lots of other criticals before that [06:38] sorry, s/useless/useful/ [06:54] pitti, I see so it might not be that then, I just find it strange it being the last thing that occurs before the segafult. [06:54] DanChapman: oh, it certainly might be that [06:54] DanChapman: i. e. it's well worth investigating why that critical happens [07:36] morning === davmor2_ is now known as davmor2 === chihchun is now known as chihchun_afk [11:35] jibel: hm, wazn unhappy again? I already retried on i386, amd64 works fine (http://d-jenkins.ubuntu-ci:8080/job/trusty-adt-shotwell/ARCH=i386,label=adt/61/console) [11:35] * pitti logs in and check [11:38] jibel: indeed, I can reproduce with "./auto-package-testing/bin/run-adt-test -sU -a i386 shotwell" === _salem is now known as salem_ [11:38] pitti, there is no particular activity on this machine [11:38] with -a amd64 it works [11:38] pitti, I'll reprovision the base vm [11:38] broken i386 base image? [11:39] jibel: OOI, how do you do that? run prepare-testbed, or trigger the jenkins job? [11:40] pitti, either trigger the job trusty-adt-setup-testbed or run $HOME/auto-package-testing/bin/prepare-testbed -r trusty $ARCH [11:40] jibel: ok, I can do the latter (unless you already do) [11:42] pitti, it is running [11:42] jibel: I see you started prepare-testbed for amd64 (but that's the good one) [11:42] jibel: ack, thanks; I'll retry shotwell and spamassassin after that [11:42] pitti, I'll redo both [11:42] archs [11:58] pitti, vms have been rebuilt and shotwell is running on i386 [11:59] that's http://d-jenkins.ubuntu-ci:8080/job/trusty-adt-shotwell/ARCH=i386,label=adt/61/console ? [11:59] looks fine [11:59] jibel: merci [12:00] * pitti kicks spamassassin, too === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === zequence_ is now known as zequence [15:03] good morning QA. [15:04] morning [15:04] morning elopio [15:11] afternoon :p [15:18] xnox, ping [15:18] balloons: hey! [15:19] xnox, so playing with the emulator finally :-) I noticed a few things.. 1) It's REALLY slow for me 2) The core apps don't seem to be a part of the base image you are using [15:20] balloons: 1) emulator is slow, it's a single core 700 Mhz cpu 2) like what ? [15:20] I'm mostly concerned about #2. Should I need to install them? Your run script runs things like weather and calc, but they aren't on my image [15:32] balloons: i'm still confused what you are missing. my scripts as posted latest are to do automated autopilot testing. [15:33] balloons: for general instructions of using emulator see https://wiki.ubuntu.com/Touch/Emulator ; which is ./build-emulator-sdcard.sh & ./run-emulator.sh [15:33] balloons: you should have mostly everything as a normal device, with same sets of apps etc. [15:33] balloons: what's missing? what's different? [15:44] xnox, sorry, got dc'd for a min. I followed your instructions.. I'll try again from the top. The resulting images were simply missing the core apps and I don't know why [15:44] elopio: Good morning [15:44] browsing, dialer, friends.. all there.. no core apps like weather or calc tho [15:45] balloons: which instructions? [15:45] from your email [15:45] balloons: can you give me a link to the post? [15:46] * xnox sent many emails with different instructions.... =) [15:46] balloons: or title of the email post. [15:46] https://lists.launchpad.net/ubuntu-phone/msg05784.html [15:46] balloons: none of that should be used to "play with emulator" or "use it" [15:47] balloons: that's only for automated, hand-off, execution of autopilot tests. [15:47] I'm trying to run core apps tests.. I tried the steps to do that and it failed. [15:47] balloons: if you want to start emulator and use it, please see https://wiki.ubuntu.com/Touch/Emulator#Instructions [15:48] xnox, right.. I don't care about using the emulator, I was looking to reproduce test runs.. then run my own [15:48] balloons: do you have logs of the failures? [15:48] xnox, do you build the sdcard before you run your tests? [15:48] balloons: it does everything. [15:49] ok, that's what I did then.. [15:49] if I provision again, will it wipe the old stuff.. [15:49] * balloons looks [15:50] balloons: calculator autopilot tests can crash the emulator at the moment, what else have you tried. friends pass, dialer partioally passes. weather should be ok as well, let me run that one now. [15:52] xnox, after taking a boot to boot, the loop checking for unity status goes on forever.. that's how every test ended [15:52] anyways, trying again.. let's see what happens [15:52] elopio: https://bugs.launchpad.net/ubuntuone-credentials/+bug/1248576 [15:52] Ubuntu bug 1248576 in ubuntuone-credentials "StateNotFoundError running autopilot tests" [High,Incomplete] [15:53] balloons: i should add watchdog for that. basically sometimes some click hooks fail in the emulator and thus unity8 never starts =( [15:53] elopio: I'm getting a similar bug in another test. Did you discover why StateNotFoundError happens? [15:54] rvr: in that case, it seems QML doesn't find a module. [15:54] so the window is actually missing the component you are looking for. [15:54] take a look at the QML errors to see if that's your case. [16:00] xnox, ok, same thing happens. Here's the log: http://paste.ubuntu.com/6721438/ [16:01] currently, the emulator displays unity with the intro screen [16:01] balloons: the log does tell you what's wrong. [16:01] balloons: "error: more than one device and emulator" over and over and over again =) [16:01] yes :-) [16:02] lol.. I do have a device plugged in [16:02] d'oh [16:02] balloons: maybe you want to fix that =) as raised in the email thread.... [16:02] it was the tablet, didn't even notice.. I'm sorry [16:11] balloons: [16:11] balloons: so weather app does work. [16:12] xnox, seems to be running.. is there an easy way to get logs, or should I simply redirect output? [16:12] balloons: just wait for it to finish. [16:12] balloons: you will have logs, as done by phablet-test-run in the "clientlogs/" folder at the end. [16:12] perfect, ty. I ran the full run-tests.sh so this looks like it's going to be quite a bit [16:17] balloons: and will most-lickely fail a lot and get your emulator stuck =) [16:17] hah.. maybe I should quite while I'm ahead [16:18] or be quiet or quit :p [16:18] good day to you balloons :) [16:19] hallo [16:24] howdy elfy and knome [16:26] xubuntu people, they come in numbers [16:27] always in a pack :-) [16:42] balloons - we hunt in packs [16:42] slickymaster is sneaking round the back to catch you out [16:43] * balloons looks around nervously [16:45] hey balloons [16:45] * slickymaster thinks that balloons should always be afraid... very afraid [16:46] and looking over his shoulder [16:46] hey pitti [16:47] balloons: your prodding sufficiently convinced me about bug 1033932, I just committed a fix :) [16:47] bug 1033932 in apport (Ubuntu Precise) "Error alerts appear on login for problems in previous session" [Undecided,Triaged] https://launchpad.net/bugs/1033932 [16:47] oO pitti ! [16:47] http://bazaar.launchpad.net/~apport-hackers/apport/trunk/revision/2746 [16:47] that's excellent news! [16:48] I really wasn't aware that we have so many "crash on logout" cases; I almost never get those [16:48] last cycle was especially bad for me.. this cycle hasn't been so bad. not seen many [16:49] it seems like the longer I run the same install the more buildup I get, the more I see them :-) [16:50] so pitti plans to SRU this at all? This should hit trusty at least soon [16:50] balloons: let's let it bake in trusty for a bit; the bug has a precise task [16:50] balloons: yes, stgraber is working on fixing apport for containers, then we'll land this in trusty (by next week) [16:50] ofc.. not rushing anything out, heh ;-) [16:51] still, that's very good news. This should really help out trsuty [17:04] balloons, DanChapman, can I ping you guys in something offtopic? [17:04] slickymaster, sure :-) [17:04] indeed [17:05] * DanChapman waves to balloons [17:05] * balloons waves back [17:05] I'm applying for ubuntu membership so I would like your testimonials if you think I'm worthy [17:06] balloons, DanChapman ^^^ [17:06] slickymaster, I'd be more than happy too :-) [17:07] thanks, DanChapman. I sure appreciate it [17:07] DanChapman: my wiki page is https://wiki.ubuntu.com/slickymaster [17:08] slickymaster, cool I will get that done :-) [17:08] DanChapman: once again, thanls [17:08] *thanks [17:28] xnox, ok, so last question :-) If I wanted to run an arbitrary lp branch in the emulator, can I do it? If not, can we get it added? [17:29] balloons: there is no support for that at the moment, so patches welcome. [17:29] xnox, k, I'll give it a whirl [17:29] balloons: it will most-likely be patches to phablet-test-run [17:29] balloons: or options to run-test.sh [17:30] I want to let phablet-test-run do it.. so with that it should be simple on your end eh? [17:30] balloons: right, yeah anything that phablet-test-run supports is easy. [18:08] slickymaster, nice work on wiki page. I want to get mine together and apply also [18:09] thanks cgoldberg. I trully hope that you'll manage to be successful [18:10] cgoldberg: if you'll need any help, just ping me [18:12] jibel, you still about? [18:13] balloons: slickymaster is going for membership - he might not mention it to you ;) [18:13] elfy, indeed. I'm on it :-) [18:13] :p [18:14] if most of my contributions to Ubuntu (besides being a fanatic user since Warty) are as a Canonical employee, do those count? [18:15] cgoldberg: I think elfy would be the best person to address that question [18:15] elfy: ^^ [18:16] * elfy read the question and has no answer - balloons might have, being one [18:17] not sure I could think of any reason why not - but I've never had anything to do with the membership boards [18:17] contributions are contributions :-) The keys are sustained contributions and testimony from people about said contributions :-) [18:18] that [18:18] fair enough [18:36] xnox, does the size of a disk image affect whether you get the auto-resize options. I'm using a 40G win7.qcow2 image and I only have the options to erase it or 'something else' [18:37] DanChapman: auto-resize option is tricky, it depends on many things. [18:39] DanChapman: so first the drive needs to be big enough for two fit two OS, second the original OS needs to be "defragmented enough" to be able to shrink, then shrinking should be possible (so Windws may not be suspsended / hybernated / mix hybernated / quick start disabled), there should be enough primary partitions left if MBR is used (some OEM configurations use up all primary partitions). [18:39] DanChapman: so there are plenty of reasons why auto-resize may not be offered. [18:39] DanChapman: at the moment there is no good logging in ubiquity to log why autoresize was not available. [18:40] DanChapman: also we at times fail to detect Windows 7 / WIndows 8 which also leads ubiquity into believing there "is a large data partition of unknown stuff, not sure what to do with" ===> wipe&install or "advanced" ? =) [18:42] xnox: related: http://askubuntu.com/a/401543/612 - saw this on AU recently. check out the screenshot of his partitions.. [18:43] popey: well bdmurray wants to remove all upgrade options from inside ubiquity =) [18:44] popey: as "upgrade manager is the only support option" [18:47] xnox, ok I see thanks, it seems this could be quite tricky to get an image to work with the ubiquity tests then. THis image has sda1 -> ntfs MBR and sda2 -> ntfs, It detects its windows 7 ok i'll try a larger disk image [18:47] xnox: makes sense === josepht_ is now known as josepht [20:29] Letozaf_, evening to you [20:29] balloons, hello [20:33] balloons, howzit ? [20:33] Letozaf_, getting into using the emulator to run tests [20:34] since it runs so much slower, it can show finicky tests [20:34] balloons, but is it better to test on a device or on the emulator ? [20:35] balloons, saw the email about emulator but was wondering to use it as I have a device [20:35] Letozaf_, device is much faster testing if you have it :-) But if I can't reproduce the issue, the emulator might be helpful [20:36] balloons, so it's better to have it just in case of need [20:36] Letozaf_,right.. and for example I'm looking at these failures; http://ci.ubuntu.com/smokeng/trusty/touch/maguro/114:20140106:20131223.2/5905/ubuntu-clock-app-autopilot/642628/ and http://ci.ubuntu.com/smokeng/trusty/touch/maguro/114:20140106:20131223.2/5905/calendar-app-autopilot/ [20:36] they seem to only happen on the slow maguro device [20:38] balloons, oh I see, so if one does not have that device he can use the emulator to reproduce it [20:38] Letozaf_, potentially yes :-) That's the idea [20:40] * DanChapman has just come across 'indicator-pitti', haha BRILLIANT! [20:41] balloons, well maybe I can try this emulator too, then :) [20:42] Letozaf_, give it a whirl: https://lists.launchpad.net/ubuntu-phone/msg05784.html [20:42] Just make sure you unplug any devices from your pc first [20:42] or it won't work and you'll scratch your head like me [20:43] it will take 5 mins or so to boot [20:43] balloons, thanks for telling me, had my device attached just now :P [20:44] lol.. trust me I spent awhile confused [20:48] DanChapman, lolwut. Was that his fanclub again? [20:48] balloons, the emulator started, but suppose I will have to wait for it to boot [20:49] Letozaf_, yes, it takes a bit.. no worries [20:49] balloons, ok [21:02] balloons, I get a ubuntu-phablet login: but no tests start :? [21:05] keep waiting [21:05] lol [21:05] balloons, ok :) [21:12] Letozaf_, it work? [21:13] balloons, I'm a bit confused.. after some logs on the screen got another ubuntu-phablet login: now [21:13] what did you run? [21:13] balloons, but in the emulator screen is still black [21:14] balloons, I ran ./scripts/run-tests.sh [21:14] run-all takes 1.5 hours [21:14] balloons, what ? [21:14] yes, looks like it iterated [21:14] you can find the logs in the clientlogs subfolder of the directory [21:14] so, now to run something specific do it like this [21:14] balloons, wow I better stop it and try to run just one test [21:15] balloons, ok tell me [21:15] ./scripts/run-test.sh click com.ubuntu.calendar calendar_app [21:15] or even just one test: ./scripts/run-test.sh click com.ubuntu.calendar calendar_app.tests.test_monthview.TestMonthView.test_monthview_go_to_today_prev_year [21:15] that's a 10 min task, hah :-p [21:16] balloons, but now first I must shut it down and restart it with that script, right ? [21:16] yep [21:17] just end the current process [21:18] * Letozaf_ is running the calendar test now [21:23] it fails :-) [21:24] balloons, yes the emulator closed and got emulator-arm: no process found [21:24] balloons, ended just now [21:24] ohh.. that's no good [21:25] balloons, :( what a pitty was curious to see it [21:25] it can fail, it's still wip.. but I've had good success [21:25] try again [21:25] if the emulator itself died you mena [21:25] balloons, ok [21:25] the test should run, just fail to pass is all [21:25] balloons, I will try it again, didn't see anything on the emulator === salem_ is now known as _salem [21:32] balloons, the test does not run on my box: [21:32] Removing directory /tmp/tmplRjlEL [21:32] + /home/letozaf/autopilot-tests/touch-emulator/scripts/run-autopilot-tests.sh -s emulator-5554 -a calendar_app.tests.test_monthview.TestMonthView.test_monthview_go_to_today_prev_year -Q -S [21:32] error: device not found [21:33] weird [21:33] balloons, my device is not attached to my PC :) [21:35] * Letozaf_ is trying agai [21:35] again [21:36] :-) [21:41] balloons, sorry same error :( [21:46] balloons, it's late for me now, be back tomorrow [21:46] Letozaf_, sorry to hear that [21:46] thanks for trying.. but you got to see the run-all work I suppose :-) [21:46] so it's another option [21:47] balloons, what, where ? [21:47] Letozaf_, the initial run output stuff [21:47] balloons, how do I do that ? [21:47] check clientlogs [21:47] the 1.5 hour run you did.. run-tests :-) [21:48] balloons, ooooh! [21:48] balloons, well maybe tomorrow I will try that and start early :) [21:49] balloons, bye see you tomorrow [21:49] Letozaf_, ciao