/srv/irclogs.ubuntu.com/2015/02/17/#ubuntu-quality.txt

elfyjibel: hey - any reason why lookiing at jenkins the last result for ubiquity appears to be 3rd December?11:23
elfyhttps://jenkins.qa.ubuntu.com/view/Ubiquity/view/All/?11:23
jibelelfy, the tests have been dropped by mistake, they'll be back soon (before Beta 2) Sorry about that.11:24
elfyjibel: ok - thanks :)11:25
elfyI don't often look tbh - only when we appear to have boot issues11:25
elfywhich of course we've got right now :)11:25
jibelI start looking around 1st beta, when things start to calm down and false positives are rare.11:26
elfyoh this is a real positive - boot 32 bit with vbox and it doesn't :)11:26
elfywaiting for today's builds and I'll check again and try it on hardware as well11:27
=== pgraner-hol is now known as pgraner
=== josepht_ is now known as josepht
elopiogood morning15:08
brendandelopio, fgimenez - a branch i'd really like to merge before i start working on the camera test - https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/launch_app_base_class/+merge/25001715:08
dobeypitti: hey. could we get the ubuntu-touch-session script in autopkgtest changed to use a resolution that matches an actual phone (like 540x960, 720x1280, 1080x1920 or such), and have a window manager running (so that --maximize and --fullscreen work properly with qml/qt)?15:09
elopiobrendand: I would prefer it as a fixture instead of a base class.15:10
pittidobey: that sounds like a nice idea; would you mind filing a bug? (can't do it right now, sorry)15:10
pittidobey: I'll get that done tomorrow15:10
dobeypitti: sure15:10
elopiobrendand: also, that bug link seems wrong.15:11
elopiotakes me to page not found.15:12
brendandelopio, mm, there's a small typo15:13
brendandelopio, i'm not so sure about it being a fixture15:13
elopiopitti: ping. fgimenez found that dpkg-architecture, when installed in a read-only phone, gives an error because it can't find the perl libraries. Shouldn't adt-run set the env var for perl libraries path too?15:13
dobeyelopio: that's weird15:15
dobeyelopio: i don't buy that, because we use dpkg-architecture in the store scope to get the architecture to query for15:15
dobeyelopio: so either that's just a new bug, or something else is wrong15:16
elopiobrendand: I think it will be clearer. Call it OpenAppFromDash, and then it's also clear that the click scope will be opened.15:16
brendandelopio, but i don't want it to open the app immediately15:17
brendandelopio, the test author needs to control that part15:17
brendandelopio, whether to run some code before actually launching the app15:17
fgimenezdobey: in our case the dpkg-architecture script was complaining about the Dpkg.pm module, it wasn't able to find it15:18
brendandelopio, so then the launch_app part will have to be a seperate helper15:18
elopiodobey: I think that in jenkins, that dependency is installed making the partition writable. I'm not sure about that.15:18
brendandelopio, i'll have a look at it though to see how it works15:19
elopiobrendand: then you could make a method launch_app that's not called from the fixture setup. Or do not open the click scope on the fixture, just launch the app.15:19
elopiowe can make a fixture that opens the scope and launches the app. And one that only launches the app.15:20
fgimenezdobey: dpkg-architecture is installed through dpkg-dev and is located under the writable /tmp dir where the dependencies are installed15:20
elopiobut do we have a test that needs to do something between opening the scope and launching the app?15:20
pittielopio: yes, it does set $PERLPATH15:21
pittielopio: but you really don't want dpkg-architecture; dpkg-dev has a *huuuge* dependency list15:21
pittielopio: you want dpkg --print-architecture15:21
pitti(which requires no dependencies)15:21
pittidobey: really? you don't use dpkg --print-architecture?15:22
pitti(most of the time that's all that an app wants/needs to know)15:22
elopiopitti: what we need is dpkg-architecture -qDEB_HOST_MULTIARCH15:22
dobeypitti: oh maybe15:22
elopiois there a way to get that string without dpkg-architecture?15:22
dobeyi thought we were using DEB_HOST_MULTIARCH though15:23
elopiopitti: and the envvar that we need to set is $PERL5LIB15:23
pittielopio: right, that's what it actually sets15:24
pittiso, feel free to open a bug about dpkg-archtiecture not working15:24
dobeyoh no, right15:24
pittibut keep in mind: this is a r/o phone; tests can't pull in a gazillion debs :)15:24
dobeywhenever i used ro-apt with qemu, it just hangs :-/15:25
elopiopitti: yes, it's taking a long time. I would like not to install dpkg-architecture, but I know no other way to get the path to the installed libraries.15:26
pittielopio: which libs?15:27
pitti(ldd is a nice tool for that)15:27
elfyballoons: you got any idea why vivid dailies have stopped for some of us?15:27
elopiopitti: /usr/lib/{}/oxide-qt/chromedriver'.format(architecture)15:28
pittior dpkg -L even15:28
pittianyway, sorry guys; I need to run early today15:28
pittiTTY tomorrow!15:28
elopiobye, thanks pitti.15:28
pittielopio: installing 50 MB of dependencies just for that seems excessive; dpkg -L might be a lot ligher?15:29
pittilighter15:29
pittielopio: or some globbing perhaps?15:29
elopiofgimenez: what do you think? dpkg -L sounds good to me.15:30
fgimenezelopio: pitti: of course, we can give it a try and remove a lot of download time15:31
elopiofgimenez: could you please file the bug about dpkg-architecture not working on r-o?15:33
elopiohttps://bugs.launchpad.net/ubuntu/+source/autopkgtest/15:33
dobeydpkg-architecture requires installing 50 MB of deps?15:35
fgimenezelopio: ok thanks15:35
balloonselfy, I see nothing today for xubuntu15:40
balloonsubuntu gnome and lubuntu also seem affected15:41
balloonsbut no I don't have an answer15:41
paulliuelopio: https://code.launchpad.net/~paulliu/unity8/notification_helper/+merge/24921115:43
paulliuelopio: just fix all the issues you mentioned. Also run flake8. I didn't fix the errors not related to this commit.15:43
elopiopaulliu: +115:52
elopiothank you.15:52
elfyballoons: ok15:54
elfyballoons: seems kylin and studio as well15:56
brendandelopio, we don't run --dbus-probe-enable in the sanity tests at the moment right?15:57
elopiobrendand: yes we do15:57
brendandelopio, hmm i wonder why autopilot can't introspect the camera-app then?15:57
brendandmaybe it didn't load properly15:57
elopiobrendand: do you have a screenshot?15:58
elopioI was talking with the camera guys some minutes ago, and on their test they are launching with upstart.15:58
elopioif it works for them, should work for us.15:58
brendandelopio, nope the app is launched but still nothing16:15
brendandhmmm16:15
* brendand afk for a bit16:19
paulliualesage: https://code.launchpad.net/~paulliu/unity8/notification_helper/+merge/24921116:28
alesagepaulliu, ok will get to today16:28
elopiojibel: the last question for today, for when you return from your meeting17:07
elopiojibel: there is a QML element that takes care of blurring the camera screen when the controls are open. To do that check, the easy solution is to check for the presence of that QML element.17:08
elopiothen we are not really checking that things are blurred. We leave that to lower level of testing, probably qml tests.17:09
elopiois that ok for you?17:09
jibelelopio, that's fine. I don't think we check it manually either :)17:15
elopioubuntu-qa: we need to think a sustainable way to say: this high level test assumes that all these low level tests exist and are green.17:17
elopioif the low level tests don't exist, are not running, or are not passing, somebody has to suffer.17:17
balloonsoO, a lovely whole and not the first or last I suspect17:17
davmor2Test All the Tests17:18
elopiothe problem is, how to notice about that without reading all code.17:18
alesageelopio presumably a software wouldn't make it onto the phone if its low-level tests were failing?17:19
* alesage notes that it's necessary to put a question mark there17:19
alesageelopio, or maybe I'm not following17:19
alesageelopio, we can't take on the responsibility of running those tests *again*, correct?17:20
* alesage taps microphone17:22
alesagean audit of some kind?17:22
elopioalesage: I think it's not often that we release something with low level tests failing or skipped, you are right there.17:27
elopiobut we often releasea things with low coverage.17:27
alesageelopio, so you're proposing a way for ops to see the "testing health check" for the release or silo17:29
elopioalesage: yes, maybe. It will help if they see test coverage and test results.17:30
elopioI also think it would be useful if they inspect the branches that are landing, to get a feeling of the new tests being added.17:31
elopiobut maybe that's too much.17:31
alesageelopio, would those stories split like: "a review of lower-level relevant tests for ops", "lower-level tests added", "test coverage total for project"17:34
elopioI don't follow.17:34
alesageelopio just trying to split up what you're proposing into specific stories, with the ops team as the user17:35
elopioalesage: ok, I think it should have a historic view of number of tests run, passed and skipped.17:36
elopioit shouldn't arrive to silo testing if there are failed tests, at any level.17:36
elopioand a silo must come with links to bug reports and MPs, so we can take a look at specific code details that seem important.17:37
alesageelopio, right, would be good to trace back to the relevant CI builds and validate that they happened, etc.17:37
alesage(as a first step toward other coverage, tests introduced, etc.)17:38
alesagethe edges demo changed, eh?17:41
=== roadmr is now known as roadmr_afk
elopioping veebers: there is this bug: https://bugs.launchpad.net/autopilot/+bug/142279720:06
ubot5Launchpad bug 1422797 in Autopilot "MIR refuses the app to connect" [Undecided,New]20:06
elopiowhich I think it's either the camera app taking too long to start, or unity failing to see it, or ubuntu-app-launch doing something wrong.20:07
elopiothat is, I think it's not an autopilot bug.20:07
elopioveebers: if you have some time, could you take a quick look at it and tell us if you notice something useful for the tests?20:08
veeberselopio: right, I don't think it's an autopilot bug (unless something has changed and we weren't told and launch isn't doing the right thing any more)20:08
elopiobut only if it's quick. If it will take more time, we'll have to schedule it for next week.20:08
veeberselopio: I'm not sure I follow your request? Ah you mean take a look at the issue and see if I can find out anything?20:09
elopioveebers: yes. As far as I can see, it can't be autopilot's fault because I read the code for launching and stoping and it looks correct to me.20:09
veebersbrendand: do you think the issue you're seeing in your test is related to the bug elopio linked? ^^20:09
elopiolike, it can't fail if ual is behaving correctly. But I might be missing something.20:09
veebersbrendand: if you're seeing a testability load error I don't think this bug is what you're seeing :-)20:13
=== roadmr_afk is now known as roadmr
elopiough.20:30
elopiothe new error in the wizard test it's because we have a class also called LanguagePage in system settings.20:30
elopiowe need to improve this autopilot cache. I have no idea how.20:30
balloonsubuntu-qa, Letozaf_ has an interesting problem to work on I suspect a solution already exists for. She needs to test mounting and unmounting an sdcard inside file manager20:31
elopioballoons: we haven't solved that yet. We talked about calling the mount and umount to fake the insertion of the card.20:32
veeberselopio: yeah it's becoming more an issue20:33
elopioLetozaf_: we haven't tried that yet, and we don't know if there's a better solution than that either.20:33
elopiooh, and there's the issue that the lab will need to have a phone with a card.20:33
dobeyelopio: mount/unmount won't simulate insertion of the card. you'd need to do some fake thing through udev i think20:33
elopiofor our case, we are assuming that. For a general case, the sdcard itself needs to be faked.20:34
Letozaf_elopio, ah, ok20:34
dobeybecause, really, 'insertion' would normally result in automounting20:34
elopiodobey: yes, that sounds right.20:34
elopioballoons: do you know who worked on the sdcard? He needs to provide a fake for us.20:35
dobeyor permanently affix an sd card to an arm, attached to a stepper motor, and use an arduino to insert/remove it physically in a slot20:35
balloonselopio, ohh good point. No idea, but here is pitti's thoughts on the specific bug. I was curious about what he had done with mocking anyway, looks like umockdev isn't enough: https://bugs.launchpad.net/ubuntu-filemanager-app/+bug/1406988/comments/120:36
ubot5Launchpad bug 1406988 in Ubuntu File Manager App "Autopilot tests for SDCard support" [High,New]20:36
dobeymay damage the sd card or slot though20:36
elopioveebers: do you know if autopilot has a bug for that thing in the registry?20:36
veeberselopio: let me check20:37
veeberselopio: no there isn't a bug for it20:42
elopioveebers: I'll make one.20:42
veeberselopio: awesome, thanks20:44
balloonsdobey, indeed, but the test fails on the desktop then :-)20:45
balloonselopio, so is there an open bug / issue for this? I guess if not we'll need a meta bug for it, assuming you think there should be a helper / provider for this20:46
dobeyballoons: not if the device is present and being used :)20:46
dobeybut will fail for many other reasons20:46
elopioballoons: no bug that I know of.20:48
elopioveebers: how did you workaround the MRO issue?20:48
elopioI'm looking for a quick fix for this LanguagePage problem, hitting funnier problems on every try.20:49
veeberselopio: By doing the import at the top of the file and not within the method20:49
elopiooh, I know what to do 8-)20:49
veeberselopio: oh?20:56
=== pgraner is now known as pgraner-afk
veeberselopio: hey, follow up, is this MP ready for review (was going through cards)? https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/fix1422568-wizard/+merge/25007121:26
veebersI ask because it's not in the review lane ;-)21:26
ianorlinuh 14.04.2 download info for ubuntu server generates a 404 error and zsync fails21:31

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!