/srv/irclogs.ubuntu.com/2015/03/11/#ubuntu-quality.txt

balloonselfy, excellent!00:06
=== chihchun_afk is now known as chihchun
pittibarry: we just got bug 1430479 which sounds fairly common05:51
ubot5bug 1430479 in systemd (Ubuntu) "switching init systems together with a libc upgrade kills X and disrupts the upgrade" [Critical,In progress] https://launchpad.net/bugs/143047905:51
barrypitti: that looks exactly like what happened to my vm the first time13:21
pittibarry: ah, good; that workaround is in vivid now13:45
barrypitti: awesome.  i have one more vivid vm to try to break now :)13:47
pittithat was a fun bug13:47
barrypitti: the best ones are :)13:51
elopiogood morning.13:59
kalikianaubuntu-qa: ping, I'm looking to change a property from Autopilot - preferrably w/o relying on visible UI in the QML14:07
elopiokalikiana: thanks :)14:10
elopiokalikiana: so, I assume that you want this to change the orientation, is that right?14:10
elopioI ask because generally that's a bad idea, we should rely only on UI visible components. But faking an orientation change is one of the exceptions to that rule.14:12
kalikianaelopio: yes. I have the QML support for it now. but it only works via a button; see here https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/orientationManual/+merge/252270 something like "root.__helper.orientation = Qt.LandscapeOrientation" is called by the OptionSelector14:12
kalikianathe thing is, no real app would ever have a button to flip orientation14:13
kalikianaand I'm hoping this could become generally useful for apps14:14
elopiokalikiana: I see two options. I imagine there must be a way to link a QML property to an environment variable. If so, we could change the env var when we want to change the orientation.14:14
elopiothe other options is https://code.launchpad.net/~canonical-platform-qa/ubuntu-ui-toolkit/autopilot-listview/+merge/25258614:14
elopiosorry, wrong link.14:14
elopiokalikiana: the other option is http://bazaar.launchpad.net/~autopilot/autopilot/trunk/view/head:/autopilot/introspection/qt.py14:15
elopioyou can call a slot from autopilot.14:15
elopioif you make a function to change the orientation, we could trigger it using from dbus through autopilot.14:15
kalikianathe env var wouldn't be sufficient, because it would only work for an initial value, not at runtime14:16
kalikianaelopio: hmm so I add a "function rotate(int orientation){this.orientation=orientation}" basically?14:17
kalikianais there an example of how to call it?14:18
elopiokalikiana: no examples. This is the first time I've seen a good use case for this.14:20
elopiokalikiana: I'm not sure how to define a slot in QML. I'm reading about that. But if you open autopilot3 vis while your app is running with testabililty, you can see the slots defined for each qml type in the tree.14:20
elopioI'll take a look at your branch.14:21
kalikianaelopio: indeed a function shows up in "slots"14:28
elopiokalikiana: that's nice. Now, I'm not quite sure how to call them from autopilot. My guess would be just to call select_single('OrientationHelper').slot()14:34
elopioI'm slower than you, just looking at the app with the orientation button now.14:35
kalikianaelopio: no worries, I appreciate your taking the time to guide me, that branch didn't pop up in 5 minutes either :-D14:44
elopiokalikiana: you access the slots through the slots property.14:46
elopioself.main_view.select_single('OrientationHelper').slots.test()14:46
kalikianaokay, lemme try that14:47
elopiocool! that works :)14:47
elopiokalikiana: rhuddie: I defined test like this: http://paste.ubuntu.com/10580419/14:49
elopioit just needs to be smarter to switch from the current orientation instead of just setting it to portrait.14:50
elopiothis is really nice kalikiana, thanks for working on it.14:50
rhuddieelopio, kalikiana, that looks like it could be really useful14:52
elopiorhuddie: lets start a vanguard journal. Do you think the scrum journal is too big already?14:53
rhuddieelopio, yes. I would definitely start a new doc. Maybe this is something to put in autopilot docs too.14:55
elopiorhuddie: this class has been undocumented on purpose. It would be good to discuss about it.14:56
elopioping pitti. Would it be a good idea to add a reset() to the adb ssh script that does ubuntu-device-flash?14:58
elopionow that we are upgrading, every test breaks the testbed.14:58
balloonselopio, rhuddie slots are definitely interesting. I can see how we might abuse them14:59
pittielopio: hm, upgrade tests need some scriptery around them anyway (install an old version, etc.); it seems a bit excessive to me to wrap that into the script, as it doesn't need to run while adt-runs (just afterwards)15:00
elopioballoons: we found our first use case for them in 3 years :D15:00
pittielopio: but if it's useful to you for some reason, and you have a way to make that work without extra proxy settings and the like, feel free to add a --reflash15:00
elopiopitti: I think that the first flash is the responsibility of whatever is calling adt-run. But it would be nice if after every test, it gets back to the version it had at the start.15:01
elopiothat will need parameters for version number and channel. So we need to think more about it. I'll make a card for this sprint.15:02
pittielopio: so you'd need to save them at the start, and re-flash at the end15:03
elopiopitti: I'm guessing you would do it by writting the values to files like with adt_reboot_target.15:04
brendandelopio, mardy mentioned you might know how to create a u1 account from the command line?15:07
brendandelopio, the bug that we're trying to write a test for requires the account to be authenticated, but i can't find how to do it in any scripted way15:08
elopiobrendand: : yes, take a look at the unity-scope-click project.15:08
brendandelopio, you mean account-console?15:08
elopioin autopilot/credentials.py15:08
elopiobrendand: in there, we are using dummy oauth tokens. I think you would need to get a real one. There's a sso client python library that might help15:09
brendandelopio, you just call add_u1_credentials on AccountManager to create the account?15:11
elopiobrendand: that add_u1_credentials is implemented in there.15:12
elopioI don't remember what it does. Adds a provider, a service, sets some properties.15:12
elopiofgimenez: nice work on the reboot branch. I'll give it a try later.15:33
fgimenezelopio, thanks, let's see how it works with the rest of the pieces15:34
fgimenezi'm currently adding a test for the noop case (no update available), to make sure that all goes fine until then15:35
elopiofgimenez: rhuddie or brendand: please review https://code.launchpad.net/~veebers/autopilot/fix_for_config_init/+merge/25186215:39
elopiofgimenez: a noop case should raise an exception, right?15:39
fgimenezelopio, yes, i think that TimeOutError should be raised15:40
rhuddieelopio, I'll take a look15:42
elopiofgimenez: sounds good.15:43
elopiowe need to start thinking about reflashing the phone after a successful upgrade.15:43
elopioI didn't had that in mind before.15:43
fgimenezelopio, to leave it in the initial state right?15:44
elopiofgimenez: yes. We currently have only one test, but we will add more. We need each test to finish in the same version it started.15:45
fgimenezelopio, we could keep the first value returned by get_current_build_number to use it later15:47
=== chihchun is now known as chihchun_afk
=== roadmr_ is now known as roadmr
=== meetingology` is now known as meetingology
=== roadmr is now known as roadmr_afk
fgimenezelopio, there's no exception on noop, the Rebooting signal is received with status set to false16:24
fgimenezit's already pushed16:24
elopiofgimenez: I think that we can raise the exception based on the status of CheckUpdates.16:24
elopiolike, check update, if there is update, upgrade, if there i none, raise error.16:25
elopioI'm not sure if that's what you have just tried.16:25
fgimenezelopio, no, just calling ApplyUpdate without having downloaded one16:25
fgimenezfor testing the reactor, what you say about check updates makes sense16:28
=== roadmr_afk is now known as roadmr
brendandelopio, what am i supposed to do with the account returned by add_u1_credentials? it created an account but that account isn't authenticated17:34
brendandelopio, it's important that's done otherwise the bug doesn't reproduct17:34
brendandreproduce17:34
elopiobrendand: I don't know about that. For the scope tests we just needed an account, as we were using a fake server that ignores the tokens.17:35
elopiobrendand: what are you trying to reproduce?17:36
brendandelopio, the bug we're writing the test to cover17:36
brendandelopio, if you upgrade then it deletes your u1 account17:36
brendandelopio, but it only happens if the u1 account was created in the ui17:36
brendandelopio, so it needs to be created the same way17:37
elopiobrendand: how is it different creating the account from the UI to creating it with the glib service?17:37
brendandelopio, i don't know exactly, but with the glib service there is nothing in the 'ID' field on the details screen17:38
brendandelopio, the one created in the ui has  token, token_secret, consumer_key etc17:40
elopiobrendand: maybe you are not setting some properties. I think that there is nothing that the UI does that we can't do with the service, but mardy or dobey are the ones who can answer that question.17:40
brendandelopio, yeah i agree - i just don't know what to do17:41
dobeythe U1 UI does a lot17:41
elopiobrendand: take a look at http://bazaar.launchpad.net/~ubuntuone-control-tower/unity-scope-click/trunk/view/head:/autopilot/unityclickscope/test_click_scope.py#L24217:41
dobeyif you want a real account that works against a live server, you'll pretty much have to use the UI at this point17:42
elopiodobey: can't we take the tokens from the sso client?17:42
brendanddobey, no i don't think we do need that17:43
dobeyelopio: you mean ubuntu-sso-client? no17:43
brendanddobey, or maybe we do, i'm currently confused17:43
elopiodobey: and can we sign with the UI in one phone, and then copy the credentials to other?17:44
dobeyelopio: at least, i don't think so. ubuntu-sso-client uses the v1 REST API, and i'm not sure those tokens will work with v2 api using things (and v1 api doesn't support 2fa). ubuntu-sso-client also depends on gnome-keyring which will only make things more complex.17:44
brendanddobey, we're actually working on a test to cover https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/141365517:45
ubot5Ubuntu bug 1413655 in ubuntu-system-settings (Ubuntu RTM) "Updates panel does not prompt for login when U1 account is invalid/deleted" [High,Triaged]17:45
dobeyelopio: it's theoretically possible to copy the credentials, but it's not trivial at all, and i wouldn't advise doing it17:45
brendanddobey, so we'd need an account which would reproduce that bug17:45
dobeybrendand: it will be very hard to do that at this point, as there's also now a migration script (in rtm), so even if you had an account that didn't have the ACL, upon upgrade it will be migrated when the phablet user is logged in, and so things will just work17:46
brendanddobey, well that's ok :)17:47
dobeyso a test that forces that behavior to happen would be very diffcult to do, and way too complex i think17:47
brendanddobey, no we just want a test that would make it happen *if* the fix wasn't there17:47
brendanddobey, i already reproduced it by upgrading from rtm 16 to rtm 1917:48
brendanddobey, but it only happens with the u1 account i created via the ui17:48
brendandif it's not possible then i need to fall back to using autopilot for this, which is what i was doing17:49
brendandbut it would be nice if there was a simple, few lines of python/bash way17:49
dobeyyou're using https://bazaar.launchpad.net/~ubuntuone-control-tower/unity-scope-click/trunk/view/head:/autopilot/unityclickscope/credentials.py right?17:51
dobeyoh, since i'm in here17:52
dobeywhat do you guys know about this "boottest" thing?17:52
dobeyoh, nevermind. seems that corrected itself now17:53
dobeybrendand: if you do what's done in https://bazaar.launchpad.net/~ubuntuone-control-tower/unity-scope-click/trunk/view/head:/autopilot/unityclickscope/test_click_scope.py#L242 then you should see the issue, as that account is not being created with the ACL18:00
brendanddobey, ok, cool18:00
elopiojibel: in case you want to review it: https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/fix1422569-wait_greeter/+merge/25263018:33
alesagervr, hey Victor, do you have a few min to tutorialize the new flashing rules?18:35
rvralesage: Yes18:35
rvralesage: ops-standup?18:36
alesagervr, ok!18:36
alesagervr, that seems to have worked thx :)19:06
alesagervr, so use that recovery image for what?19:06
rvralesage: Nice, you don't need to do anything else to the phone for regression testing19:06
alesagervr, use for silos too, etc.?19:07
rvralesage: For silo testing, installing the silo with the citrain tool19:07
rvrcitrain device-upgrade <silo number> <password>19:08
rvrphablet-tools-citrain19:08
rvrThat's the package which ships it19:08
alesagervr, but always start from that recovery image, ok19:37
dobeyelopio: you were asking the other day about making url-dispatcher work in a chroot or whatever under X right?19:45
elopiodobey: right.19:46
dobeyelopio: did you figure out a solution for that?19:47
elopiodobey: there were some options. It needs a qpa19:48
elopiocurrently, the qpa is define in mirclient.19:48
dobeyoh right :-/19:48
elopioted suggested that instead of writing the qpa, it would be more useful to get mir working on the virtual machines.19:48
elopiowe have a way to run mir on kvm, http://unity.ubuntu.com/mir/setup_kvm_for_mir.html19:49
dobeywell, i don't need it for running autopilot tests19:49
dobeyso mirvfb wouldn't help me really19:49
elopiobut we don't have those vms available on the ci lab, so ci is investigating what would be better for them.19:50
elopiodobey: if you can start the mirxfb on the machine that you are writting the tests, you could run them in the mirvfb19:50
dobeyi think mirvfb would be best for ci19:50
dobeyelopio: yeah, i'm not doing this for tests. i want it for an actual application :)19:51
elopiodobey: ah, so, yes, we need a qpa for unity7 desktop.19:52
dobeyso i can have my app open up instead of firefox/chromium/whatever for certain urls19:52
elopiodobey: I would like that very much. Shouldn't this qpa be installed when url-dispatcher is isntalled?19:53
dobeyelopio: i would presume so, yes; but it's difficult to control that, as we wouldn't want to have it installed on the phone i don't think19:54
rvralesage: The only thing different for silo testing in vivid is that you need to flash the custom tarball too, using jibel's original script19:56
Letozaf_balloons, hey :D20:08
balloonsLetozaf_, hello20:08
Kaleoelopio, are you aware of any way to fake device orientation for test purposes?20:08
elopioKaleo: thanks.20:08
elopiowe have two options, both in progress.20:08
Kaleo- in progress as in I can't use them yet?20:08
elopioone is faking the sensors, https://code.launchpad.net/~canonical/unity8/fake_platform_sensors_module/+merge/24733420:09
Letozaf_balloons, I fixed this if you want to take a look: https://code.launchpad.net/~carla-sella/ubuntu-filemanager-app/fixed-test-cut-directory/+merge/25220520:09
Letozaf_:D20:09
elopioKaleo: the other is extending the orientation helper to be able to overwrite it: https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/orientationManual/+merge/25227020:10
elopioKaleo: no, you can't use them yet. But I think kalikiana is close to finishing his branch.20:10
Kaleoelopio, thanks, that will be nice; will there be away to trigger the orientation lock as well?20:11
veeberselopio: hey, who and when did you ping re: autopilot CI issues? I would like to follow it up20:11
elopioKaleo: we can do that through dbus.20:11
elopioveebers: fginther, like 30 minutes ago. Said fixed it, it is running, and I should just wait.20:12
elopioKaleo: let me show you some code that I think we have...20:13
elopioKaleo: http://bazaar.launchpad.net/~charlesk/unity8/lp-1410915-re-enable-rotation-lock-test/view/head:/tests/autopilot/unity8/fixture_setup.py#L9120:15
elopioyou would need to add the unity8-autopilot dependency.20:16
elopioif that is not correct for your scenario, we would need to find a better place to put it.20:16
veeberselopio: awesome, thanks for that20:16
Kaleoelopio, I see, is that available already?20:17
balloonsLetozaf_, I saw that but didn't review20:17
balloonsI suppose I can now :-)20:18
elopioKaleo: that one is on unity trunk, yes.20:18
elopiosorry, I linked using a branch.20:18
Letozaf_balloons, :) yes please :)20:18
balloonsLetozaf_, so adding the cleanup means it should always cleanup. The only issue that would remain then is if the test doesn't exist cleanly (which can happen I guess)20:18
Kaleoelopio, ok20:19
Kaleoelopio, thank you20:19
Letozaf_balloons, oh! I understand...20:19
elopioKaleo: np. next time you have a question, please ping the team as mentioned on the topic and whoever is on vanguard will help you.20:19
Letozaf_balloons, I will fix20:20
Kaleoelopio, ok20:21
balloonsLetozaf_, also, the cleanup step is run as part of the teardown the test. so it's run at the end20:21
balloonsthe order then doesn't matter, and your logger line then doesn't make sense20:21
Letozaf_balloons, ok thanks, I will cancel it then20:22
balloonsLetozaf_, all that said, I see that the temporary home is only run on non-devices20:22
balloonspresumably the issue is on the device or no?20:22
balloonsLetozaf_, but notice it uses fixtures.TempDir().. which already cleans up after20:23
balloonsso again presumably this is ONLY an issue when testing the click on the device20:24
balloonsin which case the answer there is a bit muddled20:24
Letozaf_balloons, I will check better20:24
balloonsLetozaf_, it's ok, I didn't give it a good look until now20:24
Letozaf_balloons, :) it's ok I will fix :)20:25
balloonsLetozaf_, SO, with all that in mind, I would suggest implementing something a little different. For click, see if you can create a temporary directory inside of home. Give it a random name. Then for the start of each test as needed, change to that directory. That will let you make sure you can add/remove the folders20:26
balloonsLetozaf_, alternatively you can simply just issue removals for any files / folders you need before creating them20:26
balloonsyou need to make sure things are setup as you want20:26
balloonsthe latter option is "ok", and I might even suggest it20:26
Letozaf_balloons, ok I will do so20:29
balloonsI think that goes back to your original implementation. I would just make sure it's done then for everything20:29
balloonsI hope my rambling makes sense heh20:29
Letozaf_balloons, yes more or less, but yes I think I got the point20:29
nuclearbobso who's the vanguard over here?20:30
nuclearbobI'm having problems with package updates for sanity tests20:31
balloonsnuclearbob, say the magic words20:32
nuclearbobballoons: abracadabra? please? open sesame?20:32
elfysize of a buffalo20:32
balloonsnuclearbob, speak friend and enter20:32
nuclearbobor is a problem with sanity tests something I need to ask ubuntu-qa about?20:33
nuclearbobI'm getting this http://paste.ubuntu.com/10582078/20:33
balloonsnuclearbob, that's the magic20:34
nuclearbobballoons: that's what I was afraid of. I'm not sure if I'm recursively summoning myself now20:34
balloonsnuclearbob, however I can help. Notice the error at the bottom.. That was hopefully supposed to be more informative than in the past20:34
balloonsbasically it means your image index is out of date with the archive20:35
nuclearbobballoons: how do I updated that? I went to the system settings, and it doesn't offer me an update20:35
balloonsthis is common on older images, but possible even on the latest image. the hack listed as a workaround might work for you. If it doesn't the only answer is flash a newer image20:35
nuclearbobballoons: I'm not invoking adt-run directly, unfortunately, so I'm not sure how to pass that argument20:36
veebersnuclearbob: looking at your log now20:36
nuclearbobveebers: thanks. I tried adding -u, and that changed the error, but it's still erroring20:36
nuclearbobis error a verb?20:36
veebersnuclearbob: not with that attitude ;-)20:36
veebersnuclearbob: um, when did you last flash your phone?20:36
nuclearbobveebers: today?20:37
balloonsnuclearbob, ahh gotcha. Well, good luck with that. presumably the archive updated after the image was built20:37
balloonsyou should check and see when those packages landed..20:37
veebersnuclearbob: hmm yeah, well the -u should help with that20:37
nuclearbobveebers: would the log when I run it without -u be useful?20:37
nuclearbobthe main different error there is: E: Packages need to be removed but remove is disabled.20:38
veebersnuclearbob, balloons has a good idea. Its def. something to to with the packages in the archive20:38
nuclearbobveebers: okay. hmm20:38
veebershmm20:38
nuclearbobI see we had the same idea20:38
veebersnuclearbob: yeah sure a log without the -u might be good20:38
nuclearbobhttp://paste.ubuntu.com/10582114/20:39
nuclearbobso I guess oxideqt-codecs-extra is the problem20:40
balloonshttps://launchpad.net/ubuntu/+source/ubuntu-touch-meta/1.21220:40
veebersnuclearbob: yeah I would say so, deciding to remove ubuntu-touch to fix the problem sounds wrong20:40
nuclearbobis ubuntu-touch an important package? it doesn't have autopilot in the name, maybe I don't need it20:41
nuclearbobso do I just wait until a new image is built?20:41
veebersnuclearbob: lol :-) Um I imagine so, it's worth mentioning in #u-ci-eng to see if someone there has the downlow20:42
nuclearbobthe dependency refresh happened today, but the image is from yesterday20:42
balloonsyea, package hit today20:42
balloonsfun isn't it?20:42
balloonsso nuclearbob veebers can we get something on the backlog to talk about this?20:42
balloonsadt isn't the magic sauce I once thought it was20:43
nuclearbobmaybe I can build that functionality into utah...20:43
nuclearbobveebers: I guess we need to propose a card to jfunk?20:43
veebersballoons: what is it that you want to talk about? It seems that something screwy has happened with the archive + image20:43
nuclearbobveebers: I think the problem is, if a package hits the archive, we can't test until the new image is built20:44
balloonsveebers, basically it's not possible to recreate test results20:44
balloonsand further, it's impossible to test old images20:44
balloonsor in this case, the latest image, heh20:44
veebersballoons, nuclearbob, right once we have a story formulated we'll talk to jfunk. (He's in the process of setting up a template to use etc.)20:44
balloonsthis popped up a couple weeks ago, and has been mulling in my mind. pitti and I discussed it at length20:45
balloonsthe error message you see is thanks to my bantering and filing a bug so it would evident and martin updated adt20:45
balloonsI guess I'll start a thread and evolve it from there20:48
veebersballoons: sorry needed coffee. That's def a big problem. Is it obvious whose backlog it would exist on?20:50
dkesselballoons: hey, good day - why would plymouthd consume 10% cpu all the time while i am logged into the desktop?21:09
dkesselmake that around 12,5% and that might be one cpu thread...21:10
balloonsdkessel, howdy21:13
balloonsveebers, umm.. I suppose qa.. really ci isn't too affected, but both are indeed affected, as is the community21:14
dkesselwell... so i just killed plymouthd :p21:15
veebersballoons: right, it affects qa but it's not really us causing it right? But yes putting it with qa is a good start then we could go from there21:18
balloonsdkessel, hmm.. maybe it's crashing on startup?21:18
dkesselmaybe it was...21:18
balloonsveebers, well, I mean it's a tool that your team maintains.. And yes, it's not something you can fix in the tool; but presumably your team is the ones with the know-how and ability to create a better solution21:19
balloonsdkessel, I would point you here: https://wiki.ubuntu.com/Plymouth21:21
Kaleoelopio, is there a special trick for wait_select_single to find an object inheriting from QObject only?21:22
Kaleoelopio, in a QML app21:22
dkesselballoons: ok so i got this in syslog: Mar 11 20:18:25 daniel-laptop systemd[1]: Received SIGRTMIN+20 from PID 6282 (plymouthd).21:23
veebersballoons: oh right you mean fix adt or the sanity suite? I thought you meant that it's possible that there is an image that has uninstallable packages due to image and archive release times21:23
dkesseli'll see if i can reproduce after reboot21:23
balloonsveebers, ohh that.. yea, that lies in CI and I suppose is quite related21:23
Kaleoelopio, not an object inheriting from QQuickItem21:24
veebersballoons: right sorry looks like we had our wires a little crossed there. Yeah, qa for getting the tools working which will probably spawn further work for teams re: the archive/image issue21:26
veebersnuclearbob: you have any luck with running the tests?21:26
elopioveebers: can you give a hand to Kaleo?21:34
Kaleoelopio, thanks21:34
veeberselopio, Kaleo sure thing21:35
* veebers reads backlog21:35
Kaleoveebers, so, let's back up a little, elopio earlier pointed me to how to fake orientation sensor values: https://code.launchpad.net/~canonical/unity8/fake_platform_sensors_module/+merge/24733421:35
Kaleoveebers, but now I'd like to be able to _read_ what the current orientation is21:36
veebersKaleo: Ah right, there should be code as an example there. One moment while I find it21:37
Kaleothere is a "# TODO how to get the shell orientation?" :)21:37
Kaleoveebers, I would have a not too  hard way to do it that works for me because I have in my app an instance of OrientationSensor available http://doc.qt.io/qt-5/qml-qtsensors-orientationsensor.html21:38
Kaleoveebers, but my wait_select_single on it always fails21:38
veebershmm, now was that on this laptop or my old one. Kaleo, as an aside while I'm looking for what I'm sure exists, from memory there is a property in the Unity shell, I print_tree-ed the shell21:38
veebersKaleo: I wonder if its not exported for introspection (not everything is, i.e. most non-visible elements)21:39
Kaleoveebers, note that I don't want the orientation of the apps (that is influenced by the orientation lock) but the actual orientation of the device21:39
veebersKaleo: ah right, hmm.21:39
veebersKaleo: A quick test re: your select single is to do a print_tree on the root element and grep the log to make sure what you're looking for is there (or has a different name or something like that)21:40
veebersOn that note, the example I was looking at was the properties that the unity shell provide21:40
Kaleoveebers, good idea21:41
veebersKaleo: how much longer are you online? I've just packed up my devices etc. as I was out the door to the other office, do you have ~5-10 minutes? Then I can go there and get my device out etc21:41
elopioKaleo: veebers: that TODO is in there because the orientation property will change. Not because it's not possible to get it.21:42
elopiothe shell has an "orientation" property. Currently it doesn't change as the shell ignores orientation changes, but when they land the orientation PPA, there will be something similar you can query.21:43
veeberselopio: right, the shell property is there somewhere (I need to have a look for it as I've lost my notes) the question is if the shell exports the details that Kaleo is looking for21:43
Kaleoveebers, I can wait a bit21:43
veebersin the past I've been able to see some orientation change details, but I can't remember exactly where or what (will need to boot up my device)21:44
veebersKaleo: awesome cheers, sorry for the delay :-P I'll be back shortly.21:44
veebersKaleo: right, back now22:03
veebersgetting my device ready to introspect22:03
elopiohum, kiwi doesn't work really well for more than a minute.22:20
veebersKaleo: so I can see 'deviceOrientationAngle:' in the unity8 shell, not sure if that helps you at all?22:21
veeberselopio: kiwi?22:21
elopioit's the irc client on the phone.22:21
Kaleoveebers, it might22:21
ianorlinjoin #ubuntu-offtopic22:21
Kaleoveebers, how can I read it from my app's AP test?22:21
Kaleoveebers, does it return the device orientation without the orientation lock interfering though?22:21
veebersKaleo: good question, I'm not too sure22:27
veebersKaleo:  would have to play around withit22:27
Kaleoveebers, k22:28
elopioKaleo: veebers: in order to access the property from the shell, you would have to launch unity with testability.22:45
elopioKaleo: what is it that you are testing?22:45
Kaleoelopio, I would need to access it from my app22:45
Kaleoelopio, the camera app22:45
veeberselopio: ah right good point22:46
veebersKaleo: how does the camera app (or any app)know if it's rotated or not, surely it should know something about that?22:47
elopioveebers: is it because of this bug that autopilot can't see the orientationsensor object? https://bugs.launchpad.net/autopilot-qt/+bug/127395622:47
ubot5Ubuntu bug 1273956 in Autopilot Qt Support "Re-enable non-visual component exporting in autopilot-qt" [Critical,Triaged]22:47
veeberselopio: potentially22:47
veeberselopio: is the orientationsensor in the app?22:48
Kaleoveebers, I use OrientationSensor22:49
Kaleoelopio, yes it is :)22:49
veeberselopio: if that bug is the reason it should be easier to resolve it for this instance as we changed a little how autopilot-qt did it's introspection so we can do it piecemeal instead of for the whole lot (which is where we ran into major issues and thus couldn't get it released)22:49
elopioveebers: Kaleo: we have been working around that issue with non-visual components adding the value you are looking for as a property.22:50
veebersKaleo, elopio: Right, looks like we could resolve this by updating libautopilot-qt. We should get a card to fix it22:50
elopioyou will see on the shell that the property is:22:50
elopioreadonly property int deviceOrientationAngle: Screen.angleBetween(Screen.primaryOrientation, Screen.orientation)22:50
Kaleoelopio, the underlying issue really needs fixing :/22:50
elopioyou could do something similar on your app, with Screen or the sensor.22:50
elopioveebers: if that bug can be fixed, that would be awesome.22:50
elopioI had already forgotten about it.22:51
Kaleoelopio, the Screen.orientation value is wrong for me22:51
Kaleoelopio, it is influenced by the orientation lock22:51
veeberselopio, Kaleo: I think the bug could be quickly resolved for this instance (and a couple of other defined non-visual) it's a bigger bug for all non-visual22:51
Kaleoelopio, yes I could add a property22:51
veebershmm, odd that bugs title and the contents don't match (the description is re: issues with a newer version of qt :-P)22:53
elopioit looks recycled :)22:54
veebersyeah :-) elopio, Kaleo I might file a specific bug re: the orientationsensor not being exported to get a card in the backlog, the card should include investigation around getting all elements exported (but focus on just the sensor for now)22:56
Kaleoveebers, ok, cool22:57
elopiosounds good. Thanks veebers.22:58
elopioveebers: btw, I pushed on today's meeting to keep the proxy base classes bug in the suggested backlog.23:00
elopioI had no idea how to make a good story for it, so the title might not be the best.23:00
veeberselopio: ack, I can eyeball it if you like23:01
elopioif you have good reasons to fix it in the next sprint, it would be nice if you left them in the comments.23:01
veebersfyi bug #143105723:01
ubot5bug 1431057 in Autopilot Qt Support "Orientation Sensor not exported" [Undecided,New] https://launchpad.net/bugs/143105723:02
veebersack, will do23:02
veeberselopio: thoughts? http://paste.ubuntu.com/10582689/'23:07
veeberselopio: (man, I keep bugging you today :-)) hey you have a moment to discuss the upgrade MP quickly? https://code.launchpad.net/~canonical-platform-qa/ubuntu-ota-tests/upgrade/+merge/25237223:11
elopioveebers: you are not bugging me.23:14
* veebers increases effort to bug elopio23:14
elopioveebers: the paste looks good. jfunk: http://paste.ubuntu.com/10582689/ (in case you want to change the title)23:15
elopioveebers: so, upgrade...23:15
veeberselopio: so to confirm the file 'adb-reboot-to-recovery' is a forked/patched copy of what's in adt-run already? Followup: 'autopkgtest-reboot' is provided by adt-run?23:15
veeberscoolio re: card, I'll email jfunk in a bit about this card and getting a new one in for the ap-qt/sensor bug23:16
jfunkthx elopio, veebers -- plz put in email23:17
veebersjfunk: do you have a template yet? :-)23:17
elopioveebers: yes, it is. It has two line changes for adt_reboot_target23:17
elopioand we might need two more lines if we want to reboot with /sbin/reboot instead of adb reboot.23:18
veeberselopio: any possibility of this being upstreamed into autopkgtest ?23:18
elopioveebers: it must be upstreamed.23:19
veeberselopio: awesome23:19
elopioI haven't done it because I think it would be good to make the two changes in the same patch. But we can do it anytime.23:19
elopioand /tmp/autopkgtest-reboot is provided by adt-run, yes.23:19
veeberselopio: sweetbix, that's my 'big' questions asked then :-) I'll give it a run and I have a couple of comments re: non-functional parts of the diff (needed comments/clarifications etc.) otherwise it's looking good. Always a good experience reading your code23:20
elopioveebers: thanks man. Please leave all the suggestions in there and I'll implement them in the night.23:22
veeberselopio: will do23:23
veebersubuntu-qa: I'm having flashing troubles. trying this command using u-d-f from ppa: ubuntu-device-flash touch --developer-mode --password 0000 --channel="ubuntu-touch/devel-proposed" --bootloader --wipe23:53
veebersand getting the error: unknown flag bootloader23:54
veebersI also get that error if I try to use the --revision=138 flag, which is documented in the help text23:54
veebersany ideas?23:54
alesageveebers hi23:55
veebershey alesage o/23:56
alesageveebers, what does --bootloader do?  not seeing it in help23:56
veebersalesage: I'm told by elopio it's what I need to do to fix my issue :-)23:57
alesageveebers, ok I'm stumped :) , can you describe the issue?23:59

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