17:24 #startmeeting Weather App meeting 17:24 Meeting started Fri Jan 29 17:24:54 2016 UTC. The chair is popey. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 17:24 17:24 Available commands: action commands idea info link nick 17:24 Go go go! 17:25 So there's two topics 17:25 It is cold, wet and windy 17:25 It is! 17:25 I closed the blind it was so bad 17:25 jenkins? 17:25 We need that in the app 17:25 haha 17:25 "Make the weather go away" 17:25 Right, what's up with jenkins? 17:25 "Should you open the blinds today? Yes/No" 17:25 * popey waves at balloons 17:25 yeah, App, done. 17:25 ahayzen, popey: the offline caching needs to be explicitly enabled by gst-spotify for tracks (don't know if it is already or not) 17:25 ah okay. 17:25 jenkins is failing due to the location allow/deny dialog 17:25 Elleo, ok thanks :-) 17:26 like the test pass on my device, as i've already selected 'allow' for weather 17:26 So, one way around that, might be to have a command line option -nolocation 17:27 which completely disables all location lookups 17:27 and relies on manually adding locations 17:27 we need to either find a way of selecting the allow button (which the webbrowser tests seem todo) or running something in command line which just allows the app 17:27 they do!? 17:27 popey, but that sortof kills the points of some of the tests? maybe 17:27 i'd have to have a look 17:27 right, but if they can't be made to work, it's a last resort 17:28 yeah :-/ 17:28 I thought this was un-trivially-fixable 17:28 lets look at how webbrowser-app does it 17:28 it should be simple just to say click allow though 17:28 i tried copying parts of their code 17:28 but it didn't work 17:28 really this should be a uitk autopilot helper 17:29 (the same issue is breaking clock in some of its tests) 17:29 balloons, aware of this ^ (location popup clicking in weather) 17:29 yes 17:29 the other issue, which music even hits sometimes, is bug 1536361 17:29 ubot5> bug 1536361 in ubuntu-ui-toolkit (Ubuntu) "When using the sdk helper for the keyboard to type in autopilot, it reports "Keyboard is not on screen"" [Undecided,Confirmed] https://launchpad.net/bugs/1536361 17:29 their test_geolocation looks lean 17:29 it turns out it wasn't trivial like we thought I did the work but it didn't work on jenkin 17:29 it does, and it 'should' work 17:30 self.pointing_device.click_object(self.dialog.get_allow_button()) 17:30 balloons, you can't run Vis on the device can you? 17:30 i'd have to use print_tree 17:30 popey, then look at the helper 17:30 no, no vis on device 17:31 but yea, we can try again and come up with a solution 17:31 I just took what should work (and apparently does if you aren't Andrew or myself) 17:32 * ahayzen tries to find the helper he found before 17:32 balloons, popey, this is what the webbrowser uses http://bazaar.launchpad.net/~phablet-team/webbrowser-app/trunk/view/head:/tests/autopilot/webbrowser_app/emulators/browser.py#L423 17:33 which is called from http://bazaar.launchpad.net/~phablet-team/webbrowser-app/trunk/view/head:/tests/autopilot/webbrowser_app/emulators/browser.py#L119 17:33 and then from where popey said in http://bazaar.launchpad.net/~phablet-team/webbrowser-app/trunk/view/head:/tests/autopilot/webbrowser_app/tests/test_geolocation.py 17:33 balloons, do you think that 'should' work? 17:34 like is the dialog that comes up part of the app itself ? 17:34 ahayzen, ohh.. Interesting. I thought the tweaks we made prevented the popup 17:34 or is webbrowser doing the dialog differently 17:34 But accepting the request works too I guess -- insomuch as we are on a real device 17:34 balloons, that is what i initially tried to copy and it doesn't work 17:34 balloons, we get stuck when the dialog appears 17:35 IIRC the tweaks we made the dialog still came up 17:35 ahayzen, ahh.. that's coming from the app itself 17:35 src/app/GeolocationPermissionRequest.qml 17:35 ah 17:36 they have a custom dialog? 17:36 so it's not the same 17:36 yeah 17:36 yea, for sites that want your location 17:36 as they are maybe unconfined they can then do the accept manually ? 17:36 its a thing that says "This page wants to know your location" 17:36 so implemented in the app itself 17:36 oh 17:36 hmm 17:36 just tested with maps.google.co.uk on the phone 17:36 so anyways, what we did should have made a dummy location 17:36 but still the app itself at some point would have had to say allow/deny ? 17:36 and no popup 17:36 balloons, maybe check my code 17:36 ahayzen, anyways, I want to test it out manually on my phone -- did you try that? 17:37 balloons, i didn't as i'd rather not clean my phone :-/ 17:37 the whole turn on the fake location and restart the service 17:37 * ahayzen hasn't reflashed his phone for like 2years 17:37 ahayzen, ohh.. I don't remember it being invasive, but that's what I have a test device for 17:38 oh i thought you meant do a --bootstrap flash and then do it ;-) 17:38 no, do you have the wiki page I sent you actually? 17:38 yeah i do 17:39 just run those commands? 17:39 can you link again so I don't have to dig it out? 17:39 and yea 17:39 balloons, https://wiki.ubuntu.com/Process/Merges/TestPlan/location-service#Dummy_provider 17:39 i would have thought the app would still have been accepted though? 17:39 it this not trust store stuff? 17:41 balloons, this part is interesting :-) https://wiki.ubuntu.com/Process/Merges/TestPlan/location-service#Trust_Store_Integration 17:42 we can always ask tvoss 17:43 if we still get stuck 17:45 we either need a way to make that dialog not come up (just force it to be accepted at the command-line) or an autopilot helper to click the allow/deny button :-) 17:45 preferably the latter ;-) 17:46 yeah 17:47 ahayzen, I believe the issue with the latter is that we can't see those objects 17:47 surely an UITK level helper would be able to see it? 17:47 or come other magical thing 17:47 and ahayzen, do you need a proper location in general? 17:47 are there _no_ other apps which do this? 17:47 we just need the dialog to be accepted 17:48 camera? 17:48 as it then puts extra delegates in for the currentLocation 17:48 so it screws all the delegate locations for AP 17:48 popey, yeah, i can't remember what camera did 17:48 i went through the ones i could think of that used location 17:48 so you want a real location ahayzen, right? 17:48 no for it to be denied 17:48 if you want to test it properly, sure? 17:49 balloons, i want *a* location don't really care where, just have it accepted 17:49 ok, so 1) a location of some sort 2) for the dialog to be accepted to use it 17:49 yup 17:49 the dummy provider does #1 17:49 but we are still stuck at #2 17:53 So what's our conclusion? 17:54 either something needs to be done in the fixture at the command line, so accept the dialog 17:54 or somehow we need to be able to access it from autopilot (which seems unlikely) 17:54 so probably the former 17:55 I propose we switch the tests off until they're fixed 17:55 because they're just slowing things down until then. 17:57 :-/ 17:57 i guess 17:57 balloons, jenkins is broken with music as well now :-) https://core-apps-jenkins.ubuntu.com/job/run-ap-tests-ci/202/console "adt-run [17:25:07]: ERROR: unexpected error: Test dependencies are unsatisfiable. A common reason is that your testbed is out of date with respect to the archive, and you need to use a current testbed, or try "--setup-commands ro-apt-update".Exit request sent." 17:58 looks like you need to run apt update on the device :-) 17:59 There seems to be a theme here with jenkins at the moment :( 17:59 heh 17:59 balloons, just needs to add what it says to the adt-run command :-) 18:00 yeah 18:00 Ok. Other than jenkins, what else to discuss? 18:00 The "redesign" from design with new logo placement 18:00 erm, did the sidestage stuff work? 18:00 That's in progress right? 18:00 oh yeah, i haven't head since that email ? 18:00 ok 18:00 Will chase on monday 18:00 they didn't respond yet right? 18:00 correct 18:00 cool 18:01 did sidestage hint land in trunk? 18:01 as davmor showed me we need to get it working nicely on a tablet and converged 18:01 dunno 18:01 ahayzen, it has that. The device was offline. Probably needs a re-flash 18:01 did it even work? 18:01 i took a pot shot, as i don't have a tablet 18:01 balloons, ah ok :-) 18:02 i dont see it in trunk 18:02 you can use marvin :) 18:02 heh 18:02 i just ask Victor to try it :-) 18:02 okay 18:02 The other thing was the API transition which we need to do once we get the design icons out of the way IMO 18:02 (and jenkins) 18:02 yup 18:03 and migration to the new listitems etc 18:03 will make the scrolling super smooth, like music :-) 18:03 yay 18:03 but the expandable listitems could make it a bit more interesting 18:03 listitem could happen with logo? :) 18:03 not a huge change is it? 18:03 heh, listitems would require a bit more work and testing 18:04 logo should be a super simple branch 18:04 we'll see 18:04 not sure how todo the expandable stuff with the new listitems yet 18:04 i can have a look next week 18:04 ok 18:05 Lets wrap up, we're over already. 18:05 okies :-) 18:05 So essentially, jenkins (fix or disable), logo, listitems, api 18:05 yup 18:05 and sidestage 18:06 ok. 18:07 sidestage is already done though 18:07 get it landed ;-) 18:07 see also, jenkins 18:07 ok well that's all from me 18:07 Ok, lets wrap then. 18:07 Thanks guys! 18:07 Have a great weekend. 18:07 thanks :-) 18:07 have a good weekend :-) 18:07 #endmeeting