[08:13] <dholbach> 0good morning
[08:28] <nerochiaro> oSoMoN: good morning. did you have time yesterday to look into that test I mentioned ?
[08:29] <oSoMoN> nerochiaro, morning! no, but I’m on it now, sorry for the delay
[08:29] <nerochiaro> oSoMoN: no problem
[09:31] <JamesTait> Good morning all; happy International Day of People with Disability! :-D
[11:28] <t1mp> oSoMoN: hello
[11:29] <t1mp> oSoMoN: in webbrowser-app, does the address bar currently use a Panel or Toolbar?
[11:29] <oSoMoN> t1mp, it uses a Panel
[11:29] <t1mp> oSoMoN: ok. And there is no automatic hiding of that toolbar after inactivity, right?
[11:30]  * t1mp is checking that he doesn't break anything with some minor implementation changes to toolbar and panel
[11:31] <oSoMoN> t1mp, no
[11:32] <t1mp> oSoMoN: ok, thanks. final question: are all webbrowser autopilot tests supposed to pass on maguro?
[11:33] <t1mp> I'm running ubuntuuitoolkit tests now, but I'll run webbrowser tests also afterwards.
[11:33] <oSoMoN> t1mp, yes, they are
[11:34] <oSoMoN> t1mp, check the latest results at http://ci.ubuntu.com/smokeng/trusty/touch/
[11:34] <mzanetti> nerochiaro: hey. do you know what I need to do in order to make autopilot print the applications debug output?
[11:35] <nerochiaro> mzanetti: it seems to do that for me without me needing to do anything extra
[11:35] <nerochiaro> mzanetti: oh, wait, you mean the app output, not the test output
[11:35] <mzanetti> nerochiaro: yeah
[11:36] <nerochiaro> mzanetti: then AFAIK you can't. but you can tail -f the logfile that the app generates, let me find the path for you
[11:36] <mzanetti> nerochiaro: ah right...
[11:36] <nerochiaro> mzanetti: logs are in ~/.cache/upstart
[11:37] <nerochiaro> mzanetti: do an ls and you should be able to find your app's log (the names seems different if it's a click app, or legacy, or something else)
[11:38] <t1mp> oSoMoN: cool, thanks.
[11:40] <oSoMoN> nerochiaro, I’m running the test on the device with phablet-test-run -v, and here is some interesting output I’m getting: http://pastebin.ubuntu.com/6514021/
[11:41] <oSoMoN> nerochiaro, clearly, it’s tapping at the wrong position on screen, the center of the only image in my gallery is more like at 230,640 on the screen
[11:41] <oSoMoN> nerochiaro, so I would say you’re not selecting the right object
[11:41] <nerochiaro> oSoMoN: hmm, i kind of copied that code from a gallery tests
[11:42] <nerochiaro> oSoMoN: does autopilot vis work for you ? that would simplify things a lot if i could poke around the object tree that way
[11:42] <oSoMoN> nerochiaro, wild guess: photo = grid.select_many("OrganicItemInteraction")[0] selects the item with the date, not the picture
[11:45] <nerochiaro> oSoMoN: if you check test_picker_mode.py in the gallery's tests, you have test_pick_first_photo(), which is where i took the code from. am i doing something totally different ?
[11:45] <nerochiaro> oSoMoN: or subtly different but wrong ?
[11:46] <oSoMoN> nerochiaro, I don’t know, I’m not familiar with the gallery-app tests, the code seems to make sense, but from my observations it’s clearly not selecting the right object
[11:46] <oSoMoN> nerochiaro, I’d say subtly different but wrong
[11:47] <oSoMoN> nerochiaro, if you dump the tree of objects to a file, you should be able to determine which item is the photo, by looking at the global rects
[11:52] <nerochiaro> oSoMoN: i tried that, dumps way too much stuff. i'll give it another shot and try to find the right item
[11:52] <nerochiaro> oSoMoN: it's really frustrating though that not even autopilot vis works
[11:53] <oSoMoN> yeah
[11:54] <oSoMoN> nerochiaro, try grepping on the globalRects, since you know more or less where the photo lives on screen (the dimensions of maguro’s screen are 720×1280)
[11:54] <nerochiaro> oSoMoN: i'll give that a shot
[11:58] <nerochiaro> om26er: any idea why running apps on the desktop with -testability fails with this error: "Library qttestability load failed!"
[11:59] <om26er> nerochiaro, probably due to incompatible versions of python-autopilot and autopilot-qt
[12:00] <nerochiaro> om26er: can that be fixed ? it's been around for weeks
[12:01] <om26er> nerochiaro, sorry, i got shifted to doing totally different things. I am going to email the autopilot guys, or can you send an email to the autopilot mailing list ?
[12:02] <nerochiaro> om26er: i'd appreciate if you could do that, as you know the problem and the tool better than i do
[12:02] <om26er> nerochiaro, Ok
[12:07] <nerochiaro> om26er: thanks
[12:22] <nerochiaro> oSoMoN: what was the function to dump the object tree ?
[13:57] <nerochiaro> oSoMoN: i don't know what's going on in the gallery, but printing the tree confirms that that OrganicItemInteraction is the right children of the grid to click on (it has inside the mousearea that reacts to the click and the overlay that signals the selection)
[13:58] <nerochiaro> oSoMoN: so either autopilot is reportig the wrong globalRect or the move function is moving the cursor to the wrong place
[14:01] <oSoMoN> nerochiaro, I notice that you’re first moving the cursor to the item, then clicking, how about just clicking on the item? not that I expect it will make a difference, but it would make for shorter code anyway
[14:07] <nerochiaro> oSoMoN: doesn't make any difference indeed
[14:08] <nerochiaro> oSoMoN: i'm kind of blocked by this to be honest
[14:09] <oSoMoN> nerochiaro, how many OrganicItemInteraction does the select_many(…) call returns? and how many images do you have in your library
[14:12] <nerochiaro> oSoMoN: i have one, and select_many returns one
[14:13] <oSoMoN> nerochiaro, and what’s the values of its global rect?
[14:18] <nerochiaro> oSoMoN: 24,195 216x216
[14:19] <oSoMoN> nerochiaro, is this consistent with where the photo appears on screen?
[14:20] <nerochiaro> oSoMoN: i would say so
[14:20] <oSoMoN> nerochiaro, just a random idea: add an image to your library, and print information about all the instances of OrganicItemInteraction returned by select_many(…)
[14:21] <nerochiaro> oSoMoN: ok
[14:25] <nerochiaro> oSoMoN: hmm, 24, 195 and 252, 195... that's fishy, i thought 24 was the x
[14:26] <oSoMoN> nerochiaro, yes, globalRect is x,y,w,h
[14:30] <nerochiaro> oSoMoN: ok, this is really weird: if i have 2 images, and click on the first, it works
[14:30] <oSoMoN> nerochiaro, could it be a race condition with the position of the photo? try adding a sleep(2) before clicking on the photo, and see if that makes a difference
[14:31] <nerochiaro> oSoMoN: already done, no difference
[14:31] <nerochiaro> oSoMoN: and if i delete one photo i have the same issue again, adding one more fixes it
[14:32] <nerochiaro> oSoMoN: but with three pictures it is broken again
[14:32] <nerochiaro> wtf ?
[17:04] <nerochiaro> bfiller: sent you an email with a lot of details about what is going wrong in these integration tests for browser content picker, and CC'd om26er . If you need any other detail to be able to take it from there let me know
[17:06] <bfiller> nerochiaro: great, thanks