[00:53] balloons: still around? I'm trying the tmp home for reminders but I still get: [00:53] http://paste.ubuntu.com/8295014/ [00:54] elopio, I am [00:55] invalid option -s.. how are you launching? [00:55] balloons: we can ignore that one. I am writing a desktop file that add the -s [00:55] it gives the error, but launches with sandbox, so that part is working. [00:56] this is the unity log http://paste.ubuntu.com/8295034/ [00:56] it complaints about a missing desktop file hint. [00:57] all the logs are left on /home/phablet/.cache, so I think that ubuntu-app-launch is not using my tmp dir. [00:59] it's weird, the -s comes from the desktop file. So why does it need a hint if it's finding it. [01:04] elopio, hmm [01:05] can you share the code? [01:16] balloons: sorry, I have to go and most of it is just hacked into my phone. I'll get it to a branch when I return. [01:19] elopio, we'll chat tomorrow :-) === _salem is now known as salem_ === salem_ is now known as _salem [05:37] Good morning [07:23] has someone been playing with the image tracker site - if so can they unplay :( [07:23] filtering doesn't work here and you can't stop the testcase showing in results page [07:30] http://tinyurl.com/pvrox6p note the products turned off in filter, being ignored in the main frame === jibel_ is now known as jibel === _salem is now known as salem_ [13:15] elfy, you can blame knome and me for the playing [14:39] balloons: well I didn't like to apportion blame - but I had a fairly good idea :p [14:41] then you'll forgive me when it's still broken until tomorrow yes? :-) === salem_ is now known as _salem [14:47] I can do that - however no forgiveness will be forthcoming on Thursday :p === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [15:04] balloons: not sure if you saw the bug for vbox today, vboxvideo is the culprit seemingly - not loaded === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === _salem is now known as salem_ [17:26] balloons: this makes all the reminders tests pass: [17:26] https://code.launchpad.net/~canonical-platform-qa/reminders-app/fix1363599-upstart_and_sandbox/+merge/233832 [17:26] we still have the temp dir to solve, and we might simplify the desktop file, ted is thinking about that. [17:26] but those can be done on following branches. [17:28] elopio, we really require a desktop file eh? [17:29] balloons: that or to convince ted to pass arguments from ubuntu-app-launch to the exec. [17:29] I take it you tried? [17:55] balloons: sorry. Tried to convince him? [17:55] yes, he said that at best, it won't happen soon. But he doesn't like graphical apps to receive arguments on the command line. [17:59] balloons: you told me that the most recent version of the calendar was on the store, right? [18:08] elopio, when I told you that, it was correct :-) [18:08] elopio, let me share something with you [18:08] heck brendand might find this useful too [18:08] elopio, brendand https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AnZdnhOl8MU5dDJseW1vT1N5RkJvLUJHZTdhalRVd1E&usp=sharing#gid=1 [18:09] I'm sorry it never dawned on me before to share that. It automagically tracks everything core apps [18:09] so don't update any of the fields please :-) [18:10] it's weird that the logs show a version lower than 0.4.440 [18:10] but anyway we need a new release to get updated results. [18:11] elopio, sure, let me build and push one [18:11] thanks balloons. [18:14] popey pushed one this morning though looks like === roadmr is now known as roadmr_afk [18:35] elopio, can we brainstorm on fixing tests that have PAM authentication requirements on them (file manager and terminal) [18:36] balloons: sure. [18:36] this is a problem only on desktop, for now. [18:36] oh, no, filemanager and terminal. I was thinking just about credentials. [18:38] elopio, yea, ogra mentioned dev mode hits rtm tomorrow, so rtm will start blowing up too [18:39] balloons: I'm not sure what's the problem, I haven't payed a lot of attention to file manager and terminal. [18:39] do we need to unlock after opening them or something like that? [18:43] elopio, me neither :-) But PAM support was added. So if a device is locked, you don't get access to terminal at all, and file manager is REALLY limited access unless you enter the password or pin into the authentication box [18:44] balloons: I see a button called unlock for full access. We just need to know the password that CI is setting on the phones. [18:45] or alternatively, launch the app on an unsafe mode to run most of the tests. [18:45] btw, this button is really hard to click. It triggers the bottom edge gesture. [18:45] those are more or less the two options [18:45] the file manager requires more setup yea [18:45] the terminal throws it in your face :-) [18:46] I'm not sure how we could launch the app in an unsafe mode [18:46] balloons: I would like a combination of both. For tests that are not related to security, use the unsafe mode. And then do a small suite of tests with security. [18:46] elopio, yep, you mirror my thoughts [18:46] we should test the unlocking works [18:47] but everything else makes sense to run outside of it. The trouble is, I don't know if that's possible at all [18:47] I mean, not unless we send something to the app to not trigger PAM [18:50] I have to refresh my knowledge on things.. maybe it would be possible to mock that we don't have a pin on the device when we actually do [18:50] and vice versa of course [18:57] elopio, rev 441 of calendar is in the store [18:57] thanks balloons [18:58] balloons: we need help from the devs here. [18:58] maybe use an laternate desktop file that launches the app for testing purposes and works as before, if that's doable. [18:58] or a way to disable the password. We are running the tests with python, so the tests are not confined and they can do anything. [18:59] I don't know. But we need testability helpers for this. [18:59] I agree we need helpers :-) [19:01] balloons: do you agree it's not us the once who should define and implement them? [19:01] I think we should just push to get somebody to give us an app in a state that we can automate. [19:01] the helpers? well sadly it falls on my plate at least [19:01] you might be in the clear :- [19:02] this is a stop-gap solution until the PAM is implemented on a wider scale than inside individual apps [19:03] in theory we could then argue for the same reason those apps are required to have a testing mode if you will in lieu of use mocking something [19:03] elopio, I agree it's not QA who should decide how PAM is implemented [19:06] balloons: but what I'm saying is that if developers keep adding features like this locking mechanism, and they don't think of how to test them [19:06] we need to stop adding patches or solving the problems you and me. [19:06] next time they need to involve us on the planning, and now they need to add a testability mode for things. [19:15] testing - something that everyone agrees should be done - actions rarely speak as loud as the words [19:30] balloons: so according to ted, now we need a different approach instead of using a temp home dir. [19:31] one more thing to the bucket. [19:31] why don't we list all the testability issues that we are finding, and push high management to get somebody to fix them in a clever way? [19:33] elopio, you see the conversation happening in #ubuntu-touch? [19:33] I think you are right.. perhaps we should have pushed harder sooner on this [19:34] I mean in general on the testability stuff.. we've instead hacked and worked around it [19:42] balloons: more or less. I'm trying to have lunch. [19:42] I think one of the big mistakes is to run the tests with unity running. [19:42] but I haven't been able to convince everybody about this [19:42] elopio, I think this is a session for dc [19:42] I only hope we can wait so long eh? [19:43] it's a long wait. [19:43] but I would prefer to wait than to work it around again just to find a new wall. [19:46] elopio, we should still start a conversation internally.. even if we 'solve' it before then, it will still be a topic === roadmr_afk is now known as roadmr === salem_ is now known as _salem