=== steve__ is now known as sbeattie === chihchun_afk is now known as chihchun [06:17] Good morning [07:32] Good morning [08:09] jibel: I'm scared to say that slowly, but steadily I understand the autopkgtest code.. [08:10] pitti, argh, that's horrible, does it hurt? ;) [08:10] jibel: it does.. [08:11] jibel: I spent several hours trying to understand what all these AutoFile classes do, adding comments to document it, and now I remove their usage step by step :) [08:13] pitti, oh, I'm sorry, I understand your pain. [08:13] I already got rid of some of the confusing output/logging/Errplumb stuff [08:13] jibel: but it seems this is becoming a more and more important infrastructure piece, so it's probably better if more than just two people in the world know how this works :) [08:14] s/know/can understand/ [08:15] pitti, agreed. From what I understood of the code, the logic is not really complicated but the code is somewhat obscure [08:15] and it's an understatement [09:14] good morning all [09:26] hey DanChapman, how are you? [09:44] pitti, simplejson was the package TheMuso was referring to in his patch pilot summary this morning? [09:44] Morning DanChapman [09:44] jibel: I don't know, I'm afraid [09:44] it's sad because this package has the testsuite of +110 tests [09:45] I'll enable autopkgtest and reply to the list [09:46] jibel: I guess even a 'python -c "import simpleson"' test would already help a lot :) [09:46] jibel, in this case definitely [09:46] pitti: but running the full test suite is even better, of course [09:46] (SCNR) [09:48] pitti, do we have a preferred runner for python? [09:48] jibel: what do you mean with runner? [09:48] pitti, nose, pytest, unittest, ... [09:49] not for autopkgtest [09:49] pitti, in simplejson if I call the test script directly it'll use the source tree [09:49] whatever upstream uses [09:49] pitti, but if you use another runner it will use the package as installed [09:49] oh, surprising; usually that's the easiest way to make it run against the system libs [09:50] jibel: whatever is easiest [09:50] they do a sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))) [09:50] which point to the root of the source tree [09:50] +s [09:51] oh, eww [09:51] how ugly; that sohuld be done by setting PYTHONPATH in the source tree test runner [10:18] ls [10:28] Morning all [11:02] good morning all === chihchun is now known as chihchun_afk [11:26] hey pitti, sorry I didn't see your message pop up. I'm good thanks how are you? [11:26] morning jibel o/ [11:27] DanChapman: heh, no problem; I'm fine, thanks! [11:31] pitti, with UDD should I update changelog or it is done automatically from the commit message when the branch is merged? [11:34] jibel: you should still write debian/changelog, and then use debcommit [11:34] (that'll build the bzr changelog from debian/changelog) [13:25] pitti, https://code.launchpad.net/~jibel/ubuntu/trusty/simplejson/enable_autopkgtest/+merge/195380 when you have a minute [13:26] jibel: I haven't seen py.test yet, that's a runner like nose? [13:26] jibel: how does that work around the sys.path() setting? [13:27] pitti, yes it is similar to nose [13:27] I guess I would just have copied the tests to $ADTTMP and run them from there, but if that's working as well, it's fine [13:27] pitti, path is set in main of __init__ in the tests which is not called if you use the runner [13:27] ah [13:28] jibel: can you please forward it to Debian? [13:28] otherwise setup.py just calls __init__.py to execute them [13:28] right [13:28] and uses unittest [13:28] I'll merge/upload in the meantime [13:29] I'll forward to debian and reply to luke [13:29] thanks [13:31] pitti, FW is finally open, I'll reenable cron in the lab which has not been copied from the previous machine and autopkgtest should start again [13:31] yay [13:31] jibel: merci [13:31] well, that's theory of course :) [15:52] balloons: I've just been informed that my day's of next week have been cancelled, I see no way to unregister from uds sessions - so you'll just have to ignore my name there [15:53] elfy, ugh [15:53] well, no worries about your name being there === wylde is now known as Guest64913 [20:10] balloons, hello [20:11] balloons, I saw you fixed rssreader-app test, thanks, its nice to see it work both on the dekstop and device [20:11] Letozaf_, howdy [20:11] balloons, fine what about you ? [20:13] Letozaf_, good.. there was a package issue that prevented the merge bot from running till just now [20:13] so i'm reviewing the run [20:14] looks like a dependency issue.. i fixed and pushing again :-) [20:14] balloons, ok [20:18] balloons, it passed! [20:18] hehe ;-) [20:20] Letozaf_, you can do the approval honors ;-) [20:20] balloons, ok thanks [20:21] Letozaf_, note I did disable the test_view_mode_and_feed_item test. I couldn't get it to switch tabs properly. Need to ask elopio about it [20:21] balloons, yes I saw that [20:22] dat timing [20:22] * balloons waves to elopio [20:24] elopio, I wanted to ask you about why I couldn't get the emulator to switch tabs in rssreader for a test. the tabs are dynamically generated.. the emulator sees no tabs present [20:24] also, I can't switch to a tab by name anymore, only objectname. naturally, these will only have dynamic names, not objectnames