[00:07] <balloons> rpadovani, that time it failed though.. before the app never launched
[00:10] <balloons> rpadovani, wow, look at how slow it runs ;-) http://91.189.93.70:8080/job/generic-mediumtests-vivid/940/artifact/ubuntu_calculator_app.tests.test_main.MainTestCase.test_operation_on_infinite_number.ogv. Still these look correct. You'll have to change the asserts in the tests to match the new significant figures you calculate too
[15:59] <MacSlow> Why are dash and unity8 just two small square windows when an autopilot-test is started on the desktop? What's missing to have them use the usual 40x71 GU dimensions?
[16:17] <MacSlow> elopio, ^
[16:18] <elopio> MacSlow: hum, that I don't know. Does the same happen when you do initctl start unity8?
[16:18] <MacSlow> elopio, yes
[16:19] <elopio> MacSlow: when I do that I get the two windows, with space enough for three columns of apps.
[20:31] <thomi> hey barry, got time for me to hassle you about packaging *again*? :D
[20:34] <barry> thomi: it's a brand new day, so sure!
[20:34] <thomi> barry: so, I get one question a day? :D
[20:34] <thomi> curses! did I just use it up?
[20:35] <barry> bzzt!  that's two
[20:36] <thomi> heh
[20:36] <thomi> barry: so, I've used pybuild before several times, and never had this issue...
[20:36] <thomi> selenium ships some binary files which are listed in the data_files section of setup.py
[20:37] <thomi> but pybuild can't see them - I suspect because they're not copied into the temporary directory perhaps
[20:37] <thomi> Any ideas? The branch I'm working from is lp:~thomir/+junk/selenium-packaging
[20:38] <barry> thomi: make sure there's a MANIFEST.in file that names those data files, otherwise they may not get into the orig.tar.gz
[20:39] <barry> that's always been what bites me
[20:39] <thomi> ahhhh
[20:39] <thomi> they're not :D
[20:40] <thomi> barry: actually, they are, I missed them
[20:40] <thomi> "recursive-include py/selenium/webdriver/firefox/x86 *.so"
[20:40] <barry> thomi: so if you `python setup.py sdist` they end up in the .tar.gz?
[20:41] <thomi> barry: yes
[20:41] <thomi> barry: the error message I get is:
[20:41] <thomi> error: can't copy 'py/selenium/webdriver/firefox/x86/x_ignore_nofocus.so': doesn't exist or not a regular file
[20:42] <barry> oh, .so files.  those might be getting filtered out by other bits of the debian build tools.  `export DH_VERBOSE=1` in your d/rules file.  so it is (almost?) never correct to include .so files source packages, those just might be getting filtered out
[20:42] <thomi> I assume that a .so counts as a 'regular' file - it's not a socket or anything
[20:42] <barry> thomi: where do you get that error?
[20:42] <thomi> I have verbose set, I don't see any filter messages
[20:42] <thomi> let me pastebin, one second
[20:43] <thomi> barry: http://pastebin.ubuntu.com/10041069/
[20:43] <thomi> further up I get:
[20:43] <thomi> warning: no files found matching '*.so' under directory 'py/selenium/webdriver/firefox/x86'
[20:44] <thomi> which I guess is a sign of the root cause...
[20:44] <barry> yep, i see that
[20:57] <thomi> I assume that, if those files were being filtered there'd be some sort of log message?
[22:33] <thomi> barry: I guess you're stumped too? Who else should I bug? I'm tempted to email the DPMT ML, but I suspect I'll get the "don't put .so files in your source package" answer :(
[22:41] <barry> thomi: probably so.  but iirc there's a debian/binary-file-override-magic file that might do the trick.  right now i can't remember what it's called and am putting out some other fires :(
[22:41] <thomi> barry: ahh, ok, no worries
[22:41] <barry> thomi: debian-python@ is your best bet