[02:00] <veebers> thomi: you have a moment?
[02:00] <thomi> sure
[02:00] <thomi> what's up?
[02:01] <veebers> thomi: Do you agree that this test does what we described the other day (WRT the datetime + tz tests)? http://bazaar.launchpad.net/~veebers/autopilot/adding_timezone_timestamp_tests/revision/514
[02:02] <veebers> thomi: this is against trunk and passes, it fails when used on the datetime changes of balloons (that I'm taking over)
[02:03] <thomi> veebers: no, that's not right
[02:03] <thomi> I don't think
[02:03] <thomi> hmmm
[02:03] <veebers> thomi: my thought process:
[02:04] <veebers> given a known timestamp, create a qml datetime object and compare what we get back down the line against a locally created datetime w/ the same timestamp
[02:04] <thomi> veebers: ok, a few things.
[02:05] <thomi> first, I'm not convinced that setting 'TZ' does what you expect - maybe it does, but I'd like to see a test proving that first.
[02:05] <thomi> Second, that's not how users will ever compare the times in their autopilot tests, so it's not a great representation of what we're trying to fix
[02:05] <thomi> Third, this doesn't mention times outside the 32 biut boundary - shoul dit?
[02:05] <thomi> *should it
[02:05] <thomi> that's all I have right now :D
[02:07] <veebers> thomi: Right, I did some manual testing for the TZ part (but nothing seen here) and it does what I expect. 2nd good point re: user comparison. 3rd No doesn't test outside the 32bit yet, but is intended to prove the correctness of the DateTime types (and thus soon to prove the >32bit datetime stuff)
[02:07] <veebers> I'll update it now
[02:07] <veebers> thanks thomi :-)
[02:08] <thomi> hmm, yes
[02:08] <thomi> I think also we need to do some manual testing before we try and land this
[02:08] <thomi> get balloons to try it
[02:08] <thomi> and maybe others in different timezones
[02:12] <veebers> thomi: yeah sounds good.
[03:20] <veebers> thomi: if you have a moment, are these changes a step in the right direction (ignoring pt3 of yours at this stage); http://bazaar.launchpad.net/~veebers/autopilot/adding_timezone_timestamp_tests/revision/515
[03:23] <thomi> veebers: that looks to me like you're testing the incorrect behavior
[03:23] <thomi> veebers: if the qml property says:
[03:24] <thomi> property date testingTime: new Date('2014-01-15 12:34:52');
[03:24] <thomi> then in autopilot I should be able to do:
[03:24] <thomi> date_object.day == 15
[03:24] <thomi> date_object.hour == 12
[03:24] <thomi> tht should *never* change
[03:24] <veebers> thomi: hmm, good point, will check that out