[14:18] <meena> looks like even the simplest growpart test will fail, unless I dig deeper
[20:31] <holman> falcojr: nice job on that tz fix 
[20:31] <falcojr> thanks!
[20:31] <holman> falcojr: i completely overlooked the fact that date was emitting in non-utc
[20:32] <holman> I saw before that we were parsing without tz, but I didn't put 2 and 2 together
[20:32] <holman> falcojr: putting up a ci change now to run unittests against alpine with a tz set
[20:32] <falcojr> sounds good
[20:46] <minimal> I'm in UTC (at least for part of the year) so it would have been hit-and-miss whether I'd see in in my local builds
[21:26] <meena> heh, i got all of my servers in UTC, but my laptop is in… minimal's TZ
[22:01] <minimal> this TZ ain't big enough for the both of us! ;-)
[22:52]  * meena should really have her own TZ
[23:34] <meena> holman: i'm not sure right now we ever actually end up in this codepath, but this seems wrong: https://gist.github.com/igalic/3c343d2766ad910ede38b7ecfea8e05e (this is caused by this line: https://github.com/canonical/cloud-init/blob/main/cloudinit/distros/bsd.py#L172
[23:52] <meena> we should either just return the input at that point (bail) or raise some exception (bail very loudly)
[23:52] <meena> as it stands, we're making a call to the parent class, which is… to say, we're calling out to Linux
[23:59] <meena> to be clear, this is bogus input for that function that should never get there, but it's good I tried it, or i wouldn't have discovered that the function itself is broken in unexpected ways