Hmmm... This is an untested branch and seems obviously right, but I'm not sure how this has never come up before.

@lapointexavier Did this come up for you "in the wild", or did you figure it out from the docs?

CI failure is fixed in #602, when that is merged you can rebase against master to get a passing CI run.

It actually came out since we switched our dev environment to kubernetes. I imagine something's different in the environment as we were using vagrant before.

This block of code started failing and I had to drop in a debugger to see what was going on.

(code edited for simplicity)

LOCAL_TZ = dateutil.tz.gettz()

dt = dateutil.parser.parse('<some datetime>')
dt = dt.astimezone(LOCAL_TZ).replace(tzinfo=None)

dateutil.tz.gettz() would return None, and realized that the environment variable was the following:

echo $TZ
:/etc/localtime

Interesting, seems like just no one was hitting that branch before, which is interesting in itself.

For your current situation, here's a workaround:

LOCAL_TZ = dateutil.tz.tzlocal()

dt = dateutil.parser.parse('<some datetime>')
dt = dt.astimezone(LOCAL_TZ).replace(tzinfo=None)

Since you throw away the time zone immediately anyway, it doesn't really matter if you get a tzfile or tz.tzlocal.

@lapointexavier I merged #602, so if you want to rebase against master that will fix CI, then I can merge this one.

ah great, thanks!

@pganssle Should be good to go 🚀

@lapointexavier Oh, one other thing (probably should add this to the PR template when I get a chance), can you add yourself to the AUTHORS.md file? Goes in alphabetical order by first name. Put a D next to your name to indicate that your contributions have always been dual-licensed Apache/BSD. You can use any name you want, all fields are optional.

Nice catch.

@lapointexavier By the way, I masked those e-mail address domains because I hadn't asked permission from those authors to actually publish their e-mail address. If you want your e-mail address included, I can include it next time I make a change to AUTHORS.MD. I assume you realize this and didn't want your e-mail included, but I thought I'd make that clear.

@pganssle yeah I was a little ambivalent. I guess it wouldn't hurt anyway. Here it is: [email protected]