DateTime.toZone().strftime() does not apply to returned call of toZone()

[Charles Choi]

The call toZone('GMT') should return a DateTime object whose setting reflects the new time zone. However calling strftime() on the new DateTime object returns the old time value.

            expireDate = DateTime(t, 'US/Pacific').toZone('GMT')
            expires = expireDate.strftime('%A, %d-%b-%y %H:%M:%S ') + 'GMT'

[Paul Winkler (slinkp)]

If I understand https://bugs.launchpad.net/zope2/+bug/142784 correctly, this is intended behavior.

I'm not sure *why* that's intended though; I find it highly counterintuitive.
I guess the idea was that it should work the same as python's time.strftime(), but that seems like a specious model to me, since the latter is working with struct_time inputs that carry no timezone information, so interpreting "%Z" as "local system timezone" seems reasonable for them.

I also notice that DateTime.strftime() prints nothing for the "%Z" character; maybe a clue to avoid using it :-p

IMO, a better model would have been datetime.datetime.strftime() which *does* respect the input's zone:

>>> import datetime
>>> import pytz
>>> d2 = datetime.datetime.now(pytz.timezone('US/Eastern'))
>>> print d2.strftime('%H:%M %Z')
18:01 EDT
>>> d3 = datetime.datetime.now(pytz.timezone('US/Pacific'))
>>> print d3.strftime('%H:%M %Z')
15:02 PDT

This seems much more sane than the analogous operations from DateTime:

>>> d4 = DateTime(d3).toZone('US/Pacific')
>>> print d4
2008/04/25 15:02:19.079246 US/Pacific
>>> print d4.strftime('%H:%M %Z')
18:02

Are we stuck with this bizarre behavior for legacy reasons?

If so, Charles, you might have to give up and use something other than DateTime.strftime().
Here's one handy hack - convert to a datetime.datetime:

>>> print d4.asdatetime().strftime('%H:%M %Z')
15:02 PST