bug#72570: 31.0.50; Regression in date-to-time

[Eli Zaretskii <eliz@gnu.org>]

> From: Ulrich Mueller <ulm@gentoo.org>
> Cc: eggert@cs.ucla.edu,  72570@debbugs.gnu.org
> Date: Mon, 12 Aug 2024 15:48:04 +0200
> 
> >>>>> On Mon, 12 Aug 2024, Eli Zaretskii wrote:
> 
> >> Well, parse-time-string doesn't make any assumptions about the timezone.
> >> It will return a list with nil in its TZ element, if it cannot extract
> >> any timezone info from the string.
> >> 
> >> Same for iso8601-parse, the returned TZ is nil if it is not present in
> >> the argument string.
> 
> > Yes, but what does nil as timezone mean, when you later interpret the
> > time values?  Is it UTC or is it local time?
> 
> AFAICS it has its usual meaning (same as in ZONE of encode-time), i.e.
> nil for local time.

If this is true in all use cases (and I'm not sure, as we just
discovered in this case), it should be documented, and we should have
tests in our test suite to make sure we never regress in this aspect.