Closed vincentdephily closed 5 years ago
Actually, testing further shows that 0.1.2 didn't support 2018-04-03T00:00:00
either (it can't make sense of the T
, but parses a space correctly) and ignored the +00:00
(it assumes the local timezone). Parsing T
and the timezone are desirable features, but not a regression (should a separate issue be opened ?).
Ok there's arguably no regression in the support of YYYY-MM-DD HH:MM:SS
: it's just the timezone part that used to be ignored but now raises an error. I'd say 0.1.3 has the correct behaviour, because we don't want the user to think that the tz has been parsed when it hasn't. I'll close this and open different issues for parsing T
and +HH:MM
.
This prints
Ok(2018-04-03T00:00:00+01:00)
with chrono-english 0.1.2 butErr(DateError { details: "iden expected, got Char(\'+\')" })
with chrono-english 0.1.3.Even adding a
T
in the input string (making it strictly iso86010-compliant)... I know this crate's purpose is to parse weird dates, but it should also be able to parse standard ones ;)This happens with my system configured with
LC_TIME=en_DK.UTF-8
as well asLC_TIME=en_US.utf8
.For what it's worth, I caught this with the unit test at https://github.com/vincentdephily/emlop/blob/master/src/main.rs#L302