[PATCH] Allow microseconds in parse_date() input

[Alex Willmer <al.willmer@…>]

The function trac.util.datefmt.parse_date() accepts ​ISO:8601 strings. ISO-8601 allows for fractional seconds, but parse_date() cannot handle these. Attached is a patch to remedy this - against trunk, with tests.

The change was useful for converting between the _ts used by ​th:XmlRpcPlugin ticket.update() and the ts used by TicketModule

[Christian Boos]

Looks good!

As you introduced [T ] for the date/time separator, maybe you should also add a test for that…

[Alex Willmer <al.willmer@…>]

Also test space as date/time seperator

[Alex Willmer <al.willmer@…>]

Reading ​http://en.wikipedia.org/wiki/ISO_8601#Time_offsets_from_UTC more closely

The offset can also be used in the case where the UTC time is known, but the local offset is not. In this case the offset is "-00:00", which is semantically different from "Z" or "+00:00", as these imply that UTC is the preferred reference point for those times.

I'm sure any more that the following tests are valid

self.assertEqual(t_utc, datefmt.parse_date('2009-12-01T11:00:00Z'))
self.assertEqual(t_utc, datefmt.parse_date('2009-12-01T11:00:00+00:00'))
self.assertEqual(t_utc, datefmt.parse_date('2009-12-01T11:00:00-00:00'))

[Christian Boos]

Replying to Alex Willmer <al.willmer@…>:

(about) ​http://en.wikipedia.org/wiki/ISO_8601#Time_offsets_from_UTC … I'm sure any more that the following tests are valid

self.assertEqual(t_utc, datefmt.parse_date('2009-12-01T11:00:00Z'))
self.assertEqual(t_utc, datefmt.parse_date('2009-12-01T11:00:00+00:00'))
self.assertEqual(t_utc, datefmt.parse_date('2009-12-01T11:00:00-00:00'))

I think the tests are correct. IIUC, the semantic difference is about the fact that for the first two formats, UTC was a deliberate choice, but for the last format, UTC was used by default, lacking information about the local time. But all three refer to the same time.

​RFC:2822#section-3.3 has perhaps a better wording when it explains its similar convention:

The form "+0000" SHOULD be used to indicate a time zone at Universal Time. Though "-0000" also indicates Universal Time, it is used to indicate that the time was generated on a system that may be in a local time zone other than Universal Time and therefore indicates that the date-time contains no information about the local time zone.

[Remy Blank]

Preparing for 1.0.

[Christian Boos]

Will apply for 1.0.

[Jun Omae]

It would be good to allow also less than microsecond precision (e.g. 2012-07-11T12:34:56.098). repos:jomae.git:ticket10644

[Christian Boos]

Looks good, but at this occasion I noticed the following:

 (?:[,.]?(\d{1,6}))?)?)?             # microseconds
        ^

I think the decimal mark shouldn't be optional (​ISO_8601#Times).

[Jun Omae]

Thanks for the pointer! [4280a17d/jomae.git].

[Jun Omae]

Applied in [11107]. Thanks, Alex Willmer!