On 7/28/2014 11:49 AM, Jan Nijtmans wrote:
2014-07-28 17:10 GMT+02:00 Igor Tandetnik <i...@tandetnik.org>:
All your fix does is have the parser accept "60" as valid seconds field.
That's not very interesting.
Yes, that's exactly all that I'm after. ISO 8601 does not specify how
leap seconds are handled
... whereas SQLite does need to handle them, somehow, once they are allowed.
It's perfectly OK for fossil/SQLite to ignore the leap second, it's not OK that
an error-message is produced
for a valid ISO 8601 time-stamp.
What do you mean "ignore"? Can you specify precisely what the semantics
of such a timestamp should be, in all the date/time manipulation
operations SQLite supports? Say, what should be the value of
julianday('2012-06-30T23:59:60'), and how should it compare with
julianday('2012-06-30T23:59:59') and julianday('2012-07-01T00:00:00') ?
--
Igor Tandetnik
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users