Re: Timestamp Conversion Woes Redux

On Tue, 19 Jul 2005, Dave Cramer wrote:

> What about creating two extension classes
>
> PGTimestamp, and PGTimestamptz
>
> then allowing setObject to recognize these internally and bind to
> Oid.Timestamp, and Oid.Timestamptz repectively
>

Who is going to really use these unportable things instead of standard
Timestamps?  People who are desperate to get their application to work
around driver bugs?  I don't believe timestamp/timestamptz is an
intractable problem and in the previous discussion I recall suggesting
setting the server time zone to match the client's would fix this problem.
I also seem to recall a number of other timezone related issues.  Let's
focus on trying to fix what we've got before we go down this unportable
route.  I'll put together at test cases for the bugs I'm aware of tonight
so we can focus this discussion on the real problems.

Kris Jurka