Re: Infinity and beyond

On Thu, 5 May 2005, Donald Fraser wrote:

> A while back I noted in this thread that some people were having issues
> with timestamps and infinity. I have been using a modified version of
> the 7.4 driver for some time now and have had no real issues with these
> modifications. The modifications that I made involve adding two
> additional classes to the driver package. They are simple classes that
> extend java.sql.Date and java.sql.Timestamp. In my opinion this solution
> is much more elegant than relying on some magic number used in the
> underlying java.util.Date class. The main beauty of these modification
> are that you do not need to rely any PostgreSQL driver specifics to
> detect values of infinity, and on a similar note the driver doesn't have
> to rely on its own implementations of infinity, it can handle third
> party implementations. If anyone is interested in these modification I
> will post them on this thread, just ask. Cheers Donald Fraser.

How can clients not have any reliance on pg driver specifics? If these
new classes are added to the driver, won't that just move the dependence
from PGStatement to PGTimestamp? Have you seen what the 8.0+ driver
does for infinity or are you only familiar with 7.4's handling? In any
case, I'm interested to see what you have done.

Kris Jurka