Re: timestamp (minor bug)

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: "And. Andruikhanov" <andy@euinf.dp.ua>
Cc: pgsql-bugs@postgresql.org
Date: 2001-07-22T15:51:05Z
Lists: pgsql-bugs
"And. Andruikhanov" <andy@euinf.dp.ua> writes:
> Wrong result in timestamp value between "winter" and "summer" time:

Could you trace through the timestamp conversion routines and determine
exactly what's going wrong?  The problem is most likely in or near
DetermineLocalTimeZone() in src/backend/utils/adt/datetime.c.  We've
heard some reports of misbehavior of the C library's mktime() routine
on certain platforms, when given a time near a DST boundary in certain
timezones (what timezone do you use, anyway?).  DetermineLocalTimeZone
attempts to defend itself against this, but perhaps is not completely
successful.  Unfortunately, none of the developers have a machine on
which mktime() misbehaves, so it's hard for us to figure out much more
about the issue...

			regards, tom lane