> Ok, I now understand more about how Postgres handles these older dates but...
It isn't quite as nice and clean as we'd like. For example, the cutoff
dates for assuming that time zone is correctly supported are arbitrarily
chosen to be near the Unix system time range boundaries of 1901 and
2038, on the theory that this is, well, an arbitrary boundary.
I'm not sure how to test for "correctness" of time zone support in a way
that gives better results than this assumption. We do know that some
timezone databases are incorrect before 1970, for example (AIX), and
that there are variations in content between different vendors (Sun does
a painstakingly accurate job afaict).
- Thomas
---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
