"Jim Buttafuoco" <jim@contactbda.com> writes:
> Why doesn't the regression test set the timezone to GMT instead of PST. I believe the horology test would just work
then.
... and lose any semblance of testing DST-aware behavior. Lobotomizing
the regression test isn't the answer.
(What is the answer, IMHO, is to make the interval datatype able to
distinguish "1 day" from "24 hours".)
regards, tom lane