Robert Hentosh <hentosh(at)io(dot)com> writes:
> After the daylight savings time change I have had similar failures
> with 7.1RC2 and the latest CVS of 7.1. It failed on both timestamp
> and horology regression tests. This happened on both my RH7.0 and
> OpenBSD 2.8 (RH7.0 on an Athlon and BSD on a P3 laptop)
Hm. The timestamp diffs are an expected behavior near DST transition
days --- see
http://www.postgresql.org/devel-corner/docs/postgres/regress.html#AEN14359
However, the horology diffs are not, and I can't reproduce them here.
Did anyone else see that?
regards, tom lane