Tom Lane writes:
> No, that is *not* what happens. For dates outside the range known to
> the local platform's timezone database, we effectively assume that all
> times are given in GMT; no timezone corrections are applied on either
> input or output.
Oh, I wasn't sure whether Bruce meant the behaviour in 1850 exactly.
--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter