Chris Studholme <cvs(at)cs(dot)utoronto(dot)ca> writes:
> What follows is not necessarily a bug, but may be a misinterpretation of
> the SQL standard.
Yeah, it's a bug; the implementation of row comparisons in PG is
completely bogus. (The parser just expands it out to an AND clause
of scalar comparisons, which works for = but not for < or >.) This
is (or should be) on the TODO list, but no one has gotten around to
it yet.
regards, tom lane