Alvaro Herrera <alvherre@dcc.uchile.cl> writes:
> Right. But note that Tom wants to distinguish between statements
> created via PREPARE (which would rollback) from those created via a
> Prepare message (which wouldn't).
Actually, no, I'd prefer not to make such a distinction; I'd be happy
with SQL-level PREPARE being nontransactional. I'd be willing to put up
with that distinction if someone shows it's needed, but so far there's
not been a really good argument advanced for it, has there?
regards, tom lane