Andres Freund <> writes:
> FWIW, I think we're getting pretty close to the point, or are there
> even, where we can remove default_with_oids. So not adding complications
> because of it sounds good to me.

Well, pg_dump uses it --- so the argument about not breaking old dump
scripts would apply to any attempt to remove it entirely.  But I don't
have a problem with saying that its semantics are frozen.

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to