Álvaro Herrera <alvherre@kurilemu.de> wrote:
> Hmm, this is actually a problem in 1bd6f22f43ac. I added a preliminary
> patch that should fix this.
I've reviewed this patch set too. The only question that occurs to me is
whether INSERT_SKIP_FSM and INSERT_FROZEN shouldn't actually be prefixed with
HEAP_, as these are rather low-level and thus might be considered
AM-specific.
(I like the idea to use the high bits to avoid collisions.)
--
Antonin Houska
Web: https://www.cybertec-postgresql.com
