Re: Reject access to unlogged tables in contrib modules

On Wed, Nov 04, 2026 at 09:06:14AM +0100, Jim Jones wrote:
> Since the message is not wrong, I'd vote to keep it as is. This patch
> set keeps growing and I'd not like to take too much time from the
> committers for yet another cosmetic change.
>
> PFA v4. 0002-0005 are unchanged.

Note that for pg_visibility, read_stream has some protections when it
comes to temp tables of other sessions, but that does not fly under
!include_pd. Merged 0004 and 0005 together, tweaked a bit the whole,
and applied.

pgrowlocks (uses read_stream, safe for temp from other settions,
complains about recovery) has a single point of failure, with a weird
error message. Done as well.

Regarding pg_freespacemap, I am not convinced. Returning 0 reflects
also the reality: there is no FSM for the temp-from-other-session case
and temp/unlogged at recovery.

For pageinspect, we have a pageinspect.h, where I'd suggest to have a
common routine for the whole to remove the duplication. For
pgstattuple, we don't have a "common" path, but that seems worth
having to avoid duplicating this kind of check all over the place.
--
Michael