Re: CompactCheckpointerRequestQueue versus pad bytes

Robert Haas <robertmhaas@gmail.com> writes:
> With respect to this chunk:

> +  * We do not need to go through this dance for temp relations, though, because
> +  * we never make WAL entries for temp rels, and so a temp rel poses no threat
> +  * to the health of a regular rel that has taken over its relfilenode number.

> ...I would say that a clearer way to put this is that temporary
> relations use a different file naming convention than permanent
> relations and therefore there can never be any confusion between the
> two.

Yeah, that's an entirely independent reason why there's probably no
issue in recent releases.  The rationale as stated is back-patchable
to earlier releases, though.

BTW, I wonder whether the code that checks for relfilenode conflict
when selecting a pg_class or relfilenode OID tries both file naming
conventions?  If not, should we make it do so?
        regards, tom lane