> Tatsuo Ishii <t-ishii@sra.co.jp> writes:
> > I'm going to change contrib/pgstattuple so that it returns a tuple
> > rather than a NOTICE using new SRF interface. I believe this way is
> > much more convenient for users. Any objection?
>
> Sure. I like the way Joe did show_all() better than the way Neil
> did lock_status() --- I'm in the middle of getting rid of zero
> type OIDs throughout the system, and so I'm not happy with anything
> that re-introduces them, even if it's only transient during initdb.
> Other than that minor point, go for it...
I think pgstattuple() never uses OPAQUE...
--
Tatsuo Ishii