Robert Haas <robertmhaas@gmail.com> writes:
> So, I think that hacking on psql's query generation rules may well be
> a good idea, but shouldn't we also be bumping procost for the
> pg_whatever_is_visible functions? I mean, Stephen's information
> suggests that those values are pretty clearly wrong, regardless of
> anything else.
Yeah, I think we discussed that once before. I have no particular
objection to doing that in HEAD, just think it's a bit late for 9.2.
In any case, it will only help for medium-size numbers of entries;
once you get to enough tables/functions/whatever that a seqscan of the
catalog is bad news, only fixing the name matching is going to help.
regards, tom lane