Peter Eisentraut <peter_e@gmx.net> writes:
> I can't believe we're operating in ground-breaking
> territory here.
We're not. This functionality has always been in Postgres, right back
to the PostQUEL days. Joe is trying to clean it up to the point where
it has explainable, defensible semantics. But he's not adding something
that wasn't there before.
> What concerns me is that we're inserting table-generating
> syntax elements into the select list, which is where they've never
> belonged.
Do you see another way to pass non-constant arguments to the
table-generating function?
regards, tom lane