Joe Abbate <jma@freedomcircle.com> writes:
> Yes, I suspected that an OID was stored. What I'd still quibble with is
> the use of the ambiguous regproc in pg_operator (also pg_type) and the
> still-ambiguous schema-qualified proc name. I guess it's not feasible
> (at least, short term), but it'd be preferable to store a "raw" OID and
> let the user cast to regprocedure (or change the 'regproc' to
> 'regprocedure').
Yeah, ideally those columns would be regprocedure. There are
bootstrapping problems involved though with populating the initial
contents of the catalogs during initdb --- the regprocedure input
function doesn't work in that environment. (It might be possible to
hack something for pg_operator, but the circularity is rather
fundamental for loading pg_type, since the input function would need to
consult pg_type to make sense of argument types.)
In the meantime I'd suggest casting the columns to regprocedure when
querying, if you want unambiguous results. That's what pg_dump does.
Or you can cast to OID if you like numbers.
regards, tom lane