Re: [HACKERS] update_pg_pwd

wieck(at)debis(dot)com (Jan Wieck) writes:
> One last point though. The comment says it's using lower case
> name now to be callable from SQL, what it isn't because of
> it's Opaque return type in pg_proc.

> pgsql=> select update_pg_pwd();
> ERROR: typeidTypeRelid: Invalid type - oid = 0

> Is that a wanted (needed) capability or should I better
> change the comment to reflect it's real nature?

What would you expect the SELECT to produce here? I think the error
message is pretty poor, but I can't really see OPAQUE functions being
allowed in expression contexts...

I don't really like the description of these functions as returning
something "OPAQUE", anyway, particularly when that is already being
(mis) used for user-defined type input/output functions. I wish
they were declared as returning something like "TUPLE".

regards, tom lane