Re: ALTER TYPE 2: skip already-provable no-work rewrites

On Mon, Feb 07, 2011 at 12:04:02AM -0500, Robert Haas wrote:
> On Sun, Feb 6, 2011 at 8:18 PM, Noah Misch <noah(at)leadboat(dot)com> wrote:
> >> Or how about passing an ObjectType? ?Then we could specify
> >> OBJECT_TABLE, OBJECT_FOREIGN_TABLE, or OBJECT_TYPE.
> >
> > Could this be done without a several-line blob of code at each call site to
> > determine the answer? ?If and only if so, this sounds better.
>
> Yeah, that's a problem. New thought: how about we go back more or
> less to the original coding, except replacing the second argument
> (only) with a Relation? In other words, callers will pass either a
> Relation (which might be a table or foreign table) or a type name.
> Not particularly elegant, but no worse than what we had before.

Sounds good. Thanks.