On 8/6/08, Florian Pflug <fgp.phlo.org@gmail.com> wrote:
> Merlin Moncure wrote:
> > you missed the point...if your return type is a composite type that is
> > backed by the table (CREATE TABLE, not CREATE TYPE), then you can
> > 'alter' the type by altering the table. This can be done without full
> > drop recreate of the function.
>
> Which - at least IMHO - clearly shows that we ought to support
> ALTER TYPE for composite types ;-)
>
> Is there anything fundamental standing in the way of that, or is it just
> that nobody yet cared enough about this?
Yes, that would be really good idea. Although as I mentioned in previous
email, even for tables it works by accident, thus we need some invalidation
mechanism for both tables and types in PLs.
And then maybe even rettype (OUT param) change with CREATE OR REPLACE?
--
marko