Re: Wishlist?

# mail(at)etolnay(dot)com(dot)ar / 2005-07-22 12:41:25 +1000:
> Tom Lane wrote:
> >Bruno Wolff III <bruno(at)wolff(dot)to> writes:
> >>What happens if there is more than one existing function with that name
> >>already. Do all of the old functions get deleted?
> >
> >What happens if there are existing references to the function using the
> >old parameter list?
> >
> >The current approach is that if you want to change the parameter list or
> >result type, you have to issue DROP FUNCTION, with the resulting cascade
> >of any dependent objects. I think trying to change that would be a very
> >bad idea. It'd create a real mess in terms of consistency of views, and
> >what does it really buy you notationally over DROP/CREATE?
>
> What I see is that the overloading is a very cool feature, but is not
> necessary in most cases, and it introduces unnecessary administration
> hassles. If there would be an attribute stating if a procedure has been
> created overlodable (enabled by default, for backwards-compabitility),
> then CREATE OR REPLACE would always replace the current one, regardless
> of the parameters, and DROP would not require any parameters at all, and
> a creation of a function with the NOOVERLOAD attribute would not be
> allowed if there is already a function marked for overloading.

What would your hassle-free CREATE OR REPLACE do in this situation?
What should the fiew look like after you replace foo() with
foo(int4)?

CREATE TYPE t1 AS (
a INTEGER,
b INTEGER,
c INTEGER
);
CREATE TYPE t1 AS (
a TEXT,
c TEXT
);

CREATE FUNCTION foo() RETURNS SETOF t1 ...

CREATE VIEW fiew AS SELECT a * b, c FROM SELECT foo();

CREATE OR REPLACE FUNCTION foo(INTEGER) RETURNS SETOF t2 ...

--
How many Vietnam vets does it take to screw in a light bulb?
You don't know, man. You don't KNOW.
Cause you weren't THERE. http://bash.org/?255991