Tom Lane escribió:
> Kohei KaiGai <kaigai@kaigai.gr.jp> writes:
> > About ALTER FUNCTION towards aggregate function, why we should raise
> > an error strictly?
>
> I agree we probably shouldn't --- traditionally we have allowed that,
> AFAIR, so changing it would break existing applications for little
> benefit.
Okay, I have pushed the version I posted last week.
> Similarly, you should not be throwing error when ALTER TABLE is applied
> to a view, sequence, etc, and the command would otherwise be sensible.
As far as ALTER some-obj RENAME goes, this is already working, so I
haven't changed anything.
Thanks,
--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services