On Sun, 11 Aug 2002, Tom Lane wrote:
> Gavin Sherry <swm@linuxworld.com.au> writes:
> > On Sun, 11 Aug 2002, Tom Lane wrote:
> >> Why would we need that, given that we now have CREATE OR REPLACE
> >> FUNCTION? It's not like anything could depend on a trigger
> >> (other than FK constraints, and I don't think I want people replacing
> >> trigger definitions that are part of an FK constraint...)
>
> > I thought it conceivable that one may want to point the trigger at a
> > different function, change BEFORE or AFTER, or change the event(s) upon
> > which the trigger fires.
>
> Given the lack of popular demand (I don't think I've yet heard anyone
> ask for this feature), I can't get excited about expending code bulk
> and complexity on it --- the patch seems to make CreateTrigger
> significantly messier, and that will cost us in maintenance effort
> down the road.
From memory, I think some of the pgadmin guys jumped up and down about it
for a while: http://fts.postgresql.org/db/mw/msg.html?mid=1353289.
Gavin