On Fri, 2002-08-16 at 01:15, Tom Lane wrote:
> Rod Taylor <rbt@zort.ca> writes:
> > It should be noted that my patch in the queue displays the actual
> > foreign key definition and suppresses 'Constraint' Triggers.
>
> > This is a touch better than the below with the side effect that if a
> > user issues a CREATE CONSTRAINT TRIGGER manually it is mistakenly
> > hidden.
>
> I was thinking of looking into pg_depend to discover whether a trigger
> is associated with a constraint, and hiding the trigger if and only if
> we have a constraint to show instead. Anyone want to pursue that?
I had assumed that since the keyword CONSTRAINT isn't documented on the
CREATE TRIGGER page, that the check for isconstraint would be enough?
Erm, except legacy stuff would be hidden.
Sure, I'll fix it up.
