Well, actually the question of whether failing referential actions
due to permission deficits of the user doing the delete/update
on the pk table is a bug or feature still stands. It would be
fairly trivial to extend Peter's patch to effectively setuid on
the actions, but the question is whether or not that's useful and
correct.
Stephan Szabo
sszabo@bigpanda.com
On Sat, 14 Oct 2000, Bruce Momjian wrote:
> Oh, OK. I will forget it.
>
> >
> > Actually Peter did a patch for this fairly recently I
> > believe. I haven't grabbed CVS recently enough to know
> > if it got committed. There's a related question of what
> > permissions you need to follow referential actions (currently
> > it's the same permission as if you were doing the implied
> > statement).