On Sat, 3 Jul 2004, Tom Lane wrote:
> trigger.c: not at all sure about the model for handling trigger firing
> status. It looks like a subtrans could fire triggers that were pending
> from an outer xact, which is dubious at best.
Well, SET CONSTRAINTS potentially needs to check the state of any
outstanding constraints whose state changes from deferred to immediate. I
don't think we can say that it sets a constraint to immediate, but doesn't
check outstanding instances because they were from an outer transaction
(unless the constraint state reverted on commit which seems really odd).