"Jim C. Nasby" <jnasby@pervasive.com> writes:
> That could also break some partitioning schemes; I don't think it's a
> given that parents and children have matching constraints, and afaik a
> parent can have constraints that a child doesn't.
Not unless you drop the inherited constraint; the fact that you can is a
bug we are going to fix someday soon.
regards, tom lane