Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Sounds good to me, and I don't think we need to require RESTRICT just
> because the standard says so. Does the standard require RESTRICT for
> every DROP or just drops that have foreign keys?
Syntactically, it requires RESTRICT or CASCADE on *every* form of DROP.
I don't think we're willing to do that...
regards, tom lane