Christopher Kings-Lynne <chriskl@familyhealth.com.au> writes:
>> Hmm ... I hadn't thought of that approach, but it seems pretty
>> reasonable offhand ... comments?
> What if they change the owner of the serial sequence independently
> anyway?
I suppose a complete solution would involve forbidding that. We don't
allow you to alter the owner of an index independently of its parent
table ...
regards, tom lane