Remembering past discussions: Yes OID's wrap around. When this happens,
they will not nesecarily be unique anymore. Apparently this should not
be a problem unless one has queries that specifically use OID's. INT8
OID's have been talked about and will probably be implemented some day.
Jeff
>>>>>
Someone mentioned the OIDs are unique for every row for an entire
database cluster. I also notice that they are not being reused (like a
sequence). Will they wrap around as soon as the 2^32 boundary is hit?