When I change an id (primary key serial) in a table, the next value
returned by the sequence for the id can conflict with that id (e.g.,
change the id to be id + 1). MySQL seems to handle this transparently
by skipping conflicting values, but with PostgreSQL I get primary key
conflicts. It seems rather bad if a user can modify an id in a row and
cause failures for all future inserts - it's just too fragile. What's
the proper way to handle this in PostgreSQL?