Re: An Idea for OID conflicts

Gregory Stark <stark@enterprisedb.com> writes:
> Those types, functions and operators that aren't used by system tables could
> be created by a simple SQL script instead.

Only if you don't need to know their OIDs anywhere in the C code.  I'm
not certain offhand how many of the non-core objects are so referenced.
We have in any case got a problem with changing OIDs for any built-in
types, because there are a number of clients out there with hard-wired
knowledge of what the type OIDs returned in query results mean.

T'would be nice to simplify the operator class setup though ...
        regards, tom lane