Brendan Jurd escribió:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Sat, Apr 26, 2008 at 6:02 AM, Tom Lane wrote:
> > "Brendan Jurd" writes:
> > > Has anyone had a close look at how hard it would be allow just the
> > > "add to the end" capability?
> >
> > The problem is you can't guarantee anything about the ordering of the
> > new value relative to the old ones. The OID it's assigned might be
> > after them, or before them if the OID counter has wrapped around,
> > or (with much smaller probability) between two existing ones.
>
> I see. So to add a value on to the end with guaranteed ordering, you
> would have to rebuild the enum from scratch anyway.
The other alternative is to make the system generate the new OIDs in
such a way that the ordering is preserved. This, of course, has a lot
of problems of its own.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support