On Tue, 24 Feb 1998, Bruce Momjian wrote:
> The only problem with that is that the database administrator now should
> deal with pg_shadow, and not pg_user, and pg_user is not a real table
> anymore. Actually, in Unix, this is true too. I don't think we can
> change the real table to pg_shadow this close to a release, can we?
I like the pg_shadow idea.
One thing though, although the v6.3 driver is not backward compatible with
pre 6.3 backends (because of the network protocol changes), it should be
forward compatible. However, if we change the real table to pg_shadow
after 6.3, it would break the DatabaseMetaData class.
If we change the table's name now, we won't have this problem, and
existing code won't break either.
--
Peter T Mount petermount@earthling.net or pmount@maidast.demon.co.uk
Main Homepage: http://www.demon.co.uk/finder
Work Homepage: http://www.maidstone.gov.uk Work EMail: peter@maidstone.gov.uk