Bruce Momjian writes:
> OK, I propose that we do the index part of this patch too. No one seems
> to know if seting the index user id to zero in pg_class will break
> things so we may as keep it consistent. Can you send that over?
What would it break? Joins against pg_user should use outer joins anyway.
I don't think this is the right direction. It's just a waste of code.
--
Peter Eisentraut peter_e(at)gmx(dot)net