Robert,
> However, since we're on that tangent, I'm not completely convinced
> that additional lists of search paths that get prepended or appended
> to the main search path are the right way to go. It seems like that's
> just chopping up the problem into smaller bits without really fixing
> anything. I wonder if the right solution might be to associate with
> each schema a list of other schemas to be searched if the object isn't
> found in that schema. This means that the contents of search_path
> would really become the roots of the trees of schemas to be searched.
See, that strikes me a completely unmanageable and likely to give rise
to application security holes. But you're a smart guy ... so, *why*
would that be a better idea than some superuser settings? What am I not
thinking of?
--
Josh Berkus
PostgreSQL Experts Inc.
www.pgexperts.com
