On Wed, Dec 8, 2010 at 4:19 AM, Dimitri Fontaine <dimitri@2ndquadrant.fr> wrote:
> Robert Haas <robertmhaas@gmail.com> writes:
>> I think this so-called two-step approach is pretty ugly.
>
> Well it does not need to be exposed to the user, thinking about it, as
> proposed in the other thread. Other than that, you're argument here is
> exactly the same as the ones saying that VACUUM or Hint Bints are
> bad. It's just that if you want correctness, you don't have anything
> better.
Exposing it to the user is what I think is ugly.
It's also worth noting that ALTER EXTENSION .. SET SCHEMA does NOT
guarantee a correct relocation, because someone might have done ALTER
FUNCTION .. SET search_path = @extschema@, and that's not going to get
properly fixed up. I'm coming to the conclusion more and more that
ALTER EXTENSION .. SET SCHEMA just can't work reliably.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company