On Tue, Dec 17, 2013 at 1:27 PM, Simon Riggs <simon@2ndquadrant.com> wrote:
> Not sure I'd say required, but its certainly desirable to have
> updateable security barrier views in themselves. And it comes across
> to me as a cleaner and potentially more performant way of doing the
> security checks for RLS.
Yes, that's how I'm thinking of it. It's required in the sense that
if we don't do it as a separate patch, we'll need to fold many of
changes into the RLS patch, which IMV is not desirable. We'd end up
with more complexity and less functionality with no real upside that I
can see.
But I think we are basically saying the same thing.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
