On Wed, Apr 13, 2016 at 6:53 PM, Stephen Frost <sfrost@snowman.net> wrote:
> Requiring that SET ROLE be allowed will mean that many more paths must
> be checked and adjusted, such as in all of the CreateObject statements
> and potentionally many other paths that I'm not thinking of here, not
> the least of which are all of the *existing* checks near
> get_rolespec_oid/tuple which will require additional checks for the
> CURRENT_USER case to see if the current user is a default role.
I don't get it. I think that these new roles should work just like
any other roles, except for existing at initdb time. I don't see why
they would require checking or adjusting any code-paths at all. It
would presumably have made the patch you committed smaller and
simpler. The only thing you'd need to do is (approximately) not dump
them.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
