Re: Privileges and inheritance

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> So let's get rid of that. Selecting (or in general, operating) on a
> table with children only checks the privileges on that table, not the
> children.

+1

> We could use a GUC variable to ease the transition, perhaps like
> sql_inheritance = no | yes_without_privileges | yes

If we're gonna do it, let's just do it. I think adding a GUC just
complicates matters, especially since it would have to be superuser-only
(and thus effectively installation-wide). There would also be issues
with when it takes effect. The only simple way to implement this is
going to be to modify the planner's expansion of the range table, but
privilege checks should happen at execution; so the GUC would take
effect at the wrong time.

regards, tom lane