Re: Inherited constraints and search paths (was Re: Preserving

Berend Tober <btober(at)seaworthysys(dot)com> writes:
> If the pg_dump output produced "SET search_path" statement with the
> complete actual path required to find all objects in subsequent DDL
> statements, my world would be at peace.

We're not doing that, because it's demonstrably impossible :-(.
You can't guarantee that every object in a random collection can be
referenced with an unqualified name. Consider
... CHECK (a.foo(f1) AND b.foo(f1))
ie identically named functions in different schemas.

Even in the cases where it is possible, the infrastructure for it isn't
available to pg_dump, which means there's no hope of a solution like
that being back-ported to 7.3.

regards, tom lane