Stephen Frost wrote:
> While I agree that raising a warning makes sense I don't believe it
> should be forced. There may be cases where, even in security definer
> functions, the current search_path should be used (though, of course,
> care must be taken in writing such functions).
I really wonder whether such a use case exists. What would it be?
--
Peter Eisentraut
http://developer.postgresql.org/~petere/