Re: BUG #16137: pg_upgrade fails with an index over nesting function

PG Bug reporting form <noreply@postgresql.org> writes:
> CREATE TABLE test(i INT PRIMARY KEY);
> CREATE FUNCTION func2(i INT) RETURNS INT IMMUTABLE LANGUAGE SQL
>     AS 'SELECT $1';
> CREATE FUNCTION func1(c INT) RETURNS INT IMMUTABLE LANGUAGE SQL
>     AS 'SELECT func2($1)';

This function is unreliable for lack of schema qualification.

I tend to agree that it'd be nice if pg_upgrade wouldn't be
so picky about that, but we haven't really found a way that
doesn't create security issues (cf CVE-2018-1058).

            regards, tom lane