This should not be a violation of the rule:
BEGIN v_tab(1) := v_another_tab(1); END;
Thanks Peter. This is a no-go indeed.
(QP-2887)