Some remarks on integration with respect to weak martingales