On the use of evidence theory in belief base revision