I have updated the cronjob to read from SVN instead from CVS. Details are explained at http://trac.osgeo.org/grass/wiki/WebPagesManagement including the link to the new Drupal test site. Markus