yet another "trunc*" bug fixed - and the fundamental one! (now, when next trunc* word is less than curdoc it doesn't led to wrong results)