That's beside the point. I had this problem on an earlier version (reported somewhere in this thread) and epage corrected it very soon afterward. Maybe this is a regression issue?