something is majorly broken with goto-char, goto-line

sorry if it's already been complained about, as i'm not currently on xemacs-beta. goto-line doesn't work any more. its implementation uses forward-line; if you do ((progn (forward-line 55) (point)) you get the correct results, but the cursor doesn't end up there. obviously something in redisplay is changing where the cursor is -- it appears that it's refusing to allow window-start to change when a scroll command hasn't been invoked. this occurs in both my major workspaces, which have no common code, so i assume it's something that was recently checked in. my guess is one of the two recent changes by Mike S, the window config change or the frame-size change, since they both affect code in the right place. Mike, can you verify this, and if you see it too, check around a bit to see if you could have accidentally messed something up? thanks, ben