That no longer crashes and generally seems to work better as far as the cursor itself is concerned...however, now I'm observing a different problem...whenever I pass a spot that causes a cursor shape change, that area isn't properly repainted once the cursor moves away again, leaving a dirty image of the cursor shape at that spot (i.e. if the mouse passes over a splitter and the cursor correspondingly changes to a resize indicator and back, the spot where it shifted to a resize indicator has a ghost of the latter remaining until repaint). Regards, Rene
Okay, that is probably just a clipping issue - that should resolve itself when I chase down the performance-robbing code (or when radeon uses bitmap cursors, most likely).
Glad we're finally on the right track ;-) --The loon