> In any case, I think the series looks good. I let you decide if you > want to make fixed_point_scaling_factor a method or not, it doesn't > make a big difference in the end. Thanks Simon. I pushed all 6 patches to master (including patch #5, which I think brings better consistency). -- Joel