Bugfix: should use the full bit-range for randomised data in `TestFrame` Bugfix: prevent division by zero for approximate floatingpoint equality ...and use the new zip()-itertor to simplify the loops