cut test option to choose between shell and .py script as obsolete

Now that slider.sh invokes slider.py, the need to make the choice of which to run configurable is moot -there's no need to test the .py script separately any more on unix.

Script choice can just be done in the test code based purely on OS.

Culling the code simplifes the test client (slightly) and allows a section to be cut from the functional testing docs