Before, there were 1+n calls of df where n is the number of output values. I introduced some script magic to use the values from the
first call. Motivation was, there was a very complex sed call which failed to process some of my df output lines. The new code is much
safer.
Actually, the original problem obviously was that the sed regex didn't cover capital letters which I had in my mountpoints.
Two fixes:
1) Use same regexp to skip df partitions in "config" and "fetch".
2) Use same function (was two different sed) to cleanup partition name.
This fixes bug #5.
