[GRASSLIST:2997] r.param.scale curvatures fixed

[Markus Neteler]

Hi,

interesting for those using r.param.scale:

My colleagues Stefano Menegon and Lorenzo Potrich applied fixes for
curvature calculations (have been wrong earlier).
The curvatures should be o.k., we have tested with hemisphere and other
synthetic surfaces.

Cheers,

 Markus