Visualisation: vacancies (vectors of length zero) create black isosurfaces

[GPMueller]

Can this behaviour somehow be avoided?

It is a hindrance in some cases to have the barrier between non-zero and zero vectors a black surface. Can we somehow make the isosurface function ignore vectors of zero length?

[GPMueller]

In some cases thiy may actually be desirable, e.g. when a part of the system is cut out, so that the black isosurface may visualize this:

[GPMueller]

With the addition of disorder (see #143) with 524413f, this actually represents a problem, as isosurfaces become unusable for a cube with many vacancies.