polyhedra self-intersection and T1 transitions

[glyg]

A question by mail (thanks Ali!) motivated me to address this in some details.

The question basically is how to avoid self intersections like this one:

Notice that the edge lengths are never 0. I think the proposed solution is the physically relevant one.

[glyg]

so we can adopt the same strategy as for type 1 transitions, when this angle is bellow a certain threshold, perform the topology changed as a solution.