[Bf-python] more NAN in AngleBetweenVec's

[Campbell Barton]

I was testing my tri2quads script and traced a bug back to 
AngleBetweenVecs returning NAN.

The issue is with 2 face normals being compared that are the same.

When I contrive the situation it geves me fine results.
print AngleBetweenVecs(Vector(-0.0, -0.0, 1.0), Vector(-0.0, -0.0, 1.0)) 
# < This works
    0.0

Maybe AngleBetweenVecs could do a simple vector comparison first.

If anyone wants I can package a blend with the problem

- Cam,