[Fluxus] commits

[gabor papp]

hi,

i replaced the c vector operations with faster and more flexible scheme 
versions (vadd, vsub, vmul, vdiv). they support arbitrary vector size 
now. renamed the old versions to vaddc, vsubc, vmulc, vdivc.

generated normals from the gradient for (voxels->poly) and 
(blobby->poly). these normals are probably not as nice as calling 
(recalc-normals 1), but much faster when converting a big voxel primitive.

best,
gabor