Non-adjacent faces can affect DivideAlongCreases output

The underlying surface of the right face splits with _DivideAlongCreases _SplitAtTangents=Yes only when the face on the left is part of the polysurface. This happens in V5, V6, & V7.


DivideAlongCreases.3dm (95.2 KB)

Hello - it looks like it changes but does not split but rearranges the knot lines - the little arc surface has a pile of control points along the upper edge - it looks like the trim is almost but not right on, a knot line there. So there is a tiny span that extends out past the trim.


In the U direction,
Before the split: Domain:[0.000000,0.546267] Knot multiplicities:2,2,2
After the split: Domain:[0.005409,0.546267] Knot multiplicities:2,2
The narrow span is removed.

I just created a degree 3 surface with both a crease and tangent and trimmed it so that both conditions are outside of the face. _DivideAlongCreases does not modify either condition, so what happens in the provided model may be due to a conflict of whether Rhino sees the narrow span being inside or outside of the face.

I often see full multiplicity knots that overlap an edge this way in STEP imports from NX. In fact, that is where this data came from.

By the way, _ShrinkTrimmedSrf does not shrink the surface, and _ShrinkTrimmedSrfToEdge packs, not stacks, the control points along the side.

More important than the _DivideAlongCreases result for this edge condition is why it occurs in the 3-face brep and not the other two.