I just updated this code in another thread (explained there) to avoid using EdgeSrf:
Trial 2_2023Sep21a.gh (67.5 KB) (UPDATED below)
P.S. I found a component to get face centroids directly instead of using the average of face boundary discontinuity points. The result appears to be identical though, so I don’t know why a few faces have different colors (Z values) from the Sep17a version using EdgeSrf.
Trial 2_2023Sep22a.gh (71.1 KB) (UPDATED below)
Face Normals ‘C’ output (FaceN) in cyan group.
P.P.S. Not that it matters in this case but I added a fix internal to the CMesh cluster to correctly handle multiple meshes in a data tree. I’ll replace two earlier versions in this post with this one:
Trial 2_2023Sep24a.gh (70.4 KB)
A demo of multiple meshes in a data tree:
cmesh_demo_2023Sep24a.gh (27.7 KB)