vertProperties behavior

[pca006132]

1. We should definitely document them better.
2. Should we interpolate them according to barycentric coordinate after subdivide? It seems that we forgot to implement it.

   manifold/src/manifold/src/smoothing.cpp

   Lines 458 to 472 in 1514c15

   	// Make original since the subdivided faces are intended to be warped into
   	// being non-coplanar, and hence not being related to the original faces.
   	meshRelation_.originalID = ReserveIDs(1);
   	InitializeOriginal();
   	if (meshRelation_.numProp > 0) {
   	meshRelation_.properties.resize(meshRelation_.numProp * numVert);
   	meshRelation_.triProperties.resize(meshRelation_.triRef.size());
   	// Fill properties according to barycentric.
   	// Set triProp to share properties on continuous edges.
   	// Duplicate properties will be removed during sorting.
   	}
   	return vertBary;
   	}

3. Should the decimator update copy vertex properties after duplicating the vertices?

   manifold/src/manifold/src/edge_op.cpp

   Lines 324 to 342 in 1514c15

   	void Manifold::Impl::FormLoop(int current, int end) {
   	int startVert = vertPos_.size();
   	vertPos_.push_back(vertPos_[halfedge_[current].startVert]);
   	int endVert = vertPos_.size();
   	vertPos_.push_back(vertPos_[halfedge_[current].endVert]);
   	int oldMatch = halfedge_[current].pairedHalfedge;
   	int newMatch = halfedge_[end].pairedHalfedge;
   	UpdateVert(startVert, oldMatch, newMatch);
   	UpdateVert(endVert, end, current);
   	halfedge_[current].pairedHalfedge = newMatch;
   	halfedge_[newMatch].pairedHalfedge = current;
   	halfedge_[end].pairedHalfedge = oldMatch;
   	halfedge_[oldMatch].pairedHalfedge = end;
   	RemoveIfFolded(end);
   	}

[elalish]

Ah yes, indeed I haven't gotten around to 2 yet. There's a bunch of stuff I want to do to improve smoothing as well.

For 3, no it should not copy the properties because they're identical. Better to reference them and keep the memory down. I don't think we make any assumptions about how verts and property verts relate to each other.

Agreed, regarding the need for better docs.

[pca006132]

oh ok, for 3 I think I mixed up the MeshGL interface and our internal representation a bit, I thought we have properties for each vertex.

[elalish]

2 is fixed by #741