Not having a good way to get the reference element is a hindrance (although not blocking) supporting quadratic coordinate fields for FFD and MeshMovement2. What happens inside PDESolver is this:
face_verts = SummationByParts.SymCubatures.getfacevertexindices(sbp.cub)
edge_verts = [1 2 1 1 2 3; # TODO: SBP should provide this
2 3 3 4 4 4]
topo2 = ElementTopology2()
topo = ElementTopology{3}(face_verts, edge_verts, topo2=topo2)
But there should be some kind of API to get the edge_verts (or perhaps to supply the ElementTopology object itself, which is defined in ODLCommonTools?).
Not having a good way to get the reference element is a hindrance (although not blocking) supporting quadratic coordinate fields for FFD and MeshMovement2. What happens inside PDESolver is this:
But there should be some kind of API to get the
edge_verts(or perhaps to supply theElementTopologyobject itself, which is defined in ODLCommonTools?).