Explode and join loses top surface in rhino 5 but not 8

[davidcockey]

I am defining “non-manifold” in this thread and the associated thread as the presence of Rhino “non-manifold” edges, not a more general geometric one which considers an object to be non-manifold if three or more surfaces have coincident edge independent of how many edges Rhino considers to be joined. It appears that @jeremy is using the latter.

The model I uploaded has exactly the same geometry as Steve’s model. It has not “broken” Steve’s model. The difference is corners which coincide in Rhino are not joined.

@jeremy5

I don’t know how Steve got from your exploded and joined model to the misbehaving surface. In my case,

One way to do so, probably not how Steve did it, is to use NonmanifoldMerge on each of the sets of coincident corners. The result should be identical to Steve’s model. The Explode and Join proceedure I used essentially reversed NonmanifoldMerge for this particular topology which is non-manifold solely due to coincident corners. Please note that in the previous thread where I described this proceed I explicitly said:

@davidcockey

Note: This proceedure works with the topology of this object. It may not work work with other non-manifold topologies.