OK from what I can tell: there are several places where we do indeed call IntersectWithLine and FindClosestPoint in our vtkCellLocator child class, and routinely use the returned cellID, which in the original method signature is vtkIdType& cellId but we have it as vtkIdType cellId[1] in our child class. And we’ve been using this for years (well before I started on the project even), so I can only conclude that it works as intended. We don’t seem to use the subId and dist2 values, instead just providing properly sized arrays (both size 1) to have the wrapping system accept them (and checking the values of those in a few places, they look sensical, I think).
We don’t currently use the other classes I listed above, but I believe the original author updated those signatures as well just so we could use them if we wanted to in the future.
Given that - any further thoughts?