Read / support non-conformal meshes by introducing polygons / polyhedras

[GiudGiud]

Just an idea. If when we have non-conformality we could just decide the element that is faced with a node in the middle of its face is a poly, then we can support it with the poly support.

Can you see a major flaw with that?
It would go a long way to be able to support non-conformal meshes, would help us verify flow solutions against starccm on the same meshes

[roystgnr]

So if we have a mesh that looks like a coarse quad with two fine quads on the same side, we can't support that as an adaptively refined mesh because we need four fine quads to create a coarse parent, but we could instead reinterpret it as a flat mesh in which the "coarse quad" was turned into a degenerate pentagon?

That's not crazy. IMHO it's a great idea for FVM and piecewise-constant DG. It's more questionable for C0 variables, though, where it really would be a reinterpretation.

[GiudGiud]

which the "coarse quad" was turned into a degenerate pentagon?

yup