I might be able to work on this in a couple weeks, the sort would only be for one mesh, not between meshes and it would happen on the CPU. It would also require all rotate, scale, transform operations be done on the CPU too (normally they are done on the GPU). This can put a heavy load on CPU. It is a difficult problem with how we can render using C3.
If we could get at the pre-draw shade we could do screen door.
I am open to other ideas on it, also since it is open source other PRs (other people submitting code/changes).