Add string deconstruction into contiguous UTF-8 (allocating if needed). Comments and docs to come soon, for now doing early testing.