Currently, the canvas-based implementation can only layout plain text. We want to add support for rich text so that it can be laid out with the same fidelity as plain text.