To be clear, it's rejected until I find a good way to implement it that doesn't involve hitting the API for no reason 99% of the time, or overly optimistic caching. Technically this has been on the TODO list since the day I started working on the reader.