I would bet this was done for performance reasons and they don’t consider it a bug. It’s much slower to have to transfer data pixel by pixel checking the alpha channel than just a bulk memory copy.
↧