I have recently revisited the bdecoder in libtorrent, and ended up implementing a new bdecoder that is two orders of magnitude faster than the original (naive) parser. This is the 3rd decoder in libtorrent’s history, and I would like to cover its evolution of parsing bencoded data.