This package implements fast seamless encoding/decoding to/from bencode format for many native datatypes. To achieve more performance we use bytestring builders and hand optimized attoparsec parser so this library is considered as replacement for BEncode and AttoBencode packages.
Bencode is pretty similar to JSON: it has dictionaries(JSON objects), lists(JSON arrays), strings and integers. However bencode has a few advantages:
- Compactness: no spaces in between any values — nor lists nor dicts nor anywhere else.
- Dictionaries always sorted lexicographically by the keys. This allow us to test data on equality without decoding from raw bytestring. Moreover this allow to hash encoded data (this property is heavily used by core bittorrent protocol).
- All strings prefixed with its length. This simplifies and speed up string parsing.
Hovewer there are some disadvantages comparing with JSON:
- Bencode is certainly less human readable.
- Bencode is rarely used, except bittorrent protocol of course.
For documentation see package hackage page.
Feel free to report bugs and suggestions via issue tracker or the mail.