Kotlin Bencode Decoding and Encoding

As a fun weekend project I tried my hand at rolling an encoder/decoder for Bencode (the encoding format used in torrent files). Code for the project is available.

The specification for how Bencoding works is defined in The BitTorrent Protocol Specification v2 (which supersedes the v1 specification). The section defining bencoding states:

Decoding

  • Strings are length-prefixed base ten followed by a colon and the string. For example 4:spam corresponds to ‘spam’.
  • Integers are represented by an ‘i’ followed by the number in base 10 followed by an ‘e’. For example i3e corresponds to 3 and i-3e corresponds to -3. Integers have no size limitation. i-0e is invalid. All encodings with a leading zero, such as i03e, are invalid, other than i0e, which of course corresponds to 0.
  • Lists are encoded as an ‘l’ followed by their elements (also bencoded) followed by an ‘e’. For example l4:spam4:eggse corresponds to [‘spam’, ‘eggs’].
  • Dictionaries are encoded as a ‘d’ followed by a list of alternating keys and their corresponding values followed by an ‘e’. For example, d3:cow3:moo4:spam4:eggse corresponds to {‘cow’: ‘moo’, ‘spam’: ‘eggs’} and d4:spaml1:a1:bee corresponds to {‘spam’: [‘a’, ‘b’]}. Keys must be strings and appear in sorted order (sorted as raw strings, not alphanumerics).

This relatively straight forward specification lends itself to a simple decoding process:

  • read the next character to parse and see if it matches one of the markers in the specification:
  • The digits 0 through nine mark a bencoded string
  • i marks a bencoded integer
  • l marks a bencoded listed
  • d marks a bencoded dictionary
  • try to decode the next part of the sequence based on the rules for that particular marker, recursing for lists and dictionaries

It is a sequence of transformations over the Result type. Each of these steps could possibly fail:

  • read digits off the iterator. This could fail if the iterator does not have a next value.
  • convert them to an integer (representing the string’s length). This could result in a parse failure.
  • try to consume the ‘:’ separator.
  • try to read the correct number of characters from the iterator

The upside of this approach is it requires no backtracking and only needs to look ahead a single character at a time. I chose to use Guava’s PeekingIterator to assist this. It allows you to peek at the next value in an iterator without actually removing it (the same way one might peek at the top value in a stack). Additionally I implemented some helper extension methods over PeekingIterator for common operations I was using to read characters off the iterator. For example reading characters while a condition is true:

Internal Representation

Internally I chose to represent the becoded information as an ADT using a sealed class. Here is a simplified version of it (omitting methods):

Additionally I wanted to try and produce meaningful error information when decoding failed. I enjoyed using Results in Rust to do this before and wanted to see how it could work in Kotlin. I found a Result library and refactored the decoder to use it. On the whole I found it easier to propagate error messages but I’m not sure if it created idiomatic Kotlin code. For example the Bencoded String parsing looks like:

Encoding

In comparison to decoding the encoding process is even simpler:

  • for a string prepend the length on the string with a ‘:’ separating them
  • for a integer wrap the base 10 string representation of the number in ‘i’ and ‘e’
  • for a list recursively encode the list elements, concatenate them and wrap them in an ‘l’ and ‘e’
  • for a dictionary recursively encode the key value pairs, concatenate them and wrap them in an ‘d’ and ‘e’

Testing

As far as testing the decoding and encoding I found example torrent files from the Webtorrent Testing Fixtures project which are available for use under Creative Commons. This was a great way to test real world example of bencoded data. On large file in particular even revealed a performance problem around bencode string parsing.

Updated:

Comments