2.2.1

What's Changed

• Updates by @NelsonVides in #12

stringprep new version loads processone/stringprep#14, super important! 😄
Then the optimised from_binary splits the binary parts in around 6x less reductions for inputs of any size (note, BEAM reductions). Honestly in wall-clock measurements, the difference is unnoticeable unless the input gets very big. At the maximum allowed size for JIDs, that is, 1023 chars per each of the three parts, we get ~10% faster at best. Not a big diff, but well, not a big PR either 🤷🏽

Full Changelog: 2.2.0...2.2.1