To help you get started, we've collected the most common ways that rfc4648 is being used within popular public projects.

EdgeApp / edge-core-js / src / core / login / edge.js View on Github

emit(ai.props.output.context.api, 'loginStart', {
    username: stashTree.username
  })

  // Find the appropriate child:
  const child = searchTree(stashTree, stash => stash.appId === appId)
  if (child == null) {
    throw new Error(`Cannot find requested appId: "${appId}"`)
  }

  // The Airbitz mobile will sometimes send the pin2Key in base58
  // instead of base64 due to an unfortunate bug. Fix that:
  if (child.pin2Key != null && child.pin2Key.slice(-1) !== '=') {
    io.console.warn('Fixing base58 pin2Key')
    child.pin2Key = base64.stringify(base58.parse(child.pin2Key))
  }
  await saveStash(ai, stashTree)

  // This is almost guaranteed to blow up spectacularly:
  const loginKey = base64.parse(reply.loginKey)
  const loginTree = makeLoginTree(stashTree, loginKey, appId)
  const login = searchTree(loginTree, login => login.appId === appId)
  if (login == null) {
    throw new Error(`Cannot find requested appId: "${appId}"`)
  }
  const newLoginTree = await syncLogin(ai, loginTree, login)
  const account = await makeAccount(ai, appId, newLoginTree, 'edgeLogin', opts)
  emit(ai.props.output.context.api, 'login', account)
}

View more ways to use rfc4648

rfc4648.js

This library implements encoding and decoding for the data formats specified in rfc4648:

• base64
• base64url
• base32
• base32hex
• base16

Each encoding has a simple API inspired by Javascript's built-in JSON object:

import { base32 } from "rfc4648";

base32.stringify([42, 121, 160]); // -> 'FJ42A==='
base32.parse("FJ42A==="); // -> Uint8Array([42, 121, 160])

The library has tree-shaking support, so tools like rollup.js or Webpack 2+ can automatically trim away any encodings you don't use.

• Zero external dependencies
• 100% test coverage
• Built-in types for Typescript & Flow
• 0.8K minified + gzip (can be even smaller with tree shaking)

API details

The library provides the following top-level modules:

• base64
• base64url
• base32
• base32hex
• base16
• codec

Each module exports a parse and stringify function.

const string = baseXX.stringify(data, opts)

Each stringify function takes array-like object of bytes and returns a string.

If you pass the option { pad: false } in the second parameter, the encoder will not output padding characters (=).

const data = baseXX.parse(string, opts)

Each parse function takes a string and returns a Uint8Array of bytes. If you would like a different return type, such as plain Array or a Node.js Buffer, pass its constructor in the second argument:

base64.parse("AOk=", { out: Array });
base64.parse("AOk=", { out: Buffer.allocUnsafe });

The constructor will be called with new, and should accept a single integer for the output length, in bytes.

If you pass the option { loose: true } in the second parameter, the parser will not validate padding characters (=):

base64.parse("AOk", { loose: true }); // No error

The base32 codec will also fix common typo characters in loose mode:

base32.parse("He1l0==", { loose: true }); // Auto-corrects as 'HELLO==='

Custom encodings

To define your own encodings, use the codec module:

const codec = require("rfc4648").codec;

const myEncoding = {
  chars: "01234567",
  bits: 3
};

codec.stringify([220, 10], myEncoding); // '670050=='
codec.parse("670050", myEncoding, { loose: true }); // [ 220, 10 ]

The encoding structure should have two members, a chars member giving the alphabet and a bits member giving the bits per character. The codec.parse function will extend this with a third member, codes, the first time it's called. The codes member is a lookup table mapping from characters back to numbers.