convert a string to a byte array

to_byte.js

function to_bytes(str) {
    var arr = []
      , byte_count
      , padding
      , binary
      , prefix
      , bytes
      , code
   
    for(var i = 0, len = str.length; i < len; ++i) {
      code = fixedCharCodeAt(str, i)
      if(code === false) {
          // non BMP chars will have a length of 2
          // if you try to get the charCode of the second element, it will be false
          continue
      } 
      // if it is a multibyte character
      if(code > 0x7F) {
        binary = code.toString(2)
        // each aditional byte gives 5 additional usable bits
        byte_count = Math.ceil((binary.length - 1) / 5)

        // create leading 1s that represent number of bytes used to represent this character 
        prefix = new Array(byte_count + 1).join('1')

        // create leading 0s to fill before the begining of the code point
        padding = new Array(((byte_count - 1) * 6) + (9 - prefix.length) - binary.length)
        binary = padding.join('0') + binary

        // fill the remaining bits in the leading byte
        prefix += binary.slice(0, 8 - byte_count)
        binary = binary.slice(8 - byte_count)

        // each byte has room for six bits from the code point
        bytes = [prefix].concat(binary.match(/.{6}/g))
        for(var j = 0, jlen = bytes.length; j <  jlen; ++j) {
          // ORing with 0x80 ensure that the non leading bytes are formated as 10xxxxxx
          arr[arr.length] = parseInt(bytes[j], 2) | 0x80
        }
      } else {
        arr[arr.length] = code
      }
    }
    return arr
  }

  // from https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/String/charCodeAt#Example_2.3A_Fixing_charCodeAt_to_handle_non-Basic-Multilingual-Plane_characters_if_their_presence_earlier_in_the_string_is_unknown
  // a version of charCodeAt that also handles non-Basic-Multilingual-Plane characters
  function fixedCharCodeAt(str, idx) {
    idx = idx || 0

    var code = str.charCodeAt(idx)
      , low
      , hi

    if(0xD800 <= code && code <= 0xDBFF) {
      low = str.charCodeAt(idx+1)
      hi = code

      if(isNaN(low)) {
        throw new Error('High surrogate not followed by low surrogate in fixedCharCodeAt()')
      }
      return ((hi - 0xD800) * 0x400) + (low - 0xDC00) + 0x10000
    }
    if(0xDC00 <= code && code <= 0xDFFF) {
      return false
    }
    return code
  }