Function lengthBytesUTF8

mkbitmap/mkbitmap.js:1262–1281

(str)

  
  
  function lengthBytesUTF8(str) {
      var len = 0;
      for (var i = 0; i < str.length; ++i) {
        // Gotcha: charCodeAt returns a 16-bit word that is a UTF-16 encoded code
        // unit, not a Unicode code point of the character! So decode
        // UTF16->UTF32->UTF8.
        // See http://unicode.org/faq/utf_bom.html#utf16-3
        var c = str.charCodeAt(i); // possibly a lead surrogate
        if (c <= 0x7F) {
          len++;
        } else if (c <= 0x7FF) {
          len += 2;
        } else if (c >= 0xD800 && c <= 0xDFFF) {
          len += 4; ++i;
        } else {
          len += 3;
        }
      }
      return len;
    }
  
  function stringToUTF8Array(str, heap, outIdx, maxBytesToWrite) {
      assert(typeof str === 'string');