Javascript: encoding conversion

conversion.js

// escape don't: * @ – _ + . /
// encodeURI don't: ; , / ? : @ & = + $ - _ . ! ~ * ' ( ) #
// encodeURIComponent don't:  - _ . ! ~ * ' ( )
/*
// 部分转义
// text -> unicode
unicode = escape(text);
// unicode -> text
text = unescape(unicode);
// text -> utf8
utf8 = encodeURIComponent(text);
// utf8 -> text
text = decodeURIComponent(utf8);
// unicode -> utf8
utf8 = encodeURIComponent(unescape(unicode));
// utf8 -> unicode
unicode = escape(decodeURIComponent(utf8));
*/
// 全部转义
// text -> unicode
var toUnicode = function(str) {
  var buf = [];
  for (var i=str.length-1;i>=0;i--) {
    buf.unshift( '%u' + str.charCodeAt(i).toString(16) );
  }
  return buf.join('');
};

// unicode -> text
var fromUnicode = function(unicode) {
  return unicode.replace(/%u([0-9a-fA-F]+)/g, function(match, hex) {
    return String.fromCharCode(parseInt(hex, 16));
  });
};
// unicode -> utf8
var unicode2utf8 = function(unicode) {
  return unicode.replace(/%u([0-9a-fA-F]+)/g, function(match, hex) {
    var utf8CharCodes = [];
    c = parseInt(hex, 16);
    if (c < 128) {
      utf8CharCodes.push(c);
    } else if (c < 2048) {
      utf8CharCodes.push((c >> 6) | 192, (c & 63) | 128);
    } else if (c < 65536) {
      utf8CharCodes.push((c >> 12) | 224, ((c >> 6) & 63) | 128, (c & 63) | 128);
    } else {
      utf8CharCodes.push((c >> 18) | 240, ((c >> 12) & 63) | 128, ((c >> 6) & 63) | 128, (c & 63) | 128);
    }
    for (var i=utf8CharCodes.length-1;i>=0;i--) {
      utf8CharCodes[i] = '%' + utf8CharCodes[i].toString(16);
    }
    return utf8CharCodes.join('');
  });
};
// utf8 -> unicode
var utf82unicode = function(utf8) {
  var parts = utf8.split('%').slice(1),
  i = 0,
  buf = [];
  while (i<parts.length) {
    var part = parseInt(parts[i], 16);
    if ((part & 128) === 0) {
      buf.push(part);
      i = i+1;
    } else if ((part & 240) === 240) {
      var part2 = parseInt(parts[i+1], 16);
      var part3 = parseInt(parts[i+2], 16);
      var part4 = parseInt(parts[i+3], 16);
      buf.push(
        ((part & 7) << 18) + ((part2 & 63) << 12) + ((part3 & 63) << 6) + (part4 & 63)
      );
      i = i+4;
    } else if ((part & 224) === 224) {
      var part2 = parseInt(parts[i+1], 16);
      var part3 = parseInt(parts[i+2], 16);
      buf.push(
        ((part & 15) << 12) + ((part2 & 63) << 6) + (part3 & 63)
      );
      i = i+3;
    } else if ((part & 192) === 192) {
      var part2 = parseInt(parts[i+1], 16);
      buf.push(
        ((part & 31) << 6) + (part2 & 63)
      );
      i = i+2;
    }
  }
  for (var i=buf.length-1;i>=0;i--) {
    buf[i] = '%u' + buf[i].toString(16);
  }
  return buf.join('');
};

// text -> utf8
var toUtf8 = function(str){
  return unicode2utf8(toUnicode(str));
};
// utf8 -> text
var fromUtf8 = function(utf8) {
  return fromUnicode(utf82unicode(utf8));
};
// text -> html entity
var toHtmlEntity = function(str) {
  var buf = [];
  for (var i=str.length-1;i>=0;i--) {
    buf.unshift('&#'+str.charCodeAt(i)+';');
  }
  return buf.join('');
};
// html entity -> text
var fromHtmlEntity = function(entities) {
  return entities.replace(/&#(\d+);/g, function(match, dec) {
    return String.fromCharCode(dec);
  });
};

(function(){
  var str = 'JavaScript高级程序设计';
  var unicode = toUnicode(str);
  var utf8 = toUtf8(str);
  var entities = toHtmlEntity(str);

  var utf82 = unicode2utf8(unicode);
  var unicode2 = utf82unicode(utf8);

  var str2 = fromUnicode(unicode);
  var str3 = fromUtf8(utf8);
  var str4 = fromHtmlEntity(entities);

  console.log(str);
  console.log(str2);
  console.log(str3);
  console.log(str4);

  console.log(utf8);
  console.log(utf82);

  console.log(unicode);
  console.log(unicode2);

  console.log(entities);
  // Output:
  // JavaScript高级程序设计
  // JavaScript高级程序设计
  // JavaScript高级程序设计
  // JavaScript高级程序设计
  // %4a%61%76%61%53%63%72%69%70%74%e9%ab%98%e7%ba%a7%e7%a8%8b%e5%ba%8f%e8%ae%be%e8%ae%a1
  // %4a%61%76%61%53%63%72%69%70%74%e9%ab%98%e7%ba%a7%e7%a8%8b%e5%ba%8f%e8%ae%be%e8%ae%a1
  // %u4a%u61%u76%u61%u53%u63%u72%u69%u70%u74%u9ad8%u7ea7%u7a0b%u5e8f%u8bbe%u8ba1
  // %u4a%u61%u76%u61%u53%u63%u72%u69%u70%u74%u9ad8%u7ea7%u7a0b%u5e8f%u8bbe%u8ba1
  // &#74;&#97;&#118;&#97;&#83;&#99;&#114;&#105;&#112;&#116;&#39640;&#32423;&#31243;&#24207;&#35774;&#35745;
}());