Java Decode double encoded utf-8 char

Question

I am parsing a websocket message and due do a bug in a specific socket.io version (Unfortunately I don't have control over the server side), some of the payload is double encoded as utf-8:

The correct value would be Wrocławskiej (note the l letter which is LATIN SMALL LETTER L WITH STROKE) but I actually get back WrocÅawskiej .

I already tried to decode/encode it again with java

String str = new String(wrongEncoded.getBytes(StandardCharsets.UTF_8), StandardCharsets.UTF_8);

Unfortunately the string stays the same. Any idea on how to do a double decoding in java? I saw a python version where they convert it to raw_unicode first and then parse it again, but I don't know this works or if there is a similar solution for Java. I already read through a couple of posts on that topic, but none helped.

Edit: To clarify in Fiddler I receive the following byte sequence for the above mentionend word:

WrocÃÂawskiej

byte[] arrOutput = { 0x57, 0x72, 0x6F, 0x63, 0xC3, 0x85, 0xC2, 0x82, 0x61, 0x77, 0x73, 0x6B, 0x69, 0x65, 0x6A };

Answer 1

You text was encoding to UTF-8, those bytes were then interpreted as ISO-8859-1 and re-encoded to UTF-8.

Wrocławskiej is unicode: 0057 0072 006f 0063 0142 0061 0077 0073 006b 0069 0065 006a
Encoding to UTF-8 it is: 57 72 6f 63 c5 82 61 77 73 6b 69 65 6a

In ISO-8859-1 , c5 is Å and 82 is undefined .
As ISO-8859-1, those bytes are: WrocÅawskiej
Encoding to UTF-8 it is: 57 72 6f 63 c3 85 c2 82 61 77 73 6b 69 65 6a
Those are likely the bytes you are receiving.

So, to undo that, you need:

String s = new String(bytes, StandardCharsets.UTF_8);

// fix "double encoding"
s = new String(s.getBytes(StandardCharsets.ISO_8859_1), StandardCharsets.UTF_8);

Answer 2

I had the problem that sometimes I received double encoded strings and sometimes proper encoded strings. The following method fixDoubleUTF8Encoding will handle both properly:

public static void main(String[] args) {
  String input = "werewräüèö";
  String result = fixDoubleUTF8Encoding(input);
  System.out.println(result); // werewräüèö
  
  input = "üäöé";
  result = fixDoubleUTF8Encoding(input);
  System.out.println(result); // üäöé
}

private static String fixDoubleUTF8Encoding(String s) {
  // interpret the string as UTF_8
  byte[] bytes = s.getBytes(StandardCharsets.UTF_8);
  // now check if the bytes contain 0x83 0xC2, meaning double encoded garbage
  if(isDoubleEncoded(bytes)) {
    // if so, lets fix the string by assuming it is ASCII extended and recode it once
    s = new String(s.getBytes(StandardCharsets.ISO_8859_1), StandardCharsets.UTF_8);            
  }
  return s;
}

private static boolean isDoubleEncoded(byte[] bytes) {
  for (int i = 0; i < bytes.length; i++) {
    if(bytes[i] == -125 && i+1 < bytes.length && bytes[i+1] == -62) {
      return true;
    }
  }
  return false;
}

Answer 3

Well, double encoding may not be the only issue to deal with. Here is a solution that counts for more then one reason

String myString = "heartbroken ð";
                myString = new String(myString.getBytes(StandardCharsets.ISO_8859_1), StandardCharsets.UTF_8);
                String cleanedText = StringEscapeUtils.unescapeJava(myString);
                byte[] bytes = cleanedText.getBytes(StandardCharsets.UTF_8);
                String text = new String(bytes, StandardCharsets.UTF_8);
                Charset charset = Charset.forName("UTF-8");
                CharsetDecoder decoder = charset.newDecoder();
                decoder.onMalformedInput(CodingErrorAction.IGNORE);
                decoder.onUnmappableCharacter(CodingErrorAction.IGNORE);
                CharsetEncoder encoder = charset.newEncoder();
                encoder.onMalformedInput(CodingErrorAction.IGNORE);
                encoder.onUnmappableCharacter(CodingErrorAction.IGNORE);
                try {
                    // The new ByteBuffer is ready to be read.
                    ByteBuffer bbuf = encoder.encode(CharBuffer.wrap(text));
                    // The new ByteBuffer is ready to be read.
                    CharBuffer cbuf = decoder.decode(bbuf);
                    String str = cbuf.toString();
                } catch (CharacterCodingException e) {
                    logger.error("Error Message if you want to");

                } 

A