InvalidClassException thrown in deep cloning

Question

We are trying to deep-clone an object by using serialization. We want to read the object bytes to create a String ( using new String(byte[]) ), store that String in a memory cache for later retrieving, get its bytes (using string.getBytes() ) and rebuild the original object with an ObjectInputStream.
We are using ByteArrayInput/Output Streams to read and write the bytes from and to byte arrays, but when we call objectInputStream.readObject() we get an IllegalClassException. We have found that the byte arrays have in fact a couple of different elements after and before the retrieval. Here is the code we are using:
public class ObjectCloner {
public ObjectCloner(){}
static public Object deepCopy(Object oldObj) throws Exception
ObjectOutputStream oos = null;
ObjectInputStream ois = null;
try {
ByteArrayOutputStream bos = new ByteArrayOutputStream();
oos = new ObjectOutputStream(bos);
// serialize and pass the object
oos.writeObject(oldObj);
byte[] b = bos.toByteArray();
oos.flush();
String s = new String(b); //turn byte array into String
byte[] b2 = s.getBytes(); //read bytes again
ByteArrayInputStream bin = new ByteArrayInputStream(b2);
ois = new ObjectInputStream(bin);
catch(Exception e){
System.out.println("Exception in ObjectCloner = " + e);
finally{
oos.close();
ois.close();
return ois.readObject(); //recreate the object
We make a call from a client class, like this:
Integer inte = (Integer)ObjectCloner.deepCopy(new Integer(7));
Any comments will be very helpful. Thanks.

Answer 1

The reason is this, you create a new String with the byte array and use getBytes() to get a new byte array using the default encoding. First I don't understand why you use these two statements, because you can apply the bytearray b directly to your ByteArrayOutputStream.
If you want to make a copy of b, then use System.arraycopy instead of creating a String and extract the bytes from that.
It would have worked if you used getBytes("8859_1"), that is the ISO 8859-1 encoding, which maps characters 0-255 to the same values. But UTF-8 will only map 0-127 to 0-127, and 128-255 will be mapped to at least two bytes.
One more comment, you shouldnt try to read from your ObjectInputStream after you closed it. Move that line to just before you exit the try block, and in your exception block you should return null or throw an exception. The finally block will still be executed, so dont worry about that (remove the return statement from finally).