Character decoding

Question

For character decoding an incoming message , I am retrieving the text string from a bodypart and character-decoding it using the "charset" obtained from body part. Then I am creating a new body part with this decoded string . Then I am removing the old bodypart and adding this new bodypart at that place. I am creating a new ParameterList and setting "charset" as a parameter . Then I set this list to the Content Type of new bodypart. The sample code of the method is as follows:-
mimeBodyPart = (MimeBodyPart) (((BodyPart) multiPart.getBodyPart(i))); //Creating mime Bodypart object from multipart.
String strCharset = (new ContentType(mimeBodyPart.getContentType())).getParameter("charset"); //Retrieving the charset
InputStream is = mimeBodyPart.getInputStream();
is.read(b); //Reading the inputstream into a byte array.
decodedString = new String(b,strCharset); //Creating the decoded string.
MimeBodyPart newBodyPart = new MimeBodyPart(new InternetHeaders(is),decodedString.getBytes()); //Creating the new bodypart here.
ContentType newContentType = new ContentType(newBodyPart[i].getContentType()); //Getting Content Type object of new bodypart.
ParameterList parameterList=new ParameterList(); //Creating a new parameter list.
parameterList.set("charset",strCharset); //Setting 'charset' as a parameter in parameter list.
newContentType.setParameterList(parameterList);
multiPart.removeBodyPart(1); //Removing the old body part at position 1
multiPart.addBodyPart(newBodyPart[1], 1); //Adding the new body part.
But when I am retrieving the text in bodypart at later stages of my application ,I am getting corrupted data and if I am retrieving the 'charset' parameter later I am getting a different value.e.g. if initially the value is "iso-8859-1" of 'charset', i am getting the 'charset' parameter value as "us-ascii" later. Why is it so?

Answer 1

OK.. I am really confused by what you are doing here. Either I am looking at it completely wrong, or you are doing a whole lot of unneccessary stuff here.
When you get the content of a Part in JavaMail, it should be automatically decoding it unless the encoding type is not supported. So, what I normally do is:
Get the Part content (this is an Object) using getContent();
Get the Part content type
Get the Part content disposition
Test to see if the content is a Multipart using an instanceof test. If so, recursively call this function for each part
Otherwise, if the content type is text/plain or text/html then assume the content Object is a String and just call toString() or cast it.
Otherwise if the content disposition is "inline" or "attachment" assume the content is a binary stream and use the Part.getDataHandler().getInputStream() method
This will fail is the encoding of the Part is not one supported by Java. For example, this happens with UTF-7 encoding (often used for Delivery Status Notifications) so you would have to write your own decoder for this and any other charsets you expect which are outside the range supported by Java.