Problem with JAXB Unmarshaller

I have the following code in a servlet, and all works great from either a servlet parameter or the servlet input stream.....as long as there are no characters such as & or #. How can I get the unmarshaller to handle special characters? Can I add something to the xml to fix this?

String requestPost = request.getParameter("xmlString");
        Master masterBinding = null;
        try {
            JAXBContext jc = JAXBContext.newInstance( "myjaxb.bindings" );
            Unmarshaller u = jc.createUnmarshaller();
            if (requestPost != null){
              StringBuffer datapost = new StringBuffer(requestPost);
              logger.info("Datapost content: " + datapost + ":end");
              masterBinding = (Master)u.unmarshal( new StreamSource(
                  new StringReader( datapost.toString() ) ) );
            }
            else{
              masterBinding = (Master)u.unmarshal(siStream);
            }
        } 
        catch (JAXBException jbe) {
            jbe.printStackTrace(response.getWriter());
            logger.error("Exception encountered creating bindings.", jbe);
            acknowledgement.setCode("-1");
            acknowledgement.setReason("Exception Parsing datapost");
            acknowledgement.setStatus("fail");
            logger.warn("Datapost missing");
            sendResponseXML(acknowledgement, response.getWriter());
            return;
        }

TIA!

Hi Tia,

I'd say the better question is where do these special characters come from! Although you can turn on explicitly XML validation against a DTD or XML schema during the unmarshalling process, I think in this case the problem is most probably that the XML data aren't even well-formed and with correctly escaped special characters like <, > or &.

I must admit that I can't tell ad-hoc how the JAXB unmarshaller handles such kind of invalid XML especially without knowing more about the XML data your servlet gets. But I can tell you that I've successfully used JAXB for very big and even binary data and I was very happy with its ease of use. And perhaps a good idea for your problem would be to look for the source of the trouble, i.e. where these data come from! Maybe you can simply fix the input data and you're done with the rest of the problem!

It would be helpful, too, if you would provide some more details about the XML input data and what error you get from it!

Marco

Maybe this one will help you handling the said "special" characters:

http://commons.apache.org/lang/api-2.1/org/apache/commons/lang/StringEscapeUtils.html

I've used it to escape and unescape the few reserved characters during the marshalling and unmarshalling process of an application and it worked just fine. This utility class for example correctly converts a "&" character to "&" which you then can use in XML data!

Marco

Thanks Marco! What we're running into is someting like AT&T in a text node and the parser chokes on it. I will give the commons utility a try.