Jnlp file shown as XML in Internet Explorer

Question

My problem is somewhat complex and have given me a headache while it seem to be a completely erratic behaviour. (Erratic, not over time, but over different computers.)
The main trouble is that I've got two separate webservers serving jnlp files for a production and a test environment. A user can access both servers and there no troubles finding and clicking at jnlp-files. There is however troubles while the jnlp-files in the test environment are shown as XML ("xmltext" in the Internet Explorer way) files.
But only when using Internet Explorer. The problem do not occur when using Netscape and never in the production environment (webserver).
I've triple checked the mimetype configuration of both the webservers and they are ok.
The result of a "get", from both webservers, is (except for the timestamps):
HTTP/1.1 200 OK
Server: Netscape-Enterprise/4.1
Date: Tue, 11 Jun 2002 14:05:32 GMT
Content-type: application/x-java-jnlp-file
Last-modified: Tue, 26 Mar 2002 15:58:27 GMT
Content-length: 1763
Accept-ranges: bytes
Connection: close
The client computer is running Win2000, Internet Explorer v5.50 and a newly installed JavaWebStart.
Well, there are more to it. This behaviour is for some users at the company. There are users that can, with no problem, run the application deployed at the webserver in the test environment.
A user who can not load jnlp-files in the test environment using his computer, can be able to load the program at another computer as he logs in.
Is there anyone who can help me out with this one?
How come the user(s) at a computer (and one login account) have different Internet Explorer behaviour when getting a jnlp file from two seperate webservers (using same webserver mime setup) when other users (computers) don't?

Answer 1

I've seen a lot of messages concerning the response of the webserver in posts of this JWS&JNLP forum. Some people gives the advise to use telnet to get the server response. I've used this code (based upon some free sun code in some tutorial trail I think it was).
import java.io.*;
import java.net.MalformedURLException;
import java.net.Socket;
import java.net.URL;
* Programmed using the beautiful Java programming tool
* <a href="http://www.intellij.com/idea/">Idea</a>.
* If you have not used it, try their free 30day trial version out
* (clean and nice (un)install, no nags, fully functional). I can not work without it.
* (This is not ment to be an advertisment but merely my personal recomendation.)
* <p>
* Usage:
* <code>
* % HttpResponse http://www.ibitec.se/index.html
* % HttpResponse http://www.intellij.com/idea
* % HttpResponse http://www.anyhost.com/anyjnlpfile.jnlp
* </code>
public class HttpResponse {
    private static int HTTP_PORT = 80;
    public static void main(String[] args) {
        for (int i = 0; i < args.length; i++) {
            System.out.println("Getting server response for url:\n\"" + args[i] + "\"\n");
            try {
                URL url = new URL(args);
if (url.getPort() != -1) HTTP_PORT = url.getPort();
if (!(url.getProtocol().equalsIgnoreCase("http"))) {
System.err.println("Sorry. I only understand Http not protocol " + url.getProtocol());
continue;
Socket socket = new Socket(url.getHost(), HTTP_PORT);
OutputStream theOutput = socket.getOutputStream();
PrintWriter pw = new PrintWriter(theOutput, false); // no auto-flushing
String EOL = "\r\n"; // native line endings are uncertain so add them manually
pw.print("GET " + url.getFile() + " HTTP/1.0" + EOL);
pw.print("Accept: application/x-java-jnlp-file, text/plain, text/html, text/*" + EOL);
pw.print(EOL);
pw.flush();
InputStreamReader isr = new InputStreamReader(socket.getInputStream());
BufferedReader br = new BufferedReader(isr);
BufferedWriter bw = new BufferedWriter(new OutputStreamWriter(System.out));
int character;
while ((character = br.read()) != -1) bw.write((char)character);
bw.flush();
} catch (MalformedURLException e) {
System.err.println("\"" + args[i] + "\" is not a valid URL.");
} catch (IOException e) {
System.err.println(e);
if (args.length == 0)
System.err.println("Usage: " + HttpResponse.class.getName() + " <url1> [url2 url3 ...]");