Http Server in Java

Question

Hi everyone, I'm having a little problem with this code:
import java.net.*;
import java.io.*;
class httpz {
public static void main(String args[]) {
     ServerSocket ss = null;
     try { ss = new ServerSocket(80); }
     catch(IOException ioe) { ioe.printStackTrace(); }
     System.out.println("httpz running on port 80");
     while (true) {
          try {
               Socket s = ss.accept();
               s.setSoTimeout(300000);
               System.out.println("Recieved Request\n");
               getRequest(s);
          } catch (Exception ioe) { ioe.printStackTrace(); }
public static void getRequest(Socket s) throws Exception {
     DataInputStream dis = new DataInputStream(s.getInputStream());
     byte[] b = new byte[dis.available()];
     dis.read(b);
     String str = new String(b);
     int extract1 = str.indexOf("GET");
     int extract2 = str.indexOf("HTTP");
     String filename = str.substring(extract1+5, extract2);
     System.out.println(str);
     System.out.println("Requested File: " + filename + " length = " + filename.length());
     if (filename.length() <= 1) filename = "index.html";
     try { setResponse(s, filename);} catch(Exception e) {}
public static void setResponse(Socket s, String filename) throws Exception {
     if (new File(filename).exists()) {
          FileInputStream fis = new FileInputStream(filename);
          byte b[] = new byte[fis.available()];
          fis.read(b);
          PrintWriter pw = new PrintWriter(s.getOutputStream(), true);
          pw.println(new String(b));
          pw.flush();
          pw.close();
     } else {
          PrintWriter pw = new PrintWriter(s.getOutputStream());     
          pw.println("Sorry, but the requested file seems to be unavailable");
          pw.flush();
          pw.close();
There doesn't seem to be anything wrong with this code. As you can see, this code accepts a request from the web browser. Then it passes the socket connection to the method getRequest, to extract the filename from the GET request.It also prints out the entire request in the console. Then it sends the filename to the setResponse method. It also checks to see if there is no file name, and if there is no file name, it sets the filename to index.html, the default page. In the setResponse method, it checks to see whether or not the file exists. If yes it reads from the file and then sends the page to the web browser. If not it prints out a message saying that the page cannot be found.
So this all works. When I run the server, and then type localhost in the browser, IE, it displays the proper page. I put a link from that page to a another page. When I get to the next page it is displayed properly. But the thing is, when I press 'back' it goes to the error message about not finding the page, and their is no request from the browser displayed in the console. And then If i press refresh, the proper page is displayed.
Anyone have an idea of what is going on here?
Thanx in advance.

Answer 1

After looking at your code...
while (true) {
try {
Socket s = ss.accept();
s.setSoTimeout(300000);
System.out.println("Recieved Request\n");
getRequest(s);
} catch (Exception ioe) {
ioe.printStackTrace();
... I am a little more sure that the problem is the multiple-requests-on-one-connection thing. Your code attempts to accept a new connection before processing a request. Perhaps you should do something like:
    Socket s;
    while (true) {
        System.out.println("Listening for incoming connections...");
        try {
            s = ss.accept();
        } catch (Exception x) {
            x.printStackTrace();
            s = null;
        if (s != null) {
            try {
                while (true) {
                    getRequest(s);
            } catch (IOException iox) {
                System.out.println("A connection error occured (client closed connection?)");
    };Now, keep in mind that your server will still have trouble handling multiple clients if one of the clients decides to stay connected after sending a request. The way servers like this are usually implemented is that, after a connection is accepted, a new Thread is spawned that handles all incoming requests from that client, and terminates when the connection closes. That way, each client has it's own thread on the server, so the server doesn't have to worry about keeping track of multiple connections and which socket belongs to which client, and which request came from who, and who to send the next response to, etc...
Jason