Question

server

 public void HandleConnection(Socket socket) {
    OutputStream out = socket.getOutputStream();
    InputStream in = socket.getInputStream();
              BufferedReader br = new BufferedReader(new InputStreamReader(in));
              String strLine;
              try {
                while ((strLine = br.readLine()) != null)   {
                  System.out.println (strLine);
                  }
            } catch (IOException e) {
                e.printStackTrace();
            }
        }

client

public void httpPostTest() {
        HttpClient httpclient = new DefaultHttpClient();
        HttpPost httppost = new HttpPost("http://192.168.0.2:8080/");

        try {
            httppost.setEntity(new StringEntity("test"));
            HttpResponse response = httpclient.execute(httppost);
        } catch (ClientProtocolException e) {
         e.printStackTrace();
        } catch (IOException e) {
         e.printStackTrace();
        }

    }

Any thoughts on this? This is all I get back, just the header:

POST / HTTP/1.1
Content-Length: 4
Content-Type: text/plain; charset=ISO-8859-1
Host: 192.168.0.2:8080
Connection: Keep-Alive
User-Agent: Apache-HttpClient/4.1.2 (java 1.5)




public String readLine(InputStream inputStream) {
        StringWriter writer = new StringWriter();
        try {
            IOUtils.copy(inputStream, writer, "ASCII");
        } catch (IOException e) {
            e.printStackTrace();
        }
        String theString = writer.toString();
        return theString.trim();
    }

Answer 1

From what I understand after doing some reading, BufferedReader is a bad way to read inputStreams from http post. What I was experiencing was that the client making the post would hang and so what the connection handler thread on every post, when I would kill the client the server would stop hanging and I would see the message body. From what I gathered on the internet, the best way to read the input stream is to step through it byte by byte, and exit the loop after the sum of the bytes == the content-length value.

That said, I’m planning on using Apache HTTPCore instead to handle this as that seems like a better solution