HTTP 1.1 request message cannot get last chunk of

2019-09-02 12:44发布

问题:

Here is my scratch code:

using System;
using System.Linq;
using System.Net;
using System.Net.Sockets;
using System.Text;

namespace SocketsDemo
{
    class Program
    {
        static void Main(string[] args)
        {
            Uri uri;
            if (args.Any()) uri = new Uri(args[0]);
            else uri = new Uri("http://odetocode.com/Articles/473.aspx"); //http://www.odetocode.com
            var result = GetResource(uri);
            Console.WriteLine(result);
            Console.ReadLine();
        }

        private static string GetResource(Uri uri)
        {
            var host = uri.Host;
            var resource = uri.PathAndQuery;
            var hostEntry = Dns.GetHostEntry(host);

            var socket = CreateSocket(hostEntry);
            SendRequest(socket, host, resource);
            return GetResponse(socket);
        }

        private static Socket CreateSocket(IPHostEntry hostEntry)
        {
            const int HTTP_PORT = 80;
            var endPoint = new IPEndPoint(hostEntry.AddressList[0], HTTP_PORT);
            var socket = new Socket(endPoint.AddressFamily, SocketType.Stream, ProtocolType.Tcp);
            socket.Connect(endPoint);
            return socket.Connected ? socket : null;
        }

        private static void SendRequest(Socket socket, string host, string resource)
        {
            var requestMessage = String.Format(
                "GET {0} HTTP/1.1\r\n" +
                "HOST: {1}\r\n" +
                "\r\n", resource, host);

            var requestBytes = Encoding.ASCII.GetBytes(requestMessage);
            socket.Send(requestBytes);
        }

        private static string GetResponse(Socket socket)
        {
            int bytesCount = 0;
            var buffer = new byte[256];
            var result = new StringBuilder();

            do
            {
                bytesCount = socket.Receive(buffer);
                result.Append(Encoding.ASCII.GetString(buffer, 0, bytesCount));
            } while (bytesCount > 0);
            return result.ToString();
        }
    }
}

When i change HTTP part of request message in SendRequest method to HTTP/1.0 - everything working. But when I try to repeat this on HTTP/1.1, this block do { .. } while (bytesCount > 0) hangs out on 181 cycle. Looks like server or client cannot deal with last chunk of bytes. Can anybody explain what stays behind this and how i can repair this with smallest changes to existing code.

回答1:

That's the keep-alive "feature" of http. The server does not disconnect after sending its message, keeping you waiting for further data.

You'll have to parse the HTTP headers. If there's a Transfer-Encoding: chunked, you'll have to parse the chunks you're receiving until you receive a chunk of 0 bytes.

If it isn't sent as chunks, you'll have to parse a Content-Length header to see how many bytes total to read.

HTTP/1.0 did not support chunks and some clients did not support keeping the connection alive, so for HTTP/1.0 the default server behavior should be keep-alive off, while for HTTP/1.1 the default is on.

See http://en.wikipedia.org/wiki/Chunked_transfer_encoding