Now the implementation of the reverse proxy is mainly using Nginx, google version of the implementation of Python, most of them are implemented by Flask+requests. Too young. What should I do if I encounter a post request?

The following example shows how to implement reverse proxy with no brain tcp forwarding. The instance website is 1024, based on Python 3.

First, open a tcp server server and listen on port 1024:

import socket
s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
s.bind(('0.0.0.0', 1024))
s.listen(1500)

Then, accept a connection from a browser, no brain to receive the http request packet:

while 1:
  conn ,addr = s.accept()
  headers = ''
  while 1:
    buf = conn.recv(2048).decode('utf-8')
    headers += buf
    if len(buf) < 2048:
      break

Modify the headers, change the target host (127.0.0.1:1024) to 1024 domain name, close the long connection and compression, and modify the web page returned by the server:

headers = headers.replace('127.0.0.1:1024', 't66y.com')\
                 .replace('keep-alive', 'close')\
                 .replace('gzip','')

Open a new tcp, connect 1024, send http request:

s1 = socket.socket()
s1.connect(('t66y.com', 80))
s1.sendall(headers.encode())

Receive webpage:

while 1:
  try:
    buf = s1.recv(1024*8)
  except socket.timeout as e:
    print(e)
    break
  resp += buf
  if not buf or\
    buf.startswith(b'WebSocket') and buf.endswith(b'\r\n\r\n'):
    break

Modify the content of the webpage and replace the host as the local machine:

resp = resp.replace(b'Content-Encoding: gzip\r\n', b'')\
           .replace(b'Transfer-Encoding: chunked\r\n', b'')\
           .replace(b't66y.com', b'bjgong.tk:1024')

Finally, return the page to the browser:

conn.sendall(resp)
conn.close()

After the code runs, the browser openshttp://127.0.0.1:1024/index.php The one that is presented is the 1024 homepage, excited!
The code is as follows:

import socket

def main():
    s = socket.socket()
    s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
    s.bind(('0.0.0.0', 1024))
    s.listen(1500)
    while 1:
        conn, addr = s.accept()
        print(addr)
        headers = ''
        while 1:
            buf = conn.recv(2048).decode('utf-8')
            headers += buf
            if len(buf) < 2048:
                break

        headers = headers.replace('127.0.0.1:1024', 't66y.com')\
                         .replace('keep-alive', 'close')\
                         .replace('gzip','')
        print(headers)
        s1 = socket.socket()
        s1.connect(('t66y.com', 80))
        s1.sendall(headers.encode())

        resp = b''
        while 1:
            try:
                buf = s1.recv(1024*8)
            except socket.timeout as e:
                print(e)
                break
                
            resp += buf
            if not buf or\
               buf.startswith(b'WebSocket') and buf.endswith(b'\r\n\r\n'):
                break

        resp = resp.replace(b'Content-Encoding: gzip\r\n', b'')\
                   .replace(b't66y.com', b'bjgong.tk:1024')
        print('send to', addr)
        conn.sendall(resp)
        conn.close()


main()

This simple reverse proxy does not support https website, its performance is not good, it needs to be optimized. Please change the domain name of 1024 (t66y.com) to other websites when you demonstrate at home, otherwise it will be 2333.
If you want to hang on vps, change 127.0.0.1 to your domain name or ip (no domain name).

EOF