1. Introduction (Translation)

Universal socket server class

This module tries its best to define the server from various aspects:

For socket-based servers:
--address family：
- AF_INET{,6}: IP socket (default)
- AF_UNIX: Unix domain sockets
- others, such as AF_DECNET (see ) (not commonly used)
--socket type：
- SOCK_STREAM (reliable connection TCP)
- SOCK_DGRAM (UDP)

For request-based servers:
-- The client address needs to be authenticated before making further requests (this effectively blocks all processes that need to make requests until they are authenticated)
-- How to handle multiple requests:
- Synchronous (only one request can be processed at a time)
- forking (fork a new process to handle a request)
- threading (creating a new thread to handle a request)

Among the various classes in this module, the simplest server type is synchronous TCP/IP server. This is a poor class design, but it also preserves some of the design's type concepts.

The following is an inheritance diagram of five classes, four of which represent four types of synchronization servers:
                                                                                                                                                                                                                                                                                                +--------------+
|BaseServer |
                                                                                                                                                                                                                                                                                                +--------------+
                                            v
| TCPServer |------->| UnixStreamServer |
                                             v
| UDPServer |------->| UnixDatagramServer |
           +-------------+                                                                                                                                                                    Note: UnixDatagramServer inherits from UDPServer, not UnixStreamServer. The only difference between IP and Unix stream server is the address family. Most of the contents of the two server classes are simple repetitions.

forking and threading can be created for ForkingMixIn and TreadingMixIn mix-in classes. For example: threading UDP server class will be created as follows:
class ThreadingUDPServer(ThreadingMixIn, UDPServer): pass (See examples below for details)

Mix-in This class must be implemented first because it overrides the method that defines UDPServer. Setting different member variables also changes the basic server construction method.

In order to implement a service, you must redefine its handle method from the base class BaseRequestHandler. Then run the new service class by combining the service class with your overridden Handle method class.

The TCP and UDP methods of the request processing class are different. This can be hidden by using the request processing subclass StreamRequestHandler or DatagramRequestHandler.

Of course, you can also think of other methods.
For example, if the service contains the state of the memory requested to be modified, then there is no point in using the forking server (because modifications in the child process will not affect the initialization state of the parent process, and the parent process will not pass this modified parameter to other child process). In this case, you can use a threading server, and you are more likely to need to use a "lock" to avoid two requests arriving at the same time and causing a conflict in the server state.
In addition, if you are building an HTTP server, etc., all data will be stored externally (such as in the file system). When a request from the client is processed, and the client is slow to read data, the synchronous class will Make the service unresponsive, which may take a long time.
In some cases, request synchronization may require appropriate methods, but be affected by the request data in order to complete the request in the child process. This can be achieved by using a synchronization server and explicitly specifying the process for fork in the Handle method in the request handling class.
Another way to handle multiple simultaneous requests is to maintain an explicit form of completed requests, using the select() method to determine which request should be responded to next (or to determine whether to process the new request) , which is very important for streaming services when each client needs to establish a connection for a long time. (The premise is that threads and sub-processes are not used)

2. All class methods used

<span style="color: #0000ff;">import</span><span style="color: #000000;"> socket
</span><span style="color: #0000ff;">import</span><span style="color: #000000;"> select
</span><span style="color: #0000ff;">import</span><span style="color: #000000;"> sys
</span><span style="color: #0000ff;">import</span><span style="color: #000000;"> os
</span><span style="color: #0000ff;">import</span><span style="color: #000000;"> errno
</span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
    </span><span style="color: #0000ff;">import</span><span style="color: #000000;"> threading
</span><span style="color: #0000ff;">except</span><span style="color: #000000;"> ImportError:
    </span><span style="color: #0000ff;">import</span><span style="color: #000000;"> dummy_threading as threading

</span><span style="color: #800080;">__all__</span> = [<span style="color: #800000;">"</span><span style="color: #800000;">TCPServer</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">UDPServer</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">ForkingUDPServer</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">ForkingTCPServer</span><span style="color: #800000;">"</span><span style="color: #000000;">,
           </span><span style="color: #800000;">"</span><span style="color: #800000;">ThreadingUDPServer</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">ThreadingTCPServer</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">BaseRequestHandler</span><span style="color: #800000;">"</span><span style="color: #000000;">,
           </span><span style="color: #800000;">"</span><span style="color: #800000;">StreamRequestHandler</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">DatagramRequestHandler</span><span style="color: #800000;">"</span><span style="color: #000000;">,
           </span><span style="color: #800000;">"</span><span style="color: #800000;">ThreadingMixIn</span><span style="color: #800000;">"</span>, <span style="color: #800000;">"</span><span style="color: #800000;">ForkingMixIn</span><span style="color: #800000;">"</span><span style="color: #000000;">]
</span><span style="color: #0000ff;">if</span> hasattr(socket, <span style="color: #800000;">"</span><span style="color: #800000;">AF_UNIX</span><span style="color: #800000;">"</span><span style="color: #000000;">):
    </span><span style="color: #800080;">__all__</span>.extend([<span style="color: #800000;">"</span><span style="color: #800000;">UnixStreamServer</span><span style="color: #800000;">"</span>,<span style="color: #800000;">"</span><span style="color: #800000;">UnixDatagramServer</span><span style="color: #800000;">"</span><span style="color: #000000;">,
                    </span><span style="color: #800000;">"</span><span style="color: #800000;">ThreadingUnixStreamServer</span><span style="color: #800000;">"</span><span style="color: #000000;">,
                    </span><span style="color: #800000;">"</span><span style="color: #800000;">ThreadingUnixDatagramServer</span><span style="color: #800000;">"</span>])

Copy after login

3. BaseServer and BaseRequestHandler

Python abstracts network services into two main classes. One is the Server class, which is used to handle connection-related network operations, and the other is the RequestHandler class, which is used to handle data-related operations. And provides two MixIn classes for extending Server to implement multi-process or multi-threading. When building network services, Server and RequestHandler are not separated. The instance object of RequestHandler works in conjunction with Server.

3.1 BaseServer Analysis

BaseServer methods that can be called externally:

- __init__(server_address, RequestHandlerClass)
- serve_forever(poll_interval=0.5)
- shutdown()
- handle_request() # if you do not use serve_forever()
- fileno() -> int # for select()

That is, we can initialize through init and provide the serve_forever() and handle_request() methods to the outside world

3.1.1 init initialization

timeout =<span style="color: #000000;"> None

    </span><span style="color: #0000ff;">def</span> <span style="color: #800080;">__init__</span><span style="color: #000000;">(self, server_address, RequestHandlerClass):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Constructor.  May be extended, do not override.</span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.server_address </span>=<span style="color: #000000;"> server_address
        self.RequestHandlerClass </span>=<span style="color: #000000;"> RequestHandlerClass
        self.</span><span style="color: #800080;">__is_shut_down</span> =<span style="color: #000000;"> threading.Event()
        self.</span><span style="color: #800080;">__shutdown_request</span> = False

Copy after login

The main function of __init__ is to create the server object and initialize server_address and RequestHandlerClass.

server_address is a tuple containing the host address and port.

3.1.2 serve_forever

    <span style="color: #0000ff;">def</span> serve_forever(self, poll_interval=0.5<span style="color: #000000;">):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Handle one request at a time until shutdown.

        Polls for shutdown every poll_interval seconds. Ignores
        self.timeout. If you need to do periodic tasks, do them in
        another thread.
        </span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.</span><span style="color: #800080;">__is_shut_down</span><span style="color: #000000;">.clear()
        </span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
            </span><span style="color: #0000ff;">while</span> <span style="color: #0000ff;">not</span> self.<span style="color: #800080;">__shutdown_request</span><span style="color: #000000;">:
                
                r, w, e </span>=<span style="color: #000000;"> _eintr_retry(select.select, [self], [], [],
                                       poll_interval)
                </span><span style="color: #0000ff;">if</span> self <span style="color: #0000ff;">in</span><span style="color: #000000;"> r:
                    self._handle_request_noblock()
        </span><span style="color: #0000ff;">finally</span><span style="color: #000000;">:
            self.</span><span style="color: #800080;">__shutdown_request</span> =<span style="color: #000000;"> False
            self.</span><span style="color: #800080;">__is_shut_down</span>.set()

Copy after login

The select() function is used here, that is, server_forever accepts a poll_interval=0.5 parameter, which represents the time used for select polling, and then enters an infinite loop. In this loop, select every poll_interval seconds Poll once (block here) to monitor network IO. Once a new network connection request arrives, the _handle_request_noblock() method will be called to handle the new connection.

3.1.3 _handle_request_noblock()

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> _handle_request_noblock(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Handle one request, without blocking.

        I assume that select.select has returned that the socket is
        readable before this function was called, so there should be
        no risk of blocking in get_request().
        </span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">try</span><span style="color: #000000;">:
            request, client_address </span>=<span style="color: #000000;"> self.get_request()
        </span><span style="color: #0000ff;">except</span><span style="color: #000000;"> socket.error:
            </span><span style="color: #0000ff;">return</span>
        <span style="color: #0000ff;">if</span><span style="color: #000000;"> self.verify_request(request, client_address):
            </span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
                self.process_request(request, client_address)
            </span><span style="color: #0000ff;">except</span><span style="color: #000000;">:
                self.handle_error(request, client_address)
                self.shutdown_request(request)</span>

Copy after login

英文说明已经说的很明确，该方法处理的是一个非阻塞请求，首先通过get_request()方法获取连接，具体实现在其子类，一旦获取了连接，立即调用verify_request认证连接信息，通过认证，则调用process_request()方法处理请求，如果中途出现错误，则调用handle_error处理错误，同时，调用shutdown_request()方法结束这个连接。

那下面我们就先看上面提到的其他几个函数调用：

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> verify_request(self, request, client_address):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Verify the request.  May be overridden.

        Return True if we should proceed with this request.

        </span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">return</span><span style="color: #000000;"> True


    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> process_request(self, request, client_address):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Call finish_request.

        Overridden by ForkingMixIn and ThreadingMixIn.

        </span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.finish_request(request, client_address)
        self.shutdown_request(request)

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> handle_error(self, request, client_address):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Handle an error gracefully.  May be overridden.

        The default is to print a traceback and continue.

        </span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">print</span> <span style="color: #800000;">'</span><span style="color: #800000;">-</span><span style="color: #800000;">'</span>*40
        <span style="color: #0000ff;">print</span> <span style="color: #800000;">'</span><span style="color: #800000;">Exception happened during processing of request from</span><span style="color: #800000;">'</span><span style="color: #000000;">,
        </span><span style="color: #0000ff;">print</span><span style="color: #000000;"> client_address
        </span><span style="color: #0000ff;">import</span><span style="color: #000000;"> traceback
        traceback.print_exc() </span><span style="color: #008000;">#</span><span style="color: #008000;"> XXX But this goes to stderr!</span>
        <span style="color: #0000ff;">print</span> <span style="color: #800000;">'</span><span style="color: #800000;">-</span><span style="color: #800000;">'</span>*40


    <span style="color: #0000ff;">def</span><span style="color: #000000;"> shutdown_request(self, request):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Called to shutdown and close an individual request.</span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.close_request(request)</span><pre name=<span style="color: #800000;">"</span><span style="color: #800000;">code</span><span style="color: #800000;">"</span> <span style="color: #0000ff;">class</span>=<span style="color: #800000;">"</span><span style="color: #800000;">python</span><span style="color: #800000;">"</span>>    

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> finish_request(self, request, client_address):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Finish one request by instantiating RequestHandlerClass.</span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.RequestHandlerClass(request, client_address, self)</span>

Copy after login

verify_request()方法对request进行验证，通常会被子类重写。

process_request需要注意一下，它被ForkingMixIn 和 ThreadingMixIn重写，因此是mixin的入口，ForkingMixIn和ThreadingMixIn分别进行多进程和多线程的配置，并且调用finish_request()完成请求，调用shutdown_request()结束请求。

handle_error也可以被子类重写，打印错误的信息和客户端地址。

finish_request()处理完毕请求，在__init__中创建requestHandler对象，并通过requestHandler做出具体的处理。

3.1.4 handle_request()

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> handle_request(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Handle one request, possibly blocking.

        Respects self.timeout.
        </span><span style="color: #800000;">"""</span>
        <span style="color: #008000;">#</span><span style="color: #008000;"> Support people who used socket.settimeout() to escape</span>
        <span style="color: #008000;">#</span><span style="color: #008000;"> handle_request before self.timeout was available.</span>
        timeout =<span style="color: #000000;"> self.socket.gettimeout()
        </span><span style="color: #0000ff;">if</span> timeout <span style="color: #0000ff;">is</span><span style="color: #000000;"> None:
            timeout </span>=<span style="color: #000000;"> self.timeout
        </span><span style="color: #0000ff;">elif</span> self.timeout <span style="color: #0000ff;">is</span> <span style="color: #0000ff;">not</span><span style="color: #000000;"> None:
            timeout </span>=<span style="color: #000000;"> min(timeout, self.timeout)
        fd_sets </span>=<span style="color: #000000;"> _eintr_retry(select.select, [self], [], [], timeout)
        </span><span style="color: #0000ff;">if</span> <span style="color: #0000ff;">not</span><span style="color: #000000;"> fd_sets[0]:
            self.handle_timeout()
            </span><span style="color: #0000ff;">return</span><span style="color: #000000;">
        self._handle_request_noblock()</span>

Copy after login

上面已经提到，如果你没有用到server_forever()方法，说明你希望使用的是阻塞请求来处理连接，如英文描述所说，该方法只是处理一个阻塞的请求，仍然使用select()方法轮询来监听网络连接，但是需要考虑时间超时影响，一旦超时，调用handle_timeout()方法处理超时，一般在子类重写该方法；如果在超时之前监听到了网络的连接请求，则同server_forever一样，调用_handle_request_noblock()方法，完成对新的连接的请求处理。

3.2 BaseRequestHandler分析

<span style="color: #0000ff;">class</span><span style="color: #000000;"> BaseRequestHandler:

    </span><span style="color: #800000;">"""</span><span style="color: #800000;">Base class for request handler classes.

    This class is instantiated for each request to be handled.  The
    constructor sets the instance variables request, client_address
    and server, and then calls the handle() method.  To implement a
    specific service, all you need to do is to derive a class which
    defines a handle() method.

    The handle() method can find the request as self.request, the
    client address as self.client_address, and the server (in case it
    needs access to per-server information) as self.server.  Since a
    separate instance is created for each request, the handle() method
    can define arbitrary other instance variariables.

    </span><span style="color: #800000;">"""</span>

    <span style="color: #0000ff;">def</span> <span style="color: #800080;">__init__</span><span style="color: #000000;">(self, request, client_address, server):
        self.request </span>=<span style="color: #000000;"> request
        self.client_address </span>=<span style="color: #000000;"> client_address
        self.server </span>=<span style="color: #000000;"> server
        self.setup()
        </span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
            self.handle()
        </span><span style="color: #0000ff;">finally</span><span style="color: #000000;">:
            self.finish()

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> setup(self):
        </span><span style="color: #0000ff;">pass</span>

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> handle(self):
        </span><span style="color: #0000ff;">pass</span>

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> finish(self):
        </span><span style="color: #0000ff;">pass</span>

Copy after login

以上描述说明，所有requestHandler都继承BaseRequestHandler基类，该类会处理每一个请求。在__init__中初始化实例变量request、client_address、server，然后调用handle()方法完成请求处理。那么，我们唯一需要做的就是重写好Handle()方法，处理所有的请求。

总结：构建一个网络服务，需要一个BaseServer用于处理网络IO，同时在内部创建requestHandler对象，对所有具体的请求做处理。

四、各种子类

4.1 由BaseServer衍生的子类

4.1.1 TCPServer

<span style="color: #0000ff;">class</span><span style="color: #000000;"> TCPServer(BaseServer):
    address_family </span>=<span style="color: #000000;"> socket.AF_INET

    socket_type </span>=<span style="color: #000000;"> socket.SOCK_STREAM

    request_queue_size </span>= 5<span style="color: #000000;">

    allow_reuse_address </span>=<span style="color: #000000;"> False

    </span><span style="color: #0000ff;">def</span> <span style="color: #800080;">__init__</span>(self, server_address, RequestHandlerClass, bind_and_activate=<span style="color: #000000;">True):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Constructor.  May be extended, do not override.</span><span style="color: #800000;">"""</span><span style="color: #000000;">
        BaseServer.</span><span style="color: #800080;">__init__</span><span style="color: #000000;">(self, server_address, RequestHandlerClass)
        self.socket </span>=<span style="color: #000000;"> socket.socket(self.address_family,
                                    self.socket_type)
        </span><span style="color: #0000ff;">if</span><span style="color: #000000;"> bind_and_activate:
            </span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
                self.server_bind()
                self.server_activate()
            </span><span style="color: #0000ff;">except</span><span style="color: #000000;">:
                self.server_close()
                </span><span style="color: #0000ff;">raise</span>

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> server_bind(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Called by constructor to bind the socket.

        May be overridden.

        </span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">if</span><span style="color: #000000;"> self.allow_reuse_address:
            self.socket.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, </span>1<span style="color: #000000;">)
        self.socket.bind(self.server_address)
        self.server_address </span>=<span style="color: #000000;"> self.socket.getsockname()

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> server_activate(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Called by constructor to activate the server.

        May be overridden.

        </span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.socket.listen(self.request_queue_size)

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> server_close(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Called to clean-up the server.

        May be overridden.

        </span><span style="color: #800000;">"""</span><span style="color: #000000;">
        self.socket.close()

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> fileno(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Return socket file number.

        Interface required by select().

        </span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">return</span><span style="color: #000000;"> self.socket.fileno()

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> get_request(self):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Get the request and client address from the socket.

        May be overridden.

        </span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">return</span><span style="color: #000000;"> self.socket.accept()

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> shutdown_request(self, request):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Called to shutdown and close an individual request.</span><span style="color: #800000;">"""</span>
        <span style="color: #0000ff;">try</span><span style="color: #000000;">:
            </span><span style="color: #008000;">#</span><span style="color: #008000;">explicitly shutdown.  socket.close() merely releases</span>
            <span style="color: #008000;">#</span><span style="color: #008000;">the socket and waits for GC to perform the actual close.</span>
<span style="color: #000000;">            request.shutdown(socket.SHUT_WR)
        </span><span style="color: #0000ff;">except</span><span style="color: #000000;"> socket.error:
            </span><span style="color: #0000ff;">pass</span> <span style="color: #008000;">#</span><span style="color: #008000;">some platforms may raise ENOTCONN here</span>
<span style="color: #000000;">        self.close_request(request)

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> close_request(self, request):
        </span><span style="color: #800000;">"""</span><span style="color: #800000;">Called to clean up an individual request.</span><span style="color: #800000;">"""</span><span style="color: #000000;">
        request.close()</span>

Copy after login

在BaseServer基础上增加了一个TCP的socket连接，使用server_bind、server_activate、server_close处理TCP启停等操作，同时增加了get_request、shutdown_request、close_request处理客户端请求。

4.1.2 UDPServer

<span style="color: #0000ff;">class</span><span style="color: #000000;"> UDPServer(TCPServer):

    </span><span style="color: #800000;">"""</span><span style="color: #800000;">UDP server class.</span><span style="color: #800000;">"""</span><span style="color: #000000;">

    allow_reuse_address </span>=<span style="color: #000000;"> False

    socket_type </span>=<span style="color: #000000;"> socket.SOCK_DGRAM

    max_packet_size </span>= 8192

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> get_request(self):
        data, client_addr </span>=<span style="color: #000000;"> self.socket.recvfrom(self.max_packet_size)
        </span><span style="color: #0000ff;">return</span><span style="color: #000000;"> (data, self.socket), client_addr

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> server_activate(self):
        </span><span style="color: #008000;">#</span><span style="color: #008000;"> No need to call listen() for UDP.</span>
        <span style="color: #0000ff;">pass</span>

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> shutdown_request(self, request):
        </span><span style="color: #008000;">#</span><span style="color: #008000;"> No need to shutdown anything.</span>
<span style="color: #000000;">        self.close_request(request)

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> close_request(self, request):
        </span><span style="color: #008000;">#</span><span style="color: #008000;"> No need to close anything.</span>
        <span style="color: #0000ff;">pass</span>

Copy after login

继承自TCPServer，将socket改为了SOCK_DGRAM型，并修改了get_request，用于从SOCK_DGRAM中获取request。同时server_activate、shutdown_request、close_request都改成了空（UDP不需要），比TCP简单一些。

4.2 由BaseRequestHandler衍生的子类

4.2.1 StreamRequestHandler

<span style="color: #0000ff;">class</span><span style="color: #000000;"> StreamRequestHandler(BaseRequestHandler):
    rbufsize </span>= -1<span style="color: #000000;">
    wbufsize </span>=<span style="color: #000000;"> 0
    timeout </span>=<span style="color: #000000;"> None
    disable_nagle_algorithm </span>=<span style="color: #000000;"> False
    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> setup(self):
        self.connection </span>=<span style="color: #000000;"> self.request
        </span><span style="color: #0000ff;">if</span> self.timeout <span style="color: #0000ff;">is</span> <span style="color: #0000ff;">not</span><span style="color: #000000;"> None:
            self.connection.settimeout(self.timeout)
        </span><span style="color: #0000ff;">if</span><span style="color: #000000;"> self.disable_nagle_algorithm:
            self.connection.setsockopt(socket.IPPROTO_TCP,
                                       socket.TCP_NODELAY, True)
        self.rfile </span>= self.connection.makefile(<span style="color: #800000;">'</span><span style="color: #800000;">rb</span><span style="color: #800000;">'</span><span style="color: #000000;">, self.rbufsize)
        self.wfile </span>= self.connection.makefile(<span style="color: #800000;">'</span><span style="color: #800000;">wb</span><span style="color: #800000;">'</span><span style="color: #000000;">, self.wbufsize)

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> finish(self):
        </span><span style="color: #0000ff;">if</span> <span style="color: #0000ff;">not</span><span style="color: #000000;"> self.wfile.closed:
            </span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
                self.wfile.flush()
            </span><span style="color: #0000ff;">except</span><span style="color: #000000;"> socket.error:
                </span><span style="color: #008000;">#</span><span style="color: #008000;"> An final socket error may have occurred here, such as</span>
                <span style="color: #008000;">#</span><span style="color: #008000;"> the local error ECONNABORTED.</span>
                <span style="color: #0000ff;">pass</span><span style="color: #000000;">
        self.wfile.close()
        self.rfile.close()</span>

Copy after login

最主要的功能是根据socket生成了读写socket用的两个文件对象（可以理解为句柄）rfile和wfile

4.2.2 DatagramRequestHandler

<span style="color: #0000ff;">class</span><span style="color: #000000;"> DatagramRequestHandler(BaseRequestHandler):

    </span><span style="color: #008000;">#</span><span style="color: #008000;"> XXX Regrettably, I cannot get this working on Linux;</span>
    <span style="color: #008000;">#</span><span style="color: #008000;"> s.recvfrom() doesn't return a meaningful client address.</span>

    <span style="color: #800000;">"""</span><span style="color: #800000;">Define self.rfile and self.wfile for datagram sockets.</span><span style="color: #800000;">"""</span>

    <span style="color: #0000ff;">def</span><span style="color: #000000;"> setup(self):
        </span><span style="color: #0000ff;">try</span><span style="color: #000000;">:
            </span><span style="color: #0000ff;">from</span> cStringIO <span style="color: #0000ff;">import</span><span style="color: #000000;"> StringIO
        </span><span style="color: #0000ff;">except</span><span style="color: #000000;"> ImportError:
            </span><span style="color: #0000ff;">from</span> StringIO <span style="color: #0000ff;">import</span><span style="color: #000000;"> StringIO
        self.packet, self.socket </span>=<span style="color: #000000;"> self.request
        self.rfile </span>=<span style="color: #000000;"> StringIO(self.packet)
        self.wfile </span>=<span style="color: #000000;"> StringIO()

    </span><span style="color: #0000ff;">def</span><span style="color: #000000;"> finish(self):
        self.socket.sendto(self.wfile.getvalue(), self.client_address)</span>

Copy after login

同样是生成rfile和wfile，但UDP不直接关联socket。这里的rfile是直接由从UDP中读取的数据生成的，wfile则是新建了一个StringIO，用于写数据。

(题目起的有点大，部分剖析的不好，等之后再往祖坟上刨。。。。^-^)

参考博客：http://www.cnblogs.com/tuzkee/p/3573210.html

http://www.jianshu.com/p/357e436936bf

Python in-depth analysis of the SocketServer module (1) (V2.7.11)

1. Introduction (Translation)