python高并发异步服务器核心库forkcore使用方法

1 拷贝下面的代码到一个文件，并命名为forkcore.py

import os
import threading
import select
import socket

class ds_forkcore(object):

    #async IO(epoll)
    def ds_epoll(self):
        epoll=select.epoll()
        epoll.register(self.s.fileno(),select.EPOLLIN|select.EPOLLET)
        while 1:
            epoll_list=epoll.poll()
            for fd,_events in epoll_list:
                if fd==self.s.fileno():
                    conn,addr=self.s.accept()
                    print "Current process's pid is "+str(os.getpid())
                    self.worker(conn,addr)

    #multi_thread
    def ds_thread(self,thread_num=100):
        for _ in range(0,thread_num):
            t=threading.Thread(target=self.ds_epoll)
            t.setDaemon(1)
            t.start()
            t.join()

    #multi_process
    def ds_process(self,child_process_num=8):
        pid=os.getpid()
        print "Main process start, pid is "+str(pid)
        for _ in range(0,child_process_num):
            if pid==os.getpid():
                if os.fork():
                    pass
                else:
                    print "Worker process start, pid is "+str(os.getpid())
                    self.ds_thread()

    #init function
    def __init__(self,worker,port=3333):
        s=socket.socket(socket.AF_INET,socket.SOCK_STREAM)
        s.setsockopt(socket.SOL_SOCKET,socket.SO_REUSEADDR,1)
        s.bind(("",port))
        s.listen(50000)
        self.s=s
        self.worker=worker
        self.ds_process()

2 编写自己的代码

1> 导入forkcore库。

2> 定义worker函数，worker函数需要两个参数，conn代表客户端连接的socket，addr是(ip,port)的元组。

3> 直接使用forkcore.ds_forecore(worker,port=5555)即可，port用于指定监听端口。

import forkcore

if __name__=="__main__":
    def worker(conn,addr):
        print "Message from ("+str(addr[0])+":"+str(addr[1])+"): "+conn.recv(1024)[0:-1]

    forkcore.ds_forkcore(worker,port=5555)

注：需要linux 2.6以上的内核