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
