python进程池

近日学习python的多进程，遇到了一个坑，在此记录一下。
代码如下：

from multiprocessing import Manager, Pool, Lock

def worker(l):
    for i in range(5):
        l.append(i)
    print(l)

def callback(arg):
    print('......exec done ')

if __name__ == '__main__':
    pool = Pool()
    m = Manager()
    l = m.list()
    for _ in range(4):
        pool.apply_async(worker, args=(l,), callback=callback)
    pool.close()
    pool.join()
    print('Main Process End.')

其中一次输出：

[0, 1, 2, 3, 0, 4, 1]
......exec done: 
[0, 1, 2, 3, 0, 4, 1, 2, 3, 0, 4]
......exec done: 
[0, 1, 2, 3, 0, 4, 1, 2, 3, 0, 4, 1, 2, 3, 0, 4, 1]
......exec done: 
[0, 1, 2, 3, 0, 4, 1, 2, 3, 0, 4, 1, 2, 3, 0, 4, 1, 2, 3, 4]
......exec done: 
Main Process End.

emmm....好像哪里有些不对....
于是就想到了加锁，修改代码如下:

def worker(l, lock):
    with lock:
        for i in range(5):
            l.append(i)
        print(l)
....
if __name__ == '__main__':
    pool = Pool()
    m = Manager()
    l = m.list()
    lock = Lock()
    for _ in range(4):
        pool.apply_async(worker, args=(l,lock), callback=callback)
    pool.close()
    pool.join()
    print('Main Process End.')

结果发现程序直接就执行了最后一句print('Main Process End.')...
查了一下发现似乎要用Manage，再次修改代码将lock = Lock()替换为lock=m.Lock()

[0, 1, 2, 3, 4]
......exec done: 
[0, 1, 2, 3, 4, 0, 1, 2, 3, 4]
......exec done: 
[0, 1, 2, 3, 4, 0, 1, 2, 3, 4, 0, 1, 2, 3, 4]
......exec done: 
[0, 1, 2, 3, 4, 0, 1, 2, 3, 4, 0, 1, 2, 3, 4, 0, 1, 2, 3, 4]
......exec done: 
Main Process End.

这次执行结果就和我们想要的一样了。