为什么Python的multiprocessing.Queue不会阻塞

Question

I am new to Python, but I want to write a multiproccesing program that uses multiprocessing.Queue as Go's channel. However, it seems that multiprocessing.Queue.put(, block=True) doesn't block as documented.

Here is my Python example program:

import multiprocessing
import select
import time


def f(q):
    while True:
        time.sleep(1)
        print("Put")
        q.put("hello world")


if __name__ == "__main__":
    q = multiprocessing.Queue(maxsize=0)
    f(q)

I expect that it will print only one "Put" and blocks forever. However, what I got is that it prints "Put" repeatedly.

Here is the Go program I think almost equivalent, except that goroutine is a green thread but not a process.

package main

import (
    "fmt"
    "time"
)

func main() {
    ch := make(chan int)
    for {
        time.Sleep(time.Second)
        fmt.Println("Put")
        ch <- 1
    }
}

When I type go run csp.go, it prints only one "Put" and the Go runtime prompts that the program is in deadlock.

Why multiprocessing.Queue.put() doesn't block?

Answer 1

Give a queue length of 1 and it will block as required

   Python 2.7.12 (default, Nov 12 2018, 14:36:49) 
    [GCC 5.4.0 20160609] on linux2
    Type "help", "copyright", "credits" or "license" for more information.
    >>> import multiprocessing
    >>> import select
    >>> import time
    >>> 
    >>> q = multiprocessing.Queue(maxsize=1)
    >>> q.put(9)
    >>> q.put(10)  // blocking here, have to ctrl-c to escape
    ^CTraceback (most recent call last):
      File "<stdin>", line 1, in <module>
      File "/usr/lib/python2.7/multiprocessing/queues.py", line 101, in put
        if not self._sem.acquire(block, timeout):
    KeyboardInterrupt
    >>>