I have trouble understanding locks in Python

Question

I'm playing with threading.Lock and wrote the following test to further my understanding.

import threading
from time import sleep

LOCK = threading.Lock()

class Printer(threading.Thread):
    def __init__(self, *args, **kwargs):
        super(Printer, self).__init__(*args, **kwargs)
        self.i = 0

    def run(self):            
        while True:
            print(self.i)
            self.i += 1
            sleep(1)

p = Printer()
p.start()

raw_input() # block main thread until return pressed
LOCK.acquire()

# stop printer for 6 seconds
for _ in range(3):
    print('holding...')
    sleep(2)

# let printer continue
LOCK.release()

I expected the following output (comment not part of output):

0
1
2
# user pressed enter
holding...
holding...
holding...
3
4
5
...

However, the printer does not stop - the actual output I am getting is:

0
1
2
# user pressed enter
holding...
3
4
holding...
5
6
holding...
7
8
...

Obviously I am misunderstanding something here? How can I temporarily suspend the printer?

The printer thread never attempts to acquire the lock, as far as I can see. — Dave Costa, Oct 14 '19 at 12:28
`threading` is kind of deprecated, since it's not true multiprocessing. The `multiprocessing` module may be easier to deal with (https://pymotw.com/2/multiprocessing/basics.html) or maybe `mpi4py` (https://pypi.org/project/mpi4py/)? — RightmireM, Oct 14 '19 at 12:33
@DaveCosta oh, I see. [This](https://stackoverflow.com/questions/33640283/thread-that-i-can-pause-and-resume) is probably what I need, then. — actual_panda, Oct 14 '19 at 12:34
`threading` is not deprecated. Threading and multiprocessing solve two different problems. A real argument can be made for using threads versus asynchronous programming. — chepner, Oct 14 '19 at 12:56

score 1 · Accepted Answer · answered Oct 14 '19 at 13:47

You must acquire the lock in your thread as well. A lock by itself does not do anything at all. Its only purpose is to make it so that you cannot acquire the lock twice - thus making two or more parts of the code mutually exclusive. Whoever gets the lock, can proceed while every other thread block while trying to acquire the lock.

Consider this version of the code (I am using with but you can as well replace it with acquire/release.

import threading
from time import sleep

LOCK = threading.Lock()

class Printer(threading.Thread):
    def __init__(self, *args, **kwargs):
        super(Printer, self).__init__(*args, **kwargs)
        self.i = 0

    def run(self):
        while True:
            with LOCK:
                print(self.i)
                self.i += 1
            sleep(1)

p = Printer()
p.start()

raw_input() # block main thread until return pressed
LOCK.acquire()

# stop printer for 6 seconds
for _ in range(3):
    print('holding...')
    sleep(2)

# let printer continue
LOCK.release()

This does now what you want it to do.

This still is not "right" as your main thread exits as soon as it releases the lock. Normally this would be considered bad programming but in a demo script it probably does not matter. You just do not have any control over your thread anymore - it will run there until killed with a signal.

I have trouble understanding locks in Python

1 Answers1