Python：有没有办法检查“是否请求停止”？在线程中的每一行之前？

Question

我希望能够同时中断多个线程。一种解决方案可能是：

stop_all_threads = False

def abort_if_requested():
    global stop_all_threads

    if stop_all_threads:
        raise Exception()


def thread_target():
    abort_if_requested()
    do_some_work()
    abort_if_requested()
    do_more_work()
    abort_if_requested()
    last_work_piece()


my_thread = threading.Thread(target=thread_target)
my_thread.start()

...

stop_all_threads = True

但显然，遵循起来有点烦人，并且如果工作功能需要很长时间，也不会立即停止它们。有没有更简单的方法？我知道 multiprocessing.Process.terminate 但我必须在我的用例中使用 threading.Thread，或者其他提供低延迟共享内存的方法。

Answer 1

我不太确定你在这里问什么，但这是我唯一能想到的：

stop_all_threads = False

def abort_if_requested():
    global stop_all_threads
    
    if stop_all_threads:
        raise Exception()

def thread_target():
    def check():
        while True:
            abort_if_requested()
            time.sleep(0.75)
    checkThread = threading.Thread(target=check)
    checkThread.start()
    do_some_work()

尝试重新编码，因为我不太明白你在这里需要什么。

Answer 2

您可以传递 Event来自 threading 库的类作为线程目标函数的参数，并在您想要停止所有线程上的处理时设置它。当事件被设置时，while 循环将停止，线程将正常终止。

我还建议您通过调用 < 等待线程终止Thread 对象上的 code>join 方法。如果您需要处理结果，这可以确保结果可用并且线程终止。如果您使用多个线程，请对所有线程调用join。

from threading import Thread, Event

def do_work_1(args):
    # Do your work here

def do_work_2(args):
    # Do your work here

def thread_target(stop_event, funcs, funcs_args):
    # Perform work until the event is set
    for idx, func in enumerate(funcs):
        func(funcs_args[idx])
        if stop_event.is_set():
           break

stop_event = Event()

sequential_work_funcs = [do_work_1, do_work_2, do_work_2]
# Add arguments you want to pass to functions 
# (args on the index get passed to func of same index)
sequential_work_funcs_args = [{}, {}, {}] 

# Pass the event as an argument to the target function
my_thread = Thread(target=thread_target, args=(stop_event, sequential_work_funcs, sequential_work_funcs_args))
my_thread.start()

# When you want to stop the threads
stop_event.set()

# Wait for thread to gracefully terminate if you need results of work
# join() is a blocking call that returns when the thread terminates
my_thread.join()