Python threading.current_thread() and callback

Question

我在下面有这个小的线程测试代码。线程名称有时是执行程序的名称，有时是主线程的名称。根据 3.7 Python 文档add_done_callback（注）...添加的可调用对象按添加顺序调用，并且始终在属于添加它们的进程的线程中调用。

为什么我随机获得不同的线程名称？

from concurrent.futures import ThreadPoolExecutor
import threading
def task(n):
    print(f'task - Thread: {threading.current_thread().name}')
    print("Processing {}".format(n))
    ttask()

def ttask():
    print(f'ttask - Thread: {threading.current_thread().name}')
    print("Processing ")

def taskDone(fn):
    print(f'taskDone - Thread: {threading.current_thread().name}')
    if fn.cancelled():
        print("Our {} Future has been cancelled".format(fn.arg))
    elif fn.done():
        print("Our Task has completed")

def secondTaskDone(fn):
    print(f'secondaTaskDone - Thread: {threading.current_thread().name}')
    print("I didn't think this would work")

def main():
    print("Starting ThreadPoolExecutor")
    print(f'Thread: {threading.current_thread().name}')
    with ThreadPoolExecutor(max_workers=3) as executor:
        print(f'Thread: {threading.current_thread().name}')
        future = executor.submit(task, (2))
        future.add_done_callback(taskDone)
        future.add_done_callback(secondTaskDone)
    print("All tasks complete")

if __name__ == '__main__':
    main()

输出是随机的：

输出 1：

Starting ThreadPoolExecutor
Thread: MainThread
Thread: MainThread
task - Thread: ThreadPoolExecutor-0_0
Processing 2
ttask - Thread: ThreadPoolExecutor-0_0
Processing
taskDone - Thread: ThreadPoolExecutor-0_0
Our Task has completed
secondaTaskDone - Thread: ThreadPoolExecutor-0_0
I didn't think this would work
All tasks complete

产出2：

Starting ThreadPoolExecutor
Thread: MainThread
Thread: MainThread
task - Thread: ThreadPoolExecutor-0_0
Processing 2
ttask - Thread: ThreadPoolExecutor-0_0
Processing
taskDone - Thread: MainThread
Our Task has completed
secondaTaskDone - Thread: MainThread
I didn't think this would work
All tasks complete

Answer 1

本主题中回答了问题。谢谢你的@dano。Python ThreadPoolExecutor - 回调是否保证与提交的函数在同一线程中运行？

为 ttask（） 添加了 0.2s 延迟，回调始终在 "ThreadPoolExecutor-0_0" 中运行。

此问题可以与上述主题重复关闭。