Python多线程详解

在Python中我们常用threading来处理多线程，我们先看一个例子：

# 代码段1
import threading
from time import sleep

def counter(n):
    print("The current thread [%s] is running." % threading.current_thread().getName())
    for i in range(n):
        print(i)
        sleep(1)
    print("The current thread [%s] ended." % threading.current_thread().getName())

print("The current thread [%s] is running." % threading.current_thread().getName())

t = threading.Thread(target=counter, args=(5,))
t.start()
t.join()

print("The current thread [%s] ended." % threading.current_thread().getName())

输出如下：

The current thread [MainThread] is running.
The current thread [Thread-1] is running.
0
1
2
3
4
The current thread [Thread-1] ended.
The current thread [MainThread] ended.

分析一下上面代码的一些关键点：

1. t = threading.Thread(target=counter, args=(5,))创建了一个线程，传入的target参数是线程要执行的对象，这里是一个函数，args是给函数传递的参数。
2. t.start()启动线程。
3. t.join()会阻塞调用线程的那个线程直到被调用线程结束，所以我们看到MainThread在Thread-1退出后才退出。

再看下面一段代码：

# 代码段2
import threading
from time import sleep

def counter(n):
    print("The current thread [%s] is running." % threading.current_thread().getName())
    for i in range(n):
        print(i)
        sleep(1)
    print("The current thread [%s] ended." % threading.current_thread().getName())

print("The current thread [%s] is running." % threading.current_thread().getName())

t = threading.Thread(target=counter, args=(5,))
t.start()

print("The current thread [%s] ended." % threading.current_thread().getName())

上面这段代码唯一的区别是去掉了t.join()，输出如下：

The current thread [MainThread] is running.
The current thread [Thread-1] is running.
0
The current thread [MainThread] ended.
1
2
3
4
The current thread [Thread-1] ended.

我们发现MainThread没有等Thread-1执行就结束了，这之后Thread-1继续执行直到结束。这印证了刚才提到的一个事实：t.join()会阻塞调用线程的那个线程直到被调用线程结束。

接下来看另外一段代码：

# 代码段3
import threading
from time import sleep

def daemon():
    print("The current thread [%s] is running." % threading.current_thread().getName())
    while True:
        print('in daemon')
        sleep(1)
    print("The current thread [%s] ended." % threading.current_thread().getName())

def counter(n):
    print("The current thread [%s] is running." % threading.current_thread().getName())
    for i in range(n):
        print(i)
        sleep(1)
    print("The current thread [%s] ended." % threading.current_thread().getName())

print("The current thread [%s] is running." % threading.current_thread().getName())

t1 = threading.Thread(target=daemon)
t2 = threading.Thread(target=counter, args=(5,))
t1.setDaemon(True)
t2.setDaemon(False)
t1.start()
t2.start()

print("The current thread [%s] ended." % threading.current_thread().getName())

输出如下：

The current thread [MainThread] is running.
The current thread [Thread-1] is running.
in daemon
The current thread [Thread-2] is running.
0
The current thread [MainThread] ended.
in daemon
1
in daemon
2
in daemon
3
in daemon
4
in daemon
The current thread [Thread-2] ended.

在代码段3中，线程t1被我们设置成守护线程，里面是一个无限循环，线程t2是用户线程，从0计数到4，每秒计数一次。可以观察到，当t2结束时，t1也跟着结束。在threading.py中，有这么一段注释：
The entire Python program exits when no alive non-daemon threads are left.
也就是说，当一个Python进程中没有任何非守护线程时，这个进程就会退出。在一个进程中，守护线程主要是用来在后台做一些幕后工作，是为用户线程服务的，用户线程都退出了，守护线程也失去了守护对象，所以也退出了。