Java中创建线程有两种方式:继承Thread重写run()与实现Runnable()接口通过Thread构造。
继承Thread重写run()
创建一个TestThread继承自Thread,然后调用start()运行线程
TestThread testThread = new TestThread();
testThread.start();
public class TestThread extends Thread{
@Override
public void run() {
super.run();
Log.i("thread","TestThread");
}
}
打印的log为:TestThread
I/thread: TestThread
实现Runnable接口
定义一个类TestRunnable实现Runnable的run方法,通过Thread的构造函数传入一个TestRunnable对象,调用start()开启线程
Thread runnable = new Thread(new TestRunnable());
runnable.start();
public class TestRunnable implements Runnable{
@Override
public void run() {
Log.i("thread","testRunnable");
}
}
打印的log为:testRunnable
I/thread: testRunnable
继承Thread类与实现Runnable接口的原理和区别
通过上面的两种方式都可以实现多线程,那么将TestRunnable对象传给TestThread会是怎样的呢?代码走起
TestThread testThread = new TestThread(new TestRunnable());
testThread.start();
public class TestThread extends Thread{
public TestThread(Runnable target) {
super(target);
}
@Override
public void run() {
super.run();
Log.i("thread","TestThread");
}
}
public class TestRunnable implements Runnable{
@Override
public void run() {
Log.i("thread","testRunnable");
}
}
运行代码,发现log分别打印了TestThread和testRunnable
I/thread: TestThread
I/thread: testRunnable
而调用Thread的start()函数会创建一个新的线程以及调用线程的run()方法,而在TestThread的run()方法中,调用了super.run(),说明super.run()中调用了TestRunnable的run()方法,那看看Thread中构造函数传入Runnable对象是如何实现的。
public class Thread implements Runnable {
public Thread(Runnable target) {
init(null, target, "Thread-" + nextThreadNum(), 0);
}
private void init(ThreadGroup g, Runnable target, String name,
long stackSize, AccessControlContext acc) {
/*省略其他的一些代码*/
this.target = target;
setPriority(priority);
tid = nextThreadID();
}
@Override
public void run() {
if (target != null) {
target.run();
}
}
}
可以发现Thread也是实现了Runnable接口,在构造函数时将传入的Runnable对象赋值给了target,而在实现的run()函数中调用了target.run(),即我们构造函数传入的Runnable对象。因此我们可以得到,通过继承Thread和实现Runnable本质是一样的,Thread是通过重写run()函数,而Runnable的方式间接通过Thread的target属性来调用run()函数,都依托于Thread,最终执行run()函数。
start()与run()的区别
通过上面的分析可以得知道Thread实现了Runnable接口,最终执行的也是run()函数,而我们在调用启动线程的时候是start()函数,那么他们有什么区别呢?首先看段代码
TestThread testThread = new TestThread();
testThread.start();
try {
Thread.sleep(200);
} catch (InterruptedException ignored) {
}
Log.i("thread","TestThread start");
public class TestThread extends Thread{
@Override
public void run() {
super.run();
for (int i = 0; i < 10 ; i ++) {
Log.i("thread","TestThread" + i);
try {
Thread.sleep(200);
} catch (InterruptedException ignored) {
}
}
}
}
在TestThread的run()打印10行log,然后在调用start()打印TestThread start,sleep 200毫秒方便看打印顺序,看看打印的结果如何
thread: TestThread0
thread: TestThread start
thread: TestThread1
thread: TestThread2
thread: TestThread3
thread: TestThread4
thread: TestThread5
thread: TestThread6
thread: TestThread7
thread: TestThread8
thread: TestThread9
因为TestThread是多线程执行的,因此TestThread start打印在TestThread0之后是正常的,既然start()间接调用了run(),那么我们直接调用run()会怎样呢?
TestThread testThread = new TestThread();
testThread.run();
try {
Thread.sleep(200);
} catch (InterruptedException ignored) {
}
Log.i("thread","TestThread start");
输出的log如下
thread: TestThread0
thread: TestThread1
thread: TestThread2
thread: TestThread3
thread: TestThread4
thread: TestThread5
thread: TestThread6
thread: TestThread7
thread: TestThread8
thread: TestThread9
thread: TestThread start
发现TestThread是最后打印的,在执行完TestThread的run()方法之后才执行,变成了同步执行。
看看Thread的start()函数是怎么实现的
/**
* Starts the new Thread of execution. The <code>run()</code> method of
* the receiver will be called by the receiver Thread itself (and not the
* Thread calling <code>start()</code>).
*
* @throws IllegalThreadStateException - if this thread has already started.
* @see Thread#run
*/
public synchronized void start() {
checkNotStarted();
hasBeenStarted = true;
nativeCreate(this, stackSize, daemon);
}
private native static void nativeCreate(Thread t, long stackSize, boolean daemon);
在start()函数中调用了nativeCreate来创建一个线程,而start()的注释中也写明,创建一个新的线程,run()函数的代码在新的线程里面执行,而start()函数的线程在调用start()的线程里面执行。
根据上面的分析可以得到,直接调用run()函数没有创建新的线程,而是在调用的线程中执行,因此变成了同步执行;而调用start()函数,start()函数会在调用的线程中执行,还是属于同步执行,但是start()函数会创建一个新的线程来执行run()函数,因此我们通过继承Thread还是实现Runnable重写run()函数都会在新的线程执行,从而变成异步执行。