Android消息机制(Handler原理)-完全解析

一 概述

Handler主要被用来在子线程中访问UI线程，在ViewRootImpl中有一个checkThread()方法，对UI的操作都会有此验证。所以操作UI只能在主线程中进行。
概念：
Handler的运行由底层MessageQueue和Looper支撑。
MessageQueue消息队列，单链表存储消息列表。它是消息的存储单元。
Looper循环，它会无限循环查找是否有消息，有就处理掉。Looper创建一般被保存在TheadLoacal中。
创建：
Handler创建采用当前线程的Looper来构造消息循环系统。线程默认没有Looper
主线程(UI线程)即：ActivityThread,主线程创建时会初始化Looper,所以主线程可直接使用Handler

二 运行机制

2.1 TheadLoacal

TheadLoacal的用来保存数据，以线程作用域，不同线程具有不同的数据副本。
以一个例子来简单粗暴的说明：

//定义一个Integer泛型的ThreadLocal
private ThreadLocal<Integer> mIntTheadLocal = new ThreadLocal<Integer>();

mIntTheadLocal.set(0);
Log.d("yink","UI Thead, mIntTheadLocal = " + mIntTheadLocal.get());//UI线程

new Thread("thead 1") {
    @Override
    public void run() {
        mIntTheadLocal.set(1);
        Log.d("yink","thead 1, mIntTheadLocal = " + mIntTheadLocal.get());//线程1
    }
}.start();

new Thread("thead 2") {
    @Override
    public void run() {
        Log.d("yink","thead 2, mIntTheadLocal = " + mIntTheadLocal.get());//线程2
    }
}.start();

mIntTheadLocal是同一个对象，我们分别在UI线程、线程1和线程2中对其进行set,get并打印Log如下：

07-06 23:24:33.151 12661-12661/com.example.android.myapplication D/yink: UI Thead, mIntTheadLocal = 0
07-06 23:24:33.153 12661-16454/com.example.android.myapplication D/yink: thead 1, mIntTheadLocal = 1
07-06 23:24:33.154 12661-16455/com.example.android.myapplication D/yink: thead 2, mIntTheadLocal = null

由于TheadLoacal的特性，它其中之一的功能就是被用来保存Looper对象

2.2 MessageQueue

MessageQueue,单链表数据结构，主要包含两个操作：插入(enqueueMessage)和读取(next)
先看enqueueMessage，它之做了单链表插入数据操作

boolean enqueueMessage(Message msg, long when) {
    ...
    if (p == null || when == 0 || when < p.when) {
        // New head, wake up the event queue if blocked.
        msg.next = p;
        mMessages = msg;
        needWake = mBlocked;
    } else {
        // Inserted within the middle of the queue.  Usually we don't have to wake
        // up the event queue unless there is a barrier at the head of the queue
        // and the message is the earliest asynchronous message in the queue.
        needWake = mBlocked && p.target == null && msg.isAsynchronous();
        Message prev;
    //单链表插入数据操作。
        for (;;) {
            prev = p;
            p = p.next;
            if (p == null || when < p.when) {
                break;
            }
            if (needWake && p.isAsynchronous()) {
                needWake = false;
            }
        }
        msg.next = p; // invariant: p == prev.next
        prev.next = msg;
    }
    ...
}

再看next，查询到msg后，移除这个msg,并返回这个msg

Message next() {
        ...
    if (msg != null && msg.target == null) {
        // Stalled by a barrier.  Find the next asynchronous message in the queue.
        //被阻塞，在队列中查询下一个消息
        do {
        prevMsg = msg;
        msg = msg.next;
        } while (msg != null && !msg.isAsynchronous());
    }
    
    if (msg != null) {
        if (now < msg.when) {
        // Next message is not ready.  Set a timeout to wake up when it is ready.
        nextPollTimeoutMillis = (int) Math.min(msg.when - now, Integer.MAX_VALUE);
        } else {
        // Got a message.
        mBlocked = false;
        //msg不为空，移除msg
        if (prevMsg != null) {
            prevMsg.next = msg.next;
        } else {
            mMessages = msg.next;
        }
        msg.next = null;
        if (DEBUG) Log.v(TAG, "Returning message: " + msg);
        msg.markInUse();
        //返回msg
        return msg;
        }
    } else {
        // No more messages.
        nextPollTimeoutMillis = -1;
    }
        ...
}

2.3 Looper

Looper作用：不停的从MessageQueue中查询是否有消息，有就处理，没有就一直阻塞。
我们先来一个demo,为子线程创建Looper，来理解Looper的作用：

//主线程中定义
private Handler mHandler;

new Thread("Thread 1") {
    @Override
    public void run() {
        //创建Looper
        Looper.prepare();
        //在Thread 1中创建Handler
        mHandler =  new Handler(){
            @Override
            public void handleMessage(Message msg) {
                Bundle b = new Bundle();
                b = msg.getData();
                Log.d("yink","receive the number = " + b.getInt(KEY));
                switch (b.getInt(KEY)) {
                    case 3:
                        //子线程处理完事情后，退出循环
                        //Looper还提供一个quitSafely()，这个方法是设置标记位，把队列中消息处理完后再退出。
                        getLooper().quit();
                        break;
                    default:
                        break;
                }
                super.handleMessage(msg);
            }
        };
        //循环,关键函数
        Looper.loop();
    }
}.start();

new Thread("Thread 2") {
    @Override
    public void run() {
        for (int i = 0 ; i < 5 ; i++) {
            Message msg = new Message();
            Bundle b = new Bundle();
            b.putInt(KEY,i);
            msg.setData(b);
            //如果线程1的run没执行的话，mHandler还是null,所以加个判断
            if (mHandler != null) {
                Log.d("yink","send key i = " + i);
                //发送消息
                mHandler.sendMessage(msg);
            } else {
                Log.d("yink","mHandler is null, i = " + i);
            }
            SystemClock.sleep(1000);
        }
    }
}.start();

结果如下：

07-07 02:29:24.219  1151 12404 D yink    : send key i = 0
07-07 02:29:24.220  1151 12403 D yink    : receive the number = 0
07-07 02:29:25.221  1151 12404 D yink    : send key i = 1
07-07 02:29:25.223  1151 12403 D yink    : receive the number = 1
07-07 02:29:26.223  1151 12404 D yink    : send key i = 2
07-07 02:29:26.223  1151 12403 D yink    : receive the number = 2
07-07 02:29:27.225  1151 12404 D yink    : send key i = 3
07-07 02:29:27.226  1151 12403 D yink    : receive the number = 3
07-07 02:29:28.227  1151 12404 D yink    : send key i = 4
07-07 02:29:29.231  1151 12404 D yink    : send key i = 5

ActivityThread中的Looper比较特殊，它的创建是：prepareMainLooper(),获得：getMainLooper,可在任何地方获取到主线程的Looper

接下来我们来看Looper中最重要的loop()方法：

public static void loop() {
        ...
    
    for (;;) {
        //无限循环获取queue.next(),由前面描述的MesageQueue的next方法，只有获取到消息才会返回，否则一直会阻塞，所以looper也会在这里阻塞
        Message msg = queue.next(); // might block
        if (msg == null) {
        // No message indicates that the message queue is quitting.
        //只有当Looper调用退出后，queue.next()才会返回null,因为不退出，MesageQueue会阻塞在那儿一直查询
        //此无限循环中，只有当msg == null才退出循环
        return;
        }
    ...
    try {
                //target实际就是Handler,所以最终调用到创建Handler线程中去。
                msg.target.dispatchMessage(msg);
            } 
    ....

2.4 Handler

Handler主要也是两个动作：发送和接收
先看发送：

public final boolean sendMessage(Message msg){
    return sendMessageDelayed(msg, 0);
}

public final boolean sendEmptyMessageDelayed(int what, long delayMillis) {
    Message msg = Message.obtain();
    msg.what = what;
    return sendMessageDelayed(msg, delayMillis);
}

public final boolean sendEmptyMessageAtTime(int what, long uptimeMillis) {
    Message msg = Message.obtain();
    msg.what = what;
    return sendMessageAtTime(msg, uptimeMillis);
}

private boolean enqueueMessage(MessageQueue queue, Message msg, long uptimeMillis) {
    msg.target = this;
    if (mAsynchronous) {
        msg.setAsynchronous(true);
    }
    return queue.enqueueMessage(msg, uptimeMillis);
}

发送过程最后调用到了enqueueMessage，最后是向队列中插入了一条消息。
最后通过Looper返回给了Handler的dispatchMessage，转向接收处理。

public void dispatchMessage(Message msg) {
        if (msg.callback != null) {
            handleCallback(msg);
        } else {
            if (mCallback != null) {
                if (mCallback.handleMessage(msg)) {
                    return;
                }
            }
            handleMessage(msg);
        }
    }

第一种接收if (msg.callback != null) ：
如果msg.callback != null成立，则调用handleCallback(msg)来处理

private static void handleCallback(Message message) {
        message.callback.run();
    }

这个callback其实就是Handler.post中的runable
post过程如下：

public final boolean post(Runnable r)
    {
       return  sendMessageDelayed(getPostMessage(r), 0);//调用getPostMessage
    }

private static Message getPostMessage(Runnable r) {
        Message m = Message.obtain();
        m.callback = r;//把post中的Runnable赋值给msg.callback
        return m;
    }

证实:

        new Thread("Thead 1") {
            @Override
            public void run() {
                Looper.prepare();
                mHandler =  new Handler(){
            @Override
                    public void handleMessage(Message msg) {
            Log.d("yink","handleMessage ");
                        super.handleMessage(msg);
            }

                    @Override
                    public void dispatchMessage(Message msg) {
                        Log.d("yink","dispatchMessage ");
                        super.dispatchMessage(msg);
                    }
                };
                Looper.loop();
            }
        }.start();

        new Thread("Thead 2") {
            @Override
            public void run() {
                mHandler.post(new Runnable() {
                    @Override
                    public void run() {
                        Log.d("yink","post runable");
                    }
                });
            }
        }.start();

log如下，先是调用那个dispatchMessage然后走到调用了message.callback.run();

07-07 03:42:43.862 10475 10515 D yink    : dispatchMessage 
07-07 03:42:43.862 10475 10515 D yink    : post runable

第二种接收else：
如果 if (mCallback != null) 则调用mCallback.handleMessage(msg)
这里的mCallBack为：

public interface Callback {
        public boolean handleMessage(Message msg);
    }

//Handler的一个构造函数
public Handler(Callback callback) {
        this(callback, false);
    }

可以在创建Handler的时候传递进来，这样就可以回调监听。就是一个接口回调，这里就不举例了。

如果mCallback == null就调用handleMessage，在源码中handleMessage是一个空函数，可以继承Handler实现handleMessage方法来进行监听。Looper中的demo就是实现handleMessage方法实现监听的。也可继承Handler实现，道理相同

//Handler中的handleMessage方法，一个空函数
public void handleMessage(Message msg) {
}

总结

Android的消息机制整体来说也就是Handler的消息机制，Handler、Looper和MessageQueue三者结合，实现了线程间的通信。明白这种机制，使用更加如鱼得水。

Read the fucking source code!