AsyncTask 的两个线程池原理探究
熟悉AsyncTask的人都知道，AsyncTask中维护了两个线程池
//第一个线程池，用作执行具体任务
public static final Executor THREAD_POOL_EXECUTOR;
static {
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
CORE_POOL_SIZE, MAXIMUM_POOL_SIZE, KEEP_ALIVE_SECONDS, TimeUnit.SECONDS,
sPoolWorkQueue, sThreadFactory);
threadPoolExecutor.allowCoreThreadTimeOut(true);
THREAD_POOL_EXECUTOR = threadPoolExecutor;
}
/**
* An {@link Executor} that executes tasks one at a time in serial
* order. This serialization is global to a particular process.
*/
//第二个线程池，用作阻塞队列，进行任务的存储
public static final Executor SERIAL_EXECUTOR = new SerialExecutor();
private static final int MESSAGE_POST_RESULT = 0x1;
private static final int MESSAGE_POST_PROGRESS = 0x2;
//注意sDefaultExecutor ，马上就会提到
private static volatile Executor sDefaultExecutor = SERIAL_EXECUTOR;其中第二个线程池SERIAL_EXECUTOR 只用作排队，具体任务的执行仍由THREAD_POOL_EXECUTOR进行处理；
让我们从AsyncTask的execute函数查看执行原理；
@MainThread
public final AsyncTask execute(Params... params) {
return executeOnExecutor(sDefaultExecutor, params);
} 可以看到，execute函数默认调用了executeOnExecutor 方法，默认执行线程是sDefaultExecutor，即默认使用SERIAL_EXECUTOR来执行任务；
那么这个SERIAL_EXECUTOR到底做了什么呢
@MainThread
public final AsyncTask executeOnExecutor(Executor exec,
Params... params) {
if (mStatus != Status.PENDING) {
switch (mStatus) {
case RUNNING:
throw new IllegalStateException("Cannot execute task:"
+ " the task is already running.");
case FINISHED:
throw new IllegalStateException("Cannot execute task:"
+ " the task has already been executed "
+ "(a task can be executed only once)");
}
}
mStatus = Status.RUNNING;
onPreExecute();
mWorker.mParams = params;
exec.execute(mFuture);
return this;
} 可以看到这个executeOnExecutor函数，首先确保该AsyncTask只执行一次，然后执行onPreExecute，真正的执行逻辑是调用了SERIAL_THREADPOOL的execute方法；
private static class SerialExecutor implements Executor {
final ArrayDeque mTasks = new ArrayDeque();
Runnable mActive;
//加锁，确保只有一个任务得到执行
public synchronized void execute(final Runnable r) {
mTasks.offer(new Runnable() {
public void run() {
try {
r.run();
} finally {
scheduleNext();
}
}
});
//当前没有任务在被执行，则尝试加入新任务
if (mActive == null) {
scheduleNext();
}
}
//任务队列中如果有新任务，则将该任务从队列中取出，交给THREAD_POOL线程池执行；
protected synchronized void scheduleNext() {
if ((mActive = mTasks.poll()) != null) {
THREAD_POOL_EXECUTOR.execute(mActive);
}
}
} 再找到这个类的execute方法。可以非常清楚的知道，这个线程池仅仅用作排队使用，并且保证了THREAD_POOL一次只执行一个任务；
单看这一步，其实可以认为 两个线程池是没有必要的，一个执行，一个排队，就是标准的newSingleThreadExecutor ，那么为什么设计时候要使用两个线程池提高资源浪费呢；
答案就在马上揭晓：
我们正常使用asyncTask，往往是AysnTask.execute();内部帮我们调用了executeOnExecutor(sDefaultExecutor, params);
但是Google同样给我们留下了并行的方法，我们可以自己调用executeOnExecutor(THREAD_POOL);来并行执行任务；而且这种策略有一个好处；默认的队列执行是线性的，但是通过源码看出，线性的意思是，只有SERIAL_EXECUTOR提交的任务需要线性执行，我们通过THREAD_POOL执行的任务是不受影响的，即使正在运行，只要mActive为空，那么SERIAL_EXECUTOR将会提交新任务给THREAD_POOL执行；因此，这种设计，能够让线性的保持线性，不需要线性的任务就直接插队进行并行处理；
下面展示THREAD_POOL的构造源码，一次最多只能并发执行4个任务；
// We want at least 2 threads and at most 4 threads in the core pool,
// preferring to have 1 less than the CPU count to avoid saturating
// the CPU with background work
private static final int CORE_POOL_SIZE = Math.max(2, Math.min(CPU_COUNT - 1, 4));
private static final int MAXIMUM_POOL_SIZE = CPU_COUNT * 2 + 1;
private static final int KEEP_ALIVE_SECONDS = 30;
private static final ThreadFactory sThreadFactory = new ThreadFactory() {
private final AtomicInteger mCount = new AtomicInteger(1);
public Thread newThread(Runnable r) {
return new Thread(r, "AsyncTask #" + mCount.getAndIncrement());
}
};
private static final BlockingQueue sPoolWorkQueue =
new LinkedBlockingQueue(128);
/**
* An {@link Executor} that can be used to execute tasks in parallel.
*/
public static final Executor THREAD_POOL_EXECUTOR;
static {
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
CORE_POOL_SIZE, MAXIMUM_POOL_SIZE, KEEP_ALIVE_SECONDS, TimeUnit.SECONDS,
sPoolWorkQueue, sThreadFactory);
threadPoolExecutor.allowCoreThreadTimeOut(true);
THREAD_POOL_EXECUTOR = threadPoolExecutor;
}
回到开始，我们能看到，AsyncTask的任务提交函数都通过注解设定了MainThread，这是为什么呢?
首先，我们回到使用AsyncTask的使用上，首先，他先在主线程执行onPreExecute()，可以做一些加载图标，对话框，进度条之类的。再在子线程执行doInBackground(),最后，我们希望在执行结束后，在主线程执行onPostExecute()，取消对话框 进度条，或者其他的View的更新。
因此，我们需要在切换子线程和主线程；
private static Handler getMainHandler() {
synchronized (AsyncTask.class) {
if (sHandler == null) {
sHandler = new InternalHandler(Looper.getMainLooper());
}
return sHandler;
}
}
private Handler getHandler() {
return mHandler;
}
.........
.........
.........
public AsyncTask(@Nullable Looper callbackLooper) {
mHandler = callbackLooper == null || callbackLooper == Looper.getMainLooper()
? getMainHandler()
: new Handler(callbackLooper);
mWorker = new WorkerRunnable() {
public Result call() throws Exception {
mTaskInvoked.set(true);
Result result = null;
try {
Process.setThreadPriority(Process.THREAD_PRIORITY_BACKGROUND);
//noinspection unchecked
result = doInBackground(mParams);
Binder.flushPendingCommands();
} catch (Throwable tr) {
mCancelled.set(true);
throw tr;
} finally {
postResult(result);
}
return result;
}
}; 我们在主线程创建，保证了handler在主线程创建，便于onProgressUpdate， onPostExecute等 可以正常更新UI。
private Result postResult(Result result) {
@SuppressWarnings("unchecked")
Message message = getHandler().obtainMessage(MESSAGE_POST_RESULT,
new AsyncTaskResult(this, result));
message.sendToTarget();
return result;
} 下面给出一个简单的Demo，展示一下并发AsyncTask和串行AsyncTask的使用结果；
//创建10个串行AysnTask ，延迟500ms输出
for(int i =0;i<10;i++) {
new AsyncTask() {
@Override
protected void onPreExecute() {
super.onPreExecute();
}
@Override
protected void onPostExecute(Void aVoid) {
}
@Override
protected Void doInBackground(Void... voids) {
try {
Thread.sleep(500);
} catch (InterruptedException e) {
}
Log.e("AsynTask" +
"", "finished");
return null;
}
}.execute(null, null, null);
}
//每隔200ms 创建1个并行AysnTask ，延迟500ms输出
for(int i =0;i<10;i++) {
new AsyncTask() {
@Override
protected void onPreExecute() {
super.onPreExecute();
}
@Override
protected void onPostExecute(Void aVoid) {
}
@Override
protected Void doInBackground(Void... voids) {
try {
Thread.sleep(500);
} catch (InterruptedException e) {
}
Log.e("AsynTask" +
"", "finishedCon");
return null;
}
}.executeOnExecutor(AsyncTask.THREAD_POOL_EXECUTOR);
try {
Thread.sleep(200);
} catch (InterruptedException e) {
} 忽略指令执行耗时，我们一次性创建了10个串行和1个并行，之后每0.2s创建了1个并行，创建完10个后 花费1.8秒；此时打印了3个串行；
我们看一下打印结果
//创建了第10个串行和一个并行，等待500ms打印第一个串行和第一个并行，然后第二个串行进入等待
08-16 20:38:40.202 26282-26346/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:40.202 26282-26347/com.example.administrator.singletoptest E/AsynTask: finishedCon
//再过200ms，此时第700ms，打印第二个并行
08-16 20:38:40.405 26282-26348/com.example.administrator.singletoptest E/AsynTask: finishedCon
//900ms，第三个并行
08-16 20:38:40.607 26282-26349/com.example.administrator.singletoptest E/AsynTask: finishedCon
//1000ms， 第二个并行等待500ms后，打印
08-16 20:38:40.707 26282-26350/com.example.administrator.singletoptest E/AsynTask: finished
//1100，第三个并行，打印；
08-16 20:38:40.810 26282-26351/com.example.administrator.singletoptest E/AsynTask: finishedCon
//1300
08-16 20:38:41.013 26282-26352/com.example.administrator.singletoptest E/AsynTask: finishedCon
//1500，第三个串行打印完成，第5个并行打印完成，此时第4个串行进入等待；
08-16 20:38:41.211 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:41.213 26282-26354/com.example.administrator.singletoptest E/AsynTask: finishedCon
//1700,不再分析
08-16 20:38:41.415 26282-26346/com.example.administrator.singletoptest E/AsynTask: finishedCon
08-16 20:38:41.616 26282-26347/com.example.administrator.singletoptest E/AsynTask: finishedCon
08-16 20:38:41.712 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:41.816 26282-26349/com.example.administrator.singletoptest E/AsynTask: finishedCon
08-16 20:38:42.019 26282-26350/com.example.administrator.singletoptest E/AsynTask: finishedCon
08-16 20:38:42.213 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:42.714 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:43.216 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:43.718 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:44.220 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
08-16 20:38:44.721 26282-26353/com.example.administrator.singletoptest E/AsynTask: finished
总结：
1、AsyncTask，默认的两个线程池可以用一个newSingleThreadExecutor 单任务线程池代替
2、使用两个线程池，可以做到需要串行的任务保持串行提交任务，不需要串行完成的任务可以插队进行并行提交；
3、串行的任务只需要SERIAL_EXECUTOR提交的任务完成即可准备执行，不需要等待并行提交的任务完成；