AsyncTask分析（二）—Future、Callable、FutureTask
之前分析AsyncTask源码的时候，在其中提到了Future,Callable,FutureTask三个类或接口，今天就揭开他们的神秘面纱，探究他们的运用。
先看一下Future接口源码：
//Future是一个接口，他提供给了我们方法来检测当前的任务是否已经结束
//，还可以等待任务结束并且拿到一个结果，说白了他可以用来进行线程同步
public interface Future<V> {
//如果任务已经完成或者已经停止了或者这个任务无法停止，则会返回一个false
boolean cancel(boolean mayInterruptIfRunning);
//判断任务是否取消，
boolean isCancelled();
//判断任务是否完成
boolean isDone();
//取值，任务未完成之前，阻塞当前线程，直到任务完成,
//如果已经调用cancel()，在调用gete,会抛异常
V get() throws InterruptedException, ExecutionException;
//设置最大等待时间（线程阻塞最长时间），其余同get()
V get(long timeout, TimeUnit unit)
throws InterruptedException, ExecutionException, TimeoutException;
}干说，可能不太好理解，我们来个Demo,测试一下
private static void executor() throws Exception{
ExecutorService service = Executors.newSingleThreadExecutor();
CountRunnable count = new CountRunnable();
Future f = service.submit(count);
long start = System.currentTimeMillis();
System.out.println("start:"+start);
try{
System.out.println("result:"+f.get(10*1000, TimeUnit.MILLISECONDS));
}catch(Exception e){
System.out.println(e.toString());
}finally{
service.shutdown();
}
long end = System.currentTimeMillis();
System.out.println("任务结束 end:"+end+",共耗时："+(end-start));
}
static class CountRunnable implements Runnable{
private int sum;
@Override
public void run() {
for(int i=1 ; i<11 ; i++){
sum = sum+i;
try {
TimeUnit.SECONDS.sleep(2);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
System.out.println("run 结束 sum="+sum);
}
}
public static void main(String[] args) {
try {
executor();
} catch (Exception e) {
e.printStackTrace();
}
}看下打印结果：
Future.get()方法的的确确是阻塞了当前线程（因此如果在Android上使用Future不能再UIThread调用Future.get();），任务线程的完成时间是20s,但是我们的get()设置的阻塞时间是10s,在10s内无法返回结果，就会报TimeOut异常。为什么 任务线程最后输出还是sum=55?那是因为我们并没有取消任务，任务继续执行。
继续来看Callable:
//同样是一个泛型接口，只有一个call()
public interface Callable<V> {
V call() throws Exception;
}这个是用来做什么的呢？看过文档注释就会明白，在普通的Runnable执行过程中，也就是run()执行是没有返回值的，那我们也就无法判断任务的执行情况，也没法决定在什么时候停止任务，取消任务，但是Callable call()就给我们提供了这个判断。
看一个Demo
static void callTest(){
ExecutorService pool = Executors.newSingleThreadExecutor();
CountCallable count = new CountCallable();
Future<Number> f = pool.submit(count);
long start = System.currentTimeMillis();
System.out.println("start:"+start);
try{
System.out.println("result:"+f.get().number);
}catch(Exception e){
System.out.println(e.toString());
}finally{
pool.shutdown();
}
long end = System.currentTimeMillis();
System.out.println("任务结束 end:"+end+",共耗时："+(end-start));
}
static class CountCallable implements Callable<Number>{
@Override
public Number call() throws Exception {
Number number = new Number();
number.setNumber(100);
System.out.println("call...");
TimeUnit.SECONDS.sleep(2);
return number;
}
}
static class Number{
public int number;
public void setNumber(int number){
this.number = number;
}
}看下结果：
当pool.submit()之后，就会立即回调call(),当任务完成之后，就会返回我们在call()中设置的Number值。
刚刚说了，Future还有个功能是取消任务，不妨来试一下：
static void cancelTest() throws Exception{
ExecutorService pool = Executors.newSingleThreadExecutor();
CountCallable count = new CountCallable();
Future<Number> f = pool.submit(count);
System.out.println("任务开始于:"+System.currentTimeMillis());
TimeUnit.SECONDS.sleep(4);
f.cancel(true);
if(f.isCancelled()){
System.out.println("任务被取消于："+System.currentTimeMillis());
}else{
Number b = f.get();
System.out.println("任务继续执行：");
if(f.isDone()){
System.out.println("任务执行完毕："+System.currentTimeMillis());
pool.shutdown();
}
}
}看一下打印结果：
，cancel()之后，是不能再调get(),不然会报异常。
到这里Future和Callable基本就介绍完了，下面再看一下FutureTask
FutureTask登场
Future是一个接口，他的唯一实现类就是FutureTask，其实FutureTask的一个很好地特点是他有一个回调函数done（）方法，当一个任务执行结束后，会回调这个done（）方法，我们可以在done（）方法中调用FutureTask的get（）方法来获得计算的结果。为什么我们要在done（）方法中去调用get（）方法呢？ 这是有原因的，我在Android开发中，如果我在主线程去调用futureTask.get()方法时，会阻塞我的UI线程，如果在done（）方法里调用get（），则不会阻塞我们的UI线程。
先看一下
public class FutureTask<V> implements RunnableFuture<V>{}
public interface RunnableFuture<V> extends Runnable, Future<V> {}这就知道了FutureTask,即可作为Runnable,也可作为Future;
那就看下FutureTask中的代码：
public void run() {
if (state != NEW ||
!U.compareAndSwapObject(this, RUNNER, null, Thread.currentThread()))
return;
try {
Callable<V> c = callable;
if (c != null && state == NEW) {
V result;
boolean ran;
try {
result = c.call();
ran = true;
} catch (Throwable ex) {
result = null;
ran = false;
setException(ex);
}
if (ran)
set(result);
}
} finally {
// runner must be non-null until state is settled to
// prevent concurrent calls to run()
runner = null;
// state must be re-read after nulling runner to prevent
// leaked interrupts
int s = state;
if (s >= INTERRUPTING)
handlePossibleCancellationInterrupt(s);
}
}
protected void setException(Throwable t) {
if (U.compareAndSwapInt(this, STATE, NEW, COMPLETING)) {
outcome = t;
U.putOrderedInt(this, STATE, EXCEPTIONAL); // final state
finishCompletion();
}
}当执行FutureTask中的run()时，真正起作用的其实是Callable接口在调用call(),这里最终也是调用了finishCompletion();
再来看一段：
public boolean cancel(boolean mayInterruptIfRunning) {
if (!(state == NEW &&
U.compareAndSwapInt(this, STATE, NEW,
mayInterruptIfRunning ? INTERRUPTING : CANCELLED)))
return false;
try { // in case call to interrupt throws exception
if (mayInterruptIfRunning) {
try {
Thread t = runner;
if (t != null)
t.interrupt();
} finally { // final state
U.putOrderedInt(this, STATE, INTERRUPTED);
}
}
} finally {
finishCompletion();
}
return true;
}
private void finishCompletion() {
// assert state > COMPLETING;
for (WaitNode q; (q = waiters) != null;) {
if (U.compareAndSwapObject(this, WAITERS, q, null)) {
for (;;) {
Thread t = q.thread;
if (t != null) {
q.thread = null;
LockSupport.unpark(t);
}
WaitNode next = q.next;
if (next == null)
break;
q.next = null; // unlink to help gc
q = next;
}
break;
}
}
done();
callable = null; // to reduce footprint
}这里我们看到，当调用cancel(boolean)后，如果mayInterruptIfRunning = true,就会去执行t.interrupt();不论成功与否，最后执行finishCompletion(),执行done(),置空callable;
所以不论是cancel(返回true的时候),还是run最终都会调用done()
来个Demo验证下：
static class CountCallable implements Callable<Number>{
@Override
public Number call() throws Exception {
Number n = new Number();
n.number = 100;
System.out.println("call....");
TimeUnit.SECONDS.sleep(10);
return n;
}
}
static class Number{
public int number;
}
FutureTask<Number> task ;
ExecutorService es;
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.activity_main);
CountCallable callable = new CountCallable();
task = new FutureTask<Number>(callable){
@Override
protected void done() {
try {
Number result = task.get();
System.out.println("任务结束 number:"+result.number+",结束："+System.currentTimeMillis()+",isDone:"+task.isCancelled()+",isCancel:"+task.isCancelled()+",theadName:"+Thread.currentThread().getName());
} catch (Exception e) {
e.printStackTrace();
}
}
};
es = Executors.newFixedThreadPool(2);
es.execute(task);
System.out.println("任务开始："+System.currentTimeMillis());
findViewById(R.id.btn).setOnClickListener(new OnClickListener() {
@Override
public void onClick(View v) {
if(task.isDone()||task.isCancelled()){
return;
}
System.out.println(task.cancel(true)+",isCanceled:"+task.isCancelled());
}
});
}看下结果：
1，不点击cancel:
为毛要在done()里面调用get(),因为done()是运行在子线程的，不会阻塞UIThread,
2,点击cancel:
task.cancel()返回true;
那在回过头来看下AsyncTask中的运用：
public AsyncTask() {
//Callable接口的实现类，也就是FutureTask run()中真正工作的部分
mWorker = new WorkerRunnable<Params, Result>() {
public Result call() throws Exception {
mTaskInvoked.set(true);
Process.setThreadPriority(Process.THREAD_PRIORITY_BACKGROUND);
// 暴露到程序中的就是doInBackGround();
Result result = doInBackground(mParams);
Binder.flushPendingCommands();
return postResult(result);
}
};
mFuture = new FutureTask<Result>(mWorker) {
@Override
protected void done() {
try {
//这里获取的就是call()中的result
postResultIfNotInvoked(get());
} catch (InterruptedException e) {
android.util.Log.w(LOG_TAG, e);
} catch (ExecutionException e) {
throw new RuntimeException("An error occurred while executing doInBackground()",
e.getCause());
} catch (CancellationException e) {
postResultIfNotInvoked(null);
}
}
};
}
public final boolean cancel(boolean mayInterruptIfRunning) {
mCancelled.set(true);
//同样交给了Future去处理
return mFuture.cancel(mayInterruptIfRunning);
}我们重点追踪execute()这个流程：
@MainThread
public final AsyncTask<Params, Progress, Result> execute(Params... params) {
return executeOnExecutor(sDefaultExecutor, params);
}
@MainThread
public final AsyncTask<Params, Progress, Result> 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;
//这里其实就是SerialExecutor的execute()
exec.execute(mFuture);
return this;
}
private static volatile Executor sDefaultExecutor = SERIAL_EXECUTOR;
public static final Executor SERIAL_EXECUTOR = new SerialExecutor();
private static class SerialExecutor implements Executor {
final ArrayDeque<Runnable> mTasks = new ArrayDeque<Runnable>();
Runnable mActive;
public synchronized void execute(final Runnable r) {
mTasks.offer(new Runnable() {
public void run() {
try {
r.run();
} finally {
scheduleNext();
}
}
});
if (mActive == null) {
scheduleNext();
}
}
protected synchronized void scheduleNext() {
if ((mActive = mTasks.poll()) != null) {
THREAD_POOL_EXECUTOR.execute(mActive);
}
}
}
一步一步追踪下来，发现最终都是在SerialExecutor里面了，exec.execute(mFuture)其实就是SerialExecutor执行的execute(mFuture);
这里面我们看到直接调用了mFuture.run();
执行的流程图如下：
