Android源码解析之截屏事件流程

今天这篇文章我们主要讲一下Android系统中的截屏事件处理流程。用过android系统手机的同学应该都知道，一般的android手机按下音量减少键和电源按键就会触发截屏事件（国内定制机做个修改的这里就不做考虑了）。那么这里的截屏事件是如何触发的呢？触发之后android系统是如何实现截屏操作的呢？带着这两个问题，开始我们的源码阅读流程。

我们知道这里的截屏事件是通过我们的按键操作触发的，所以这里就需要我们从android系统的按键触发模块开始看起，由于我们在不同的App页面，操作音量减少键和电源键都会触发系统的截屏处理，所以这里的按键触发逻辑应该是Android系统的全局按键处理逻辑。

在android系统中，由于我们的每一个Android界面都是一个Activity，而界面的显示都是通过Window对象实现的，每个Window对象实际上都是PhoneWindow的实例，而每个PhoneWindow对象都一个PhoneWindowManager对象，当我们在Activity界面执行按键操作的时候，在将按键的处理操作分发到App之前，首先会回调PhoneWindowManager中的dispatchUnhandledKey方法，该方法主要用于执行当前App处理按键之前的操作，我们具体看一下该方法的实现。

/** {@inheritDoc} */
 @Override
 public KeyEvent dispatchUnhandledKey(WindowState win, KeyEvent event, int policyFlags) {
 ...
 KeyEvent fallbackEvent = null;
 if ((event.getFlags() & KeyEvent.FLAG_FALLBACK) == 0) {
 final KeyCharacterMap kcm = event.getKeyCharacterMap();
 final int keyCode = event.getKeyCode();
 final int metaState = event.getMetaState();
 final boolean initialDown = event.getAction() == KeyEvent.ACTION_DOWN
 && event.getRepeatCount() == 0;

 // Check for fallback actions specified by the key character map.
 final FallbackAction fallbackAction;
 if (initialDown) {
 fallbackAction = kcm.getFallbackAction(keyCode, metaState);
 } else {
 fallbackAction = mFallbackActions.get(keyCode);
 }

 if (fallbackAction != null) {
 ...
 final int flags = event.getFlags() | KeyEvent.FLAG_FALLBACK;
 fallbackEvent = KeyEvent.obtain(
 event.getDownTime(), event.getEventTime(),
 event.getAction(), fallbackAction.keyCode,
 event.getRepeatCount(), fallbackAction.metaState,
 event.getDeviceId(), event.getScanCode(),
 flags, event.getSource(), null);

 if (!interceptFallback(win, fallbackEvent, policyFlags)) {
 fallbackEvent.recycle();
 fallbackEvent = null;
 }

 if (initialDown) {
 mFallbackActions.put(keyCode, fallbackAction);
 } else if (event.getAction() == KeyEvent.ACTION_UP) {
 mFallbackActions.remove(keyCode);
 fallbackAction.recycle();
 }
 }
 }
 ...
 return fallbackEvent;
 }

这里我们关注一下方法体中调用的：interceptFallback方法，通过调用该方法将处理按键的操作下发到该方法中，我们继续看一下该方法的实现逻辑。

private boolean interceptFallback(WindowState win, KeyEvent fallbackEvent, int policyFlags) {
 int actions = interceptKeyBeforeQueueing(fallbackEvent, policyFlags);
 if ((actions & ACTION_PASS_TO_USER) != 0) {
 long delayMillis = interceptKeyBeforeDispatching(
 win, fallbackEvent, policyFlags);
 if (delayMillis == 0) {
 return true;
 }
 }
 return false;
 }

然后我们看到在interceptFallback方法中我们调用了interceptKeyBeforeQueueing方法，通过阅读我们我们知道该方法主要实现了对截屏按键的处理流程，这样我们继续看一下interceptKeyBeforeWueueing方法的处理：

@Override
 public int interceptKeyBeforeQueueing(KeyEvent event, int policyFlags) {
 if (!mSystemBooted) {
 // If we have not yet booted, don't let key events do anything.
 return 0;
 }

 ...
 // Handle special keys.
 switch (keyCode) {
 case KeyEvent.KEYCODE_VOLUME_DOWN:
 case KeyEvent.KEYCODE_VOLUME_UP:
 case KeyEvent.KEYCODE_VOLUME_MUTE: {
 if (mUseTvRouting) {
 // On TVs volume keys never go to the foreground app
 result &= ~ACTION_PASS_TO_USER;
 }
 if (keyCode == KeyEvent.KEYCODE_VOLUME_DOWN) {
 if (down) {
 if (interactive && !mScreenshotChordVolumeDownKeyTriggered
 && (event.getFlags() & KeyEvent.FLAG_FALLBACK) == 0) {
 mScreenshotChordVolumeDownKeyTriggered = true;
 mScreenshotChordVolumeDownKeyTime = event.getDownTime();
 mScreenshotChordVolumeDownKeyConsumed = false;
 cancelPendingPowerKeyAction();
 interceptScreenshotChord();
 }
 } else {
 mScreenshotChordVolumeDownKeyTriggered = false;
 cancelPendingScreenshotChordAction();
 }
 }
 ...

 return result;
 }

可以发现这里首先判断当前系统是否已经boot完毕，若尚未启动完毕，则所有的按键操作都将失效，若启动完成，则执行后续的操作，这里我们只是关注音量减少按键和电源按键组合的处理事件。另外这里多说一句想安卓系统的HOME按键事件，MENU按键事件，进程列表按键事件等等都是在这里实现的，后续中我们会陆续介绍这方面的内容。

回到我们的interceptKeyBeforeQueueing方法，当我用按下音量减少按键的时候回进入到：case KeyEvent.KEYCODE_VOLUME_MUTE分支并执行相应的逻辑，然后同时判断用户是否按下了电源键，若同时按下了电源键，则执行：

if (interactive && !mScreenshotChordVolumeDownKeyTriggered
 && (event.getFlags() & KeyEvent.FLAG_FALLBACK) == 0) {
 mScreenshotChordVolumeDownKeyTriggered = true;
 mScreenshotChordVolumeDownKeyTime = event.getDownTime();
 mScreenshotChordVolumeDownKeyConsumed = false;
 cancelPendingPowerKeyAction();
 interceptScreenshotChord();
 }

可以发现这里的interceptScreenshotChrod方法就是系统准备开始执行截屏操作的开始，我们继续看一下interceptcreenshotChord方法的实现。

private void interceptScreenshotChord() {
 if (mScreenshotChordEnabled
 && mScreenshotChordVolumeDownKeyTriggered && mScreenshotChordPowerKeyTriggered
 && !mScreenshotChordVolumeUpKeyTriggered) {
 final long now = SystemClock.uptimeMillis();
 if (now <= mScreenshotChordVolumeDownKeyTime + SCREENSHOT_CHORD_DEBOUNCE_DELAY_MILLIS
 && now <= mScreenshotChordPowerKeyTime
 + SCREENSHOT_CHORD_DEBOUNCE_DELAY_MILLIS) {
 mScreenshotChordVolumeDownKeyConsumed = true;
 cancelPendingPowerKeyAction();

 mHandler.postDelayed(mScreenshotRunnable, getScreenshotChordLongPressDelay());
 }
 }
 }

在方法体中我们最终会执行发送一个延迟的异步消息，请求执行截屏的操作而这里的延时时间，若当前输入框是打开状态，则延时时间为输入框关闭时间加上系统配置的按键超时时间，若当前输入框没有打开则直接是系统配置的按键超时处理时间，可看一下getScreenshotChordLongPressDelay方法的具体实现。

private long getScreenshotChordLongPressDelay() {
 if (mKeyguardDelegate.isShowing()) {
 // Double the time it takes to take a screenshot from the keyguard
 return (long) (KEYGUARD_SCREENSHOT_CHORD_DELAY_MULTIPLIER *
 ViewConfiguration.get(mContext).getDeviceGlobalActionKeyTimeout());
 }
 return ViewConfiguration.get(mContext).getDeviceGlobalActionKeyTimeout();
 }

回到我们的interceptScreenshotChord方法，发送了异步消息之后系统最终会被我们发送的Runnable对象的run方法执行，这里关于异步消息的逻辑可参考：android源码解析之（二）– 异步消息机制

这样我们看一下Runnable类型的mScreenshotRunnable的run方法的实现:

private final Runnable mScreenshotRunnable = new Runnable() {
 @Override
 public void run() {
 takeScreenshot();
 }
 };

好吧，方法体中并未执行其他操作，直接就是调用了takeScreenshot方法，这样我们继续看一下takeScreenshot方法的实现。

private void takeScreenshot() {
 synchronized (mScreenshotLock) {
 if (mScreenshotConnection != null) {
 return;
 }
 ComponentName cn = new ComponentName("com.android.systemui",
 "com.android.systemui.screenshot.TakeScreenshotService");
 Intent intent = new Intent();
 intent.setComponent(cn);
 ServiceConnection conn = new ServiceConnection() {
 @Override
 public void onServiceConnected(ComponentName name, IBinder service) {
 synchronized (mScreenshotLock) {
 if (mScreenshotConnection != this) {
 return;
 }
 Messenger messenger = new Messenger(service);
 Message msg = Message.obtain(null, 1);
 final ServiceConnection myConn = this;
 Handler h = new Handler(mHandler.getLooper()) {
 @Override
 public void handleMessage(Message msg) {
 synchronized (mScreenshotLock) {
  if (mScreenshotConnection == myConn) {
  mContext.unbindService(mScreenshotConnection);
  mScreenshotConnection = null;
  mHandler.removeCallbacks(mScreenshotTimeout);
  }
 }
 }
 };
 msg.replyTo = new Messenger(h);
 msg.arg1 = msg.arg2 = 0;
 if (mStatusBar != null && mStatusBar.isVisibleLw())
 msg.arg1 = 1;
 if (mNavigationBar != null && mNavigationBar.isVisibleLw())
 msg.arg2 = 1;
 try {
 messenger.send(msg);
 } catch (RemoteException e) {
 }
 }
 }
 @Override
 public void onServiceDisconnected(ComponentName name) {}
 };
 if (mContext.bindServiceAsUser(
 intent, conn, Context.BIND_AUTO_CREATE, UserHandle.CURRENT)) {
 mScreenshotConnection = conn;
 mHandler.postDelayed(mScreenshotTimeout, 10000);
 }
 }
 }

可以发现这里通过反射机制创建了一个TakeScreenshotService对象然后调用了bindServiceAsUser，这样就创建了TakeScreenshotService服务并在服务创建之后发送了一个异步消息。好了，我们看一下TakeScreenshotService的实现逻辑。

public class TakeScreenshotService extends Service {
 private static final String TAG = "TakeScreenshotService";

 private static GlobalScreenshot mScreenshot;

 private Handler mHandler = new Handler() {
 @Override
 public void handleMessage(Message msg) {
 switch (msg.what) {
 case 1:
 final Messenger callback = msg.replyTo;
 if (mScreenshot == null) {
 mScreenshot = new GlobalScreenshot(TakeScreenshotService.this);
 }
 mScreenshot.takeScreenshot(new Runnable() {
 @Override public void run() {
 Message reply = Message.obtain(null, 1);
 try {
 callback.send(reply);
 } catch (RemoteException e) {
 }
 }
 }, msg.arg1   0, msg.arg2   0);
 }
 }
 };

 @Override
 public IBinder onBind(Intent intent) {
 return new Messenger(mHandler).getBinder();
 }
}

可以发现在在TakeScreenshotService类的定义中有一个Handler成员变量，而我们在启动TakeScreentshowService的时候回发送一个异步消息，这样就会执行mHandler的handleMessage方法，然后在handleMessage方法中我们创建了一个GlobalScreenshow对象，然后执行了takeScreenshot方法，好吧，继续看一下takeScreentshot方法的执行逻辑。

/**
 * Takes a screenshot of the current display and shows an animation.
 */
 void takeScreenshot(Runnable finisher, boolean statusBarVisible, boolean navBarVisible) {
 // We need to orient the screenshot correctly (and the Surface api seems to take screenshots
 // only in the natural orientation of the device :!)
 mDisplay.getRealMetrics(mDisplayMetrics);
 float[] dims = {mDisplayMetrics.widthPixels, mDisplayMetrics.heightPixels};
 float degrees = getDegreesForRotation(mDisplay.getRotation());
 boolean requiresRotation = (degrees   0);
 if (requiresRotation) {
 // Get the dimensions of the device in its native orientation
 mDisplayMatrix.reset();
 mDisplayMatrix.preRotate(-degrees);
 mDisplayMatrix.mapPoints(dims);
 dims[0] = Math.abs(dims[0]);
 dims[1] = Math.abs(dims[1]);
 }

 // Take the screenshot
 mScreenBitmap = SurfaceControl.screenshot((int) dims[0], (int) dims[1]);
 if (mScreenBitmap == null) {
 notifyScreenshotError(mContext, mNotificationManager);
 finisher.run();
 return;
 }

 if (requiresRotation) {
 // Rotate the screenshot to the current orientation
 Bitmap ss = Bitmap.createBitmap(mDisplayMetrics.widthPixels,
 mDisplayMetrics.heightPixels, Bitmap.Config.ARGB_8888);
 Canvas c = new Canvas(ss);
 c.translate(ss.getWidth() / 2, ss.getHeight() / 2);
 c.rotate(degrees);
 c.translate(-dims[0] / 2, -dims[1] / 2);
 c.drawBitmap(mScreenBitmap, 0, 0, null);
 c.setBitmap(null);
 // Recycle the previous bitmap
 mScreenBitmap.recycle();
 mScreenBitmap = ss;
 }

 // Optimizations
 mScreenBitmap.setHasAlpha(false);
 mScreenBitmap.prepareToDraw();

 // Start the post-screenshot animation
 startAnimation(finisher, mDisplayMetrics.widthPixels, mDisplayMetrics.heightPixels,
 statusBarVisible, navBarVisible);
 }

可以看到这里后两个参数：statusBarVisible，navBarVisible是否可见，而这两个参数在我们PhoneWindowManager.takeScreenshot方法传递的：

if (mStatusBar != null && mStatusBar.isVisibleLw())
 msg.arg1 = 1;
 if (mNavigationBar != null && mNavigationBar.isVisibleLw())
 msg.arg2 = 1;

可见若果mStatusBar可见，则传递的statusBarVisible为true，若mNavigationBar可见，则传递的navBarVisible为true。然后我们在截屏的时候判断nStatusBar是否可见，mNavigationBar是否可见，若可见的时候则截屏同样将其截屏出来。继续回到我们的takeScreenshot方法，然后调用了：

// Take the screenshot
mScreenBitmap = SurfaceControl.screenshot((int) dims[0], (int) dims[1]);

方法，看注释，这里就是执行截屏事件的具体操作了，然后我看一下SurfaceControl.screenshot方法的具体实现，另外这里需要注意的是，截屏之后返回的是一个Bitmap对象，其实熟悉android绘制机制的童鞋应该知道android中所有显示能够显示的东西，在内存中表现都是Bitmap对象。

public static Bitmap screenshot(int width, int height) {
 // TODO: should take the display as a parameter
 IBinder displayToken = SurfaceControl.getBuiltInDisplay(
 SurfaceControl.BUILT_IN_DISPLAY_ID_MAIN);
 return nativeScreenshot(displayToken, new Rect(), width, height, 0, 0, true,
 false, Surface.ROTATION_0);
 }

好吧，这里调用的是nativeScreenshot方法，它是一个native方法，具体的实现在JNI层，这里就不做过多的介绍了。继续回到我们的takeScreenshot方法，在调用了截屏方法screentshot之后，判断是否截屏成功：

if (mScreenBitmap == null) {
 notifyScreenshotError(mContext, mNotificationManager);
 finisher.run();
 return;
 }

若截屏之后，截屏的bitmap对象为空，这里判断截屏失败，调用了notifyScreenshotError方法，发送截屏失败的notification通知。

static void notifyScreenshotError(Context context, NotificationManager nManager) {
 Resources r = context.getResources();

 // Clear all existing notification, compose the new notification and show it
 Notification.Builder b = new Notification.Builder(context)
 .setTicker(r.getString(R.string.screenshot_failed_title))
 .setContentTitle(r.getString(R.string.screenshot_failed_title))
 .setContentText(r.getString(R.string.screenshot_failed_text))
 .setSmallIcon(R.drawable.stat_notify_image_error)
 .setWhen(System.currentTimeMillis())
 .setVisibility(Notification.VISIBILITY_PUBLIC) // ok to show outside lockscreen
 .setCategory(Notification.CATEGORY_ERROR)
 .setAutoCancel(true)
 .setColor(context.getColor(
 com.android.internal.R.color.system_notification_accent_color));
 Notification n =
 new Notification.BigTextStyle(b)
 .bigText(r.getString(R.string.screenshot_failed_text))
 .build();
 nManager.notify(R.id.notification_screenshot, n);
 }

然后继续看takeScreenshot方法，判断截屏的图像是否需要旋转，若需要的话，则旋转图像：

if (requiresRotation) {
 // Rotate the screenshot to the current orientation
 Bitmap ss = Bitmap.createBitmap(mDisplayMetrics.widthPixels,
 mDisplayMetrics.heightPixels, Bitmap.Config.ARGB_8888);
 Canvas c = new Canvas(ss);
 c.translate(ss.getWidth() / 2, ss.getHeight() / 2);
 c.rotate(degrees);
 c.translate(-dims[0] / 2, -dims[1] / 2);
 c.drawBitmap(mScreenBitmap, 0, 0, null);
 c.setBitmap(null);
 // Recycle the previous bitmap
 mScreenBitmap.recycle();
 mScreenBitmap = ss;
 }

在takeScreenshot方法的最后若截屏成功，我们调用了：

// Start the post-screenshot animation
 startAnimation(finisher, mDisplayMetrics.widthPixels, mDisplayMetrics.heightPixels,
 statusBarVisible, navBarVisible);

开始截屏的动画，好吧，看一下动画效果的实现：

/**
 * Starts the animation after taking the screenshot
 */
 private void startAnimation(final Runnable finisher, int w, int h, boolean statusBarVisible,
 boolean navBarVisible) {
 // Add the view for the animation
 mScreenshotView.setImageBitmap(mScreenBitmap);
 mScreenshotLayout.requestFocus();

 // Setup the animation with the screenshot just taken
 if (mScreenshotAnimation != null) {
 mScreenshotAnimation.end();
 mScreenshotAnimation.removeAllListeners();
 }

 mWindowManager.addView(mScreenshotLayout, mWindowLayoutParams);
 ValueAnimator screenshotDropInAnim = createScreenshotDropInAnimation();
 ValueAnimator screenshotFadeOutAnim = createScreenshotDropOutAnimation(w, h,
 statusBarVisible, navBarVisible);
 mScreenshotAnimation = new AnimatorSet();
 mScreenshotAnimation.playSequentially(screenshotDropInAnim, screenshotFadeOutAnim);
 mScreenshotAnimation.addListener(new AnimatorListenerAdapter() {
 @Override
 public void onAnimationEnd(Animator animation) {
 // Save the screenshot once we have a bit of time now
 saveScreenshotInWorkerThread(finisher);
 mWindowManager.removeView(mScreenshotLayout);

 // Clear any references to the bitmap
 mScreenBitmap = null;
 mScreenshotView.setImageBitmap(null);
 }
 });
 mScreenshotLayout.post(new Runnable() {
 @Override
 public void run() {
 // Play the shutter sound to notify that we've taken a screenshot
 mCameraSound.play(MediaActionSound.SHUTTER_CLICK);

 mScreenshotView.setLayerType(View.LAYER_TYPE_HARDWARE, null);
 mScreenshotView.buildLayer();
 mScreenshotAnimation.start();
 }
 });
 }

好吧，经过着一些列的操作之后我们实现了截屏之后的动画效果了，这里暂时不分析动画效果，我们看一下动画效果之后做了哪些？还记不记的一般情况下我们截屏之后都会收到一个截屏的notification通知？这里应该也是在其AnimatorListenerAdapter的onAnimationEnd方法中实现的，也就是动画执行完成之后，我们看一下其saveScreenshotInWorkerThread方法的实现：

/**
 * Creates a new worker thread and saves the screenshot to the media store.
 */
 private void saveScreenshotInWorkerThread(Runnable finisher) {
 SaveImageInBackgroundData data = new SaveImageInBackgroundData();
 data.context = mContext;
 data.image = mScreenBitmap;
 data.iconSize = mNotificationIconSize;
 data.finisher = finisher;
 data.previewWidth = mPreviewWidth;
 data.previewheight = mPreviewHeight;
 if (mSaveInBgTask != null) {
 mSaveInBgTask.cancel(false);
 }
 mSaveInBgTask = new SaveImageInBackgroundTask(mContext, data, mNotificationManager,
 R.id.notification_screenshot).execute(data);
 }

好吧，这里主要逻辑就是构造了一个SaveImageInBackgroundTask对象，看样子发送截屏成功的通知应该是在这里实现的，我们看一下SaveImageInBackgroundTask构造方法的实现逻辑：

SaveImageInBackgroundTask(Context context, SaveImageInBackgroundData data,
 NotificationManager nManager, int nId) {
 ...

 // Show the intermediate notification
 mTickerAddSpace = !mTickerAddSpace;
 mNotificationId = nId;
 mNotificationManager = nManager;
 final long now = System.currentTimeMillis();

 mNotificationBuilder = new Notification.Builder(context)
 .setTicker(r.getString(R.string.screenshot_saving_ticker)
 + (mTickerAddSpace ? " " : ""))
 .setContentTitle(r.getString(R.string.screenshot_saving_title))
 .setContentText(r.getString(R.string.screenshot_saving_text))
 .setSmallIcon(R.drawable.stat_notify_image)
 .setWhen(now)
 .setColor(r.getColor(com.android.internal.R.color.system_notification_accent_color));

 mNotificationStyle = new Notification.BigPictureStyle()
 .bigPicture(picture.createAshmemBitmap());
 mNotificationBuilder.setStyle(mNotificationStyle);

 // For "public" situations we want to show all the same info but
 // omit the actual screenshot image.
 mPublicNotificationBuilder = new Notification.Builder(context)
 .setContentTitle(r.getString(R.string.screenshot_saving_title))
 .setContentText(r.getString(R.string.screenshot_saving_text))
 .setSmallIcon(R.drawable.stat_notify_image)
 .setCategory(Notification.CATEGORY_PROGRESS)
 .setWhen(now)
 .setColor(r.getColor(
 com.android.internal.R.color.system_notification_accent_color));

 mNotificationBuilder.setPublicVersion(mPublicNotificationBuilder.build());

 Notification n = mNotificationBuilder.build();
 n.flags |= Notification.FLAG_NO_CLEAR;
 mNotificationManager.notify(nId, n);

 // On the tablet, the large icon makes the notification appear as if it is clickable (and
 // on small devices, the large icon is not shown) so defer showing the large icon until
 // we compose the final post-save notification below.
 mNotificationBuilder.setLargeIcon(icon.createAshmemBitmap());
 // But we still don't set it for the expanded view, allowing the smallIcon to show here.
 mNotificationStyle.bigLargeIcon((Bitmap) null);
 }