Android Lifecycle源码解析(一)
首先我们看HomeActivity中我们添加到一行代码
public class HomeActivity extends AppCompatActivity {
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.activity_main);
//here
getLifecycle().addObserver(new HomeObserver());
}
}
先看下类的继承关系
public class AppCompatActivity extends FragmentActivity
public class FragmentActivity extends SupportActivity
我们看下getLifecycle()是个什么东西？点击去看发现getLifecycle()方法是定义在SupportActivity，下面是其部分源码
SupportActivity
public class SupportActivity extends Activity implements LifecycleOwner, Component {
...
private LifecycleRegistry mLifecycleRegistry = new LifecycleRegistry(this);
public SupportActivity() {
}
protected void onCreate(@Nullable Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
ReportFragment.injectIfNeededIn(this);
}
@CallSuper
protected void onSaveInstanceState(Bundle outState) {
this.mLifecycleRegistry.markState(State.CREATED);
super.onSaveInstanceState(outState);
}
public Lifecycle getLifecycle() {
return this.mLifecycleRegistry;
}
...
}
可以看到getLifecycle返回的其实是个LifecycleRegistry对象；该类继承了Lifecycle抽象类；
Lifecycle类可以理解为具备获取生命周期状态的抽象类，为什么？看下面代码
public abstract class Lifecycle {
@MainThread
public abstract void addObserver(@NonNull LifecycleObserver observer);
@MainThread
public abstract void removeObserver(@NonNull LifecycleObserver observer);
@MainThread
@NonNull
public abstract State getCurrentState();
}
Lifecycle声明了添加观察者，移除观察者，获取当前生命周期状态三个抽象方法
而LifecycleRegistry则是其子类，他是观察者和被观察者(HomeActivity)的纽带，能够管理多个观察者；HomeActivity的生命周期事件都是通过该类handleLifecycleEvent方法进行下发的，那么是谁将Activity的生命周期的事件转给LifecycleRegistry类中呢？答案是ReportFragment
这里有点像Glide框架中到设计一样使用一个空Fragment来感知Activity生命周期，避免图片加载出现内存泄漏
ReportFragment这家伙是在哪出现的？我们在SupportActivity类中的onCreate找到答案
class SupportActivity...
protected void onCreate(@Nullable Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
ReportFragment.injectIfNeededIn(this);
}
ReportFragment
@RestrictTo(RestrictTo.Scope.LIBRARY_GROUP)
public class ReportFragment extends Fragment {
private static final String REPORT_FRAGMENT_TAG = "android.arch.lifecycle"
+ ".LifecycleDispatcher.report_fragment_tag";
/**
**在Activity上附上一个空视图的ReportFragment
**/
public static void injectIfNeededIn(Activity activity) {
// ProcessLifecycleOwner should always correctly work and some activities may not extend
// FragmentActivity from support lib, so we use framework fragments for activities
android.app.FragmentManager manager = activity.getFragmentManager();
if (manager.findFragmentByTag(REPORT_FRAGMENT_TAG) == null) {
manager.beginTransaction().add(new ReportFragment(), REPORT_FRAGMENT_TAG).commit();
// Hopefully, we are the first to make a transaction.
manager.executePendingTransactions();
}
}
private ActivityInitializationListener mProcessListener;
private void dispatchCreate(ActivityInitializationListener listener) {
if (listener != null) {
listener.onCreate();
}
}
...
@Override
public void onActivityCreated(Bundle savedInstanceState) {
super.onActivityCreated(savedInstanceState);
dispatchCreate(mProcessListener);
dispatch(Lifecycle.Event.ON_CREATE);
}
@Override
public void onStart() {
super.onStart();
dispatchStart(mProcessListener);
dispatch(Lifecycle.Event.ON_START);
}
...
@Override
public void onDestroy() {
super.onDestroy();
dispatch(Lifecycle.Event.ON_DESTROY);
// just want to be sure that we won't leak reference to an activity
mProcessListener = null;
}
//可以看到Activtity的onCreate方法，以及ResportFragment中的onStart等各种生命周期方法都调用该方法，
private void dispatch(Lifecycle.Event event) {
Activity activity = getActivity();
//v7中LifecycleRegistryOwner类不建议使用
if (activity instanceof LifecycleRegistryOwner) {
((LifecycleRegistryOwner) activity).getLifecycle().handleLifecycleEvent(event);
return;
}
//因为SupportActivity实现了LifecycleOwner接口，所以会走此处分支
if (activity instanceof LifecycleOwner) {
Lifecycle lifecycle = ((LifecycleOwner) activity).getLifecycle();
if (lifecycle instanceof LifecycleRegistry) {
//关键点：生命周期事件转发给了LifecycleRegistry中了
((LifecycleRegistry) lifecycle).handleLifecycleEvent(event);
}
}
}
...
}
至此我们明白了Activity的生命周期事件是如何被转发到LifecycleRegistry中了，以一张流程图总结下
接下来我们看下LifecycleRegistry类是如何处理事件的
LifecycleRegistry
我理解的是LifecycleRegistry类和java中的Observable类有点像；因为Lifecycle的设计思想有部分就是使用了观察者模式，LifecycleRegistry类中可以注册或反注册多个观察者，一旦生命周期发生变更LifecycleRegistry会及时通知所有的观察者，这样就达到了观察者能够感知带生命周期组件的变化了。
/**
* An implementation of {@link Lifecycle} that can handle multiple observers.
* <p>
* It is used by Fragments and Support Library Activities. You can also directly use it if you have
* a custom LifecycleOwner.
*/
public class LifecycleRegistry extends Lifecycle {
private static final String LOG_TAG = "LifecycleRegistry";
/**
* Custom list that keeps observers and can handle removals / additions during traversal.
*
* Invariant: at any moment of time for observer1 & observer2:
* if addition_order(observer1) < addition_order(observer2), then
* state(observer1) >= state(observer2),
*/
private FastSafeIterableMap<LifecycleObserver, ObserverWithState> mObserverMap =
new FastSafeIterableMap<>();
/**
* Current state
*/
private State mState;
/**
* The provider that owns this Lifecycle.
* Only WeakReference on LifecycleOwner is kept, so if somebody leaks Lifecycle, they won't leak
* the whole Fragment / Activity. However, to leak Lifecycle object isn't great idea neither,
* because it keeps strong references on all other listeners, so you'll leak all of them as
* well.
*/
//使用弱引用避免Activity/Fragment内存泄漏
private final WeakReference<LifecycleOwner> mLifecycleOwner;
private int mAddingObserverCounter = 0;
private boolean mHandlingEvent = false;
private boolean mNewEventOccurred = false;
// we have to keep it for cases:
// void onStart() {
// mRegistry.removeObserver(this);
// mRegistry.add(newObserver);
// }
// newObserver should be brought only to CREATED state during the execution of
// this onStart method. our invariant with mObserverMap doesn't help, because parent observer
// is no longer in the map.
private ArrayList<State> mParentStates = new ArrayList<>();
/**
* Creates a new LifecycleRegistry for the given provider.
* <p>
* You should usually create this inside your LifecycleOwner class's constructor and hold
* onto the same instance.
*
* @param provider The owner LifecycleOwner
*/
public LifecycleRegistry(@NonNull LifecycleOwner provider) {
mLifecycleOwner = new WeakReference<>(provider);
mState = INITIALIZED;
}
/**
* Moves the Lifecycle to the given state and dispatches necessary events to the observers.
*
* @param state new state
*/
@SuppressWarnings("WeakerAccess")
@MainThread
public void markState(@NonNull State state) {
moveToState(state);
}
/**
* Sets the current state and notifies the observers.
* <p>
* Note that if the {@code currentState} is the same state as the last call to this method,
* calling this method has no effect.
*
* @param event The event that was received
*/
public void handleLifecycleEvent(@NonNull Lifecycle.Event event) {
State next = getStateAfter(event);
moveToState(next);
}
private void moveToState(State next) {
//1. 如果状态和当前状态一样，则不需要下发事件通知
if (mState == next) {
return;
}
//2. 更新最新生命周期事件状态
mState = next;
if (mHandlingEvent || mAddingObserverCounter != 0) {
mNewEventOccurred = true;
// we will figure out what to do on upper level.
return;
}
//3. 表示当前正在处理事件
mHandlingEvent = true;
//4. 同步事件到各Observer
sync();
//5. 同步结束，重置标志
mHandlingEvent = false;
}
// happens only on the top of stack (never in reentrance),
// so it doesn't have to take in account parents
private void sync() {
//获取具备生命周期的组件，如果没有直接返回
LifecycleOwner lifecycleOwner = mLifecycleOwner.get();
if (lifecycleOwner == null) {
Log.w(LOG_TAG, "LifecycleOwner is garbage collected, you shouldn't try dispatch "
+ "new events from it.");
return;
}
//如果未同步，那就同步状态直到完成
while (!isSynced()) {
mNewEventOccurred = false;
// no need to check eldest for nullability, because isSynced does it for us.
if (mState.compareTo(mObserverMap.eldest().getValue().mState) < 0) {
backwardPass(lifecycleOwner);
}
Entry<LifecycleObserver, ObserverWithState> newest = mObserverMap.newest();
if (!mNewEventOccurred && newest != null
&& mState.compareTo(newest.getValue().mState) > 0) {
forwardPass(lifecycleOwner);
}
}
mNewEventOccurred = false;
}
//判断最新的和最旧的观察者状态是否一致，如果一致说明同步状态成功；如果无观察者默认已同步
private boolean isSynced() {
if (mObserverMap.size() == 0) {
return true;
}
State eldestObserverState = mObserverMap.eldest().getValue().mState;
State newestObserverState = mObserverMap.newest().getValue().mState;
return eldestObserverState == newestObserverState && mState == newestObserverState;
}
private void forwardPass(LifecycleOwner lifecycleOwner) {
Iterator<Entry<LifecycleObserver, ObserverWithState>> ascendingIterator =
mObserverMap.iteratorWithAdditions();
while (ascendingIterator.hasNext() && !mNewEventOccurred) {
Entry<LifecycleObserver, ObserverWithState> entry = ascendingIterator.next();
ObserverWithState observer = entry.getValue();
while ((observer.mState.compareTo(mState) < 0 && !mNewEventOccurred
&& mObserverMap.contains(entry.getKey()))) {
pushParentState(observer.mState);
//可以看到最终事件是在这里分发给了observer
observer.dispatchEvent(lifecycleOwner, upEvent(observer.mState));
popParentState();
}
}
}
private void backwardPass(LifecycleOwner lifecycleOwner) {
Iterator<Entry<LifecycleObserver, ObserverWithState>> descendingIterator =
mObserverMap.descendingIterator();
while (descendingIterator.hasNext() && !mNewEventOccurred) {
Entry<LifecycleObserver, ObserverWithState> entry = descendingIterator.next();
ObserverWithState observer = entry.getValue();
while ((observer.mState.compareTo(mState) > 0 && !mNewEventOccurred
&& mObserverMap.contains(entry.getKey()))) {
Event event = downEvent(observer.mState);
pushParentState(getStateAfter(event));
observer.dispatchEvent(lifecycleOwner, event);
popParentState();
}
}
}
...
}
通过上面代码我们知道handleLifecycleEvent事件最后传递给observer的dispatchEvent方法中了；那么这个observer是什么？
首先看下它的类型是ObserverWithState，它是在什么时候被创建的呢？答案是addObserver方法
public class LifecycleRegistry extends Lifecycle {
private static final String LOG_TAG = "LifecycleRegistry";
private FastSafeIterableMap<LifecycleObserver, ObserverWithState> mObserverMap =
new FastSafeIterableMap<>();
/**
* Current state
*/
private State mState;
//使用弱引用避免Activity/Fragment内存泄漏
private final WeakReference<LifecycleOwner> mLifecycleOwner;
...
private ArrayList<State> mParentStates = new ArrayList<>();
...
@Override
public void addObserver(@NonNull LifecycleObserver observer) {
State initialState = mState == DESTROYED ? DESTROYED : INITIALIZED;
//创建一个带状态的Observer，可以理解为是observer的包装器
ObserverWithState statefulObserver = new ObserverWithState(observer, initialState);
//将observer以及带状态的observer保存到map中
ObserverWithState previous = mObserverMap.putIfAbsent(observer, statefulObserver);
//如果observer存在则无需添加，直接返回
if (previous != null) {
return;
}
//生命周期组件校验
LifecycleOwner lifecycleOwner = mLifecycleOwner.get();
if (lifecycleOwner == null) {
// it is null we should be destroyed. Fallback quickly
return;
}
boolean isReentrance = mAddingObserverCounter != 0 || mHandlingEvent;
State targetState = calculateTargetState(observer);
mAddingObserverCounter++;
while ((statefulObserver.mState.compareTo(targetState) < 0
&& mObserverMap.contains(observer))) {
pushParentState(statefulObserver.mState);
statefulObserver.dispatchEvent(lifecycleOwner, upEvent(statefulObserver.mState));
popParentState();
// mState / subling may have been changed recalculate
targetState = calculateTargetState(observer);
}
if (!isReentrance) {
// we do sync only on the top level.
sync();
}
mAddingObserverCounter--;
}
//带状态的Observer
static class ObserverWithState {
State mState;
GenericLifecycleObserver mLifecycleObserver;
ObserverWithState(LifecycleObserver observer, State initialState) {
//此处划重点，感兴趣可以进去仔细研究
mLifecycleObserver = Lifecycling.getCallback(observer);
mState = initialState;
}
void dispatchEvent(LifecycleOwner owner, Event event) {
State newState = getStateAfter(event);
mState = min(mState, newState);
mLifecycleObserver.onStateChanged(owner, event);
mState = newState;
}
}
}
可以看到在ObserverWithState的dispatchEvent方法最终将事件委托给mLifecycleObserver，
运行demo可以看到这个mLifecycleObserver类型其实是ReflectiveGenericLifecycleObserver类型；其实也好蛮理解，因为我们的观察者的方法标注了注解，而要想调用它，其实此处是通过反射来查找观察者的带生命周期注解的方法，一旦状态改变时，通过反射直接调用就行；
因为反射代价较大，所以不太建议此种写法，建议通过注解处理器，动态生成辅助类来提高性能，这个后面再说
class ReflectiveGenericLifecycleObserver implements GenericLifecycleObserver {
private final Object mWrapped;
private final CallbackInfo mInfo;
ReflectiveGenericLifecycleObserver(Object wrapped) {
mWrapped = wrapped;
mInfo = ClassesInfoCache.sInstance.getInfo(mWrapped.getClass());
}
@Override
public void onStateChanged(LifecycleOwner source, Event event) {
mInfo.invokeCallbacks(source, event, mWrapped);
}
}
GenericLifecycleObserver的实现类有五种，感兴趣可以看看
推荐写法
上面一开始写法涉及到反射，为提高性能我们使用注解处理器方式来实现，google已经为我们想好了
1. 集成
annotationProcessor "android.arch.lifecycle:compiler:1.1.1" //java
kapt "android.arch.lifecycle:compiler:1.1.1" //如果用kotlin
重新clean，rebuild下可以看到,注解处理器帮我创建了辅助类，通过调用辅助类来提高运行期性能
辅助类代码
public class HomeObserver_LifecycleAdapter implements GeneratedAdapter {
final HomeObserver mReceiver;
HomeObserver_LifecycleAdapter(HomeObserver receiver) {
this.mReceiver = receiver;
}
@Override
public void callMethods(LifecycleOwner owner, Lifecycle.Event event, boolean onAny,
MethodCallsLogger logger) {
boolean hasLogger = logger != null;
if (onAny) {
return;
}
if (event == Lifecycle.Event.ON_CREATE) {
if (!hasLogger || logger.approveCall("onCreate", 1)) {
mReceiver.onCreate();
}
return;
}
if (event == Lifecycle.Event.ON_RESUME) {
if (!hasLogger || logger.approveCall("onResume", 1)) {
mReceiver.onResume();
}
return;
}
if (event == Lifecycle.Event.ON_DESTROY) {
if (!hasLogger || logger.approveCall("onDestroy", 1)) {
mReceiver.onDestroy();
}
return;
}
}
}
给大家的问题
- ReflectiveGenericLifecycleObserver使用了反射机制类调用方法，为了减少反射带来的性能问题，它做了哪些优化？
- 我们写的HomeObserver中onCreate方法的参数有哪些限制？为什么
- 使用推荐写法时Lifecycling.getCallback(observer);返回的类型是什么？为什么？
- 使用推荐写法时到底有没使用到反射？
664
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
