什么是Lifecycle
Lifecycle是生命周期感知型组件，它可以通过响应另一个组件（Activity或Fragment）的生命周期的变化来执行相应的操作，它有助于产出更有条理、更轻量级、更易于维护的代码。
Lifecycle是LiveData和ViewModel的基础，下面就先介绍为何使用Lifecycle及如何使用Lifecycle。
使用Lifecycle之前
假设我们有一个在屏幕上显示设备位置信息的Activity。常见的实现可能如下所示：
java
class MyLocationListener {
public MyLocationListener(Context context, Callback callback) {
// ...
}
void start() {
// 开启定位服务
}
void stop() {
// 停止定位服务
}
}
class MyActivity extends AppCompatActivity {
private MyLocationListener myLocationListener;
@Override
public void onCreate(...) {
myLocationListener = new MyLocationListener(this, (location) -> {
// 更新 UI
});
}
@Override
public void onStart() {
super.onStart();
myLocationListener.start();
}
@Override
public void onStop() {
super.onStop();
myLocationListener.stop();
}
}上面的示例看起来好像没问题，但是在实际的应用中，你需要在Activity对应的生命周期方法（如上例中的onStart()、onStop()方法）中写很多代码来管理UI和其他组件，这时候你就会发现这些代码很难维护。
此外，无法保证Activity在执行onStop()方法之前执行myLocationListener的start()方法，例如下面这种情况：
java
class MyActivity extends AppCompatActivity {
private MyLocationListener myLocationListener;
public void onCreate(...) {
myLocationListener = new MyLocationListener(this, location -> {
// 更新 UI
});
}
@Override
public void onStart() {
super.onStart();
Util.checkUserStatus(result -> {
//如果checkUserStatus()方法耗时较长，在activity执行完onStop()方法后才回调，
//那么myLocationListener执行完start()方法后就没办法走stop()方法了，
//又因为myLocationListener持有activity，会造成内存泄漏。
if (result) {
myLocationListener.start();
}
});
}
@Override
public void onStop() {
super.onStop();
myLocationListener.stop();
}
}这样就导致myLocationListener不能按照预期执行stop()方法，又因为myLocationListener持有Activity的引用，从而导致内存泄漏。
上面出现了2个问题：
- Activity的生命周期方法中存在大量管理其他组件的代码，难以维护；
- 无法保证在退出Activity前执行相应的回收方法，容易导致内存泄露问题；
而Lifecycle可以帮助你解决这些问题。
Lifecycle的使用
引入依赖
- 非androidx项目引入:
arduino
implementation "android.arch.lifecycle:extensions:1.1.1"- androidx项目引入:
arduino
implementation 'androidx.appcompat:appcompat:1.2.0'因为appcompat依赖了androidx.fragment，而androidx.fragment下依赖了ViewModel和LiveData，LiveData内部又依赖了Lifecycle。
- 如果想要单独引入依赖，官方给出的依赖如下：
kotlin
//app的build.gradle
dependencies {
def lifecycle_version = "2.2.0"
def arch_version = "2.1.0"
// ViewModel
implementation "androidx.lifecycle:lifecycle-viewmodel:$lifecycle_version"
// LiveData
implementation "androidx.lifecycle:lifecycle-livedata:$lifecycle_version"
// 只有Lifecycles (不带 ViewModel or LiveData)
implementation "androidx.lifecycle:lifecycle-runtime:$lifecycle_version"
// Saved state module for ViewModel
implementation "androidx.lifecycle:lifecycle-viewmodel-savedstate:$lifecycle_version"
// lifecycle注解处理器
annotationProcessor "androidx.lifecycle:lifecycle-compiler:$lifecycle_version"
// 替换 - 如果使用Java8,就用这个替换上面的lifecycle-compiler
implementation "androidx.lifecycle:lifecycle-common-java8:$lifecycle_version"
//以下按需引入
// 可选 - 帮助实现Service的LifecycleOwner
implementation "androidx.lifecycle:lifecycle-service:$lifecycle_version"
// 可选 - ProcessLifecycleOwner给整个 app进程 提供一个lifecycle
implementation "androidx.lifecycle:lifecycle-process:$lifecycle_version"
// 可选 - ReactiveStreams support for LiveData
implementation "androidx.lifecycle:lifecycle-reactivestreams:$lifecycle_version"
// 可选 - Test helpers for LiveData
testImplementation "androidx.arch.core:core-testing:$arch_version"
}看着有很多，实际上如果只使用Lifecycle，只需要引入lifecycle-runtime即可，但通常都是和 ViewModel、 LiveData 配套使用的，所以lifecycle-viewmodel、lifecycle-livedata 一般也会引入。另外，lifecycle-process是给整个app进程提供一个lifecycle，后面也会提到。
基本使用
Lifecycle的使用方法很简单，主要分为2步：
- 生命周期拥有者使用getLifecycle()获取Lifecycle实例，然后调用addObserver()添加观察者；
- 观察者实现LifecycleObserver，方法上使用@OnLifecycleEvent注解关注对应的生命周期，生命周期触发时就会执行相应的方法；
代码如下：
java
class LifecycleTestActivity : AppCompatActivity() {
private val TAG = LifecycleTestActivity::class.java.simpleName
override fun onCreate(savedInstanceState: Bundle?) {
super.onCreate(savedInstanceState)
lifecycle.addObserver(MyObserver())
}
override fun onResume() {
super.onResume()
Log.i(TAG, "onResume: ")
}
override fun onPause() {
super.onPause()
Log.i(TAG, "onPause: ")
}
}
class MyObserver : LifecycleObserver {
private val TAG = "LifeCycleTest_MyObserver"
@OnLifecycleEvent(Lifecycle.Event.ON_RESUME)
fun connect() = Log.d(TAG,"connect run ...")
@OnLifecycleEvent(Lifecycle.Event.ON_PAUSE)
fun disconnect() = Log.d(TAG,"disconnect run ...")
}运行后打印如下：
arduino
LifecycleTestActivity com.example.test I onResume:
LifeCycleTest_MyObserver com.example.test D connect run ...
LifeCycleTest_MyObserver com.example.test D disconnect run ...
LifecycleTestActivity com.example.test I onPause: 这里看到connect run ...在onResume:之前打印，而disconnect run ...在onPause:之后打印，这是为什么呢？这个问题我们留到后面分析源码的时候再来解释。
自定义LifecycleOwner
直接在Activity中使用getLifecycle()就能获取到Lifecycle实例，getLifecycle()方法来自接口LifecycleOwner：
java
public interface LifecycleOwner {
@NonNull
Lifecycle getLifecycle();
}如果一个类实现了LifecycleOwner就表示这个类具有生命周期，你也可以自定义类实现LifecycleOwner接口。
support library 26.1.0及以上的Fragment和Activity已实现LifecycleOwner接口，所以我们在Activity中可以直接使用getLifecycle()方法。
如果你想要自己定义一个类实现LifecycleOwner接口，你可以使用LifecycleRegistry，它是Lifecycle的实现类，但你需要将事件转发到LifecycleRegistry：
java
public class MyActivity extends Activity implements LifecycleOwner {
private LifecycleRegistry lifecycleRegistry;
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
lifecycleRegistry = new LifecycleRegistry(this);
lifecycleRegistry.markState(Lifecycle.State.CREATED);
}
@Override
public void onStart() {
super.onStart();
lifecycleRegistry.markState(Lifecycle.State.STARTED);
}
@NonNull
@Override
public Lifecycle getLifecycle() {
return lifecycleRegistry;
}
}MyActivity实现了LifecycleOwner接口，getLifecycle()返回lifecycleRegistry实例。lifecycleRegistry实例是在onCreate()方法中创建的，并且在各个生命周期内调用markState()方法完成生命周期事件的传递。这就完成了LifecycleOwner的自定义，也即MyActivity变成了LifecycleOwner，然后就可以和实现了LifecycleObserver接口的类配合使用了。
补充一点，观察者的方法可以接受一个参数LifecycleOwner，就可以用来获取当前状态、或者继续添加观察者。 若注解的是ON_ANY还可以接收Event，用于区分是哪个事件。如下：
java
class TestObserver implements LifecycleObserver {
@OnLifecycleEvent(Lifecycle.Event.ON_CREATE)
void onCreated(LifecycleOwner owner) {
//owner.getLifecycle().addObserver(anotherObserver);
//owner.getLifecycle().getCurrentState();
}
@OnLifecycleEvent(Lifecycle.Event.ON_ANY)
void onAny(LifecycleOwner owner, Lifecycle.Event event) {
//event.name()
}
}MVP架构中使用Lifecycle
在MVP架构中可以把presenter作为观察者：
java
public class LifecycleTestActivity extends AppCompatActivity implements IView {
private String TAG = "Lifecycle_Test";
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.activity_lifecycle_test);
//Lifecycle 生命周期
//getLifecycle().addObserver(new MyObserver());
//MVP中使用Lifecycle
getLifecycle().addObserver(new MyPresenter(this));
Log.i(TAG, "onCreate: ");
}
@Override
protected void onResume() {
super.onResume();
Log.i(TAG, "onResume: ");
}
@Override
protected void onPause() {
super.onPause();
Log.i(TAG, "onPause: ");
}
@Override
public void showView() {}
@Override
public void hideView() {}
}
//Presenter
class MyPresenter implements LifecycleObserver {
private static final String TAG = "Lifecycle_Test";
private final IView mView;
public MyPresenter(IView view) {mView = view;}
@OnLifecycleEvent(value = Lifecycle.Event.ON_START)
private void getDataOnStart(LifecycleOwner owner){
Log.i(TAG, "getDataOnStart: ");
Util.checkUserStatus(result -> {
//checkUserStatus是耗时操作，回调后检查当前生命周期状态
if (owner.getLifecycle().getCurrentState().isAtLeast(STARTED)) {
start();
mView.showView();
}
});
}
@OnLifecycleEvent(value = Lifecycle.Event.ON_STOP)
private void hideDataOnStop(){
Log.i(TAG, "hideDataOnStop: ");
stop();
mView.hideView();
}
}
//IView
interface IView {
void showView();
void hideView();
}这里是让Presenter实现LifecycleObserver接口，同样在方法上注解要触发的生命周期，最后在Activity中作为观察者添加到Lifecycle中。
这样做好处是啥呢？ 当Activity生命周期发生变化时，MyPresenter就可以感知并执行方法，不需要在MainActivity的多个生命周期方法中调用MyPresenter的方法了。
Presenter类自动感知生命周期，如果需要在其他的Activity或Fragment也使用这个Presenter，只需添加其为观察者即可。让Presenter存储自己的逻辑，减轻Activity或Fragment中代码，更易于管理；
另外，注意到 getDataOnStart()中耗时校验回调后，对当前生命周期状态进行了检查：至少处于STARTED状态才会继续执行start()方法，也就是保证了Activity停止后不会走start()方法；
这样，使用LifeCycle后，文章前面遇到的2个问题得到了解决。
Application生命周期 ProcessLifecycleOwner
之前对App进入前后台的判断是通过registerActivityLifecycleCallbacks(callback)方法，然后在callback中利用一个全局变量做计数，在onActivityStarted()中计数加1，在onActivityStopped方法中计数减1，从而判断前后台切换。
而使用ProcessLifecycleOwner可以直接获取应用前后台切换状态。（记得先引入lifecycle-process依赖）
使用方式和Activity中类似，只不过要使用ProcessLifecycleOwner.get()获取ProcessLifecycleOwner，代码如下：
java
public class MyApplication extends Application {
@Override
public void onCreate() {
super.onCreate();
//注册App生命周期观察者
ProcessLifecycleOwner.get().getLifecycle().addObserver(new ApplicationLifecycleObserver());
}
/**
* Application生命周期观察，提供整个应用进程的生命周期
*
* Lifecycle.Event.ON_CREATE只会分发一次，Lifecycle.Event.ON_DESTROY不会被分发。
*
* 第一个Activity进入时，ProcessLifecycleOwner将分派Lifecycle.Event.ON_START, Lifecycle.Event.ON_RESUME。
* 而Lifecycle.Event.ON_PAUSE, Lifecycle.Event.ON_STOP，将在最后一个Activit退出后后延迟分发。如果由于
* 配置更改而销毁并重新创建活动，则此延迟足以保证ProcessLifecycleOwner不会发送任何事件。
*
* 作用：监听应用程序进入前台或后台
*/
private static class ApplicationLifecycleObserver implements LifecycleObserver {
@OnLifecycleEvent(Lifecycle.Event.ON_START)
private void onAppForeground() {
Log.w(TAG, "ApplicationObserver: app moved to foreground");
}
@OnLifecycleEvent(Lifecycle.Event.ON_STOP)
private void onAppBackground() {
Log.w(TAG, "ApplicationObserver: app moved to background");
}
}
}看到确实很简单，和前面Activity的Lifecycle用法几乎一样，而我们使用ProcessLifecycleOwner就显得很优雅了，生命周期分发逻辑已在注释里说明。
源码分析
本文源码基于androidx.lifecycle:lifecycle-runtime:2.3.1
Lifecycle持有另一个组件（比如Activity或Fragment）的生命周期状态有关的信息，主要使用两个枚举类型来追踪与其关联的组件的生命周期状态，Lifecycle代码如下：
java
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();
//对应LifecycleOwner的Event
public enum Event {
ON_CREATE,
ON_START,
ON_RESUME,
ON_PAUSE,
ON_STOP,
ON_DESTROY,
ON_ANY;
}
//对应LifecycleOwner的状态
public enum State {
//此事件后，Lifecycle将不再分发事件。对于Activity来说，
//在Activity的onDestroy()方法调用之前到达这个状态。
DESTROYED,
//对于Activity来说，这是Activity已经构建但是onCreate()方法调用前的状态。
INITIALIZED,
//对于Activity来说，这是Activity调用onCreate()后，调用onStop()前的状态。
CREATED,
//对于Activity来说，这是Activity调用onStart()后，调用onPause()前的状态。
STARTED,
//对于Activity来说，是调用Activity的onResume()方法后到达的状态。
RESUMED;
//比较当前状态是否>=传入的状态
public boolean isAtLeast(@NonNull State state) {
return compareTo(state) >= 0;
}
}
}这两个枚举类型分别是：
- Event：分发的生命周期事件，对应于Activity或Fragment中的生命周期方法。
- State：LifecycleOwner当前的状态。
关于Activity中State与Event的关系，官网有张图很清晰： 
其中State是图中的一个个节点，Event是这些节点之间的过渡。
被观察者流程
前面提到Activity实现了LifecycleOwner，所以才能直接使用getLifecycle()方法，具体是在androidx.activity.ComponentActivity中，我们来看看ComponentActivity中是怎么处理的:
java
public class ComponentActivity extends androidx.core.app.ComponentActivity implements LifecycleOwner{
private final LifecycleRegistry mLifecycleRegistry = new LifecycleRegistry(this);
@Override
protected void onCreate(@Nullable Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
...
//使用ReportFragment监听Activity的生命周期状态
ReportFragment.injectIfNeededIn(this);
...
}
@CallSuper
@Override
protected void onSaveInstanceState(@NonNull Bundle outState) {
mLifecycleRegistry.markState(Lifecycle.State.CREATED);
super.onSaveInstanceState(outState);
}
@NonNull
@Override
public Lifecycle getLifecycle() {
return mLifecycleRegistry;
}
}乍一看，并没有在ComponentActivity的生命周期方法中找到对应的代码，实际监听生命周期的处理在 ReportFragment的injectIfNeededIn()方法里中：
java
public class ReportFragment extends android.app.Fragment {
public static void injectIfNeededIn(Activity activity) {
if (Build.VERSION.SDK_INT >= 29) {
//API 29+，直接使用LifecycleCallbacks监听生命周期
LifecycleCallbacks.registerIn(activity);
}
//API 29之前，添加ReportFragment监听生命周期
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();
}
}
@Override
public void onActivityCreated(Bundle savedInstanceState) {
super.onActivityCreated(savedInstanceState);
dispatch(Lifecycle.Event.ON_CREATE);
}
@Override
public void onStart() {
super.onStart();
dispatch(Lifecycle.Event.ON_START);
}
@Override
public void onResume() {
super.onResume();
dispatch(Lifecycle.Event.ON_RESUME);
}
@Override
public void onPause() {
super.onPause();
dispatch(Lifecycle.Event.ON_PAUSE);
}
@Override
public void onStop() {
super.onStop();
dispatch(Lifecycle.Event.ON_STOP);
}
@Override
public void onDestroy() {
super.onDestroy();
dispatch(Lifecycle.Event.ON_DESTROY);
}
private void dispatch(@NonNull Lifecycle.Event event) {
if (Build.VERSION.SDK_INT < 29) {
// API 29之前才需要这样分发
dispatch(getActivity(), event);
}
}
static void dispatch(@NonNull Activity activity, @NonNull Lifecycle.Event event) {
if (activity instanceof LifecycleRegistryOwner) {
((LifecycleRegistryOwner) activity).getLifecycle().handleLifecycleEvent(event);
return;
}
//走这里
if (activity instanceof LifecycleOwner) {
Lifecycle lifecycle = ((LifecycleOwner) activity).getLifecycle();
if (lifecycle instanceof LifecycleRegistry) {
((LifecycleRegistry) lifecycle).handleLifecycleEvent(event);
}
}
}
//在API 29及以上，使用的生命周期回调
static class LifecycleCallbacks implements Application.ActivityLifecycleCallbacks {
static void registerIn(Activity activity) {
activity.registerActivityLifecycleCallbacks(new LifecycleCallbacks());
}
@Override
public void onActivityPostCreated(@NonNull Activity activity,
@Nullable Bundle savedInstanceState) {
dispatch(activity, Lifecycle.Event.ON_CREATE);
}
@Override
public void onActivityPostStarted(@NonNull Activity activity) {
dispatch(activity, Lifecycle.Event.ON_START);
}
@Override
public void onActivityPostResumed(@NonNull Activity activity) {
dispatch(activity, Lifecycle.Event.ON_RESUME);
}
@Override
public void onActivityPrePaused(@NonNull Activity activity) {
dispatch(activity, Lifecycle.Event.ON_PAUSE);
}
@Override
public void onActivityPreStopped(@NonNull Activity activity) {
dispatch(activity, Lifecycle.Event.ON_STOP);
}
@Override
public void onActivityPreDestroyed(@NonNull Activity activity) {
dispatch(activity, Lifecycle.Event.ON_DESTROY);
}
}
}在API 29及以上，使用LifecycleCallbacks来监听Activity的生命周期，由于dispatch(activity, Lifecycle.Event.ON_RESUME)在onActivityPostResumed()方法中调用，而dispatch(activity, Lifecycle.Event.ON_PAUSE)onActivityPrePaused()方法中调用，这就是前面的打印为什么一个在onResume()之后，一个在onPause()之前的原因。
一般我们会通过重写onCreateView()方法来给Fragment添加布局，注意这里ReportFragment中没有重写这个方法，所以ReportFragment是没有布局的透明Fragment。ReportFragment的作用就是获取生命周期而已，因为fragment生命周期是依附于Activity的。好处就是把这部分逻辑抽离出来，实现activity的无侵入。如果你对Glide比较熟，就会知道它也是使用透明Fragment获取生命周期的。
上面的代码中无论是使用LifecycleCallbacks还是使用ReportFragment最后都会在对应的生命周期方法中执行对应的dispatch(@NonNull Activity activity, @NonNull Lifecycle.Event event)方法，最后使用LifecycleRegistry的handleLifecycleEvent(event)方法来对对应的事件进行处理。这样就把生命周期事件的处理转移到了LifecycleRegistry中：
java
public class LifecycleRegistry extends Lifecycle {
//LifecycleObserver当前状态
private State mState;
//LifecycleOwner的弱引用
private final WeakReference<LifecycleOwner> mLifecycleOwner;
public LifecycleRegistry(@NonNull LifecycleOwner provider) {
this(provider, true);
}
private LifecycleRegistry(@NonNull LifecycleOwner provider, boolean enforceMainThread) {
mLifecycleOwner = new WeakReference<>(provider);
//默认初始化为INITIALIZED
mState = INITIALIZED;
mEnforceMainThread = enforceMainThread;
}
public void handleLifecycleEvent(@NonNull Lifecycle.Event event) {
State next = getStateAfter(event);//获取event发生之后的将要处于的状态
moveToState(next);//移动到这个状态
}
private void moveToState(State next) {
if (mState == next) {
return;
}
//更新状态
mState = next;
//正在sync()或正在addObserver()
if (mHandlingEvent || mAddingObserverCounter != 0) {
mNewEventOccurred = true;
// we will figure out what to do on upper level.
return;
}
mHandlingEvent = true;
sync();
mHandlingEvent = false;
}
//对应于官网那张图
static State getStateAfter(Event event) {
switch (event) {
case ON_CREATE:
case ON_STOP:
return CREATED;
case ON_START:
case ON_PAUSE:
return STARTED;
case ON_RESUME:
return RESUMED;
case ON_DESTROY:
return DESTROYED;
case ON_ANY:
break;
}
throw new IllegalArgumentException("Unexpected event value " + event);
}
}在moveToState()方法中将当前状态mState与下一个状态next比较，如果相等，直接return；如果不相等，进行sync()同步:
java
public class LifecycleRegistry extends Lifecycle {
//存储observers，遍历时可以删除或添加
//后添加的observer的state<=先添加的observer的state
private FastSafeIterableMap<LifecycleObserver, ObserverWithState> mObserverMap
= new FastSafeIterableMap<>();
// happens only on the top of stack (never in reentrance),
// so it doesn't have to take in account parents
private void sync() {
LifecycleOwner lifecycleOwner = mLifecycleOwner.get();
//LifecycleOwner已经被回收了
if (lifecycleOwner == null) {
throw new IllegalStateException("LifecycleOwner of this LifecycleRegistry is already"
+ "garbage collected. It is too late to change lifecycle state.");
}
while (!isSynced()) { //isSynced()表示所有观察者都同步完了
mNewEventOccurred = false;
//mObserverMap中存储的是观察者，状态按照从大到小排列，mState比最小的还小
if (mState.compareTo(mObserverMap.eldest().getValue().mState) < 0) {
backwardPass(lifecycleOwner);
}
Entry<LifecycleObserver, ObserverWithState> newest = mObserverMap.newest();
//mState比最大的还大
if (!mNewEventOccurred && newest != null
&& mState.compareTo(newest.getValue().mState) > 0) {
forwardPass(lifecycleOwner);
}
}
mNewEventOccurred = false;
}
private boolean isSynced() {
if (mObserverMap.size() == 0) {
return true;
}
//mObserverMap中最早添加的观察者的状态
State eldestObserverState = mObserverMap.eldest().getValue().mState;
//mObserverMap中最新添加的观察者的状态
State newestObserverState = mObserverMap.newest().getValue().mState;
return eldestObserverState == newestObserverState && mState == newestObserverState;
}
}这里出现了mObserverMap，mObserverMap是FastSafeIterableMap类型的，其实是把数据存储在其内部的成员mHashMap中，代码如下：
typescript
//简陋版的LinkedHashMap，支持遍历时的元素修改，比SafeIterableMap占用的内存多，非线程安全
public class FastSafeIterableMap<K, V> extends SafeIterableMap<K, V> {
private HashMap<K, Entry<K, V>> mHashMap = new HashMap<>();
@Override
protected Entry<K, V> get(K k) {
return mHashMap.get(k);
}
@Override
public V putIfAbsent(@NonNull K key, @NonNull V v) {
//从mHashMap中获取key对应的current(current类型是Entry<K,V>)，
//如果存在，返回current.mValue；
Entry<K, V> current = get(key);
if (current != null) {
return current.mValue;
}
//如果不存在，存入mHashMap
mHashMap.put(key, put(key, v));
return null;
}
@Override
public V remove(@NonNull K key) {
V removed = super.remove(key);
mHashMap.remove(key);
return removed;
}
public boolean contains(K key) {
return mHashMap.containsKey(key);
}
//获取mHashMap中k对应的节点在链表中的上一个节点
public Map.Entry<K, V> ceil(K k) {
if (contains(k)) {
return mHashMap.get(k).mPrevious;
}
return null;
}
}FastSafeIterableMap的eldest()、newest()方法来自父类SafeIterableMap，下面来看看SafeIterableMap：
typescript
//看起来是map,实际是LinkedList,支持遍历时修改元素，非线程安全
public class SafeIterableMap<K, V> implements Iterable<Map.Entry<K, V>> {
//最早添加的entry
Entry<K, V> mStart;
//最新添加的entry
private Entry<K, V> mEnd;
private int mSize = 0;
//返回最早添加的entry或null
public Map.Entry<K, V> eldest() {
return mStart;
}
//返回最新添加的entry或null
public Map.Entry<K, V> newest() {
return mEnd;
}
static class Entry<K, V> implements Map.Entry<K, V> {
@NonNull
final K mKey;
@NonNull
final V mValue;
//下一个节点
Entry<K, V> mNext;
//上一个节点
Entry<K, V> mPrevious;
}
}分析发现SafeIterableMap实际上是一个Entry<K, V>组成的双链表，FastSafeIterableMap是对SafeIterableMap的进一步封装，在Entry<K, V>的基础上添加了K，存储在HashMap<K, Entry<K, V>>中，这样方便快速通过key获取value，就不需要通过链表一个个遍历了。
mObserverMap是用来存储观察者的，mObserverMap添加观察者的地方就在getLifecycle().addObserver()中，代码如下：
java
public class LifecycleRegistry extends Lifecycle {
@Override
public void addObserver(@NonNull LifecycleObserver observer) {
//检查是否是主线程
enforceMainThreadIfNeeded("addObserver");
//初始状态
State initialState = mState == DESTROYED ? DESTROYED : INITIALIZED;
//将observer与initialState一起封装到ObserverWithState中，ObserverWithState是一个内部类
ObserverWithState statefulObserver = new ObserverWithState(observer, initialState);
//observer作为key，ObserverWithState作为value，存到mObserverMap
ObserverWithState previous = mObserverMap.putIfAbsent(observer, statefulObserver);
if (previous != null) {
return;//已经添加过，直接返回
}
LifecycleOwner lifecycleOwner = mLifecycleOwner.get();
if (lifecycleOwner == null) {
// it is null we should be destroyed. Fallback quickly
// lifecycleOwner被回收了，直接返回
return;
}
//isReentrance为true表示正在sync()或正在addObserver()
boolean isReentrance = mAddingObserverCounter != 0 || mHandlingEvent;
//获取目标状态
State targetState = calculateTargetState(observer);
mAddingObserverCounter++;
//关注1，这里后面再分析
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--;
}
static class ObserverWithState {
State mState;
LifecycleEventObserver mLifecycleObserver;
//构造函数
ObserverWithState(LifecycleObserver observer, State initialState) {
//初始化mLifecycleObserver
mLifecycleObserver = Lifecycling.lifecycleEventObserver(observer);
mState = initialState;
}
void dispatchEvent(LifecycleOwner owner, Event event) {
State newState = getStateAfter(event);
mState = min(mState, newState); //mState的作用是：新的事件触发后遍历通知所有观察者时，判断是否已经通知这个观察者了，即防止重复通知。
mLifecycleObserver.onStateChanged(owner, event);
mState = newState;
}
}
}用observer创建带状态的观察者ObserverWithState，observer作为key，ObserverWithState作为value，存入mObserverMap。
关注1我们后面再详细分析，这里的目的就是通过分发Event把当前添加的观察者的State一步一步同步到最新状态targetState。这就导致，比如你在Activity的onResume()方法中通过addObserver()添加观察者，但是onResume()之前的事件也会一一分发给观察者，即粘性。
回到最前面的sync()方法，mObserverMap.eldest().getValue()返回的是最早添加的到mObserverMap中的带状态的观察者ObserverWithState，mObserverMap.eldest().getValue().mState就是最早添加到mObserverMap中的观察者的状态。
这里为什么要使用while循环？
因为如果LifecycleOwner的当前状态为RESUMED，mObserverMap中观察者的状态为INITIALIZED，观察者的状态会从INITIALIZED->CREATED->STARTED->RESUMED一步一步转变，需要多次状态同步。
进入循环的条件是!isSynced()，若最早和最新添加的观察者的状态一致，且都是当前状态mState，说明已经同步完了。
没有同步完就进入循环体：
- mState比最早添加的观察者状态小，走backwardPass()方法，从新到老分发，循环使用downEvent()和observer.dispatchEvent()，连续分发事件；
- mState比最新添加的观察者状态大，走forwardPass()方法，从老到新分发，循环使用upEvent()和observer.dispatchEvent()，连续分发事件。
java
public class LifecycleRegistry extends Lifecycle {
private void forwardPass(LifecycleOwner lifecycleOwner) {
//正序遍历链表，从老到新
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);
//观察者分发事件
observer.dispatchEvent(lifecycleOwner, upEvent(observer.mState));
popParentState();
}
}
}
private void backwardPass(LifecycleOwner lifecycleOwner) {
//倒序遍历链表，从新到老
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));
//观察者分发事件
observer.dispatchEvent(lifecycleOwner, event);
popParentState();
}
}
}
//回退
private static Event downEvent(State state) {
switch (state) {
case INITIALIZED:
throw new IllegalArgumentException();
case CREATED:
return ON_DESTROY;
case STARTED:
return ON_STOP;
case RESUMED:
return ON_PAUSE;
case DESTROYED:
throw new IllegalArgumentException();
}
throw new IllegalArgumentException("Unexpected state value " + state);
}
//前进
private static Event upEvent(State state) {
switch (state) {
case INITIALIZED:
case DESTROYED:
return ON_CREATE;
case CREATED:
return ON_START;
case STARTED:
return ON_RESUME;
case RESUMED:
throw new IllegalArgumentException();
}
throw new IllegalArgumentException("Unexpected state value " + state);
}
}有没有发现上面代码12-15行与前面关注1处while循环里面的代码基本一致，也是同样的进行事件分发、状态同步。
observer.dispatchEvent()最终還是执行ObserverWithState的dispatchEvent()方法：
java
static class ObserverWithState {
State mState;
LifecycleEventObserver mLifecycleObserver;
//构造函数
ObserverWithState(LifecycleObserver observer, State initialState) {
//初始化mLifecycleObserver
mLifecycleObserver = Lifecycling.lifecycleEventObserver(observer);
mState = initialState;
}
void dispatchEvent(LifecycleOwner owner, Event event) {
State newState = getStateAfter(event);
//mState的作用是：新的事件触发后遍历通知所有观察者时，
//判断是否已经通知这个观察者了，即防止重复通知。
mState = min(mState, newState);
mLifecycleObserver.onStateChanged(owner, event);
mState = newState;
}
}ObserverWithState的dispatchEvent()方法又调用了mLifecycleObserver的onStateChanged()方法，在ObserverWithState的构造方法中通过Lifecycling.lifecycleEventObserver(observer)初始化了mLifecycleObserver：
java
public class Lifecycling {
@NonNull
static LifecycleEventObserver lifecycleEventObserver(Object object) {
...
return new ReflectiveGenericLifecycleObserver(object);
}
}其中，ReflectiveGenericLifecycleObserver代码如下：
java
class ReflectiveGenericLifecycleObserver implements LifecycleEventObserver {
private final Object mWrapped;
private final CallbackInfo mInfo;
ReflectiveGenericLifecycleObserver(Object wrapped) {
//mWrapped即为LifecycleObserver
mWrapped = wrapped;
//mInfo为LifecycleObserver对应的CallbackInfo
mInfo = ClassesInfoCache.sInstance.getInfo(mWrapped.getClass());
}
@Override
public void onStateChanged(LifecycleOwner source, Event event) {
mInfo.invokeCallbacks(source, event, mWrapped);
}
}mLifecycleObserver的onStateChanged()方法又会调用mInfo.invokeCallbacks(source, event, mWrapped)，其中mWrapped是我们添加的LifecycleObserver，而mInfo来自ClassesInfoCache：
ini
final class ClassesInfoCache {
//单例
static ClassesInfoCache sInstance = new ClassesInfoCache();
//使用mCallbackMap缓存观察者类对应的CallbackInfo
private final Map<Class<?>, CallbackInfo> mCallbackMap = new HashMap<>();
CallbackInfo getInfo(Class<?> klass) {
CallbackInfo existing = mCallbackMap.get(klass);
//如果存在，直接返回
if (existing != null) {
return existing;
}
//如果不存在，创建
existing = createInfo(klass, null);
return existing;
}
private CallbackInfo createInfo(Class<?> klass, @Nullable Method[] declaredMethods) {
Class<?> superclass = klass.getSuperclass();
//方法对应的Lifecycle.Event
Map<MethodReference, Lifecycle.Event> handlerToEvent = new HashMap<>();
if (superclass != null) {
CallbackInfo superInfo = getInfo(superclass);
if (superInfo != null) {
handlerToEvent.putAll(superInfo.mHandlerToEvent);
}
}
Class<?>[] interfaces = klass.getInterfaces();
for (Class<?> intrfc : interfaces) {
for (Map.Entry<MethodReference, Lifecycle.Event> entry : getInfo(
intrfc).mHandlerToEvent.entrySet()) {
verifyAndPutHandler(handlerToEvent, entry.getKey(), entry.getValue(), klass);
}
}
Method[] methods = declaredMethods != null ? declaredMethods : getDeclaredMethods(klass);
boolean hasLifecycleMethods = false;
//遍历实现了mLifecycleObserver接口的类中的方法
for (Method method : methods) {
//获取有@OnLifecycleEvent注解的方法
OnLifecycleEvent annotation = method.getAnnotation(OnLifecycleEvent.class);
if (annotation == null) {
continue;
}
hasLifecycleMethods = true;
Class<?>[] params = method.getParameterTypes();//获取方法参数
int callType = CALL_TYPE_NO_ARG;
//如果有参数
if (params.length > 0) {
callType = CALL_TYPE_PROVIDER;
//第1个参数必须是LifecycleOwner
if (!params[0].isAssignableFrom(LifecycleOwner.class)) {
throw new IllegalArgumentException(
"invalid parameter type. Must be one and instanceof LifecycleOwner");
}
}
Lifecycle.Event event = annotation.value();
//如果有2个参数
if (params.length > 1) {
callType = CALL_TYPE_PROVIDER_WITH_EVENT;
//第2个参数必须是Lifecycle.Event
if (!params[1].isAssignableFrom(Lifecycle.Event.class)) {
throw new IllegalArgumentException(
"invalid parameter type. second arg must be an event");
}
//有2个参数，注解值只能是ON_ANY
if (event != Lifecycle.Event.ON_ANY) {
throw new IllegalArgumentException(
"Second arg is supported only for ON_ANY value");
}
}
//如果参数超过2个
if (params.length > 2) {
throw new IllegalArgumentException("cannot have more than 2 params");
}
MethodReference methodReference = new MethodReference(callType, method);
//将methodReference对应的event存入handlerToEvent
verifyAndPutHandler(handlerToEvent, methodReference, event, klass);
}
CallbackInfo info = new CallbackInfo(handlerToEvent);
//将观察者类对应的CallbackInfo存入mCallbackMap中
mCallbackMap.put(klass, info);
mHasLifecycleMethods.put(klass, hasLifecycleMethods);
return info;
}
static class CallbackInfo {
//Lifecycle.Event对应的方法
final Map<Lifecycle.Event, List<MethodReference>> mEventToHandlers;
//方法对应的Lifecycle.Event
final Map<MethodReference, Lifecycle.Event> mHandlerToEvent;
CallbackInfo(Map<MethodReference, Lifecycle.Event> handlerToEvent) {
mHandlerToEvent = handlerToEvent;
mEventToHandlers = new HashMap<>();
//遍历handlerToEvent，将Lifecycle.Event对应的方法存入mEventToHandlers
for (Map.Entry<MethodReference, Lifecycle.Event> entry : handlerToEvent.entrySet()) {
Lifecycle.Event event = entry.getValue();
List<MethodReference> methodReferences = mEventToHandlers.get(event);
if (methodReferences == null) {
methodReferences = new ArrayList<>();
mEventToHandlers.put(event, methodReferences);
}
methodReferences.add(entry.getKey());
}
}
}
}因为反射是比较昂贵的，上面使用了ClassesInfoCache类来缓存实现了LifecycleObserver接口的类以及其中的Lifecycle.Event和对应的方法，主要使用了3个HashMap:
- mCallbackMap：缓存观察者类对应的CallbackInfo。
- mEventToHandlers：CallbackInfo的成员变量，缓存Lifecycle.Event对应的方法。
- mHandlerToEvent：CallbackInfo的成员变量，缓存方法对应的Lifecycle.Event。
接着回到ReflectiveGenericLifecycleObserver的onStateChanged()方法，onStateChanged()方法会调用CallbackInfo的invokeCallbacks()方法：
java
final class ClassesInfoCache {
static class CallbackInfo {
void invokeCallbacks(LifecycleOwner source, Lifecycle.Event event, Object target) {
invokeMethodsForEvent(mEventToHandlers.get(event), source, event, target);
invokeMethodsForEvent(mEventToHandlers.get(Lifecycle.Event.ON_ANY), source, event,
target);
}
private static void invokeMethodsForEvent(List<MethodReference> handlers,
LifecycleOwner source, Lifecycle.Event event, Object mWrapped) {
if (handlers != null) {
//遍历Lifecycle.Event对应的MethodReference
for (int i = handlers.size() - 1; i >= 0; i--) {
//执行MethodReference的invokeCallback()方法
handlers.get(i).invokeCallback(source, event, mWrapped);
}
}
}
}
}最终通过遍历该event对应的List<MethodReference>，然后一个个执行MethodReference的invokeCallback()方法：
arduino
final class ClassesInfoCache {
static final class MethodReference {
final int mCallType;
final Method mMethod;
void invokeCallback(LifecycleOwner source, Lifecycle.Event event, Object target) {
switch (mCallType) {
case CALL_TYPE_NO_ARG:
mMethod.invoke(target);//没有参数的
break;
case CALL_TYPE_PROVIDER:
mMethod.invoke(target, source);//一个参数的：LifecycleOwner
break;
case CALL_TYPE_PROVIDER_WITH_EVENT:
mMethod.invoke(target, source, event);//两个参数的：LifecycleOwner，Event
break;
}
}
}
}上面就是分发相应的事件，就会调用观察者中对应的方法的流程。
前面的calculateTargetState()方法再解释一下：
java
// 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<>();
private State calculateTargetState(LifecycleObserver observer) {
Entry<LifecycleObserver, ObserverWithState> previous = mObserverMap.ceil(observer);
State siblingState = previous != null ? previous.getValue().mState : null;
State parentState = !mParentStates.isEmpty() ? mParentStates.get(mParentStates.size() - 1)
: null;
return min(min(mState, siblingState), parentState);
}previous是新观察者的前一个entry，即原先的队尾，新观察者此时变成队尾。获取目标状态，previous即为当前状态的最小值。
假如没有mParentStates，好像还正常，然后注释给了一个反例：
- Observer1 在 onStart() 回调中把自己从集合中移除，然后添加了新的Observer2；
- 假如集合中只有Observer1这个观察者，移除后集合就是空的，会导致Observer2直接更新到LifecycleRegister的STARTED状态；
- 但，此时Observer1的 onStart() 回调还未执行完，而 Observer2 的 ON_START就回调执行完了，显然就违背了LifecycleRegistry的设计------观察者的同步是按照顺序执行的（这里没懂）；
添加了这个属性，在执行观察者回调前 pushParentState() 暂存当前观察者，回调完后 popParentState() 移除观察者，然后执行 calculateTargetState() 时判断是否为空，不为空取出最后一个缓存的观察者，然后取：LifecycleRegister当前状态、previous当前状态、缓存观察者状态中的最小值，作为当前观察者的状态。
感谢与参考