一、我们想象中的生命周期感应应该是怎样的?
1、设想
监听过程如果让我们开始自己去实现一个简单的生命周期感应框架,刚开始很可能是这样的。
自己实现一个Data<T>,里面有一个集合,保存所有的Observer。同时BaseActivity包含一个Data的集合,当生命周期变化的时候去动态注销UI和Data的订阅关系。
2、google的实现
google把生命周期处理逻辑抽离到一个单独的LifecycleRegistry中,并将数据的逻辑放在LiveData中,职责更加清晰明了。
在多线程处理,LiveData状态的保持,以及事件的变化传递控制更加细致具体。
二、Observer监听过程以及生命周期感应
1、监听过程
(1·1)LiveData.observe()
- 将Observer对象添加到LiveData中(后续Data变化,声明周期变化通知Observer)
//LiveData::observe()
public void observe(@NonNull LifecycleOwner owner, @NonNull Observer<T> observer) {
if (owner.getLifecycle().getCurrentState() == DESTROYED) {
// ignore
return;
}
//(装饰者模式)封装成一个能一个能接收并处理生命周期事件的LifecycleBoundObserver类,
//LifecycleBoundObserver会当FragmentActivity/Fragment的onDestroy()调用之后将所有的监听关系移除掉
//LifecycleBoundObserver wrapper = new LifecycleBoundObserver(owner, observer);
//将observer作为key添加到Map中(注意:同一个liveData能被多个observer监听)
//注意防止重复监听导致逻辑可能出错,例如重复执行网络请求的时候很容易多次发生多次监听
ObserverWrapper existing = mObservers.putIfAbsent(observer, wrapper);
//同一个Observer对象不能绑定多个不同拥有生命周期的对象
if (existing != null && !existing.isAttachedTo(owner)) {
throw new IllegalArgumentException("Cannot add the same observer"
+ " with different lifecycles");
}
if (existing != null) {
return;
}
//参见2
owner.getLifecycle().addObserver(wrapper);
}
LifecycleBoundObserver类
class LifecycleBoundObserver extends ObserverWrapper implements GenericLifecycleObserver {
@NonNull final LifecycleOwner mOwner;
LifecycleBoundObserver(@NonNull LifecycleOwner owner, Observer<T> observer) {
super(observer);
mOwner = owner;
}
@Override
boolean shouldBeActive() {
return mOwner.getLifecycle().getCurrentState().isAtLeast(STARTED);
}
@Override
public void onStateChanged(LifecycleOwner source, Lifecycle.Event event) {
if (mOwner.getLifecycle().getCurrentState() == DESTROYED) {
removeObserver(mObserver);
return;
}
activeStateChanged(shouldBeActive());
}
@Override
boolean isAttachedTo(LifecycleOwner owner) {
return mOwner == owner;
}
@Override
void detachObserver() {
mOwner.getLifecycle().removeObserver(this);
}
}
(1·2)owner.getLifecycle().addObserver(wrapper);
- 将Observer对象添加到Lifecycle中(Lifecycle能接收FragmentActivity/Fragment生命周期变化)
@Override
public void addObserver(@NonNull LifecycleObserver observer) {
State initialState = mState == DESTROYED ? DESTROYED : INITIALIZED;
ObserverWithState statefulObserver = new ObserverWithState(observer, initialState);
//将Observer添加到mObserverMap中,此处被封装成一个ObserverWithState类(具有状态属性)
ObserverWithState previous = mObserverMap.putIfAbsent(observer, statefulObserver);
if (previous != null) {
return;
}
//省略observer与LifecycleRegistry状态同步逻辑(后面有分析).....
}
ObserverWithState类
//
//ObserverWithState::dispatchEvent()
//接收生命周期事件,并进行状态的变更,同时将事件传递给GenericLifecycleObserver处理
//(其实也就是上面第一步的LifecycleBoundObserver),方便进行监听关系的注销
void dispatchEvent(LifecycleOwner owner, Event event) {
//计算接收Event事件之后应该处于哪种状态
State newState = getStateAfter(event);
mState = min(mState, newState);//此句多余
mLifecycleObserver.onStateChanged(owner, event);
//更新并记录新状态
mState = newState;
}
2、生命周期的感应
(2·1)生命周期的捕获
记得在步骤(1·1)中需要传入一个实现了LifecycleOwner接口的对象。而我们一般传入的是FragmentActivity/Fragment,此时我们直接点开FragmentActivity源码之后一层层剥掉其外壳,将会发现它的父类SupportActivity其实实现了LifecycleOwner接口,并且会通过接口的getLifecycle()方法返回一个Lifecycle的子类(也就是LifecycleRegstry)。
SupportActivity这个类代码并不多,大致看一下就能明白其意思
重点: 你会发现这行代码
protected void onCreate(@Nullable Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
//在FragmentActivity上面添加一个无界面的Fragment(利用Fragment生命周期和Activity/Fragment绑定在一起的特点,
//实现hook到Activity/Fragment的生命周期事件,将其整合到一起处理,达到低耦合高内聚的目的)。有兴趣的可以去分
//析glide的动态取消加载图片请求逻辑,甚至可以自己开发一个动态取消网络请求的网络库开源项目。
ReportFragment.injectIfNeededIn(this);
}
在ReportFragment的各个生命周期回调方法中会调用dispatch(Lifecycle.Event event)方法,将生命周期事件传递给前面提到过的Lifecycle的子类(也就是LifecycleRegstry)。
private void dispatch(Lifecycle.Event event) {
Activity activity = getActivity();
//LifecycleRegistryOwner目前已标记弃用,可不管
if (activity instanceof LifecycleRegistryOwner) {
((LifecycleRegistryOwner) activity).getLifecycle().handleLifecycleEvent(event);
return;
}
if (activity instanceof LifecycleOwner) {
Lifecycle lifecycle = ((LifecycleOwner) activity).getLifecycle();
if (lifecycle instanceof LifecycleRegistry) {
//将事件传递给LifecycleRegstry对象
((LifecycleRegistry) lifecycle).handleLifecycleEvent(event);
}
}
}
分析到这里我们大致总结一下:(同时可以加入一些猜想和自己的理解)
现在FragmentActivity/Fragment的生命周期事件会通过ReportFragment传递给LifecycleRegistry。
然后我们大胆猜想下,在LifecycleRegistry在处理事件的时候是否当收到on_destroy事件的时候就通过前面步骤(1·2)中的ObserverWithState类通知Observer,然后进行Observer的监听注销呢?
带着疑问和猜想,我们继续去分析LifecycleRegistry中的逻辑。
(2·2)LifecycleRegistry生命周期事件的处理以及状态变化
在讲这部分逻辑之前,需要将里面涉及到的几个重要的状态变化方法逻辑图理一遍。
其实这部分状态机的变化图你也可以不用花这么大精力去理解,因为我们真正关心的只有on_destroy事件,只有收到它我们才会去注销监听关系。所以本部分逻辑你可以不需要深入理解。
public void handleLifecycleEvent(@NonNull Lifecycle.Event event) {
//接收生命周期事件,计算得出应该处于什么状态,参照逻辑图理解
State next = getStateAfter(event);
moveToState(next);
}
private void moveToState(State next) {
if (mState == next) {
return;
}
//保存接收事件后的状态
mState = next;
if (mHandlingEvent || mAddingObserverCounter != 0) {
mNewEventOccurred = true;
// we will figure out what to do on upper level.
return;
}
//根据当前的状态mState,去更新Map中Observer的状态
//mHandlingEvent在addObserver(@NonNull LifecycleObserver observer)方法会用到
//用来判断协同是否有正在进行相同的sync逻辑,避免重复执行
mHandlingEvent = true;
//根据当前的状态mState,去更新Map中Observer的状态
sync();
mHandlingEvent = false;
}
moveToState() -----> sync()
//同步所有的Observer与LifecycleResitry的状态(也就是Activity/Fragment的状态)
private void sync() {
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;
}
//isSynced()会判断当前map中头尾的Observer状态与LifecycleRegistry中mState一致
//不一致表示在sync同步状态过程中有新的Observer加入map中了
while (!isSynced()) {
mNewEventOccurred = false;
//Activity/Fragment状态从被销毁或启动其它界面时,与其绑定的Observer状态的同步
if (mState.compareTo(mObserverMap.eldest().getValue().mState) < 0) {
backwardPass(lifecycleOwner);
}
//Activity/Fragment启动时,与其绑定的Observer状态的同步
Entry<LifecycleObserver, ObserverWithState> newest = mObserverMap.newest();
if (!mNewEventOccurred && newest != null
&& mState.compareTo(newest.getValue().mState) > 0) {
forwardPass(lifecycleOwner);
}
}
mNewEventOccurred = false;
}
moveToState() -----> backwardPass()
或者
moveToState() -----> forwardPass()
//Activity/Fragment启动时,与其绑定的Observer状态的同步
//注意: 多个Observer监听同一个LiveData,并且Observer 中有相同的逻辑处理,
//最终的处理结果以最后添加的Observer中的逻辑为准。
private void forwardPass(LifecycleOwner lifecycleOwner) {
//mObserverMap从Start到End进行遍历(最早添加-->最近添加,越早越靠前)。
//这里的ascendingIterator就是上面现象的原因,具体翻阅源码分析下
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);
//while state in observer.mState到mState,逐次计算出触发状态变化的事件,传递给
//ObserverWithState处理
observer.dispatchEvent(lifecycleOwner, upEvent(observer.mState));
popParentState();
}
}
}
//Activity/Fragment状态从被销毁或启动其它界面时,与其绑定的Observer状态的同步
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));
//将生命周期shijian传递给ObserverWithState
observer.dispatchEvent(lifecycleOwner, event);
popParentState();
}
}
}
ObserverWithState
ObserverWithState是LifecycleRegistry的内部类
static class ObserverWithState {
State mState;
GenericLifecycleObserver mLifecycleObserver;
ObserverWithState(LifecycleObserver observer, State initialState) {
mLifecycleObserver = Lifecycling.getCallback(observer);
mState = initialState;
}
void dispatchEvent(LifecycleOwner owner, Event event) {
State newState = getStateAfter(event);
mState = min(mState, newState);
//改变ObserverWithState状态,同时将生命周期Event传递给LifecycleBoundObserver
mLifecycleObserver.onStateChanged(owner, event);
mState = newState;
}
}
LifecycleBoundObserver
LifecycleBoundObserver是LiveData的内部类
class LifecycleBoundObserver extends ObserverWrapper implements GenericLifecycleObserver {
@NonNull final LifecycleOwner mOwner;
LifecycleBoundObserver(@NonNull LifecycleOwner owner, Observer<T> observer) {
super(observer);
mOwner = owner;
}
//状态为Started或Resumed才处于激活状态
@Override
boolean shouldBeActive() {
return mOwner.getLifecycle().getCurrentState().isAtLeast(STARTED);
}
@Override
public void onStateChanged(LifecycleOwner source, Lifecycle.Event event) {
//注意:如果Observer绑定的Activity/Fragment处于DESTROYED(onDestroy方法调用之后)状态,就将所有的观察订阅关系注销
if (mOwner.getLifecycle().getCurrentState() == DESTROYED) {
removeObserver(mObserver);
return;
}
activeStateChanged(shouldBeActive());
}
@Override
boolean isAttachedTo(LifecycleOwner owner) {
return mOwner == owner;
}
@Override
void detachObserver() {
mOwner.getLifecycle().removeObserver(this);
}
}
LiveData:
@MainThread
public void removeObserver(@NonNull final Observer<T> observer) {
assertMainThread("removeObserver");
//移除LiveData中的Observer
ObserverWrapper removed = mObservers.remove(observer);
if (removed == null) {
return;
}
//调用LifecycleBoundObserver的detachObserver()方法
//也就是移除LifecycleRegistry中的Observer引用
removed.detachObserver();
//更改observer的状态
removed.activeStateChanged(false);
}
(2·3)关键方法状态变化逻辑图
回顾下Activity的生命周期:
LifecycleRegistry的状态转换图:
LifecycleRegistry的状态转换关键控制方法:
3、Observer状态与Registry的同步
LifecycleRegistry的方法sync()只在两个地方调用到了:
- LifecycleRegistry::addObserver(@NonNull LifecycleObserver observer)
- LifecycleRegistry::moveToState(State next)
@Override
public void addObserver(@NonNull LifecycleObserver observer) {
......
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--;
}
private void moveToState(State next) {
......
//根据当前的状态mState,去更新Map中Observer的状态
//mHandlingEvent在addObserver(@NonNull LifecycleObserver observer)方法会用到
//用来判断协同是否有正在进行相同的sync逻辑,避免重复执行
mHandlingEvent = true;
//根据当前的状态mState,去更新Map中Observer的状态
sync();
mHandlingEvent = false;
}
因为涉及到多线程操作(如addObserver可能在其他线程执行)。这里的总体逻辑就是如果当前只要有一个地方在执行sync操作,另一个都无须再次执行sync方法进行同步。
这和sync方法的逻辑有关
while(!isSynced()){
...
}
isSynced方法就是判断当前LifecycleRegistry状态mState是否和Map中的首位的Observer状态是否一致(因为每次添加新的Observer都会触发sync,所以每次只要判断首尾Observer就行)。
意思就是只要首尾不一致,就肯定不同步。
首尾一致,因为同步是遍历的,那么肯定同步。
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;
}
4、ViewModel的生命周期管理
这部分相对简单很多。逻辑主要在FragmentActivity的以下三个方法中
- onCreate()
- onRetainNonConfigurationInstance()
- onDestroy()
protected void onCreate(@Nullable Bundle savedInstanceState) {
...
// 恢复mViewModelStore
FragmentActivity.NonConfigurationInstances nc = (FragmentActivity.NonConfigurationInstances)this.getLastNonConfigurationInstance();
if (nc != null && nc.viewModelStore != null && this.mViewModelStore == null) {
this.mViewModelStore = nc.viewModelStore;
}
...
}
protected void onDestroy() {
super.onDestroy();
//Configurations发生变化而引起的onDestroy不会销毁ViewModel
if (this.mViewModelStore != null && !this.isChangingConfigurations()) {
this.mViewModelStore.clear();
}
this.mFragments.dispatchDestroy();
}
//mViewModelStore的持久化
public final Object onRetainNonConfigurationInstance() {
Object custom = this.onRetainCustomNonConfigurationInstance();
FragmentManagerNonConfig fragments = this.mFragments.retainNestedNonConfig();
if (fragments == null && this.mViewModelStore == null && custom == null) {
return null;
} else {
FragmentActivity.NonConfigurationInstances nci = new FragmentActivity.NonConfigurationInstances();
nci.custom = custom;
nci.viewModelStore = this.mViewModelStore;
nci.fragments = fragments;
return nci;
}
}
5、总结
- 不同Observe监听相同LiveData,生命周期没控制好,导致崩溃
- 多个不同Observer监听同一个LiveData,并且Observer 中有相同的逻辑处理,最终的处理结果以最后添加的Observer中的逻辑为准。
- 默认的Observer的只会在Activity或Fragment处于onStarted/onResumed状态下才会接收到LiveData的数据;或者重新处于onStarted/onResumed时。
6、思考
- 如何在不同的Activity之间共享ViewModel?
答:自己去实现一个ViewModelStore,自己控制它的生命周期。 - ViewModel是否帮我们做了数据持久化?
答:ViewModel没做数据的持久化,只是内存变量的复用。 - 怎样让observer处于onCreated状态也能处理数据?
答:目前并没有提供对应API去自定义Observer的逻辑触发时机。