根据上一篇文章 浅析EventBus 3.0实现思想 对EventBus的概括,本文针对其中一些重要且比较有意思的知识点,做一下如下的汇总整理 :
在EventBus中,会根据class信息,来获取 SubscriberMethod
,这里会在 SubscriberMethodFinder
中进行处理,提供了两种方式来进行获取:
findUsingInfo(Class<?> subscriberClass)
在apt中进行查找获取 FindState
,来看一下它的结构: static class FindState { final List<SubscriberMethod> subscriberMethods = new ArrayList<>(); Class<?> subscriberClass; Class<?> clazz; boolean skipSuperClasses; SubscriberInfo subscriberInfo; }
这里对查找的状态值做了一些封装,其中有订阅类 subscriberClass
,事件对象 clazz
,以及查找的结果 subscriberMethods
、 subscriberInfo
等,另外,还有一个判断的标志量 skipSuperClasses
,用来标记是否需要进行父类的查看查找。
同时,我们可以看出在使用EventBus定义订阅方法的时候,有些通用的逻辑,是可以抽象放置在父类中的。
为什么要使用FindState呢?首先是面向对象封装的采用,那么看看它给我们提供了哪些方法?
void initForSubscriber(Class<?> subscriberClass) { ... } boolean checkAdd(Method method, Class<?> eventType) { ... } private boolean checkAddWithMethodSignature(Method method, Class<?> eventType) { ... } void moveToSuperclass() { ... }
方法中的 initForSubscriber
是用来初始化传入订阅类的,两个check方法则是用来检查方法信息的,这样用来保证获取的订阅方法都是合法的。 moveToSuperClass
则是需要查看父类中的订阅方法。这样对方法检查的逻辑,我们就把它们抽象在了FindState中。
使用java的,应该要知道频繁地创建对象,是非常消耗资源的,在jvm垃圾回收时候,会出现内存抖动的问题。所以,我们在这里,一定要注意缓存的使用。
上文中提到的中间器FindState,就采用了缓存:
private static final int POOL_SIZE = 4; private static final FindState[] FIND_STATE_POOL = new FindState[POOL_SIZE];
指定了FindState的缓存大小为4,并使用一维的静态数组,所以这里需要注意线程同步的问题:
private FindState prepareFindState() { synchronized (FIND_STATE_POOL) { for (int i = 0; i < POOL_SIZE; i++) { FindState state = FIND_STATE_POOL[i]; if (state != null) { FIND_STATE_POOL[i] = null; return state; } } } return new FindState(); }
这段是用来获取FindState, 可以看到的是对这段缓存的获取使用了 synchronized
关键字,来将缓存中FindState的获取,变为同步块。 而在subscriberMethod的获取的同时,则对FindState的缓存做了添加的操作,同样是也必须是同步代码块:
private List<SubscriberMethod> getMethodsAndRelease(FindState findState) { List<SubscriberMethod> subscriberMethods = new ArrayList<>(findState.subscriberMethods); findState.recycle(); synchronized (FIND_STATE_POOL) { for (int i = 0; i < POOL_SIZE; i++) { if (FIND_STATE_POOL[i] == null) { FIND_STATE_POOL[i] = findState; break; } } } return subscriberMethods; }
另外,EventBus也对subsciberMethod的获取,也做了缓存的操作,这样进行SubscriberMethod查找的时候,则优先进行缓存的查找:
private static final Map<Class<?>, List<SubscriberMethod>> METHOD_CACHE = new ConcurrentHashMap<>();
这里,使用的是数据结构是 ConcurrentHashMap
,就可以不必写大量的同步代码块了。
反射虽然是比较浪费性能的,但对我们Java开发者来说,这又是必须掌握的一个技能,现在来熟悉一下EventBus中通过 @Subscribe
注解对 SubscriberMethod
的查找:
private void findUsingReflectionInSingleClass(FindState findState) { Method[] methods; // 优先使用getDeclareMethods方法,如注释中所说,比getMethods方法块。 try { // This is faster than getMethods, especially when subscribers are fat classes like Activities methods = findState.clazz.getDeclaredMethods(); } catch (Throwable th) { // Workaround for java.lang.NoClassDefFoundError, see https://github.com/greenrobot/EventBus/issues/149 methods = findState.clazz.getMethods(); findState.skipSuperClasses = true; } for (Method method : methods) { int modifiers = method.getModifiers(); // 通过访问符只获取public if ((modifiers & Modifier.PUBLIC) != 0 && (modifiers & MODIFIERS_IGNORE) == 0) { Class<?>[] parameterTypes = method.getParameterTypes(); // 方法的参数(事件类型)长度只能为1 if (parameterTypes.length == 1) { Subscribe subscribeAnnotation = method.getAnnotation(Subscribe.class); if (subscribeAnnotation != null) { Class<?> eventType = parameterTypes[0]; // 获取到annotation中的内容,进行subscriberMethod的添加 if (findState.checkAdd(method, eventType)) { ThreadMode threadMode = subscribeAnnotation.threadMode(); findState.subscriberMethods.add(new SubscriberMethod(method, eventType, threadMode, subscribeAnnotation.priority(), subscribeAnnotation.sticky())); } } } else if (strictMethodVerification && method.isAnnotationPresent(Subscribe.class)) { //抛出方法参数只能为1的异常 String methodName = method.getDeclaringClass().getName() + "." + method.getName(); throw new EventBusException("@Subscribe method " + methodName + "must have exactly 1 parameter but has " + parameterTypes.length); } } else if (strictMethodVerification && method.isAnnotationPresent(Subscribe.class)) { //抛出方法访问符只能为public的异常 String methodName = method.getDeclaringClass().getName() + "." + method.getName(); throw new EventBusException(methodName + " is a illegal @Subscribe method: must be public, non-static, and non-abstract"); } } }
其中,最核心的类便是 Method
和 Class
,通过 Class
的 getDeclaredMethods
及 getMethods
来进行方法信息的获取;使用 Method
类的 getParameterTypes
获取方法的参数及 getAnnotation
获取方法的注解类。
在EventBus类中,定义了4种线程处理的策略:
public enum ThreadMode { POSTING, MAIN, BACKGROUND, ASYNC }
POSTING
采用与事件发布者相同的线程, MAIN
指定为主线程, BACKGROUND
指定为后台线程,而 ASYNC
相比前三者不同的地方是可以处理耗时的操作,其采用了线程池,且是一个异步执行的过程,即事件的订阅者可以立即得到执行。
这里,我们主要看两个Poster, BackgroundPoster
和 AsyncPoster
:
final class BackgroundPoster implements Runnable { private final PendingPostQueue queue; private final EventBus eventBus; private volatile boolean executorRunning; BackgroundPoster(EventBus eventBus) { this.eventBus = eventBus; queue = new PendingPostQueue(); } public void enqueue(Subscription subscription, Object event) { PendingPost pendingPost = PendingPost.obtainPendingPost(subscription, event); synchronized (this) { queue.enqueue(pendingPost); if (!executorRunning) { executorRunning = true; eventBus.getExecutorService().execute(this); } } } @Override public void run() { try { try { while (true) { PendingPost pendingPost = queue.poll(1000); if (pendingPost == null) { synchronized (this) { // Check again, this time in synchronized pendingPost = queue.poll(); if (pendingPost == null) { executorRunning = false; return; } } } eventBus.invokeSubscriber(pendingPost); } } catch (InterruptedException e) { Log.w("Event", Thread.currentThread().getName() + " was interruppted", e); } } finally { executorRunning = false; } } }
代码中,主要通过 enqueue
方法,将当前的订阅者添加至队列 PendingPostQueue
中,是否立即执行,则需要判断当前队列是否还有正在执行的任务,若没有的话,则立即执行,若还有执行任务的话,则只进行队列的添加。这样,保证了后台任务永远只会在一个线程执行。
class AsyncPoster implements Runnable { private final PendingPostQueue queue; private final EventBus eventBus; AsyncPoster(EventBus eventBus) { this.eventBus = eventBus; queue = new PendingPostQueue(); } public void enqueue(Subscription subscription, Object event) { PendingPost pendingPost = PendingPost.obtainPendingPost(subscription, event); queue.enqueue(pendingPost); eventBus.getExecutorService().execute(this); } @Override public void run() { PendingPost pendingPost = queue.poll(); if(pendingPost == null) { throw new IllegalStateException("No pending post available"); } eventBus.invokeSubscriber(pendingPost); } }
这段代码就很简单了,直接通过线程池调用执行,相比 BackgroundPoster
执行来说,则没有等待的过程。
EventBus对事件的执行,采用队列的数据结构:
final class PendingPostQueue { private PendingPost head; private PendingPost tail; synchronized void enqueue(PendingPost pendingPost) { if (pendingPost == null) { throw new NullPointerException("null cannot be enqueued"); } if (tail != null) { tail.next = pendingPost; tail = pendingPost; } else if (head == null) { head = tail = pendingPost; } else { throw new IllegalStateException("Head present, but no tail"); } notifyAll(); } synchronized PendingPost poll() { PendingPost pendingPost = head; if (head != null) { head = head.next; if (head == null) { tail = null; } } return pendingPost; } synchronized PendingPost poll(int maxMillisToWait) throws InterruptedException { if (head == null) { wait(maxMillisToWait); } return poll(); } }
而对 PendingPost
的封装,使用了数据缓存池:
final class PendingPost { private final static List<PendingPost> pendingPostPool = new ArrayList<PendingPost>(); Object event; Subscription subscription; PendingPost next; // 对PendingPost的获取,优先从缓存池中拿 private PendingPost(Object event, Subscription subscription) { this.event = event; this.subscription = subscription; } static PendingPost obtainPendingPost(Subscription subscription, Object event) { synchronized (pendingPostPool) { int size = pendingPostPool.size(); if (size > 0) { PendingPost pendingPost = pendingPostPool.remove(size - 1); pendingPost.event = event; pendingPost.subscription = subscription; pendingPost.next = null; return pendingPost; } } return new PendingPost(event, subscription); } // 对PendingPost释放时,将其添加到缓存池中 static void releasePendingPost(PendingPost pendingPost) { pendingPost.event = null; pendingPost.subscription = null; pendingPost.next = null; synchronized (pendingPostPool) { // Don't let the pool grow indefinitely if (pendingPostPool.size() < 10000) { pendingPostPool.add(pendingPost); } } } }
可以看到其对缓存的大小限制到10000,好任性啊。。
EventBus给我们提供了相当强大的功能,同时它的写法也相当有味道,值得我们深深地去研究。总的来说,其中EventBus采用了Facade模式,方便开发者的统一调用;另外不同的线程策略,以及反射代码,Apt处理代码生成以及缓存的大量使用。