alighters

程序、写作、人生

EventBus后续深入知识点整理

| Comments

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 :

FindState的妙用

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:

  • 通过findUsingInfo(Class<?> subscriberClass)在apt中进行查找获取
  • 使用'findUsingReflection(Class<?> subscriberClass)‘方法,进行反射来获取 而在这里,EventBus采用了一个中间器FindState,来看一下它的结构:
1
2
3
4
5
6
7
8
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subscriberMethodssubscriberInfo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?

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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。

缓存的使用

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:

1
2
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:

1
2
3
4
5
6
7
8
9
10
11
12
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:

1
2
3
4
5
6
7
8
9
10
11
12
13
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:

1
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。

反射类方法的使用

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:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
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MethodClass,通过ClassgetDeclaredMethodsgetMethods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。

线程处理类信息的使用

在EventBus类中,定义了4种线程处理的策略:

1
2
3
4
5
6
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, BackgroundPosterAsyncPoster

BackgroundPoster - 后台任务执行

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
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。

AsyncPoster - 异步任务执行

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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。

事件执行队列 PendingPostQueue

EventBus对事件的执行,采用队列的数据结构:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
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。。

总结

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。

版权归作者所有,转载请注明原文链接:/blog/2016/05/24/eventbus-3-dot-0-indepth-knowledge/

给 Ta 个打赏吧...

Comments