RxJava Observer与Subscriber的关系

栏目：综合技术时间：2016-06-21 08:56:32

在说Observer与Subscriber的关系之前，我们下重温下相干概念。

RxJava 的视察者模式

RxJava 有4个基本概念：Observable (可视察者，即被视察者)、 Observer (视察者)、 subscribe (定阅)、事件。Observable 和 Observer 通过 subscribe() 方法实现定阅关系，从而 Observable 可以在需要的时候发失事件来通知 Observer。

与传统视察者模式不同， RxJava 的事件回调方法除普通事件 onNext() （相当于 onClick() / onEvent()）以外，还定义了两个特殊的事件：onCompleted() 和 onError()。

onCompleted(): 事件队列完结。RxJava 不但把每一个事件单独处理，还会把它们看作1个队列。RxJava 规定，当不会再有新的 onNext() 发出时，需要触发 onCompleted() 方法作为标志。
onError(): 事件队列异常。在事件处理进程中出异常时，onError() 会被触发，同时队列自动终止，不允许再有事件发出。
在1个正确运行的事件序列中, onCompleted() 和 onError() 有且只有1个，并且是事件序列中的最后1个。需要注意的是，onCompleted() 和 onError() 2者也是互斥的，即在队列中调用了其中1个，就不应当再调用另外一个。

RxJava 的视察者模式大致以下图：
这里写图片描述

RxJava的实现

基于以上的概念， RxJava 的基本实现主要有3点：

1) 创建 Observer

Observer 即视察者，它决定事件触发的时候将有怎样的行动。 RxJava 中的 Observer 接口的实现方式：

Observer<Apps> observer = new Observer<Apps>() {
        @Override
        public void onCompleted() {
            listView.onRefreshComplete();
        }

        @Override
        public void onError(Throwable e) {
            listView.onRefreshComplete();
        }

        @Override
        public void onNext(Apps appsList) {
            listView.onRefreshComplete();
            appLists.addAll(appsList.apps);
            adapter.notifyDataSetChanged();
        }
    };

除 Observer 接口以外，RxJava 还内置了1个实现了 Observer 的抽象类：Subscriber。 Subscriber 对 Observer 接口进行了1些扩大，但他们的基本使用方式是完全1样的：

Subscriber subscriber = new Subscriber<Apps>() {
        @Override
        public void onCompleted() {
            listView.onRefreshComplete();
        }

        @Override
        public void onError(Throwable e) {
            listView.onRefreshComplete();
        }

        @Override
        public void onNext(Apps appsList) {
            listView.onRefreshComplete();
            appLists.addAll(appsList.apps);
            adapter.notifyDataSetChanged();
        }
    };

不但基本使用方式1样，实质上，在 RxJava 的 subscribe 进程中，Observer 也总是会先被转换成1个 Subscriber 再使用。所以如果你只想使用基本功能，选择 Observer 和 Subscriber 是完全1样的。

Subscriber是Observer的实现类

public abstract class Subscriber<T> implements Observer<T>, Subscription

而onStart()方法是Subscriber中的1个方法。它也属于回调级别的。

subscribe(Subscriber)方法中有以下代码：

// if not already wrapped   包裹1层
        if (!(subscriber instanceof SafeSubscriber)) {
            // assign to `observer` so we return the protected version
            subscriber = new SafeSubscriber<T>(subscriber);
        }

他将subscriber包装起来，这个具体甚么意思有待研究，继续下看。

hook.onSubscribeStart(observable, observable.onSubscribe).call(subscriber);
   return hook.onSubscribeReturn(subscriber);

hook是甚么呢？

private static final RxJavaObservableExecutionHook hook = RxJavaPlugins.getInstance().getObservableExecutionHook();

RxJavaObservableExecutionHook.java源码：

/**
 * Copyright 2014 Netflix, Inc.
 * 
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 * 
 * http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package rx.plugins;

import rx.Observable;
import rx.Observable.OnSubscribe;
import rx.Observable.Operator;
import rx.Subscriber;
import rx.Subscription;
import rx.functions.Func1;

/**
 * Abstract ExecutionHook with invocations at different lifecycle points of {@link Observable} execution with a
 * default no-op implementation.
 * <p>
 * See {@link RxJavaPlugins} or the RxJava GitHub Wiki for information on configuring plugins:
 * <a href="https://github.com/ReactiveX/RxJava/wiki/Plugins">https://github.com/ReactiveX/RxJava/wiki/Plugins</a>.
 * <p>
 * <b>Note on thread-safety and performance:</b>
 * <p>
 * A single implementation of this class will be used globally so methods on this class will be invoked
 * concurrently from multiple threads so all functionality must be thread-safe.
 * <p>
 * Methods are also invoked synchronously and will add to execution time of the observable so all behavior
 * should be fast. If anything time-consuming is to be done it should be spawned asynchronously onto separate
 * worker threads.
 * 
 */
public abstract class RxJavaObservableExecutionHook {
    /**
     * Invoked during the construction by {@link Observable#create(OnSubscribe)}
     * <p>
     * This can be used to decorate or replace the <code>onSubscribe</code> function or just perform extra
     * logging, metrics and other such things and pass-thru the function.
     * 
     * @param f
     *            original {@link OnSubscribe}<{@code T}> to be executed
     * @return {@link OnSubscribe}<{@code T}> function that can be modified, decorated, replaced or just
     *         returned as a pass-thru
     */
    public <T> OnSubscribe<T> onCreate(OnSubscribe<T> f) {
        return f;
    }

    /**
     * Invoked before {@link Observable#subscribe(rx.Subscriber)} is about to be executed.
     * <p>
     * This can be used to decorate or replace the <code>onSubscribe</code> function or just perform extra
     * logging, metrics and other such things and pass-thru the function.
     * 
     * @param onSubscribe
     *            original {@link OnSubscribe}<{@code T}> to be executed
     * @return {@link OnSubscribe}<{@code T}> function that can be modified, decorated, replaced or just
     *         returned as a pass-thru
     */
    public <T> OnSubscribe<T> onSubscribeStart(Observable<? extends T> observableInstance, final OnSubscribe<T> onSubscribe) {
        // pass-thru by default
        return onSubscribe;
    }

    /**
     * Invoked after successful execution of {@link Observable#subscribe(rx.Subscriber)} with returned
     * {@link Subscription}.
     * <p>
     * This can be used to decorate or replace the {@link Subscription} instance or just perform extra logging,
     * metrics and other such things and pass-thru the subscription.
     * 
     * @param subscription
     *            original {@link Subscription}
     * @return {@link Subscription} subscription that can be modified, decorated, replaced or just returned as a
     *         pass-thru
     */
    public <T> Subscription onSubscribeReturn(Subscription subscription) {
        // pass-thru by default
        return subscription;
    }

    /**
     * Invoked after failed execution of {@link Observable#subscribe(Subscriber)} with thrown Throwable.
     * <p>
     * This is <em>not</em> errors emitted via {@link Subscriber#onError(Throwable)} but exceptions thrown when
     * attempting to subscribe to a {@link Func1}<{@link Subscriber}{@code <T>}, {@link Subscription}>.
     * 
     * @param e
     *            Throwable thrown by {@link Observable#subscribe(Subscriber)}
     * @return Throwable that can be decorated, replaced or just returned as a pass-thru
     */
    public <T> Throwable onSubscribeError(Throwable e) {
        // pass-thru by default
        return e;
    }

    /**
     * Invoked just as the operator functions is called to bind two operations together into a new
     * {@link Observable} and the return value is used as the lifted function
     * <p>
     * This can be used to decorate or replace the {@link Operator} instance or just perform extra
     * logging, metrics and other such things and pass-thru the onSubscribe.
     * 
     * @param lift
     *            original {@link Operator}{@code <R, T>}
     * @return {@link Operator}{@code <R, T>} function that can be modified, decorated, replaced or just
     *         returned as a pass-thru
     */
    public <T, R> Operator<? extends R, ? super T> onLift(final Operator<? extends R, ? super T> lift) {
        return lift;
    }
}

RxJavaObservableExecutionHook类的作用很特殊，仿佛没有甚么太大的作用，传进去甚么（类型）参数，返回甚么（类型）参数。

以下代码所示：

public <T> OnSubscribe<T> onCreate(OnSubscribe<T> f) {
        return f;
    }

public <T> OnSubscribe<T> onSubscribeStart(Observable<? extends T> observableInstance, final OnSubscribe<T> onSubscribe) {
        // pass-thru by default
        return onSubscribe;
    }

至于最后关键的返回结果：

public <T> Subscription onSubscribeReturn(Subscription subscription) {
        // pass-thru by default
        return subscription;
    }

说白了，就是返回定阅的Observer对象。

Observer与Subscriber的区分

它们的区分对使用者来讲主要有两点：

onStart(): 这是 Subscriber 增加的方法。它会在 subscribe 刚开始，而事件还未发送之前被调用，可以用于做1些准备工作，例如数据的清零或重置。这是1个可选方法，默许情况下它的实现为空。需要注意的是，如果对准备工作的线程有要求（例如弹出1个显示进度的对话框，这必须在主线程履行）， onStart() 就不适用了，由于它总是在 subscribe 所产生的线程被调用，而不能指定线程。要在指定的线程来做准备工作，可使用 doOnSubscribe() 方法，具体可以在后面的文中看到。

unsubscribe(): 这是 Subscriber 所实现的另外一个接口 Subscription 的方法，用于取消定阅。在这个方法被调用后，Subscriber 将不再接收事件。1般在这个方法调用前，可使用 isUnsubscribed() 先判断1下状态。 unsubscribe() 这个方法很重要，由于在 subscribe() 以后， Observable 会持有 Subscriber 的援用，这个援用如果不能及时被释放，将有内存泄漏的风险。所以最好保持1个原则：要在不再使用的时候尽快在适合的地方（例如 onPause() onStop() 等方法中）调用 unsubscribe() 来消除援用关系，以免内存泄漏的产生。

2) 创建 Observable

Observable 即被视察者，它决定甚么时候触发事件和触发怎样的事件。 RxJava 使用 create() 方法来创建1个 Observable ，并为它定义事件触发规则：

Observable observable = Observable.create(new Observable.OnSubscribe<String>() {
    @Override
    public void call(Subscriber<? super String> subscriber) {
        subscriber.onNext("Hello");
        subscriber.onNext("John");
        subscriber.onCompleted();
    }
});

可以看到，这里传入了1个 OnSubscribe 对象作为参数。OnSubscribe 会被存储在返回的 Observable 对象中，它的作用相当于1个计划表，当 Observable 被定阅的时候，OnSubscribe 的 call() 方法会自动被调用，事件序列就会依照设定顺次触发（对上面的代码，就是视察者Subscriber 将会被调用两次 onNext() 和1次 onCompleted()）。这样，由被视察者调用了视察者的回调方法，就实现了由被视察者向视察者的事件传递，即视察者模式。

create() 是 RxJava 最基本的创造事件序列的操作符。基于这个操作符， RxJava 还提供了1些方法用来快捷创建事件队列，详见RxJava操作符系列文章：http://blog.csdn.net/jdsjlzx/article/details/51485861

3) Subscribe (定阅)
创建了 Observable 和 Observer 以后，再用 subscribe() 方法将它们联结起来，整条链子就能够工作了。代码情势很简单：

observable.subscribe(observer);
// 或：
observable.subscribe(subscriber);

Observable.subscribe(Subscriber) 的内部实现是这样的（仅核心代码）：

public final Subscription subscribe(Subscriber<? super T> subscriber) {
        return Observable.subscribe(subscriber, this);
    }

    private static <T> Subscription subscribe(Subscriber<? super T> subscriber, Observable<T> observable) {
     // validate and proceed
        if (subscriber == null) {
            throw new IllegalArgumentException("observer can not be null");
        }
        if (observable.onSubscribe == null) {
            throw new IllegalStateException("onSubscribe function can not be null.");
            /*
             * the subscribe function can also be overridden but generally that's not the appropriate approach
             * so I won't mention that in the exception
             */
        }

        // new Subscriber so onStart it
        subscriber.onStart();

        /*
         * See https://github.com/ReactiveX/RxJava/issues/216 for discussion on "Guideline 6.4: Protect calls
         * to user code from within an Observer"
         */
        // if not already wrapped
        if (!(subscriber instanceof SafeSubscriber)) {
            // assign to `observer` so we return the protected version
            subscriber = new SafeSubscriber<T>(subscriber);
        }

        // The code below is exactly the same an unsafeSubscribe but not used because it would 
        // add a significant depth to already huge call stacks.
        try {
            // allow the hook to intercept and/or decorate
            hook.onSubscribeStart(observable, observable.onSubscribe).call(subscriber);
            return hook.onSubscribeReturn(subscriber);
        } catch (Throwable e) {
            // special handling for certain Throwable/Error/Exception types
            Exceptions.throwIfFatal(e);
            // if an unhandled error occurs executing the onSubscribe we will propagate it
            try {
                subscriber.onError(hook.onSubscribeError(e));
            } catch (Throwable e2) {
                Exceptions.throwIfFatal(e2);
                // if this happens it means the onError itself failed (perhaps an invalid function implementation)
                // so we are unable to propagate the error correctly and will just throw
                RuntimeException r = new RuntimeException("Error occurred attempting to subscribe [" + e.getMessage() + "] and then again while trying to pass to onError.", e2);
                // TODO could the hook be the cause of the error in the on error handling.
                hook.onSubscribeError(r);
                // TODO why aren't we throwing the hook's return value.
                throw r;
            }
            return Subscriptions.unsubscribed();
        }
    }

可以看到，subscriber() 做了3件事：

调用 Subscriber.onStart() 。这个方法在前面已介绍过，是1个可选的准备方法。
调用 Observable 中的 OnSubscribe.call(Subscriber) 。在这里，事件发送的逻辑开始运行。从这也能够看出，在 RxJava 中， Observable 其实不是在创建的时候就立即开始发送事件，而是在它被定阅的时候，即当 subscribe() 方法履行的时候。
将传入的 Subscriber 作为 Subscription 返回。这是为了方便 unsubscribe().

全部进程中对象间的关系以下图：
这里写图片描述

除 subscribe(Observer) 和 subscribe(Subscriber) ，subscribe() 还支持不完全定义的回调，RxJava 会自动根据定义创建出 Subscriber 。情势以下：

Action1<String> onNextAction = new Action1<String>() {
    // onNext()
    @Override
    public void call(String s) {
        Log.d(tag, s);
    }
};
Action1<Throwable> onErrorAction = new Action1<Throwable>() {
    // onError()
    @Override
    public void call(Throwable throwable) {
        // Error handling
    }
};
Action0 onCompletedAction = new Action0() {
    // onCompleted()
    @Override
    public void call() {
        Log.d(tag, "completed");
    }
};

// 自动创建 Subscriber ，并使用 onNextAction 来定义 onNext()
observable.subscribe(onNextAction);
// 自动创建 Subscriber ，并使用 onNextAction 和 onErrorAction 来定义 onNext() 和 onError()
observable.subscribe(onNextAction, onErrorAction);
// 自动创建 Subscriber ，并使用 onNextAction、 onErrorAction 和 onCompletedAction 来定义 onNext()、 onError() 和 onCompleted()
observable.subscribe(onNextAction, onErrorAction, onCompletedAction);

注：正如前面所提到的，Observer 和 Subscriber 具有相同的角色，而且 Observer 在 subscribe() 进程中终究会被转换成 Subscriber 对象，因此，从某种程度上来讲用 Subscriber 来代替 Observer ，这样会更加严谨。

根据目前的经验来看，Observer与Subscriber的主要区分在于onNext方法履行终了后是不是取消了定阅。

首先，我们分别定义mSubscriber 和 mObserver 。

以下代码：

protected Subscriber<D> mSubscriber = new Subscriber<D>() {
        @Override
        public void onCompleted() {
            executeOnLoadFinish();
        }

        @Override
        public void onError(Throwable e) {
            TLog.error("onError " + e.toString());
            executeOnLoadDataError(null);
        }

        @Override
        public void onNext(D d) {
            TLog.log("onNext " );
            List<T> list = d;
            TLog.log("entity " + list.size());
            executeOnLoadDataSuccess(list);

            TLog.log("onSuccess totalPage " + totalPage);
        }
    };
    protected Observer<D> mObserver = new Observer<D>() {
        @Override
        public void onCompleted() {
            executeOnLoadFinish();
        }

        @Override
        public void onError(Throwable e) {
            TLog.error("onError " + e.toString());
            executeOnLoadDataError(null);
        }

        @Override
        public void onNext(D d) {
            TLog.log("onNext " );
            List<T> list = d;
            TLog.log("entity " + list.size());
            executeOnLoadDataSuccess(list);

            TLog.log("onSuccess totalPage " + totalPage);
        }
    };

observable.subscribeOn(Schedulers.io())
                .map(new Func1<Response<D>,D>() {

                    @Override
                    public D call(Response<D> response) {
                        if(response == null){
                            throw new ApiException(100);
                        }
                        totalPage = response.total;
                        return response.result;
                    }
                })
                .observeOn(AndroidSchedulers.mainThread())
                //.subscribe(mObserver);
                .subscribe(mSubscriber);

subscribe(mObserver)和subscribe(mSubscriber)履行结果就会有区分：

subscribe(mSubscriber)这类定阅方式在第2次要求数据时就不会履行了，缘由就是onCompleted后自动取消了定阅
（详见文章：http://blog.csdn.net/jdsjlzx/article/details/51542003）；
subscribe(mObserver)则不出现此问题。

提示：个人以为subscribe(mObserver)这个方式更合适分页加载。

请注意，如果你每次都使用subscribe(new Subscriber< T>() {｝）方式实现定阅，就不会出现上面的问题。

以下代码：

private void toSubscribe(Observable<Response<D>> observable) {
        observable.subscribeOn(Schedulers.io())
                .map(new Func1<Response<D>,D>() {

                    @Override
                    public D call(Response<D> response) {
                        if(response == null){
                            throw new ApiException(100);
                        }
                        totalPage = response.total;
                        return response.result;
                    }
                })
                .observeOn(AndroidSchedulers.mainThread())
                .subscribe(new Subscriber<D>() {
                    @Override
                    public void onCompleted() {
                        executeOnLoadFinish();
                    }

                    @Override
                    public void onError(Throwable e) {
                        TLog.error("onError " + e.toString());
                        executeOnLoadDataError(null);
                    }

                    @Override
                    public void onNext(D d) {
                        TLog.log("onNext " );
                        List<T> list = d;
                        TLog.log("entity " + list.size());
                        executeOnLoadDataSuccess(list);

                        TLog.log("onSuccess totalPage " + totalPage);
                    }
                });
    }

固然，这个方式实现分页加载也是可以的。至于哪一个更好，还需要再验证。

------分隔线----------------------------

上一篇 hdu1548 A strange lift(bfs)

下一篇 Python--python基础知识(1)

------分隔线----------------------------