在网上很少看到有解释 RemoteCallbackList类的,没办法要搞懂只能去看源码,大致理解:一个容器,容纳的对象是一些接口,用于执行列表中对象的回调函数,主要用于服务调用activity函数,或者解释为服务端和客户端通信。

/* * Copyright (C) 2008 The Android Open Source Project * * 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 android.os;import java.util.HashMap;/** * Takes care of the grunt work of maintaining a list of remote interfaces, * typically for the use of performing callbacks from a * {@link android.app.Service} to its clients.  In particular, this: * * 
    *
  • Keeps track of a set of registered {@link IInterface} callbacks, * taking care to identify them through their underlying unique {@link IBinder} * (by calling {@link IInterface#asBinder IInterface.asBinder()}. *
  • Attaches a {@link IBinder.DeathRecipient IBinder.DeathRecipient} to * each registered interface, so that it can be cleaned out of the list if its * process goes away. *
  • Performs locking of the underlying list of interfaces to deal with * multithreaded incoming calls, and a thread-safe way to iterate over a * snapshot of the list without holding its lock. *
* *

To use this class, simply create a single instance along with your * service, and call its {@link #register} and {@link #unregister} methods * as client register and unregister with your service. To call back on to * the registered clients, use {@link #beginBroadcast}, * {@link #getBroadcastItem}, and {@link #finishBroadcast}. * *

If a registered callback's process goes away, this class will take * care of automatically removing it from the list. If you want to do * additional work in this situation, you can create a subclass that * implements the {@link #onCallbackDied} method. */public class RemoteCallbackList<E extends IInterface> { /*package*/ HashMap mCallbacks = new HashMap(); private Object[] mActiveBroadcast; private int mBroadcastCount = -1; private boolean mKilled = false; private final class Callback implements IBinder.DeathRecipient { final E mCallback; final Object mCookie; Callback(E callback, Object cookie) { mCallback = callback; mCookie = cookie; } public void binderDied() { synchronized (mCallbacks) { mCallbacks.remove(mCallback.asBinder()); } onCallbackDied(mCallback, mCookie); } } /** * Simple version of {@link RemoteCallbackList#register(E, Object)} * that does not take a cookie object. */ public boolean register(E callback) { return register(callback, null); } /** * Add a new callback to the list. This callback will remain in the list * until a corresponding call to {@link #unregister} or its hosting process * goes away. If the callback was already registered (determined by * checking to see if the {@link IInterface#asBinder callback.asBinder()} * object is already in the list), then it will be left as-is. * Registrations are not counted; a single call to {@link #unregister} * will remove a callback after any number calls to register it. * * @param callback The callback interface to be added to the list. Must * not be null -- passing null here will cause a NullPointerException. * Most services will want to check for null before calling this with * an object given from a client, so that clients can't crash the * service with bad data. * * @param cookie Optional additional data to be associated with this * callback. * * @return Returns true if the callback was successfully added to the list. * Returns false if it was not added, either because {@link #kill} had * previously been called or the callback's process has gone away. * * @see #unregister * @see #kill * @see #onCallbackDied */ public boolean register(E callback, Object cookie) { synchronized (mCallbacks) { if (mKilled) { return false; } IBinder binder = callback.asBinder(); try { Callback cb = new Callback(callback, cookie); binder.linkToDeath(cb, 0); mCallbacks.put(binder, cb); return true; } catch (RemoteException e) { return false; } } } /** * Remove from the list a callback that was previously added with * {@link #register}. This uses the * {@link IInterface#asBinder callback.asBinder()} object to correctly * find the previous registration. * Registrations are not counted; a single unregister call will remove * a callback after any number calls to {@link #register} for it. * * @param callback The callback to be removed from the list. Passing * null here will cause a NullPointerException, so you will generally want * to check for null before calling. * * @return Returns true if the callback was found and unregistered. Returns * false if the given callback was not found on the list. * * @see #register */ public boolean unregister(E callback) { synchronized (mCallbacks) { Callback cb = mCallbacks.remove(callback.asBinder()); if (cb != null) { cb.mCallback.asBinder().unlinkToDeath(cb, 0); return true; } return false; } } /** * Disable this callback list. All registered callbacks are unregistered, * and the list is disabled so that future calls to {@link #register} will * fail. This should be used when a Service is stopping, to prevent clients * from registering callbacks after it is stopped. * * @see #register */ public void kill() { synchronized (mCallbacks) { for (Callback cb : mCallbacks.values()) { cb.mCallback.asBinder().unlinkToDeath(cb, 0); } mCallbacks.clear(); mKilled = true; } } /** * Old version of {@link #onCallbackDied(E, Object)} that * does not provide a cookie. */ public void onCallbackDied(E callback) { } /** * Called when the process hosting a callback in the list has gone away. * The default implementation calls {@link #onCallbackDied(E)} * for backwards compatibility. * * @param callback The callback whose process has died. Note that, since * its process has died, you can not make any calls on to this interface. * You can, however, retrieve its IBinder and compare it with another * IBinder to see if it is the same object. * @param cookie The cookie object original provided to * {@link #register(E, Object)}. * * @see #register */ public void onCallbackDied(E callback, Object cookie) { onCallbackDied(callback); } /** * Prepare to start making calls to the currently registered callbacks. * This creates a copy of the callback list, which you can retrieve items * from using {@link #getBroadcastItem}. Note that only one broadcast can * be active at a time, so you must be sure to always call this from the * same thread (usually by scheduling with {@link Handler}) or * do your own synchronization. You must call {@link #finishBroadcast} * when done. * *

A typical loop delivering a broadcast looks like this: * *

     * int i = callbacks.beginBroadcast();     * while (i > 0) {     *     i--;     *     try {     *         callbacks.getBroadcastItem(i).somethingHappened();     *     } catch (RemoteException e) {     *         // The RemoteCallbackList will take care of removing     *         // the dead object for us.     *     }     * }     * callbacks.finishBroadcast();
* * @return Returns the number of callbacks in the broadcast, to be used * with {@link #getBroadcastItem} to determine the range of indices you * can supply. * * @see #getBroadcastItem * @see #finishBroadcast */
public int beginBroadcast() { synchronized (mCallbacks) { if (mBroadcastCount > 0) { throw new IllegalStateException( "beginBroadcast() called while already in a broadcast"); } final int N = mBroadcastCount = mCallbacks.size(); if (N <= 0) { return 0; } Object[] active = mActiveBroadcast; if (active == null || active.length < N) { mActiveBroadcast = active = new Object[N]; } int i=0; for (Callback cb : mCallbacks.values()) { active[i++] = cb; } return i; } } /** * Retrieve an item in the active broadcast that was previously started * with {@link #beginBroadcast}. This can only be called after * the broadcast is started, and its data is no longer valid after * calling {@link #finishBroadcast}. * *

Note that it is possible for the process of one of the returned * callbacks to go away before you call it, so you will need to catch * {@link RemoteException} when calling on to the returned object. * The callback list itself, however, will take care of unregistering * these objects once it detects that it is no longer valid, so you can * handle such an exception by simply ignoring it. * * @param index Which of the registered callbacks you would like to * retrieve. Ranges from 0 to 1-{@link #beginBroadcast}. * * @return Returns the callback interface that you can call. This will * always be non-null. * * @see #beginBroadcast */ public E getBroadcastItem(int index) { return ((Callback)mActiveBroadcast[index]).mCallback; } /** * Retrieve the cookie associated with the item * returned by {@link #getBroadcastItem(int)}. * * @see #getBroadcastItem */ public Object getBroadcastCookie(int index) { return ((Callback)mActiveBroadcast[index]).mCookie; } /** * Clean up the state of a broadcast previously initiated by calling * {@link #beginBroadcast}. This must always be called when you are done * with a broadcast. * * @see #beginBroadcast */ public void finishBroadcast() { if (mBroadcastCount < 0) { throw new IllegalStateException( "finishBroadcast() called outside of a broadcast"); } Object[] active = mActiveBroadcast; if (active != null) { final int N = mBroadcastCount; for (int i=0; inull; } } mBroadcastCount = -1; } /** * Returns the number of registered callbacks. Note that the number of registered * callbacks may differ from the value returned by {@link #beginBroadcast()} since * the former returns the number of callbacks registered at the time of the call * and the second the number of callback to which the broadcast will be delivered. *

* This function is useful to decide whether to schedule a broadcast if this * requires doing some work which otherwise would not be performed. *

* * @return The size. */
public int getRegisteredCallbackCount() { synchronized (mCallbacks) { if (mKilled) { return 0; } return mCallbacks.size(); } }}

通过查看上面的源码,我们可以知道主要是维护 mCallbacks ,mCallbacks中包含我们的客户端接口

    /*package*/ HashMap mCallbacks            = new HashMap();

其中常用的几个函数 register

    public boolean register(E callback) {        return register(callback, null);    }    public boolean register(E callback, Object cookie) {        synchronized (mCallbacks) {            if (mKilled) {                return false;            }            IBinder binder = callback.asBinder();            try {                Callback cb = new Callback(callback, cookie);                binder.linkToDeath(cb, 0);                mCallbacks.put(binder, cb);                return true;            } catch (RemoteException e) {                return false;            }        }    }

向 mCallbacks中添加接口对象,主要使用 mCallbacks.put(binder, cb);方法,

unregister()方法

    public boolean unregister(E callback) {        synchronized (mCallbacks) {            Callback cb = mCallbacks.remove(callback.asBinder());            if (cb != null) {                cb.mCallback.asBinder().unlinkToDeath(cb, 0);                return true;            }            return false;        }    }

通过 mCallbacks.remove(callback.asBinder()); 从HashMap中删除

kill()方法

    public void kill() {        synchronized (mCallbacks) {            for (Callback cb : mCallbacks.values()) {                cb.mCallback.asBinder().unlinkToDeath(cb, 0);            }            mCallbacks.clear();            mKilled = true;        }    }

通过 mCallbacks.clear(); 清空集合。

beginBroadcast() ;

    public int beginBroadcast() {        synchronized (mCallbacks) {            if (mBroadcastCount > 0) {                throw new IllegalStateException(                        "beginBroadcast() called while already in a broadcast");            }            final int N = mBroadcastCount = mCallbacks.size();            if (N <= 0) {                return 0;            }            Object[] active = mActiveBroadcast;            if (active == null || active.length < N) {                mActiveBroadcast = active = new Object[N];            }            int i=0;            for (Callback cb : mCallbacks.values()) {                active[i++] = cb;            }            return i;        }    }

从注释的解释来说大致是:准备开始调用当前注册的回调。这将创建一个回调列表的副本,你可以从使用getBroadcastItem检索条目,注意:一次只能激活一个广播,所以你必须确保总是从同一个线程调用这个或者是自己做同步。完成后必须调用finishBroadcast。
返回值:callbacks的大小;

如果 mBroadcastCount > 0 说明之前调用过beginBroadcast(),则会抛出异常;

getBroadcastItem()方法 从副本中获取我们之前存入的接口对象

    public E getBroadcastItem(int index) {        return ((Callback)mActiveBroadcast[index]).mCallback;    }

finishBroadcast(),清理我们之前建立的active[] 副本

    public void finishBroadcast() {        if (mBroadcastCount < 0) {            throw new IllegalStateException(                    "finishBroadcast() called outside of a broadcast");        }        Object[] active = mActiveBroadcast;        if (active != null) {            final int N = mBroadcastCount;            for (int i=0; inull;            }        }        mBroadcastCount = -1;    }

如果 mBroadcastCount <0 说明之前掉用过finishBroadcast() 则会抛出异常。

常用的函数大概就是这个样子,如果我理解有误欢迎大家留言指正,也欢迎大家补充,一起学习

更多相关文章

  1. android如何完全退出应用程序
  2. Android培训班(68)dex文件打开流程
  3. Android两种相机的调用方式
  4. Unity3d与Android的相互调用
  5. 关于android WebViewClient 的方法解释
  6. android 事件传递机制
  7. Android进程通信-AIDL
  8. android 解锁,锁屏流程
  9. 面试题及答案

随机推荐

  1. android透明状态栏效果
  2. Android 桌面部件
  3. android实现共享数据
  4. Duplicate class XXXXX library-2.4.0.ja
  5. android 自定义控件pickview
  6. 转 android背景选择器selector用法汇总
  7. Android工程运行 ADB server didn't ACK
  8. Android crop image size
  9. android 图片放大缩小
  10. Android获取本周本月本年的第一天和最后