1 背景##
在Android中任何耗时的操作都不能放在UI主线程中，所以耗时的操作都需要使用异步实现。同样的，在ContentProvider中也可能存在耗时操作，这时也该使用异步操作，而3.0之后最推荐的异步操作就是Loader。它可以方便我们在Activity和Fragment中异步加载数据，而不是用线程或AsyncTask，他的优点如下：
提供异步加载数据机制；
对数据源变化进行监听，实时更新数据；
在Activity配置发生变化（如横竖屏切换）时不用重复加载数据；
适用于任何Activity和Fragment；
PS：由于在我们现在的项目中都使用了Loader来处理数据加载（而且由于粗心跳过几个坑，譬如Loader ID重复导致数据逻辑异常、多线程中restartLoader导致Loader抛出异常（最后保证都在UI线程中执行即可）等），所以接下来我们进行下使用及源码浅析。
既然接下来准备要说说他的使用强大之处了，那不妨我们先来一张图直观的感性认识下不用Loader（左）与用Loader（右）对我们开发者及代码复杂度和框架的影响吧，如下：
loaders.png
2 基础使用实例##
Loader API 摘要###
在应用中使用加载器时，可能会涉及到多个类和接口。 下表汇总了这些类和接口：
| 类/接口 | 说明 |
|---|---|
| LoaderManager | 一种与 Activity 或 Fragment 相关联的的抽象类，用于管理一个或多个 Loader 实例。 这有助于应用管理与 Activity 或 Fragment 生命周期相关联的、运行时间较长的操作。它最常见的用法是与 CursorLoader 一起使用，但应用可自由写入其自己的加载器，用于加载其他类型的数据。 每个 Activity 或片段中只有一个 LoaderManager。但一个 LoaderManager 可以有多个加载器。 |
| LoaderManager.LoaderCallbacks | 一种回调接口，用于客户端与 LoaderManager 进行交互。例如，您可使用 onCreateLoader() 回调方法创建新的加载器。 |
| Loader | 一种执行异步数据加载的抽象类。这是加载器的基类。 您通常会使用 CursorLoader，但您也可以实现自己的子类。加载器处于活动状态时，应监控其数据源并在内容变化时传递新结果。 |
| AsyncTaskLoader | 提供 AsyncTask 来执行工作的抽象加载器。 |
| CursorLoader | AsyncTaskLoader 的子类，它将查询 ContentResolver 并返回一个 Cursor。此类采用标准方式为查询游标实现 Loader 协议。它是以 AsyncTaskLoader 为基础而构建，在后台线程中执行游标查询，以免阻塞应用的 UI。使用此加载器是从 ContentProvider 异步加载数据的最佳方式，而不用通过片段或 Activity 的 API 来执行托管查询。 |
上表中的类和接口是您在应用中用于实现加载器的基本组件。 并非您创建的每个加载器都要用到上述所有类和接口。但是，为了初始化加载器以及实现一个 Loader 类（如 CursorLoader），您始终需要要引用 LoaderManager。 下文将为您展示如何在应用中使用这些类和接口。
在应用中使用加载器###
此部分描述如何在 Android 应用中使用加载器。使用加载器的应用通常包括：
- Activity 或 Fragment。
- LoaderManager 的实例。
- 一个 CursorLoader，用于加载由 ContentProvider 支持的数据。您也可以实现自己的 Loader 或 AsyncTaskLoader 子类，从其他源中加载数据。
- 一个 LoaderManager.LoaderCallbacks 实现。您可以使用它来创建新加载器，并管理对现有加载器的引用。
- 一种显示加载器数据的方法，如 SimpleCursorAdapter。
- 使用 CursorLoader 时的数据源，如 ContentProvider。
启动加载器###
LoaderManager 可在 Activity 或 Fragment 内管理一个或多个 Loader 实例。每个 Activity 或片段中只有一个 LoaderManager。
通常，您会在 Activity 的 onCreate() 方法或片段的onActivityCreated() 方法内初始化 Loader。您执行操作如下：
// Prepare the loader. Either re-connect with an existing one,
// or start a new one.
getLoaderManager().initLoader(0, null, this);
initLoader() 方法采用以下参数：
-
用于标识加载器的唯一 ID。在此示例中，ID 为 0。
-
在构建时提供给加载器的可选参数（在此示例中为 null）。
-
LoaderManager.LoaderCallbacks 实现， LoaderManager 将调用此实现来报告加载器事件。在此示例中，本地类实现 LoaderManager.LoaderCallbacks 接口，因此它会传递对自身的引用 this。
initLoader() 调用确保加载器已初始化且处于活动状态。这可能会出现两种结果： -
如果 ID 指定的加载器已存在，则将重复使用上次创建的加载器。
-
如果 ID 指定的加载器不存在，则 initLoader() 将触发 LoaderManager.LoaderCallbacks 方法 onCreateLoader()。在此方法中，您可以实现代码以实例化并返回新加载器。有关详细介绍，请参阅 onCreateLoader 部分。
无论何种情况，给定的 LoaderManager.LoaderCallbacks 实现均与加载器相关联，且将在加载器状态变化时调用。如果在调用时，调用程序处于启动状态，且请求的加载器已存在并生成了数据，则系统将立即调用 onLoadFinished()（在 initLoader() 期间），因此您必须为此做好准备。 有关此回调的详细介绍，请参阅 onLoadFinished。
请注意，initLoader() 方法将返回已创建的 Loader，但您不必捕获其引用。LoaderManager 将自动管理加载器的生命周期。LoaderManager 将根据需要启动和停止加载，并维护加载器的状态及其相关内容。 这意味着您很少直接与加载器进行交互（有关使用加载器方法调整加载器行为的示例，请参阅 LoaderThrottle 示例）。当特定事件发生时，您通常会使用 LoaderManager.LoaderCallbacks 方法干预加载进程。有关此主题的详细介绍，请参阅使用 LoaderManager 回调。
重启加载器###
当您使用 initLoader() 时（如上所述），它将使用含有指定 ID 的现有加载器（如有）。如果没有，则它会创建一个。但有时，您想舍弃这些旧数据并重新开始。
要舍弃旧数据，请使用 restartLoader()。例如，当用户的查询更改时，此 SearchView.OnQueryTextListener 实现将重启加载器。 加载器需要重启，以便它能够使用修订后的搜索过滤器执行新查询：
public boolean onQueryTextChanged(String newText) {
// Called when the action bar search text has changed. Update
// the search filter, and restart the loader to do a new query
// with this filter.
mCurFilter = !TextUtils.isEmpty(newText) ? newText : null;
getLoaderManager().restartLoader(0, null, this);
return true;
}
使用 LoaderManager 回调###
LoaderManager.LoaderCallbacks 是一个支持客户端与 LoaderManager 交互的回调接口。
加载器（特别是 CursorLoader）在停止运行后，仍需保留其数据。这样，应用即可保留 Activity 或片段的 onStop() 和 onStart() 方法中的数据。当用户返回应用时，无需等待它重新加载这些数据。您可使用 LoaderManager.LoaderCallbacks 方法了解何时创建新加载器，并告知应用何时停止使用加载器的数据。
LoaderManager.LoaderCallbacks 包括以下方法：
- onCreateLoader()：针对指定的 ID 进行实例化并返回新的 Loader
- onLoadFinished() ：将在先前创建的加载器完成加载时调用
- onLoaderReset()：将在先前创建的加载器重置且其数据因此不可用时调用
下文更详细地描述了这些方法。
onCreateLoader
当您尝试访问加载器时（例如，通过 initLoader()），该方法将检查是否已存在由该 ID 指定的加载器。 如果没有，它将触发 LoaderManager.LoaderCallbacks 方法 onCreateLoader()。在此方法中，您可以创建新加载器。 通常，这将是 CursorLoader，但您也可以实现自己的 Loader 子类。
在此示例中，onCreateLoader() 回调方法创建了 CursorLoader。您必须使用其构造函数方法来构建 CursorLoader。该方法需要对 ContentProvider 执行查询时所需的一系列完整信息。具体地说，它需要：
- uri：用于检索内容的 URI
- projection：要返回的列的列表。传递 null 时，将返回所有列，这样会导致效率低下
- selection：一种用于声明要返回哪些行的过滤器，采用 SQL WHERE 子句格式（WHERE 本身除外）。传递 null 时，将为指定的 URI 返回所有行
- selectionArgs：您可以在 selection 中包含 ?s，它将按照在 selection 中显示的顺序替换为 selectionArgs 中的值。该值将绑定为字串符
- sortOrder：行的排序依据，采用 SQL ORDER BY 子句格式（ORDER BY 自身除外）。传递 null 时，将使用默认排序顺序（可能并未排序）
例如：
// If non-null, this is the current filter the user has provided.
String mCurFilter;
...
public Loader<Cursor> onCreateLoader(int id, Bundle args) {
// This is called when a new Loader needs to be created. This
// sample only has one Loader, so we don't care about the ID.
// First, pick the base URI to use depending on whether we are
// currently filtering.
Uri baseUri;
if (mCurFilter != null) {
baseUri = Uri.withAppendedPath(Contacts.CONTENT_FILTER_URI,
Uri.encode(mCurFilter));
} else {
baseUri = Contacts.CONTENT_URI;
}
// Now create and return a CursorLoader that will take care of
// creating a Cursor for the data being displayed.
String select = "((" + Contacts.DISPLAY_NAME + " NOTNULL) AND ("
+ Contacts.HAS_PHONE_NUMBER + "=1) AND ("
+ Contacts.DISPLAY_NAME + " != '' ))";
return new CursorLoader(getActivity(), baseUri,
CONTACTS_SUMMARY_PROJECTION, select, null,
Contacts.DISPLAY_NAME + " COLLATE LOCALIZED ASC");
}
onLoadFinished###
当先前创建的加载器完成加载时，将调用此方法。该方法必须在为此加载器提供的最后一个数据释放之前调用。 此时，您应移除所有使用的旧数据（因为它们很快会被释放），但不要自行释放这些数据，因为这些数据归其加载器所有，其加载器会处理它们。
当加载器发现应用不再使用这些数据时，即会释放它们。 例如，如果数据是来自 CursorLoader 的一个游标，则您不应手动对其调用 close()。如果游标放置在 CursorAdapter 中，则应使用 swapCursor() 方法，使旧 Cursor 不会关闭。例如：
// This is the Adapter being used to display the list's data.
SimpleCursorAdapter mAdapter;
...
public void onLoadFinished(Loader<Cursor> loader, Cursor data) {
// Swap the new cursor in. (The framework will take care of closing the
// old cursor once we return.)
mAdapter.swapCursor(data);
}
onLoaderReset###
此方法将在先前创建的加载器重置且其数据因此不可用时调用。 通过此回调，您可以了解何时将释放数据，因而能够及时移除其引用。
此实现调用值为 null 的swapCursor()：
// This is the Adapter being used to display the list's data.
SimpleCursorAdapter mAdapter;
...
public void onLoaderReset(Loader<Cursor> loader) {
// This is called when the last Cursor provided to onLoadFinished()
// above is about to be closed. We need to make sure we are no
// longer using it.
mAdapter.swapCursor(null);
}
3 源码浅析###
和上面的基本使用介绍一样，关于Loader的源码浅析过程会涉及到Activity、Fragment、LoaderManager、Loader、AsyncLoader、CursorLoader等类。所以我们分析的过程还是和以前一样，依据使用顺序进行分析。
我们在分析之前先来看一个Loader框架概要图,如下：
loadermanager.jpg
通过上面图和前面的基础实例我们可以发现Loader的框架和各个类的职责都很明确。Activity和Fragment管理LoaderManager，LoaderManager管理Loader，Loader得到数据后触发在LoaderManager中实现的Loader的callback接口，LoaderManager在接收到Loader的callback回传调运时触发我们Activity或Fragment中实现的LoaderManager回调callback接口，就这样就实现了Loader的所有功能，而我们平时写代码一般只用关心LoaderManager的callback实现即可；对于自定义Loader可能还需要关心AsyncTaskLoader子类的实现。
看看Activity中的LoaderManager#####
image.png
恢复所有LoaderManagers的保存状态。 给定的LoaderManager列表是跨配置更改保留的LoaderManager实例。
onCreateLoaderConfig.png
onStart.png
performOnstop.png
performDestroy.png
--------------------上面是Activity里对Loader的生命周期管理---------------------
onstart2.png
onstart3.png
LoaderManagerStart.png
通过上面的分析可以发现，Activity其实真正的管理了Activity及Fragment的LoaderManager（Fragment也会管理一部分自己LoaderManager的周期），而LoaderManager又管理了Loader，可以发现他们各自的管理范围都是十分的清晰明了的。
LoadManager及其实现类LoadManagerImpl的浅析#####
public abstract class LoaderManager {
/**
* Callback interface for a client to interact with the manager.
*/
public interface LoaderCallbacks<D> {
/**
* Instantiate and return a new Loader for the given ID.
*
* @param id The ID whose loader is to be created.
* @param args Any arguments supplied by the caller.
* @return Return a new Loader instance that is ready to start loading.
*/
public Loader<D> onCreateLoader(int id, Bundle args);
/**
* Called when a previously created loader has finished its load. Note
* that normally an application is <em>not</em> allowed to commit fragment
* transactions while in this call, since it can happen after an
* activity's state is saved. See {@link FragmentManager#beginTransaction()
* FragmentManager.openTransaction()} for further discussion on this.
*
* <p>This function is guaranteed to be called prior to the release of
* the last data that was supplied for this Loader. At this point
* you should remove all use of the old data (since it will be released
* soon), but should not do your own release of the data since its Loader
* owns it and will take care of that. The Loader will take care of
* management of its data so you don't have to. In particular:
*
* <ul>
* <li> <p>The Loader will monitor for changes to the data, and report
* them to you through new calls here. You should not monitor the
* data yourself. For example, if the data is a {@link android.database.Cursor}
* and you place it in a {@link android.widget.CursorAdapter}, use
* the {@link android.widget.CursorAdapter#CursorAdapter(android.content.Context,
* android.database.Cursor, int)} constructor <em>without</em> passing
* in either {@link android.widget.CursorAdapter#FLAG_AUTO_REQUERY}
* or {@link android.widget.CursorAdapter#FLAG_REGISTER_CONTENT_OBSERVER}
* (that is, use 0 for the flags argument). This prevents the CursorAdapter
* from doing its own observing of the Cursor, which is not needed since
* when a change happens you will get a new Cursor throw another call
* here.
* <li> The Loader will release the data once it knows the application
* is no longer using it. For example, if the data is
* a {@link android.database.Cursor} from a {@link android.content.CursorLoader},
* you should not call close() on it yourself. If the Cursor is being placed in a
* {@link android.widget.CursorAdapter}, you should use the
* {@link android.widget.CursorAdapter#swapCursor(android.database.Cursor)}
* method so that the old Cursor is not closed.
* </ul>
*
* @param loader The Loader that has finished.
* @param data The data generated by the Loader.
*/
public void onLoadFinished(Loader<D> loader, D data);
/**
* Called when a previously created loader is being reset, and thus
* making its data unavailable. The application should at this point
* remove any references it has to the Loader's data.
*
* @param loader The Loader that is being reset.
*/
public void onLoaderReset(Loader<D> loader);
}
/**
* Ensures a loader is initialized and active. If the loader doesn't
* already exist, one is created and (if the activity/fragment is currently
* started) starts the loader. Otherwise the last created
* loader is re-used.
*
* <p>In either case, the given callback is associated with the loader, and
* will be called as the loader state changes. If at the point of call
* the caller is in its started state, and the requested loader
* already exists and has generated its data, then
* callback {@link LoaderCallbacks#onLoadFinished} will
* be called immediately (inside of this function), so you must be prepared
* for this to happen.
*
* @param id A unique identifier for this loader. Can be whatever you want.
* Identifiers are scoped to a particular LoaderManager instance.
* @param args Optional arguments to supply to the loader at construction.
* If a loader already exists (a new one does not need to be created), this
* parameter will be ignored and the last arguments continue to be used.
* @param callback Interface the LoaderManager will call to report about
* changes in the state of the loader. Required.
*/
public abstract <D> Loader<D> initLoader(int id, Bundle args,
LoaderManager.LoaderCallbacks<D> callback);
/**
* Starts a new or restarts an existing {@link android.content.Loader} in
* this manager, registers the callbacks to it,
* and (if the activity/fragment is currently started) starts loading it.
* If a loader with the same id has previously been
* started it will automatically be destroyed when the new loader completes
* its work. The callback will be delivered before the old loader
* is destroyed.
*
* @param id A unique identifier for this loader. Can be whatever you want.
* Identifiers are scoped to a particular LoaderManager instance.
* @param args Optional arguments to supply to the loader at construction.
* @param callback Interface the LoaderManager will call to report about
* changes in the state of the loader. Required.
*/
public abstract <D> Loader<D> restartLoader(int id, Bundle args,
LoaderManager.LoaderCallbacks<D> callback);
/**
* Stops and removes the loader with the given ID. If this loader
* had previously reported data to the client through
* {@link LoaderCallbacks#onLoadFinished(Loader, Object)}, a call
* will be made to {@link LoaderCallbacks#onLoaderReset(Loader)}.
*/
public abstract void destroyLoader(int id);
/**
* Return the Loader with the given id or null if no matching Loader
* is found.
*/
public abstract <D> Loader<D> getLoader(int id);
/**
* Print the LoaderManager's state into the given stream.
*
* @param prefix Text to print at the front of each line.
* @param fd The raw file descriptor that the dump is being sent to.
* @param writer A PrintWriter to which the dump is to be set.
* @param args Additional arguments to the dump request.
*/
public abstract void dump(String prefix, FileDescriptor fd, PrintWriter writer, String[] args);
/**
* Control whether the framework's internal loader manager debugging
* logs are turned on. If enabled, you will see output in logcat as
* the framework performs loader operations.
*/
public static void enableDebugLogging(boolean enabled) {
LoaderManagerImpl.DEBUG = enabled;
}
/** @hide for internal testing only */
public FragmentHostCallback getFragmentHostCallback() { return null; }
}
/**
* Returns a {@link LoaderManager}.
*/
public LoaderManager getLoaderManager() {
return mHost.getLoaderManagerImpl();
}
我们获取的LoaderManager实例其实就是LoaderManagerImpl对象，而LoaderManagerImpl又是LoaderManager类的子类.
这十分类似于我们现在的框架结构，VRL和具体的ViewMng的实现关系，把各自的功能单一化。
CursorLoader###
关于用到的这个，就得层层剥茧去一探究竟了
public class CursorLoader extends AsyncTaskLoader<Cursor>
CursorLoader继承于AsyncTaskLoader
public abstract class AsyncTaskLoader<D> extends Loader<D> {
而AsyncTaskLoader是抽象类并继承于Loader
我们就先看看这个Loader基类吧，该类核心方法及内部类结构图如下：
Loader.png
public class Loader<D> {
int mId;
OnLoadCompleteListener<D> mListener;
OnLoadCanceledListener<D> mOnLoadCanceledListener;
Context mContext;
boolean mStarted = false;
boolean mAbandoned = false;
boolean mReset = true;
boolean mContentChanged = false;
boolean mProcessingChange = false;
/**
* An implementation of a ContentObserver that takes care of connecting
* it to the Loader to have the loader re-load its data when the observer
* is told it has changed. You do not normally need to use this yourself;
* it is used for you by {@link CursorLoader} to take care of executing
* an update when the cursor's backing data changes.
*/
//数据源变化监听器（观察者模式），实现了ContentObserver类
public final class ForceLoadContentObserver extends ContentObserver {
public ForceLoadContentObserver() {
super(new Handler());
}
@Override
public boolean deliverSelfNotifications() {
return true;
}
@Override
public void onChange(boolean selfChange) {
//实质是调运Loader的forceLoad方法
onContentChanged();
}
}
/**
* Interface that is implemented to discover when a Loader has finished
* loading its data. You do not normally need to implement this yourself;
* it is used in the implementation of {@link android.app.LoaderManager}
* to find out when a Loader it is managing has completed so that this can
* be reported to its client. This interface should only be used if a
* Loader is not being used in conjunction with LoaderManager.
* Loader加载完成接口，当加载完成时Loader通知
* loaderManager,loaderManager再回调我们initLoader方法的callback
*/
public interface OnLoadCompleteListener<D> {
/**
* Called on the thread that created the Loader when the load is complete.
*
* @param loader the loader that completed the load
* @param data the result of the load
*/
public void onLoadComplete(Loader<D> loader, D data);
}
/**
* Interface that is implemented to discover when a Loader has been canceled
* before it finished loading its data. You do not normally need to implement
* this yourself; it is used in the implementation of {@link android.app.LoaderManager}
* to find out when a Loader it is managing has been canceled so that it
* can schedule the next Loader. This interface should only be used if a
* Loader is not being used in conjunction with LoaderManager.
*/
public interface OnLoadCanceledListener<D> {
/**
* Called on the thread that created the Loader when the load is canceled.
*
* @param loader the loader that canceled the load
*/
public void onLoadCanceled(Loader<D> loader);
}
/**
* Stores away the application context associated with context.
* Since Loaders can be used across multiple activities it's dangerous to
* store the context directly; always use {@link #getContext()} to retrieve
* the Loader's Context, don't use the constructor argument directly.
* The Context returned by {@link #getContext} is safe to use across
* Activity instances.
*
* @param context used to retrieve the application context.
*/
public Loader(Context context) {
mContext = context.getApplicationContext();
}
/**
* Sends the result of the load to the registered listener. Should only be called by subclasses.
*
* Must be called from the process's main thread.
*
* @param data the result of the load
*/
//加载完成时回调传递加载数据结果，实质是对OnLoadCompleteListener接口方法的封装
public void deliverResult(D data) {
if (mListener != null) {
mListener.onLoadComplete(this, data);
}
}
/**
* Informs the registered {@link OnLoadCanceledListener} that the load has been canceled.
* Should only be called by subclasses.
*
* Must be called from the process's main thread.
*/
public void deliverCancellation() {
if (mOnLoadCanceledListener != null) {
mOnLoadCanceledListener.onLoadCanceled(this);
}
}
/**
* @return an application context retrieved from the Context passed to the constructor.
*/
public Context getContext() {
return mContext;
}
/**
* @return the ID of this loader
*/
public int getId() {
return mId;
}
/**
* Registers a class that will receive callbacks when a load is complete.
* The callback will be called on the process's main thread so it's safe to
* pass the results to widgets.
*
* <p>Must be called from the process's main thread.
*/
public void registerListener(int id, OnLoadCompleteListener<D> listener) {
if (mListener != null) {
throw new IllegalStateException("There is already a listener registered");
}
mListener = listener;
mId = id;
}
/**
* Remove a listener that was previously added with {@link #registerListener}.
*
* Must be called from the process's main thread.
*/
public void unregisterListener(OnLoadCompleteListener<D> listener) {
if (mListener == null) {
throw new IllegalStateException("No listener register");
}
if (mListener != listener) {
throw new IllegalArgumentException("Attempting to unregister the wrong listener");
}
mListener = null;
}
/**
* Registers a listener that will receive callbacks when a load is canceled.
* The callback will be called on the process's main thread so it's safe to
* pass the results to widgets.
*
* Must be called from the process's main thread.
*
* @param listener The listener to register.
*/
public void registerOnLoadCanceledListener(OnLoadCanceledListener<D> listener) {
if (mOnLoadCanceledListener != null) {
throw new IllegalStateException("There is already a listener registered");
}
mOnLoadCanceledListener = listener;
}
/**
* Unregisters a listener that was previously added with
* {@link #registerOnLoadCanceledListener}.
*
* Must be called from the process's main thread.
*
* @param listener The listener to unregister.
*/
public void unregisterOnLoadCanceledListener(OnLoadCanceledListener<D> listener) {
if (mOnLoadCanceledListener == null) {
throw new IllegalStateException("No listener register");
}
if (mOnLoadCanceledListener != listener) {
throw new IllegalArgumentException("Attempting to unregister the wrong listener");
}
mOnLoadCanceledListener = null;
}
/**
* Return whether this load has been started. That is, its {@link #startLoading()}
* has been called and no calls to {@link #stopLoading()} or
* {@link #reset()} have yet been made.
*/
public boolean isStarted() {
return mStarted;
}
/**
* Return whether this loader has been abandoned. In this state, the
* loader <em>must not</em> report any new data, and <em>must</em> keep
* its last reported data valid until it is finally reset.
*/
public boolean isAbandoned() {
return mAbandoned;
}
/**
* Return whether this load has been reset. That is, either the loader
* has not yet been started for the first time, or its {@link #reset()}
* has been called.
*/
public boolean isReset() {
return mReset;
}
/**
* This function will normally be called for you automatically by
* {@link android.app.LoaderManager} when the associated fragment/activity
* is being started. When using a Loader with {@link android.app.LoaderManager},
* you <em>must not</em> call this method yourself, or you will conflict
* with its management of the Loader.
*
* Starts an asynchronous load of the Loader's data. When the result
* is ready the callbacks will be called on the process's main thread.
* If a previous load has been completed and is still valid
* the result may be passed to the callbacks immediately.
* The loader will monitor the source of
* the data set and may deliver future callbacks if the source changes.
* Calling {@link #stopLoading} will stop the delivery of callbacks.
*
* <p>This updates the Loader's internal state so that
* {@link #isStarted()} and {@link #isReset()} will return the correct
* values, and then calls the implementation's {@link #onStartLoading()}.
*
* <p>Must be called from the process's main thread.
*/
public final void startLoading() {
mStarted = true;
mReset = false;
mAbandoned = false;
onStartLoading();
}
/**
* Subclasses must implement this to take care of loading their data,
* as per {@link #startLoading()}. This is not called by clients directly,
* but as a result of a call to {@link #startLoading()}.
*/
//真正开始加载数据的地方-----空方法，子类实现
protected void onStartLoading() {
}
/**
* Attempt to cancel the current load task.
* Must be called on the main thread of the process.
*
* <p>Cancellation is not an immediate operation, since the load is performed
* in a background thread. If there is currently a load in progress, this
* method requests that the load be canceled, and notes this is the case;
* once the background thread has completed its work its remaining state
* will be cleared. If another load request comes in during this time,
* it will be held until the canceled load is complete.
*
* @return Returns <tt>false</tt> if the task could not be canceled,
* typically because it has already completed normally, or
* because {@link #startLoading()} hasn't been called; returns
* <tt>true</tt> otherwise. When <tt>true</tt> is returned, the task
* is still running and the {@link OnLoadCanceledListener} will be called
* when the task completes.
*/
public boolean cancelLoad() {
return onCancelLoad();
}
/**
* Subclasses must implement this to take care of requests to {@link #cancelLoad()}.
* This will always be called from the process's main thread.
*
* @return Returns <tt>false</tt> if the task could not be canceled,
* typically because it has already completed normally, or
* because {@link #startLoading()} hasn't been called; returns
* <tt>true</tt> otherwise. When <tt>true</tt> is returned, the task
* is still running and the {@link OnLoadCanceledListener} will be called
* when the task completes.
*/
//真正取消的地方-----，子类实现！！！！！！return false表示取消失败（已完成或未开始）
protected boolean onCancelLoad() {
return false;
}
/**
* Force an asynchronous load. Unlike {@link #startLoading()} this will ignore a previously
* loaded data set and load a new one. This simply calls through to the
* implementation's {@link #onForceLoad()}. You generally should only call this
* when the loader is started -- that is, {@link #isStarted()} returns true.
*
* <p>Must be called from the process's main thread.
*/
public void forceLoad() {
onForceLoad();
}
/**
* Subclasses must implement this to take care of requests to {@link #forceLoad()}.
* This will always be called from the process's main thread.
*/
protected void onForceLoad() {
}
/**
* This function will normally be called for you automatically by
* {@link android.app.LoaderManager} when the associated fragment/activity
* is being stopped. When using a Loader with {@link android.app.LoaderManager},
* you <em>must not</em> call this method yourself, or you will conflict
* with its management of the Loader.
*
* <p>Stops delivery of updates until the next time {@link #startLoading()} is called.
* Implementations should <em>not</em> invalidate their data at this point --
* clients are still free to use the last data the loader reported. They will,
* however, typically stop reporting new data if the data changes; they can
* still monitor for changes, but must not report them to the client until and
* if {@link #startLoading()} is later called.
*
* <p>This updates the Loader's internal state so that
* {@link #isStarted()} will return the correct
* value, and then calls the implementation's {@link #onStopLoading()}.
*
* <p>Must be called from the process's main thread.
*/
public void stopLoading() {
mStarted = false;
onStopLoading();
}
/**
* Subclasses must implement this to take care of stopping their loader,
* as per {@link #stopLoading()}. This is not called by clients directly,
* but as a result of a call to {@link #stopLoading()}.
* This will always be called from the process's main thread.
*/
protected void onStopLoading() {
}
/**
* This function will normally be called for you automatically by
* {@link android.app.LoaderManager} when restarting a Loader. When using
* a Loader with {@link android.app.LoaderManager},
* you <em>must not</em> call this method yourself, or you will conflict
* with its management of the Loader.
*
* Tell the Loader that it is being abandoned. This is called prior
* to {@link #reset} to have it retain its current data but not report
* any new data.
*/
public void abandon() {
mAbandoned = true;
onAbandon();
}
/**
* Subclasses implement this to take care of being abandoned. This is
* an optional intermediate state prior to {@link #onReset()} -- it means that
* the client is no longer interested in any new data from the loader,
* so the loader must not report any further updates. However, the
* loader <em>must</em> keep its last reported data valid until the final
* {@link #onReset()} happens. You can retrieve the current abandoned
* state with {@link #isAbandoned}.
*/
protected void onAbandon() {
}
/**
* This function will normally be called for you automatically by
* {@link android.app.LoaderManager} when destroying a Loader. When using
* a Loader with {@link android.app.LoaderManager},
* you <em>must not</em> call this method yourself, or you will conflict
* with its management of the Loader.
*
* Resets the state of the Loader. The Loader should at this point free
* all of its resources, since it may never be called again; however, its
* {@link #startLoading()} may later be called at which point it must be
* able to start running again.
*
* <p>This updates the Loader's internal state so that
* {@link #isStarted()} and {@link #isReset()} will return the correct
* values, and then calls the implementation's {@link #onReset()}.
*
* <p>Must be called from the process's main thread.
*/
public void reset() {
onReset();
mReset = true;
mStarted = false;
mAbandoned = false;
mContentChanged = false;
mProcessingChange = false;
}
/**
* Subclasses must implement this to take care of resetting their loader,
* as per {@link #reset()}. This is not called by clients directly,
* but as a result of a call to {@link #reset()}.
* This will always be called from the process's main thread.
*/
protected void onReset() {
}
/**
* Take the current flag indicating whether the loader's content had
* changed while it was stopped. If it had, true is returned and the
* flag is cleared.
*/
public boolean takeContentChanged() {
boolean res = mContentChanged;
mContentChanged = false;
mProcessingChange |= res;
return res;
}
/**
* Commit that you have actually fully processed a content change that
* was returned by {@link #takeContentChanged}. This is for use with
* {@link #rollbackContentChanged()} to handle situations where a load
* is cancelled. Call this when you have completely processed a load
* without it being cancelled.
*/
public void commitContentChanged() {
mProcessingChange = false;
}
/**
* Report that you have abandoned the processing of a content change that
* was returned by {@link #takeContentChanged()} and would like to rollback
* to the state where there is again a pending content change. This is
* to handle the case where a data load due to a content change has been
* canceled before its data was delivered back to the loader.
*/
public void rollbackContentChanged() {
if (mProcessingChange) {
onContentChanged();
}
}
/**
* Called when {@link ForceLoadContentObserver} detects a change. The
* default implementation checks to see if the loader is currently started;
* if so, it simply calls {@link #forceLoad()}; otherwise, it sets a flag
* so that {@link #takeContentChanged()} returns true.
*
* <p>Must be called from the process's main thread.
*/
//上面ForceLoadContentObserver内部类的onChange方法调运
public void onContentChanged() {
if (mStarted) {
forceLoad();
} else {
// This loader has been stopped, so we don't want to load
// new data right now... but keep track of it changing to
// refresh later if we start again.
mContentChanged = true;
}
}
/**
* For debugging, converts an instance of the Loader's data class to
* a string that can be printed. Must handle a null data.
*/
public String dataToString(D data) {
StringBuilder sb = new StringBuilder(64);
DebugUtils.buildShortClassTag(data, sb);
sb.append("}");
return sb.toString();
}
@Override
public String toString() {
StringBuilder sb = new StringBuilder(64);
DebugUtils.buildShortClassTag(this, sb);
sb.append(" id=");
sb.append(mId);
sb.append("}");
return sb.toString();
}
/**
* Print the Loader's state into the given stream.
*
* @param prefix Text to print at the front of each line.
* @param fd The raw file descriptor that the dump is being sent to.
* @param writer A PrintWriter to which the dump is to be set.
* @param args Additional arguments to the dump request.
*/
public void dump(String prefix, FileDescriptor fd, PrintWriter writer, String[] args) {
writer.print(prefix); writer.print("mId="); writer.print(mId);
writer.print(" mListener="); writer.println(mListener);
if (mStarted || mContentChanged || mProcessingChange) {
writer.print(prefix); writer.print("mStarted="); writer.print(mStarted);
writer.print(" mContentChanged="); writer.print(mContentChanged);
writer.print(" mProcessingChange="); writer.println(mProcessingChange);
}
if (mAbandoned || mReset) {
writer.print(prefix); writer.print("mAbandoned="); writer.print(mAbandoned);
writer.print(" mReset="); writer.println(mReset);
}
}
}
通过上面粗略的分析可以发现，Loader基类无非也就是一个方法接口的定义类，组织预留了一些方法供LoaderManager去调运处理，同时需要子类实现其提供的一些onXXX方法，以便LoaderManager调运Loader的方法时可以触发Loader子类的实现逻辑。
AsyncTaskLoader抽象子类源码#####
public abstract class AsyncTaskLoader<D> extends Loader<D> {
static final String TAG = "AsyncTaskLoader";
static final boolean DEBUG = false;
//LoadTask内部类是对AsyncTask的封装，实现了Runnable接口
final class LoadTask extends AsyncTask<Void, Void, D> implements Runnable {
private final CountDownLatch mDone = new CountDownLatch(1);
// Set to true to indicate that the task has been posted to a handler for
// execution at a later time. Used to throttle updates.
boolean waiting;
//AsyncTask的子线程中执行AsyncTaskLoader的onLoadInBackground方法！！！！
/* Runs on a worker thread */
@Override
protected D doInBackground(Void... params) {
if (DEBUG) Log.v(TAG, this + " >>> doInBackground");
try {
D data = AsyncTaskLoader.this.onLoadInBackground();
if (DEBUG) Log.v(TAG, this + " <<< doInBackground");
//把执行结果数据D返回到UI线程
return data;
} catch (OperationCanceledException ex) {
if (!isCancelled()) {
// onLoadInBackground threw a canceled exception spuriously.
// This is problematic because it means that the LoaderManager did not
// cancel the Loader itself and still expects to receive a result.
// Additionally, the Loader's own state will not have been updated to
// reflect the fact that the task was being canceled.
// So we treat this case as an unhandled exception.
throw ex;
}
if (DEBUG) Log.v(TAG, this + " <<< doInBackground (was canceled)", ex);
return null;
}
}
/* Runs on the UI thread */
@Override
protected void onPostExecute(D data) {
if (DEBUG) Log.v(TAG, this + " onPostExecute");
try {
//AsyncTask子线程执行完毕后回调AsyncTaskLoader的dispatchOnLoadComplete方法
AsyncTaskLoader.this.dispatchOnLoadComplete(this, data);
} finally {
mDone.countDown();
}
}
/* Runs on the UI thread */
@Override
protected void onCancelled(D data) {
if (DEBUG) Log.v(TAG, this + " onCancelled");
try {
//取消AsyncTask时调运
AsyncTaskLoader.this.dispatchOnCancelled(this, data);
} finally {
mDone.countDown();
}
}
/* Runs on the UI thread, when the waiting task is posted to a handler.
* This method is only executed when task execution was deferred (waiting was true). */
//Runnable的实现方法
@Override
public void run() {
waiting = false;
AsyncTaskLoader.this.executePendingTask();
}
/* Used for testing purposes to wait for the task to complete. */
public void waitForLoader() {
try {
mDone.await();
} catch (InterruptedException e) {
// Ignore
}
}
}
private final Executor mExecutor;
volatile LoadTask mTask;
volatile LoadTask mCancellingTask;
long mUpdateThrottle;
long mLastLoadCompleteTime = -10000;
Handler mHandler;
public AsyncTaskLoader(Context context) {
this(context, AsyncTask.THREAD_POOL_EXECUTOR);
}
/** {@hide} */
public AsyncTaskLoader(Context context, Executor executor) {
super(context);
mExecutor = executor;
}
/**
* Set amount to throttle updates by. This is the minimum time from
* when the last {@link #loadInBackground()} call has completed until
* a new load is scheduled.
*
* @param delayMS Amount of delay, in milliseconds.
*/
public void setUpdateThrottle(long delayMS) {
mUpdateThrottle = delayMS;
if (delayMS != 0) {
mHandler = new Handler();
}
}
@Override
protected void onForceLoad() {
super.onForceLoad();
cancelLoad();
mTask = new LoadTask();
if (DEBUG) Log.v(TAG, "Preparing load: mTask=" + mTask);
executePendingTask();
}
@Override
protected boolean onCancelLoad() {
if (DEBUG) Log.v(TAG, "onCancelLoad: mTask=" + mTask);
if (mTask != null) {
if (mCancellingTask != null) {
// There was a pending task already waiting for a previous
// one being canceled; just drop it.
if (DEBUG) Log.v(TAG,
"cancelLoad: still waiting for cancelled task; dropping next");
if (mTask.waiting) {
mTask.waiting = false;
mHandler.removeCallbacks(mTask);
}
mTask = null;
return false;
} else if (mTask.waiting) {
// There is a task, but it is waiting for the time it should
// execute. We can just toss it.
if (DEBUG) Log.v(TAG, "cancelLoad: task is waiting, dropping it");
mTask.waiting = false;
mHandler.removeCallbacks(mTask);
mTask = null;
return false;
} else {
boolean cancelled = mTask.cancel(false);
if (DEBUG) Log.v(TAG, "cancelLoad: cancelled=" + cancelled);
if (cancelled) {
mCancellingTask = mTask;
cancelLoadInBackground();
}
mTask = null;
return cancelled;
}
}
return false;
}
/**
* Called if the task was canceled before it was completed. Gives the class a chance
* to clean up post-cancellation and to properly dispose of the result.
*
* @param data The value that was returned by {@link #loadInBackground}, or null
* if the task threw {@link OperationCanceledException}.
*/
public void onCanceled(D data) {
}
//LoadTask的Runnable方法run中执行
void executePendingTask() {
if (mCancellingTask == null && mTask != null) {
if (mTask.waiting) {
mTask.waiting = false;
mHandler.removeCallbacks(mTask);
}
if (mUpdateThrottle > 0) {
long now = SystemClock.uptimeMillis();
if (now < (mLastLoadCompleteTime+mUpdateThrottle)) {
// Not yet time to do another load.
if (DEBUG) Log.v(TAG, "Waiting until "
+ (mLastLoadCompleteTime+mUpdateThrottle)
+ " to execute: " + mTask);
mTask.waiting = true;
mHandler.postAtTime(mTask, mLastLoadCompleteTime+mUpdateThrottle);
return;
}
}
if (DEBUG) Log.v(TAG, "Executing: " + mTask);
//真正的触发执行AsyncTask方法
mTask.executeOnExecutor(mExecutor, (Void[]) null);
}
}
void dispatchOnCancelled(LoadTask task, D data) {
onCanceled(data);
if (mCancellingTask == task) {
if (DEBUG) Log.v(TAG, "Cancelled task is now canceled!");
rollbackContentChanged();
mLastLoadCompleteTime = SystemClock.uptimeMillis();
mCancellingTask = null;
if (DEBUG) Log.v(TAG, "Delivering cancellation");
//触发Loader的接口方法onLoadCanceled，在LoaderManager中实现
deliverCancellation();
executePendingTask();
}
}
void dispatchOnLoadComplete(LoadTask task, D data) {
if (mTask != task) {
if (DEBUG) Log.v(TAG, "Load complete of old task, trying to cancel");
dispatchOnCancelled(task, data);
} else {
if (isAbandoned()) {
// This cursor has been abandoned; just cancel the new data.
onCanceled(data);
} else {
commitContentChanged();
mLastLoadCompleteTime = SystemClock.uptimeMillis();
mTask = null;
if (DEBUG) Log.v(TAG, "Delivering result");
//触发Loader的接口方法onLoadComplete，在LoaderManager中实现
deliverResult(data);
}
}
}
/**
* Called on a worker thread to perform the actual load and to return
* the result of the load operation.
*
* Implementations should not deliver the result directly, but should return them
* from this method, which will eventually end up calling {@link #deliverResult} on
* the UI thread. If implementations need to process the results on the UI thread
* they may override {@link #deliverResult} and do so there.
*
* To support cancellation, this method should periodically check the value of
* {@link #isLoadInBackgroundCanceled} and terminate when it returns true.
* Subclasses may also override {@link #cancelLoadInBackground} to interrupt the load
* directly instead of polling {@link #isLoadInBackgroundCanceled}.
*
* When the load is canceled, this method may either return normally or throw
* {@link OperationCanceledException}. In either case, the {@link Loader} will
* call {@link #onCanceled} to perform post-cancellation cleanup and to dispose of the
* result object, if any.
*
* @return The result of the load operation.
*
* @throws OperationCanceledException if the load is canceled during execution.
*
* @see #isLoadInBackgroundCanceled
* @see #cancelLoadInBackground
* @see #onCanceled
*/
//需要子类实现！！！！！在子线程中执行
public abstract D loadInBackground();
/**
* Calls {@link #loadInBackground()}.
*
* This method is reserved for use by the loader framework.
* Subclasses should override {@link #loadInBackground} instead of this method.
*
* @return The result of the load operation.
*
* @throws OperationCanceledException if the load is canceled during execution.
*
* @see #loadInBackground
*/
//LoadTask（AsyncTask的子线程中回调）中调运
protected D onLoadInBackground() {
return loadInBackground();
}
/**
* Called on the main thread to abort a load in progress.
*
* Override this method to abort the current invocation of {@link #loadInBackground}
* that is running in the background on a worker thread.
*
* This method should do nothing if {@link #loadInBackground} has not started
* running or if it has already finished.
*
* @see #loadInBackground
*/
public void cancelLoadInBackground() {
}
/**
* Returns true if the current invocation of {@link #loadInBackground} is being canceled.
*
* @return True if the current invocation of {@link #loadInBackground} is being canceled.
*
* @see #loadInBackground
*/
public boolean isLoadInBackgroundCanceled() {
return mCancellingTask != null;
}
/**
* Locks the current thread until the loader completes the current load
* operation. Returns immediately if there is no load operation running.
* Should not be called from the UI thread: calling it from the UI
* thread would cause a deadlock.
* <p>
* Use for testing only. <b>Never</b> call this from a UI thread.
*
* @hide
*/
public void waitForLoader() {
LoadTask task = mTask;
if (task != null) {
task.waitForLoader();
}
}
@Override
public void dump(String prefix, FileDescriptor fd, PrintWriter writer, String[] args) {
super.dump(prefix, fd, writer, args);
if (mTask != null) {
writer.print(prefix); writer.print("mTask="); writer.print(mTask);
writer.print(" waiting="); writer.println(mTask.waiting);
}
if (mCancellingTask != null) {
writer.print(prefix); writer.print("mCancellingTask="); writer.print(mCancellingTask);
writer.print(" waiting="); writer.println(mCancellingTask.waiting);
}
if (mUpdateThrottle != 0) {
writer.print(prefix); writer.print("mUpdateThrottle=");
TimeUtils.formatDuration(mUpdateThrottle, writer);
writer.print(" mLastLoadCompleteTime=");
TimeUtils.formatDuration(mLastLoadCompleteTime,
SystemClock.uptimeMillis(), writer);
writer.println();
}
}
}
上面继承Loader的AsyncTaskLoader其实质是提供了一个基于AsyncTask工作机制的Loader（子类LoadTask继承AsyncTask<Void, Void, D>，并且实现了Runable接口，功能十分强大。），但是不可直接用，因为其为abstract抽象类，所以我们需要继承实现它才可以使用，然而好在系统API已经帮我们提供了他现成的子类CursorLoader，但CursorLoader同时也限制了Loader的泛型数据为Cursor类型。
public class CursorLoader extends AsyncTaskLoader<Cursor> {
final ForceLoadContentObserver mObserver;
Uri mUri;
String[] mProjection;
String mSelection;
String[] mSelectionArgs;
String mSortOrder;
Cursor mCursor;
CancellationSignal mCancellationSignal;
/* Runs on a worker thread */
//最核心的实现方法，在这里查询获取数据
@Override
public Cursor loadInBackground() {
synchronized (this) {
if (isLoadInBackgroundCanceled()) {
throw new OperationCanceledException();
}
mCancellationSignal = new CancellationSignal();
}
try {
//耗时的查询操作
Cursor cursor = getContext().getContentResolver().query(mUri, mProjection, mSelection,
mSelectionArgs, mSortOrder, mCancellationSignal);
if (cursor != null) {
try {
// Ensure the cursor window is filled.
cursor.getCount();
/*给Cursor设置观察者；
* ContentProvider通知Cursor的观察者数据发生了改变，
* Cursor通知CursorLoader的观察者数据发生了改变，CursorLoader通过
* ContentProvider重新加载新的数据*/
cursor.registerContentObserver(mObserver);
} catch (RuntimeException ex) {
cursor.close();
throw ex;
}
}
return cursor;
} finally {
synchronized (this) {
mCancellationSignal = null;
}
}
}
@Override
public void cancelLoadInBackground() {
super.cancelLoadInBackground();
synchronized (this) {
if (mCancellationSignal != null) {
mCancellationSignal.cancel();
}
}
}
/* Runs on the UI thread */
@Override
public void deliverResult(Cursor cursor) {
if (isReset()) {
// An async query came in while the loader is stopped
if (cursor != null) {
cursor.close();
}
return;
}
Cursor oldCursor = mCursor;
mCursor = cursor;
if (isStarted()) {
super.deliverResult(cursor);
}
if (oldCursor != null && oldCursor != cursor && !oldCursor.isClosed()) {
oldCursor.close();
}
}
/**
* Creates an empty unspecified CursorLoader. You must follow this with
* calls to {@link #setUri(Uri)}, {@link #setSelection(String)}, etc
* to specify the query to perform.
*/
public CursorLoader(Context context) {
super(context);
mObserver = new ForceLoadContentObserver();
}
/**
* Creates a fully-specified CursorLoader. See
* {@link ContentResolver#query(Uri, String[], String, String[], String)
* ContentResolver.query()} for documentation on the meaning of the
* parameters. These will be passed as-is to that call.
*/
public CursorLoader(Context context, Uri uri, String[] projection, String selection,
String[] selectionArgs, String sortOrder) {
super(context);
/*新建一个当前类（Loader）的内部类对象，数据库变化时调运
* ForceLoadContentObserver的onChange方法，onChange调运Loader的
* onContentChanged方法，onContentChanged调运Loader的forceLoad方法*/
mObserver = new ForceLoadContentObserver();
mUri = uri;
mProjection = projection;
mSelection = selection;
mSelectionArgs = selectionArgs;
mSortOrder = sortOrder;
}
/**
* Starts an asynchronous load of the contacts list data. When the result is ready the callbacks
* will be called on the UI thread. If a previous load has been completed and is still valid
* the result may be passed to the callbacks immediately.
*
* Must be called from the UI thread
*/
@Override
protected void onStartLoading() {
if (mCursor != null) {
deliverResult(mCursor);
}
if (takeContentChanged() || mCursor == null) {
forceLoad();
}
}
/**
* Must be called from the UI thread
*/
@Override
protected void onStopLoading() {
// Attempt to cancel the current load task if possible.
cancelLoad();
}
@Override
public void onCanceled(Cursor cursor) {
if (cursor != null && !cursor.isClosed()) {
cursor.close();
}
}
@Override
protected void onReset() {
super.onReset();
// Ensure the loader is stopped
onStopLoading();
if (mCursor != null && !mCursor.isClosed()) {
mCursor.close();
}
mCursor = null;
}
public Uri getUri() {
return mUri;
}
public void setUri(Uri uri) {
mUri = uri;
}
public String[] getProjection() {
return mProjection;
}
public void setProjection(String[] projection) {
mProjection = projection;
}
public String getSelection() {
return mSelection;
}
public void setSelection(String selection) {
mSelection = selection;
}
public String[] getSelectionArgs() {
return mSelectionArgs;
}
public void setSelectionArgs(String[] selectionArgs) {
mSelectionArgs = selectionArgs;
}
public String getSortOrder() {
return mSortOrder;
}
public void setSortOrder(String sortOrder) {
mSortOrder = sortOrder;
}
@Override
public void dump(String prefix, FileDescriptor fd, PrintWriter writer, String[] args) {
super.dump(prefix, fd, writer, args);
writer.print(prefix); writer.print("mUri="); writer.println(mUri);
writer.print(prefix); writer.print("mProjection=");
writer.println(Arrays.toString(mProjection));
writer.print(prefix); writer.print("mSelection="); writer.println(mSelection);
writer.print(prefix); writer.print("mSelectionArgs=");
writer.println(Arrays.toString(mSelectionArgs));
writer.print(prefix); writer.print("mSortOrder="); writer.println(mSortOrder);
writer.print(prefix); writer.print("mCursor="); writer.println(mCursor);
writer.print(prefix); writer.print("mContentChanged="); writer.println(mContentChanged);
}
}
CursorLoader的封装大大简化了应用开发者代码的复杂度；它完全就是一个异步的数据库查询工具。
Loaders相关总结###
-
一次完整的数据加载流程为Activity调用LoaderManager的doStart()方法，然后LoaderManager调用Loader的startLoading()方法，然后Loader调运AsyncTaskLoader的doingBackground()方法进行耗时数据加载，然后AsyncTaskLoader回调LoaderManager的complete数据加载完成方法，接着LoaderManager回调我们在Activity中实现的callback中的onLoadFinish()方法。
-
Acivity和Fragment的生命周期主动管理了LoaderManager，每个Activity用一个ArrayMap的mAllLoaderManager来保存当前Activity及其附属Frament的唯一LoaderManager；在Activity配置发生变化时，Activity在destory前会保存mAllLoaderManager，当Activity再重新创建时，会在Activity的onAttcach()、onCreate()、performStart()方法中恢复mAllLoaderManager。
-
由于整个Activity和Fragment主动管理了Loader，所以关于Loader的释放（譬如CursorLoader的Cursor关闭等）不需要我们人为处理，Loader框架会帮我们很好的处理的；同时特别注意，对于CursorLoader，当我们数据源发生变化时Loader框架会通过ContentObserver调用onContentChanged的forceLoad方法重新请求数据进行回调刷新。
细节点睛
- 对于毕竟敏感操作的cursor，AsyncTaskLoader在封装中onPostExecute方法内，完成数据操作之后就对应的通过dispatchOnLoadComplete调用父类Loader中的deliverResult方法，CursorLoader子类实现了这个方法，便在里面做出了对cursor的关闭操作。
——未完待续#
参考以及引用资料：
1、Google Loader官方指导文档
2、Loader源码解析
3、LoaderManager详解
网友评论