Android——Serializable & Parcelable

Serializable & Parcelable这两种序列化方法是Android中经常使用的方法，Serializable是Android从Java中继承过来的，Parcelable是Android自己提供的方法，Google是推荐使用Parcelable，至于这两种方法的区别，下面通过对源码的分析来慢慢的了解。

在分析源码之前，首先还是说一下序列化在Android中使用的场景：

1）我们在四大组件之间使用intente来传递数据，intente中所传递的数据是需要序列化的

2）binder，binder是IPC机制中是很重要的，binder中所传递的数据也是需要序列化的

3）还有我们经常要把一些数据持久化到存储设备上，这些数据是需要序列化的

4）在网络中传递对象

好啦，现在我们开始对两种序列化方式的源码进行了解：

Serializable：

因为Serializable是一个标示接口，所以我们主要是对它的注解进行阅读

* Marks classes that can be serialized by {@link ObjectOutputStream} and
* deserialized by {@link ObjectInputStream}.

这一句告诉我们，serializable是通过ObjectOutputStream、ObjectInputStream来完成序列化过程的

* <p><strong>Warning:</strong> this interface limits how its implementing
* classes can change in the future. By implementing {@code Serializable} you
* expose your flexible in-memory implementation details as a rigid binary
* representation. Simple code changes--like renaming private fields--are
* not safe when the changed class is serializable.

这一句告诉我们，当实现Serializable接口的类中有代码变化，可能会导致序列化失败

<p>Every serializable class is assigned a version identifier called a {@code
* serialVersionUID}. By default, this identifier is computed by hashing the
* class declaration and its members. This identifier is included in the
* serialized form so that version conflicts can be detected during
* deserialization. If the local {@code serialVersionUID} differs from the
* {@code serialVersionUID} in the serialized data, deserialization will fail
* with an {@link InvalidClassException}.
*
* <p>You can avoid this failure by declaring an explicit {@code
* serialVersionUID}. Declaring an explicit {@code serialVersionUID} tells the
* serialization mechanism that the class is forward and backward compatible
* with all versions that share that {@code serialVersionUID}. Declaring a
* {@code serialVersionUID} looks like this: <pre>   {@code
*
*     private static final long serialVersionUID = 0L;
* }</pre>
* If you declare a {@code serialVersionUID}, you should increment it each
* time your class changes incompatibly with the previous version. Typically
* this is when you add, change or remove a non-transient field.

序列化和反序列化其实是通过serialVersionUID来进行工作的，默认情况下，serialVersionUID是通过hash计算类的成员变量和类的方法来得到的，serialVersionUID是存储在系列化的文件中的，当要反序列化时，就从序列化文件中能拿到serialVersionUID，如果拿到的serialVersionUiD与通过类中的成员变量和方法hash计算出的serialVersionUID一样的话，就反序列化成功，反序列化成功返回的结果与序列化的结果不是同一个对象，但是值是相同的，如果反序列化失败的话就会抛出一个
IncaildClassException。

当然我们也可以在类中自定义serialVersionUID，比如：

private static final long serialVersionUID = 0L；

这样的好处是，当类中的成员变量或者是方法增加、改变、移除的话，我们的反序列化依旧会成功，但是当类的结构有了翻天覆地的变化的时候，也会反序列化失败的。

上面有一句是：

Typically this is when you add, change or remove a non-transient field.

告诉我们，被transient修饰的成员变量是不会参与序列化的。

* <p>You can take control of your serialized form by implementing these two
* methods with these exact signatures in your serializable classes:
* <pre>   {@code
*
*   private void writeObject(java.io.ObjectOutputStream out)
*       throws IOException {
*     // write 'this' to 'out'...
*   }
*
*   private void readObject(java.io.ObjectInputStream in)
*       throws IOException, ClassNotFoundException {
*     // populate the fields of 'this' from the data in 'in'...
*   }
* }</pre>

这段注解告诉我们：默认的序列化过程和反序列化过程是可以改变的，只要重写wireteObject、readObject这两个方法就可以啦，但是一般的情况下我们是不会重写这两个方法的，下面就让我们看看默认它们是怎么实现的：
先看writeObject：

它是ObjectOutputStream中的方法：

public final void writeObject(Object object) throws IOException {
writeObject(object, false);
}

private void writeObject(Object object, boolean unshared) throws IOException {
boolean setOutput = (primitiveTypes == output);
if (setOutput) {
primitiveTypes = null;
}
// This is the specified behavior in JDK 1.2. Very bizarre way to allow
// behavior overriding.
if (subclassOverridingImplementation && !unshared) {
writeObjectOverride(object);
return;
}

try {
// First we need to flush primitive types if they were written
drain();
// Actual work, and class-based replacement should be computed
// if needed.
writeObjectInternal(object, unshared, true, true);
if (setOutput) {
primitiveTypes = output;
}
} catch (IOException ioEx1) {
// This will make it pass through until the top caller. Only the top caller writes the
// exception (where it can).
if (nestedLevels == 0) {
try {
writeNewException(ioEx1);
} catch (IOException ioEx2) {
// If writing the exception to the output stream causes another exception there
// is no need to propagate the second exception or generate a third exception,
// both of which might obscure details of the root cause.
}
}
throw ioEx1; // and then we propagate the original exception
}
}

这段中会判断是否在子类中实现啦writeObject，如果实现了就直接返回，否则就会调用

writeObjectInternal(object, unshared, true, true);

private int writeObjectInternal(Object object, boolean unshared,
boolean computeClassBasedReplacement,
boolean computeStreamReplacement) throws IOException {
...
f (!(enableReplace && computeStreamReplacement)) {
// Is it a Class ?
if (objClass == ObjectStreamClass.CLASSCLASS) {
return writeNewClass((Class<?>) object, unshared);
}
// Is it an ObjectStreamClass ?
if (objClass == ObjectStreamClass.OBJECTSTREAMCLASSCLASS) {
return writeClassDesc((ObjectStreamClass) object, unshared);
}
}
''''
}

这里面我只复制了一些代码，像writeNewClass、wirteClassDesc还有好多这种方法，他们的内部都是在output中存储类型，在objectWritten中存储对象。

再看readObject：

readObject是ObjectInoutStream中的方法：

public final Object readObject() throws OptionalDataException,
ClassNotFoundException, IOException {
return readObject(false);
}

private Object readObject(boolean unshared) throws OptionalDataException,
ClassNotFoundException, IOException {
...
Object result;
try {
// We need this so we can tell when we are returning to the
// original/outside caller
if (++nestedLevels == 1) {
// Remember the caller's class loader
callerClassLoader = VMStack.getClosestUserClassLoader(bootstrapLoader, systemLoader);
}

result = readNonPrimitiveContent(unshared);
if (restoreInput) {
primitiveData = input;
}
}
...
}

调用readNonPrimitiveContent();

private Object readNonPrimitiveContent(boolean unshared)
throws ClassNotFoundException, IOException {
checkReadPrimitiveTypes();
if (primitiveData.available() > 0) {
OptionalDataException e = new OptionalDataException();
e.length = primitiveData.available();
throw e;
}

do {
byte tc = nextTC();
switch (tc) {
case TC_CLASS:
return readNewClass(unshared);
case TC_CLASSDESC:
return readNewClassDesc(unshared);
case TC_ARRAY:
return readNewArray(unshared);
case TC_OBJECT:
return readNewObject(unshared);
case TC_STRING:
return readNewString(unshared);
case TC_LONGSTRING:
return readNewLongString(unshared);
case TC_ENUM:
return readEnum(unshared);
case TC_REFERENCE:
if (unshared) {
readNewHandle();
throw new InvalidObjectException("Unshared read of back reference");
}
return readCyclicReference();
case TC_NULL:
return null;
case TC_EXCEPTION:
Exception exc = readException();
throw new WriteAbortedException("Read an exception", exc);
case TC_RESET:
resetState();
break;
case TC_ENDBLOCKDATA: // Can occur reading class annotation
pushbackTC();
OptionalDataException e = new OptionalDataException();
e.eof = true;
throw e;
default:
throw corruptStream(tc);
}
// Only TC_RESET falls through
} while (true);
}

这里面就是在一个do—while循环里，通过nextTC()来获取输入流中的字段的类型，然后根据不同的类型得到不同的对象

private byte nextTC() throws IOException {
if (hasPushbackTC) {
hasPushbackTC = false; // We are consuming it
} else {
// Just in case a later call decides to really push it back,
// we don't require the caller to pass it as parameter
pushbackTC = input.readByte();
}
return pushbackTC;
}

这里input其实就是上面那个output，在do-while循环里就是不断的从input中取，然后获取字段，最终把字段封装成一个对象返回，这就是反序列化的过程。

可以看出在序列化中要进行大量的I\O的操作，而且里面还有用到反射的机制，对资源的消耗也比较大

Parcelable：

老规矩，先看一下注解：

* Interface for classes whose instances can be written to
* and restored from a {@link Parcel}.  Classes implementing the Parcelable
* interface must also have a static field called <code>CREATOR</code>, which
* is an object implementing the {@link Parcelable.Creator Parcelable.Creator}
* interface.

这句话意思是，对象能够通过Parcel来进行序列化和反序列化，而且重要的是，在实现Parcelable中必须要有一个成员变量，而且还要被 static field来修饰，类型为Parcelable.Creator，它是Parcelable中的一个内部接口

<p>A typical implementation of Parcelable is:</p>
*
* <pre>
* public class MyParcelable implements Parcelable {
*     private int mData;
*
*     public int describeContents() {
*         return 0;
*     }
*
*     public void writeToParcel(Parcel out, int flags) {
*         out.writeInt(mData);
*     }
*
*     public static final Parcelable.Creator<MyParcelable> CREATOR
*             = new Parcelable.Creator<MyParcelable>() {
*         public MyParcelable createFromParcel(Parcel in) {
*             return new MyParcelable(in);
*         }
*
*         public MyParcelable[] newArray(int size) {
*             return new MyParcelable[size];
*         }
*     };
*
*     private MyParcelable(Parcel in) {
*         mData = in.readInt();
*     }
* }</pre>

这段注解是Android给我们提供的一个例子，因为Parcelable相比于Serializable实现起来比较复杂，可能处于这种考虑，就给我们一个实例，确实给我们带来啦一些好处，我们就不用那么刻意的去记，当使用的时候看看源码就OK了！

public static final int CONTENTS_FILE_DESCRIPTOR = 0x0001;

/**
* Describe the kinds of special objects contained in this Parcelable's
* marshalled representation.
*
* @return a bitmask indicating the set of special object types marshalled
* by the Parcelable.
*/
public int describeContents();

文件描述符，当对象有文件描述符时就返回1(CONTENTS_FILE_DESCRIPTOP),否则就返回0，大多数情况就返回0

/**
* Flatten this object in to a Parcel.
*
* @param dest The Parcel in which the object should be written.
* @param flags Additional flags about how the object should be written.
* May be 0 or {@link #PARCELABLE_WRITE_RETURN_VALUE}.
*/
public void writeToParcel(Parcel dest, int flags);

该方法是将对象写入到序列化结构中，dest：对象应该要被写入的序列化结构的容器

flags：有两种结果，为1(PARCELABLE_WRITE_RETURN_VALUE)从字面意 思可以看到就是要求要返回对象，大多数情况下是为0的。

/**
* Interface that must be implemented and provided as a public CREATOR
* field that generates instances of your Parcelable class from a Parcel.
*/
public interface Creator<T> {
}

这是Parcelable中的内部接口，它的作用是从一个Parcel中构造出一个实现了Parcelable的类的实例，也就是承担了反序列化的过程。

然后看Creator中的两个接口：

/**
* Create a new instance of the Parcelable class, instantiating it
* from the given Parcel whose data had previously been written by
* {@link Parcelable#writeToParcel Parcelable.writeToParcel()}.
*
* @param source The Parcel to read the object's data from.
* @return Returns a new instance of the Parcelable class.
*/
public T createFromParcel(Parcel source);

这个方法就是从Parcel中构造出一个实现了Parcelable的类的实例，一般我们实现是通过类的构造方法中实现的：

return new T(source)；

/**
* Create a new array of the Parcelable class.
*
* @param size Size of the array.
* @return Returns an array of the Parcelable class, with every entry
* initialized to null.
*/
public T[] newArray(int size);

创建一个指定长度的原始对象的数组，一般我们就直接返回：

return new T[size]；

在Parcelable中实现序列化和反序列化主要是通过Parcel来实现的，对于Parcel，可以去看

Android中的Parcel机制(上)

最后进行一下总结：

	优点	缺点
Serializable	1）实现起来简单 2）通过自定义serialVersionUID 可以 减少反序列化失败的发生	1）反射 2）大量的I\O操作，反序列化速度慢 2）耗资源
parcelable	1）速度快	1）实现难，难以阅读 2）难以维护

至于选择哪种方式序列化，还需酌情选择。

最后，谢谢大家的观看！