Analysis of Hessian serialization principle

1. hessian Serialization example

1.1 hessian Protocol Brief

1.1.1 characteristic

Refer to the description in the official document , Here are some brief descriptions ,http://hessian.caucho.com/doc/hessian-serialization.html.

• It must self describe the serialization type , That is, no external schema or interface definition is required
• It must be language independent , Including support for scripting languages
• It must be able to read and write in a single way
• It must be as compact as possible
• It must be simple , Only in this way can we effectively test and implement
• Must be as fast as possible
• It has to support Unicode character string
• It has to support 8 Bit binary data , There is no need to escape or use attachments
• It must support encryption 、 Compress 、 Signature and transaction context envelopes

1.1.2 hessian Syntactic introduction

The following examples , More reference ：http://hessian.caucho.com/doc/hessian-serialization.html.

        #boolean true/false
boolean ::= 'T'
        ::= 'F'
          
        # list/vector
list    ::= 'V' type? length? value* 'z'
        ::= 'v' int int value*          # first int Represents a type reference ,  the second int Length 
          

        #32-bit  signed int （ such as 0x90 Encoded as 0）
int     ::= 'I' b3 b2 b1 b0
        ::= [x80-xbf]                   #-x10 to x3f
        ::= [xc0-xcf] b0                #-x800 to x7ff
        ::= [xd0-xd7] b1 b0             #-x40000 to x3ffff

1.2 hessian Examples of serialization and deserialization

public static void serialize1(Student student){
    
  FileOutputStream fileOutputStream;
  try {
    
    fileOutputStream = new FileOutputStream("/Users/zhuqiuhui/Desktop/studentHession.txt");
    //  Get byte stream from object 
    ByteArrayOutputStream os = new ByteArrayOutputStream();
    Hessian2Output output = new Hessian2Output(os);
    output.writeObject(student);
    output.getBytesOutputStream().flush();
    output.completeMessage();
    output.close();

    //  Write to a file 
    fileOutputStream.write(os.toByteArray());
  } catch (Exception e) {
    
    e.printStackTrace();
  }
}

public static Student deserialize1(){
    
  FileInputStream fileInputStream;
  Object result = null;
  try {
    
    fileInputStream = new FileInputStream("/Users/zhuqiuhui/Desktop/studentHession.txt");
    byte[] data = new byte[1024];
    int len = fileInputStream.read(data);
    System.out.println("read byte length:" + len);

    //  Read the object from the stream 
    ByteArrayInputStream is = new ByteArrayInputStream(data);
    Hessian2Input input = new Hessian2Input(is);
    result = input.readObject();
  } catch (Exception e) {
    
    e.printStackTrace();
  }
  return (Student)result;
}

public static void main(String[] args) {
    

  Student student = new Student();
  student.setId("123");
  student.setName(" Fang Chen ");

  //  serialize 
  // serialize1(student);

  //  Deserialization 
  Student serializestudent = deserialize1();
  System.out.println("deserialize result entity id is "+serializestudent.getId());
  System.out.println("deserialize result entity name is "+serializestudent.getName());
}

1.3 hessian Deal with the jdk difference

• Difference one ：java Serialization cannot cross language
• Difference two ： Versions of old and new objects Java Through one serialVersionUID To relate , Developers need to pay attention to the semantics of serialization
• Difference three ：java Serialization does not support encryption
• Difference 4 ：Java Serialized content ratio hessian Big

2. hessian Serialization analysis

2.1 hessian Serialization must serialVersionUID Do you ？

hessian Serialized objects do not need to be added serialVersionUID,hessian When serializing, write the description information of the class to byte[] in

2.2 hessian Serialization and deserialization source code analysis

2.2.1 （ back ） The corresponding relationship of serializers

type	Serializer	Deserializer
Collection	CollectionSerializer	CollectionDeserializer
Map	MapSerializer	MapDeserializer
Iterator	IteratorSerializer	IteratorDeserializer
Annotation	AnnotationSerializer	AnnotationDeserializer
Interface	ObjectSerializer	ObjectDeserializer
Array	ArraySerializer	ArrayDeserializer
Enumeration	EnumerationSerializer	EnumerationDeserializer
Enum	EnumSerializer	EnumDeserializer
Class	ClassSerializer	ClassDeserializer
Default	JavaSerializer	JavaDeserializer
Throwable	ThrowableSerializer
InputStream	InputStreamSerializer	InputStreamDeserializer
InetAddress	InetAddressSerializer

2.2.2 Why serialize objects implements Serializable Interface ？

// com.caucho.hessian.io.SerializerFactory  The default serializer will be obtained when serializing 
protected Serializer getDefaultSerializer(Class cl) {
    
  if (_defaultSerializer != null)
    return _defaultSerializer;

  //  If the serialized object is not implemented  Serializable  Interface , It will be thrown out. IllegalStateException
  if (! Serializable.class.isAssignableFrom(cl)
      && ! _isAllowNonSerializable) {
    
    throw new IllegalStateException("Serialized class " + cl.getName() + " must implement java.io.Serializable");
  }

  if (_isEnableUnsafeSerializer
      && JavaSerializer.getWriteReplace(cl) == null) {
    
    return UnsafeSerializer.create(cl);
  }
  else
    return JavaSerializer.create(cl);
}

2.2.3 Serialization process and deserialization process

• Serialization process ：Hessian2Output -> SerializerFactory -> Serializer
• Deserialization process ：Hessian2Input -> SerializerFactory -> Deserializer

2.3 summary

• Serialize objects to achieve Serializable Interface , Otherwise, it will report “must implement java.io.Serializable” abnormal
• If the serialized object passes hessian After serialization , Serialized object without serialVersionUID when , Re change （ Add object properties 、 Delete object properties ） Will not generate deserialization exceptions , namely hessian Serializing objects no longer requires serialVersionUID.
• hessian It stores all the attributes of a complex object in one Map Serialization in . So in the parent class 、 When a subclass has a member variable with the same name , Hessian Serialization , Serialize subclasses first , Then serialize the parent class , Therefore, the result of deserialization will cause the member variable with the same name of the child class to be overwritten by the value of the parent class .
• hessian Medium writeReplace Methods and readResolve The method works the same , If the serialized object has this method , The instance returned by this method will be used to replace the serialized instance , To ensure the singleness of the object .