一、基本案例

1、new String("helloworld") 与 "helloworld"

public static void main(String[] args) throws Exception {
        String s0 = new String("helloworld");
        String s1 = s0.intern(); // 此时"helloworld"已经存在常量池中,现在只是通过intern方法取出而已
        String s2 = "helloworld"; // 
        System.out.println(s0 == s1); // false
        System.out.println(s2 == s1); // true
}

// 学过java编译过程的都知道编译会进行热点代码的优化,如:方法内联、常量传播、空值检查消除、寄存器分配等等,热点代码一般通过热点探测得出,而HotSpotIntrinsicCandidate注解能够直接手段将某个方法直接指定为热点代码,jvm尽快优化它(非绝对优化,优化时机不确定)。

// 注释简言之:new String("helloworld") 是"helloworld"的一个复制;因为String是不可变的,除非需要显示复制"hellworld",不然使用构造器来复制字符串是不必要的。

/**
* Initializes a newly created {@code String} object so that it represents
* the same sequence of characters as the argument; in other words, the
* newly created string is a copy of the argument string. Unless an
* explicit copy of {@code original} is needed, use of this constructor is
* unnecessary since Strings are immutable.
*
* @param  original
*         A {@code String}
*/
@HotSpotIntrinsicCandidate
public String(String original) {
    this.value = original.value;
    this.coder = original.coder;
    this.hash = original.hash;
}
public static void main(String[] args) throws Exception {
        String s0 = new String("helloworld");
        String s1 = s0.intern(); // 此时"helloworld"已经存在常量池中,现在只是通过intern方法取出而已
        String s2 = "helloworld"; // 
        System.out.println(s0 == s1); // false
        System.out.println(s2 == s1); // true
}

// 学过java编译过程的都知道编译会进行热点代码的优化,如:方法内联、常量传播、空值检查消除、寄存器分配等等,热点代码一般通过热点探测得出,而HotSpotIntrinsicCandidate注解能够直接手段将某个方法直接指定为热点代码,jvm尽快优化它(非绝对优化,优化时机不确定)。

// 注释简言之:new String("helloworld") 是"helloworld"的一个复制;因为String是不可变的,除非需要显示复制"hellworld",不然使用构造器来复制字符串是不必要的。

/**
* Initializes a newly created {@code String} object so that it represents
* the same sequence of characters as the argument; in other words, the
* newly created string is a copy of the argument string. Unless an
* explicit copy of {@code original} is needed, use of this constructor is
* unnecessary since Strings are immutable.
*
* @param  original
*         A {@code String}
*/
@HotSpotIntrinsicCandidate
public String(String original) {
    this.value = original.value;
    this.coder = original.coder;
    this.hash = original.hash;
}
(1)对比new String("helloworld")与"helloworld"的反编译源码
// 对应的反编译源码
public static void main(java.lang.String[]) throws java.lang.Exception;
    Code:
       // new 创建一个对象,并将其引用值压入栈顶
       0: new           #2                  // class java/lang/String
       // 复制栈顶数值(数值不能是long或double类型的)并将复制值压入栈顶
       3: dup
       // 从常量池中取出字面量(常量值)
       4: ldc           #3                  // String helloworld
       6: invokespecial #4                  // Method java/lang/String."<init>":(Ljava/lang/String;)V
       9: astore_1
      10: aload_1
      11: invokevirtual #5                  // Method java/lang/String.intern:()Ljava/lang/String;
      14: astore_2
      15: ldc           #3                  // String helloworld
      17: astore_3
      18: getstatic     #6                  // Field java/lang/System.out:Ljava/io/PrintStream;
      21: aload_1
      22: aload_2
      23: if_acmpne     30
      26: iconst_1
      27: goto          31
      30: iconst_0
      31: invokevirtual #7                  // Method java/io/PrintStream.println:(Z)V
      34: getstatic     #6                  // Field java/lang/System.out:Ljava/io/PrintStream;
      37: aload_3
      38: aload_2
      39: if_acmpne     46
      42: iconst_1
      43: goto          47
      46: iconst_0
      47: invokevirtual #7                  // Method java/io/PrintStream.println:(Z)V
      50: return
}

// String s0 = new String("helloworld") 的反编译源码
0: new           #2                  // class java/lang/String
3: dup
4: ldc           #3                  // String helloworld
6: invokespecial #4                  // Method java/lang/String."<init>":(Ljava/lang/String;)V
9: astore_1

//  String s2 = "helloworld" 的反编译源码 
15: ldc           #3                 // String helloworld
17: astore_3

// 对比可知,"helloworld"不会在堆中创建对象,即不调用new指令和String的构造器方法
// 但是只要看到"helloworld",就会在'静态常量池'中生成"helloworld"字面量,后面还有个例子可以比较看看。
// 对应的反编译源码
public static void main(java.lang.String[]) throws java.lang.Exception;
    Code:
       // new 创建一个对象,并将其引用值压入栈顶
       0: new           #2                  // class java/lang/String
       // 复制栈顶数值(数值不能是long或double类型的)并将复制值压入栈顶
       3: dup
       // 从常量池中取出字面量(常量值)
       4: ldc           #3                  // String helloworld
       6: invokespecial #4                  // Method java/lang/String."<init>":(Ljava/lang/String;)V
       9: astore_1
      10: aload_1
      11: invokevirtual #5                  // Method java/lang/String.intern:()Ljava/lang/String;
      14: astore_2
      15: ldc           #3                  // String helloworld
      17: astore_3
      18: getstatic     #6                  // Field java/lang/System.out:Ljava/io/PrintStream;
      21: aload_1
      22: aload_2
      23: if_acmpne     30
      26: iconst_1
      27: goto          31
      30: iconst_0
      31: invokevirtual #7                  // Method java/io/PrintStream.println:(Z)V
      34: getstatic     #6                  // Field java/lang/System.out:Ljava/io/PrintStream;
      37: aload_3
      38: aload_2
      39: if_acmpne     46
      42: iconst_1
      43: goto          47
      46: iconst_0
      47: invokevirtual #7                  // Method java/io/PrintStream.println:(Z)V
      50: return
}

// String s0 = new String("helloworld") 的反编译源码
0: new           #2                  // class java/lang/String
3: dup
4: ldc           #3                  // String helloworld
6: invokespecial #4                  // Method java/lang/String."<init>":(Ljava/lang/String;)V
9: astore_1

//  String s2 = "helloworld" 的反编译源码 
15: ldc           #3                 // String helloworld
17: astore_3

// 对比可知,"helloworld"不会在堆中创建对象,即不调用new指令和String的构造器方法
// 但是只要看到"helloworld",就会在'静态常量池'中生成"helloworld"字面量,后面还有个例子可以比较看看。

关于JVM指令,请参考:

(2)通过Classpy查看静态常量池:

关于Classpy,请参考:https://github.com/zxh0/classpy

StringDemo.class
    magic:0xCAFEBABE
    ...
    ...
    constant_pool:
        #01 (Methodref): java/lang/Object.<init>
        #02 (Class): java/lang/String
        // 符号引用
        #03 (String): helloworld
            tag:8
            string_index:26
        ...
        ...
        // 字面量
        #26 (Utf8):hellowrold
            tag:1
            length:10
            bytes:helloworld
(3)intern方法(返回字符串的引用)
// 注释简言之:intern会判断字符串常量池是否拥有该字符串对象,拥有则返回,反之添加到常量池并返回该字符串对象的引用。
// 其中所有的字面量字符串和字符数值的常量表达式(请参考java规范)都会被常量池保存起来。
// 后面再详细讲解intern方法的内部实现

/**
* Returns a canonical representation for the string object.
* <p>
* A pool of strings, initially empty, is maintained privately by the
* class {@code String}.
* <p>
* When the intern method is invoked, if the pool already contains a
* string equal to this {@code String} object as determined by
* the {@link #equals(Object)} method, then the string from the pool is
* returned. Otherwise, this {@code String} object is added to the
* pool and a reference to this {@code String} object is returned.
* <p>
* It follows that for any two strings {@code s} and {@code t},
* {@code s.intern() == t.intern()} is {@code true}
* if and only if {@code s.equals(t)} is {@code true}.
* <p>
* All literal strings and string-valued constant expressions are
* interned. String literals are defined in section 3.10.5 of the
* <cite>The Java™ Language Specification</cite>.
*
* @return  a string that has the same contents as this string, but is
*          guaranteed to be from a pool of unique strings.
* @jls 3.10.5 String Literals
*/
public native String intern();
// 注释简言之:intern会判断字符串常量池是否拥有该字符串对象,拥有则返回,反之添加到常量池并返回该字符串对象的引用。
// 其中所有的字面量字符串和字符数值的常量表达式(请参考java规范)都会被常量池保存起来。
// 后面再详细讲解intern方法的内部实现

/**
* Returns a canonical representation for the string object.
* <p>
* A pool of strings, initially empty, is maintained privately by the
* class {@code String}.
* <p>
* When the intern method is invoked, if the pool already contains a
* string equal to this {@code String} object as determined by
* the {@link #equals(Object)} method, then the string from the pool is
* returned. Otherwise, this {@code String} object is added to the
* pool and a reference to this {@code String} object is returned.
* <p>
* It follows that for any two strings {@code s} and {@code t},
* {@code s.intern() == t.intern()} is {@code true}
* if and only if {@code s.equals(t)} is {@code true}.
* <p>
* All literal strings and string-valued constant expressions are
* interned. String literals are defined in section 3.10.5 of the
* <cite>The Java™ Language Specification</cite>.
*
* @return  a string that has the same contents as this string, but is
*          guaranteed to be from a pool of unique strings.
* @jls 3.10.5 String Literals
*/
public native String intern();

java规范的下载地址:https://docs.oracle.com/javase/specs/index.html

2、new String("hello") + new String("world") 与 “helloworld”

public static void main(String[] args) throws Exception {
        // new StringBuilder().append("hello").append("world").toString();
        String s0 = new String("hello") + new String("world");
        String s1 = s0.intern(); // 此时常量池没有helloworld,此时放入,放入的是s0的地址。
        String s2 = "helloworld"; // 从常量池中取出s0的地址
        System.out.println(s0 == s1); // true
        System.out.println(s2 == s1); // true
}

public static void main(String[] args) throws Exception {
        String s2 = "helloworld";
        String s0 = new String("hello") + new String("world");
        String s1 = s0.intern(); // 此时常量池有helloworld,直接取出,为s2的地址。
        System.out.println(s0 == s1); // false
        System.out.println(s2 == s1); // true
}

// 此例证明new StringBuilder().append("hello").append("world").toString()没有intern的功能
public static void main(String[] args) throws Exception {
        String s0 = new String("hello") + new String("world");
        String s2 = "helloworld"; // 具有intern的功能
        System.out.println(s2 == s0); // false
}


// StringBuilder的toString方法
@Override
@HotSpotIntrinsicCandidate
public String toString() {
    // Create a copy, don't share the array
    return isLatin1() ? StringLatin1.newString(value, 0, count): StringUTF16.newString(value, 0, count);
}

// StringLatin1.newString方法 ==> String的重载构造器
public static String newString(byte[] val, int index, int len) {
    return new String(Arrays.copyOfRange(val, index, index + len),LATIN1);
}   
public static void main(String[] args) throws Exception {
        // new StringBuilder().append("hello").append("world").toString();
        String s0 = new String("hello") + new String("world");
        String s1 = s0.intern(); // 此时常量池没有helloworld,此时放入,放入的是s0的地址。
        String s2 = "helloworld"; // 从常量池中取出s0的地址
        System.out.println(s0 == s1); // true
        System.out.println(s2 == s1); // true
}

public static void main(String[] args) throws Exception {
        String s2 = "helloworld";
        String s0 = new String("hello") + new String("world");
        String s1 = s0.intern(); // 此时常量池有helloworld,直接取出,为s2的地址。
        System.out.println(s0 == s1); // false
        System.out.println(s2 == s1); // true
}

// 此例证明new StringBuilder().append("hello").append("world").toString()没有intern的功能
public static void main(String[] args) throws Exception {
        String s0 = new String("hello") + new String("world");
        String s2 = "helloworld"; // 具有intern的功能
        System.out.println(s2 == s0); // false
}


// StringBuilder的toString方法
@Override
@HotSpotIntrinsicCandidate
public String toString() {
    // Create a copy, don't share the array
    return isLatin1() ? StringLatin1.newString(value, 0, count): StringUTF16.newString(value, 0, count);
}

// StringLatin1.newString方法 ==> String的重载构造器
public static String newString(byte[] val, int index, int len) {
    return new String(Arrays.copyOfRange(val, index, index + len),LATIN1);
}
(1)对应的反编译源码
public static void main(java.lang.String[]) throws java.lang.Exception;
    Code:
       0: new           #2                  // class java/lang/StringBuilder
       3: dup
       4: invokespecial #3                  // Method java/lang/StringBuilder."<init>":()V
       7: new           #4                  // class java/lang/String
      10: dup
      11: ldc           #5                  // String hello
      13: invokespecial #6                  // Method java/lang/String."<init>":(Ljava/lang/String;)V
      16: invokevirtual #7                  // Method java/lang/StringBuilder.append:(Ljava/lang/String;)Ljava/lang/StringBuilder;
      19: new           #4                  // class java/lang/String
      22: dup
      23: ldc           #8                  // String world
      25: invokespecial #6                  // Method java/lang/String."<init>":(Ljava/lang/String;)V
      28: invokevirtual #7                  // Method java/lang/StringBuilder.append:(Ljava/lang/String;)Ljava/lang/StringBuilder;
      31: invokevirtual #9                  // Method java/lang/StringBuilder.toString:()Ljava/lang/String;
      34: astore_1
      35: aload_1
      36: invokevirtual #10                 // Method java/lang/String.intern:()Ljava/lang/String;
      39: astore_2
      40: ldc           #11                 // String helloworld
      42: astore_3
      43: getstatic     #12                 // Field java/lang/System.out:Ljava/io/PrintStream;
      46: aload_1
      47: aload_2
      48: if_acmpne     55
      51: iconst_1
      52: goto          56
      55: iconst_0
      56: invokevirtual #13                 // Method java/io/PrintStream.println:(Z)V
      59: getstatic     #12                 // Field java/lang/System.out:Ljava/io/PrintStream;
      62: aload_3
      63: aload_2
      64: if_acmpne     71
      67: iconst_1
      68: goto          72
      71: iconst_0
      72: invokevirtual #13                 // Method java/io/PrintStream.println:(Z)V
      75: return
}
(2)通过Classpy查看静态常量池:
  • 检查静态常量池可知,只有"hello"和"world",但是上述结果为true,代表执行intern时,运行常量池中放入了s0指向的地址。
(3)通过HSDB查看一下main线程的状态
  • 代码通过jdb打断点停止在 String s1 = s0.intern();这一行。
  • cmd中,输入java -classpath "%JAVA_HOME%/lib/sa-jdi.jar" sun.jvm.hotspot.HSDB
  • 点击右上角File -> Attach Hotspot process
  • 点击Java Threads的main线程,点击Stack Memory...(左上角第二个)
  • 查看Stack Memory for main,看到PSYoungGen java/langString 的地址:0x00000000d61515b8
  • 点击Windows -> Console,输入inspect 0x00000000d61515b8,返回了instance of [C @ 0x00000000d61515d0 @ 0x00000000d61515d0 (size = 40)
  • Console Line中,继续输入inspect 0x00000000d61515d0 ,得到"helloworld",代表new StringBuilder().toString()不会再常量池中放入字符串的引用。

resultType怎么返回json_常量池

  • 断开HSDB的连接,代码通过jdb继续运行至System.out.println(s0 == s1);该行(HSDB连接时线程将阻塞,gdb无法直接调试,需要断开调试,再连接查看)
  • 连接HSDB到虚拟机,明显可以看出main线程中的3个本地变量指向同一个字符串,对应s0、s1、s2.

resultType怎么返回json_常量池_02

HSDB,请参考:

GBD,请参考:

二、String.intern()源码

由于intern方法是native方法,采用了JNI技术。

关于JNI技术,请参考:

为了理解简单,下面源码省略了加锁、内存管理和重哈希的代码,感兴趣可以下载HotSpot的源码进行研读。

通过源码可以知道,String的常量池其实就是C++版本的HashMap而已。

下载源码,请参考:

// \openjdk10\jdk\src\share\native\java\lang\String.c
// 第二个参数为返回值
JNIEXPORT jobject JNICALL
// jni命名规范(声明为native自动生成):java.lang.String:intern => Java_java_lang_String_intern(Java前缀 + 包名 + 方法名,分隔符号采用_)
Java_java_lang_String_intern(JNIEnv *env, jobject this){
    //(1)JVM_InternString调用
    return JVM_InternString(env, this);
}

// \openjdk10\hotspot\src\share\vm\prims\jvm.h
/*
 * java.lang.String
 */
JNIEXPORT jstring JNICALL
JVM_InternString(JNIEnv *env, jstring str);

// \openjdk10\hotspot\src\share\vm\prims\jvm.cpp
// String support ///
// (2)JVM_InternString的实现
JVM_ENTRY(jstring, JVM_InternString(JNIEnv *env, jstring str))
  JVMWrapper("JVM_InternString");
  JvmtiVMObjectAllocEventCollector oam;
  if (str == NULL) return NULL;
  oop string = JNIHandles::resolve_non_null(str);
  // (3)StringTable::intern调用
  oop result = StringTable::intern(string, CHECK_NULL);
  return (jstring) JNIHandles::make_local(env, result);
JVM_END

// \openjdk10\hotspot\src\share\vm\classfile\stringTable.cpp 
// (4)StringTable::intern的实现 StringTable是HashTable的子类
oop StringTable::intern(oop string, TRAPS){
  if (string == NULL) return NULL;
  int length;
  Handle h_string (THREAD, string); // 创建Handle
  jchar* chars = java_lang_String::as_unicode_string(string, length, CHECK_NULL);
  // (5)StringTable::intern的重载方法
  return intern(h_string, chars, length, CHECK_NULL);
}

// \openjdk10\hotspot\src\share\vm\classfile\stringTable.cpp
oop StringTable::intern(Handle string_or_null, jchar* name,int len, TRAPS) {
  // shared table always uses java_lang_String::hash_code
  // 个人理解: java_lang_String属于工具类,提供一些操作string的方法
  unsigned int hashValue = java_lang_String::hash_code(name, len);
  // (6)查询共享数组
  oop found_string = lookup_shared(name, len, hashValue);
  if (found_string != NULL) return found_string;
 
  // the_table()返回StringTable的引用
  int index = the_table() -> hash_to_index(hashValue);// 其实就是hashValue % _table_size
  found_string = the_table() -> lookup_in_main_table(index, name, len, hashValue);
  if (found_string != NULL) return found_string;

  Handle string;
  if (!string_or_null.is_null()) string = string_or_null;
  else string = java_lang_String::create_from_unicode(name, len, CHECK_NULL);
    
  // 前面常量池存在该字符串就返回了,不存在则进行添加操作
  oop added_or_found = the_table()->basic_add(index,string,name,len,hashValue,CHECK_NULL);
  return added_or_found;
}

// \openjdk10\hotspot\src\share\vm\classfile\javaClasses.cpp
// hash_code的实现,与jdk源码String类的HashCode()方法类似
unsigned int java_lang_String::hash_code(oop java_string) {
  int length = java_lang_String::length(java_string);
  if (length == 0) return 0;

  typeArrayOop value  = java_lang_String::value(java_string);
  bool is_latin1 = java_lang_String::is_latin1(java_string);

  if (is_latin1) { 
    // \openjdk10\hotspot\src\share\vm\classfile\javaClasses.hpp中static修饰的类方法
    return java_lang_String::hash_code(value->byte_at_addr(0), length);
  } else {
    // \openjdk10\hotspot\src\share\vm\classfile\javaClasses.hpp中static修饰的类方法
    return java_lang_String::hash_code(value->char_at_addr(0), length);
  }
}

// \openjdk10\hotspot\src\share\vm\classfile\javaClasses.hpp
static unsigned int hash_code(const jbyte* s, int len) {
    unsigned int h = 0;
    while (len-- > 0) {
      h = 31*h + (((unsigned int) *s) & 0xFF);
      s++;
    }
    return h;
}

// package java.lang.StringLatin1类中的hashCode与Openjdk中保持一致
public static int hashCode(byte[] value) {
    int h = 0;
    for (byte v : value) {
        h = 31 * h + (v & 0xff);
    }
    return h;
}

// \openjdk10\hotspot\src\share\vm\classfile\stringTable.cpp
oop StringTable::lookup_shared(jchar* name, int len, unsigned int hash) {
    //(7)共享数组是一个HashTable的子类, CompactHashtable<oop, char> StringTable::_shared_table;
    return _shared_table.lookup((const char*)name, hash, len);
}

// \openjdk10\hotspot\src\share\vm\classfile\compactHashtable.inline.hpp
template <class T, class N>
inline T CompactHashtable<T,N>::lookup(const N* name, unsigned int hash, int len) {
  if (_entry_count > 0) {// 
    int index = hash % _bucket_count;// _bucket_count为_buckets数组大小
    u4 bucket_info = _buckets[index];// bucket_info为32位,高2位代表类型,低30为代表偏移量
    u4 bucket_offset = BUCKET_OFFSET(bucket_info);// 取出低30位
    int bucket_type = BUCKET_TYPE(bucket_info);// 取出高2位
    u4* entry = _entries + bucket_offset;// 根据偏移量取出entries数组中值
    
    if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
      // 只存值的entry,包含一个偏移量
      T res = decode_entry(this, entry[0], name, len);// 获取存放的值,代码就不贴了
      if (res != NULL) return res;
    } else {
      // This is a regular bucket, which has more than one
      // entries. Each entry is a pair of entry (hash, offset).
      // Seek until the end of the bucket.
        
      // 常规bucket,索引0放着hash值,索引1放着偏移量
      u4* entry_max = _entries + BUCKET_OFFSET(_buckets[index + 1]);// 获取下一个_buckets的偏移量作为寻找entry的最大值
      while (entry < entry_max) {
        if ((unsigned int)(entry[0]) == hash) {
          T res = decode_entry(this, entry[1], name, len);
          if (res != NULL) return res;
        }
        entry += 2;
      }
    }
  }
  return NULL;
}

// \openjdk10\hotspot\src\share\vm\classfile\stringTable.cpp
oop StringTable::lookup_in_main_table(int index, jchar* name,int len, unsigned int hash) {
  // bucket方法位于hashtable.inline.hpp中,与java的HashMap类似,取出HashtableEntry,类比Map.Entry.单向链表形式。
  // hash碰撞导致index相同,存放形式为链表。所以需要取出来对比hash值和内部值是否相等。
  // bucket(index) ==>  _buckets[i].get_entry();
  for (HashtableEntry<oop, mtSymbol>* l = bucket(index); l != NULL; l = l->next()) {
    // hash方法 ==> unsigned int hash() const { return _hash; }
    if (l->hash() == hash) {
      // literal方法取出oop,即String字面量 ==> T literal() const { return _literal;}
      if (java_lang_String::equals(l->literal(), name, len)) return l->literal();
    }
  }
  return NULL;
}

// \openjdk10\hotspot\src\share\vm\classfile\stringTable.cpp
oop StringTable::basic_add(int index_arg, Handle string, jchar* name,int len, unsigned int hashValue_arg, TRAPS) {
  unsigned int hashValue = hashValue_arg;
  int index = index_arg;

  oop test = lookup_in_main_table(index, name, len, hashValue);
  if (test != NULL)  return test;

  // \openjdk10\hotspot\src\share\vm\utilities\hashtable.cpp
  // StringTable继承了HashTable,()是Handle的运算符重载,返回string的对象值
  HashtableEntry<oop, mtSymbol>* entry = new_entry(hashValue, string());
  add_entry(index, entry);
  return string();
}

// \openjdk10\hotspot\src\share\vm\utilities\hashtable.inline.cpp
template <MEMFLAGS F> inline void BasicHashtable<F>::add_entry(int index, BasicHashtableEntry<F>* entry) {
  entry->set_next(bucket(index));
  _buckets[index].set_entry(entry);
  ++_number_of_entries;
}

// \openjdk10\hotspot\src\share\vm\runtime\handles.hpp
class Handle VALUE_OBJ_CLASS_SPEC {
    private: 
        oop* _handle;
    protected:
        oop obj() const { return _handle == NULL ? (oop)NULL : *_handle; }// ()运算符重载
}