菜瓜:水稻,这次我特意去看了java的循环依赖
水稻:哟,有什么收获
菜瓜:两种情况,构造器循环依赖,属性循环依赖
- 构造器循环依赖在逻辑层面无法通过。对象通过构造函数创建时如果需要创建另一个对象,就会存在递归调用。栈内存直接溢出
- 属性循环依赖可以解决。在对象创建完成之后通过属性赋值操作。
package club.interview.base; /**
* 构造器循环依赖 - Exception in thread "main" java.lang.*Error
* toString()循环打印也会异常 - Exception in thread "main" java.lang.*Error
* @author QuCheng on 2020/6/18.
*/
public class Circular { class A {
B b; // public A() {
// b = new B();
// } // @Override
// public String toString() {
// return "A{" +
// "b=" + b +
// '}';
// }
} class B {
A a; // public B() {
// a = new A();
// } // @Override
// public String toString() {
// return "B{" +
// "a=" + a +
// '}';
// }
} private void test() {
B b = new B();
A a = new A();
a.b = b;
b.a = a;
System.out.println(a);
System.out.println(b);
} public static void main(String[] args) {
new Circular().test();
}
}
水稻:厉害啊,Spring也不支持构造函数的依赖注入,而且也不支持多例的循环依赖。同样的,它支持属性的依赖注入。
- 看效果 - 如果toString()打印同样会出现栈内存溢出。
package com.vip.qc.circular; import org.springframework.stereotype.Component; import javax.annotation.Resource; /**
* @author QuCheng on 2020/6/18.
*/
@Component("a")
public class CircularA { @Resource
private CircularB circularB; // @Override
// public String toString() {
// return "CircularA{" +
// "circularB=" + circularB +
// '}';
// }
} package com.vip.qc.circular; import org.springframework.stereotype.Component; import javax.annotation.Resource; /**
* @author QuCheng on 2020/6/18.
*/
@Component("b")
public class CircularB { @Resource
private CircularA circularA; // @Override
// public String toString() {
// return "CircularB{" +
// "circularA=" + circularA +
// '}';
// }
} @Test
public void testCircular() {
String basePackages = "com.vip.qc.circular";
new AnnotationConfigApplicationContext(basePackages);
}
菜瓜:看来spring的实现应该也是通过属性注入的吧
水稻:你说的对。先给思路和demo,之后带你扫一遍源码,follow me !
- spring的思路是给已经初始化的bean标记状态,假设A依赖B,B依赖A,先创建A
- 先从缓存容器(总共三层,一级拿不到就拿二级,二级拿不到就从三级缓存中拿正在创建的)中获取A,未获取到就执行创建逻辑
- 对象A在创建完成还未将属性渲染完之前标记为正在创建中,放入三级缓存容器。渲染属性populateBean()会获取依赖的对象B。
- 此时B会走一次getBean逻辑,B同样会先放入三级缓存,然后渲染属性,再次走getBean逻辑注入A,此时能从三级缓存中拿到A,并将A放入二级容器。B渲染完成放入一级容器
- 回到A渲染B的方法populateBean(),拿到B之后能顺利执行完自己的创建过程。放入一级缓存
-
为了证实结果,我把源码给改了一下,看结果
-
package com.vip.qc.circular; import org.springframework.stereotype.Component; import javax.annotation.Resource; /**
* @author QuCheng on 2020/6/18.
*/
@Component("a")
public class CircularA { @Resource
private CircularB circularB; @Override
public String toString() {
return "CircularA{" +
"circularB=" + circularB +
'}';
}
} package com.vip.qc.circular; import org.springframework.stereotype.Component; import javax.annotation.Resource; /**
* @author QuCheng on 2020/6/18.
*/
@Component("b")
public class CircularB { @Resource
private CircularA circularA; @Override
public String toString() {
return "CircularB{" +
"circularA=" + circularA +
'}';
}
} 测试代码
@Test
public void testCircular() {
String basePackages = "com.vip.qc.circular";
new AnnotationConfigApplicationContext(basePackages);
} 测试结果(我改过源码了)
----
将a放入三级缓存
将b放入三级缓存
将a放入二级缓存
将b放入一级缓存
从二级缓存中拿到了a
将a放入一级缓存
-
- 再看源码
- 关键类处理getSingleton逻辑 - 缓存容器
public class DefaultSingletonBeanRegistry /** Cache of singleton objects: bean name to bean instance. */
// 一级缓存
private final Map<String, Object> singletonObjects = new ConcurrentHashMap<>(256); /** Cache of singleton factories: bean name to ObjectFactory. */
// 三级缓存
private final Map<String, ObjectFactory<?>> singletonFactories = new HashMap<>(16); /** Cache of early singleton objects: bean name to bean instance. */
// 二级缓存
private final Map<String, Object> earlySingletonObjects = new HashMap<>(16);
- 主流程 AbstractApplicationContext#refresh() -> DefaultListableBeanFactory#preInstantiateSingletons() -> AbstractBeanFactory#getBean() & #doGetBean()
protected <T> T doGetBean(final String name, @Nullable final Class<T> requiredType,
@Nullable final Object[] args, boolean typeCheckOnly) throws BeansException {
/**
* 处理FactoryBean接口名称转换 {@link BeanFactory#FACTORY_BEAN_PREFIX }
*/
final String beanName = transformedBeanName(name);
...
// ①从缓存中拿对象(如果对象正在创建中且被依赖注入,会放入二级缓存)
Object sharedInstance = getSingleton(beanName);
if (sharedInstance != null && args == null) {
...
}else {
...
if (mbd.isSingleton()) {
// ② 将创建的对象放入一级缓存
sharedInstance = getSingleton(beanName, () -> {
try {
// ③ 具体创建的过程,每个bean创建完成之后都会放入三级缓存,然后渲染属性
return createBean(beanName, mbd, args);
}catch (BeansException ex) {
...
...
return (T) bean;
}
- ①getSingleton(beanName) - 二级缓存操作
protected Object getSingleton(String beanName, boolean allowEarlyReference) {
// 实例化已经完成了的放在singletonObjects
Object singletonObject = this.singletonObjects.get(beanName);
// 解决循环依赖
if (singletonObject == null && isSingletonCurrentlyInCreation(beanName)) {
synchronized (this.singletonObjects) {
singletonObject = this.earlySingletonObjects.get(beanName);
if (singletonObject == null && allowEarlyReference) {
ObjectFactory<?> singletonFactory = this.singletonFactories.get(beanName);
if (singletonFactory != null) {
singletonObject = singletonFactory.getObject();
this.earlySingletonObjects.put(beanName, singletonObject);
if(beanName.equals("a")||beanName.equals("b")||beanName.equals("c"))
System.out.println("将"+beanName+"放入二级缓存");;
this.singletonFactories.remove(beanName);
}
}else if(singletonObject != null){
System.out.println("从二级缓存中拿到了"+beanName);
}
}
}
return singletonObject;
}
- ② getSingleton(beanName,lamdba) - 一级缓存操作
public Object getSingleton(String beanName, ObjectFactory<?> singletonFactory) {
Assert.notNull(beanName, "Bean name must not be null");
synchronized (this.singletonObjects) {
Object singletonObject = this.singletonObjects.get(beanName);
if (singletonObject == null) {
if (this.singletonsCurrentlyInDestruction) {
...
// 正在创建的bean加入singletonsCurrentlyInCreation - 保证只有一个对象创建,阻断循环依赖
beforeSingletonCreation(beanName);
...
try {
singletonObject = singletonFactory.getObject();
...
finally {
...
// 从singletonsCurrentlyInCreation中移除
afterSingletonCreation(beanName);
}
if (newSingleton) {
// 对象创建完毕 - 放入一级缓存(从其他缓存移除)
addSingleton(beanName, singletonObject);
}
}
return singletonObject;
}
} // ----- 内部调用一级缓存操作
protected void addSingleton(String beanName, Object singletonObject) {
synchronized (this.singletonObjects) {
if(beanName.equals("a")||beanName.equals("b")||beanName.equals("c"))
System.out.println("将"+beanName+"放入一级缓存");;
this.singletonObjects.put(beanName, singletonObject);
this.singletonFactories.remove(beanName);
this.earlySingletonObjects.remove(beanName);
this.registeredSingletons.add(beanName);
}
}
- ③createBean(beanName, mbd, args) - 三级缓存操作
protected Object doCreateBean(final String beanName, final RootBeanDefinition mbd, final @Nullable Object[] args)throws BeanCreationException {
...
if (instanceWrapper == null) {
// 5* 实例化对象本身
instanceWrapper = createBeanInstance(beanName, mbd, args);
}
...
boolean earlySingletonExposure = (mbd.isSingleton() && this.allowCircularReferences &&
isSingletonCurrentlyInCreation(beanName));
if (earlySingletonExposure) {
...
// 将创建好还未渲染属性的bean 放入三级缓存
addSingletonFactory(beanName, () -> getEarlyBeanReference(beanName, mbd, bean));
} Object exposedObject = bean;
try {
// 渲染bean自身和属性
populateBean(beanName, mbd, instanceWrapper);
// 实例化之后的后置处理 - init
exposedObject = initializeBean(beanName, exposedObject, mbd);
}
catch (Throwable ex) {
...
return exposedObject;
} // ------------- 内部调用三级缓存操作
protected void addSingletonFactory(String beanName, ObjectFactory<?> singletonFactory) {
Assert.notNull(singletonFactory, "Singleton factory must not be null");
synchronized (this.singletonObjects) {
if (!this.singletonObjects.containsKey(beanName)) {
if(beanName.equals("a")||beanName.equals("b")||beanName.equals("c"))
System.out.println("将"+beanName+"放入三级缓存");;
this.singletonFactories.put(beanName, singletonFactory);
this.earlySingletonObjects.remove(beanName);
this.registeredSingletons.add(beanName);
}
}
}
- 关键类处理getSingleton逻辑 - 缓存容器
菜瓜:demo比较简单,流程大致明白,源码我还需要斟酌一下,整体有了概念。这个流程好像是掺杂在bean的创建过程中,结合bean的生命周期整体理解可能会更深入一点
水稻:是的。每个知识点都不是单一的,拿着bean的生命周期再理解一遍可能会更有收获。
讨论
- 为什么是三级缓存,两级不行吗?
- 猜测:理论上两级也可以实现。多一个二级缓存可能是为了加快获取的速度。加入A依赖B,A依赖C,B依赖A,C依赖A,那么C在获取A的时候只需要从二级缓存中就能拿到A了
总结
- Spring的处理方式和java处理的思想一致,构造器依赖本身是破坏语义和规范的
- 属性赋值--> 依赖注入 。 先创建对象,再赋值属性,赋值的时候发现需要创建便生成依赖对象,被依赖对象需要前一个对象就从缓存容器中拿取即可