Spring @Async 的使用与实现的示例代码

时间:2022-09-23 08:26:16

首先Spring AOP有两个重要的基础接口,Advisor和PointcutAdvisor,接口声明如下:

Advisor接口声明:

?
1
2
3
4
5
public interface Advisor {
  Advice getAdvice();
  boolean isPerInstance();
 
}

PointcutAdvisor的接口声明:

?
1
2
3
4
5
6
7
8
public interface PointcutAdvisor extends Advisor {
 
  /**
   * Get the Pointcut that drives this advisor.
   */
  Pointcut getPointcut();
 
}

PointcutAdvisor用来获取一个切点以及这个切点的处理器(Advise)。

@Async注解使用后置处理器BeanPostProcessor的子类AsyncAnnotationBeanPostProcessor来实现bean处理 :

AsyncAnnotationAdvisor继承了PointcutAdvisor接口。并且在AsyncAnnotationBeanPostProcessor实现了其父类接口的BeanFactoryAware中的setBeanFactory初始化。Spring一旦创建beanFactory回调成功,就会回调这个方法。保证Advisor对象最先被初始化。

?
1
2
3
4
5
6
7
8
9
10
11
12
13
  @Override
  public void setBeanFactory(BeanFactory beanFactory) {
    super.setBeanFactory(beanFactory);
 
    AsyncAnnotationAdvisor advisor = new AsyncAnnotationAdvisor(this.executor, this.exceptionHandler);
    if (this.asyncAnnotationType != null) {
      advisor.setAsyncAnnotationType(this.asyncAnnotationType);
    }
    advisor.setBeanFactory(beanFactory);
    this.advisor = advisor;
  }
 
}

具体的后置处理是通过AsyncAnnotationBeanPostProcessor的后置bean处理是通过其父类AbstractAdvisingBeanPostProcessor来实现的。AbstractAdvisingBeanPostProcessor提供的后置bean处理方法对所有的自定义注解的bean处理方法时通用的。其具体的代码如下:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
@Override
  public Object postProcessAfterInitialization(Object bean, String beanName) {
    if (bean instanceof AopInfrastructureBean) {
      // Ignore AOP infrastructure such as scoped proxies.
      return bean;
    }
   
   /*
   
   * bean对象如果是一个ProxyFactory对象。ProxyFactory继承了AdvisedSupport,而    AdvisedSupport又继承了Advised接口。这个时候就把不同的Advisor添加起来。
   *
    if (bean instanceof Advised) {
      Advised advised = (Advised) bean;
      if (!advised.isFrozen() && isEligible(AopUtils.getTargetClass(bean))) {
        // Add our local Advisor to the existing proxy's Advisor chain...
        if (this.beforeExistingAdvisors) {
          advised.addAdvisor(0, this.advisor);
        }
        else {
          advised.addAdvisor(this.advisor);
        }
        return bean;
      }
    }
 
    if (isEligible(bean, beanName)) {
      ProxyFactory proxyFactory = prepareProxyFactory(bean, beanName);
      if (!proxyFactory.isProxyTargetClass()) {
        evaluateProxyInterfaces(bean.getClass(), proxyFactory);
      }
      proxyFactory.addAdvisor(this.advisor);
      customizeProxyFactory(proxyFactory);
      return proxyFactory.getProxy(getProxyClassLoader());
    }

可以看得出来,isEligible用于判断这个类或者这个类中的某个方法是否含有注解。这个方法最终进入到AopUtils的canApply方法中间:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
public static boolean canApply(Advisor advisor, Class<?> targetClass, boolean hasIntroductions) {
    if (advisor instanceof IntroductionAdvisor) {
      return ((IntroductionAdvisor) advisor).getClassFilter().matches(targetClass);
    }
    else if (advisor instanceof PointcutAdvisor) {
      PointcutAdvisor pca = (PointcutAdvisor) advisor;
      return canApply(pca.getPointcut(), targetClass, hasIntroductions);
    }
    else {
      // It doesn't have a pointcut so we assume it applies.
      return true;
    }
  }

这里的advisor就是AsyncAnnotationAdvisor对象。然后调用AsyncAnnotationAdvisor对象的getPointcut()方法,得到了Pointcut对象。在AOP规范中间,表示一个具体的切点。那么在方法上注释@Async注解,就意味着声明了一个切点。

然后再根据Pointcut判断是否含有指定的注解。

切点的执行

由于生成了JDK动态代理对象,那么每一个方法的执行必然进入到JdkDynamicAopProxy中的invoke方法中间去执行:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
@Override
  public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
    MethodInvocation invocation;
    Object oldProxy = null;
    boolean setProxyContext = false;
 
    TargetSource targetSource = this.advised.targetSource;
    Class<?> targetClass = null;
    Object target = null;
 
    try {
      if (!this.equalsDefined && AopUtils.isEqualsMethod(method)) {
        // The target does not implement the equals(Object) method itself.
        return equals(args[0]);
      }
      else if (!this.hashCodeDefined && AopUtils.isHashCodeMethod(method)) {
        // The target does not implement the hashCode() method itself.
        return hashCode();
      }
      else if (method.getDeclaringClass() == DecoratingProxy.class) {
        // There is only getDecoratedClass() declared -> dispatch to proxy config.
        return AopProxyUtils.ultimateTargetClass(this.advised);
      }
      else if (!this.advised.opaque && method.getDeclaringClass().isInterface() &&
          method.getDeclaringClass().isAssignableFrom(Advised.class)) {
        // Service invocations on ProxyConfig with the proxy config...
        return AopUtils.invokeJoinpointUsingReflection(this.advised, method, args);
      }
 
      Object retVal;
 
      if (this.advised.exposeProxy) {
        // Make invocation available if necessary.
        oldProxy = AopContext.setCurrentProxy(proxy);
        setProxyContext = true;
      }
 
      // May be null. Get as late as possible to minimize the time we "own" the target,
      // in case it comes from a pool.
      target = targetSource.getTarget();
      if (target != null) {
        targetClass = target.getClass();
      }
 
      // Get the interception chain for this method.
      List<Object> chain = this.advised.getInterceptorsAndDynamicInterceptionAdvice(method, targetClass);
 
      // Check whether we have any advice. If we don't, we can fallback on direct
      // reflective invocation of the target, and avoid creating a MethodInvocation.
      if (chain.isEmpty()) {
        // We can skip creating a MethodInvocation: just invoke the target directly
        // Note that the final invoker must be an InvokerInterceptor so we know it does
        // nothing but a reflective operation on the target, and no hot swapping or fancy proxying.
        Object[] argsToUse = AopProxyUtils.adaptArgumentsIfNecessary(method, args);
        retVal = AopUtils.invokeJoinpointUsingReflection(target, method, argsToUse);
      }
      else {
        // We need to create a method invocation...
        invocation = new ReflectiveMethodInvocation(proxy, target, method, args, targetClass, chain);
        // Proceed to the joinpoint through the interceptor chain.
        retVal = invocation.proceed();
      }
 
      // Massage return value if necessary.
      Class<?> returnType = method.getReturnType();
      if (retVal != null && retVal == target && returnType.isInstance(proxy) &&
          !RawTargetAccess.class.isAssignableFrom(method.getDeclaringClass())) {
        // Special case: it returned "this" and the return type of the method
        // is type-compatible. Note that we can't help if the target sets
        // a reference to itself in another returned object.
        retVal = proxy;
      }
      else if (retVal == null && returnType != Void.TYPE && returnType.isPrimitive()) {
        throw new AopInvocationException(
            "Null return value from advice does not match primitive return type for: " + method);
      }
      return retVal;
    }
    finally {
      if (target != null && !targetSource.isStatic()) {
        // Must have come from TargetSource.
        targetSource.releaseTarget(target);
      }
      if (setProxyContext) {
        // Restore old proxy.
        AopContext.setCurrentProxy(oldProxy);
      }
    }
  }

重点的执行语句:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
// 获取拦截器
      List<Object> chain = this.advised.getInterceptorsAndDynamicInterceptionAdvice(method, targetClass);
 
      // Check whether we have any advice. If we don't, we can fallback on direct
      // reflective invocation of the target, and avoid creating a MethodInvocation.
      if (chain.isEmpty()) {
        // We can skip creating a MethodInvocation: just invoke the target directly
        // Note that the final invoker must be an InvokerInterceptor so we know it does
        // nothing but a reflective operation on the target, and no hot swapping or fancy proxying.
        Object[] argsToUse = AopProxyUtils.adaptArgumentsIfNecessary(method, args);
        retVal = AopUtils.invokeJoinpointUsingReflection(target, method, argsToUse);
      }
      else {
      
        // 根据拦截器来执行
        invocation = new ReflectiveMethodInvocation(proxy, target, method, args, targetClass, chain);
        // Proceed to the joinpoint through the interceptor chain.
        retVal = invocation.proceed();
      }

@Async注解的拦截器是AsyncExecutionInterceptor,它继承了MethodInterceptor接口。而MethodInterceptor就是AOP规范中的Advice(切点的处理器)。

自定义注解

由于其bean处理器是通用的,所以只要实现PointcutAdvisor和具体的处理器就好了。首先自定义一个注解,只要方法加入了这个注解,就可以输出这个方法的开始时间和截止时间,注解的名字叫做@Log:

?
1
2
3
4
5
6
@Target({ElementType.METHOD, ElementType.TYPE})
@Retention(RetentionPolicy.RUNTIME)
@Documented
public @interface Log {
 
}

定义一个简单的方法用于测试:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
public interface IDemoService {
 
  void add(int a, int b);
  String getName();
}
@Service
public class DemoServiceImpl implements IDemoService {
 
  
  @Log
  
  public void add(int a, int b) {
    System.out.println(Thread.currentThread().getName());
    System.out.println(a + b);
 
  }
 
  @Override
  public String getName() {
    System.out.println("DemoServiceImpl.getName");
    return "DemoServiceImpl";
  }
 
}

定义Advisor:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
public class LogAnnotationAdvisor extends AbstractPointcutAdvisor {
 
  private Advice advice;
 
  private Pointcut pointcut;
 
  public LogAnnotationAdvisor() {
 
    this.advice = new LogAnnotationInterceptor();
  }
 
 
  @Override
  public Advice getAdvice() {
 
    return this.advice;
  }
 
  @Override
  public boolean isPerInstance() {
 
    return false;
  }
 
  @Override
  public Pointcut getPointcut() {
 
    return this.pointcut;
  }
 
  public void setAsyncAnnotationType(Class<? extends Annotation> asyncAnnotationType) {
    Assert.notNull(asyncAnnotationType, "'asyncAnnotationType' must not be null");
    Set<Class<? extends Annotation>> asyncAnnotationTypes = new HashSet<Class<? extends Annotation>>();
    asyncAnnotationTypes.add(asyncAnnotationType);
    this.pointcut = buildPointcut(asyncAnnotationTypes);
  }
 
  protected Pointcut buildPointcut(Set<Class<? extends Annotation>> asyncAnnotationTypes) {
    ComposablePointcut result = null;
    for (Class<? extends Annotation> asyncAnnotationType : asyncAnnotationTypes) {
      Pointcut cpc = new AnnotationMatchingPointcut(asyncAnnotationType, true);
      Pointcut mpc = AnnotationMatchingPointcut.forMethodAnnotation(asyncAnnotationType);
      if (result == null) {
        result = new ComposablePointcut(cpc).union(mpc);
      } else {
        result.union(cpc).union(mpc);
      }
    }
    return result;
  }
 
}

定义具体的处理器:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
public class LogAnnotationInterceptor implements MethodInterceptor, Ordered {
 
  @Override
  public int getOrder() {
 
    return Ordered.HIGHEST_PRECEDENCE;
  }
 
  @Override
  public Object invoke(MethodInvocation invocation) throws Throwable {
    System.out.println("开始执行");
    Object result = invocation.proceed();
    System.out.println("结束执行");
    return result;
  }
 
}

定义@Log专属的BeanPostProcesser对象:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
@SuppressWarnings("serial")
@Service
public class LogAnnotationBeanPostProcesser extends AbstractBeanFactoryAwareAdvisingPostProcessor {
 
  @Override
  public void setBeanFactory(BeanFactory beanFactory) {
    super.setBeanFactory(beanFactory);
    LogAnnotationAdvisor advisor = new LogAnnotationAdvisor();
    advisor.setAsyncAnnotationType(Log.class);
    this.advisor = advisor;
  }
 
 
 
}

对bean的后置处理方法直接沿用其父类的方法。当然也可以自定义其后置处理方法,那么就需要自己判断这个对象的方法是否含有注解,并且生成代理对象:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
@Override
  public Object postProcessAfterInitialization(Object bean, String beanName) {
 
    Method[] methods = ReflectionUtils.getAllDeclaredMethods(bean.getClass());
    for (Method method : methods) {
      if (method.isAnnotationPresent(Log.class)) {
        ProxyFactory proxyFactory = prepareProxyFactory(bean, beanName);
        System.out.println(proxyFactory);
        if (!proxyFactory.isProxyTargetClass()) {
          evaluateProxyInterfaces(bean.getClass(), proxyFactory);
        }
        proxyFactory.addAdvisor(this.advisor); 
        customizeProxyFactory(proxyFactory);
        return proxyFactory.getProxy(getProxyClassLoader());
      }
    }
    return bean;
 
  }

测试注解是否是正常运行的:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
public class Main {
  public static void main(String[] args) {
    @SuppressWarnings("resource")
    ClassPathXmlApplicationContext context = new ClassPathXmlApplicationContext("application-context.xml");
     IDemoService demoService = context.getBean(IDemoService.class);  
     demoService.add(1, 2);
     demoService.getName();
////     AsyncAnnotationAdvisor
//    AsyncAnnotationBeanPostProcessor
     
    
  }
}

输出:

?
1
2
3
4
5
开始执行
main
3
结束执行
DemoServiceImpl.getName

功能一切正常。

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持服务器之家。

原文链接:http://www.jianshu.com/p/3b2b881cd78f?utm_source=tuicool&utm_medium=referral