Spring Spring AOP support for prototype-scoped advice

If we use prototype-scoped aspect and use args pointcut expression, it will throw warning no match for this type name: ss [Xlint:invalidAbsoluteTypeName]，the arg havn't be binding.

Code like bellow:

@Aspect("perthis(execution(* com.tom.test.demo..*(..)))")
@Component
@Scope("prototype")
public class AspectJ {

    @Before("execution(* com.tom.test.demo..*(..)) && args(ss)")
    public <T extends String> void pointCut(JoinPoint joinPoint,String ss) {
        System.out.println("????????????????????????");
    }
}

This will throw exception

Caused by: java.lang.IllegalArgumentException: warning no match for this type name: ss [Xlint:invalidAbsoluteTypeName]

If we use the follow code it will run well，but advise can't obtain parameter

@Aspect("perthis(execution(* com.tom.test.demo..*(..)))")
@Component
@Scope("prototype")
public class AspectJ {

    @Before("execution(* com.tom.test.demo..*(..)) && args(java.lang.String)")
    public <T extends String> void pointCut(JoinPoint joinPoint) {
        System.out.println("????????????????????????");
    }
}

Is this a bug, or is it meant to work so?

Comment From: rstoyanchev

I'm not sure if prototype scope applies to an @Aspect and I'm not sure what it would mean. Can you clarify what your expectation with that declaration? @jhoeller is this even an expected combination?

Comment From: ZhongGuoDragon

see code class org.springframework.aop.aspectj.annotation.InstantiationModelAwarePointcutAdvisorImpl you will see if you use @Aspect(perthis()) or @Aspect((pertarget)) it will create PerTargetInstantiationModelPointcut

if (aspectInstanceFactory.getAspectMetadata().isLazilyInstantiated()) { // Static part of the pointcut is a lazy type. Pointcut preInstantiationPointcut = Pointcuts.union( aspectInstanceFactory.getAspectMetadata().getPerClausePointcut(), this.declaredPointcut);

      // Make it dynamic: must mutate from pre-instantiation to post-instantiation state.
      // If it's not a dynamic pointcut, it may be optimized out
      // by the Spring AOP infrastructure after the first evaluation.
      this.pointcut = new PerTargetInstantiationModelPointcut(
              this.declaredPointcut, preInstantiationPointcut, aspectInstanceFactory);
      this.lazy = true;
  }

it will create tow pointcut but the second pointcut without args binding, it will throw exception Caused by: java.lang.IllegalArgumentException: warning no match for this type name: ss [Xlint:invalidAbsoluteTypeName] if you use like below:

@Aspect("perthis(execution(* com.tom.test.demo..*(..)))")
@Component
@Scope("prototype")
public class AspectJ {

    @Before("execution(* com.tom.test.demo..*(..)) && args(ss)")
    public <T extends String> void pointCut(JoinPoint joinPoint,String ss) {
        System.out.println("????????????????????????");
    }
}

Comment From: ZhongGuoDragon

I'm not sure if prototype scope applies to an @Aspect and I'm not sure what it would mean. Can you clarify what your expectation with that declaration? @jhoeller is this even an expected combination?

the args bingding only create advise invoking. see org.springframework.aop.aspectj.annotation.ReflectiveAspectJAdvisorFactory method: public Advice getAdvice(Method candidateAdviceMethod, AspectJExpressionPointcut expressionPointcut, MetadataAwareAspectInstanceFactory aspectInstanceFactory, int declarationOrder, String aspectName) ............ if (argNames != null) { springAdvice.setArgumentNamesFromStringArray(argNames); } springAdvice.calculateArgumentBindings(); ............. if you use @Aspect(perthis()) or @Aspect(pertarget) args pointcut expression and need advise method arg binding mode,it will create union pointcut but the second AspectJExpressionPointcut havn't args binding will throw Caused by: java.lang.IllegalArgumentException: warning no match for this type name: ss [Xlint:invalidAbsoluteTypeName]

Comment From: vishalsingh2972

@ZhongGuoDragon were you able to solve this issue, @jhoeller any update on this can we submit a PR on this or is this an internal issue?

Comment From: jhoeller

From my side, there is no concrete plan for this. If someone gets to the bottom of the problem and submits a PR for it, I'm happy to consider it but I'm not actively working on it myself.