Re: [Springframework-developer] StaticMethodPointcut

[Colin S. <col...@ex...>]

+1

Rod Johnson wrote:

>Guys,
>
>I think we should change the signature on the StaticMethodPointcut applies()
>method to introduce a new argument, targetClass, as below:
>
>boolean applies(Method method, Class targetClass, AttributeRegistry
>attributeRegistry);
>
>The reason is that sometimes we need to know not just the method, but the
>target class we're invoking.
>
>Consider the getAge() method on class TestBean. A subclass, SpecialTestBean
>adds a class-level metadata attribute that should cause auto-proxying. With
>the old signature, the method argument would be TestBean.getAge() and
>without knowledge of the target class we would miss this important
>attribute.
>
>Does everyone agree this makes sense? Is there a better solution?
>
>This will break existing static pointcuts, although it's trivial to fix.
>I've already revised the source and test tree (although not committed yet).
>Hence I didn't commit it before M2.
>
>Regards,
>Rod
>  
>