Re: [aspectj-users] AspectJ Performance toward the pointcut definition

[Andy Clement <andrew.clement@xxxxxxxxx>]

There is no extra runtime performance impact for choosing the pointcut with wildcards in.  The difference will be in the compile/weaving time as it will be comparing more possible candidates against the wildcarded pointcut. 

If you were using a pointcut where we could not statically (at compile time) work out if it was an exact match, we would insert a runtime test to confirm whatever we couldn't statically determine. For example: execution(* foo(..)) && target(Foo). If Foo were an interface we might be inserting 'instanceof' checks into the bytecode.

Andy

On 11 February 2014 03:46, Muhammad Adinata <mail.dieend@xxxxxxxxx> wrote:

Hello,

I want to know whether more definitive pointcut definition have different impact toward a generic one.

For example:

after(): execution(void Hello.helloWorld()) {
}

and 

after(): execution(* Hello.hello*()) {
}

Assuming there is only one method that fulfill the definition, does they have different impact to application performance?

I used compile-time weaving and no runtime weaving.

Thanks!

--

Muhammad Adinata

13509022 - Informatika ITB 2009

_______________________________________________
aspectj-users mailing list
aspectj-users@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/aspectj-users