Re: [aspectj-users] Previously running program broke with 3.0M8/AJDT 1.1

[Wes Isberg <wes@xxxxxxxxxxxxxx>]

Thanks for pointing this out.

I believe this is a bug fix (i.e., the semantics didn't
change, but our implementation is more correct :)

  https://bugs.eclipse.org/bugs/show_bug.cgi?id=49295

The bug is mentioned in the porting guide, but not the
issue of losing these join points.

Your code should work if you change it

from: execution(Marker.new(..))
  to: execution(Marker+.new(..))

I'll add this to the porting guide.

Wes

mmonteiro@xxxxxxxxxxxxxxxxx wrote:

> Today I found out that a program that runned without any problems with 
> eclipse 2.1/AJDT 1.1.4 broke with eclipse 3.0M8/AJDT 1.1.8.
> The cause were pointcuts that targeted constructor executions based on a 
> marker interface. They worked fine with in the older environment, but 
> the pointcuts no longer captured the *.new() joinpoints in the new version.
> Next follows a simplified example that captures the situation. It runs 
> fine with eclipse 2.1/AJDT 1.1.4 but throws a RuntimeException with 
> 3.0M8/AJDT 1.1.8.
> ------------------------------------------------------------------
> public class Capsule {
> }
> ------------------------------------------------------------------
> import java.util.ArrayList;
> public aspect Generic {
>  private interface Marker {}
>  private ArrayList Marker._list;
>  pointcut markerCreation(Marker marker):
>     execution(Marker.new(..)) && this(marker);
>  after(Marker marker) returning : markerCreation(marker) {
>     marker._list = new ArrayList();
>  }
>  public int Marker.getSize() {
>     return this._list.size();
>  }
>  declare parents: Capsule implements Marker;
> }
> ------------------------------------------------------------------
> public class Run {
>  public static void main(String[] args) {
>     Capsule capsule = new Capsule();
>     System.out.println("Size on creation: " + capsule.getSize());
>  }
> }
> ------------------------------------------------------------------
> Was there any change in the semantics of AspectJ from 1.1 to 1.2?
> Miguel J. T. Pessoa Monteiro
> Ph.D. student Minho University,Portugal
> eMail: mmonteiro@xxxxxxxxxxxx
> URL:http://gec.di.uminho.pt/mpm/
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/aspectj-users