I think you're right. After modifying my pointcuts I do not seem to have the
tangle anymore.

I've not added A as aspectpath to B, I've added A as a dependency in the
manifest. Perhaps aspectpath is a better way to go about?

About the inpath, are you saying that I should put the (bin folders of the)
project (let's call it C) which I'd like to be woven by B as inpath for B?
Right now I'm putting the bin folders of B (and A) as inpath for C. It seems
to work in my Eclipse development environment (how ever I still don't get it
to work when I deploy everything to a clean Eclipse installation).

Thanks for your help.

On Nov 29, 2007 3:47 PM, Andy Clement <[EMAIL PROTECTED]> wrote:

> inpath is code that will be subject to weaving, aspectpath is code
> that is input to the weaving process but should not be woven.  I could
> imagine A on the aspectpath for B so that it can reference aspects on
> it (concretize them).
>
> references from A to B.aspectOf() suggest to me that you have
> accidentally woven A in the pointcuts/advice you are using?
>
> Andy.
>
>
> On 29/11/2007, Johan Haleby <[EMAIL PROTECTED]> wrote:
> > B has A added to its manifest file as a Required Plugin whereas A has no
> > dependency to B in its manifest file. Both A & B are added as binary
> inpath
> > to the AspectJified projects which should be woven by the aspect(s). Is
> it
> > perhaps enough to add B to the inpath?
> >
> > Structure 101 reports the following:
> > From                       Usage               To
> > A.<init>                   references         .B
> > A.<init>                   calls                 .B.aspectOf
> > A.formatJoinPoint     references         .B
> > A.formatJoinPoint     calls                 .B.aspectOf
> > A.logWarning           references         .B
> > A.logWarning           calls                 .B.aspectOf
> >
> > (I have of course renamed the aspects to A & B in my mail)
> >
> >
> >
> > On Nov 29, 2007 2:12 PM, Andy Clement <[EMAIL PROTECTED]> wrote:
> > > Are you simply using project dependencies so that B can see A? Or
> > > aspectpath? Or inpath?
> > >
> > > I cannot see a reason for A to reference B unless you have explicitly
> > > directly mentioned the types of B in your abstract aspect.  If you
> > > supply more information, maybe we can work out why
> > >
> > > Andy.
> > >
> > >
> > >
> > >
> > > On 29/11/2007, Johan Haleby <[EMAIL PROTECTED]> wrote:
> > > > Hi,
> > > >
> > > > I have two aspects and one of them is abstract (A) and the other one
> (B)
> > > > extends A and and provides the concrete implementation. A & B are in
> two
> > > > separate projects. When I examine my project in Structure 101 it
> says
> > that I
> > > > have a package tangle between A & B. It seems like A is referencing
> and
> > > > calling B and B must obviously know of A since it extends it. Is
> this
> > the
> > > > correct behavior? It's sad that you have to live with an extra
> tangle
> > that I
> > > > can do nothing about.
> > > >
> > > > Thanks,
> > > > Johan
> > > >
> > > > _______________________________________________
> > > > aspectj-users mailing list
> > > > [email protected]
> > > > https://dev.eclipse.org/mailman/listinfo/aspectj-users
> > > >
> > > >
> > > _______________________________________________
> > > aspectj-users mailing list
> > > [email protected]
> > > https://dev.eclipse.org/mailman/listinfo/aspectj-users
> > >
> >
> >
> > _______________________________________________
> > aspectj-users mailing list
> > [email protected]
> > https://dev.eclipse.org/mailman/listinfo/aspectj-users
> >
> >
> _______________________________________________
> aspectj-users mailing list
> [email protected]
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
_______________________________________________
aspectj-users mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/aspectj-users

Reply via email to