Re: [aspectj-dev] XReferences Algorithm

[Armin <funkoholic@xxxxxx>]

Hi devs!

I have another question on this XReferences. Examination of the  
weaver's source code shows that every join point shadow has a unique  
ID. Every time a new shadow is produced it gets a new ID just by  
incremanting a static field in the shadow class.
How long does this ID maintain?
Is this whole "ask-every-possible-shadow-if-it-matches-some-of-the- 
mungers"-process done every time the weaver runs?
Or can the weaver somehow maintain the ID (thus the list of matching  
mungers) of a shadow during code evolution by processing the code  
with an incremental weaving feature?
The documentation on incremental weaving per-class tells me that if  
some crosscutting specification may have been updated, then all code  
potentially affected by it may need to be woven.

thanks for advice!

Greets.
cptnuss


Am 10.01.2006 um 04:43 schrieb Mik Kersten:

> This strategy is captured by the org.aspectj.asm module.  It  
> doesn't use
> pattern matching and instead stores the static join point shadow  
> that the
> weaver computes.  To see how this is done look at the  
> implementation of
> AsmRelationshipProvider.  Note that the AsmHierarchyBuilder builds  
> up the
> declarations and statements that appear as the sources and targets  
> of the
> relationships.  To experiment with extending it take a look at the  
> following
> message:
>
> http://dev.eclipse.org/mhonarc/lists/aspectj-dev/msg01648.html
>
> Cheers,
>
> Mik
>
> > -----Original Message-----
> > From: aspectj-dev-bounces@xxxxxxxxxxx [mailto:aspectj-dev-
> > bounces@xxxxxxxxxxx] On Behalf Of Armin
> > Sent: January 9, 2006 5:16 AM
> > To: aspectj-dev@xxxxxxxxxxx
> > Subject: [aspectj-dev] XReferences Algorithm
> >
> > Hi devs!
> >
> > I searched the source code of AJDT and the mailing archive for this
> > but did not get through!
> >
> > I'd like to know what kind of strategy has been choosen to compute
> > the XReferences between the pointcuts
> > used in the advices and the JavaElements that matches the patterns
> > within the pointcuts.
> >
> > I guess that there has to be a pattern matching algorithm somewhere
> > but I did not find a clue.
> >
> > thanks for advice =)
> >
> > Greets
> > cptnuss
> >
> > ----------------
> >
> > from Bonn/Germany (the place of this years great AOSD conference.
> > Hope  to meet u there!)
> > _______________________________________________
> > aspectj-dev mailing list
> > aspectj-dev@xxxxxxxxxxx
> > https://dev.eclipse.org/mailman/listinfo/aspectj-dev
>
> _______________________________________________
> aspectj-dev mailing list
> aspectj-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-dev