Re: [aspectj-users] doesn't a handler-pointcut 'return'?

["Wes Isberg" <wes@xxxxxxxxxxxxxx>]

As the programming guide implementation notes point out, it is not 
possible to reliably determine the end of an exception handler 
in bytecode:

http://dev.eclipse.org/viewcvs/indextech.cgi/%7Echeckout%7E/aspectj-home/doc/progguide/apcs02.html#d0e6914

So this isn't really a compiler limitation (to be fixed), but is 
a limitation of bytecode weaving (which no weaver could solve).

As for workarounds, I know of half-measures, but no good 
workarounds for that use-case.  (It would be nice as a first cut 
if JQuery could evaluate predicates for handlers and throws clauses.)

Wes

> ------------Original Message------------
> From: Hermann Voßeler <hermann.vosseler@xxxxxxxxxxxxx>
> To: aspectj-users@xxxxxxxxxxx
> Date: Tue, Oct-19-2004 12:16 PM
> Subject: Re: [aspectj-users] doesn't a handler-pointcut 'return'?
>
> > Hermann Voßeler wrote:
> >>but this doesn't work due to a compiler limitaion (compiler prints
> >>a warning). Does anyone know a workaround?
> 
> Eric Bodden wrote:
> > It could help to know what the warning actually says.
> 
> "Only before advice is supported on handler join points
> (compiler limitation)"
> 
> I use the compiler in
> plugins/org.aspectj.ajde_1.2.0/aspectjtools.jar
> 
> 
> -- 
> 
> Hermann
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/aspectj-users
>