[ews.eclipse.technology.aspectj] Re: Can AspectJ identify synthetic fiel

[Adrian Colyer <adrian.colyer@xxxxxxxxx>]

Lefty A. Rivera wrote:

> I'm using a:
> Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2)
> Classic VM (build 1.4.2, J2RE 1.4.2 IBM Windows 32 build cn142-20050609 
> JIT enabled: jitc)).
>
> Let me add the synthetic fields were added by EMMA instrumentation, does 
> this make any difference?.
>
> How can I tell AspectJ to ignore a line like this  :
> boolean aflag[] = ($VRc != null ? $VRc : $VRi())[0];
>
>
> If not I'll open a Bugzilla bug. :-(.

I had a feeling it might be an IBM 1.4.2 - we've had some "known issues" 
with that VM in the past where the debug information output by the 
AspectJ compiler upsets the VM and causes it to crash. If you're 
compiling with AspectJ 1.2.1 or earlier, and passing the compiler option 
  -g:none makes the problem go away, then odds are very high that this 
is what you are running into.

AspectJ will only weave into types that have matching join points - if 
there are no pointcuts matching any join points in the line you quote 
(or if this is a local variable, any join points inside the declaring 
method), then the bytecode should pass through unchanged.

As long as EMMA is producing valid .class files there should be no 
problems passing these into AspectJ - but it is possible that EMMA 
generates valid though seldom seen patterns that may be triggering an 
aspectj bug. Without looking at the .class files, it is hard to be more 
definite than that.

Regards, Adrian.