[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [aspectj-dev] Synthetic Methods?
- From: Andrew Clement <CLEMAS@xxxxxxxxxx>
- Date: Fri, 29 Apr 2005 08:31:17 +0100
- Delivered-to: firstname.lastname@example.org
I believe we found the regular synthetic attribute wasn't sufficient as
compiler/tool/toolkit out there either supplied it or interpreted it in a
different way - so we had to switch to our own. I have heard of coverage
choking, and have even have a bug raised on one (don't think it was EMMA).
If you need the original synthetic attribute too - I guess we could add
to keep these other tools happy.
Sent by: "'AspectJ developer discussions'"
29/04/2005 00:05 [aspectj-dev] Synthetic Methods?
Please respond to
I remember that at one time AspectJ added a synthetic modifier to generated
methods like has Aspect and aspectOf. It looks like it now only adds an
AspectJ-specific attribute org.aspectj.weaver.AjSynthetic. Was there a
reason to remove the standard synthetic modifier?
I ask because I'm trying to get the EMMA coverage tool working on AspectJ
aspects, and it expects to have methods like this tagged with synthetic or
else for them to have line numbers associated with them. The good news is
works, but unfortunately it won't produce much info since it thinks the
number data is missing.
I'm also wondering if there are other notable cases where AspectJ emits
bytecode without line number information. I know I've seen stack traces
aspect code (like runtime tests for cflow or if in pointcut evaluation)
didn't have meaningful line numbers. In looking at the bytecode, it looks
like they do *have* a line number, it's just 0.
Chief Technology Officer
New Aspects of Software
w: (415) 824-4690
aspectj-dev mailing list