[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [aspectj-users] WebLogic 9.2 + AspectJ = boom?
|
One question, is how do you have the EAR Loading setup? I was wondering if
you have it setup to load from the EAR first? By using something like:
<prefer-application-packages>
<package-name>org.apache.log4j.*</package-name>
<package-name>antlr.*</package-name>
</prefer-application-packages>
In weblogic-application.xml
See:
http://e-docs.bea.com/wls/docs92/programming/classloading.html
I know the class loader architecture on WLS can be a pain especially under
9.2. In fact at a client of mine we dropped 9.2 in favor or 10.0 because of
these types of issues.
On 2/9/09 6:47 PM, "Michael Laccetti" <michael@xxxxxxxxxxxx> wrote:
> I think we have a bigger problem... I see the same error happening on debug.
> I re-ran tests against WebLogic in debug, and had nothing show up in the DB (I
> cleared the tables out beforehand). Ran it again under Tomcat and it filled
> the tables up nicely. So, I suspect that it never worked under WebLogic, that
> it was the Tomcat testing that was populating the tables.
>
> I did get a massive amount of debug logging from both the debug and staging
> WebLogic servers - if you have the time and patience to dig through them, you
> can download them from a zip here: http://laccetti.com/files/AOP tests.zip
>
> I appreciate all of the effort you are putting into this.
>
> Mike
>
>> -----Original Message-----
>> From: aspectj-users-bounces@xxxxxxxxxxx [mailto:aspectj-users-
>> bounces@xxxxxxxxxxx] On Behalf Of Michael Laccetti
>> Sent: February-09-09 6:06 PM
>> To: 'aspectj-users@xxxxxxxxxxx'
>> Subject: Re: [aspectj-users] WebLogic 9.2 + AspectJ = boom?
>>
>> I'll give it a shot. (Both turn on debug in debug and verbose in
>> staging.)
>>
>> Mike
>> --------------------------
>> Michael Laccetti
>> 416-558-9718
>>
>> -----Original Message-----
>> From: aspectj-users-bounces@xxxxxxxxxxx <aspectj-users-
>> bounces@xxxxxxxxxxx>
>> To: aspectj-users@xxxxxxxxxxx <aspectj-users@xxxxxxxxxxx>
>> Sent: Mon Feb 09 17:46:26 2009
>> Subject: Re: [aspectj-users] WebLogic 9.2 + AspectJ = boom?
>>
>> I am half wondering if the ClassNotFoundExceptions are entirely red
>> herrings. They are not nice (and indicate I think we have a bug) but
>> it is rare that anyone turns on AspectJ trace - I suspect if you turned
>> on trace on your working cluster, you might well see them! So, going
>> back to the original note, I would like to know how the AOP stuff is
>> actually not working? In your aop.xml you can turn on -verbose, and
>> additionally -showWeaveInfo as another option which will tell us if the
>> weaver is advising where you want to advise.
>>
>> cheers,
>> Andy.
>>
>>
>> 2009/2/9 Michael Laccetti <michael@xxxxxxxxxxxx>
>>
>>
>> I love straw grasping. :)
>>
>>
>>
>> I've attached the stack, from just a bit before the exceptions,
>> to just after. Let me know if there is anything else I can dig up. (I
>> forgot to turn on verbose in the config file; next release will have
>> that.)
>>
>>
>>
>> Mike
>>
>>
>>
>> From: aspectj-users-bounces@xxxxxxxxxxx [mailto:aspectj-users-
>> bounces@xxxxxxxxxxx] On Behalf Of Andy Clement
>> Sent: February-09-09 4:14 PM
>>
>> To: aspectj-users@xxxxxxxxxxx
>> Subject: Re: [aspectj-users] WebLogic 9.2 + AspectJ = boom?
>>
>>
>>
>> Hi Michael,
>>
>>
>>
>> I see you are also having a problem for another type there, and
>> it also trying multiple prefixes: java.lang, and
>> com.cryptologic.bsg.aspect:
>> com.cryptologic.bsg.aspect.com.cryptologic.bsg.service.impl.EcashPlayer
>> AccountServiceImpl
>>
>> As the -Xset option did not fix it - that suggests the
>> classloader is OK (but I guess we might have known that as it deploys
>> fine on your other system).
>>
>> Do you have the full stack traces for the ClassNotFound problems?
>> Very peculiar that the java.lang. prefixed CNFE comes out before the
>> CNFE for the non-prefixed version.
>>
>> I'm kind of grasping at straws here - since it works on one
>> system but not on the other, I'm finding it difficult to work out what
>> AJ might be doing wrong, or what assumption it is making that isn't
>> valid across all configurations.
>>
>> Is there by any chance another older copy of the code in some
>> other jar in the system configuration or on the classpath ? If an old
>> version of the application/aspect got loaded by a different loader
>> (maybe a loader higher in the loader hierarchy), then anything below
>> that loader in the classloader hierarchy would not be visible to the
>> weaver. Is the classpath different in any way across these
>> configurations?
>>
>> Andy.
>>
>> _______________________________________________
>> aspectj-users mailing list
>> aspectj-users@xxxxxxxxxxx
>> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>>
>>
>>
>>
>> _______________________________________________
>> aspectj-users mailing list
>> aspectj-users@xxxxxxxxxxx
>> https://dev.eclipse.org/mailman/listinfo/aspectj-users
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
Ron DiFrango
Manager and Architect | CapTech Ventures
(804) 855-9196-6308 | rdifrango@xxxxxxxxxxxxxxxxxxx