jboss-dev-forums

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[<kab...@jb...>]

You need to provide a mixin when your interface contains methods

http://jira.jboss.com/jira/browse/JBAOP-71

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3887003#3887003

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3887003

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[nthx <nu...@jb...>]

Hi there Kabir!

Well, I read the JIRA's issue. I completely agree with Bill here: anonymous wrote : The thing is, if you're just introducing an interface with no mixin, then an interceptor should handle the implementation.

But hear this, cause that's more funny:
'takeSnapshot' is another method from the interface. Similar pointcut exists, the same aspect, the same test case and call to ((Persistable)root).takeSnapshot(); works!

jboss-aop-1.3, no mixin for interface, aop config you see here, the same behavior for call/execution type...

This is wired.. have any possible explanation for that behaviour?

Thanks - Tomasz



View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3887259#3887259

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3887259

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[<kab...@jb...>]

Hi Tomasz,

If you're still having problems, can you please send me a scaled down version of this to demonstrate the problem?

Thanks

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3888862#3888862

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3888862

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[nthx <nu...@jb...>]

Yes, I still do have that problem.

I prepared an archive with all unnecessary files removed from my project. It's stripped down version of Pat M1.09. The build process needs to make distros for jdk5 and jdk4, so don't look at 'build.xml' ;-).

$ant tests (for jdk4, uses @@)
$ant -Dnthx.is.java5=yes tests (requires jdk5, uses @)

Archive sent on priv.

Kind regards,
Tomasz Nazar



View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3889608#3889608

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3889608

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[<kab...@jb...>]

Hi Tomasz,

The problem is your pointcut expression

<bind pointcut="execution(void @pat.root->getObjectByOID(java.lang.Long))">
  |     <advice aspect="org.nthx.pat.PersistableAspect" name="getObjectByOID"/>
  | </bind>

A bit of digging reveals that Persistable.getObjectByOID() returns Identifiable, so the pointcut is never matched, meaning interception does not take place, which causes the recursion.

If I change the pointcut to

<bind pointcut="execution(org.nthx.pat.Identifiable @pat.root->getObjectByOID(java.lang.Long))">
  |     <advice aspect="org.nthx.pat.PersistableAspect" name="getObjectByOID"/>
  | </bind>

it is fine.

Cheers,

Kabir

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3890346#3890346

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3890346

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[<kab...@jb...>]

PS sorry for the delay, I should have spotted this a lot sooner

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3890348#3890348

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3890348

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[nthx <nu...@jb...>]

Aaaaaa! ;-)

What a silly bug. Thanks for your time. :)

This happens when you "copy and paste"... 

So few points here:
1. JBAOP should warn the pointcut doesn't exist, shouldn't it? (Instead of recursively calling itself)
2. Here the IDE could help by showing "No pointcuts matched". Is there any for IDEA? There is one for Eclipse, isn't it?

One more time: "thank you" -- what a silly bug :)

Tomasz


View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3890384#3890384

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3890384

[Jboss-dev-forums] [Design of AOP on JBoss (Aspects/JBoss)] - Re: StackOverflow when trying to

[<kab...@jb...>]

1) I think if you use verbose mode you should get the warning during weaving about methods not matching any joinpoints

2) We currently have an IDE for Eclipse, but not for IDEA and I think it has some warnings like this. I was just trying out the latest IDE myself this morning, but I seemed to get hold of an older version. I'll let you know as soon as I know myself :-)

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3890390#3890390

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3890390