get* and set* pointcuts vs. set(field) get(field)

[nthx]

Hi,

In TreeCacheAop doc you state sth like this: "..Whenever there is state change ('set*' interceptors)..All 'get*' operations.. "

I wonder if it is a good idea to restrict user of JBossCache to conform to some getters-setters standard. Maybe it would be better to intercept field access (read or write) using standard JBossAOP: get/set(field expression)?

You know there might be problems with "settle(..)" method for example, and maybe someone (I :) ) doesn't like accessing their private implementation through getters and setters..

I _think_ field access joinpoint would solve some problems..

How do you think? Maybe you have that in your plans for 1.2.

And one more thing: are you going to extract jbossCache as a concern, you know.. to provide jboss-aop.xml bindings for jbossCache, for users who will need some more advanced concern management?

Thanks in advance,
Tomasz Nazar

[ben.wang]

Thomaz,

Actually, we are doing field interception already (in jboss-aop.xml all "prepare" declaration are for fields). So I will need to update this part of the document (of which I am doing it right now). :-)

Thanks for pointing it out,

-Ben

[nthx]

Ok. This was in TreeCacheAop.java's javadoc..

And what about second question?

Regards,
Tomasz

[ben.wang]

Yes, to "aspectize" JBossCache is in 2.0 roadmap. The refactoring is currently underway done by Bela now.

-Ben