> Hi!
>
> marc fleury wrote:
> > I am on top of this one (as answered in the private mails)
> > here is the skinny...
> >
> > right now some "reentrancy" is done in the cache. The call looses the
> > "invoke" "invokeHome" trail since the cache is not aware of it.
> However the
> > cache takes the decision.
> >
> > The right way is to rewire the decision in the
> InstanceInterceptor (no need
> > for a new interceptor actually, my bad) and I tried to do that.
> >
> > There is one lingering issue and that is that the "lock" information is
> > cache proprietary at this point. It would just take
> refactoring to expose
> > it to the II so that he can do essentially the same code that
> the cache is
> > doing but in the invoke and invokeHome trails so that the right
> decision can
> > be taken.
>
> Actually, the purpose of the lock was to allow reentrancy checks. If
> this is to be done by the II, then it is no longer needed. Instead, we
> could add a flag "isCalled" in the EntityEnterpriseContext instead,
> which the II could use. The semantics of the cache "get" would then
> *only* be "please give me the context for this id", and nothing more.
Ok
marc
>
> Sounds ok?
>
> /Rickard
>
> --
> Rickard �berg
>
> Email: [EMAIL PROTECTED]
> http://www.telkel.com
> http://www.jboss.org
> http://www.dreambean.com
>
>
