Pete, Mark,
2 items extracted...
On 2/2/13 2:45 AM, Pete Muir wrote:
<...>
>
>>> Chapter 2.5:
>>> "An around-invoke method must not be declared as abstract, final or static."
>>> Why so? I can see that abstract makes no sense, buy we don't need to further proxy them, don't we? So why not final nor static? Has this to do with some EJB limitations?
>> Yes, it was an EJB limitation when interceptors were added. Is it a problem for CDI?
> No, it's just uncessary. We could relax this.
Changing it will require changes to all interceptor types, and the
common annotations spec. I doubt we have time for that...
<...>
>
>>> Chapter 3.2:
>>> "If an interceptor does not declare an Interceptor annotation, it could be bound to components using Interceptors annotation or a deployment descriptor file."
>>> Does this also imply forbidding the reverse? That a class annotated with @Interceptor must not be used in @Interceptors(..) ?
>> I'm not sure we should this should be forbidden, though ch2 says "Interceptors classes may be associated with the target class using either interceptor binding (see Chapter 3) or the Interceptors annotation (see Chapter 4)."
>>
> I don't think it does imply the reverse.
Pete, (I'm confused to which question it applies) do we need to add
anything to the spec?
thanks,
-marina