Re: Question regarding Dependent scoped beans and Interceptors/Decorators

[Gurkan Erdogdu <cg...@gmail.com>]

Hi Joe;

Yeap this is the current situation at OWB. I have not found any valuable
information about this topic from the specification too. But spec. says that
@Dependent scoped beans are not proxied. From this statement, I though that
I not intercept/decorate dependent beans. Maybe this is not correct!

Mark, could you ask about it to Pete Muir :)?

Thanks;

--Gurkan

2009/10/26 Joseph Bergmark <be...@gmail.com>

> Does the 299 specification indicate that interceptors and decorators should
> not be called for dependent scoped beans?  That appears to be the current
> behavior in OWB but I'm having trouble nailing it down in the spec.
>
> For example, 8.1.2 says:
> "The decorator applies to any bean that is eligible for injection to the
> delegate injection point, according to the rules defined in Section 5.3,
> “Typesafe resolution”. "
>
> 7.2 appears to apply some additional rules around what is a business method
> invocation, and that interceptors and decorators are only called for
> business method invocations.  One restriction here is that the call must be
> to a contextual reference to a bean.  6.5.3 does talk about contextual
> instances of dependent scoped beans though.
>
> Section 5.5 talks about client proxies and that pseudo scopes (such as
> dependent) don't require a proxy.  However, this doesn't appear to get tied
> back into Decorators and Interceptors anywhere that I have been able to
> find.
>
> Sincerely,
>
> Joe Bergmark
>



-- 
Gurkan Erdogdu
http://gurkanerdogdu.blogspot.com

Re: Question regarding Dependent scoped beans and Interceptors/Decorators

[Eric Covener <co...@gmail.com>]

On Mon, Oct 26, 2009 at 4:31 PM, Gurkan Erdogdu
<cg...@gmail.com> wrote:
> Hi Joe;
>
> Yeap this is the current situation at OWB. I have not found any valuable
> information about this topic from the specification too. But spec. says that
> @Dependent scoped beans are not proxied. From this statement, I though that
> I not intercept/decorate dependent beans. Maybe this is not correct!
>
> Mark, could you ask about it to Pete Muir :)?

FWIW, my take on the proxy issue was that while normal scope
contextual instances MUST be a proxy, pseudo-scope beans are _not
necessarily_ a proxy.

IOW which of these is intended in 299:

  [ ] contextual instance of normal scope MUST be a client proxy,
contextual instance of  pseudo-scope MAY be a proxy
  [ ] contextual instance of normal scope MUST be a client proxy,
contextual instance of  pseudo-scope MUST NOT be a proxy

(might be worth kicking up to Pete at the same time)

-- 
Eric Covener
covener@gmail.com

Re: Question regarding Dependent scoped beans and Interceptors/Decorators

[Sven Linstaedt <sv...@googlemail.com>]

Afaik dependent scoped beans should inheret the lifecycle of these  
beans they are injected into. Since they only live within the  
"context" of a single other bean they do not need to proxied, which  
does not mean from my point of view they must not get a full  
interceptor/decorator stack.

Does anyone has a good use case/example when dependent scoped beans  
are useful or necessary?

br, Sven



Am 26.10.2009 um 21:31 schrieb Gurkan Erdogdu  
<cg...@gmail.com>:

> Hi Joe;
>
> Yeap this is the current situation at OWB. I have not found any  
> valuable
> information about this topic from the specification too. But spec.  
> says that
> @Dependent scoped beans are not proxied. From this statement, I  
> though that
> I not intercept/decorate dependent beans. Maybe this is not correct!
>
> Mark, could you ask about it to Pete Muir :)?
>
> Thanks;
>
> --Gurkan
>
> 2009/10/26 Joseph Bergmark <be...@gmail.com>
>
>> Does the 299 specification indicate that interceptors and  
>> decorators should
>> not be called for dependent scoped beans?  That appears to be the  
>> current
>> behavior in OWB but I'm having trouble nailing it down in the spec.
>>
>> For example, 8.1.2 says:
>> "The decorator applies to any bean that is eligible for injection  
>> to the
>> delegate injection point, according to the rules defined in Section  
>> 5.3,
>> “Typesafe resolution”. "
>>
>> 7.2 appears to apply some additional rules around what is a  
>> business method
>> invocation, and that interceptors and decorators are only called for
>> business method invocations.  One restriction here is that the call  
>> must be
>> to a contextual reference to a bean.  6.5.3 does talk about  
>> contextual
>> instances of dependent scoped beans though.
>>
>> Section 5.5 talks about client proxies and that pseudo scopes (such  
>> as
>> dependent) don't require a proxy.  However, this doesn't appear to  
>> get tied
>> back into Decorators and Interceptors anywhere that I have been  
>> able to
>> find.
>>
>> Sincerely,
>>
>> Joe Bergmark
>>
>
>
>
> -- 
> Gurkan Erdogdu
> http://gurkanerdogdu.blogspot.com