Re: POJO Style-beans

[David Blevins]

On Aug 14, 2006, at 4:15 PM, Paulo Lopes wrote:

Hi,
What Mohammed wanted to say (i think) is that we should try to the the
deployment process for openejb 3 working so the real tests could
start. In other words we'd need
to have the annotations parser working and some bootstrapping of the
server to fully
test the new EJB 3.0 spec features.

Ah, alright.  We do have deployment and runtime support for POJO  
Stateless and Stateful beans that have component interfaces (home/ 
remote).  It just uses the deployment descriptor instead of Annotations.

Anyone have any thoughts on doing OPENEJB-145 as I detail in this  
email?  http://marc.theaimsgroup.com/?l=openejb- 
development&m=115481372914101&w=2

-David


Regards,
Paulo

On 8/14/06, David Blevins <[EMAIL PROTECTED]> wrote:

On Aug 14, 2006, at 6:49 AM, Mohammed Nour wrote:

> Hi DBlevins...
>
> Concerning the iTests, I and Paulo suggest that we own the impl  
of the
> missing of EJB3.0 deployment time features, like using annotations,
> this
> will help us deploying our test beans and in the same time u did
> the impl of
> the run time so we can test it too, this will be with the impl of
> the iTests
> ofcaurse, so what do u think ?

I have to apologize, I didn't follow that sentence too well.

Maybe you can explain your game-plan in terms of JIRA items you'd
like to work on.

-David


> Thanks and best regards...
> Mohammad Nour El-Din
>
>
> On 8/4/06, David Blevins <[EMAIL PROTECTED]> wrote:
>
>>
>> On Jul 25, 2006, at 5:24 PM, David Blevins wrote:
>>
>> >
>> > On Jul 24, 2006, at 1:35 PM, David Blevins wrote:
>> >
>> >> Going to take a whack at support for EJB 3 POJO Style beans.
>> >> Thinking to start here http://jira.codehaus.org/browse/ 
OPENEJB-76
>> >
>> > Got some trivial support in the stateless container.  It longer
>> > assumes that all bean instances are subclasses of SessionBean.
>> > Considered dynamically subclassing each bean class to adapt  
it to
>> > be a SessionBean class, but that would add significant  
overhead to
>> > startup.  The other downer would be you'd never get to see your
>> > bean classname in the stacktrace, only the generated class name.
>>
>> Got something in for the stateful container now too.  What we have
>> now is the ability to support a pojo lifecycle on a stateless or
>> stateful session bean.  The bean class doesn't have to import any
>> javax.ejb classes, interfaces or annotations at all.
>>
>> The bean must be configured via the ejb-jar.xml.  We still  
don't have
>> support for the related annotations @PostConstruct, @PreDestroy,
>> @PrePassivate, and @PostActivate (OPENEJB-216).  Nor full  
dependency
>> injection (OPENEJB-98), nor for business local or business remote
>> interfaces quite yet (OPENEJB-90).
>>
>> Getting the business interfaces to work will be the next thing.
>>
>> -David
>>




--
Paulo Lopes
www.scratchydreams.com