Re: EJB contextual reference (our proxy) in creationalContext instead of instance?

2010-07-19 Thread Gurkan Erdogdu
I'm still confused by this.  Do you agree terminology wise that our
own  javasist proxy is a contextual reference and that the proxy from
EJB  is our (underlying) contextual instance? 
Yes.

For EJB at least, wouldn't keeping the contextual reference instead of
the  instance cause us to go create EJB's from our context just in
order  to @Remove them?  I
All of the things are done using contextual reference, including calling 
@Remove. Thing is that
Altough it is someway ridicilous, any method is called follows the below 
execution path, 


Context Ref -- EJB Proxy Instance -- EJB Instance


Gurkan



From: Eric Covener cove...@gmail.com
To: dev@openwebbeans.apache.org
Sent: Sun, July 18, 2010 6:19:39 PM
Subject: Re: EJB contextual reference (our proxy) in creationalContext instead  
of instance?

On Sat, Jul 17, 2010 at 6:37 PM, Gurkan Erdogdu gurkanerdo...@yahoo.com wrote:
 I think that in your case : RequestScope is ended and @RequestScoped bean is
 destroyed by the container. When this bean instance is destroyed by the
 container, container collects all of its dependencies from 
CreationalContextImpl
 and destroy all of them. In your case, it seems that this is the Stateful EJB
 instance dependency. To destroy stateful instance, container calls remove
 method.

I'm still confused by this.  Do you agree terminology wise that our
own javasist proxy is a contextual reference and that the proxy from
EJB is our (underlying) contextual instance?  The latter may or may
not need to be created, based on whether the injected reference is
actually used.  But it seems like we store the contextual reference
itself, instead of the instance, in the dependent creational context.

For EJB at least, wouldn't keeping the contextual reference instead of
the instance cause us to go create EJB's from our context just in
order to @Remove them?  I am not sure if other bean types have enough
going on in the destroyComponentInstance to care too much.

I may be way off in left field, as I am struggling with trying to get
my head around the code.  Any illumination you could provide is much
appreciated!

 at 
org.apache.webbeans.sample.ejb.Echo_$$_javassist_3.remove(Echo_$$_javassist_3.java)

 at 
org.apache.webbeans.util.ClassUtil.callInstanceMethod(ClassUtil.java:533)
 at 
org.apache.webbeans.ejb.common.component.BaseEjbBean.destroyComponentInstance(BaseEjbBean.java:149)

 at 
org.apache.webbeans.component.AbstractInjectionTargetBean.destroyInstance(AbstractInjectionTargetBean.java:179)

 at 
org.apache.webbeans.component.AbstractOwbBean.destroyCreatedInstance(AbstractOwbBean.java:265)

 at 
org.apache.webbeans.portable.creation.InjectionTargetProducer.preDestroy(InjectionTargetProducer.java:135)

 at 
org.apache.webbeans.component.InjectionTargetWrapper.preDestroy(InjectionTargetWrapper.java:104)

 at 
org.apache.webbeans.component.AbstractOwbBean.destroy(AbstractOwbBean.java:219)
 at 
org.apache.webbeans.context.creational.CreationalContextImpl.removeDependents(CreationalContextImpl.java:278)

 at 
org.apache.webbeans.context.creational.CreationalContextImpl.release(CreationalContextImpl.java:349)

 at 
org.apache.webbeans.component.AbstractOwbBean.destroy(AbstractOwbBean.java:232)

 at 
org.apache.webbeans.sample.ejb.Echo_$$_javassist_3.remove(Echo_$$_javassist_3.java)


 at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
 at 
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:48)
 at 
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:37)

 at java.lang.reflect.Method.invoke(Method.java:600)
 at 
org.apache.webbeans.util.ClassUtil.callInstanceMethod(ClassUtil.java:533)
 at 
org.apache.webbeans.ejb.common.component.BaseEjbBean.destroyComponentInstance(BaseEjbBean.java:149)

 at 
org.apache.webbeans.component.AbstractInjectionTargetBean.destroyInstance(AbstractInjectionTargetBean.java:179)


-- 
Eric Covener
cove...@gmail.com




Re: ... then no parameters are passed to the method by the container

2010-07-19 Thread Gurkan Erdogdu
You are so right that Spec is not so clear about those kind of statements


--Gurkan





From: Eric Covener cove...@gmail.com
To: dev@openwebbeans.apache.org
Sent: Mon, July 19, 2010 3:26:38 AM
Subject: ... then no parameters are passed to the method by the container

3.2.1
  If the application directly calls an EJB remove method of a
contextual instance of a session bean that is a stateful session bean
and has scope @Dependent then no parameters are passed to the method
by the container

How can the container opt to not pass parameters to a remove method
(whose signature has parameters) called by the application?


-- 
Eric Covener
cove...@gmail.com




[jira] Created: (OWB-406) BaseEjbBean.removedStatefulInstance used by multiple instances/EjbBeanProxyHandlers

2010-07-19 Thread Eric Covener (JIRA)
BaseEjbBean.removedStatefulInstance used by multiple 
instances/EjbBeanProxyHandlers
---

 Key: OWB-406
 URL: https://issues.apache.org/jira/browse/OWB-406
 Project: OpenWebBeans
  Issue Type: Bug
  Components: Enterprise Web Beans
Affects Versions: 1.0.0-alpha-1
 Environment: OWB + tomcat EJB integration
Reporter: Eric Covener
Assignee: Gurkan Erdogdu


BaseEjbBean.removedStatefulInstance is used by MethodHandlers to track when a 
remove method on a session bean has been called, but there is only one one of 
these Contextuals per bean class not per instance.

BaseEjbBean needs to manitain a map or the tracking needs to move into the 
Method Handler (EjbBeanProxyHandler)

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.



[jira] Created: (OWB-407) detailed information about exceptions

2010-07-19 Thread Gerhard Petracek (JIRA)
detailed information about exceptions
-

 Key: OWB-407
 URL: https://issues.apache.org/jira/browse/OWB-407
 Project: OpenWebBeans
  Issue Type: Improvement
Reporter: Gerhard Petracek
Assignee: Gurkan Erdogdu


owb should provide more information in case of invalid constellations.

e.g. UnproxyableResolutionException shows way to few information about the real 
problem. all required information would be available (in this case in 
WebBeansUtil#checkUnproxiableApiType).
detailed information about exceptions would help a lot during dev.-time.


-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.



[jira] Created: (OWB-408) NPE in WebBeansELResolver

2010-07-19 Thread Gerhard Petracek (JIRA)
NPE in WebBeansELResolver
-

 Key: OWB-408
 URL: https://issues.apache.org/jira/browse/OWB-408
 Project: OpenWebBeans
  Issue Type: Bug
Reporter: Gerhard Petracek
Assignee: Gurkan Erdogdu


OwbElContextListener gets called in combination with myfaces-core but it's 
ignored in combination with mojarra (tested with mojarra 2).
that leads to a NPE in WebBeansELResolver (there is no instance of 
ELContextStore)
even though it looks like an issue of mojarra it would be nice to have a 
fallback or a completely different solution for it.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.



Fwd: JCDI and Bean Validation TCKs

2010-07-19 Thread David Blevins
If you have any thoughts, post 'em! :)

Begin forwarded message:

 Resent-From: dblev...@visi.com
 From: David Blevins david.blev...@visi.com
 Date: July 19, 2010 9:44:53 PM PDT
 To: d...@geronimo.apache.org
 Subject: JCDI and Bean Validation TCKs
 Reply-To: d...@geronimo.apache.org
 
 Currently we have these setup in the private svn where the Java EE TCK 
 porting modules live.  The JCDI and Bean Validation TCKs are public Apache 
 licensed, so we could move the porting modules for those two TCKs into our 
 public svn.
 
 The part of my brain that finds esthetic pleasure in filling cabinets likes 
 everything all organized in the one VM, but the part of me that likes to be 
 more public than private thinks it's unnecessarily restrictive to make people 
 sign the Sun/Apache NDA to get access to things not under that restriction.  
 Specifically, everyone in the related communities (OpenWebBeans, OpenEJB) 
 could easily access the public TCKs.  Mark and Gurkan fall into that category 
 now.  Both are in the process of getting NDAs filled, but we could definitely 
 speed that up by opening the porting code to the public.
 
 We might even be able to work the JCDI and Bean Validation into our larger 
 test suite right in the main build as they only take a moments to run.
 
 Thoughts?
 
 -David