Re: svn commit: r985746 - in /openwebbeans/trunk: webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ webbeans-openejb/

2010-08-16 Thread Gurkan Erdogdu
Hi Eric,

Actually OWB returns proxy to client that implements clients preference 
interface when using getReference of BeanManager. This proxy internally uses 
EJB 
proxy instance that implements all of local interfaces. 


Thanks,

--Gurkan



From: Eric Covener cove...@gmail.com
To: dev@openwebbeans.apache.org
Sent: Sun, August 15, 2010 11:43:39 PM
Subject: Re: svn commit: r985746 - in /openwebbeans/trunk: 
webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ 
webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ 
webbeans-openejb/src/main/java/org/apache/webbeans/ejb/ webbeans-op

On Sun, Aug 15, 2010 at 4:25 PM,  gerdo...@apache.org wrote:
 Author: gerdogdu
 Date: Sun Aug 15 20:25:31 2010
 New Revision: 985746

 URL: http://svn.apache.org/viewvc?rev=985746view=rev
 Log:
 [OWB-439]  EjbPlugin session bean proxy creation thread safe problem


I had a little difficulty following, especially as this went into the
OpenEJB API.  Does this basically make each ENTERPRISE bean return
instances with all of the local interfaces as types?




Re: svn commit: r985746 - in /openwebbeans/trunk: webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ webbeans-openejb/

2010-08-16 Thread Gurkan Erdogdu
This proxy internally uses EJB 
proxy instance that implements all of  local interfaces. 
This may be internally handled by the OpenEJB.

Look org.apache.openejb.core.ivm.EjbHomeProxyHandler class in 
container/openejb-core project.





From: Gurkan Erdogdu gurkanerdo...@yahoo.com
To: dev@openwebbeans.apache.org
Sent: Mon, August 16, 2010 8:59:39 AM
Subject: Re: svn commit: r985746 - in /openwebbeans/trunk: 
webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ 
webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ 
webbeans-openejb/src/main/java/org/apache/webbeans/ejb/ webbeans-op

Hi Eric,

Actually OWB returns proxy to client that implements clients preference 
interface when using getReference of BeanManager. This proxy internally uses 
EJB 

proxy instance that implements all of local interfaces. 


Thanks,

--Gurkan



From: Eric Covener cove...@gmail.com
To: dev@openwebbeans.apache.org
Sent: Sun, August 15, 2010 11:43:39 PM
Subject: Re: svn commit: r985746 - in /openwebbeans/trunk: 
webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ 
webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ 
webbeans-openejb/src/main/java/org/apache/webbeans/ejb/ webbeans-op

On Sun, Aug 15, 2010 at 4:25 PM,  gerdo...@apache.org wrote:
 Author: gerdogdu
 Date: Sun Aug 15 20:25:31 2010
 New Revision: 985746

 URL: http://svn.apache.org/viewvc?rev=985746view=rev
 Log:
 [OWB-439]  EjbPlugin session bean proxy creation thread safe problem


I had a little difficulty following, especially as this went into the
OpenEJB API.  Does this basically make each ENTERPRISE bean return
instances with all of the local interfaces as types?



Re: svn commit: r985746 - in /openwebbeans/trunk: webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ webbeans-openejb/

2010-08-16 Thread Eric Covener
On Mon, Aug 16, 2010 at 2:06 AM, Gurkan Erdogdu gurkanerdo...@yahoo.com wrote:
This proxy internally uses EJB
proxy instance that implements all of  local interfaces.
 This may be internally handled by the OpenEJB.

 Look org.apache.openejb.core.ivm.EjbHomeProxyHandler class in
 container/openejb-core project.

FYI I've confirmed that the container I'm working with really needs
the local interface when there's more than 1 local interface on the
EJB class, so I will try to introduce a safe mechanism to pass the
injected iface down.


[jira] Resolved: (OWB-430) improve performance of WebBeansPhaseListener

2010-08-16 Thread Gerhard Petracek (JIRA)

 [ 
https://issues.apache.org/jira/browse/OWB-430?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Gerhard Petracek resolved OWB-430.
--

Fix Version/s: 1.0.0-alpha-2
   Resolution: Fixed

 improve performance of WebBeansPhaseListener
 

 Key: OWB-430
 URL: https://issues.apache.org/jira/browse/OWB-430
 Project: OpenWebBeans
  Issue Type: Improvement
  Components: Java EE Integration
Affects Versions: 1.0.0-alpha-1
Reporter: Gerhard Petracek
Assignee: Gurkan Erdogdu
Priority: Minor
 Fix For: 1.0.0-alpha-2

 Attachments: OWB-430.patch




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



[jira] Updated: (OWB-428) implementation of equals and hashCode for AbstractOwbBean

2010-08-16 Thread Gerhard Petracek (JIRA)

 [ 
https://issues.apache.org/jira/browse/OWB-428?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Gerhard Petracek updated OWB-428:
-

Attachment: OWB-428.patch

 implementation of equals and hashCode for AbstractOwbBean
 -

 Key: OWB-428
 URL: https://issues.apache.org/jira/browse/OWB-428
 Project: OpenWebBeans
  Issue Type: Task
  Components: Core
Affects Versions: 1.0.0-alpha-1
Reporter: Gerhard Petracek
Assignee: Gurkan Erdogdu
 Fix For: 1.0.0-alpha-2

 Attachments: OWB-428.patch, OWB-428.patch




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



Re: svn commit: r985746 - in /openwebbeans/trunk: webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/component/ webbeans-ejb/src/main/java/org/apache/webbeans/ejb/common/util/ webbeans-openejb/

2010-08-16 Thread Eric Covener
On Mon, Aug 16, 2010 at 11:20 AM, Eric Covener cove...@gmail.com wrote:
 On Mon, Aug 16, 2010 at 2:06 AM, Gurkan Erdogdu gurkanerdo...@yahoo.com 
 wrote:
This proxy internally uses EJB
proxy instance that implements all of  local interfaces.
 This may be internally handled by the OpenEJB.

 Look org.apache.openejb.core.ivm.EjbHomeProxyHandler class in
 container/openejb-core project.

 FYI I've confirmed that the container I'm working with really needs
 the local interface when there's more than 1 local interface on the
 EJB class, so I will try to introduce a safe mechanism to pass the
 injected iface down.


Realizing now that this is a bigger problem in actually retrieving the
ejb instances from the context -- they're not interchangeable in the
case that the container does not return EJBs capable of all local
interfaces.

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


[jira] Commented: (OWB-435) What is the expected result for following 2 decorators?

2010-08-16 Thread Joe Bergmark (JIRA)

[ 
https://issues.apache.org/jira/browse/OWB-435?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12899050#action_12899050
 ] 

Joe Bergmark commented on OWB-435:
--

Had a chance to dig into #1 today and there does appear to be a bug here.  At 
least in my testing, the call to 
ownerCreationalContext.getDependentDecorator(decorator); always returns null, 
so we end up creating a new Decorator instance each time another method is 
called on the longer lived class.  I'll go ahead and open a JIRA issue for this 
and try to determine why that is the case.

 What is the expected result for following 2 decorators?
 ---

 Key: OWB-435
 URL: https://issues.apache.org/jira/browse/OWB-435
 Project: OpenWebBeans
  Issue Type: Question
  Components: Interceptor and Decorators
Reporter: YING WANG
Assignee: Gurkan Erdogdu
Priority: Minor

 While I am testing 2 decorators decorate the same getName() method of 
 UserBean, I found the result is:
 1. UserDecorator1(UserDecorator2(MYNAME))==  Did call 
 UserDecorator1.getName() first, but before it finishes, it recursively 
 invokes the UserDecorator2.getName() on the calling stack.
 2. or  should the result be:
 UserDecorator2(MYNAME)   should decorator2's result overwrite 
 decorator1's?
 3. or should the result be:
 UserDecorator2(UserDecorator1(MYNAME))  should decorator1's result 
 to the one used for decorator2?
 I prefer 3, but I am not sure which result is the correct one
 ===Userbean 
 public class UserBean implements UserInterface, Serializable 
 {
 public String getName()
 {
   return MYNAME;
 }
 }
 ===UserDecorator1 
 @Decorator
 public abstract class UserDecorator1 implements UserInterface, Serializable 
 {
   @Inject @Delegate @Any UserInterface ui;
   
   public String getName() {
   return UserDecorator1( + ui.getName() + );
   }
 }
 ===UserDecorator2 
 @Decorator
 public abstract class UserDecorator2 implements UserInterface, Serializable 
 {
   @Inject @Delegate @Any UserInterface ui;
   
   public String getName() {
   return UserDecorator2( + ui.getName() + );
   }
 }
 
 decorators
 classcom.jcdi.test.UserDecorator1/class
 classcom.jcdi.test.UserDecorator2/class
 /decorators

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



[jira] Commented: (OWB-440) WebBeansDecoratorConfig.getDecoratorStack always returns new Decorators

2010-08-16 Thread Joe Bergmark (JIRA)

[ 
https://issues.apache.org/jira/browse/OWB-440?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12899082#action_12899082
 ] 

Joe Bergmark commented on OWB-440:
--

Solving this seems to be pretty straight forward.  CreationalContextImpl's   
getDependentDecorators() was previously always creating an empty list and 
iterating over that list.  Changing this to pass the instance and looking up 
the dependent creational contexts using by object instance allows us to find 
the dependent decorators.  This mimics the behavior the interceptor version of 
this method.

However, this change appears to break the MultipleDecoratorStackTest unit test, 
so I'm going to hold off until I can figure out if that test itself is broken, 
or if the DelegateHandler or surrounding code made some bad assumptions about a 
clean decorator stack each time which may no longer be true.

 WebBeansDecoratorConfig.getDecoratorStack always returns new Decorators
 ---

 Key: OWB-440
 URL: https://issues.apache.org/jira/browse/OWB-440
 Project: OpenWebBeans
  Issue Type: Bug
  Components: Interceptor and Decorators
Affects Versions: 1.0.0-alpha-1
Reporter: Joe Bergmark
Assignee: Joe Bergmark

 WebBeansDecoratorConfig.getDecoratorStack's call to 
 ownerCreationalContext.getDependentDecorator(decorator)
 always returns null, even if the owning bean is called multiple times within 
 its lifecycle.  Because of this, you always get a new instance of the 
 Decorators for each method call when you would expect Dependant scoped 
 instances to live as long as their parent.

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



[DISCusS] openwebbeans.configuration new

2010-08-16 Thread Mark Struberg
Hi!

I was on vacation over the weekend and had an idea about how to improve our 
configuration mechanism. I hacked the idea while flying back from Cologne and 
now all is working again.

The change is actually not really big, I just like to get rid of our 
openwebbeans-*.properties and replace them with the following mechanism:

There are still multiple property files for the same configuration, thus still 
allowing 'overriding' a configuration. But instead of manually looking for 
defined 'extensions' I just define a 'configuration.ordinal' inside the property

thus I define

webbeans-impl/ openwebbeans.properties with 'configuration.ordinal=10'
webbeans-web/ openwebbeans.properties with 'configuration.ordinal=11'
webbeans-jsf/ openwebbeans.properties with 'configuration.ordinal=12'

If a properties file doesn't define 'configuration.ordinal' then a value of 100 
is assumed.


The algorithm is easy:
.) load all properties you can find with the name 
.) sort them via configuration.ordinal in ascending order
.) overload them as we do already, but instead of some defined names we now 
just 
use the sorted list of properties.

WDYT?

It's all ready to get checked in :)

LieGrue,
strub


  


[jira] Resolved: (OWB-440) WebBeansDecoratorConfig.getDecoratorStack always returns new Decorators

2010-08-16 Thread Joe Bergmark (JIRA)

 [ 
https://issues.apache.org/jira/browse/OWB-440?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Joe Bergmark resolved OWB-440.
--

Fix Version/s: 1.0.0-alpha-2
   Resolution: Fixed

Fixed the last remaining problem from the unit test.  When we found the 
existing decorator in the creational context we weren't updating the delegate 
so the old one would mistakenly get reused and we would lose our position in 
the decorator stack.

 WebBeansDecoratorConfig.getDecoratorStack always returns new Decorators
 ---

 Key: OWB-440
 URL: https://issues.apache.org/jira/browse/OWB-440
 Project: OpenWebBeans
  Issue Type: Bug
  Components: Interceptor and Decorators
Affects Versions: 1.0.0-alpha-1
Reporter: Joe Bergmark
Assignee: Joe Bergmark
 Fix For: 1.0.0-alpha-2


 WebBeansDecoratorConfig.getDecoratorStack's call to 
 ownerCreationalContext.getDependentDecorator(decorator)
 always returns null, even if the owning bean is called multiple times within 
 its lifecycle.  Because of this, you always get a new instance of the 
 Decorators for each method call when you would expect Dependant scoped 
 instances to live as long as their parent.

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