JBossDeveloper

mpavleski wrote:

 

You were right Jaikiran, it was related to the no-interface view.

When I changed the SFSB to Local interface view, the problem persisted.

When I changed the SLSB to Local interface view the problem did not occur.

Right, if the bean being injected (in this case the SLSB) is a no-interface view, then you run into this problem.

mpavleski wrote:

 

However, in the later case i found out that activation does not work (it throws an exception), to test try retrieve the instance by accessin

     http://localhost:8080/TestCDI/faces/test.xhtml?cid=1 (apply the right conversation id number)

 

 

 

Here is the code and the WAR with only Local views:

     http://community.jboss.org/servlet/JiveServlet/downloadBody/16472-102-1-121207/TestCDIver2.zip

 

I can't download that attachment. It gives me some authorization error. Can you post the entire exception stacktrace or attach it to your reply? (I'm not sure how you managed to upload those .zip files to a URL like that )

mpavleski wrote:

 

About JIRA issue, I already found similar and commented on it. It was posted for JBoss 6 M4 but I think it's the same problem:

 

     https://issues.jboss.org/browse/JBAS-8384

 

 

 

Although that issue is about passivation, it isn't really the same as this one.

mpavleski wrote:

 

I wonder how didn't anyone notice this simple case and how did something trivial as this got to a `Final` release.

Even "Final" release do have bugs. In this specific case, this "simple" case wasn't ever covered in any of the testcases that the AS was run against   So it never really got fixed.

I can't download that attachment. It gives me some authorization error. Can you post the entire exception stacktrace or attach it to your reply? (I'm not sure how you managed to upload those .zip files to a URL like that  )

Try now again I forgot to set it as public. I go and do publish article in the Community page.

Although that issue is about passivation, it isn't really the same as this one.

I don't know but I got those same messages when I was batteling this issue.

Even "Final" release do have bugs. In this specific case, this "simple" case wasn't ever covered in any of the testcases that the AS was run against   So it never really got fixed.

I didn't meant the testcases, I was wondering how didn't anyone notice this long...

Update:
Stacktrace of the activation error supplied.

Message was edited by: Miroslav Pavleski

I didn't meant the testcases, I was wondering how didn't anyone notice this long...

The bug is always obvious to the one who finds it ;-)

I wondered how come noone had ever tried to use an extended persistence context in a postconstruct method.