On 10/20/11 2:13 PM, Michael Freedman (Created) (JIRA) wrote:
Add explicit exclusions for trinidad in 329 branch
---------------------------------------------------

                  Key: PORTLETBRIDGE-221
                  URL: https://issues.apache.org/jira/browse/PORTLETBRIDGE-221
              Project: MyFaces Portlet Bridge
           Issue Type: Bug
           Components: Impl
     Affects Versions: 2.0.0
             Reporter: Michael Freedman
             Assignee: Michael Freedman


We are running into problems with a user of the 329 branch codeline when using with extra 
Faces (componentkit) extensions.  The problem is the bridge is updating the bridge scope 
at the end of the render request pushing all new attributes created during the render 
into the scope.  These extensions are expecting this so numerous request scope attrs 
which are "flags" that something is initialized/done are inappropriately being 
carried forward and used causing malfunctions.  At the moment we don't want to just 
blanket reverse the code from saving the scope at the end of the render -- instead 
deciding the safest thing is to just exclude the specific packages that might exist.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira


I don't know why we wouldn't blanket reverse this change made in PORTLEETBRIDGE-221. The current bridge behavior is causing request-scoped attributes to live longer than a request. This appears to violate the spirit of section 5.1.2 of the bridge specification:

"Specifically, the bridge must implement a notion of a bridge request scope that supplies the expected Faces semantics."

And the letter of the specification:

"Upon completion of each portlet render, do not preserve any changes in the request scope back into the bridge request scope except those related to internal bridge operations (e.g. VIEW_STATE_PARAM management). In the portlet (and Faces) model, render is supposed to be idempotent."

-- Blake Sullivan

Reply via email to