It's going to be JSF 2.0/Portal 2.0. I'm not sure if the JCP is going to look at a Portal 1.0 Spec for JSF 2.0. The leanings on the current EG's were that they were not. Portal 1.0 has some pretty major issues in dealing with AJAX and whatnot so such a specification would be problematic.

As for the VDL, simply put the current JSP VDL casts to some servlet objects (at least in the R.I.) to do some things. ;) <yoda>ClassCastExceptions do not an effective bridge make..</yoda>

The current bridge overrides the ViewHandler.renderView to overcome this problem, but I thought it might be nice to synchronize some code up. As it is right now, I'm merging the code from the ViewHandler.renderView in the Portlet 2.0 bridge with the current R.I. There are some strange issues which I'm overcoming, but evenutally it might be nice to allow MyFaces development to continue to drive these view handlers.

<shrug>

Just a thought..

Scott

Ganesh wrote:
Cool, you're working on this. Just a week ago I was stuck with
the current portal - JSF2.0 problem. Are you doing portal 1.0 -
JSF 2.0 or portal 2.0 - JSF 2.0?

I not yet clear about why the bridge needs to care about the
VDL. I thought it would suffice to brige the portal lifecyle
phases and forward the requests to the faces servlet?

Best regards,
Ganesh

Scott O'Bryan schrieb:
Hey Guys,

I'm working on a preliminary version of the portlet-bridge for JSF2.0.. Looking at the current R.I. implementations, it appears as if I'm going to have to come up with my own implementations for the ViewDeclairationLanguage's for the bridge. Although the R.I. is laid out so that their implementations of the ViewDeclairationLanguage's is easily extended, everything is impl. Since the Portlet Bridge is an Apache project and should be container agnostic, I'm basically stuck with two choices:

   1. Write the bridge's own implementation of the
      ViewDeclairationLanguage for both JSF and facelets, or
   2. Just use/extend the ViewDeclairationLanguage for MyFaces from the
      shared project

I would rather do the latter so that the implementations of the ViewDeclairationLanguage becomes consistent and, ideally, would allow us to work with both the R.I. and MyFaces (albeit with the MyFaces code handling the ViewDeclairationLanguage in the portal.

Any preferences or comments of the feasibility of implementing this? I haven't looked at this in depth but wanted to gauge people's reactions before I went too far down the rabbit hole.

Scott

Reply via email to