We should also include the Enterprise Chat example.

Assigning to Jack.

The one thing that I wanted to do with this feature is make it more transparent. Right now we have the RenderManager but to switch to broadcast rendering, you have to use the BroadcastRenderer. I'd rather see us turn this into more of a switch that can turned on when necessary. It should be easier now that it's not a separate Enterprise feature.

There is also the new "SessionRenderer" API (ICE-2989). It would be nice if this API could also be used for cross-node deployments as well.

Message Service Client needs to be moved into the Core in order for the BroadcastHub to utilize its functionalities.

RenderManager API changes:

• com.icesoft.faces.async.render.broadcasted
• RenderManager.getInstance()
• RenderManager.isBroadcasted()
• RenderManager.setBroadcasted(boolean broadcasted)
• AsyncRenderer.isBroadcasted()
• AsyncRenderer.setBroadcasted(boolean broadcasted)

The property can be set inside an ICEfaces application's web.xml to enable or disable broadcasting. The RenderManager instance uses this property to set its default value of the broadcasted flag. This can be programmatically changed by using the setBroadcasted(boolean) method. Additionally, an AsyncRenderer can have its own value of the broadcasted flag for fine-grained control:

• if the RenderManager's flag is set to false broadcasting never happens, regardless of the AsyncRenderer's flag.
• if the RenderManager's flag is set to true, broadcasting happens if the AsyncRenderer's flag is set to true as well; otherwise broadcasting does not happen.

AsyncRenderer.requestRender() should still be used to request a render.

The icefacesRenderTopic JMS topic is used for broadcasting.

Marking this one as FIXED.