After more investigation, the RESTORE_VIEW hack I was thinking of wouldn't be particularly nice. Since the components are created during that phase, they would never get the beforePhase event, only the after (so they would assume construction would be the before, and after would be triggered via the lifecycle).
Anyway, what is the MyFaces stance on new additions to the JSF spec? Are we waiting until things are finalized, or are proposed standards being implemented? If the former, there's no reason we couldn't easily put the reconstitute methods into the UIComponent.
-- Jon
On Apr 12, 2005, at 6:10 PM, Jon Travis wrote:
Ok, further thinking aloud. The aliasbean I was using was managing my bindings.
It appears that the problem lies in the fact that I'm using the alias bean in a component binding= attribute.
I imagine that since there is no component method in the version of the JSF spec that MyFaces is using for processReconstitutes(), then things like binding aren't going to be able to use aliasbeans. As an interim hack (until MyFaces implements 1.2), can we just hook into the RESTORE_VIEW phase and alias the bean there as well?
-- Jon
On Apr 12, 2005, at 5:40 PM, Jon Travis wrote:
Should probably clarify that I actually have x:aliasBean nested within another x:aliasBean.
Reducing to just 1 seems to work.
-- Jon
On Apr 12, 2005, at 5:00 PM, Jon Travis wrote:
I'm experiencing a problem using the alias bean.
My JSP roughly looks like:
<h:form> <x:aliasBean> <f:subview> <t:insert page=.../> </f:subview> </x:aliasBean> </h:form>
The inserted page contains input components. When I submit the values, I get a binding error. The problem seems to lie in the fact that the alias bean has removed the binding before the EL to set the value is run, thus giving me a binding error.
Is using tiles in this scenario supported? Does aliasBean work for this? The examples only show output components.
-- Jon
