Affects Version/s: 2.0.1
Fix Version/s: None
Operating System: All
-------- Original Message --------
Subject: Re: Question about JSF issue
Date: Wed, 14 Jul 2010 09:40:30 -0600
From: David Schneider <firstname.lastname@example.org>
To: Ed Burns <email@example.com>, firstname.lastname@example.org
CC: Andy Schwartz <email@example.com>, Sheetal Vartak
Roger, my concern is this appears to be a big break in backward
compatibility between JSF v1.2 and v2.0 and one that can have a
significent negative impact on the Oracle ADF framework. If at all
possible I'd like to see this inconsistency resolved in favor of
preserving the JSF v1.2 restriction that FacesContext.setUIViewRoot()
can only be called during the INVOKE_APPLICATION phase, or at least
prohibiting it from being called during the RENDER_RESPONSE phase. ADF
relies on the JSF v.12 restriction and having the UIViewRoot change
during the RENDER_RESPONSE phase is something ADF can't support.
On 07/14/2010 08:45 AM, Ed Burns wrote:
> AS> On 7/13/10 2:08 PM, David Schneider wrote:
> DS> What ever happened with the issue about the new JSF 2.0
> DS> ConfigurableNavigationHandler allowing navigation during the
> DS> RENDER_RESPONSE phase while FacesContext.setViewRoot() is only allowed
> DS> to be called during the INVOKE_APPLICATION phase? Allowing the
> DS> UIViewRoot to change during the RENDER_RESPONSE phase is a potential
> DS> show-stopper for Oracle ADF.
>>>>>> On Tue, 13 Jul 2010 14:45:23 -0400, Andy
> AS> Hey Dave -
> AS> We've got two spec issues relating to this. I logged this one:
> AS> https://javaserverfaces-spec-public.dev.java.net/issues/show_bug.cgi?id=757
> AS> To track the fact that there is inconsistency between what the spec says
> AS> about when setViewRoot() can be called vs. what the implementation does
> AS> (and what we are encouraging end users to do with PreRenderEvents).
> AS> Dan logged this one:
> AS> https://javaserverfaces-spec-public.dev.java.net/issues/show_bug.cgi?id=758
> AS> To request the desired solution, ie. view actions that are invoked
> AS> during the invoke application phase.
> AS> I'll defer to Ed on time frame for when these might be addressed.
> Yes, I am passing the buck on this one, but you know how it is with
> respect to resources. Andy and Dave, Roger owns issue 757, and Sheetal
> owns issue 758. Roger, because you are the more senior of the two, I'll
> leave it to you to decide how best to handle this. Roger, I suggest you
> take ownership of 758 from Sheetal so that Andy and Dave only have to
> coordinate with you.
> AS> Note that, as we discussed previously, ADF will need to throw exceptions
> AS> in response to navigation attempts after invoke application.
> Andy, please let me know if this resolution is not satisfactory.