users@javaserverfaces-spec-public.java.net

[jsr344-experts mirror] [jsr344-experts] Re: [949-WindowId] Best place to specify creating the windowId?

From: Werner Punz <werner.punz_at_gmail.com>
Date: Wed, 29 Feb 2012 11:38:10 +0100

Hello Ed regarding the actions you required from me

- Update jsf.ajax.response() to cover <update id="javax.faces.WindowId:0" />

a) There I personally think it must also contain the viewroot
like we do it for the viewstate
aka viewroot:divider:windowid, because there can be scenarii with different
window ids per window (subscopes, two different scopes per page thanks to
portlets)
outside of that I am perfectly fine with a new update element.


- Add jsf.getWindowId(form)
I personally think your approach covers only a subset of all possible
usecases, take my code as discussion base. I have covered everything
Leonardo has mentioned and multiple form scenarii)

My approach simply is that I deal with an arbitrary root element and
determine the child forms, if no windowId element is found I revert to the
url als last fallback.
If two non matching window elements are found an error is raised.
The frameworks then have to deal themselves with which root element they
take (hence offloading the viewroot handling to the users in that case).
Your form approach is a subset of all this.


Werner



ACTION: I'm aware that Werner is working on this and I want him to look
at what I've done here and suggest how I can improve it.

On Thu, Feb 23, 2012 at 11:55 PM, Edward Burns <edward.burns_at_oracle.com>wrote:

> >>>>> On Sun, 19 Feb 2012 21:09:20 -0500, Leonardo Uribe <lu4242_at_gmail.com>
> said:
>
> LU> The problem here is we want to fix Flash scope with windowId
> LU> concept, and that initialization code occur before any phase
> LU> listener. The only way to fix it correctly is inserting
> LU> initialization code before flash scope init. This is one reason why
> LU> windowId should be include at spec level.
>
> I see. I interpret your above text as a core requirement of the feature
> and have added this text to the "Invariants" section of the proposal
> [1].
>
> * The windowId must be the anchor to which the Flash is hitched.
>
> Leonardo, have I accurately understood and applied your meaning?
>


> Snip....
>
> --
> | edward.burns_at_oracle.com | office: +1 407 458 0017
> | homepage: | http://ridingthecrest.com/
>