[jsr-314-open] More spec issues

[Simon Lessard]

PM > Why can't we say "if specified, the standard name will be used, if not 
PM > the backwards compatible name from Facelets will be read". I would 
PM > suggest an impl should log a WARN if the old one is used, but that is 
PM>  probably an impl detail only...

I like that.

~ Simon


________________________________

From: JSR 314 Open Mailing list on behalf of Pete Muir
Sent: Fri 3/27/2009 11:35 AM
To: JSR-314-OPEN at JCP.ORG
Subject: Re: [jsr-314-open] More spec issues



On 27 Mar 2009, at 15:15, Simon Lessard wrote:

> Hi all,
>
> I have a couple more issues I found while implementing the latest 
> spec:
>
>       * ViewHandler.FACELETS_VIEW_MAPPINGS_PARAM_NAME: The doc says that 
> the implementation should also read original Facelets' param for the 
> purpose of defining mappings. I don't think that's a good idea. 
> Facelets was not a standard up to that point and users shouldn't 
> really expect no migration effort at all. I know that enforcing that 
> param would allow xisting applications using facelet.jar to be 
> deployed in 2.0 environment, but to my knòwledge, no other spec read 
> some non standard context params (I may be wrong though). If we do 
> enforce the original parameters, then we'll be stuck with it forever 
> so at least let use only that one and not define a new one;

Why can't we say "if specified, the standard name will be used, if not 
the backwards compatible name from Facelets will be read". I would 
suggest an impl should log a WARN if the old one is used, but that is 
probably an impl detail only...


>       * ViewHandler.getPageDeclarationLanguage(): JavaDoc gives two 
> default implementations, one invoking the factory from FactoryFinder 
> (useful) and the throws UnsupportedOperationException one. 
> Personally, I'd keep only the real method here as the processing 
> doesn't involve any private/non standard API;
>       * AjaxBehavior.getHints(): JavaDoc says it should return an emprt 
> Set (like ClientBehaviorBase), making the method declaration in 
> AjaxBehavior useless. However Mojarra returns a Set containing 
> SUBMITTING, so I assume the code is right and the JavaDoc wrong here;
>       * UIComponentBase.initialStateMarked(): This method is not defined 
> in the JavaDoc while the two others from PartialStateHolder are.
>       * ExternalContext.getRequestCookieMap(): Can the return value be 
> changed from Map<String,Object> to Map<String, Cookie> please?
>       * PageDeclarationLanguageFactory.getPageDeclarationLanguage(String 
> viewId): The JavaDoc says that ClassNotFoundException should be 
> thrown when no valid PLD can be found. However, 
> ClassNotFoundException is a checked exception and the method does 
> not that this exception in its throws clause. The doc should either 
> says to wrap it in a FacesException or add ClassNotFoundException to 
> the throws clause;
>       * ExceptionHandler.handle(): The spec specifies that this method 
> should only throw a FacesException if an internal error occurs while 
> handling the exception. However, it also says to throw either 
> ServletException or PortletException depending on the environment 
> and those are checked exception not declared in the throws clause. 
> Mojarra wrap them in a FacesException, so the spec mighth ave to be 
> ajusted to reflect that.
> That's all for now,
>
>
> ~ Simon
>
>
>
> From: JSR 314 Open Mailing list on behalf of Ed Burns
> Sent: Fri 3/27/2009 10:30 AM
> To: JSR-314-OPEN at JCP.ORG
> Subject: Rename: PDL->VDL
>
> We have "View" everywhere else, so we should rename
> PageDeclarationLanguage to ViewDeclarationLanguage.
>
> I'm taking care of this now.
>
> Ed
> --
> | ed.burns at sun.com  | office: 408 884 9519 OR x31640
> | homepage:         | http://ridingthecrest.com/
>

--
Pete Muir
http://www.seamframework.org <http://www.seamframework.org/> 
http://in.relation.to/Bloggers/Pete