Hi Andy,
To put you into context, that's what I found in the spec:
PT depends on the alias "jsf:"
1. if declared as of "
http://java.sun.com/jsf" <
http://java.sun.com/jsf> -->
accept
2. if not declared: treat as of "
http://java.sun.com/jsf"
<
http://java.sun.com/jsf> and accept
3. if declared as other than "
http://java.sun.com/jsf" <
http://java.sun.com/jsf>
throw facelet exception
4. The user may not define her own alias.
My thoughts: This kind of definition is not good.
1. ok
2. might be ok for ease of usage (I guess that's the intention of current spec,
but it's not standard for namespace usage)
3.bad. the user may define such a mapping for something else, e.g. in existing
pages
4. bad. the user may define a different alias pointing to
"
http://java.sun.com/jsf" <
http://java.sun.com/jsf> and it should work for PT
either
Thanks for supporting this.
Michael
Andy Schwartz <andy.schwartz_at_oracle.com> hat am 3. März 2013 um 14:05
geschrieben:
> Hi Michael -
>
> On 3/3/13 3:00 AM, Michael Müller wrote:
> > Hi Ed,
> >
> > Once again...
> >
> > We have to consider compatibility for existing apps. If someone uses
> > "p:" as alias for PrimeFaces, then she may define "pa:" for Pass
> > Through Attributes. Same applys for "jsf:". On the other hand, I
> > understand ease of use. Someone may use "jsf:" without declaring a
> > namespace.
>
> I haven't had a chance to read up on the context for this, but I wanted
> to chime in to say that the JSF spec should never associate behavior
> with namespace prefixes. The spec should only associate behavior with
> namepsaces. Sure, the spec can recommend conventions for namespace
> prefix mappings. However, our users should always be free to define
> their own mappings.
>
> If we have cases in the spec where we require that a namespace prefix
> has some meaning regardless of how that prefix is mapped, we need to fix
> this.
>
> Andy
>