I've opened a discussion on the JSR-RI dev mailing list on this topic.
So far, the argument against it is weak -- "performance hit" -- which
is just an implementation detail. However, someone may jump in with
more compelling reasons.
At minimum, I'd love to see this as a MyFaces configurable option, for
which I'd be willing to provide a "performant" patch.
-Mike
On 8/9/05, Craig McClanahan <[EMAIL PROTECTED]> wrote:
> On 8/9/05, Mike Kienenberger <[EMAIL PROTECTED]> wrote:
>
> > Yes, the current RI does the same thing, but the only "documentation"
> > of the behavior is a comment in a private message
> >
> > // if converterType is null, String, or Object, assume no
> > conversion is needed
> >
> > It's not even stated as a rule, but merely as an "assumption."
> >
>
> I don't recall the details on this one either, but I can't imagine how
> this would matter unless you actually tried to register and use such a
> Converter. It definitely seems like something that should be reported
> as an RFE to the RI ("it should be possible to use String or Object
> converters") and/or as a request to the JSF 1.2 Expert Group to
> clarify that it should be disallowed -- with an obvious preference for
> the former.
>
> As to what MyFaces should do, it's really you guys's call. I can't
> find anything in the spec or Javadocs that says this behavior is
> disallowed. My only caution is that sometimes "gratuitous
> compatibility" with the RI where the spec is silent is *sometimes* a
> good thing, as it reduces the potential for problems when people
> migrate. On the other hand, you might look at this as an opportunity
> to provide a feature that the RI doesn't.
>
> Craig
>
