On Wed, Oct 21, 2009 at 9:36 AM, Jan-Kees van Andel <jankeesvanan...@gmail.com> wrote: > I've added those "constants". Can't remember why I made them public. I > assume because they are used by the BeanValidator class. > > The reason to make them static final with the static initializer block > was because it made them easy to use and also because it's very > efficient (performance wise).
well, we can't mess up with the dictated API (by the spec) > > Maybe they need to be refactored to some other class than UIInput? I > think so, especially since Bean Validation is not the only JSF2 > dependency. In the future we'll also have Web Beans and maybe JSR 330? > The amount of "is***Available" checks will definitely increase... we could have some package private stuff, which is fine. But we can't add other heavy dependencies to the API. As we have to pass - at some day - the TCK. -Matthias > > Regards, > Jan-Kees > > > 2009/10/21 Matthias Wessendorf <mat...@apache.org>: >> Nope, I checked the JavaDoc (I am not looking at the code, for a >> couple of reasons), >> and there is nothing like that. >> >> So, maybe there is need for such an helper, as there are already today >> - as a matter >> of practical use - some issues with the current spec. >> >> Let me file an ER for this. >> >> -Matthias >> >> On Wed, Oct 21, 2009 at 6:15 AM, Martin Marinschek >> <mmarinsc...@apache.org> wrote: >>> Well, we can definitely not add a public field to the API which is not >>> specified. The question is only if maybe the RI added some public >>> field which was not specified - then this could be a spec bug. >>> >>> regards, >>> >>> Martin >>> >>> On 10/20/09, Matthias Wessendorf <mat...@apache.org> wrote: >>>> I think the question is why there isn't a helper like >>>> BeanValidation.isBeanValidation() specified ? >>>> Implementation could be somewhat similar to what MyFaces >>>> does today. But I wonder why this wasn't done... >>>> >>>> -M >>>> >>>> On Mon, Oct 19, 2009 at 9:14 PM, Matthias Wessendorf <mat...@apache.org> >>>> wrote: >>>>> Hi, >>>>> >>>>> in order to work on Trinidad 2.0 and 303 support (yeah need to check >>>>> what Gerhard did for extVal) I >>>>> took a quick look at the UIInput. >>>>> >>>>> I noticed a few lines containing a call like this: >>>>> >>>>> => BeanValidator.isAvailable >>>>> >>>>> This is a public field which is only there in MyFaces, not in the JSF >>>>> 2.0 API ([1]). >>>>> >>>>> I know that the entire handling of checking if there is a JSR 303 >>>>> implementation is >>>>> a little bit ugly, but I am not sure if we should introduce such a >>>>> "public" field to the >>>>> BeanValidator validator class. >>>>> >>>>> => wouldn't be good if extensions would start to rely on this field; I >>>>> guess not there on the RI ;-) >>>>> => does this violate the TCK ? Or would it be ? I am not sure here >>>>> two, as it is a public field >>>>> >>>>> Maybe we should make the make the check become a private thing ? >>>>> >>>>> -Matthias >>>>> >>>>> >>>>> [1] http://java.sun.com/javaee/javaserverfaces/2.0/docs/api/index.html >>>>> >>>>> -- >>>>> Matthias Wessendorf >>>>> >>>>> blog: http://matthiaswessendorf.wordpress.com/ >>>>> sessions: http://www.slideshare.net/mwessendorf >>>>> twitter: http://twitter.com/mwessendorf >>>>> >>>> >>>> >>>> >>>> -- >>>> Matthias Wessendorf >>>> >>>> blog: http://matthiaswessendorf.wordpress.com/ >>>> sessions: http://www.slideshare.net/mwessendorf >>>> twitter: http://twitter.com/mwessendorf >>>> >>> >>> >>> -- >>> >>> http://www.irian.at >>> >>> Your JSF powerhouse - >>> JSF Consulting, Development and >>> Courses in English and German >>> >>> Professional Support for Apache MyFaces >>> >> >> >> >> -- >> Matthias Wessendorf >> >> blog: http://matthiaswessendorf.wordpress.com/ >> sessions: http://www.slideshare.net/mwessendorf >> twitter: http://twitter.com/mwessendorf >> > -- Matthias Wessendorf blog: http://matthiaswessendorf.wordpress.com/ sessions: http://www.slideshare.net/mwessendorf twitter: http://twitter.com/mwessendorf