On 8/15/06, joe_ersinghaus <[EMAIL PROTECTED]> wrote:
              <t:message id="msgError" for="inputText1" />
              <t:message id="msgError" for="inputText2" />

These would need to have different id values.  I'd recommend just
leaving the id tag off.


But I'm not sure how to handle the output to <t:message>. Is that something
the "validateRequiredIf" custom code would take care of?


No, that's the easy part.   A validator throws a ValidatorException()
and part of that exception is a FacesMessage.   So the validator would
generate the message while validating.


I'm still not sure that the design is such that all of the "how it
will work" details have been figured out.   Actually writing the
validator is easy enough.   Cloning the compareToValidator and
changing the condition checked is pretty trivial.



On 8/15/06, joe_ersinghaus <[EMAIL PROTECTED]> wrote:

Mike,

Sorry I've got back to you late on this. If you know how to create custom
validators (I haven't delved into that aspect of JSF yet), that would be
great. I can see with your dates how the "ne" operator might be beneficial,
for the radio button/text field, maybe not. This is the way I would set up
my app with the radio button/input fields:

<t:panelGroup>
            <t:selectOneRadio id="myRadio" value="#{myBean.myRadio}" >
              <f:selectItem itemValue="0" itemLabel="#{text['radio1']}" />
              <f:selectItem itemValue="1" itemLabel="#{text['radio2']}" />
            </t:selectOneRadio>

            <h:panelGrid columns="3">

              <t:radio for="myRadio" index="0" />
                  <validateRequiredIf for="inputText1" >
              </t:radio>
              <h:inputText id="inputText1" value="#{myBean.inputText1}" />
              <t:message id="msgError" for="inputText1" />

              <t:radio for="myRadio" index="1" />
                  <validateRequiredIf for="inputText2" >
              </t:radio>
              <h:inputText id="inputText2" value="#{myBean.inputText2}" />
              <t:message id="msgError" for="inputText2" />

            </h:panelGrid>
</t:panelGroup>

But I'm not sure how to handle the output to <t:message>. Is that something
the "validateRequiredIf" custom code would take care of?

Let me know if there's anything I can do. (Read up on custom validation for
one.)

- Joe



Mike Kienenberger wrote:
>
> On 8/9/06, Joe ersinghaus <[EMAIL PROTECTED]> wrote:
>> Is it possible to have more than one component validated as a group? For
>> example:  let's say I've have a radio button that when activated, text
>> must be entered/validated in that radio button's associated input field?
>>
>> In other words, only when the radio button is selected, should the input
>> filed be validated.
>
> It's occurring to me that I need something somewhat similar.  I need
> to verify that if a starting date is specified, a second ending date
> is also specified.  Up to this point, I've been doing the validation
> like this in my action method and binding the components to my backing
> bean.
>
> However, it'd really be more convenient to handle this as a real
> validator.
>
> So let's brainstorm how this would work, and I'll throw together
> something from a clone of my compareToValidator.
>
> Maybe something like validateRequiredIf for=<other component>
> operator=<an operator> value=
>
> So for my dates, I'd use
>
>   <inputCalendar id="startingDate">
>       <validateRequiredIf for="endingDate" operator="ne" value="null">
>   </inputCalendar>
>
>   <inputCalendar id="endingDate">
>       <validateRequiredIf for="startingDate" operator="ne" value="null">
>   </inputCalendar>
>
> I have to put it on both.   If I put it only on endingDate, then if
> endingDate was null, the validateRelationship would not be executed.
>
>
> So how would this work for your situation?  The validator has to be on
> a non-null input component.
>
>   <radiobutton id="radioComponent">
>       <validateRequiredIf for="associatedInput" operator="ne"
> value="null">
>   </radiobutton>
>   <inputText id="associatedInput">
>   </inputText>
>
> Seems ackward.   Maybe it needs a better name.   Maybe we could just
> drop the operator/value attributes (it'd be true for our two use
> cases).
>
>

--
View this message in context: 
http://www.nabble.com/many-components-to-one-%28or-more%29-validation--tf2081866.html#a5808674
Sent from the MyFaces - Users forum at Nabble.com.


Reply via email to