I agree with your view on keeping Wicket from being bloated. What if the
IChoiceRenderer was optional? If you provide it you have automation. If
you do not it works as it does now. 

It's strange that that kind of functionality is provided when multiple
selections is needed (i.e. CheckBoxMultipleChoice), but it is not when
only one choice is needed (i.e. RadioGroup/CheckGroup). It is an oddity
that other components that have choices (i.e. DropDownChoice, etc.)
provide a means to use an IChoiceRenderer, but some do not. What if
someone is using a RadioGroup/CheckGroup that uses some form of a
RepeatingView to generate the Radio/Check options dynamically? In the
case where they need to make a selection using a separate choice
instance than what is in the list, how would they accomplish that? If
they set the choice on the RadioGroup/CheckGroup model it will never be
selected because they are different instances, and because there is not
IChoiceRenderer it is not possible to define an ID value to determine
the equality of the two instances.

-----Original Message-----
From: Igor Vaynberg [mailto:[EMAIL PROTECTED] 
Sent: Thursday, May 15, 2008 5:50 PM
To: users@wicket.apache.org
Subject: Re: IChoiceRenderer: RadioGroup CheckBoxMultipleChoice

well, the whole point of radio/check group is to allow user complete
control over markup. the whole point of choice renderer is to automate
markup generation, so there is just a big mismatch. eg i always use
radio/check group when using a choice renderer would be inconvenient.

you can always roll your own subclass, i just dont think something like
that would belong in core, its just one more parallel way of doing
something.

-igor

On Thu, May 15, 2008 at 2:44 PM, Hoover, William <[EMAIL PROTECTED]>
wrote:
> yes... sorry CheckGroup is what I meant. It seems as though 
> RadioGroup/CheckGroup should also have the capability of using an 
> IChoiceRenderer as well, right? One reason I was thinking that is if 
> someone has a model object instance A that is a different instance 
> than any of the choices in the list, they can override the
> IChoiceRenderer#getIdValue(...) and provide the identifier. Even 
> though none of the choices are the same instance of A they can still 
> determine equality for selection purposes using the ID value.
>
> -----Original Message-----
> From: Igor Vaynberg [mailto:[EMAIL PROTECTED]
> Sent: Thursday, May 15, 2008 4:47 PM
> To: users@wicket.apache.org
> Subject: Re: IChoiceRenderer: RadioGroup CheckBoxMultipleChoice
>
> checkboxmultiplechoice uses ichoicerenderer afaik, did you mean 
> checkgroup?
>
> radiogroup/checkgroup work differently, they leave all text generation

> up to the user so no choice renderer is required.
>
> -igor
>
> On Thu, May 15, 2008 at 1:01 PM, Hoover, William <[EMAIL PROTECTED]>
> wrote:
>> For consistency sake, wouldn't it be wise to use IChoiceRenderer for 
>> RadioGroup and CheckBoxMultipleChoice? Seems like a natural solution 
>> to override IChoiceRenderer#getIdValue(...) to infer IDs for 
>> selection
>
>> comparison the same way that it is being done in DropDownChoice.
>> Otherwise, how else can you use two different model object instances 
>> that share the same ID to be selectable?
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [EMAIL PROTECTED]
>> For additional commands, e-mail: [EMAIL PROTECTED]
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to