[
https://issues.apache.org/jira/browse/WICKET-1411?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
David Shepherdson closed WICKET-1411.
-------------------------------------
Thank you -- I can confirm that the fix works as of the 1.3.3 release.
> FormTester doesn't initialise values for RadioGroups
> ----------------------------------------------------
>
> Key: WICKET-1411
> URL: https://issues.apache.org/jira/browse/WICKET-1411
> Project: Wicket
> Issue Type: Bug
> Affects Versions: 1.3.0-final
> Reporter: David Shepherdson
> Assignee: Igor Vaynberg
> Fix For: 1.3.3
>
>
> In the constructor for FormTester, a visitor runs through all the form
> components and sets their initial values in the request. However, the visitor
> only deals with components that are instances of specific classes: currently
> AbstractTextComponent, DropDownChoice, RadioChoice, CheckBox,
> ListMultipleChoice and CheckGroup.
> Notably absent from that list is RadioGroup; if a RadioGroup is present in
> the form, it won't have any value set in the request and therefore when the
> form is submitted, if it's a required field the form validation will fail.
> We're working around this by adding the following code to the
> onFormComponent(FormComponent) method of the listener:
> else if (formComponent instanceof RadioGroup)
> {
> final String value = formComponent.getValue();
> Object choiceValue =
> formComponent.visitChildren(Radio.class, new IVisitor()
> {
> public Object component(Component component)
> {
> // O-P Preserve old escaping value, then turn
> escaping off
> // so that values aren't escaped unnecessarily.
> boolean oldEscaping =
> component.getEscapeModelStrings();
> component.setEscapeModelStrings(false);
> Radio radio = (Radio)component;
> String radioValue = radio.getValue();
> // O-P Restore the previous escaping setting.
> component.setEscapeModelStrings(oldEscaping);
> if
> (radio.getModelObject().toString().equals(value))
> {
> return radioValue;
> }
> return CONTINUE_TRAVERSAL;
> }
> });
> if (choiceValue != null) {
> setFormComponentValue(formComponent, (String)
> choiceValue);
> }
> }
> (See issue WICKET-1094 for the reason for the 'O-P' comments about escaping,
> though I'm not sure that they're really needed in this case.)
> Would it also be sensible to add a fallback case; something like the
> following?
> else
> {
> setFormComponentValue(formComponent,
> formComponent.getValue());
> }
> That would deal with any user-defined components that are subclasses of
> FormComponent but not of one of the existing Wicket components. (Though I'm
> not sure how common that is; we haven't done it ourselves.)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
