I reported a Seam2 bug to Richfaces that turned out to be a Seam bug
and apparently already fixed in CVS (http://jira.jboss.org/jira/browse/JBSEAM-1947)
(thanks Sergey and Pete!)
I have a related question: another change in JSF ID names have broken our selenium test and many of our Javascript functions which refer to input controls by ID.
Given something like:
<s:decorate id="fooDecoration" template="/WEB-INF/facelets/templates/edit.xhtml"> <ui:define name="label">Foo</ui:define> <h:inputText id="foo" value="#{fooHome.instance.foo}"/> </s:decorate>