I think you're right on this. Just remember to set the Wicket-Ajax header on
all the different ajax requests (links (both ajax, ajaxfallback and
ajaxsubmit), ajaxbevahior execution).
This might break some users tests, but IMO it's worth it because those tests
might have been wrong anyway.
+1
Frank
On Dec 1, 2007 3:55 PM, Gerolf Seitz <[EMAIL PROTECTED]> wrote:
> another indication that there is something wrong is that
> AjaxRequestTargetTest#testHeaderContribution3
> expects that the wicket:id of the stylesheetreference is rendered, whereas
> according to
> Component#renderComponentTag, which sets stripWicketTags (for the
> methodcall
> tag.writeOutput) to true for ajaxrequests,
> wicket: attributes are not rendered.
>
> Gerolf
>
>
> On Dec 1, 2007 3:06 PM, Gerolf Seitz <[EMAIL PROTECTED]> wrote:
>
> > hi,
> > it seems like BaseWicketTester#clickLink does not generate an ajax
> > request.
> > excerpt:
> >
> > if (linkComponent instanceof AjaxLink)
> > {
> > // If it's not ajax we fail
> > [0] if (isAjax == false)
> > {
> > fail("Link " + path + "is an AjaxLink and will " +
> > "not be invoked when AJAX (javascript) is
> disabled.");
> > }
> >
> > AjaxLink link = (AjaxLink)linkComponent;
> >
> > setupRequestAndResponse();
> >
> > [1] // getServletRequest().addHeader("Wicket-Ajax", "Yes");
> >
> > [2] RequestCycle requestCycle = createRequestCycle();
> > if ( requestCycle.getRequest() instanceof WebRequest &&
> > !((WebRequest)requestCycle.getRequest()).isAjax())
> > {
> > fail("this is no ajax request");
> > }
> >
> >
> > [0] only uses the parameter for the check, not the actual request
> >
> > if [1] is uncommented, several other tests fail, due to assuming that
> > there is an ajax request, when actually there is none
> >
> > [2] checks whether the current WebRequest is an ajax request. with [1]
> > being commented, the test fails with "this is no ajax request".
> >
> > note: [1] and [2] were added by me and cannot be found in the code in
> > trunk
> >
> >
> >
> > either i'm missing something, or a lot of tests involving ajax are
> > seriously broken.
> > in any case, this should be clarified before releasing RC2.
> >
> > any comments?
> >
> > Gerolf
> >
>
