that would be another advantage of the simple default behavior to
render whatever (ul, table, span, ...) the component usually
renders also if there are no messages wouldn't it?
On Tue, Mar 25, 2008 at 11:51 PM, Andrew Robinson
<[EMAIL PROTECTED]> wrote:
> I would not recommend forceSpan as an attribute as the use of a span
> is renderer specific, not component specific. The component should not
> care how the renderer works
>
>
>
> On Tue, Mar 25, 2008 at 4:41 PM, Ernst Fastl <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
> > I like the idea of making that the default behavior. I think for
> > t:messages it should be save to do
> > so. As for h:messages I'm not sure how detailed the spec is there.
> > I'll better check that before
> > touching h:messages. I also don't know what mojarra does, but I would
> > guess that they do
> > the same thing as we do at the moment.
> >
> > forceSpan would also work for me I guess. But if there it is ok to
> > ajust the default behavior
> > of t:messages I'd prefer that solution because its one parameter less
> > the user needs to
> > know and think about.
> >
> > thanks for your ideas guys
> >
> > Ernst
> >
> >
> >
> > On Tue, Mar 25, 2008 at 10:27 PM, Cagatay Civici
> > <[EMAIL PROTECTED]> wrote:
> > > Hi,
> > >
> > > For t:message and t:messages, there's already a forceSpan attribute,
> will it
> > > work for you Ernst?
> > >
> > > Cagatay
> > >
> > >
> > >
> > > On Tue, Mar 25, 2008 at 11:17 PM, simon <[EMAIL PROTECTED]> wrote:
> > >
> > > >
> > > >
> > > >
> > > >
> > > > On Tue, 2008-03-25 at 21:45 +0100, simon wrote:
> > > > > On Tue, 2008-03-25 at 20:50 +0100, Ernst Fastl wrote:
> > > > > > Hi,
> > > > > >
> > > > > > I would like to add renderedIfEmpty to the t:messages component
> which
> > > > > > per default is false (current behaviour)
> > > > > > The reason for that: To update/append FacesMessages to a
> > > > > > messages-component after AJAX requests (e.g. PPR)
> > > > > > it has to be possible to locate a DOM element to which this
> messages
> > > > > > can be appended.
> > > > > >
> > > > > > One could also think of automatically adding a CSS-style
> display:none
> > > > > > if the empty table/list is rendered which is
> > > > > > reset by any JavaScript updates to the messages.
> > > > > >
> > > > > > Alternatively we could create a new messages-component which
> supports
> > > that.
> > > > > >
> > > > > > If anybody can think of reasons why this parameter should not be
> added
> > > > > > please tell. Otherwise I'll file a JIRA issue
> > > > > > and start implementing on thursday evening.
> > > > >
> > > > > Does your second suggestion ("CSS-style display:none") mean to
> always
> > > > > render the div, but when there are no messages do:
> > > > > <div id=".." style="display:none"/>
> > > > > and let javascript modify the style if messages need to be added to
> it
> > > > > after a ppr request?
> > > > >
> > > > > If so, I think that is nicer. Logically, it makes sense too; the
> > > > > h:messages has not been marked with rendered=false, it just happens
> to
> > > > > have no messages. So the div should be present, but hidden.
> > > > >
> > > > > And it means that no extra attribute is needed.
> > > > >
> > > > > I cannot imagine any application that would break because of the
> > > > > introduction of a div with display:none set.
> > > >
> > > > BTW, I was thinking about making this change in h:messages, not just
> in
> > > > t:messages.
> > > >
> > > > However even if the change is done in core, the same functionality
> would
> > > > still be needed in t:messages unless Sun Mojarra's h:messages
> component
> > > > also renders a hidden div when empty. Any idea what Mojarra does here?
> > > >
> > > > Regards,
> > > > Simon
> > > >
> > > >
> > >
> > >
> >
>
