[
https://issues.apache.org/jira/browse/WICKET-5464?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13865264#comment-13865264
]
Marcus Thiesen commented on WICKET-5464:
----------------------------------------
>From my point of view as a user: I didn't expect this behavior and needed some
>time to figure out what actually happens. I can't actually think up a case
>where you expect the behavior to be disabled due to adding/removal (especially
>when it is implicit due to other components like a Wizard). I use my own
>implementation of the behavior now in order to avoid the bug, so I wouldn't be
>affected by this change in mainstream Wicket. When you actually need the
>behavior to be disabled at a certain point one should (and guess would) call
>#stop() explicitly. So I'm +1 here.
If you decide not to fix this in 6.x., you should at least update the
documentation to explain when this Behavior will be stopped and how to
re-enable it.
> AbstractAjaxTimerBehavior does not work in combination with Wizards
> -------------------------------------------------------------------
>
> Key: WICKET-5464
> URL: https://issues.apache.org/jira/browse/WICKET-5464
> Project: Wicket
> Issue Type: Bug
> Components: wicket
> Affects Versions: 6.12.0
> Reporter: Marcus Thiesen
> Attachments: ajaxwizard-quickstart.tar.bz2
>
>
> AbstractAjaxTimerBehavior calls stop() in it's onRemove() method. When used
> in a Wizard onRemove is triggered by onActiveStepChanged. This disabled, but
> never re-enables the behavior if the component tree is reused during wizard
> stepping.
> In my opinion it is more correct to disable the frontend timer but keep the
> internal state of the Bahavior as is (or resetting headRendered to false as
> the Behavior is in the pre render state after the onRemove) in order to
> re-enable the Timer when the step using the Behavior is rendered again.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
