[
https://issues.apache.org/jira/browse/WICKET-6879?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17330207#comment-17330207
]
Sven Meier commented on WICKET-6879:
------------------------------------
Patrick and I were able to narrow down the problem:
Since Wicket 7.x a Border queues its body container instead of just adding it.
It seems there are circumstances where the container is not dequeued before the
header is rendered.
> renderHead() not called for components inside Border
> ----------------------------------------------------
>
> Key: WICKET-6879
> URL: https://issues.apache.org/jira/browse/WICKET-6879
> Project: Wicket
> Issue Type: Bug
> Components: wicket, wicket-core
> Affects Versions: 8.6.1, 8.10.0
> Reporter: Patrick Davids
> Priority: Major
> Attachments: Wicket_Missing_Scripts_0.PNG
>
>
> Hi Wicket Team,
> since we did an update from Wicket 6.29 to (8.6.0 and then) 8.10 we encounter
> a very special and unfortunately not reproducable issue.
> Currently I wasn't able to create a quickstart, which shows the problem.
> But I can see the result in the rendered HTML (see screenshot attached).
> It seems the render process (or any other) looses Behaviors or it just don't
> call the partical renderHead() method.
> As you can see at the screenshot there is a JQueryUI button, which gets its
> CSS style via script calls (JQueryUI <-> Wicket JQueryUI Project from
> Sebastien Brique).
> Those script calls which are rendered via Behaviors are not present at the
> HTML. Have a look at my search result in Web Developer Tool.
> The ID I focused is only present at the HTML Tag Element and no Javascript
> is rendered (e.g. jQuery('#id391b').button(\{...})
> What we know so far:
> * The problem only occurs at two use-cases of our application and always
> inside of Border Components
> ** So possibly this issue is somehow related to WICKET-6303 or WICKET-6222
> ** The problem is not reproducable / 95%+ of the render cases are working
> correct.
> * A reload of the page solves the problem.
> * One of the usecases / components had quite a lot problems with
> "org.apache.wicket.WicketRuntimeException: Cannot modify component hierarchy
> after render phase has started" before. We set a model-object of a child
> component in onConfigure() which is usually not a problem.
> ** To solve it, we changed the implementation, but know this issue arised.
> ** If it just an aftereffect of the changed implementation, setting the
> model-object, we don't know, yet.
> ** Maybe it is related, maybe not.
> * It seems, only Components/Behaviors are affected which are from sub-class
> type JQueryUIBehavior
> ** BUT: This is only an assumption, because here we a visual affect to our
> customers and they reporting it.
> ** If other Behaviors (e.g. Wicket Javascripts for Click-Handler bindings)
> are not rendered; we also dont't know, yet.
> I decided to open this issue here and not at GitHub for Sebastien, because of
> WICKET-6303 and WICKET-6222, and I know [~sebfz1] is also a Wicket committer.
> This issue is quite an urgent problem to us!
> [~svenmeier] or [~sebfz1] : If you could get in contact with me by phone to
> have some brainstorming, about how we can help getting further information
> for you, this would be great.
> We don't have any idea to get loggings or other information, because this
> problem just occurs silently and we only get feedback by our customers.
> And than often it is too late to see anything. And we don't see any error
> loggings matching this problem.
> kind regards
> Patrick Davids
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
