tapestry-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Howard M. Lewis Ship (JIRA)" <...@tapestry.apache.org>
Subject [jira] Commented: (TAPESTRY-1805) Reverse ordering/exceution of page render phase methods (afterXXX and cleanupRender) incorrect when short-circuiting
Date Wed, 02 Jan 2008 19:47:34 GMT

    [ https://issues.apache.org/jira/browse/TAPESTRY-1805?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12555379#action_12555379
] 

Howard M. Lewis Ship commented on TAPESTRY-1805:
------------------------------------------------

I'm glad you are making use of the mixins.  I think there's a difference in the mental model
between your view of mixins and how they are actually designed and implemented.

Your view resembles a nested procedure call.  The actual design and implementation is event
based, with fixed rules for how event notifications are propogated to the component and its
mixins.

I think neither solution can solve all problems.  

I'm not yet persuaded to change anything.  Perhaps with some more specific examples? I think
you may need to use instance variables to store flags in your Mixin to catch after events
when you've short-circuitted the corresponding begin/before event.

> Reverse ordering/exceution of page render phase methods (afterXXX and cleanupRender)
incorrect when short-circuiting
> --------------------------------------------------------------------------------------------------------------------
>
>                 Key: TAPESTRY-1805
>                 URL: https://issues.apache.org/jira/browse/TAPESTRY-1805
>             Project: Tapestry
>          Issue Type: Bug
>          Components: tapestry-core
>    Affects Versions: 5.0.6
>         Environment: Tapestry 5.0.6-SNAPSHOT
>            Reporter: Kristian Marinkovic
>            Priority: Critical
>
> ComponentPageElementImpl is responsible for calling the render 
> phase methods of the components. It calls the setupRender method 
> on the attached Mixins then the components and then the Mixins that
> have a @MixinAfter annotation. If in this phase the setupRender of a
> Mixin returns a boolean value that will be saved into the Event object and
> in case of a "false" the setupRender calls of the subsequent components 
> will be aborted. 
> During the "inverse" lifecycle the order is simply reversed and the render
> phase method of the Mixin with @MixinAfter annotation is called, then
> the render phase method of the component and then the method of the
> Mixin is called. The problem is now that the "inverse" render phase 
> methods are getting called without checking whether the corresponding 
> other render phase methods have been executed. 
> So in my case the cleanupRender method of the Form component gets
> called although i have short-circuited the beginRender method with my 
> Mixin because there is no state available that indicates that the beginRender
> method of the Form component has NOT been executed.
> private void invoke(boolean reverse, ComponentCallback callback)
> {
>            ....
>             Iterator<Component> i = reverse ? 
> InternalUtils.reverseIterator(_components)
>                     : _components.iterator();
>             while (i.hasNext())
>                 callback.run(i.next());
> A possible solution would be to create a List that contains those 
> components whose render phase methods really have been 
> executed. This list is then used in the "inverse" lifecycle.
> see also:
> http://www.nabble.com/T5%3A-Mixin-on-Form-does-not-work-as-expected-tf4562001.html
> http://www.nabble.com/T5%3A-Mixin-on-Form-does-not-work-as-expected-tf4562724.html

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
For additional commands, e-mail: dev-help@tapestry.apache.org


Mime
View raw message