[jira] Commented: (WICKET-876) Assigning Label to anchor tag in recursive ListView causes exponential performance decay

["Nick Johnson (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/WICKET-876?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12522294 ] 

Nick Johnson commented on WICKET-876:
-------------------------------------

As an aside, I was incorrect in identifying the rate of decay as exponential; it's actually polynomial, following the formula (n + n^2) / 2.  (O(n^2)) Bad enough to be very noticeable, but very easy to correct.

Thanks for the explanation!

> Assigning Label to anchor tag in recursive ListView causes exponential performance decay
> ----------------------------------------------------------------------------------------
>
>                 Key: WICKET-876
>                 URL: https://issues.apache.org/jira/browse/WICKET-876
>             Project: Wicket
>          Issue Type: Bug
>          Components: wicket
>    Affects Versions: 1.2.6
>            Reporter: Nick Johnson
>            Assignee: Igor Vaynberg
>            Priority: Trivial
>
> This isn't so much a bug as a curiosity, and I'm partially just documenting this in case someone else runs across the same problem later and searches the bug database for clues.  It -would- be interesting to figure out why it happens.  The workaround is trivial.
> If the markup for a recursive list view (like the recursive panel in the examples) contains a named anchor like this:
> <a name="x" wicket:id="anchor"> </a>
> And the anchor is (inappropriately) associated with a Label like this:
> add(new Label("anchor").add(new SimpleAttributeModifier("name","foo")));
> Then as the levels of recursion deepen, performance during rendering decays exponentially.  During this time, the Java process will also start consuming a tremendous amount of CPU.
> The solution is fairly simple: don't use a Label for situations like this; instead, use a WebMarkupContainer:
> add(new WebMarkupContainer("anchor").add(new SimpleAttributeModifier("name","foo")));
> With the component being WebMarkupContainer, performance does not degrade exponentially.  With Label, it does.  I'm afraid I don't know enough of the internal workings of Wicket to speculate as to why.

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