[
https://issues.apache.org/jira/browse/WICKET-3209?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12970512#action_12970512
]
Martin Grigorov commented on WICKET-3209:
-----------------------------------------
These methods are the same thing - addBufferedResponse() is in 1.4 and
storeBufferedResponse() is the same in 1.5. For some reason Matej (I guess)
didn't use MostRecentlyUsedMap in the new implementation of this logic. The
question is should we merge addBufferedResponse() and storeBufferedResponse()
(i.e. add usage of MostRecentlyUsedMap in storeBufferedResponse()) or not.
> WebApplication MostRecentlyUsedMap based upon Key age not Value age
> -------------------------------------------------------------------
>
> Key: WICKET-3209
> URL: https://issues.apache.org/jira/browse/WICKET-3209
> Project: Wicket
> Issue Type: Bug
> Components: wicket-auth-roles
> Affects Versions: 1.5-M3
> Environment: all
> Reporter: Richard Emberson
> Priority: Minor
>
> The WebApplication uses the MostRecentlyUsedMap as a limited size cache.
> Elements of
> the cache are removed when a new entry is added based upon the age of the
> element's keys
> not on the age of the element's data.
> Naively, I would expect a high usage element (one's whose associated data
> keeps changing)
> should be kept in the cache rather than reaped - simply because it was enter
> earlier than some
> low activity but more recent entry.
> The MostRecentlyUsedMap is based upon the Java LinkedHashMap container.
> As such, it is not a very good fit for data-based age cache.
> Of course I could be completely wrong in this case and it actually the age of
> the key that is
> important not how often the data associated with the key is changed.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
