[
https://issues.apache.org/jira/browse/TRINIDAD-891?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12563356#action_12563356
]
Gabrielle Crawford commented on TRINIDAD-891:
---------------------------------------------
In general, you can not statically cache Classes as you will blow up the
application if ADF Faces in running in a shared class loader and you have
cached an instance loaded from App1's ClassLoader and App2 asks for the samed
named class.
This kind of caching would require that the cache be maintained in the
Application Map, but the externalContext application map is backed by the
servletContext, which may have its own locking issues. I have filed issue 926
to see if we can create a different app map.
See https://issues.apache.org/jira/browse/TRINIDAD-926
> Locking issues coming from restore state
> -----------------------------------------
>
> Key: TRINIDAD-891
> URL: https://issues.apache.org/jira/browse/TRINIDAD-891
> Project: MyFaces Trinidad
> Issue Type: Bug
> Affects Versions: 1.2.4-core
> Reporter: Stevan Malesevic
> Assignee: Gabrielle Crawford
>
> org.apache.myfaces.trinidad.bean.util.StateUtils$Saver.restoreState tries to
> load a class on every call. This is causing locking issues since SUN JVM is
> doing locking before it tries to load the class.
> It would be much better to use static member of ConcurrentHashMap to map name
> to class. This will remove the locking and speed up the processing of restore
> state
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
