[
https://issues.apache.org/jira/browse/LABS-156?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Simone Gianni resolved LABS-156.
--------------------------------
Resolution: Fixed
> WEB: All handlers assigned to any bean should be cleared after a web request
> has completed
> ------------------------------------------------------------------------------------------
>
> Key: LABS-156
> URL: https://issues.apache.org/jira/browse/LABS-156
> Project: Labs
> Issue Type: Bug
> Components: Magma
> Reporter: Simone Gianni
> Assignee: Simone Gianni
>
> Currently the handler is thread local and softreferenced. Anyway, this could
> happen :
> * A get arrives, trying to modify a bean
> * The modification fails, but the exception is catched, and the handler
> still contains bad values
> * Or even, the handler is not committed, and just stays there
> * The reference to the handler is lost, the thread returns to the pool
> * Another get arrives
> * It reuses the same thread
> * For some reason (JPA cache, bean in cache, whatever) it obtains the
> same bean as the previous get
> * The garbage collector hasn't cleared the reference
> * The second get obtains the same handler, potentially committing on the
> bean data from the previous get.
> A solution would be to give access to all ThreadLocals? in a single place
> (another ThreadLocal? maybe :) ) so that at specific events they can all get
> cleared and the thread be fresh again. The first of these events would
> obviously be the arrival of a new get.
> It could be done with a single aspect intercepting new ThreadLocal? or calls
> to ThreadLocal?.set.
--
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: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
