[infinispan-issues] [JBoss JIRA] Updated: (ISPN-158) Verify eviction and marshalled values interaction

[Galder Zamarreno (JIRA)]

     [ https://jira.jboss.org/jira/browse/ISPN-158?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Galder Zamarreno updated ISPN-158:
----------------------------------

     Original Estimate: 0 minutes
    Remaining Estimate: 0 minutes


> Verify eviction and marshalled values interaction
> -------------------------------------------------
>
>                 Key: ISPN-158
>                 URL: https://jira.jboss.org/jira/browse/ISPN-158
>             Project: Infinispan
>          Issue Type: Task
>          Components: Eviction
>    Affects Versions: 4.0.0.ALPHA6
>            Reporter: Galder Zamarreno
>            Assignee: Galder Zamarreno
>             Fix For: 4.0.0.CR1
>
>   Original Estimate: 0 minutes
>  Remaining Estimate: 0 minutes
>
> Here's a further thought - we also need to test the eviction thread with marshalled values.  Well, not necessarily the eviction *thread* per-se, but eviction being invoked via the EvictionManager.processEviction() API - which is used by the eviction thread as well as any external management thread that may wish to schedule eviction.
> My reason for concern is that this retrieves keys in order of evictability by querying the data container (which is ordered) directly.  Now if these keys are marshalled values, what happens when the EvictionManager calls cache.evict(key)?  What *should* happen is that the MarshalledValueInerceptor detects that the key is a marshalled value already, and does not bother with further wrapping and unwrapping.
> However I don't think we currently have a unit test for this.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira