[ https://issues.apache.org/jira/browse/JCR-1108?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12524550 ] 

Julian Reschke commented on JCR-1108:
-------------------------------------

Turns out that the TCK indeed uses the same session (otherwise it would fail :-).

Which leaves us with the question why LockManagerImpl doesn't realize that the nodes are already gone...



> JCR2SPI: error level logging when cleaning up session locks 
> ------------------------------------------------------------
>
>                 Key: JCR-1108
>                 URL: https://issues.apache.org/jira/browse/JCR-1108
>             Project: Jackrabbit
>          Issue Type: Improvement
>          Components: SPI
>            Reporter: Julian Reschke
>            Priority: Minor
>
> LockManagerImpl.loggingOut() tries to unlock nodes that have a session lock. If, while doing so, a RepositoryException is thrown, this gets locked on error level.
> The TCK tests tearDown code removes test nodes using a separate session; thus we see RepositoryExceptions for the simple reason that the nodes have already been removed by somebody else.
> Proposal: handle ItemNotFoundExc and PathNotFoundExc separately, not logging them.

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