You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@manifoldcf.apache.org by "Karl Wright (JIRA)" <ji...@apache.org> on 2014/12/20 12:02:13 UTC

[jira] [Commented] (CONNECTORS-1126) Lock manager's lock ordering not working as designed

    [ https://issues.apache.org/jira/browse/CONNECTORS-1126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14254645#comment-14254645 ] 

Karl Wright commented on CONNECTORS-1126:
-----------------------------------------

r1646949(release-2.0-branch)

> Lock manager's lock ordering not working as designed
> ----------------------------------------------------
>
>                 Key: CONNECTORS-1126
>                 URL: https://issues.apache.org/jira/browse/CONNECTORS-1126
>             Project: ManifoldCF
>          Issue Type: Bug
>          Components: Framework core
>    Affects Versions: ManifoldCF 1.7.2
>            Reporter: Karl Wright
>            Assignee: Karl Wright
>             Fix For: ManifoldCF 1.9, ManifoldCF 2.1
>
>         Attachments: CONNECTORS-1126.patch
>
>
> The LockGate logic for flushing a LockGate is as follows:
> {code}
>     synchronized (lockObject)
>     {
>       if (lockObject.leaveReadLock())
>       {
>         if (lockPool != null)
>           lockPool.releaseObject(lockKey, this);
>       }
>     }
> {code}
> Unfortunately, this has the side effect of causing the nicely ordered thread queue to be destroyed on the first time a lock is exited.  That kills ordering and makes LockGate essentially do nothing of value.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)