You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Marshall Schor (JIRA)" <ui...@incubator.apache.org> on 2007/06/14 22:44:27 UTC

[jira] Resolved: (UIMA-303) Problems with BoundedQueue.dequeue(timeout)

     [ https://issues.apache.org/jira/browse/UIMA-303?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Marshall Schor resolved UIMA-303.
---------------------------------

    Resolution: Fixed
      Assignee: Adam Lally  (was: Marshall Schor)

The timeouts now handle multiple notifies (as in notify-all), and re-do waits with the "remaining time left".  Not sure the tests really test this - so would be good to have another pair of eyes do a code inspection here.  

> Problems with BoundedQueue.dequeue(timeout)
> -------------------------------------------
>
>                 Key: UIMA-303
>                 URL: https://issues.apache.org/jira/browse/UIMA-303
>             Project: UIMA
>          Issue Type: Bug
>          Components: Collection Processing
>            Reporter: Adam Lally
>            Assignee: Adam Lally
>            Priority: Minor
>             Fix For: 2.2
>
>
> In the CPM, the method BoundedQueue.dequeue(timeout) is supposed to wait for the specified timeout for a CAS to become available.  However from a code inspection it looks like there are problems with its synchronization.  If multiple threads are waiting, and one CAS becomes available, then all of the threads will wake up, but only one will get the CAS.  The others will return null prematurely, without having exhausted their timeout.
> Also there may be a problem with chunking enabled and a chunk timeout is specified, where the timeout does not actually do anything except log a message.
> Details here: http://www.mail-archive.com/uima-dev@incubator.apache.org/msg02033.html

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