[JBoss JIRA] Closed: (NETTY-365) HashedWheelTimer.stop() enters an infinite loop when called from TimerTask

[Trustin Lee (JIRA)]

     [ https://issues.jboss.org/browse/NETTY-365?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Trustin Lee closed NETTY-365.
-----------------------------

    Resolution: Done


Fixed - IllegalStateException is raised to avoid the infinite loop now.

> HashedWheelTimer.stop() enters an infinite loop when called from TimerTask
> --------------------------------------------------------------------------
>
>                 Key: NETTY-365
>                 URL: https://issues.jboss.org/browse/NETTY-365
>             Project: Netty
>          Issue Type: Bug
>          Components: Transport
>    Affects Versions: 3.2.3.Final
>            Reporter: Bruce Lowekamp
>            Assignee: Trustin Lee
>            Priority: Minor
>             Fix For: 3.2.4.Final
>
>
> If a HashedWheelTimer event fires and ultimately the triggered code calls HashedWheelTimer.stop(), it enters an infinite loop.
> The following code:
> boolean interrupted = false;
> 279         while (workerThread.isAlive()) {
> 280             workerThread.interrupt();
> 281             try {
> 282                 workerThread.join(100);
> 283             } catch (InterruptedException e) {
> 284                 interrupted = true;
> 285             }
> 286         }
> if called by the workerThread, this code just interrrupts itself in an infinite loop.
> I have mixed feelings about this calling pattern, but would like for it to either work (I don't see a reason the worker thread can't run stop here), or if it can't work, to throw an exception.  It's a pretty easy pattern to get into with a timeout causing a test to shutdown.  Either way, just need to check if the caller is the workerThread, I think.

-- 
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira