On Sun, 2002-12-08 at 21:07, Sunil Pandit wrote:
> James
> Its not CacheEventQueue but its in ElementEventQueue. Please check the
> QProcessor::Run method in class org.apache.jcs.engine.control.event
> .ElementEventQueue class. The current version 1.3 in the repository still
> got the code in the run method which calls Thread.Destroy on receiving the
> Interrupted exception.
>
> public void run()
> {
> AbstractElementEventRunner r = null;
>
> while ( !destroyed )
> {
> try
> {
> r = take();
>
> if ( log.isDebugEnabled() )
> {
> log.debug( "r from take() = " + r );
> }
>
> }
> catch ( InterruptedException e )
> {
> // We were interrupted, so terminate gracefully.
>
> this.destroy();
Perhaps this should be "CacheEventQueue.this.destroy()"?
OR is "destroyed = true" better?
> }
>
> if ( !destroyed && r != null )
> {
> r.run();
> }
> }
>
> log.info( "QProcessor exiting for " + ElementEventQueue.this );
> }
> }
>
>
>
>
> James Taylor
>
> <james To: Turbine JCS Users List
><[EMAIL PROTECTED]>
> @jamestaylor. cc:
>
> org> Subject: Re: Is this defect ???
>
>
>
> 12/05/2002
>
> 02:53 PM
>
> Please
>
> respond to
>
> "Turbine JCS
>
> Users List"
>
>
>
>
>
>
>
>
>
> That does not appear to be the case in the current CVS (1.4)
>
> see:
>
>
>http://cvs.apache.org/viewcvs.cgi/jakarta-turbine-jcs/src/java/org/apache/jcs/engine/CacheEventQueue.java.diff?r1=1.3&r2=1.4&diff_format=h
>
>
> On Thu, 2002-12-05 at 15:48, Sunil Pandit wrote:
> > James
> >
> > QProcessor::Run method calls this.destroy in the exception handler code
> > where it handles InterruptedException .
> >
> > Thanks
> > Sunil
> >
> >
> >
>
> > James Taylor
>
> > <james To: Turbine JCS Users List
> <[EMAIL PROTECTED]>
> > @jamestaylor. cc:
>
> > org> Subject: Re: Is this defect
> ???
> >
>
> > 12/05/2002
>
> > 01:54 PM
>
> > Please
>
> > respond to
>
> > "Turbine JCS
>
> > Users List"
>
> >
>
> >
>
> >
> >
> >
> >
> > On Thu, 2002-12-05 at 14:43, Sunil Pandit wrote:
> > > Hello JCS Users
> > >
> > > I came across an interesting defect in JCS .
> > > During an Initialization of JCS instance when it creates a
> CompositeCache
> > > instance, It creates an instance of ElementEventQueue.
> ElementEventQueue
> > > further creates a thread in form of an instance of QProcessor.
> > > This thread is never destroyed. If we test our application under heavy
> > > load we found that several of these threads remain active. This thread
> is
> > > created per instance of CompositeCache and therefore should be
> destroyed
> > > when the composite cache is disposed.
> >
> > Yes, ElementEventQueue.Destroy should definitely be called in
> > CompositeCache.Dispose
> >
> > > Also there is one more defect in JCS in same context. To correct the
> > above
> > > defect I thought I can use ElementEventQueue::Destroy method which
> calls
> > > interrupt on QProcessor. The run method for this thread calls
> > > Thread::Destroy method which is unimplemented in JDK 1.4 . So we just
> > > modified the run method to set the destroy flag to true on Interrupt
> > event
> > > so that the thread can gracefully exit.
> >
> > That sounds reasonable also, where exactly is Thread.destory called?
> > Just want to be sure I am following you.
> >
> >
> >
> > --
> > To unsubscribe, e-mail: <
> > mailto:[EMAIL PROTECTED]>
> > For additional commands, e-mail: <
> > mailto:[EMAIL PROTECTED]>
> >
> >
> >
> >
> >
> >
> > --
> > To unsubscribe, e-mail: <
> mailto:[EMAIL PROTECTED]>
> > For additional commands, e-mail: <
> mailto:[EMAIL PROTECTED]>
> >
>
>
> --
> To unsubscribe, e-mail: <
> mailto:[EMAIL PROTECTED]>
> For additional commands, e-mail: <
> mailto:[EMAIL PROTECTED]>
>
>
>
>
>
>
> --
> To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
>
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
