[
https://issues.apache.org/jira/browse/DIRMINA-1047?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15554597#comment-15554597
]
Emmanuel Lecharny commented on DIRMINA-1047:
--------------------------------------------
yeah, seems like we have a race condition here : the writeQueue is not empty
when we test it, but it's emptied when we poll from it.
Something like that should fix the NPE :
{noformat}
protected void destroy() {
if (writeRequestQueue != null) {
while (!writeRequestQueue.isEmpty(this)) {
WriteRequest writeRequest = writeRequestQueue.poll(this);
if (writeRequest != null) {
WriteFuture writeFuture = writeRequest.getFuture();
// The WriteRequest may not always have a future : The
CLOSE_REQUEST
// and MESSAGE_SENT_REQUEST don't.
if (writeFuture != null) {
writeFuture.setWritten();
}
}
}
}
}
{noformat}
Can you test it ? (it builds with no error on my machine)
> NullPointerException in AbstractIoSession.destroy()
> ---------------------------------------------------
>
> Key: DIRMINA-1047
> URL: https://issues.apache.org/jira/browse/DIRMINA-1047
> Project: MINA
> Issue Type: Bug
> Components: Core
> Affects Versions: 2.0.15
> Environment: Any
> Reporter: Carl Wicklow
> Fix For: 2.0.16
>
>
> After upgrading to *v2.0.15*, we get this NPE ...
> {noformat}
> java.lang.NullPointerException
> at
> org.apache.mina.core.session.AbstractIoSession.destroy(AbstractIoSession.java:369)
> at
> org.apache.mina.core.session.AbstractIoSession.closeNow(AbstractIoSession.java:350)
> at
> org.apache.mina.core.session.DefaultIoSessionDataStructureFactory$DefaultWriteRequestQueue.poll(DefaultIoSessionDataStructureFactory.java:222)
> at
> org.apache.mina.core.polling.AbstractPollingIoProcessor.flushNow(AbstractPollingIoProcessor.java:827)
> at
> org.apache.mina.core.polling.AbstractPollingIoProcessor.flush(AbstractPollingIoProcessor.java:767)
> at
> org.apache.mina.core.polling.AbstractPollingIoProcessor.access$700(AbstractPollingIoProcessor.java:68)
> at
> org.apache.mina.core.polling.AbstractPollingIoProcessor$Processor.run(AbstractPollingIoProcessor.java:1125)
> at
> org.apache.mina.util.NamePreservingRunnable.run(NamePreservingRunnable.java:64)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> {noformat}
> If I'm reading it right, the destroy also needs to allow that the
> writeRequest returned by the poll() of the WriteRequestQueue may return null
> if the queue is empty.
> In our case the destroy was invoked after polling a CLOSE_REQUEST caused by
> an earlier closeOnFlush(), and the write request queue was empty.
> {code:title=AbstractIoSession.java|borderStyle=solid}
> protected void destroy() {
> if (writeRequestQueue != null) {
> while (!writeRequestQueue.isEmpty(this)) {
> WriteRequest writeRequest = writeRequestQueue.poll(this);
> WriteFuture writeFuture = writeRequest.getFuture();
>
> // The WriteRequest may not always have a future : The
> CLOSE_REQUEST
> // and MESSAGE_SENT_REQUEST don't.
> if (writeFuture != null) {
> writeFuture.setWritten();
> }
> }
> }
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
