[
https://issues.apache.org/jira/browse/DIRMINA-803?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12922047#action_12922047
]
Emmanuel Lecharny commented on DIRMINA-803:
-------------------------------------------
You are right !
When I removed the flushWithoutFuture() method and merged its code, I simply
ignored the test for nulity.
My bad ! Will fix it in trunk, and probably release a MINA 2.0.1 asap
Thanks for the report !
> ProtocolCodecFilter.filterWrite() is no longer thread-safe
> ----------------------------------------------------------
>
> Key: DIRMINA-803
> URL: https://issues.apache.org/jira/browse/DIRMINA-803
> Project: MINA
> Issue Type: Bug
> Components: Filter
> Affects Versions: 2.0.0
> Reporter: Stuart Scott
>
> In revision 'r912149 | elecharny | 2010-02-20 22:04:32 +0500 (Sat, 20 Feb
> 2010)' the loop that flushes the encoded messages was modified and is no
> longer thread-safe. In the trunk the queue is checked if it is empty then
> polled for the next element. Between these calls another thread could have
> taken the last element in the queue and the element is not checked if it is
> null.
> Starting at line 328:
> while (!bufferQueue.isEmpty()) {
> Object encodedMessage = bufferQueue.poll();
> ...
> could be replaced with the previous loop:
> for (;;) {
> Object encodedMessage = bufferQueue.poll();
> if (encodedMessage == null) {
> break;
> }
> ...
> to make the method thread-safe again.
> The same issue exists in ProtocolEncoderOutputImpl.flush().
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
