Just to add my support to the inclusion of this fix. I've also reviewed the patch and agree with Robbie's characterization of the risks and benefits.
-- Rob On 4 April 2012 00:25, Robbie Gemmell <[email protected]> wrote: > Hi Justin, > > I would like to request for the following fix to be included in 0.16. > Alex reviewed the change before I commited it (updating the JIRA was > just waiting for additional testing, now added) and I also discussed > the issue with Rob, they should hopefully be happy to pipe up to that > effect in the morning :) > > QPID-3927: Java Broker Priority Queue dispatch not working correctly > https://issues.apache.org/jira/browse/QPID-3927 > > The result of the defect is that messages may get 'stuck' on a > priority queue until such time as a new consumer subsequently > subscribes to the queue or sometimes higher priority messages arrive, > after either of which points an entirely new set of messages can then > potentially get stuck. Repeat. > > The actual defect I fixed has been present for a while, but we appear > to have exposed it more with a specific change made since 0.14. The > update is pretty trivial in the end and is isolated in the priority > queue implementation, meaning it is fairly low risk and actually only > affects something that is otherwise going to be broken anyway. > > Robbie > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
