I've just reviewed the code - there was one possible code path (recovery) which could access the collection which was causing the concurrent modification exception. All other accesses are within a synchronized block so there is no possibility of the concurrent access exception.
I wonder could you retry with SVN HEAD to see if you can still produce this? James On 3/24/06, vik Dhawan <[EMAIL PROTECTED]> wrote: > > I looked at the code for method removeMessage() in the > org.apache.activemq.broker.region.Queue Class. This code is removing a > message from the fail-fast iterator obtained in the code. > > It's definitely a hot spot for "java.util.ConcurrentModificationException" > what i am seeing. > > I am not sure why developer of this code have chosen this implementation > when its clearly written in Sun Collection Java docs for iterators. > > I will really appreciate if some one can explain that. > > Thanks! > > -- > View this message in context: > http://www.nabble.com/JMX-exception-t1337321.html#a3578256 > Sent from the ActiveMQ - Dev forum at Nabble.com. > > -- James ------- http://radio.weblogs.com/0112098/
