[JBoss-dev] Re: JMS tasks from jboss-4.0.5

[Adrian Brock]

On Fri, 2006-09-15 at 06:09 -0500, Dimitris Andreadis wrote:
> We have some 11 JMS related tasks for jboss-4.0.5 that need to be dealt
> with or deferred. (I believe only Weston is doing jbas-3183?)
> 
> Since Adrian is buzy with JB5, can someone else (jboss-messaging maybe)
> help?
> 

I'm going to spend a couple of days on JBossMQ before 4.0.5 is released.
Below is a summary. Most of it should be catered for by Weston's
changes to the ServerSessionPool.

> Thanks
> 
> JBAS-3183  Integration with WebSphere MQ 5.x

Weston is working on this, he knows how to fix the final part,
which is really a JCA issue.

> JBAS-3597  JBAS-3437 Test error
> org.jboss.test.jbossmq.test.JBossMQUnitTest::testTopicSelectorChange

I doubt this will get done.
I don't have time to rewrite these crappy tests for this release.

Ideally, JBoss Messaging and JBossMQ should be sharing a testsuite
using the JBossAdmin abstraction I wrote when I integerated the
JORAM testsuite.

> JBAS-3620  JBossMQ's PersistenceManager RecoveryRetries doesn't work if
> failure is creating the schema

I will do this.

> JBAS-3627  Write test to test rollback of a NACK in 2PC  Adrian Brock
> Luc Texier

This isn't very important, but I will probably do it as part of
the related fix that was only partially fixed in 4.0.5RC1 (JBAS-3628)

> JBAS-3634  Acknowledgement of transacted message in MessageListener
> should be inside delivery lock

I will do this.

> JBAS-3631  Missing branch in StdServerSession commit logic

Weston is working on this.

> JBAS-1346  JDBC StateManager  

This won't get done, it is a feature request that keeps getting bumped.
We should probably remove it from the schedule.

> JBAS-1908  listMessageCounter issue with JMS subscriptions containing
> certain message selector characters

This won't get done. The original contributor of this code
looks to have orphaned it.
This should be removed from the schedule of fixes since the real "fix"
is JBAS-2841

> JBAS-3468  JMS Connection starts ping thread before ClientID is set
> resulting in the ClientMonitorInterceptor to close the connection after
> client timeout 

I'll apply my suggested fix since the user's suggested patch
is wrong.

> JBAS-3483  The javadoc relating to StdServerSessionPool pooling
> parameters is wrong 

Weston is working on this.

> JBAS-1485  JMS ResourceAdapter makes unnecessary temporary delete
> requests  

This won't get done. It is a none issue that just leads to
a minor inefficency at session.close()
The fix is fairly complicated in comparison, it involves
wrapping/unwrapping the destination objects to keep
track of their state.
In fact, it is already fixed by the new resource adapter
stuff I did in the jca prototype project.
-- 
xxxxxxxxxxxxxxxxxxxxxxxxxxx
Adrian Brock
Chief Scientist
JBoss a division of Red Hat
xxxxxxxxxxxxxxxxxxxxxxxxxxx