NullPointerException in Quercus 4.0.1

[ashamash]

Using Quercus 4.0.1, I'm consistently getting the following exception, approximately 30 seconds after undeploying, redeploying, and accessing the application in Glassfish. This does not happen after the initial deploy, only after redeploy.

Once the exception happens, it then happens in a very tight loop, making the app server unusable, forcing a restart (see timestamps from server log file below). This effectively means a restart on every deploy.

The same application did not have any issues with Quercus 3.1.6:

[#|2009-10-15T10:47:59.649-0400|WARNING|sun-appserver2.1|com.caucho.util.Alarm|_ThreadID=17;_T hreadName=alarm-coordinator;_RequestID=7364023f-c2ef-49ff-b191-a77c66ce3dd1;|java.lang.NullPointerException
java.lang.NullPointerException
at com.caucho.util.Alarm.extractAlarm(Alarm.java:432)
at com.caucho.util.Alarm$CoordinatorThread.run(Alarm. java:704)
|#]

[#|2009-10-15T10:47:59.654-0400|WARNING|sun-appserver2.1|com.caucho.util.Alarm|_ThreadID=17;_T hreadName=alarm-coordinator;_RequestID=7364023f-c2ef-49ff-b191-a77c66ce3dd1;|java.lang.NullPointerException
java.lang.NullPointerException
at com.caucho.util.Alarm.extractAlarm(Alarm.java:432)
at com.caucho.util.Alarm$CoordinatorThread.run(Alarm. java:704)
|#]

[#|2009-10-15T10:47:59.657-0400|WARNING|sun-appserver2.1|com.caucho.util.Alarm|_ThreadID=17;_T hreadName=alarm-coordinator;_RequestID=7364023f-c2ef-49ff-b191-a77c66ce3dd1;|java.lang.NullPointerException
java.lang.NullPointerException
at com.caucho.util.Alarm.extractAlarm(Alarm.java:432)
at com.caucho.util.Alarm$CoordinatorThread.run(Alarm. java:704)
|#]

[nam]

Thanks for the report. I've filed a bug at:

http://bugs.caucho.com/view.php?id=3711

[ashamash]

There is a discussion on the glassfish forums over this issue as well. I made some code modifications that work around this issue:

http://forums.java.net/jive/thread.jspa?threadID=67401

Regards,
Ari