|
|
When I hit escape to close the VirtualBUG when my app is running I get a null pointer exception. I'm guessing it's trying to tell my service tracker that the menu service has gone away, but my service tracker is already gone at the this point or something. It doesn't do this if I start either the GPS or LCD on their own then kill the bug; just when I start both (to satisfy the requirements of my app). I thought at first it was because I wasn't calling close() on the OSGI ServiceTracker in my activator's stop method, but adding that call didn't change anything. Will the next SDK include the source by default? That would make setting breakpoints and walking through this error case a little more straightforward? If it will, I'm inclined to hold off on troubleshooting this until then. [Sat Oct 13 11:43:12 EDT 2007] [INFO] Received new REMOVE model change event for module GPS STOPPING file:/home/dave/dev/eclipse/configuration/org.eclipse.osgi/bundles/171/1/.cp/jars/osgi.jar STOPPING file:/home/dave/dev/eclipse/configuration/org.eclipse.osgi/bundles/171/1/.cp/jars/cg_event.jar STOPPING file:/home/dave/dev/eclipse/configuration/org.eclipse.osgi/bundles/171/1/.cp/jars/cg_framework.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/ch.ethz.iks.slp.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.osgi.http.jar [Sat Oct 13 11:43:12 EDT 2007] [INFO] Unregistered servlet at /service. [Sat Oct 13 11:43:12 EDT 2007] [INFO] Unregistered servlet /program [Sat Oct 13 11:43:13 EDT 2007] [INFO] Server received shutdown-message. Shutting down. [Sat Oct 13 11:43:12 EDT 2007] [INFO] Unregistered servlet at /program. [Sat Oct 13 11:43:13 EDT 2007] [INFO] Unregistered servlet /module [Sat Oct 13 11:43:13 EDT 2007] [INFO] Unregistered servlet at /module. [Sat Oct 13 11:43:13 EDT 2007] [INFO] Unregistered servlet /event [Sat Oct 13 11:43:13 EDT 2007] [INFO] Unregistered servlet at /event. STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.service.jar [Sat Oct 13 11:43:13 EDT 2007] [INFO] Unregistered servlet at /service. STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.program.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.base.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.common.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.menu.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.module.jar STOPPING file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.event.jar STOPPING file:/home/dave/dev/workspace2/.metadata/.plugins/com.buglabs.osgi.concierge.runtime/bundles/GpsLogger_1.0.2.jar GpsLogger: activator stop GpsLogger: activator stopped java.lang.NullPointerException at com.buglabs.util.trackers.PublicWSAdminTracker.removedService(PublicWSAdminTracker.java:59) at org.osgi.util.tracker.ServiceTracker$Tracked.untrack(ServiceTracker.java:897) at org.osgi.util.tracker.ServiceTracker.close(ServiceTracker.java:269) at com.buglabs.bug.menu.Menu.stop(Menu.java:143) at com.buglabs.bug.menu.Activator.stop(Activator.java:37) at ch.ethz.iks.concierge.framework.BundleImpl.stopBundle(BundleImpl.java:502) at ch.ethz.iks.concierge.framework.Framework$SystemBundle.setLevel(Framework.java:2293) at ch.ethz.iks.concierge.framework.Framework$SystemBundle.access$0(Framework.java:2255) at ch.ethz.iks.concierge.framework.Framework.shutdown(Framework.java:514) at ch.ethz.iks.concierge.framework.Framework$1.run(Framework.java:2077) org.osgi.framework.BundleException: Error stopping bundle Bundle [10]: file:/home/dave/dev/eclipse/plugins/com.buglabs.dragonfly.bug.kernel_1.0.0.226/kernel/com.buglabs.bug.menu.jar at ch.ethz.iks.concierge.framework.BundleImpl.stopBundle(BundleImpl.java:509) at ch.ethz.iks.concierge.framework.Framework$SystemBundle.setLevel(Framework.java:2293) at ch.ethz.iks.concierge.framework.Framework$SystemBundle.access$0(Framework.java:2255) at ch.ethz.iks.concierge.framework.Framework.shutdown(Framework.java:514) at ch.ethz.iks.concierge.framework.Framework$1.run(Framework.java:2077) |
|
|
This happens for me in another from-scratch app that isn’t using a TimerTask, so it at least doesn’t appear to be code that I’m adding that’s causing it. |
|
|
finsprings, There are various NPEs that pop up on shutdown and it’s probably nothing to do with your application. Yes, hold off on debugging until a new SDK build is released. We’re having a hard time with another issue related to OS X but expect something shortly. |
|
|
I started seeing this NPE on my Linux box last month. A defect was opened but Ken+Team couldn’t reproduce the error. I will move this out to the public Bugzilla shortly. Fin, I assume you are seeing this on your Linux machine? Also I can reproduce the issue with the following steps with no Apps in my workspace: STR: |
|
|
I see NPEs randomly when shutting down the bug, but it’s not consistent by any means and doesn’t seem to follow any pattern based on installed modules or loaded applications. I didn’t pay it much attention because real hardware will never need to be shut down in the same manner as the virtual BUG, unless you’re powering it off, at which point an NPE during shutdown really doesn’t make a difference (unless it causes the shutdown process to hang). |
|
|
Yes, this is on a knoppmyth box (which is basically Debian Etch). |