LoaderClient throws NPE when a different computer is loading stuff

[balazsracz]

when there are multiple computers on a network, it is possible to have JMRI running and a firmware update be happening over the network from another computer.

I am getting NPE in the JMRI console when I'm doing this:

2017-01-14 13:19:47,559 jmrix.AbstractMRTrafficController WARN - notify: During reply dispatch to jmri.jmrix.openlcb.OlcbConfigurationManager$2@6fe0860f [AWT-EventQueue-0]
java.lang.NullPointerException
at org.openlcb.LoaderClient.sendStreamNext(LoaderClient.java:255)
at org.openlcb.LoaderClient.handleStreamDataProceed(LoaderClient.java:278)
at org.openlcb.StreamDataProceedMessage.applyTo(StreamDataProceedMessage.java:39)
at org.openlcb.MessageDecoder.put(MessageDecoder.java:34)
at org.openlcb.OlcbInterface$MessageDispatcher.put(OlcbInterface.java:189)
at org.openlcb.can.CanInterface$FrameParser.send(CanInterface.java:124)
at jmri.jmrix.openlcb.OlcbConfigurationManager$2.reply(OlcbConfigurationManager.java:385)
at jmri.jmrix.can.adapters.gridconnect.GcTrafficController.forwardReply(GcTrafficController.java:42)
at jmri.jmrix.AbstractMRTrafficController.notifyReply(AbstractMRTrafficController.java:191)
at jmri.jmrix.AbstractMRTrafficController$RcvNotifier.run(AbstractMRTrafficController.java:1040)
at java.awt.event.InvocationEvent.dispatch(InvocationEvent.java:301)

[pabender]

Sent from my iPhone

On Jan 14, 2017, at 7:30 AM, Balazs Racz ***@***.***> wrote: when there are multiple computers on a network, it is possible to have JMRI running and a firmware update be happening over the network from another computer. I am getting NPE in the JMRI console when I'm doing this:

Which computer is generating the NPE? What is happening on the computer that is not going the firmware update? Paul

[balazsracz]

The download is performed by a non-JMRI computer. The NPE is thrown by a JMRI on the same bus.

…

-- Sent from my mobile device

[dpharris]

Is this still an active issue?

[bobjacobsen]

I haven't been able to recreate it with two JMRI instances.

But reading the traceback, it looks like the Other Computer was doing stream-based I/O. I might not have been recreating it correctly because JMRI uses datagrams to do firmware download.