Since 4.18 I've been running the console all the time during my session and tonight with 6 smartphone throttles running there were repeated Engine Driver / WiTHrottle drop outs. Androids seemed to recover automatically and iPhones dit not.?

Scanning through the console log I see this:?

2020-01-30 20:40:57,487 ptionhandler.UncaughtExceptionHandler ERROR - Uncaught Exception caught by jmri.util.exceptionhandler.UncaughtExceptionHandler [DeviceServer-335]

java.util.ConcurrentModificationException

at java.util.HashMap$HashIterator.nextNode(HashMap.java:1445)

at java.util.HashMap$KeyIterator.next(HashMap.java:1469)

at java.util.AbstractCollection.toArray(AbstractCollection.java:141)

at java.util.ArrayList.<init>(ArrayList.java:178)

at jmri.jmrit.withrottle.MultiThrottle.passActionsToControllers(MultiThrottle.java:212)

at jmri.jmrit.withrottle.MultiThrottle.handleMessage(MultiThrottle.java:70)

at jmri.jmrit.withrottle.DeviceServer.run(DeviceServer.java:218)

at java.lang.Thread.run(Thread.java:748)

2020-01-30 20:41:03,499 withrottle.FacelessServer ? ? ? ? ? ? INFO? - Creating new WiThrottle DeviceServer(socket) on port 12090, waiting for incoming connection... [WiThrottleServer]

Seems like WiThrottle crashed and re-started am I correct in seeing that. Also is it correct that the DeviceServer -335 is a little high 1h40" into a session with only 6 throttles?

With the session complete I will upgrade from 4.19.1 to 4.19.2 - will I see a difference I wonder?

Also I'm wondering about how the client act when the router is not connected to the internet (it's a stand alone) some phones were trying to use cell data - and might this be resulting in the phone dropping off of the withrottle server?

Thanks,
Dave

John was able to supply me with a set of his operations files.? What we found is that one of his trains was building when the operation files were saved.? Only one train can build at a time, so when the files were loaded the train status was "Building", but the build would never complete as there isn't code in the program to restart a build in progress upon the restart of operations.? The solution was to reset the train that was "Building" before trying to build another train.? The next test release will when loading the operations files, will also check to see if a train is in the building state, and if so, reset it, rather than trying to restart a build that was in progress.

Dan

I concur with Ross,
All of my roster & the other necessary files live on Dropbox.
I share it between my PC & Mac.
Also, when a new version of JMRI comes out, I don't have to worry about losing anything.
Just do the install & go!

Jay

开云体育

开云体育

I use all Apple stuff and Macs.?

Marc,

Yup, fixed by PR #8004, available in build #3621 or later.

Once 4.19.3 is released I'll close

Steve
"Breezlys"

I have my Decoder Pro and JMRI Operations on my laptop. As I was working on building a train my computer went off and would not come back up. We I had dislodged one end of the batter. Plugged in back into the laptop and everything is fine.?

In the mean time the scare of loosing my roster of locomotive decoder setting in Decoder Pro along with my car roster and layout locations. Really gave me a scare.?

So my question is: I want to back up everything to another computer. Can someone tel me how?
I would also use that computer to do the Operations program. Keep the lap top for programing decoders but still keep the back up of decoders on both machines.?

Hi
ON a Pi, or most other Linux systems you can add a udev rule.
In?/etc/udev/rules.d? ?add a rule. Call it something like jmri.rules. It should/must finish with ".rules"

IN that file put

3) Finally there was a bug in the DCC++ code which was tagging an extra 0 byte on the end of the packet on the rails, so I corrected this.

Terry could you please elaborate on this? I am playing with SignalMasts over DCC decoder using NMRADCC library and can see the same very extra 0 at the end of extended accessory packet.

This issue has been address and resolved in Jenkins build 4.19.3,

Tested it on? LB-USB > DCS50 Zephyr?? and Track power remains off.
Tested it? on the Digitrax PR3 >DCS50 Zephyr ? and track power remains off.

Tested it on Digitrax PR3 in Standalone Programmer mode and track power remains off.

Marc

Dave in Trempealeau this question is for you.? I searched but only found thousands of sites related to 4 legged cats.

Dave...in Colorado

Solved.? Problem was with how the RPi assigns ports.? When only the DCC++ was connected it was assigned ttyACM0.? When two devices are connected the first one RPi recognizes is given port ttyACM0 and the second is ttyACM1.? When you plug in a 3rd device OR unplug a device and re-plug it that device will be given port ttyACM2, and so on.

So the RPi saw the port with the LocoNet first and assigned it port number ttyACM0.? I switched the ports in Preferences and the problem was solved.? This would make a great comment to add to JMRI help.

Ken, thanks for suggesting the ports issue.? BTW the defaults were correctly assigned to DCC++.

Thanks everyone - I created a class called runTask? and called it with the standard invocation? runTask.start() as Bob suggested and it worked for me.

Billybob - sorry did not know about the developers group so will go take a look at that for future reference.?

James,

Your issue contains some deep programming issues and may not be of interest to the average jmriusers list member.

The jmri-developers.groups.io list is more programming-oriented, and is a better fit for your issue.

Regards,
Billybob

Take a look at the jython/BackAndForth.py sample script. (In the JMRI distribution or here: )

It creates an indepedent thread via the AbstractAutomaton class, which handles those details. That then waits for sensors and commands a throttle.

For more:





Bob

On Jan 30, 2020, at 1:16 AM, James Anderson <james_anderson_999@...> wrote:

Hi, I have included a small cut down version of something I am trying to do.

I set up a propertyChangeLlistener for a sensor. When the sensor is activated it runs. I have simulated it running a long time just by putting a time.sleep in.
My problem is that when the se.setSate(ACTIVE) in the main function runs. It does not complete until the propertChangeListener completes.

Is there a way to have the propertyChageListener run without the main function waiting for it to complete.

I have looked at some of the documentation on threads but do not understand how to get this to work with jython.
I have used threading.Thread from the standard library start a thread running to do the long running task but this has some really strange behaviour and does not appear to run as thread but runs sequentially.

Some examples of why I want to do this is that when a sensor is activated I want to set a function key on the throttle a number of seconds later and not at the time the senor is activated or to start running some animation program on the layout that
could take some time.

Thanks jim.

Code
---------

import jmri
import java.beans
import time

class _SET_SENSOR(jmri.jmrit.automat.AbstractAutomaton) :

def init(self):
return


def handle(slef):
se=sensors.provideSensor("IStest")
se.addPropertyChangeListener(_LISTENER(), "IStest", "TCPDeviceMgr")
print(str(time.ctime())+" setting ACTIVE")
se.setState(ACTIVE)
print(str(time.ctime())+" setting INACTIVE")
se.setState(INACTIVE)
return


class _LISTENER(java.beans.PropertyChangeListener):
def propertyChange(self, event):
if event.newValue==2:
print(str(time.ctime())+" PropertyChange Activated...")
se=sensors.getSensor("IStest")
for oldListener in se.getPropertyChangeListenersByReference("IStest"):
if se.getListenerRef(oldListener) == "TCPDeviceMgr":
print(str(time.ctime())+" Removing old listner for sensor : IStest")
se.removePropertyChangeListener(oldListener)
time.sleep(15)
print(str(time.ctime())+" PropertyChange Complete")
return

test=_SET_SENSOR()
test.start()


OUTPUT from script
---------------------------
Thu Jan 30 08:48:24 2020 setting ACTIVE
Thu Jan 30 08:48:24 2020 PropertyChange Activated...
Thu Jan 30 08:48:24 2020 Removing old listner for sensor : IStest

--->>> 15 second gap for the time.sleep(15) inside the propertyChangeListener function.

Thu Jan 30 08:48:39 2020 PropertyChange Complete
Thu Jan 30 08:48:39 2020 setting INACTIVE

--
Bob Jacobsen
rgj1927@...

Thanks Steve, the main problem has been resolved, actually quite simple, I had a turnout MT+N0E9 as well as MT+N0E09, same for E6 and E7.
Checking the turnout table in 4.14 only showed N0E9 and not N0E09 and I guess the new address validation in 4.18 quite correctly picked up on this.
No idea why but could not delete the turnouts using the button in the table, clicking delete had no effect even though I had removed the signal heads associated with them, no error message, edited the XML and the panel now works with 4.18.

PT5/6T-11T-PT7C-PT8T-PT9T etc. are just user names, the system name follows the standard JMRI / MERG naming convention.
I use Logix to determine the state of the five turnouts, "PT5/6T-11T-PT7C-PT8T-PT9T"?will be Thrown if any of the turnouts are not set as required, setting PT5/6T-11T-PT7C-PT8T-PT9T to Closed changes the turnouts to the required position by sending a specific event e.g.+N0E103 which has been taught to each of the turnouts. Makes the Jython script a lot simpler, just one route to check and set.
I used Node 0 as I thought this was necessary as the event is produced by the PC / JMRI but since learnt that I could used any node number providing it was not a duplicate.

Thanks
Pete

The release notes already cover how to use the old CSV API by linking into the documentation for setting JMRI parameters.

The draft release notes for 4.19.3 are at

Dave,

You should do the test of watching /dev while plugging and unplugging the
devices. From the confused traffic, I would suspect that the picks might be
reversed for the ports. ttyACM1 and ttyACM0 may not be the device you think
it may be.

Make sure of which name shows up when you plug in the device.

-Ken Cameron, Member JMRI Dev Team
www.jmri.org
www.fingerlakeslivesteamers.org
www.cnymod.org
www.syracusemodelrr.org

Dave,

You might want to check the Edit->Preferences->Defaults, it may not have the
selections you think it should. It may have some things defaulting to the
LocoNet that should be on the DCC++.

-Ken Cameron, Member JMRI Dev Team
www.jmri.org
www.fingerlakeslivesteamers.org
www.cnymod.org
www.syracusemodelrr.org

One more thing, on the PR3 the green LED is mostly on and winks off and the red LED is on solid indicating no LocoNet signal.? So yesterday I unplugged the BDL and plugged a LocoNet cable from a DCS50 Zephyr into the PR3.? The red LED went out, but the problem in JMRI with Unknown throttle was still there.

Below is the info from the System Console: