Re: Experimental "Flatpak" distribution of JMRI
Linux Crew:
Playing further with my flatpak install...
One hoped for feature with a Linux release is a seamless update of a system.
The flatpak install uses the same files structure for data but I have a problem finding the "programs" the scripts...听 No big problem there were launcher icons installed in "other" in the start pane but I wanted to update desktop icons that are easily built in Mint Cinnamon.
The help show files gives me an /app/share/JMRI but that's not obvious where that actually is... there is no app directory in Home or the higher File System...
So for me... interesting... but I would ask the developers to get the install to 100% of the standard we have or not worry about the flatpak version...
I have several Linux/Mint/JMRI installs that I've gifted/built on recycled PC's that the easier an update the better... We are generally talking "noob" class computer users let alone Linux users.
Jim Albanowski
|
Re: JMRI and LogicX driving Signals
#signals
Randall, 听 What system is your layout? Like what system are you using for the turnouts? Driving a signal lamp is really the same as driving a turnout motor. At least to JMRI. Look at 鈥楽ignal Heads鈥�, most of those outputs are using a turnout entry to drive the lamps. 听 -Ken Cameron, Member JMRI Dev Team
听 听
|
Re: Signal Controller Board Suggestions
You might also look at the standalone application for yard ladder control which I also implemented.
--
Ken NYNH&H, Old Colony Division, Cape Cod Branch (1949-1959) Loconet * JMRI 5.11.4 * OSX,Win10,Ubuntu
Blog: 听
Youtube:
|
Re: Signal Controller Board Suggestions
I use QuadLN_S boards for my turnout, detection, and signal interfaces.
When used with JMRI Signal Mast Logic, the Quad translates Loconet signal messages into LED aspect indications.
I run my entire railroad with JMRI and the Quads.
听
--
Ken NYNH&H, Old Colony Division, Cape Cod Branch (1949-1959) Loconet * JMRI 5.11.4 * OSX,Win10,Ubuntu
Blog: 听
Youtube:
|
Phil G,
听
Changing CV 29 to 7 results in the loco responding to the short address in the same way as previously reported. Meaning sounds are fine, but no loco movement.
听
Regards
听
|
Peter,
听
I had changed the Brake Key to FKey 7, but I always check to confirm that all FKeys are Off.
听
Phil G,
听
Let me try that and report back.
听
Regards
|
JMRI and LogicX driving Signals
#signals
Wanting to drive a Signal based on a decision 'to set green / red' from a LogicX - already set a sensor 'light' on the screen and switch turnouts,, looking for guidance to then set a dwarf signal.
听
unable to use turnout motor switches as there are two x three-way turnouts used.
听
If there is a write up or Youtube I would appreciate in pointer, all i have seen so far are track sensor based and switched from a single turnout motor.
听
Happy to answer any questions...
听
thanks
Randall
|
Signal Controller Board Suggestions
I have JMRI LogicX working through a tree/ladder of turnout to set 'a road in a yard' to the 'Main' from a touch sensor on a Panel.听 Due to space, I am using two x three-way turnouts meaning I cannot use the turnout motor internal switches to switch leds on the Dward Signals on each road.
听
On the screen, I have a 'sensor' light next to the Road when it is clear to the Main.听 All good so far and works fine.
听
Looking now for an addressable signal controller board to receive a JMRI command to turn a 'dwarf' signal from 'red' to 'green' when the sensor on the panel is selected (when it sets the required road to main).听 听The signal controller will connect to an AUX bus connected to, in this case, a NCE PowerPro Controller.听
听
What Signal Controller Boards are you using to take the command to change the LEDs from JMRI?
听
happy to answer any clarifying questions!
听
Thanks
Randall
听
听
听
听
|
Richard,
Please post a screenshot of the About screen from EngineDriver while connected.
What address (exactly) are you typing in? NCE has some odd addressing, IIRC.
Also, try turning off the track power from EngineDriver and controlling Turnouts from EngineDriver. All of these will help narrow down the problem.
--SteveT听
|
|
|
Testing the brake is simple. F11 is the default for these decoders. Make sure F11 is off. -- Peter Ulvestad Linux Mint 22.1, JMRI 5.11.4plus, Java 21.0.6 JMRI Users Group Moderator ( /g/jmriusers ) JMRI Developers Group Moderator ( ) Tam Valley Group Moderator ( ) Sprog-DCC Group Moderator ( ) Edmonton Model Railroad Association ( )
|
Just for a sanity check, if you change cv29 to 7 does it respond to the short address??
toggle quoted message
Show quoted text
On 4 Apr 2025, at 11:46, Don Shroyer via groups.io <Donshroyer@...> wrote:
锘�
Phil G,
听
CV 1: 9, CV 17: 213, CV 18: 96, CV 29: 39, CV 19: 0. While there is a Primary Address of 9, I use the Long Address 5472. I ran the loco reversed.
听
And Acceleration and Deceleration are 25 and 15.
听
Regards
|
The 'delay to inactive' technique worked. I'm glad that I was made aware of these 'debouncing delays'. I think they might be useful to clean up some flaky sensor data that we have been dealing with.
听
George
|
Re: Experimental "Flatpak" distribution of JMRI
I understand that. The "tricky bit" for both MS-Windows and MacOS installs is getting and installing the "proper" JAVA JRE, not actually installing JMRI itself. That seems to be what non-techies seem to have the most trouble with. And probably the most confusing thing is that the version of Java at java.org -- Oracle's JAVA -- is the wrong one. This is where something like a "Flatpak" type of install might be helpful. Getting the correct version of JAVA for *Linux* is pretty trivial, by comparison, since almost all Linux distros have OpenJDK in their repo. Here having a RPM or DEB package with a proper depenency for OpenJDK could be helpful, eg: sudo apt install ./jmri-5.12.deb would then go after openjdk-17-jre (and openjdk-17-jre-headless) automagically. At Fri, 04 Apr 2025 10:29:53 -0500 "Dave Sand" <ds@...> wrote:
Robert,
The macOS install is essentially the same as the Linux install. Download a DMG, open it and drag "JMRI" to "Applications".
Dave Sand
----- Original message -----
From: "Robert Heller via groups.io" <heller辝[email protected]>
To: [email protected]
Subject: Re: [jmriusers] Experimental "Flatpak" distribution of JMRI
Date: Friday, April 04, 2025 8:34 AM
This seems like a "solution looking for a problem", at least for *Linux*.
Since JMRI is in JAVA and requires nothing outside of JAVA's JRE and
installing a suitable JRE under Linux is "trivial" (compared to MS-Windows or
MacOS), I am not sure how truely useful or needful this is.
I can see a flatpak (or similar) for MS-Windows or MacOS, since it appears
that those two O/S's have the greatest problems with installing JMRI, mostly
caused by confusion about installing the proper version of a JRE.
*I* would find a Linux "Flatpak" distribution of JMRI to be far more hassle to
deal with than the current Linux distribution methodolgy of JMRI. Actually, I
expect that the current Experimental "Flatpak" distribution of JMRI won't work
on my machine(s) anyway -- I expect it is a x86_64 Flatpak and all of my
machines use ARM processors. And this brings up a *new* issue that will
complicate things: there would need to be *four* Linux flatpaks: ix86, x86_64,
armv7l, and aarch64.
My only thought about "improvements" in the Linux distribution of JMRI would
be the creation of .deb and .rpm distrubutions of JMRI. I realize that JMRI
will never be in any Linux distro repo and I understand why that is, but a
self-made .deb and .rpm with a proper openjdk-XX-jre depenency could be
helpful. *I* could probably help with creating the necessary control files to
help with this, if anyone is interested.
At Fri, 4 Apr 2025 08:19:38 -0400 "Bob Jacobsen via groups.io" <rgj1927@...> wrote:
A user has been interested in possibly distributing JMRI as a 茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆FlatPak茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆 for Linux. I don茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆t know much about that method, but apparently the distribution file also includes the JRE, so it茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆s (at least in theory) easier to install.
As an experiment, you can find a 5.10 distribution in this form at
For more on this, see JMRI/JMRI Issue 11658:
Is anybody familiar enough with this distribution method to be able to try it?
Is this something that we should do on an ongoing basis?
Bob
茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆茂驴陆
Bob Jacobsen
rgj1927@...
--
Robert Heller -- Cell: 413-658-7953 GV: 978-633-5364
Deepwoods Software -- Custom Software Services
-- Linux Administration Services
heller@... -- Webhosting Services
-- Robert Heller -- Cell: 413-658-7953 GV: 978-633-5364 Deepwoods Software -- Custom Software Services -- Linux Administration Services heller@... -- Webhosting Services
|
Re: Sensor Query wrong for Digitrax BDL716 in 5.11.4
Thanks for the sensible explanation. I didn鈥檛 have detection in the yard until I bought a second unit so I didn鈥檛 notice.
|
Re: Sensor Query wrong for Digitrax BDL716 in 5.11.4
I don't think the BDL's have the ability to respond to a query.
听
It would imply that the BDL has the ability to read an translate a query message and the perform a status lookup
on a particular detection port.
听
The BDL creates messages based on trigger events. Only when a unit enters a block or exits a block will a message be generated.
听
Inobu听
|
Sensor Query wrong for Digitrax BDL716 in 5.11.4
Version 5.11.4+ R87a9dadc80 (Java 17.0.13)
The BDL168 is discontinued and replaced with BDL716 which I have two installed on LocoNet.
听
Upon starting JMRI,
- if the sensor is high as indicated by the LT5C LED, JMRI shows UNKNOWN then INACTIVE when I QUERY (wrong)
- if the sensor is low as indicated by the LT5C LED, JMRI shows UNKNOWN then INACTIVE when I QUERY (correct)
During use the sensors correctly show inactive and active so changed states work correctly.
I don't recall this behavior using version 5.10 but I could be mistaken and I also added a second BDL716 at the same time I upgraded JMRI.
The problem is JMRI doesn't "see" the trains sitting in the yard although the LT5C clearly shows high/active so I have to run the train out of the block and back in to initialize the sensor at the correct level.
听
//stevemac
|
Re: Experimental "Flatpak" distribution of JMRI
Robert,
The macOS install is essentially the same as the Linux install. Download a DMG, open it and drag "JMRI" to "Applications".
Dave Sand
toggle quoted message
Show quoted text
----- Original message ----- From: "Robert Heller via groups.io" <heller@...> To: [email protected]Subject: Re: [jmriusers] Experimental "Flatpak" distribution of JMRI Date: Friday, April 04, 2025 8:34 AM This seems like a "solution looking for a problem", at least for *Linux*. Since JMRI is in JAVA and requires nothing outside of JAVA's JRE and installing a suitable JRE under Linux is "trivial" (compared to MS-Windows or MacOS), I am not sure how truely useful or needful this is. I can see a flatpak (or similar) for MS-Windows or MacOS, since it appears that those two O/S's have the greatest problems with installing JMRI, mostly caused by confusion about installing the proper version of a JRE. *I* would find a Linux "Flatpak" distribution of JMRI to be far more hassle to deal with than the current Linux distribution methodolgy of JMRI. Actually, I expect that the current Experimental "Flatpak" distribution of JMRI won't work on my machine(s) anyway -- I expect it is a x86_64 Flatpak and all of my machines use ARM processors. And this brings up a *new* issue that will complicate things: there would need to be *four* Linux flatpaks: ix86, x86_64, armv7l, and aarch64. My only thought about "improvements" in the Linux distribution of JMRI would be the creation of .deb and .rpm distrubutions of JMRI. I realize that JMRI will never be in any Linux distro repo and I understand why that is, but a self-made .deb and .rpm with a proper openjdk-XX-jre depenency could be helpful. *I* could probably help with creating the necessary control files to help with this, if anyone is interested. At Fri, 4 Apr 2025 08:19:38 -0400 "Bob Jacobsen via groups.io" <rgj1927@...> wrote:
A user has been interested in possibly distributing JMRI as a 锟斤拷锟斤拷锟斤拷FlatPak锟斤拷锟斤拷锟斤拷 for Linux. I don锟斤拷锟斤拷锟斤拷t know much about that method, but apparently the distribution file also includes the JRE, so it锟斤拷锟斤拷锟斤拷s (at least in theory) easier to install.
As an experiment, you can find a 5.10 distribution in this form at
For more on this, see JMRI/JMRI Issue 11658:
Is anybody familiar enough with this distribution method to be able to try it?
Is this something that we should do on an ongoing basis?
Bob
锟斤拷锟斤拷锟斤拷
Bob Jacobsen
rgj1927@...
-- Robert Heller -- Cell: 413-658-7953 GV: 978-633-5364 Deepwoods Software -- Custom Software Services -- Linux Administration Services heller@... -- Webhosting Services
|
Re: How to delete a turnout with jython
I note that scripting and Java is listed in the layout automation section of the JMRI website. Perhaps that is why I am sensing discomfort with regard to what I thought was a simple programming question. No I do not intend to run scripts to operate an actual layout that would involve deleting turnouts or anything else. However, I do find that working with the layout editor and Panel Pro in general is frustrating and I am exploring alternative approaches to getting things done. For example I have used the AnyRail scripts to set up initial panels and found that the exported file from AnyRail has seemingly added tiny track segments with the incorrect block assignments around turnouts and that requires painstaking editing before signal masts can be placed easily. Additionally, part of the AnyRail to JMRI process involves using a script to place sensor icons. It places them directly over a track segment selection circle, making it almost impossible to select the icon to move it. I have not looked yet but have to believe it is relatively easy to iterate over the sensor icons and change their coordinates. Probably easier to edit the existing script for my own use to permit an offset when adding the sensor icons. But it might turn out to be useful to know how to delete things as part of an editing procedure and that is basically all I was thinking about.
--
-Ralph
|
Re: Cannot start to read decoders
听
.started completely from scratch. Deleted all remnants from previous installs.
.Installed java 17 as suggested in windows installation
.installed jmri - no problem so far.
.Before using DecoderPro I verified connection with the central unit and programming track with loko placed on it
.Started up DecoderPro with empty roster/lokpark- clicked on New loko
.New loko window came up (it is rather small if that is of any significance - see later)
.Klicked read decoder type
.came back with list of decoders (where 4 zimo decoder)
. selected correct decoder (highlight took some seconds!)
. Then entered the id name on the right side - took another 2-4 seconds until decoder address was read and displayed
.Saved the loko to the roster / lokpark
.Status line: Pause
.Exited new lok window - clicked on "Identify"
.correct entry in the Roster (there is just one) is highlighted.
.Klicked exit and save!
.Restarted DecoderPro and klicked new loko
. This time a almost double wide New loko window came up
."Read decoder type" grayed out as well as "open detailed programmer" !
.Staus line: Pause
.Klicked exit and save the log files
Notes:
1) This is windows 11 24H2 all updates applied, iTrain, Central sation Mgmgt software, Anyrail all are working ok.
2) I tried English language as well as my own German - no difference
3) I have tried with the JMRI 5.11.4 test release as well as the current release
4) I have the same JMRI installation runing on another laptop听 - Windows Version听 is Windwos 11 22H2
Thanks for your patience - if you need more info/tests just let me know
Erik
听
|
Ken Cameron
Ken Heywood
Peter Ulvestad
George Hofmann
One for the more learned amongst us to have a look at.
听
I have the following set up听:
听
NCE Powercab, JMRI running on a Raspberry Pi via NCE USB using M Steve Todd's files, Panel Pro with a touchscreen panel to operate turnouts, turnout motors are DCC Supplies Cobalt Digital IP.
听
The touchscreen panel works exactly as intended to control the turnouts.
听
I would like the ability to use my phone to become a second throttle.
听
With Wi-Throttle server running, I connected my phone to the WiFi network created by the Raspberry Pi, opened up the Engine Driver app & connected to the server.
听
And here is where it goes wrong. Typing a loco address and then trying to move something , absolutely no response from the loco. If I try to operate a turnout motor via the accessory address , it works perfectly. The loco isn't stuck in consist mode as it works off the bat from the NCE handset using the same DCC address so as far as I know I'm not using the wrong address.
听
On the Wi Throttle server screen I can see the phone throttle is connected and even the address of the locomotive I'm trying to control. The Track power is on.
听
Can anyone give any pointers here?
听
thanks in advance