Release Notes for JMRI 3.7.1 release

Date: January 14, 2014

From: Bob Jacobsen

Subject: Test version 3.7.1 of JMRI/DecoderPro is available for download.

Notes:

This is a test release. Please post a note if you encounter any new or old bugs! And please back up your JMRI files before installing this, in case you want to go back to an earlier version.

This is the first in a series of test versions. We expect this series to end in a JMRI 3.8 production release around the middle of 2014.

Some of the changes involved are quite extensive. They may require a certain amount of experience before they are working well. Therefore, this test version should be considered experimental.

JMRI is now only available under the GNU General Public License. For more information, please see our copyright and licensing page.

New warnings for this version:

The internals for handling CV read and write addressing have been extensively reworked. The change should be transparent (invisible) to users. If you do notice anything problems involving CV reading, writing, storage or anything else, please let us know.

The configuration structure for JMRI is changing to use Configuration Profiles. Configuration Profiles allow a single JMRI application to have multiple, user selectable, configurations that can be changed as application startup. For some users, this change will have no impact; other users will be visibly impacted by this change. See http://jmri.org/help/en/html/profiles.shtml for more information.

Older warnings

This release, and all other releases in this cycle, require Java 1.6 or newer.

Known problems with this release

Do not use this release for programming any ESU Select or V4.0 decoder. Instead, please use test release 3.7.7 or later.

Download links:

Please use these links for downloading:

Changes since production release 3.6:

OpenLCB library updated to better handle hub operations.

Several internal updates for how CV reading and writing is changed to improve handling of indexed CVs. If you have your own code or scripts that write or read CVs directly, check to see if we've broken anything, please.

Improved the handling of decoders that contain only long addresses, no short addresses.

Peter Ulvestad updated the Hornby 8249 default values.

Nigel Cliffe and Dave Heap fixed problems with international characters in four Python scripts: AAR105.py
RailDriver.py
RocoCrane46902.py
Zimo_pseudo_programmer.py

Added sample scripts:

• jython/SensorToTurnout.py: takes Sensor inputs from an external system and maps them to Internal turnouts
• jython/SignalHeadFromSensors.py: takes Sensor inputs from e.g. an external signal system and drives SignalHead objects

Dan Boudreau made the following changes to operations:

• The data in operations files has changed with this release. Previously the files contained backward compatibility elements so that you could load a file created with the latest release and read it back using one of the very early releases of operations. Now the backward compatibility is limited to releases 3.4 and 3.6. In any case, you can always upgrade to a newer release of operations, it is only the ability to go back to an older version of the program using the latest data files that has changed.
• The lists that operations provides for scripts has changed significantly. Instead of lists containing ids of the objects, the actual object is now passed in the list. For example a list of cars in operations was a list of car ids, and not the cars themselves. If you use scripts you will need to edit your scripts. Here's the changes you need to do for cars:

  In each of your script files, search for "carId". You should find something like this:

  for carId in carList:
  car = cm.getById(carId)

  Change the first line to:
  for car in carList:

  And delete the line that has "car = cm.getById(carId)".

  For scripts that deal with trains, search for "trainId". You should find something like this:

  for trainId in tList:
  train = self.tm.getTrainById(trainId)

  Change the first line to:
  for train in tList:

  And delete the line that has "train = self.tm.getTrainById(trainId)".

• New feature, you can now let the program create your locomotive consists for you. See Train Build Options "Build loco consist from single locos if needed" for more details. The program when building the consist can adjust the number of locos based on Horsepower per Ton (HPT) and the route's grades.
• New feature, you can now adjust all of your train departure times by a selected number of hours. See under "Tools" in the "Trains" window "Change Train Departure Times".
• New feature, you can now edit or view two locations at the same time.
• The train edit window now shows if there are road or load restrictions for the train.
• Fixed a bug in track copy where the car types serviced weren't copied to the new track.
• Fixed a bug in train copy where the train build options weren't copied to the new train.
• Fixed a bug in the router when the option to return a car to staging is enabled, and the car get's a custom load from staging.

Allow Turnout icons to be "momentary". (Sensor icons could already do this)

Dave Heap improved the handling of DCC addresses. The decoder terminology has been standardized. There's more checking of values, to prevent erroneous behavior when an out-of-range value is entered.

Paul Bender updated the XPressNet network connections (LIUSB-Server, LIUSB-Ethernet, and XnTCP) to allow automatic reconnection attempts in the event of a dropped network connection.

Paul Bender improved support in the generic network client for automatic configuration. This includes improvements to the generic mDNS/ZeroConf/Bonjour automatic configuration tools.

Paul Bender added mDNS support to the Lenz LIUSBEthernet adapter.

Suzie Tall updated the BN 1989 signal definitions. Changes include:

1. Additional panel icons to support the new signal mast definitions.
2. A modified icon which was stopping the original 2-head mast from being added to a panel.
3. Twelve new signal mast definitions.
4. An updated index.shtml with a bit more information, reference to BN source and links to the new masts.
5. An updated aspects.xml with the addition of an unlit appearance and corrections to the appearance definitions.

Suzie Tall contributed a new signal system definition for the LMS speed signaling that was installed in parts of the UK in 1932 and continued in use until 1988. The definition is as complete as is possible from the available documentation and covers all variations over the entire period of operation.

Pierre Billon contributed four additional TAMS decoder definitions, and more translation to French and German.

Peter Ulvestad contributed new decoder definitions for MRC 1641,1642, 1659 and 1664 N scale motor control only decoders.

John Lang updated the NTC 1956 signal files.

Dave Heap improved the ESU V4/Select function mapping pane

Dave Heap improved handling of long addresses in NCE consists

Enzo Fortuna added additional Italian translation terms, and fixed a couple of existing ones.

Suzie Talk added a new BNSF-1996 signal system.

Randall Wood added configuration profiles.

The internals for handling CV read and write addressing have been extensively reworked. The change should be transparent (invisible) to users. If you do notice anything problems involving CV reading, writing, storage or anything else, please let us know.

Matthew Harris:

• Updated the Windows Launcher to allow for a configuration profile to be specified. For further details, see technical details of how the startup scripts work
• Updated the Windows Launcher to bring memory calculations in-line with Linux and OS X

Suzie Tall added non-rotating icons to the BNSF-1996 signal set.

Jim Duncan provided a set of icons for building an Armstrong tower control panel.