Robert Heller

Since at least 4.20. I'm downloading the source and have a look...

> The form in the OpenLCB Turnout tests never try with leading zeros. In fact all tests us single digits in each byte of the address. Well, that was dumb.....

After a very little bit groveling in the source code and inserting some printf()s it appears that the system names are getting double prefixed: [java] *** Manager.makeSystemName(MS02.01.57.00.01.A2.00.66;02.01$ [java] *** OlcbAddress.svalidateSystemNameFormat(M2SMS02.01.57.00$...

This is using the "Make Sensor" functionality in the CDI configuration tool, not from the PanelPro Sensor table Add button, but I expect there is a similar problem -- there...

> The M2S prefix indicates the second CAN connection. The first connection would be MS. > > It is possible that the CDI tool is assuming one CAN connection which...

Here is the fix. It appears that the OpenLCB ClientActions is adding in its own prefix, that the rest of program does not expect or know to deal with. [ClientActions.patch.txt](https://github.com/JMRI/JMRI/files/6722283/ClientActions.patch.txt)

At Sun, 27 Jun 2021 13:54:29 -0700 JMRI/JMRI ***@***.***> wrote: > > > > No, M2S indicates a two event sensor. The MS prefix is exranous. > > Then that's...

At Sun, 27 Jun 2021 14:27:03 -0700 JMRI/JMRI ***@***.***> wrote: > > `SensorManager` and `TurnoutManager` instances know about the proper prefix > for a connection; the GUI (usually) doesn't unless...

At Sun, 27 Jun 2021 15:25:55 -0700 JMRI/JMRI ***@***.***> wrote: > > > Neither (should) have a prefix. The "address" is two event ids separated by a > > semicolon....

> The input validation (while you type) accepts the values as copied from most uses of the event ids and shows green that it is ok. The hover text correctly...