Memory Read space is sometimes incorrect

[Poikilos]

This PR #96 is almost ready. There seems to be only one problem (at d60cc42). When I read from space 251 I am getting XML instead of the value there.

• The issue does not occur on every segment.

Relevant part of CS-105's CDI:

    <segment space='251' origin='1'>
        <name>User Info Settings</name>
        <description>Change name and description of this Command Station for LCC network browsers.</description>
        <string size='63'>
            <name>User Name</name>
            <description>This name will appear in network browsers for this device.</description>
        </string>
        <string size='64'>
            <name>User Description</name>
            <description>This description will appear in network browsers for this device.</description>
        </string>
    </segment>

Steps to reproduce issue:

• Open examples_gui.py
• For IP type "Detect".
• Manually enter the destination Node ID & IP address/hostname, or do the following detection steps:
  • Choose a command station from the TCP service name list.
  • Click "Copy digits to Far Node ID"
• Connect
• Refresh
• Expand "User Info Settings"
• Select "User Name"
• Click "Read"

Expected result: Read same value as PanelPro's LCC Configuration dialog

Actual result: We get the XML according to the offset and length given (wrong space). In this case, '?xml version="1.0" encoding="utf-8"?>\n<cdi xmlns:xsi="http://ww'.

• Console output:

read: space=251=0xfb address=1=0x1 size=63=0x3f
No null terminator in bytearray(b'?xml version="1.0" encoding="utf-8"?>\n<cdi xmlns:xsi="http://ww')

• JMRI gets the expected value (a string previously set by the user), demonstrating that value itself is not mangled on the device.

An example of a space that works fine is:

    <segment space='253' origin='128'>
        <name>System Settings</name>
        <description>Configuration options related to the DCC functionality, including Track Output.</description>
        <group offset='4'/>
        <group>
            <name>Track Output</name>
            <description>Configuration options for the Command Station's Mainline Output, such as short circuit limits.</description>
            <int size='2'>
                <name>Booster Trip Current</name>
                <description>Current limit in milliamps at which the booster will detect an overload and shutdown (1000 to 5250 milliamps). Inrush current above this threshold is allowed.</description>
                <min>1000</min>
                <max>5250</max>
                <default>5250</default>
            </int>
            <int size='2'>
                <name>Booster Trip Inrush Delay</name>
                <description>This is the delay (10 to 320 milliseconds) to hold off on tripping the booster if the "Booster Trip Current" limit is reached.</description>
                <min>10</min>
                <max>320</max>
                <default>320</default>
            </int>
            <int size='2'>
                <name>Booster Kill Current</name>
                <description>Current limit in milliamps at which the booster will detect a short circuit and shutdown (2000 to 9000 milliamps). This is an absolute limit regardless of inrush current. Typically this value is equal to or larger than the "Booster Trip Current".</description>
                <min>2000</min>
                <max>9000</max>
                <default>9000</default>
            </int>

[bobjacobsen]

it would be interesting to get an LCC trace of what happens when User Info Settings -> User Name is read. My hypothesis is that the code is messing up the address space in the read command. Space 253 is one of the spaces that are handled via 0x41 in byte one of the Read command, but space 251 has to be conveyed with an 0x40 in byte 1 and 251 in byte 6. Maybe the logic for going from space number to needing byte 6 has a problem.