[SHARP X68000模拟器]XM6 Pro-68k Release L (230530)

疾风之狼

The Num Lock LED now retains its state when it is being used as an X68000 key.
(By default, Num Lock is mapped to the CLR key.) In other words, if you press
Num Lock, and it is interpreted as an X68000 key, the LED will revert to its
previous state after a fraction of a second. The LED will still toggle as usual
within dialogs and similar contexts. If you dislike this feature, remap CLR
(key 3F) in the keyboard settings to any other key, or remove its mapping.

The above feature can fail, but it only happens rarely and certain keyboards
never seem to have an issue. The problem is unlikely to occur during typical
usage of the emulated CLR key.

For consistency, the default directory is now set to the current directory at
the time the program starts, unless the current directory is someplace stupid.

The disassembler now flags a number of encoding errors that are ignored by the
68000. These typically arise when data is misinterpreted as instructions, but
can also be due to assembler bugs or the use of modes exclusive to newer CPUs.

The disassembly of bit manipulation instructions (BCHG, BCLR, BSET, and BTST)
now displays the bit number in decimal form, which is generally more useful.
The number is also masked to indicate how it will be interpreted by the CPU.
If the original bit number was too large, one or two question marks will be
appended, the latter being the case if it exceeded the legal 8-bit range.

SWAP was being disassembled with a .W extension, which could be misleading.
The extension is now omitted. (It swaps word-sized halves of a long-word
register, and the condition codes are set based on the 32-bit result.)

Some efficiency improvements were made concerning file I/O.


Bug fixes:

Certain configuration items were being updated even when unused. Now that the
configuration file is only updated if different, this caused unnecessary writes.

Automatic Controller Mode wouldn't activate for certain unusual configurations
if the Prefer Real setting was enabled, which it is by default.

Fixed a few minor decoding errors in the disassembler. It will still blithely
decode many invalid combinations, however.

MOVE to/from CCR was being disassembled with a .B extension which is just wrong.
(The CCR is 8-bit -- 'tis true! -- but these two instructions are word-sized.)

http://mijet.eludevisibility.org/XM6%20Pro-68k/XM6%20Pro-68k.html