Disk image generation

A loadable disk image for the Q1 emulator is a python list

data = [
# track 0
[
  [0x9e, 0x00, 0x00],
  [0x9b, 0x00, 0x00, ... ],
  ...
  [0x9e, 0x00, 0x57],
  [0x9b, 0x00, 0x00, ... ]
# track 1
[
  [0x9e, 0x01, 0x00],
  [0x9b, 0x0c, 0x00, ... ],
  ...
  [0x9e, 0x01, 0x12],
  [0x9b, 0x0c, 0x41, ... ]
],
...
# track 73
[
  [0x9e, 0x01, 0x00],
  [0x9b, 0x0c, 0x00, ... ],
  ...
  [0x9e, 0x01, 0x12],
  [0x9b, 0x0c, 0x41, ... ]
],
]

And can be converted into a filesys object

fs = filesys.FileSys()
fs.loadtracks(disk.data)

which can then be provided to the emulator in the constructor (hardcoded so far)

self.io = z80io.IO(self.cpu.m, fs)

Disk images are currently created from two sources using two different methods: a modified q1decode program from Mattis Lind and a homemade utility binread.py.

q1decode

The disk ‘images’ found here were generated by a modified code gotten from https://github.com/MattisLind/q1decode

The modified code generates an output.py file with a nearly useful python list in it.

The procedure is as follows

> gcc q1decode.c && cat Q1DISKS/PL1/C_S0T01.000 | ./a.out 255
> rm -f output.py
> cat Q1DISKS/PL1/C_S0T01.000 | ./a.out 255

The argument to a.out, in this case 255, must match the record size for that track. This might not always be known but in theory can be found on the file index on track 0.

The generated python code needs to be modified to

ensure that records start at 0
remove trailing ‘,’s
wrap the lists elements in a named outer list

So far I have chosen name the list ‘data’ and put each track in its own python file (see disks/debugdisk).

binread.py

For images produced by datamuseum.dk we use the tool in utils/binread.py to convert from .BIN files to python data structures that can be loaded by the emulator. The utility is compatible with the geometry specifications found from datamuseum’s FloppyTools:

binread.py -f Q1_FLOPPY.BIN -g '1c 1h 88s 40b, 29c 1h 19s 255b, 43c 1h 126s 20b, 1c 1h 19s 255b'