aboutsummaryrefslogtreecommitdiff
path: root/fontrip.txt
blob: 0fa8fea92f9c4f258b694372de17bfaf150db214 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41

process of ripping the font was something like...

look at a screenshot of the Apple 2 version, pick a byte that occurs on
2 successive lines in a couple of the characters (I chose the 10100000
that shows up in capital C and O).

write perl script to find occurrences of 2 of that byte (findfont.pl)
and print offsets of each one.

use bitmapdump.pl to eyeball the file at those locations, until I found
the letter C, then look up & down to find where the visible characters
start and end. Use dd to grab just those bytes into another file.

..except, the above was fruitless until I got the idea to search for the
byte with its bits reversed (00000101). Turns out, the font is stored
in the disk image with the bits reversed. Maybe if I knew the Apple
architecture, I'd have already known that?

Anyway. To create an atari font, I mapped out the order the characters
appear in the dump, and wrote convfont.c to overlay them onto the Atari
ROM font (created from atariosb.rom with dd)... and for some reason,
a few of the characters have vertical bars in the rightmost position,
caused by bit 7 (in the original reversed format) being set. convfont.c
just zeroes those out as needed.

The end result is taifont.xex, containing 1024 bytes of data to be
loaded at $BB00.

To get it to work with a cc65 program, I had to create a linker script
(custom.cfg). Ran this:

ld65 --dump-config atari > custom.cfg

Then edited it... all I changed was the ending address, which is
now $B7FF (one byte below the font's load address). After compiling a
taipan.xex with this config, I append taifont.xex and taipan.xex to make
taipantext.xex (the finished game).

The only thing the C code needs to do is POKE(756, 0xbb) in main()
before printing any text.