Hi.

I've read in Twitter the existance of the source code for the Furby toy of 1998.

And guess what? it's written in 6502

http://www.seanriddle.com/furbysource.pdf

_________________
Author, IOCCC+JS1K winner, Princess Quest in Evercade Intellivision, MSX/Atari/CV/Intellivision/SMS game dev, Knight of boot sector games, IntyBASIC/CVBasic creator.

Cool. Thanks for sharing.

_________________
Michael A.

An excellent find!

First thing to note is that it's not quite a 6502 - there's no Y register and some addressing modes and many opcodes are missing. But the source certainly looks like 6502 and would probably run on a 6502.

I noticed a couple of oddities:

- they always CLC before SBC, so maybe SEC is missing? It makes the result out-by-one, unless maybe SBC has the opposite sense of carry on this chip. Or, witness the comment on the front page, the programmers hadn't quite got the idea of borrow and SEC.

- they PHA and PHP in their interrupt routine. Is that PHP a mistake, or does RTI work differently on this chip? If it does, then PLA needs not to affect the flags.

Also available at the Internet Archive in their book viewer.

See also the discussion on Hacker News - with links to more readings.

See also this previous post.

(Text above mostly recycled from the recent gplus post by mos6502.)

The Sunplus SPC81A microcontroller was mentioned here in the forum in 2017.
The datasheet doesn't explicitly tell you that there is an Y register, so we could assume that this register is missing.
The data sheet doesn't tell you anything about the instruction set, either.


A manual for 6502 related Sunplus instruction sets seems to be here.
//'Google Analytics' again, you better go with AdBlocker, NoScript etc.
Starting at page 52, there seems to be a list of the instructions for the different instruction sets...
...but I just fail to spot an instruction set where the Y register is missing.

Hmm... that Chinese Sunplus MCU page wants you to log in with a password when klicking "English".

Thanks for the link!


Yes, see for example TAY, the instruction has cycle counts for two flavours of core but an X for the reduced flavours.

When taking a closer look at the addressing modes for LDA...


Aha: _all_ variations of LDY are missing for 65r02, 65s02 and CPU12... plus quite a few other instructions.