How does one deal with the different register sizes in a 65816 disassembler? So far all I can think of is to try to keep track of SEP/REP instructions.

There isn't going to be a foolproof way - that's the nature of mode bits. So whatever you do by way of guessing, be sure to allow an override.

That's how I do it in Supermon 816.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

Yeah, this is a problem. For the Forth disassembler, I have these routines: In other words, we check the byte after REP and SEP and set internal flags depending (xy:8, a:16, etc). To help with debugging, the disassembler adds a comment to the output so the user knows what the program thinks it knows. If REP or SEP are not used for normal register switches, print a warning.

Thanks. I feel that some (if unreliable) detection strategies are needed. The register size detection seems pretty straightforward, but trying to detect the emulation mode could have problems:


This seems unlikely though?

Yes, that's unlikely!

More like pointless. Why use three bytes to do a two-byte job?

It's a rare case where switching between native and emulation modes occurs as a regular activity. About the only one I can think of off-hand is when a eight bit Apple is retrofitted with an '816.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I don't know of any way to increase the accumulator by 29 or 30 and potentially switch modes in only two bytes of code, so I personally would be more inclined to call it "obscure" or maybe "obtuse", rather than "pointless". Some other "ob-" words also come to mind, like "obfuscated" and "obnoxious".


I installed a 65c802 in my old Apple ][+ back in about 1990, to experiment with the 16-bit registers. It was fun for awhile, but a bit cumbersome, because the monitor was completely clueless about what I was trying to do. I thought about adding the necessary monitor functionality myself, but eventually moved on to a different interest instead. I grew weary of all the mode-switching, which couldn't be avoided to get my experiments to "play nice" with the Apple ][ firmware. I went through my attic recently, but only found a //e and a Franklin Ace up there ... I hope I didn't lose that '802 in one of my moves. Meh ... easy come, easy go ...

Mike B.

I came across an article in Apple Assembly Lines archive on just this topic: http://txbobsc.com/aal/1986/aal8603.html#a4
I also have a 65802 installed in a Laser 128EX that I just got working today after being in storage and moving again and again and again...

Regards,
Andy