Re: [stella] disassembly requests

Subject: Re: [stella] disassembly requests
From: Thomas Jentzsch <tjentzsch@xxxxxx>
Date: Sat, 18 Aug 2001 09:44:55 +0200
Bob Colbert wrote:
> Please send me the bugfixes so that I can include them on my DiStella web
> page!

See attachment :)
(Now that I've dropped my bankswitch attempts, the
changes have become quite minimal)

> As for bankswitch support, it's not really feasible to use DiStella
> for that.  There isn't really much more automation that you can do without
> creating an emulator.  Since bankswitching can be performed by indirect
> addressing, the only way that DiStella could do that is to emulate a game to
> know which indirct addressing calls actually hit the bankswitching trigger.

Yes, your are right, unfortunately :(
I did some very basic attempts, tested with Asteroids,
and it showed that indirect addressing is indeed a
major problem which I can not solve.

But maybe there's a chance to build disassembly into
existing emulators. (I mean real source code, like it's
produced by DiStella, not trace files like z26 or
runtime debugging like PCAE.)

I'm not quite sure, how this could work. Maybe all
informations could be gathered while playing the game.
Then you would need some information, how much of the
code and data has already been covered during play.
And it would be necessary to be able to merge the
generated output, since for larger games, it should
become impossible to cover everything in one go.

> You can manually disassemble games larger than 4k by breaking the ROM image
> into 2k or 4k chunks and identifying the entry points into each bank.
> That's what I did with Solaris, and I also successfully disassembled Rabbit
> Transit (recompilable with DASM).

I know now, but it would have nice to have this feature
automatically. Not all dreams core true :(

Have fun!
Thomas Jentzsch         | *** Every bit is sacred ! ***
tjentzsch at web dot de |

Description: Zip compressed data

Current Thread