Feedback from ROM Hackers please

[Nach]

For those of you who use ZSNES' debugger for ROM hacking, please vote.

[Nightcrawler]

I think most people don't use the ZSNES debugger anymore since it was frozen in time and better solutions became available.

Beyond that, I don't remember that option anymore, but I think I'd say yes to effective address reporting. I assume this option gives the effective address operated on for instructions that don't explicity use a full 24-bit address? If that's the case, that's extremley useful.

[Louis Cypher]

ZSNES has a debugger? Oh wait, ZSNES DOS, right - I stopped using that as soon as Geiger's SNES9X debugger became available. It was only useful for stepping through code anyway.

[SamB]

I assume this option gives the effective address operated on for instructions that don't explicity use a full 24-bit address? If that's the case, that's extremley useful.

Well, that is certainly what it is supposed to do, but it is actually very buggy at the moment.

ZSNES has a debugger? Oh wait, ZSNES DOS, right - I stopped using that as soon as Geiger's SNES9X debugger became available. It was only useful for stepping through code anyway.

I'm actually porting the debugger so it will work on Linux and Windows as well (and be written in C and not ASM), and I couldn't help but notice that the effective address calculation code in the old debugger is really messed up, especially after talking to byuu about it. So, I wanted to see if anyone would care if I just left it out (pagefault didn't, and so far he seems to be my only user).

I have to admit that, at the moment, the debugger does seem to be pretty useless for much more than single stepping (you *can* set breakpoints in both 65816 and SPC code, sort of. Though really it is more of a "run to here" command than a breakpoint). I hope to fix this after ditching the DOS code, so that you can set multiple break conditions of various types, maybe even adding the ability to break on access to a given address, though I wouldn't count on it. (Though I suppose it probably isn't that much more difficult than calculating effective addresses properly.)

What I have so far is available from SVN (if you want to build with the old DOS debugger, you can do "make -f makefile.ms PLATFORM=dos DEBUG=yes OLDDEBUGGER=yes", though you will need the "unix curses emulator" from the DJGPP zip picker either way. note that the old debugger could be disappearing any time now...)

I wonder if anyone but pagefault will care?

[Nightcrawler]

If the idea is to bring the ZSNES debugger up to speed to be a viable alternative to use, then I'd say yes, people will care. If you're only enhancing it for the ZSNES team to use, then probably not many will as they will continue to use Geiger's or BSNES debuggers which are far beyond the ZSNES Debugger.

To expand on why effective addresses are useful to ROM hackers, that's basically what we do. Take some known RAM or ROM locations and reverse engineer how the data got there and/or what's being done with it. I'm constantly scanning for addresses. I want to know when something was stored in 7e:2345 whether it was done via a sta $7e2345 instruction or a sta $22 instruction which relies on the databank and page registers.

Geiger's and BSNES both handle this very well. So, I'd say it depends on what your goal is with the debugger. If you only want to target ZSNES devs, that's one thing, but if you intend on trying to enhance the ZSNES debugger to be useful and worthwhile again, I'd say you'd have to make something with features comparable with the other debuggers I mentioned before people would switch to ZSNES as an alternative.

Effective addressing is a must if you're going in that direction in my opinion.

[creaothceann]

Btw. the debugger wouldn't use the current GUI, right? So it would have to create additional windows to show all that data. Are you going to use SDL for that, or platform-specific code (WinApi etc.)? Just curious.

</offtopic>

[grinvader]

SDL port uses curses, win/dos just uses a terminal... the code is in C, so anyone more experienced on multiple-windows projects can try to help and add onto it so it turns into a debugger worthy of use.

[creaothceann]

Ah, I see. Thanks.

[SamB]

SDL port uses curses, win/dos just uses a terminal... the code is in C, so anyone more experienced on multiple-windows projects can try to help and add onto it so it turns into a debugger worthy of use.

Um, all ports use curses! The windows code has to allocate a console, the dos code has to set the video mode to the normal text mode, and the linux port does nothing special at the moment, but they all use curses.

[byuu]

I don't use ZSNES' debugger -- 100% -- [ 10 ]

Heh, ouch. I use it on occasion. I've learned to make it do what I want over time.

I'm sorry the effective address decoding doesn't work correctly in ZSNES. I didn't really know all that much about the SNES back then.

[IcecreamzIII]

I don't use ZSNES' debugger -- 100% -- [ 11 ]

[xamenus]

I hope those results aren't misleading; hopefully no non-ROM hackers voted.

This should probably be posted at romhacking.net if it isn't already.