What's new
By:

Is there a 3rd Strike ROM from 30th Anniversary around? (Fatal bug fixed)

MetalliC said:
plasia said:
Bad news guys, this wont happen right away.
Click to expand...
errm, wtf ?I've extracted and checked bundleStreetFighterIII_3rdStrike.mbundle files by myself - they are exact match "sfiii3nr1" set from MAME (SF3 3rd Japan 990512, NO CD), no single byte of difference found.

btw, it is even not latest available 990608 version. or maybe that is the magic ? (e.g. latest version had introduced some critical bugs ?)
Click to expand...
??

Firstly they don't match, at least without splitting it.

Secondly, as original post says:
Arlieth said:
For those of you who don't know, the 30th Anniversary release of 3rd Strike keeps almost every glitch intact, except for some cosmetic ones such as "Twelelve" and 2P Yang having the wrong Super Art name (Magnetic Storm), but the one critical glitch that it fixes is Makoto getting KO'ed by Ken with a neutral throw. This often results in a hard crash of the game and necessitates a power cycle.

Does this version of the 3rd Strike ROM exist outside of 30th Anniversary? Would it be possible to load it in Darksoft?
Click to expand...
How did these fixes happen then? I'm 99.9% there is no ver A without Makoto glitch.

Either you're right and these things are patched in the emulator.
Or you're wrong and this is a patched rom.
Or you have a different version of the game.
 
Aurich said:
xodaraP said:
That's version A which is the tournament standard and what everyone wanted/expected in the collection. It has bugs but they're the bugs everyone wants there because tournaments?

Someone earlier was saying it actually had some game breaking bugs fixed which is why they want to run it on original hardware.
Click to expand...
To summarize: version B fixed some bugs, but it also took out some things (specifically a few unblockable setups for a couple characters). The unblockables being removed was not a popular decision for a couple reasons, one they were fun, and two, the characters with them needed those setups, it made them more viable and balanced in the cast to have the bullshit so they could keep up with other stronger characters. So players rejected B, and stuck with A as the standard.
It's been that way for two decades now, so it's long been a defacto standard with players.
Click to expand...
it also fixes Chuns c.mk late cancel bug( yes it's a bug), which severely nerfs her as well and further makes the balance of the game worse
 
The whole 30th Anniversary set is emulated, no worries there :)

Everything has already been dumped it's just a matter of working out how it's been compiled for the emulator

Edit: My mistake. Alpha Anthology is a different thing. Interesting they ported rather than emulated. What platform? Was it not powerful enough to run the emulator?
 
Hmm I dunno if I buy the idea that alpha anthology is a port. The fact that you can load every rom set ever available on alpha anthology, makes me think its emulation
 
plasia said:
Firstly they don't match, at least without splitting it.
Click to expand...
sure, there was required to join MAME dumps (or split files from Anniversary), but huh, its not a rocket science.


plasia said:
How did these fixes happen then? I'm 99.9% there is no ver A without Makoto glitch.

Either you're right and these things are patched in the emulator.
Click to expand...
this^^^
also possible this glitch happen on real hardware only (exception due to unaligned 16/32bit memory access or smth like that), but not in emulation. may anyone check and deny or confirm ?
add: glitch seems happens in MAME/etc as well, so I'd guess the emulator uses some runtime patches to fix the bug(s).
 
Last edited:
So I checked 990512 on my Darksoft setup and crashed the game on my second try with Ken killing Makoto with neutral throw.

I also checked 30th Anniversary Collection's 3rd Strike and checked two things:

* 2P Yang's SAII reading as Magnetic Storm (this is fixed)
* Makoto killed by Ken's neutral throw crashes the game (tried 8 times, never triggered any glitch or crash at all)

If the ROMs is actually 990512 then I have *no* idea how they managed to fix it. You might be right about runtime patches. Hmm.
 
Runtime would make sense. They would have the original source code and it means they don't need to recompile the whole ROM.

Especially when it's well known exactly how to trigger the fault and fixing it in the source may break something else.

It may be possible to run a debugger on the PC version and try to work out what is injected and where, but not easy.
 
plasia said:
Bad news guys, this wont happen right away.

According to Darksoft:
  • Bios is original, untouched
  • Program code is encrypted with some tough stuff, the decryption must be done in the emulator.
  • Until it's decrypted we wont be able to run on original hardware

We have confirmed this is a rom and once decrypted it should work on original hardware. The problem is decrypting the roms.


Plasia's hot conjecture: Next thing I would try is to try and find the key in the program and decrypt from that. Hey we can only hope for a simple solution right? Unfortunately this low level* stuff is not something I can do.

* Low level in computer science is referring to there being a low level of abstraction, so it's harder basically.
Click to expand...
I take that back. It's exactly as @MetalliC said.

I was looking at the ISO image. The content matches exactly the sfiii3nr1 in MAME. So whatever was fixed was done at emulator level.
 
We already know it's not that simple, because they fixed some other very minor stuff like the super art name being wrong. So clearly there's code intervention happening.
 
Darksoft said:
jassin000 said:
Or is simply a non-issue... because the crash only happens on real hardware
Click to expand...
Yes, could be that as well. In that case, the bug shouldn't happen on MAME either.
Click to expand...
Not necessarily as @Aurich says above, but even without that point there is no way to know if MAME is emulating the PCB accurately (which then would have the crash) vice whatever in house emulator the company came up with.

Maybe their own home grown emulator fixes these things?
 
Aurich said:
We already know it's not that simple, because they fixed some other very minor stuff like the super art name being wrong. So clearly there's code intervention happening.
Click to expand...
xtrasmiley said:
Darksoft said:
jassin000 said:
Or is simply a non-issue... because the crash only happens on real hardware
Click to expand...
Yes, could be that as well. In that case, the bug shouldn't happen on MAME either.
Click to expand...
Not necessarily as @Aurich says above, but even without that point there is no way to know if MAME is emulating the PCB accurately (which then would have the crash) vice whatever in house emulator the company came up with.
Maybe their own home grown emulator fixes these things?
Click to expand...
Both of these are possible. A bug that's not emulated in some portion of the hardware (because it's emulation so they may have fixed it without even realizing it).

Or that they patch it in runtime because they know exactly how it's triggered so that would be easier than modifying the original ROM and potentially breaking other stuff.

Remember too that they don't want to change the bugs that people want in version A. Only the bug that causes a crash and the text bug. On the fly code injection solves that problem.
 
That's really the only bug people have any interest in seeing fixed. The incorrect text for the super art @Aurich mentioned is the only other thing that would be good

The problem is, by patching either the original code is modified and it's banned from tournament use
 
xodaraP said:
The problem is, by patching either the original code is modified and it's banned from tournament use
Click to expand...
Maybe. A phone call or text could probably solve that for the US Jazzy Circuit, which is responsible for the uptick in interest in Third Strike again here. If nothing else was touched I doubt it would be a problem to get it cleared for use.
 
It would be sweet if Alpha 3 Upper was able to run on CPS2 hardware, it would be the perfect game. I'm pretty sure I remember there being dip switch settings in the AA version where you could get Upper to retain some of A3's 'unique' characteristics. Would that not imply it was ported rather than emulated?
 
Aurich said:
xodaraP said:
The problem is, by patching either the original code is modified and it's banned from tournament use
Click to expand...
Maybe. A phone call or text could probably solve that for the US Jazzy Circuit, which is responsible for the uptick in interest in Third Strike again here. If nothing else was touched I doubt it would be a problem to get it cleared for use.
Click to expand...
This is an easy sell for Jazzy Circuit since we also allow 30th Anniversary to be run for qualifiers anyways.

Also, if this was a modified ROM available to be played via Darksoft/SuperBIOS then I actually, honestly, would be able to make it Japanese tournament standard via approval through Yasuaki Matsuda.
 
You must log in or register to reply here.
Back
Top