The DB translation is almost done, the script itself is in the editing stages.

However, there is an expansion-related bug that is causing me grief. Once you get to the second boss (the Hippogryph) and bite him to start the battle, you're not able to open the inventory menu from that point on (making it so you cannot finish the game).

How do I know it's expansion related? I checked traces between the expanded and hacked version, and a plain expanded version. No differences. So I played the Japanese game with no hacks, then the Japanese expanded version. It only happens once you expand the game.

I asked around my ASM gurus first: RedComet, DaMarsMan and Gemini. DMM and Gemini said they had never heard of such a bug. RC was able to fix the first expansion-related bug (Arma was invincible, he somehow fixed that), but the second bug itself stumped him.


DMM also suggested to ditch the expansion and go with DTE. I do eventually need to do a DTE, but I'd rather not since I've already got everything figured out for the expansion/pointers/etc in this case. I'm hoping that one of you mad scientists of the 65816 can understand what is going on, or have encountered something like this before.

Any help would be greatly appreciated.

~DS

At start up an entire bank of ROM is loaded auto-magically into RAM. Later at certain points through out the game the data in ROM is compared to the same data in RAM. For some reason, when the ROM is expanded the ROM bank isn't copied into RAM at start up. That means when the aforementioned comparisons are performed they fail and cause the game to do all sorts of weirdness. I disabled one of these comparisons to fix the Arma invincibility bug that DS mentioned.

EDIT: MathOnNapkins, just expand the ROM to 2.5MB and check out the data in RAM at $300000 or so.

I don't understand why it's doing this. Can't we just keep the bank and everything the same and use 24bit addressing?

S'alright. Yeah, there's two or three other Hippogryphs. I'm just referring to the first one ("the second boss...").

~DS

Hey Nightcrawler...Why would a rom be accessing the expanded space? The only thing I could think of is if he changed the DB right?

Well if a game is copying the first bank why would it copy the other instead? Does it overwrite the first one?

If I check out the value at $40FFC1 in the Japanese ROM, it's $65. If I check that same address after I've expanded the ROM, it's $00, though $80FFC1 is still $65. Hope that clarifies the problem a bit better.

Nightcrawler shoots and scores! 

Exactly. That's because $40:FFC1 is mapped to ROM, not RAM. It's doing exactly what I suspected. It's checking mirrored locations. According to my understanding of the the LoROM mapping model for Mode20 ROMs, bank $40 will shadow $C0 or default to the A bus. In the original ROM, there's no ROM in the traditional sense to map there because it's too small. The cart doesn't respond, and it in turn just accesses the A bus which is the equivalent of 00:FFC1 OR $80:FFC1 (And other locations, but we won't get into that).

Now when you expand the ROM, you DO have enough ROM to map to bank $40/$C0 and the cart (theoretical new cart) would respond. NOW.. $40:FFC1 is actually the same as address $C0:FFC1. Anomie could probably explain it more elegantly, but you get the picture.

Understand what's happening here? The game code as is simply will not work with an expanded ROM. I'd guess this check was either put in as some sort of copier protection (I'd imagine some copiers probably wouldn't have the correct addressing either) or hack protection (but I doubt that was an issue back then).

You need to either hack those instructions or work within the confines of the original ROM. Alternatively, you can 'trick' it by sticking the '65' in the expanded spot $C0:FFC1. That's all there is to it.