nesdev.com

I guess the error is called "missing operator" (since tiles[0] could be considered an Lvalue)... and a good compiler should give this warning.

"warning: statement with no effect"

How can you have an expression without an Lvalue?
These should throw an error. Or a warning. If they don't, this is a bug in cc65 that needs fixed.

This should work

Is "i" declared as unsigned char?

It should be able to optimize for an array of 256 bytes of less.

Or, maybe the problem is that it is an array of ints. Yes. I think it needs to be an array of chars. I don't remember. I'm also away from my keyboard.

I know this will sound stupid, but cc65 can't handle 2 arrays on the same line of code. There are workarounds, but if you aren't having slow down issues, I would just forget about it. (example of a stupid workaround) temp is a 16 bit zeropage variable... temp = playery[i]; oldy[i] = temp; another ex...

Thanks brizzo. Cool.

There are between 1 and 4 BG layers, depending on which mode you are in.

By 24 bytes, I assume you mean 3 bits per pixel (bpp), like Super Mario World uses. That's a form of software compression. Those are expanded to 4 bpp (with the upper 8 color being blank). All SNES sprites are 4 bpp (16 color).

NSF format doesn't really conform to any real NES mapper. It's sort of a mixed bag of all possible sound systems. so... I'm going to say that it can't be done.

I think you would need to redo it in Famitracker, and export it as .bin

With cc65, lower your expectations. I use fairly optimal C code, and it maxes out at around 8 objects on screen at once. x4 sprites per object = 32 sprites total.

(with music and collision code, and movement logic for each object).

Lots of things could be happening. JSR to a subroutine that changes the status flag to 8 bit. NMI or IRQ code that changes the status flag to 8 bit.

I believe there is a SEP somewhere causing the problem. You will have to step through the code in a debugger to find it or do a trace log.

If you are using NESASM, you need to indent assembler directives. I think 2-3 spaces is enough, goes before the .incbin

labels, on the other hand need no indenting.

But not reliably random that you could use initial RAM values as random numbers. It might be all zeroes. It might be all FF.

The emulators that fill RAM with random numbers is just as misleading as emulators that fill with zero.

Tiles are 8x8 pixels. The screen is 256x240 pixels. That means 32 tiles wide and 30 tiles tall. A formula to calculate a PPU address, in tiles position is... $2000 + (tile Y * 32) + tile X where tile X is 0 to 31, and tile Y is 0 to 29. You take that PPU address, send it High byte, then Low byte to ...

Looking at famitone2.s Everything between FamitoneUpdate: and .if (FT_SFX_ENABLE) ;process all sound effects can go in the bank with the music code. Slap a RTS there. You need to include _FTChannelUpdate and _FT2SetInstrument and the note table. And the DPCM functions.. (have the word "sample" in th...

The music update can be split into 2 parts, the song and the sfx. The song code only writes to a buffer, then it checks each sfx channel, then it copies the buffer to the APU registers. You can put the song part in one bank and the sfx part on another bank, and call them sequentially. By the way, if...