Properly handle unimplemented opcodes?

[st3fan]

We resolved #105 (ProDOS 2.4.1 crashing because it tried to use an unimplemented 6502 instruction for CPU detection) by simply aliasing all unimplemented instructions to nop.

But is that actually the right thing to do? Did Apple II programs rely on these instructions?

[paulhagstrom]

Some early protection schemes took advantage of undocumented opcodes. Not so much for function, I don't think, but most for "arity" (some undocumented opcodes take 1- or 2-byte arguments). So, later when the 65C02 defined some of them, the older programs would no longer boot. See, for example, 4am's writeup of Beer Run, specifically chapter 6 of the write-up txt file: https://archive.org/details/BeerRun4amCrack