As for the add and adc, there seem to be quite a few programmers (not here, of course) who misused the add instruction.  For example, the code for the Pac Man machine (yes, it was a z80 too.) has a lot of "ccf \ add b".  They still have no excuse.

Actually, you'll notice that instead of doing the stupid subtraction with B, they should have just done CP 9 since they restore A

I know nothing about ASM, and yet I still feel that faint sense of revulsion.

It is almost scary how good some of the programmers are, here. I think that if they all designed an OS, it would be ridiculously amazing and more stable (though the TI-OS is pretty stable).

Agreed.

For most z80 programmers, or a is very readable, as is xor a. Assemblers shouldn't automatically optimize because sometimes you need to use "unoptimized" code on purpose (such as ld a,0 to preserve flags or to provide a byte for SMC)

The problem is that it's not obfuscated

Quote from: tangrs on September 30, 2011, 05:14:07 am

Maybe it's deliberately obfuscated code?

Perhaps, but...
Actually, it seems just like TI to try and "protect" their code with something this stupid and heinous, but they don't seem smart enough to deliberately write bad code. I think they just produce things that make you give up all hope unintentionally. Or....

or make something called TI-NOspine that costs $230 that 'protects' it from all hardware hacks and development

Well, if you're writing on the asm level optimizations require a certain amount of interpretation, which an assembler is likely to miss, but a programmer should understand fairly easily. After all, they wrote the initial code.

*Auto-Optimizing C is simpler, since the commands arbitrate the registers and other hardware concerns into the stack.