NESASM inefficiency when referencing data by ptr-to-ptr?

[pops]

While debugging my implementation of shiru's new famitone library, I noticed that NESASM assembles this:

Code: Select all

	FT_TEMP				= $00	;3 bytes in zeropage used by the library as a scratchpad
	FT_TEMP_PTR			= FT_TEMP		;word
	FT_TEMP_PTR_L		= FT_TEMP_PTR+0
	FT_TEMP_PTR_H		= FT_TEMP_PTR+1

	lda FT_TEMP_PTR_L
	pha
	lda FT_TEMP_PTR_H
	pha

Into this:

Code: Select all

; note lda absolute where lda zp would suffice
	lda $0000
	pha
	lda $0001
	pha

I'm not sure if NESASM is still under development - hopefully someone knows who to forward this to - thanks!

[tepples]

You have to explicitly specify zero page mode in NESASM.

[zzo38]

Yes, that is correct, you have to explicitly specify zero-page mode by putting < before the operand. (This is the method I prefer over other assemblers that don't do it that way.)