Lousy 128K program crashing on 64K machine sucks.

The following routine, aimed to be short and fast, should detect a regular 128K RAM machine. For more informations about extended RAM management in the CPC/Plus, read the MMR register description.

			; Usage exemple
			;     call quick_128k_detection
			;     jr z,CPC_64K
 
			; Quick 128K Detection routine
			;
			; The test will corrupt two bytes in the extended RAM bank 0/page 0
			; when available. This should not be a problem since this routine
			; should be called before actually using any of this extended RAM.
			;
			; Output
			;	- Zero flag is set on a 64K RAM machine
			;	- All extended RAM banks (if available) will be paged out.
quick_128k_detection:
			; Initialize some CPU registers
			ld bc,&7F00
			ld hl,&C4C0
			; Make sure the possible extended RAM banks are paged out
			out (c),l
			; Preserve the two bytes in the base64K RAM we are going to modify
			ld de,(&4000)
			; Write a two bytes marker (&7F00) in the base 64K RAM at &4000
			ld (&4000),bc
			; Try to page in extended RAM bank 0/page 0 from &4000 to &7FFF
			out (c),h
			; Write another two bytes marker (&7EFF) in the extended RAM at &4000
			dec bc
			ld (&4000),bc
			; Page out extended RAM (if available)
			out (c),l
			; Read two bytes at &4000 from the base64K RAM
			ld hl,(&4000)
			; Compare them with the two bytes marker of the extended RAM
			or a
			sbc hl,bc
			; Restore the two modified bytes and exit
			ld (&4000),de
			ret