[TI-ASM/hardware?] That silly key port

[Jim e]

You should definitely reset that initially. Also I find it odd that chose to go in reverse of generally convention. Getcsc starts with the arrow group.

[driesguldolf]

Thanx, I'll reset it initially (strange it has always worked...),

I'm not a guy who thinks alot about that , I have my own conventions
The reason I do things backward is that it allows easy finish detection:

Code: Select all

	rrc c
	jr c, _grouploop

If I would do things "normal" I needed more instructions: (and a different initialization ofcourse)

Code: Select all

	rlc c
	bit 7, c ; or use a compare instr cp %01111111
	jr nz, _grouploop

Ok, because I use this little optimization I do need my own equates tough, but I've no problem with that...

[King Harold]

bit 7, c ; or use a compare instr cp %01111111

that is not the same though, bit 7 also works when other keys are pressed as well, the CP only works if exactly that combination of keys is pressed.

[Jim e]

bit 7, c ; or use a compare instr cp %01111111

that is not the same though, bit 7 also works when other keys are pressed as well, the CP only works if exactly that combination of keys is pressed.

He's talking about groups not keys.


rlc c sets the sign flag, so you can use jp m, to test bit 7.

[driesguldolf]

bit 7, c ; or use a compare instr cp %01111111

that is not the same though, bit 7 also works when other keys are pressed as well, the CP only works if exactly that combination of keys is pressed.

You're confusing things here:
I was talking about the group not the key
My routine (previous page) starts at group %10111111 and rotates right:

Code: Select all

;I check groups in this sequence:
%10111111	; 1st
%11011111	; 2nd
%11101111	; 3rd
%11110111	; 4th
%11111011	; 5th
%11111101	; 6th
%11111110	; 7th
;When I reached the final group (7th here) one more rrc c will reset the carry
;When I do things the other way around (7th to 1st) I need an additional check because there's no group %01111111

[King Harold]

You're confusing things here:
I was talking about the group not the key

oh.

so why not use the sign flag like Jim E said?

[driesguldolf]

rlc c sets the sign flag, so you can use jp m, to test bit 7

Hmmm... tought I checked that... Checks again... Oh must have overlooked it

Code: Select all

jr c, label ; 2 bytes, 7/12 clocks
jp m, label ; 3 bytes, 10 clocks

and it's even faster, thanx