> Be careful when comparing only the integer portion of a fixed point number
> because the fractional portion is considered unsigned.
>
> int range
> 2 2.000 to 2.996
> 1 1.000 to 1.996
> 0 0.000 to 0.996
> -1 -0.004 to -1.000
> -2 -1.004 to -2.000
>
> So if you clip the integer value to +/- 2 your positive range will be higher
> than your negative range.
Here's the code I came up with too make sure the abs(speed) wasn't
too big:
BALL_MAX_POS_SPEED_HI = #1
BALL_MAX_NEG_SPEED_HI = #-1
;making sure ball speed isnt to extreme (abs(speed) not too big)
LDA slowBallYSpeed+1
CMP #BALL_MAX_POS_SPEED_HI
BMI BallSpeedLessThanMax
LDA #BALL_MAX_POS_SPEED_HI
BallSpeedLessThanMax
CMP #BALL_MAX_NEG_SPEED_HI
BPL BallSpeedGreaterThanMin
LDA #BALL_MAX_NEG_SPEED_HI
BallSpeedGreaterThanMin
STA slowBallYSpeed+1
So basically I'm just looking at the high byte...
against positive values, I can see the range is
0 < val <= 0000000111111111
but by just looking at the high byte for negative
values, how much more or less negative can it be
than positive? If its just a few bits I don't care,
but I think it might be just under a whole byte worth--
an assymetry I don't like
Maybe I should just do what I came up with for
making sure abs(speed) wasn't too small...at my
friend Ranjit's suggestion, if the speed is negative,
I take it's negative to get a positive, and then do
that compare (I use Y register as a boolean about whether
the original value was negative, and then, if it's "too small",
set it to the correct a bit small postive or a bit small
negative value::
;comparison to make sure speed
;isn't too slow; it's much easier to
;do comparisons w/ positive values,
;and also we know it's a small positive
;value, so once it's positive, if the
;hi byte is not zero, it MUST be big enough
;so we make a copy into temp, and then
;change the original if we see the guaranteed
;positive version is too small
LDA slowBallYSpeed
STA tempVar
LDA slowBallYSpeed+1
STA tempVar+1
LDY #0 ; Y is boolean, 0 = postive
LDA tempVar+1
BPL ThisAintNegativeSpeed
LDY #1 ;remember that this is negative
;negate what's in temp
sec
lda #0
sbc tempVar
sta tempVar
lda #0
sbc tempVar+1
sta tempVar+1
ThisAintNegativeSpeed
;we know tempvar is 16 bit, positive value
LDA tempVar+1
BNE DoneSeeingBallSpeedIsBigEnough
LDA tempVar
CMP #BALL_MIN_POS_SPEED_LO
bcs DoneSeeingBallSpeedIsBigEnough
tya
BEQ ChangeBallSpeedToMinPositive
ChangeBallSpeedToMinNegative
LDA #BALL_MIN_NEG_SPEED_LO
sta slowBallYSpeed
LDA #BALL_MIN_NEG_SPEED_HI
sta slowBallYSpeed+1
JMP DoneSeeingBallSpeedIsBigEnough
ChangeBallSpeedToMinPositive
LDA #BALL_MIN_POS_SPEED_LO
sta slowBallYSpeed
LDA #BALL_MIN_POS_SPEED_HI
sta slowBallYSpeed+1
DoneSeeingBallSpeedIsBigEnough
It's a little brutal, and uses two bytes of tempVar,
but still, at least even *I* can understand it :-)
--
KirkIsrael@xxxxxxxxxxxxx http://kisrael.com
"Death smiles at us all. All a man can do is smile back." --Gladiator
----------------------------------------------------------------------------------------------
Archives (includes files) at http://www.biglist.com/lists/stella/archives/
Unsub & more at http://www.biglist.com/lists/stella/
