Another Question

[Appelhorne]

Hey i have a question about DJNZ

this code:

ld b, 8
label:
[...]
djnz label

how many times is this executed?

is it 8 times with
b = 8
b = 7
b = 6
b = 5
b = 4
b = 3
b = 2
b = 1

or nine times?:
b = 8
b = 7
b = 6
b = 5
b = 4
b = 3
b = 2
b = 1
b = 0

Thanks in forward for the help!

[sic]

8 times.

[Justin W.]

Yes it is executed 8 times, the reason being this. When the 8 bit register a,b,c,d,e,h,or l becomes 0, it sets the zero flag.

djnz simply decrements the b register, then performs a jr nz,Label

If the value in b has not become 0 it jumps to the label defined by the djnz instruction.

jr nz = jr "not zero" . Also can be thought of as zero flag reset.

Therefore if you have a loop such as this.

ld b,2
Label:
halt
djnz Label

One the first cycle through the loop, b is decremented once leaving 1 in the register. Then because it is not zero it jumps back to "Label". Executes the loop code again but this time when it reaches the djnz, it decrements b, but this time it will set the zero flag because b has become zero. Thus it skips the jr nz part of djnz, and the loop is complete.

[Darth Android]

a djnz expands to something like:
ld c,8
label:
[...]
dec c
ld a,c
or a
jr nz,label

[Appelhorne]

I wanted to rewrite my astroidsbelt. I first had for each astroid separate code, like this:

;Ast1
ld b, 8
ld a, (bAst1Y)
ld l, a
ld a, (bAst1X)
ld ix, sprRock
call isprite

;Ast2
ld b, 8
ld a, (bAst2Y)
ld l, a
ld a, (bAst2X)
ld ix, sprRock
call isprite

;Ast3
ld b, 8
ld a, (bAst3Y)
ld l, a
ld a, (bAst3X)
ld ix, sprRock
call isprite

etc etc

I wanted to improve this and make it all in one routine so i rewrote astroidsbelt. But it doesn't seem to work. This is my code:

#include ti83plus.inc
#include mirage.inc

.org 9D93h
.db $BB, $6D ;compiled asmprgm token

ret
.db 1 ;MirageOS token

;MirageOS Icon
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000
.db %00000000, %00000000

szDiscription:
.db "Door Jasper Smit ",0

:Astroids X's
bAstX = SafeRam1

;Astroids Y's
bAstY = bAstX + 8

;Astroid type
bAstT = bAstY + 8


xor a
ld (bAstX), a
ld (bAstX + 1), a
ld (bAstX + 2), a
ld (bAstX + 3), a
ld (bAstX + 4), a
ld (bAstX + 5), a
ld (bAstX + 6), a
ld (bAstX + 7), a
ld (bAstY), a
ld (bAstY + 1), a
ld (bAstY + 2), a
ld (bAstY + 3), a
ld (bAstY + 4), a
ld (bAstY + 5), a
ld (bAstY + 6), a
ld (bAstY + 7), a


blit:
;bcall(_grbufclr)
ld b, 8

blitloop:
push bc

blitast:
;Get astroid type
ld hl, bAstT - 1
ld c, 0
add hl, bc
ld a, (hl)

cp 0
jr Z, blitsmall
cp 1
jr Z, blitsmall
jr blitoncomming

blitsmall:
ld ix, sprSmall
jr blitnow

blitlarge:
ld ix, sprRock
jr blitnow

blitoncomming:
ld ix, sprShip

blitnow:
ld hl, bAstX - 1
add hl, bc
ld a, (hl)

ld hl, bAstY - 1
add hl, bc
ld l, (hl)

ld b, 8
;call isprite

pop bc
djnz blitloop

flip:
;call ifastcopy

CheckAst:
ld b, 8

CheckLoop:
ld c, 0
ld hl, bAstY - 1
add hl, bc
ld a, (hl)
inc a
ld (hl), a
cp 56
jr NZ, SkipReset

;Reset astroid
xor a
ld (hl), a
ld hl, bAstX - 1
add hl, bc
push bc
push hl
ld b, 87
call irandom
pop hl
pop bc
ld (hl), a

SkipReset:
djnz CheckLoop

;Loop again
jr blit

sprRock:
.db %00011000
.db %00111100
.db %01111110
.db %11111111
.db %11111111
.db %01111110
.db %00111100
.db %00011000

sprSmall:
.db %00000000
.db %00000000
.db %00010000
.db %00111000
.db %01111100
.db %01111100
.db %00111000
.db %00010000

sprShip:
.db %00011000
.db %00011000
.db %00111100
.db %00011000
.db %01011010
.db %11111111
.db %11100111
.db %11000011

.END

I read through the code for about 100 times, but i can't see anything wrong. I thought the problem was maby in the djnz, but it should work this way. Do you know what is wrong??

[Justin W.]

[Darth Android]

well, the comment after szDescription (Asteroid X's) had a : (colon) in front of it instead of a ; (semi-colon)

[sgm]

[Justin W.]

Yes you are right about the flags being preserved sigma. I was explaining how it works relatively. I was simply giving a different perspective on it.