Hi Glenn!
> >since the skipdraw branch
> >is way faster than the drawing branch. In fact you'll
> >probably even have to add NOPs or .Ws in there, to match the
> >running times of both branches.
> A problem I see is if you clear out GRPx within Skipdraw that you can't have
> the successful branch that stores graphics to GRP0 just naturally run
> through Skipdraw as it does in Thomas' example because the data that gets
> written to it will then get wiped out. So you have to add a JMP right after
> the data gets written to GRPx that skips over SkipDraw. This adds cycles to
> the LONG branch of the routine.
> Depending on the timing of the kernel that may be make or break.
Here's an example straight out of gunfight:
The famous skipdraw branch:
SkipDraw
SEC ; Same carry state after both branches
NOP ; Wasting...
LDA #$00 ; load for GRP0
STA.w COLUP0 ; Special trick, blacks the bullets
BEQ Continue ; Return...
The kernel:
TXA ; A-> Current scannline
SBC verPosP0 ;
ADC #$0B ; calc if sprite is drawn
BCC SkipDraw ; To skip or not to skip?
TAY ; not necessary when Y holds scannline
LDA (colorShapePtr00),Y;
STA COLUP0 ; Select cowboy color
LDA playerShape00,Y ; Select Cowboy shape
Continue:
CPX verPosM1 ; Do missiles
PHP ;
CPX verPosM0 ;
PHP ;
STA GRP0 ; Execute Write here!
So, you see that no matter which branch is taken,
I'm exactly on the same cycle when hitting
'Continue' - and I'm executing STA GRP0 only once,
with either the cowboy shape or '00'.
Hope all of this makes sense now...
Greetings,
Manuel
-----------------------------------------------------------------------
Bei unserem Shoppingpartner waltherliving erhalten Sie einen Gutschein im
Wert von DM 10,00 für Ihren nächsten Onlineeinkauf! Einfach hier abrufen
http://www.nexgo.de/c/pub/redir.php/living1
-----------------------------------------------------------------------
-
Archives (includes files) at http://www.biglist.com/lists/stella/archives/
Unsub & more at http://www.biglist.com/lists/stella/
