Basic Questions and Answers Thread

$1C is not an odd address.


@Deoxys: You're using bchg, use move instead.

 

$1C byte is animation number, but you need just change the frame depending on boss' health, right? Then it would be better to use $1A for this purpose.


Another weird thing about your code is using bchg instruction. It changes the bit, 0 to 1 and 1 to 0. Your routine is running every frame, so, this way everything will be quite messy.


Let's imagine how your code will work:


Frame #0, Health #8 -> bchg #0 -> $1C(a0) = %00000001 ($01)


Frame #1, Health #8 -> bchg #0 -> $1C(a0) = %00000000 ($00)


Frame #2, Health #8 -> bchg #0 -> $1C(a0) = %00000001 ($01)


(let's say Sonic has hit the boss!)


Frame #3, Health #7 -> bchg #1 -> $1C(a0) = %00000011 ($03)


Frame #4, Health #7 -> bchg #1 -> $1C(a0) = %00000001 ($01)


ect.


As you can see, you'll get absolutely random values due to bchg instruction.


And the last problem, what if you want to change boss' health to 16 for example? =P


So it would be better to write something like this:



move.b $21(a4),$1A(a0)
jmp DisplaySprite



That's it! =) So if the the health is 8, sprite will use frame #8, etc.


EDIT: Grammar

 

"lea ($FFFFD010).w,a4" ? D010 is $10(an) when (an) is Sonic's main address, so you're checking $31(an) which is Sonic's remaining invulnerability time in s1.

 

It actually was purposely to load object adresses.


ALSO: Vladik, that didn't work

 

cmpi.w #8,$21(a4)

Code:

				cmpi.w  #7,$21(a4)

Two "odd" addresses while you compare a "word", it must be compared using a byte, or using an even address.

 

Hm, what way this didn't work?


If the object is not even displayed, check out if one can be displayed correctly, i.e. if mappings, art pointers, coordinates are loaded and mapping frames 0 to 8 are correct.


If you have wrong frame displayed, check out if $21(a4) contains what it supposed to contain. I can't understand how your "lea ($FFFFD010).w,a4" works. It actually loads byte 10 of Sonic's STT, so how can $21(a4) contain boss' health counter?

 

After the label of the boss's objects (before any index), add this



move.b $21(aX),($FFFFFF99).w



(Obj3D for example uses a0, if you understand it)


Define a object ID, for this. I defined obj04 to this example:



; ---------------------------------------------------------------------------
; Object 04 - Boss Hits counter


; ---------------------------------------------------------------------------


Obj04: ; XREF: Obj_Index


moveq #0,d0 ; Borrar d0


move.b $24(a0),d0 ; Preparar index


move.w Obj04_Index(pc,d0.w),d1 ; Index


jmp Obj04_Index(pc,d1.w) ; Index


; ===========================================================================


Obj04_Index: dc.w Obj04_Main-Obj04_Index


dc.w Obj04_First-Obj04_Index


; ===========================================================================


Obj04_Main: ; XREF: Obj04_Index


addq.b #2,$24(a0) ; Terminar el resto de los codigos hasta rts y avanzar de rutina


move.l #Map_obj04,4(a0) ; Cargar mapas


move.w #$3A0,2(a0) ; Ubicacion en VRAM


move.b #0,1(a0)


move.b #1,$18(a0)


move.w #$90,8(a0)


move.w #$108,$A(a0)


Obj04_Display:


jmp DisplaySprite


rts


Obj04_First:


move.b ($FFFFFF99).w,$1A(a0)


jsr Obj04_Display


rts


; ===========================================================================


Map_obj04:


include "_maps\obj04.asm"



Note: you must define your VRAM address to make it work correctly.


I used $3A0 for this example.


(and yes, I'm sharing this code. Anyone can use this, but if credit me)


Good luck...

 

And I'm telling you you aren't loading the object's address.


D000 to D03F are Sonic's status table in s1. Object addresses in s1 are D000+40*n where n is a natural number, B000+40*n in s2, and B000+4A*n in s3 so you can't have an object starting in D010.

 

Do'l! I was gonna do that, *steals it, and runs*. Jokes aside, I echoe what was said, although not as knowledgeable, I know that ram adress aint gonna work.

 

Especially if the boss object is loaded using singleobjload (and it should be).


For example, in GHz, the boss loads through this:



loc_6EB0:
cmpi.w #$2960,($FFFFF700).w


bcs.s locret_6EE8


bsr.w SingleObjLoad ; << this finds an empty object slot in the main SST space (which goes from FFFFD800 to FFFFEFFF)


bne.s loc_6ED0


move.b #$3D,0(a1) ; << this loads GHZ boss object in the free address found by singleobjload


move.w #$2A60,8(a1)


move.w #$280,$C(a1)



So, an example to load another object AND save the boss' address would be:




Loc_6EB0:


cmpi.w #$2960,($FFFFF700).w


bcs.s locret_6EE8


bsr.w SingleObjLoad ; << this finds an empty object slot in the main SST space (which goes from FFFFD800 to FFFFEFFF)


bne.s loc_6ED0


move.b #$3D,0(a1) ; << this loads GHZ boss object ID in the free address found by singleobjload


move.w #$2A60,8(a1)


move.w #$280,$C(a1)


movea.l a1,a2 ; << this copies address of the boss in a2


bsr.w SingleObjLoad ; << find another empty object slot in the main SST space


bne.s loc_6ED0


move.b #$4F,0(a1) ; << this loads health bar object ID in the free address found by singleobjload


move.l a2,$30(a1) ; << this loads the address contained in a2 in $30(a1) (note, to do this, don't write (a2), it would move the data contained at the address pointed by a2 rather than copying the address itself)



and in your object, those first two line should look like:




obj4f_Flash:


movea.l $30(a0),a4 ; << this loads the address of the boss in a4

 

SpirituInsanum has an absolutely right idea.


I just wanted to do a small fix:



movea.l a1,a2 ; << this copies address of the boss in a2



instead of

Code:

		lea	a1,a2			; << this copies address of the boss in a2

 

Eek, you're right, I should never code when I'm tired xD


Besides, it should have been (a1),a2 not a1,a2. I'll correct it to avoid confusion.

 

Just asking... you saw my code? :'x

 

I don't know if this is a bug with the TilemapToVRAM function (ShowVDPGraphics in the old disassembly) or maybe I'm doing something wrong, but occasionally chunks of the tilemap don't draw at all. For example:


/monthly_02_2011/post-1201-1298178864_thumb.gif/monthly_02_2011/post-1201-1298178880_thumb.gif


It only happens every now and then, but often enough to be pretty annoying. Has anyone had this problem before, and worked out a solution?

 

Try this...


Before you call "TilemapToVRAM" put in "move #$2700,sr" to disable the IRQ's (interrupts), or put your mapping code somewhere during the V-Blank routines.

 

Just noticed that it does get fully drawn for maybe a single frame, something's overwriting it almost instantly though. Messing with the interrupts isn't helping. Frustrating :lol:


edit: fixed it, didn't realise I was calling TilemapToVRAM a second time :lol: I'm an idiot

 

Hi. I know, to show a map in plane A, I must made something with $C0000 (I think, anyway), but I need the map in plane B


How I do it?

 

Plane B mappings are stored to start from $E000 in Sonic 1.


Also, Plane A starts from $C000, not $C0000.

 

how do i make the rings decrease in sonic 1 as if sonic 2's super sonic? (but needing it for S monitor)


and with the eggman monitor (placed in GHZ for testing at the time) how do i make it so "Is sonic at 1 life? if yes branch" so that with 1 life you set an eggman monitor it won't take a life?


the way it's set up you get an eggman monitor, you're still alive but 1 life lost, but 1-0=0, 0-1=UNLIMITED. a glitch to be fixed.

 

Your second question is so simple, that it isn't even worth to be asked here, I think. Try putting some efforts in solving the problems before asking. To begin with, take a look at this: http://info.sonicretro.org/SCHG:Sonic_the_...hog/RAM_Editing