ROM building glitch?

Here I am again, another problem, probably the weirdest one yet, but it gets rid of a lot of bugs! Any help would be most grateful!


You can skip this story (spoiler) if you like and get on with the real problem.

Spoiler

I have put Sonic 3 sprites in my game, so as you could imagine, the Sonic running across the SEGA logo went wrong. I couldn't be bothered to fix it. But later on, I put RHS in the game, and he uses the traditional Sonic 2 sprites. So, I thought I would fix Sonic at the SEGA logo. I made it load RHS's art, his running DPLC's and mappings, but no matter what I tried, it never worked.


Here are some codes that I've done, can you see where I went wrong?


You might want to read about the problem I have first before looking into depth about this code

Code:

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

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

; Object B0 - SEGA screen? (Unknown)

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



; Sprite_3A1DC:

ObjB0:

moveq #0,d0

move.b routine(a0),d0

move.w off_3A1EA(pc,d0.w),d1

jmp off_3A1EA(pc,d1.w)



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



off_3A1EA:

dc.w loc_3A1F6-off_3A1EA

dc.w loc_3A2F4-off_3A1EA; 1

dc.w loc_3A32C-off_3A1EA; 2

dc.w loc_3A396-off_3A1EA; 3

dc.w loc_3A3CC-off_3A1EA; 4

dc.w return_3A3F6-off_3A1EA; 5



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



loc_3A1F6:

bsr.w LoadSubObject

move.w #$1E8,x_pos(a0)

move.w #$F0,objoff_A(a0)

move.w #$B,objoff_2A(a0)

move.w #2,($FFFFF662).w

bset #0,render_flags(a0)

bset #0,status(a0)

lea ($FFFFE138).w,a1

lea byte_3A76C(pc),a2

moveq #0,d0

moveq #$22,d6



loc_3A22A:

move.b (a2)+,d0

add.w d0,(a1)

addq.w #8,a1

dbf d6,loc_3A22A

lea off_3A294(pc),a1

; lea (ArtUnc_Sonic).l,a3

  lea (ArtUnc_S2sonicart).l,a3

conloc_3A22A:

lea (Metablock_Table).l,a5

moveq #3,d5



; etc, etc

Code:

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

; These next four things are pointers to Sonic's dereferenced

; DPLC entries of his "running animation" frames for the SEGA screen.

; I want that DPLC data split into a binary file for use with editors,

; but unfortunately there's no way to refer to BINCLUDE'd bytes

; from within AS, so I put an educated guess (default) here and

; run an external program (fixpointer.exe) to fix it later.

; WARNING: the build script needs editing if you rename this label



;off_3A294:

; dc.l (MapRUnc_Sonic+$33A) ;dc.l word_7181A

; dc.l (MapRUnc_Sonic+$340) ;dc.l word_71820

; dc.l (MapRUnc_Sonic+$346) ;dc.l word_71826

;  dc.l (MapRUnc_Sonic+$34C) ;dc.l word_7182C



off_3A294: dc.l MapRUnc_RHS_2DE  ; 0 ; DATA XREF: h+232C0t h+23324o ...

  dc.l MapRUnc_RHS_2E4  ; 1

  dc.l MapRUnc_RHS_2EA  ; 2

  dc.l MapRUnc_RHS_2F0  ; 3

; The above is RHS's running DPLC's, I've trippled checked



word_3A2A4:

dc.w $FFFF,	0,$FFFF, $B00, $201

dc.w $FFFF,  $C0,$FFFF, $E00, $303; 5

dc.w $FFFF, $2C0,$FFFF,$1600, $201; 10

dc.w $FFFF, $380,$FFFF,$1900, $303; 15

dc.w $FFFF, $580,$FFFF,$2100, $201; 20

dc.w $FFFF, $640,$FFFF,$2400, $303; 25

dc.w $FFFF, $840,$FFFF,$2C00, $201; 30

dc.w $FFFF, $900,$FFFF,$2F00, $303; 35

Code:

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

; sprite mappings

; Gigantic Sonic (2x size) mappings for the SEGA screen

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

;ObjB1_MapUnc_3A5A6: BINCLUDE "S3Sonic/mapsega.bin"

;ObjB1_MapUnc_3A5A6:

; BINCLUDE "mappings/sprite/objB1.bin" ;try thisArtUnc_3A5A6

ArtUnc_3A5A6:	; DATA XREF: h+23616o h+23620o ...

  dc.w word_3A5B0-ArtUnc_3A5A6; 0

  dc.w word_3A5E2-ArtUnc_3A5A6; 1

  dc.w word_3A614-ArtUnc_3A5A6; 2

  dc.w word_3A646-ArtUnc_3A5A6; 3

  dc.w word_3A678-ArtUnc_3A5A6; 4

word_3A5B0: dc.w 6   ; DATA XREF: h+23632o

  dc.w $D80F,	0,	0,$FFF0; 0

  dc.w $D807,  $10,	8,  $10; 4

  dc.w $F80F,  $18,   $C,$FFE0; 8

  dc.w $180F,  $28,  $14,$FFE0; 12

  dc.w $F80F,  $38,  $1C,	0; 16

  dc.w $180F,  $48,  $24,	0; 20

word_3A5E2: dc.w 6   ; DATA XREF: h+23632o

  dc.w $D80F,  $58,  $2C,$FFF0; 0

  dc.w $D807,  $68,  $34,  $10; 4

  dc.w $F80F,  $70,  $38,$FFE0; 8

  dc.w $180F,  $80,  $40,$FFE0; 12

  dc.w $F80F,  $90,  $48,	0; 16

  dc.w $180F,  $A0,  $50,	0; 20

word_3A614: dc.w 6   ; DATA XREF: h+23632o

  dc.w $D80F,  $B0,  $58,$FFF0; 0

  dc.w $D807,  $C0,  $60,  $10; 4

  dc.w $F80F,  $C8,  $64,$FFE0; 8

  dc.w $180F,  $D8,  $6C,$FFE0; 12

  dc.w $F80F,  $E8,  $74,	0; 16

  dc.w $180F,  $F8,  $7C,	0; 20

word_3A646: dc.w 6   ; DATA XREF: h+23632o

  dc.w $D80F, $108,  $84,$FFF0; 0

  dc.w $D807, $118,  $8C,  $10; 4

  dc.w $F80F, $120,  $90,$FFE0; 8

  dc.w $180F, $130,  $98,$FFE0; 12

  dc.w $F80F, $140,  $A0,	0; 16

  dc.w $180F, $150,  $A8,	0; 20

word_3A678: dc.w 2   ; DATA XREF: h+23632o

  dc.w $FC00,	0,	0,$FFF8; 0

  dc.w $FC00,	0,	0,	0; 4

  even

sprite glitching (well I played through my whole ROM twice this way and never came across it, even by going to the places it normally happens, it was fine).





Here is a video demonstrating what I am trying to say. As you can see, after if finally opens the ROM, Sonic is there. Resetting the ROM still shows Sonic (as you can see) But when I close the ROM then reload the same ROM, Sonic has gone again. Resetting the ROM stills shows no Sonic.



Sorry for the crap sound and video quality, the video freezes at the 0:13 point for a second for some reason





Has anyone else ever encountered this bug and no how to fix it? I'd greatly appreciate it.



P.S I do have this

Code:

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

;Uncompressed art

;

;Patterns for Sonic 2

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

align $20000

ArtUnc_S2sonicart:	 BINCLUDE "artuncompressedRHSart.bin"

With the alignment, Sonic isn't there. Without the alignment, there's garbled "Knuckles" sprites.





Loading the ROM quickly with $20000 alignment





Loading the ROM anytime with $20000 alignment





Loading the ROM quickly without alignment





Loading the ROM anytime without alignment


Cheers,


redhotsonic


EDIT: Forgot to mention, I'm using Xenowhirl's 2007 disassembly.

 

I'm really no expert, but maybe a problem with the beginning of the rom or its end?


Could you post what's from the beginning of your disassembly till the "errortrap:" routine?


How's the end btw? "EndOfRom: END"?

 

Once I plunked in S3's sprites, iirc I added 'even' instead of align and that helped it load up for me. Without that it would lock up...

 

Sure:



Beginning of ASM:

Code:

; >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

; start of ROM

StartOfRom:

if * <> 0

fatal "StartOfRom was ${*} but it should be 0"

endif

;Vectors:

dc.l System_Stack, EntryPoint, ErrorTrap, ErrorTrap; 4

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 8

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 12

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 16

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 20

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 24

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 28

dc.l H_Int, ErrorTrap, V_Int, ErrorTrap; 32

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 36

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 40

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 44

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 48

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 52

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 56

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 60

dc.l ErrorTrap, ErrorTrap, ErrorTrap, ErrorTrap; 64

; byte_200:

Header:

dc.b "SEGA GENESIS " ; Console name

dc.b "(C)SEGA 1992.SEP" ; Copyright/Date

dc.b "SONIC 2 RECREATION BETA REDHOTSONIC DATE28/01/12" ; Domestic name

dc.b "SONIC 2 RECREATION BETA REDHOTSONIC DATE28/01/12" ; International name

dc.b "GM 00001051-01" ; Version

; word_18E

Checksum:

dc.w $D951 ; Checksum (patched later if incorrect)

dc.b "J " ; I/O Support

dc.l StartOfRom ; ROM Start

; dword_1A4

ROMEndLoc:

dc.l EndOfRom-1 ; ROM End

dc.l $FF0000 ; RAM Start

dc.l $FFFFFF ; RAM End

dc.b " " ; Backup RAM ID

dc.l $20202020 ; Backup RAM start address

dc.l $20202020 ; Backup RAM end address

dc.b " " ; Modem support

dc.b " " ; Notes

dc.b "JUE " ; Country

EndOfHeader:

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

; Crash/Freeze the 68000. Note that the Z80 continues to run, so the music keeps playing.

; loc_200:

ErrorTrap:

nop

nop

bra.s ErrorTrap

End of ASM:

Code:

 align $20000

ArtUnc_S2sonicart: BINCLUDE "artuncompressedRHSart.bin"



Nem_LevSel: binclude "LS_BG/GFX2.bin"

even



Eni_LevSel: binclude "LS_BG/Map2.bin"

even



Pal_LevSel: binclude "LS_BG/Pal2.bin"

even



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



EndOfRom:

END

With the align $2000, even, or nothing, it won't work, if I put it above or underneath the art.bin

 

Maybe you could try without fixing the pointers and/or checksum :/

 

Tried that, didn't work. but...

It's now fixed, I think. Don't understand how lol


Thanks guys!

 

Touche ^_^. Very puzzling indeed, glad to see you have it sorted out, I'll certainly keep everything said here in mind. Glad you have it fixed and can't wait to see what content you have brewing!

 

I can't keep brewing, the game has been randomly freezing recently. Ever since I ported object manager, so I've gone wrong somewhere. Just don't know where. It can happen when underwater, when about to get hurt, or just when running, etc.

 

More than likely something is misaligned make sure there is an even or in your case as I assume you are using the AS version of sonic 2 use

Code:

align 2

after all bincludes and after all data labels but not after each command by the was so don't put an align 2 after every dc.l dc.w dc.b just at the end off it that should more than likely fix everything

 

It's not that, I think it's something to do with the respawn_index. It had to become a word for the S3K object manager.

Code:

anim_frame = $1B

anim = $1C

next_anim = $1D

respawn_index = $1E ; and $35, it's now a word ; was just 23.

anim_frame_duration = $23

status = $22 ; note: exact meaning depends on the object... for sonic/tails: bit 0: leftfacing. bit 1: inair. bit 2: spinning. bit 3: onobject. bit 4: rolljumping. bit 5: pushing. bit 6: underwater.

routine = $24 ; used to be 24

routine_secondary = $25 ; used to be 25

angle = $26 ; angle about the z=0 axis (360 degrees = 256)

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

; conventions followed by many objects but NOT sonic/tails:

collision_flags = $20

collision_property = $21

; $23 is free

subtype = $28

anim_frame_duration used to be a word, but I've fixed that by changing all the words to bytes. Here's a gameover example:

Code:

loc_13FEE:

btst #0,mapping_frame(a0)

bne.w BranchTo17_DisplaySprite

;tst.w (Game_paused).w ; is game already paused?

;beq.w loc_1407E ; if yes, branch

move.b (Ctrl_1_Press).w,d0

or.b (Ctrl_2_Press).w,d0

andi.b #$70,d0

bne.s loc_14014

tst.b anim_frame_duration(a0) ; anim_frame_duration - change from .w to .b

beq.s loc_14014

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

  move.b  ($FFFFFE05).w,d0 ; Move Game Frame timer to d0

  andi.b  #3,d0 ; andi by 3

  bne.s + ; if d0 does NOT equal 0, skip subtracting 1 byte from anim_frame_duration

; otherwise, subtract 1 byte from anim_frame_duration

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

subq.b #1,anim_frame_duration(a0) ; anim_frame_duration - change from .w to .b

+ ; new label, so the above can skip the subtraction of anim_frame_duration correctly

bra.w DisplaySprite

Instead of counting down every frame, it counts down every 4 frames. I tested it and it works fine.


I think it's where I stuck respawn_index. Here's a video. You'll notice I jump over the crab at first, but when I go back to him, he's gone. I touch his position, freeze. The freezing usually occurs when I'm about to touch something.


EDIT: Thanks to flamewings code, I can pinpoint why it's happening.


It is definately respawn_index. Some objects do NOT respawn when you leave and re-enter the area if you haven't destroyed them. This can happen to badniks, springs, anything. Then when you go to touch where they originally were, BOOM, error.





In the RARE case that it doesn't freeze when you touch their location where they should be, this triggers something else. As soon as you next get hurt, it will freeze.





I changed the text between both builds from flamewing to redhotsonic =P

 

341689 or 536B9 is not dividable by 2 and 12BB5 or 76725 is also not dividable by 2 so in other words it is not at an even address now there are more than likly one of two possibilities


1: something is not even aligned


2: as you said you have converted anim_frame_duration from a word to a byte maybe somewhere it is still a word Take this code for an example I believe that it wouldn't work




move.b (a1)+,d1;even though a1 started out at an even address it is now an odd address as adding one to an even number makes it an ;odd number as far as I am ;aware move.b can read odd address were as move.w needs to be even


move.w (a1)+,d1;can't read from odd address

 

I will go through my ASM file tomorrow and after all dc.b's and BINCLUDES, I will add an even at the end.


Although something is more than likely not aligned, I still think respawn_index is the cause. I will post an update tomorrow. I will spend some time on adding all EVENs, then I will test. If it still occurs, it's definitely respawn_index being moved from $23 to $1E and $1F.


anim_frame_duration is DEFINITELY all .b's, I've checked over 5 times.

 

Ok glad you got it figured I relay do hope that this works because I saw the video to your hack and looks cool ignoring the crash so don't give up on it.

 

Update. EHZ never freezes anymore, neither does any other level except ARZ. It still keeps freezing and I'm not sure how to solve it. May not be an alignment problem after all.

 

On the plus side aligning every thing did help a lot only one more level to go nice work redhotsonic!

 

Seems like ARZ is now fixed. Flamewing pinpointed the line that was causing the problem. This was all due to the new S3K object management, but now it's implemented and no more freezing =)