duet.z80

; VDP manual: http://bifi.msxnet.org/msxnet/tech/tms9918a.txt
; More on VDP and speed limits: http://map.grauw.nl/articles/vdp_tut.php

;include "msx.h"
include "svi.h"

  ; Header values for svitool to package it in a tape container
;  dw  start, end-1, start

  org 0xA000

start:
  call INIGRP

  call set_single_pattern_table
  call set_pattern_name_table
  call set_color_palette
  call clearbuffer

  call main_bg_loop
  ;call test_static

  call CHGET ; clear keyboard buffer
  call INITXT
  ret

test_static:
  ld b,0
  ld c,3
  ld d,0
  ld e,0
  call drawfigurexy

  ld b,0
  ld c,3
  ld d,32
  ld e,5
  ;call drawfigurexy

  ld b,5
  ld c,1
  ld d,0
  ld e,48
  ;call drawfigurexy

  ld b,7
  ld c,2
  ld d,32
  ld e,32
  ;call drawfigurexy

  call flushbuffer
  ret


main_bg_loop:
  halt

  ; Change backdrop for performance markers
  ;ld a,2
  ;out (VDP_CW),a
  ;ld a,128+7
  ;out (VDP_CW),a
  call .update_background
  ;ld a,8
  ;out (VDP_CW),a
  ;ld a,128+7
  ;out (VDP_CW),a
  call CHSNS
  jr z,main_bg_loop
  ret

framestate:
  db 0, 14  ; rotation frames
  dw rotation_wave
  dw (rotation_wave_end-rotation_wave)/2+rotation_wave
  db 0 ; angle

.update_background:
  call clearbuffer
  ;halt
  ;halt

  ld a,(framestate+6)
  ;and 127
  ld l,a
  ld h,0
  add hl,hl
  ld de,sintable
  add hl,de
  ld l,(hl)

  push hl
  ld a,(framestate+6)
  add a,64
  ;and 127
  ld l,a
  ld h,0
  add hl,hl
  ld de,sintable
  add hl,de
  ld a,(hl)
  pop hl
  ld h,a




  ld a,(framestate)
  ld b,a
  ld c,3
  ld de,0x0000
  ld a,l
  add d
  ld d,a
  ld a,h
  add e
  ld e,a
  call drawfigurexy

  ld a,(framestate+1)
  ld b,a
  ld c,1
  ld de,0x2004
  ld a,l
  add d
  ld d,a
  ld a,h
  add e
  ld e,a
  call drawfigurexy

  ld a,(framestate+1)
  ld b,a
  ld c,1
  ld de,0x0424
  ld a,l
  add d
  ld d,a
  ld a,h
  add e
  ld e,a
  call drawfigurexy

  ld a,(framestate)
  ld b,a
  ld c,0
  ld de,0x2424
  ld a,l
  add d
  ld d,a
  ld a,h
  add e
  ld e,a
  call drawfigurexy

  ld de,(framestate+2)
  ld hl,framestate
  ld a,(de)
  add a,(hl)
  ld (hl),a
  inc de
  ld hl,rotation_wave_end
  call compare_de_hl
  jr nz,.ub_nowrap1
  ld de,rotation_wave
.ub_nowrap1:
  ld (framestate+2),de

  ld de,(framestate+4)
  ld hl,framestate+1
  ld a,(de)
  neg
  add a,(hl)
  ld (hl),a
  inc de
  ld hl,rotation_wave_end
  call compare_de_hl
  jr nz,.ub_nowrap2
  ld de,rotation_wave
.ub_nowrap2:
  ld (framestate+4),de

  ld a,(framestate+6)
  inc a
  ld (framestate+6),a

  call flushbuffer
  ret

drawfigurexy:
  push bc
  push de
  push hl
  ; B = rotation frame (lower 4 bits will be used)
  ; C = size (0-3)
  ; D,E = x,y coordinates in drawing buffer

  ; Ensure coordinates are 0 <= x,y < 64
  ld a,d
  and 63
  ld d,a
  ld a,e
  and 63
  ld e,a

  call get_figure_ptr
  ; now HL -> figure to draw
  ld a,d
  ex de,hl
  ld h,0 ; HL is now y coordinate in buffer to draw to
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl
  ld bc,gfxbuffer
  add hl,bc
  ld c,a
  srl c
  srl c
  srl c
  ld b,0
  add hl,bc

  ex de,hl
  ; HL -> figure to draw
  ; DE -> correct position in the first row in buffer to draw to
  ; A = x coordinate

  ld b,32
  ld c,a

.processrow:
  ; B rows left
  ; HL -> current position in figure
  ; DE -> correct position in row in drawing buffer
  ; C = x offset

  push bc
  push hl
  push de

  ld b,c ; B = number of times to rotate
  ld c,(hl)
  inc hl
  ld d,(hl)
  inc hl
  ld e,(hl)
  inc hl
  ld h,(hl)
  ld l,0
  ld a,c
  or d
  or e
  or h
  jr z,.rowdone ; empty row in figure, skip it

  ld a,b
  and 7
  jr z,.rotationdone
  ld b,a
  ; If we need to rotate 5,6 or 7 times, prefer rotate 3, 2, 1 times the other way instead
  cp 5
  jr nc,.setuprotateleft
.rotateright:
  srl c
  rr d
  rr e
  rr h
  rr l
  djnz .rotateright
  jr .rotationdone
.setuprotateleft:
  ld a,8
  sub b
  ld b,a
  ; Shift the data
  ld l,h
  ld h,e
  ld e,d
  ld d,c
  ld c,0
.rotateleft:
  sla l
  rl h
  rl e
  rl d
  rl c
  djnz .rotateleft

.rotationdone:
  ; Data is now correctly rotated in BCDHL
  ld a,c
  ld c,l
  ld b,h
  pop hl
  push hl
  ; ADEBC should now be OR:ed with HL..HL+4
  ; Because of the way the drawing buffer is aligned, we can do some
  ; tricks to make the wraparound to work fast
  or (hl)
  ld (hl),a
  inc l
  res 3,l
  ld a,(hl)
  or d
  ld (hl),a
  inc l
  res 3,l
  ld a,(hl)
  or e
  ld (hl),a
  inc l
  res 3,l
  ld a,(hl)
  or b
  ld (hl),a
  inc l
  res 3,l
  ld a,(hl)
  or c
  ld (hl),a
.rowdone:
  pop de
  ld hl,16
  add hl,de
  ex de,hl
  ; DE -> next line in drawing buffer
  ld hl,gfxbufferend
  call compare_de_hl
  jr c,.cont
  ld hl,-1024
  add hl,de
  ex de,hl
.cont:
  pop hl
  inc hl
  inc hl
  inc hl
  inc hl
  ; HL -> next line in figure
  pop bc
  ; B = rows left
  djnz .processrow
  pop hl
  pop de
  pop bc
  ret

; Z will be set if DE=HL, C set if DE<HL
compare_de_hl:
  ld a,d
  cp h
  ret nz
  ld a,e
  cp l
  ret

drawbuffer:
  ; B = rotation frame (lower 4 bits will be used)
  ; C = size (0-3)
  ; DE = position in drawing buffer
  call get_figure_ptr

  ld b,32
.db2loop:
  push bc
  ldi
  ldi
  ldi
  ldi
  inc de
  inc de
  inc de
  inc de
  pop bc
  djnz .db2loop

  ret

clearbuffer:
  push bc
  push de
  push hl
  ld hl,gfxbuffer
  ld de,9
  ; Clear 64 rows
  ld b,64
  xor a
.cb_loop:
  ; Clear 8 bytes, then skip 8 bytes
  ld (hl),a
  inc l
  ld (hl),a
  inc l
  ld (hl),a
  inc l
  ld (hl),a
  inc l
  ld (hl),a
  inc l
  ld (hl),a
  inc l
  ld (hl),a
  inc l
  ld (hl),a
  add hl,de
  djnz .cb_loop
  pop hl
  pop de
  pop bc
  ret

flushbuffer:
  push bc
  push de
  push hl
  ld hl,0x0000
  call set_vdp_write_addr
  ld hl,gfxbuffer
  ld c,VDP_DW
  ld a,8 ; 8 rows
.fb2_nextrow:
  ld b,64 ; 64 bytes to write per row
.fb2_nexttile:
  ld de,15 ; 1 is added automatically after every outi
  push hl
  outi
  add hl,de
  outi
  add hl,de
  outi
  add hl,de
  outi
  add hl,de
  outi
  add hl,de
  outi
  add hl,de
  outi
  add hl,de
  outi
  pop hl
  inc hl
  jr nz, .fb2_nexttile ; flag was set from last outi
  ld de,7*16+8
  add hl,de
  dec a
  jr nz,.fb2_nextrow
  pop hl
  pop de
  pop bc
  ret

get_figure_ptr:
  ; B = rotation frame (lower 5 bits will be used)
  ; C = size (0-3)

  ; gfx start at gfx+(c*32+b)*128
  ld l,c
  ld h,0
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl ; HL = C*32
  ld a,b
  and 31
  ld c,a
  ld b,0
  add hl,bc ; HL = C*16+B
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl
  add hl,hl ; HL = (C*16+B)*128
  ld bc,gfx
  add hl,bc
  ret

draw_tile:
  ld b,8
  ld c,VDP_DW
.dt_loop:
  outi
  jr nz,.dt_loop
  ret

set_single_pattern_table:
  ; Sets all three parts of the screen to use the same pattern table
  di
  ld a,0
  out (VDP_CW),a
  ld a,128+4
  out (VDP_CW),a

  ld a,128+31
  out (VDP_CW),a
  ld a,128+3
  out (VDP_CW),a
  ei
  ret

set_color_palette:
  ; Set backdrop to blue
  ld a,4
  out (VDP_CW),a
  ld a,128+7
  out (VDP_CW),a

  ; Sets foreground to white, background to black for the 256 chars used
  ld hl,0x2000
  call set_vdp_write_addr
  ld d,8
  ld b,0
  ld a,0x54 ; FG=gray BG=black
.scp_loop:
  out (VDP_DW),a
  djnz .scp_loop
  dec d
  jr nz,.scp_loop
  ret

set_pattern_name_table:
  ; Creates an 8x8 matrix of characters 0-63 and repeats it 4x3 over the screen
  ld hl,0x1800
  call set_vdp_write_addr
  ; 01234567012345670123456701234567
  ; 89ABCDEF89ABCDEF89ABCDEF89ABCDEF
  ; etc
  ld d,0 ; row offset
  ld e,24 ; rows left
.spnt_loop2:
  ld c,0 ; col offset
  ld b,32 ; cols left
.spnt_loop1:
  ld a,c
  add a,d
  out (VDP_DW),a
  inc c
  res 3,c
  djnz .spnt_loop1
  ld a,8
  add a,d
  and 63
  ld d,a
  dec e
  jr nz,.spnt_loop2
  ret

; Sets the VDP address registers to HL and prepares for writing
; C will be equal to VDP_DW
set_vdp_write_addr:
  ld c,VDP_CW
  set 6,h
  di
  ld a,l
  out (VDP_CW),a
  ld a,h
  out (VDP_CW),a
  ei
  ret

gfx:
  incbin "squares.bin"

rotation_wave:
  db 1,3,4,5,5,5,4,3,1
  db 0,0,0,0,0,0,0,0,0
  db 0,0,0,0,0,0,0,0,0
  db 0,0,0,0,0,0,0,0,0
rotation_wave_end:
  db -1

sintable:
  incbin "sin.bin"

; The drawing buffer is 64x64 pixel in size
; Every row is 16 bytes:
;  * The first 8 bytes correspond to the 64 pixels
;  * Then their are 8 empty bytes, to make wraparound easier
gfxbuffer:    equ start-0x1000 ;8*8*8
gfxbufferend: equ gfxbuffer+2*8*8*8

end:

  end