Buffer Turns Through a Grid on Amiga | Pattern Library

A maze feels kinder when a turn pressed just before a junction happens as soon as the route opens.

Amiga shows the working diagnostic row 01 03 00 01 01 00 00 00. — A completed native caller displays 01 03 00 01 01 00 00 00. The public field order below identifies these bytes; the capture is a state diagnostic, not a recording of a finished game.

Use it in a project

CELL indexes a 4×4 row-major map; zero cells are walkable and nonzero cells block. Directions are 0 still, 1 left, 2 right, 3 up, 4 down. A nonzero REQUEST replaces PENDING. Try that turn first; if blocked, retain it and try the current DIRECTION. A successful turn clears PENDING. Each call moves at most one cell. Invalid indices/directions or a blocked current cell return VALID=0 without changing movement state.

Include component.inc once (MOS callers include the shared 6502 implementation through it). Reserve the public fields in this exact order: CELL, DIRECTION, PENDING, MOVED, VALID, REQUEST, C0, C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11, C12, C13, C14, C15. They occupy 22 contiguous bytes. Also reserve the private workspace: T0, T1, T2, T3, T4 (5 bytes). Array names in the explanation refer to the numbered consecutive bindings, such as C0, C1 and so on.

Call grid_turn_buffer_init once at the intended reset boundary. It clears inputs as well as outputs and workspace. Set your starting state and configuration afterwards, then call grid_turn_buffer_step whenever the described event occurs. The helper reads no device, waits for no frame and installs no interrupt handler.

Ownership	Fields
Retain between calls	CELL, DIRECTION and PENDING; map bytes
Supply before service	REQUEST
Read after service	CELL, DIRECTION, PENDING, MOVED and VALID

Try at most two neighbouring cells: the buffered request and the existing direction. Check column boundaries before adding or subtracting one so movement cannot wrap into the next row.

Reusable initialisation and service

component.inc

; SPDX-License-Identifier: MIT
; Buffer Turns Through a Grid. See README.md for the byte interface and ownership.
; init clears the public fields and private workspace; set inputs afterwards.
grid_turn_buffer_init:
    move.b #0,d0
    move.b d0,CELL
    move.b d0,DIRECTION
    move.b d0,PENDING
    move.b d0,MOVED
    move.b d0,VALID
    move.b d0,REQUEST
    move.b d0,C0
    move.b d0,C1
    move.b d0,C2
    move.b d0,C3
    move.b d0,C4
    move.b d0,C5
    move.b d0,C6
    move.b d0,C7
    move.b d0,C8
    move.b d0,C9
    move.b d0,C10
    move.b d0,C11
    move.b d0,C12
    move.b d0,C13
    move.b d0,C14
    move.b d0,C15
    move.b d0,T0
    move.b d0,T1
    move.b d0,T2
    move.b d0,T3
    move.b d0,T4
    rts
grid_turn_buffer_step:
; MOVED = 0
    move.b #0,d0
    move.b d0,MOVED
; VALID = 0
    move.b #0,d0
    move.b d0,VALID
; if CELL >= 16:
    move.b CELL,d0
    cmp.b #16,d0
    bcc grid_turn_buffer_L1
    jmp grid_turn_buffer_L2
grid_turn_buffer_L1:
; return
    rts
    jmp grid_turn_buffer_L3
grid_turn_buffer_L2:
grid_turn_buffer_L3:
; if DIRECTION > 4:
    move.b DIRECTION,d0
    cmp.b #4,d0
    bhi grid_turn_buffer_L4
    jmp grid_turn_buffer_L5
grid_turn_buffer_L4:
; return
    rts
    jmp grid_turn_buffer_L6
grid_turn_buffer_L5:
grid_turn_buffer_L6:
; if PENDING > 4:
    move.b PENDING,d0
    cmp.b #4,d0
    bhi grid_turn_buffer_L7
    jmp grid_turn_buffer_L8
grid_turn_buffer_L7:
; return
    rts
    jmp grid_turn_buffer_L9
grid_turn_buffer_L8:
grid_turn_buffer_L9:
; if REQUEST > 4:
    move.b REQUEST,d0
    cmp.b #4,d0
    bhi grid_turn_buffer_L10
    jmp grid_turn_buffer_L11
grid_turn_buffer_L10:
; return
    rts
    jmp grid_turn_buffer_L12
grid_turn_buffer_L11:
grid_turn_buffer_L12:
; T0 = C[CELL]
    moveq #0,d1
    move.b CELL,d1
    lea C0,a0
    move.b 0(a0,d1.w),d0
    move.b d0,T0
; if T0 != 0:
    move.b T0,d0
    cmp.b #0,d0
    bne grid_turn_buffer_L13
    jmp grid_turn_buffer_L14
grid_turn_buffer_L13:
; return
    rts
    jmp grid_turn_buffer_L15
grid_turn_buffer_L14:
grid_turn_buffer_L15:
; VALID = 1
    move.b #1,d0
    move.b d0,VALID
; if REQUEST != 0:
    move.b REQUEST,d0
    cmp.b #0,d0
    bne grid_turn_buffer_L16
    jmp grid_turn_buffer_L17
grid_turn_buffer_L16:
; PENDING = REQUEST
    move.b REQUEST,d0
    move.b d0,PENDING
    jmp grid_turn_buffer_L18
grid_turn_buffer_L17:
grid_turn_buffer_L18:
; T4 = 2
    move.b #2,d0
    move.b d0,T4
; T0 = PENDING
    move.b PENDING,d0
    move.b d0,T0
; while T4 != 0:
grid_turn_buffer_L19:
    move.b T4,d0
    cmp.b #0,d0
    bne grid_turn_buffer_L20
    jmp grid_turn_buffer_L21
grid_turn_buffer_L20:
; T1 = 255
    move.b #255,d0
    move.b d0,T1
; T2 = CELL & 3
    move.b CELL,d0
    and.b #3,d0
    move.b d0,T2
; if T0 == 1:
    move.b T0,d0
    cmp.b #1,d0
    beq grid_turn_buffer_L22
    jmp grid_turn_buffer_L23
grid_turn_buffer_L22:
; if T2 != 0:
    move.b T2,d0
    cmp.b #0,d0
    bne grid_turn_buffer_L25
    jmp grid_turn_buffer_L26
grid_turn_buffer_L25:
; T1 = CELL - 1
    move.b CELL,d0
    sub.b #1,d0
    move.b d0,T1
    jmp grid_turn_buffer_L27
grid_turn_buffer_L26:
grid_turn_buffer_L27:
    jmp grid_turn_buffer_L24
grid_turn_buffer_L23:
grid_turn_buffer_L24:
; if T0 == 2:
    move.b T0,d0
    cmp.b #2,d0
    beq grid_turn_buffer_L28
    jmp grid_turn_buffer_L29
grid_turn_buffer_L28:
; if T2 != 3:
    move.b T2,d0
    cmp.b #3,d0
    bne grid_turn_buffer_L31
    jmp grid_turn_buffer_L32
grid_turn_buffer_L31:
; T1 = CELL + 1
    move.b CELL,d0
    add.b #1,d0
    move.b d0,T1
    jmp grid_turn_buffer_L33
grid_turn_buffer_L32:
grid_turn_buffer_L33:
    jmp grid_turn_buffer_L30
grid_turn_buffer_L29:
grid_turn_buffer_L30:
; if T0 == 3:
    move.b T0,d0
    cmp.b #3,d0
    beq grid_turn_buffer_L34
    jmp grid_turn_buffer_L35
grid_turn_buffer_L34:
; if CELL >= 4:
    move.b CELL,d0
    cmp.b #4,d0
    bcc grid_turn_buffer_L37
    jmp grid_turn_buffer_L38
grid_turn_buffer_L37:
; T1 = CELL - 4
    move.b CELL,d0
    sub.b #4,d0
    move.b d0,T1
    jmp grid_turn_buffer_L39
grid_turn_buffer_L38:
grid_turn_buffer_L39:
    jmp grid_turn_buffer_L36
grid_turn_buffer_L35:
grid_turn_buffer_L36:
; if T0 == 4:
    move.b T0,d0
    cmp.b #4,d0
    beq grid_turn_buffer_L40
    jmp grid_turn_buffer_L41
grid_turn_buffer_L40:
; if CELL < 12:
    move.b CELL,d0
    cmp.b #12,d0
    bcs grid_turn_buffer_L43
    jmp grid_turn_buffer_L44
grid_turn_buffer_L43:
; T1 = CELL + 4
    move.b CELL,d0
    add.b #4,d0
    move.b d0,T1
    jmp grid_turn_buffer_L45
grid_turn_buffer_L44:
grid_turn_buffer_L45:
    jmp grid_turn_buffer_L42
grid_turn_buffer_L41:
grid_turn_buffer_L42:
; if T1 < 16:
    move.b T1,d0
    cmp.b #16,d0
    bcs grid_turn_buffer_L46
    jmp grid_turn_buffer_L47
grid_turn_buffer_L46:
; T3 = C[T1]
    moveq #0,d1
    move.b T1,d1
    lea C0,a0
    move.b 0(a0,d1.w),d0
    move.b d0,T3
; if T3 == 0:
    move.b T3,d0
    cmp.b #0,d0
    beq grid_turn_buffer_L49
    jmp grid_turn_buffer_L50
grid_turn_buffer_L49:
; CELL = T1
    move.b T1,d0
    move.b d0,CELL
; DIRECTION = T0
    move.b T0,d0
    move.b d0,DIRECTION
; if T4 == 2:
    move.b T4,d0
    cmp.b #2,d0
    beq grid_turn_buffer_L52
    jmp grid_turn_buffer_L53
grid_turn_buffer_L52:
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    jmp grid_turn_buffer_L54
grid_turn_buffer_L53:
grid_turn_buffer_L54:
; MOVED = 1
    move.b #1,d0
    move.b d0,MOVED
; return
    rts
    jmp grid_turn_buffer_L51
grid_turn_buffer_L50:
grid_turn_buffer_L51:
    jmp grid_turn_buffer_L48
grid_turn_buffer_L47:
grid_turn_buffer_L48:
; T4 = T4 - 1
    move.b T4,d0
    sub.b #1,d0
    move.b d0,T4
; T0 = DIRECTION
    move.b DIRECTION,d0
    move.b d0,T0
    jmp grid_turn_buffer_L19
grid_turn_buffer_L21:
    rts
component_end:

Machine contract

D0–D3/A0–A1/CCR may change; D4–D7/A2–A6 are preserved. A direct BSR/JSR needs four return-address stack bytes. The helper makes no nested calls and adds no other stack use. Its named scratch is private workspace, not a pointer, a ROM variable or shared zero-page space that the game may borrow concurrently. The samples use fixed symbol bindings. Scope or rename every label and binding, including the component_end measurement marker, and allocate a separate block for each independent instance. Different helpers can also use common field names, so combining them needs separate namespaces or explicitly renamed bindings. Give each block one mutation owner.

The A500 OCS diagnostic cold-boots an OFS disk, owns the Copper and bitplane and disables interrupt/DMA sources. The helper does not take over hardware or allocate Chip RAM.

Adapt the caller

demo.asm supplies the bindings and include order. exercise.inc first runs table cases, then a separate client retaining state between calls. A game can call the public step entry directly and remove the diagnostic packet and display. Do not copy the diagnostic’s hardware takeover into an existing game loop.

Build with make ASM198X=/path/to/asm198x and BUILD198X=/path/to/build198x. The Makefile produces normal, custom, deliberately wrong and repaired programs. Tools must already be installed; the sample downloads nothing.

Bindings and include order

demo.asm

    section code,code_c
CUSTOM = 0
BROKEN = 0
COYOTE_STEPS = 4
BUFFER_STEPS = 4
TEXT_CAP = 8
PROPERTY_COUNT = 16
STREAM_BYTES = 13
CRC_POLY = 7
    include "caller.inc"
    include "exercise.inc"
component_start:
    include "component.inc"
    even
    include "data.inc"
    include "cases.inc"
    even
    include "display.inc"
    include "../../framework/turn-deadline/font.inc"
code_end:
STATE:
CELL: dc.b 0
DIRECTION: dc.b 0
PENDING: dc.b 0
MOVED: dc.b 0
VALID: dc.b 0
REQUEST: dc.b 0
C0: dc.b 0
C1: dc.b 0
C2: dc.b 0
C3: dc.b 0
C4: dc.b 0
C5: dc.b 0
C6: dc.b 0
C7: dc.b 0
C8: dc.b 0
C9: dc.b 0
C10: dc.b 0
C11: dc.b 0
C12: dc.b 0
C13: dc.b 0
C14: dc.b 0
C15: dc.b 0
T0: dc.b 0
T1: dc.b 0
T2: dc.b 0
T3: dc.b 0
T4: dc.b 0
    even
ready: dc.w 0
preview: dcb.b 8,0
OUTPUT: dcb.b 1300,0
    even
signature: dc.b "198XNIGHTONE000",0
    even
copperlist:
    dc.w $08e,$2c81,$090,$2cc1,$092,$0038,$094,$00d0
    dc.w $100,$1200,$102,0,$104,0,$108,0,$10a,0
    dc.w $180,$012,$182,$fff
copbpl:
    dc.w $0e0,0,$0e2,0
    dc.w $ffff,$fffe
plane: dcb.b 10240,0

A client retaining state between calls

exercise.inc, excerpt

preview_rules:
    ifne CUSTOM
; CELL = 0
    move.b #0,d0
    move.b d0,CELL
; DIRECTION = 2
    move.b #2,d0
    move.b d0,DIRECTION
; REQUEST = 3
    move.b #3,d0
    move.b d0,REQUEST
    ifne BROKEN
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    endif
    bsr checked_step
    bsr emit_state
; REQUEST = 0
    move.b #0,d0
    move.b d0,REQUEST
    ifne BROKEN
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    endif
    bsr checked_step
    bsr emit_state
; REQUEST = 4
    move.b #4,d0
    move.b d0,REQUEST
    ifne BROKEN
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    endif
    bsr checked_step
    bsr emit_state
    rts
    else
; CELL = 8
    move.b #8,d0
    move.b d0,CELL
; DIRECTION = 2
    move.b #2,d0
    move.b d0,DIRECTION
; REQUEST = 3
    move.b #3,d0
    move.b d0,REQUEST
; C4 = 1
    move.b #1,d0
    move.b d0,C4
    ifne BROKEN
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    endif
    bsr checked_step
    bsr emit_state
; REQUEST = 0
    move.b #0,d0
    move.b d0,REQUEST
    ifne BROKEN
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    endif
    bsr checked_step
    bsr emit_state

    ifne BROKEN
; PENDING = 0
    move.b #0,d0
    move.b d0,PENDING
    endif
    bsr checked_step
    bsr emit_state
    endif
    rts

Customise it

Queue a blocked upward turn at the top edge, then replace it with a downward request. The maintained custom caller changes its inputs or client sequence rather than hiding a different implementation behind the same label. Compare every state of the two clients, not only their final display.

Custom client selection

custom.asm

    section code,code_c
CUSTOM = 1
BROKEN = 0
COYOTE_STEPS = 2
BUFFER_STEPS = 2
TEXT_CAP = 4
PROPERTY_COUNT = 16
STREAM_BYTES = 13
CRC_POLY = 49
    include "caller.inc"
    include "exercise.inc"
component_start:
    include "component.inc"
    even
    include "data.inc"
    include "cases.inc"
    even
    include "display.inc"
    include "../../framework/turn-deadline/font.inc"
code_end:
STATE:
CELL: dc.b 0
DIRECTION: dc.b 0
PENDING: dc.b 0
MOVED: dc.b 0
VALID: dc.b 0
REQUEST: dc.b 0
C0: dc.b 0
C1: dc.b 0
C2: dc.b 0
C3: dc.b 0
C4: dc.b 0
C5: dc.b 0
C6: dc.b 0
C7: dc.b 0
C8: dc.b 0
C9: dc.b 0
C10: dc.b 0
C11: dc.b 0
C12: dc.b 0
C13: dc.b 0
C14: dc.b 0
C15: dc.b 0
T0: dc.b 0
T1: dc.b 0
T2: dc.b 0
T3: dc.b 0
T4: dc.b 0
    even
ready: dc.w 0
preview: dcb.b 8,0
OUTPUT: dcb.b 1300,0
    even
signature: dc.b "198XNIGHTONE000",0
    even
copperlist:
    dc.w $08e,$2c81,$090,$2cc1,$092,$0038,$094,$00d0
    dc.w $100,$1200,$102,0,$104,0,$108,0,$10a,0
    dc.w $180,$012,$182,$fff
copbpl:
    dc.w $0e0,0,$0e2,0
    dc.w $ffff,$fffe
plane: dcb.b 10240,0

The custom diagnostic row is 06 04 00 01 01 04 00 00. The full client trace in the source explains how it gets there.

Mistake and repair

The wrong caller clears the queued turn between cells. The player continues along the row instead of turning upwards when the next junction opens. Retain PENDING until a turn succeeds or a new request replaces it.

The BROKEN branch lives in the sequential client, not the reusable helper. wrong.asm selects it; repaired.asm restores the normal call sequence. The repaired executable matches the normal program (for Spectrum, compare loaded CODE rather than the tape filename).

Select the deliberately wrong client

wrong.asm

    section code,code_c
CUSTOM = 0
BROKEN = 1
COYOTE_STEPS = 4
BUFFER_STEPS = 4
TEXT_CAP = 8
PROPERTY_COUNT = 16
STREAM_BYTES = 13
CRC_POLY = 7
    include "caller.inc"
    include "exercise.inc"
component_start:
    include "component.inc"
    even
    include "data.inc"
    include "cases.inc"
    even
    include "display.inc"
    include "../../framework/turn-deadline/font.inc"
code_end:
STATE:
CELL: dc.b 0
DIRECTION: dc.b 0
PENDING: dc.b 0
MOVED: dc.b 0
VALID: dc.b 0
REQUEST: dc.b 0
C0: dc.b 0
C1: dc.b 0
C2: dc.b 0
C3: dc.b 0
C4: dc.b 0
C5: dc.b 0
C6: dc.b 0
C7: dc.b 0
C8: dc.b 0
C9: dc.b 0
C10: dc.b 0
C11: dc.b 0
C12: dc.b 0
C13: dc.b 0
C14: dc.b 0
C15: dc.b 0
T0: dc.b 0
T1: dc.b 0
T2: dc.b 0
T3: dc.b 0
T4: dc.b 0
    even
ready: dc.w 0
preview: dcb.b 8,0
OUTPUT: dcb.b 1300,0
    even
signature: dc.b "198XNIGHTONE000",0
    even
copperlist:
    dc.w $08e,$2c81,$090,$2cc1,$092,$0038,$094,$00d0
    dc.w $100,$1200,$102,0,$104,0,$108,0,$10a,0
    dc.w $180,$012,$182,$fff
copbpl:
    dc.w $0e0,0,$0e2,0
    dc.w $ffff,$fffe
plane: dcb.b 10240,0

Amiga shows the deliberate error row 0B 02 00 01 01 00 00 00. — The wrong client completes with 0B 02 00 01 01 00 00 00; its changed inputs and state explain the result.

Cost and alternatives

The measured helper extent is 730 code bytes; public state is 22 bytes and private scratch is 5 bytes. There are no immutable helper tables. Caller data, the test cases, return-address stack, alignment and display buffers cost additional space. These are assembled sizes, not cycle measurements.

At most two map reads after validation. A larger map needs a wider index and declared dimensions; a pixel mover needs centre-line snapping and collision checks between cell boundaries.

This is logical cell movement, not pixel interpolation or a complete maze controller. Call only at a cell-step boundary; animate between those positions separately. The map is read-only during service.

Execution evidence

Four variants run on A500 OCS PAL, Kickstart1.3 OFS cold boot, using Asm198x 0.0.72, Emu198x 0.44.2 and Build198x 0.2.9. Each executes 45 boundary/varied/error table cases, dirty initialisation and the separate retained-state client. An independent mathematical/container model checks every public byte after every call. This is bounded example coverage, not exhaustive correctness for every possible state.

The records retain loaded immutable code, completion, register sentinels, full packets, source/tool/firmware hashes and captures. The public gate validates these retained observations and rejects altered evidence; it does not run restricted firmware. No firmware is distributed. The eight displayed bytes are the first eight public fields, zero-padded when shorter; the packet includes every field.

Where it fits

Four-way input: Resolve contradictory buttons before turning the result into a buffered movement request.

The Craft · Maths for Games · Source and native evidence.