flan/examples/core-input-virtual-controls.flan
Joseph Ferano 25f56c24a8 A begin/end pair that cannot come apart, and what it still cannot promise
vendor/raylib/modes.flan: five macros over the five pairs the package binds
-- with-drawing, with-mode-2d, with-mode-3d, with-texture-mode,
with-scissor-mode. A second file with no declare-c in it, split out on
vector.flan's reasoning: raylib.flan is the package's statement about C and
nothing here names C, so nothing here can be made wrong by raylib changing.

Each expands to (do (begin-... args) body... (end-...)) -- the calls the
author used to type, in the order they typed them. No let, no gensym: nothing
binds a name, so there is nothing for a caller's name to collide with.

What it removes is the End* that is missing, wrong, or no longer beside its
Begin*. What it cannot remove is a body leaving through the unwind path: a
return or an invoke-restart skips the rest of the do and the End* with it.
defer is the obvious fix and is refused inside a loop body, which is where a
pair always lives -- checked, not assumed. So sand.flan's discipline stays:
keep the restart boundary outside the pair.

35 call sites converted across examples/ and sand.flan. The one left is
core-scissor-test.flan, whose Begin and End sit in two separate `when`s with
the drawing between them -- a conditional pair is a shape a bracketing macro
cannot express.

test/programs/rl-with.flan covers with-scissor-mode, which no example can,
with a frame function unreachable from main so it needs no libraylib;
rl-with-reject.flan is the arity half.
2026-09-13 18:00:35 +07:00

194 lines
8.4 KiB
Plaintext

;;;; raylib [core] example - input virtual controls
;;;;
;;;; examples/core/core_input_virtual_controls.c. No new bindings.
;;;;
;;;; Split the way sand.flan is split: everything above the second banner is
;;;; arithmetic with no raylib calls in it, so test/programs/virtual-controls-
;;;; headless.flan can import this file as a package, drive the pointer over a
;;;; scripted path and hash where the player ended up. That is the only one of
;;;; the ten examples with a headless half, and it has one because it is the
;;;; only one whose interesting part — which D-pad button is under the pointer,
;;;; and what that does to the player — is a function of numbers rather than of
;;;; a frame buffer. The link follows what the program reaches, so the headless
;;;; driver pulls in neither a window nor libraylib.
;;;;
;;;; Two language notes, both visible below.
;;;;
;;;; **No `break`.** The C's nearest-button search is
;;;;
;;;; for (i = 0; i < BUTTON_MAX; i++) { if (...) { pressedButton = i; break; } }
;;;;
;;;; and `(break)` is refused: "break is not implemented yet (see the build
;;;; sequence in plan.org)". `return` from the function works and is what this
;;;; uses — which is why the search is its own `defn` and not written inline
;;;; the way the C has it. That is not a loss: the function is the thing the
;;;; headless case wants to call anyway. Where a loop genuinely cannot be a
;;;; function, the shape left is a flag and a compound loop condition.
;;;;
;;;; **No `fabsf`.** The prelude has sqrt-f32 and sign-f32 but no absolute
;;;; value for floats, and the comment there says why for integers. `(max v
;;;; (- 0.0 v))` is it, and unlike the integer case it has no edge: there is no
;;;; float whose negation is itself except -0.0, whose absolute value is 0.0
;;;; either way.
;;;;
;;;; The button geometry is `defconst` arrays of struct literals. That works —
;;;; and it is the only way a fixed array can be made inside anything but a
;;;; top-level `defvar`, since a `let` binding takes no type annotation.
(import rl "vendor:raylib")
(defconst screen-width 800)
(defconst screen-height 450)
;; The C's PadButton enum. -1 is BUTTON_NONE and the four directions are 0..3,
;; in the order the arrays below are written. A `defenum` would not help: the
;; value is used as an index and an enum does not convert to an integer
;; ("i32 converts a number, found ..."), so it would have to be turned back
;; into one at every use.
(defconst button-none -1)
(defconst button-up 0)
(defconst button-left 1)
(defconst button-right 2)
(defconst button-down 3)
(defconst button-max 4)
(defconst button-radius f32 30.0)
;; padPosition is {100, 350} and the buttons sit a radius and a half away from
;; it on each axis. The C computes these at run time from padPosition; they are
;; written out here because a defconst is compile-time and the numbers are.
(defconst button-positions [button-max rl/Vector2]
[(rl/Vector2 {.x 100.0 .y 305.0}) ; up
(rl/Vector2 {.x 55.0 .y 350.0}) ; left
(rl/Vector2 {.x 145.0 .y 350.0}) ; right
(rl/Vector2 {.x 100.0 .y 395.0})]) ; down
(defconst player-speed f32 75.0)
;;;; ── The part with no raylib in it ──────────────────────────────────
(defvar player rl/Vector2)
(defn reset-player [] ()
(set player (rl/Vector2 {.x (/ (f32 screen-width) 2.0)
.y (/ (f32 screen-height) 2.0)})))
(defn abs-f32 [v f32] f32
(max v (- 0.0 v)))
;; The C's search, with `return` where it had `break`. Manhattan distance and
;; not Euclidean — the C adds the two axis distances and compares against the
;; radius — so the hit area is a diamond, which is deliberate: it makes the
;; four buttons tile without overlapping.
(defn nearest-button [pos rl/Vector2] i32
(dotimes [i button-max]
(let [b (at button-positions i)
dist-x (abs-f32 (- (.x b) (.x pos)))
dist-y (abs-f32 (- (.y b) (.y pos)))]
(when (< (+ dist-x dist-y) button-radius)
(return i))))
button-none)
;; The C's switch on the pressed button. Nothing moves for button-none, which
;; is the `default: break`.
(defn move-player [button i32 dt f32] ()
(let [step (* player-speed dt)]
(cond
(= button button-up) (set (.y player) (- (.y player) step))
(= button button-left) (set (.x player) (- (.x player) step))
(= button button-right) (set (.x player) (+ (.x player) step))
(= button button-down) (set (.y player) (+ (.y player) step))
:else (do))))
;; FNV-1a over the player's two floats, so the headless driver has one number
;; to compare. The floats are read through their bit patterns rather than
;; rounded to integers: rounding would hide exactly the kind of drift — a
;; crossed x and y, a step applied twice — the case exists to catch.
(defconst fnv-offset u64 0xcbf29ce484222325)
(defconst fnv-prime u64 1099511628211)
(defn hash-f32 [h u64 v f32] u64
;; No float-to-bits cast in the language, so the value is scaled and
;; truncated instead. 1024 keeps three decimal places of a screen
;; coordinate, which is finer than any real difference between two correct
;; runs and coarser than the last bit of an f32 — so this is reproducible
;; across targets where a raw bit pattern would depend on the FPU.
(let [n (i64 (* v 1024.0))
g h]
(dotimes [b 8]
(set g (bit-xor g (u64 (bit-and (>> n (i64 (* b 8))) 255))))
(set g (* g fnv-prime)))
g))
(defn hash-player [] u64
(hash-f32 (hash-f32 fnv-offset (.x player)) (.y player)))
;;;; ── The raylib front-end ───────────────────────────────────────────
;; The arrowheads, one triangle of three points per button, in the winding
;; raylib wants — counter-clockwise, or it culls them and draws nothing. A
;; [4 [3 rl/Vector2]]: a fixed array of fixed arrays of a struct, which is the
;; deepest shape any of these ten examples asks for and which works.
(defconst arrow-tris [button-max [3 rl/Vector2]]
[[(rl/Vector2 {.x 100.0 .y 293.0})
(rl/Vector2 {.x 91.0 .y 314.0})
(rl/Vector2 {.x 109.0 .y 314.0})]
[(rl/Vector2 {.x 64.0 .y 341.0})
(rl/Vector2 {.x 43.0 .y 350.0})
(rl/Vector2 {.x 64.0 .y 359.0})]
[(rl/Vector2 {.x 157.0 .y 350.0})
(rl/Vector2 {.x 136.0 .y 341.0})
(rl/Vector2 {.x 136.0 .y 359.0})]
[(rl/Vector2 {.x 91.0 .y 386.0})
(rl/Vector2 {.x 100.0 .y 407.0})
(rl/Vector2 {.x 109.0 .y 386.0})]])
(defconst label-colors [button-max rl/Color]
[(rl/Color {.r 253 .g 249 .b 0 .a 255}) ; yellow, up
(rl/Color {.r 0 .g 121 .b 241 .a 255}) ; blue, left
(rl/Color {.r 230 .g 41 .b 55 .a 255}) ; red, right
(rl/Color {.r 0 .g 228 .b 48 .a 255})]) ; green, down
(defn main [] ()
(rl/init-window screen-width screen-height
"raylib [core] example - input virtual controls")
(defer (rl/close-window))
(reset-player)
(rl/set-target-fps 60)
(until (rl/window-should-close?)
;; Update. Touch first, mouse as the desktop stand-in — and on the desktop
;; the left button has to be held, or the player would follow the cursor
;; wherever it went.
(let [touching (> (rl/get-touch-point-count) 0)
input (if touching
(rl/get-touch-position 0)
(rl/get-mouse-position))
pressed (if (or touching (rl/mouse-button-down? :left))
(nearest-button input)
button-none)]
(move-player pressed (rl/get-frame-time))
;; Draw
(rl/with-drawing
(rl/clear-background rl/raywhite)
(rl/draw-circle-v player 50.0 rl/maroon)
(dotimes [i button-max]
(rl/draw-circle-v (at button-positions i) button-radius
(if (= i pressed) rl/darkgray rl/black))
(let [t (at arrow-tris i)]
(rl/draw-triangle (at t 0) (at t 1) (at t 2)
(at label-colors i))))
(rl/draw-text "move the player with D-Pad buttons" 10 10 20 rl/darkgray)
;; Not in the C: which button the search picked, as a number, so the
;; headless case and the window agree about the same thing.
(rl/draw-text "button: " 10 34 20 rl/lightgray)
(rl/draw-text (string (i64->bytes (i64 pressed)))
(+ 10 (rl/measure-text "button: " 20)) 34 20
rl/lightgray)))))