summaryrefslogtreecommitdiff
path: root/math.c
diff options
context:
space:
mode:
Diffstat (limited to 'math.c')
-rw-r--r--math.c856
1 files changed, 0 insertions, 856 deletions
diff --git a/math.c b/math.c
deleted file mode 100644
index db8e4f2..0000000
--- a/math.c
+++ /dev/null
@@ -1,856 +0,0 @@
-#ifdef MATH_GL
-#undef MATH_GL
-#define MATH_GL 1
-#endif
-
-#include <stdlib.h>
-#include <stdio.h>
-#include <assert.h>
-
-#define PIf 3.14159265358979f
-#define HALF_PIf 1.5707963267948966f
-#define TAUf 6.283185307179586f
-#define SQRT2f 1.4142135623730951f
-#define HALF_SQRT2f 0.7071067811865476f
-#define SQRT3f 1.7320508075688772f
-#define HALF_SQRT3f 0.8660254037844386f
-
-#include <math.h>
-
-static inline float degrees(float r) {
- return r * (180.0f / PIf);
-}
-static inline float radians(float r) {
- return r * (PIf / 180.f);
-}
-
-// map x from the interval [0, 1] to the interval [a, b]. does NOT clamp.
-static inline float lerpf(float x, float a, float b) {
- return x * (b-a) + a;
-}
-
-// opposite of lerp; map x from the interval [a, b] to the interval [0, 1]. does NOT clamp.
-static inline float normf(float x, float a, float b) {
- return (x-a) / (b-a);
-}
-
-static inline float clampf(float x, float a, float b) {
- if (x < a) return a;
- if (x > b) return b;
- return x;
-}
-
-static inline int clampi(int x, int a, int b) {
- if (x < a) return a;
- if (x > b) return b;
- return x;
-}
-
-static inline i16 clamp_i16(i16 x, i16 a, i16 b) {
- if (x < a) return a;
- if (x > b) return b;
- return x;
-}
-
-static inline u16 clamp_u16(u16 x, u16 a, u16 b) {
- if (x < a) return a;
- if (x > b) return b;
- return x;
-}
-
-static inline i32 clamp_i32(i32 x, i32 a, i32 b) {
- if (x < a) return a;
- if (x > b) return b;
- return x;
-}
-
-static inline u32 clamp_u32(u32 x, u32 a, u32 b) {
- if (x < a) return a;
- if (x > b) return b;
- return x;
-}
-
-static inline u8 ndigits_u64(u64 x) {
- u8 ndigits = 1;
- while (x > 9) {
- x /= 10;
- ++ndigits;
- }
- return ndigits;
-}
-
-// remap x from the interval [from_a, from_b] to the interval [to_a, to_b], NOT clamping if x is outside the "from" interval.
-static inline float remapf(float x, float from_a, float from_b, float to_a, float to_b) {
- float pos = (x - from_a) / (from_b - from_a);
- return lerpf(pos, to_a, to_b);
-}
-
-static inline float minf(float a, float b) {
- return a < b ? a : b;
-}
-
-static inline float maxf(float a, float b) {
- return a > b ? a : b;
-}
-
-static inline double maxd(double a, double b) {
- return a > b ? a : b;
-}
-
-static inline double mind(double a, double b) {
- return a < b ? a : b;
-}
-
-static inline u32 min_u32(u32 a, u32 b) {
- return a < b ? a : b;
-}
-
-static inline u32 max_u32(u32 a, u32 b) {
- return a > b ? a : b;
-}
-
-// set *a to the minimum of *a and *b, and *b to the maximum
-static inline void sort2_u32(u32 *a, u32 *b) {
- u32 x = *a, y = *b;
- if (x > y) {
- *a = y;
- *b = x;
- }
-}
-
-static inline i32 min_i32(i32 a, i32 b) {
- return a < b ? a : b;
-}
-
-static inline i32 max_i32(i32 a, i32 b) {
- return a > b ? a : b;
-}
-
-static inline u64 min_u64(u64 a, u64 b) {
- return a < b ? a : b;
-}
-
-static inline u64 max_u64(u64 a, u64 b) {
- return a > b ? a : b;
-}
-
-static inline i64 min_i64(i64 a, i64 b) {
- return a < b ? a : b;
-}
-
-static inline i64 max_i64(i64 a, i64 b) {
- return a > b ? a : b;
-}
-
-static inline i64 mod_i64(i64 a, i64 b) {
- i64 ret = a % b;
- if (ret < 0) ret += b;
- return ret;
-}
-static inline i32 mod_i32(i32 a, i32 b) {
- i32 ret = a % b;
- if (ret < 0) ret += b;
- return ret;
-}
-
-static inline i64 abs_i64(i64 x) {
- return x < 0 ? -x : +x;
-}
-
-static inline i64 sgn_i64(i64 x) {
- if (x < 0) return -1;
- if (x > 0) return +1;
- return 0;
-}
-
-static inline float sgnf(float x) {
- if (x < 0) return -1;
- if (x > 0) return +1;
- return 0;
-}
-
-static inline float smoothstepf(float x) {
- if (x <= 0) return 0;
- if (x >= 1) return 1;
- return x * x * (3 - 2 * x);
-}
-
-static inline float randf(void) {
- return (float)rand() / (float)((ulong)RAND_MAX + 1);
-}
-
-static float rand_gauss(void) {
- // https://en.wikipedia.org/wiki/Normal_distribution#Generating_values_from_normal_distribution
- float U, V;
- do
- U = randf(), V = randf();
- while (U == 0 || V == 0);
- return sqrtf(-2 * logf(U)) * cosf(TAUf * V);
-}
-
-static u32 rand_u32(void) {
- return ((u32)rand() & 0xfff)
- | ((u32)rand() & 0xfff) << 12
- | ((u32)rand() & 0xff) << 24;
-}
-
-static float rand_uniform(float from, float to) {
- return lerpf(randf(), from, to);
-}
-
-static float sigmoidf(float x) {
- return 1.0f / (1.0f + expf(-x));
-}
-
-// returns ⌈x/y⌉ (x/y rounded up)
-static i32 ceildivi32(i32 x, i32 y) {
- if (y < 0) {
- // negating both operands doesn't change the answer
- x = -x;
- y = -y;
- }
- if (x < 0) {
- // truncation is the same as ceiling for negative numbers
- return x / y;
- } else {
- return (x + (y-1)) / y;
- }
-}
-
-typedef struct {
- float x, y;
-} v2;
-
-static v2 const v2_zero = {0, 0};
-static v2 V2(float x, float y) {
- v2 v;
- v.x = x;
- v.y = y;
- return v;
-}
-
-static v2 v2_add(v2 a, v2 b) {
- return V2(a.x + b.x, a.y + b.y);
-}
-
-static v2 v2_add_const(v2 a, float c) {
- return V2(a.x + c, a.y + c);
-}
-
-static v2 v2_sub(v2 a, v2 b) {
- return V2(a.x - b.x, a.y - b.y);
-}
-
-static v2 v2_scale(v2 v, float s) {
- return V2(v.x * s, v.y * s);
-}
-
-static v2 v2_mul(v2 a, v2 b) {
- return V2(a.x * b.x, a.y * b.y);
-}
-
-static v2 v2_clamp(v2 x, v2 a, v2 b) {
- return V2(clampf(x.x, a.x, b.x), clampf(x.y, a.y, b.y));
-}
-
-static float v2_dot(v2 a, v2 b) {
- return a.x * b.x + a.y * b.y;
-}
-
-static float v2_len(v2 v) {
- return sqrtf(v2_dot(v, v));
-}
-
-static v2 v2_lerp(float x, v2 a, v2 b) {
- return V2(lerpf(x, a.x, b.x), lerpf(x, a.y, b.y));
-}
-
-// rotate v theta radians counterclockwise
-static v2 v2_rotate(v2 v, float theta) {
- float c = cosf(theta), s = sinf(theta);
- return V2(
- c * v.x - s * v.y,
- s * v.x + c * v.y
- );
-}
-
-static v2 v2_normalize(v2 v) {
- float len = v2_len(v);
- float mul = len == 0.0f ? 1.0f : 1.0f/len;
- return v2_scale(v, mul);
-}
-
-static float v2_dist(v2 a, v2 b) {
- return v2_len(v2_sub(a, b));
-}
-
-static float v2_dist_squared(v2 a, v2 b) {
- v2 diff = v2_sub(a, b);
- return v2_dot(diff, diff);
-}
-
-static void v2_print(v2 v) {
- printf("(%f, %f)\n", v.x, v.y);
-}
-
-static v2 v2_rand_unit(void) {
- float theta = rand_uniform(0, TAUf);
- return V2(cosf(theta), sinf(theta));
-}
-
-static v2 v2_polar(float r, float theta) {
- return V2(r * cosf(theta), r * sinf(theta));
-}
-
-typedef struct {
- float x, y, z;
-} v3;
-
-static v3 const v3_zero = {0, 0, 0};
-
-static v3 V3(float x, float y, float z) {
- v3 v;
- v.x = x;
- v.y = y;
- v.z = z;
- return v;
-}
-
-static v3 v3_from_v2(v2 v) {
- return V3(v.x, v.y, 0);
-}
-
-static v3 v3_add(v3 a, v3 b) {
- return V3(a.x + b.x, a.y + b.y, a.z + b.z);
-}
-
-static v3 v3_sub(v3 a, v3 b) {
- return V3(a.x - b.x, a.y - b.y, a.z - b.z);
-}
-
-static v3 v3_scale(v3 v, float s) {
- return V3(v.x * s, v.y * s, v.z * s);
-}
-
-static v3 v3_lerp(float x, v3 a, v3 b) {
- return V3(lerpf(x, a.x, b.x), lerpf(x, a.y, b.y), lerpf(x, a.z, b.z));
-}
-
-static float v3_dot(v3 u, v3 v) {
- return u.x*v.x + u.y*v.y + u.z*v.z;
-}
-
-static v3 v3_cross(v3 u, v3 v) {
- v3 prod = V3(u.y*v.z - u.z*v.y, u.z*v.x - u.x*v.z, u.x*v.y - u.y*v.x);
- return prod;
-}
-
-static float v3_len(v3 v) {
- return sqrtf(v3_dot(v, v));
-}
-
-static float v3_dist(v3 a, v3 b) {
- return v3_len(v3_sub(a, b));
-}
-
-static float v3_dist_squared(v3 a, v3 b) {
- v3 diff = v3_sub(a, b);
- return v3_dot(diff, diff);
-}
-
-static v3 v3_normalize(v3 v) {
- float len = v3_len(v);
- float mul = len == 0.0f ? 1.0f : 1.0f/len;
- return v3_scale(v, mul);
-}
-
-static v2 v3_xy(v3 v) {
- return V2(v.x, v.y);
-}
-
-// a point on a unit sphere
-static v3 v3_on_sphere(float yaw, float pitch) {
- return V3(cosf(yaw) * cosf(pitch), sinf(pitch), sinf(yaw) * cosf(pitch));
-}
-
-static void v3_print(v3 v) {
- printf("(%f, %f, %f)\n", v.x, v.y, v.z);
-}
-
-static v3 v3_rand(void) {
- return V3(randf(), randf(), randf());
-}
-
-static v3 v3_rand_unit(void) {
- /*
- monte carlo method
- keep generating random points in cube of radius 1 (width 2) centered at origin,
- until you get a point in the unit sphere, then extend it to find the point lying
- on the sphere.
- */
- while (1) {
- v3 v = V3(rand_uniform(-1.0f, +1.0f), rand_uniform(-1.0f, +1.0f), rand_uniform(-1.0f, +1.0f));
- float dist_squared_to_origin = v3_dot(v, v);
- if (dist_squared_to_origin <= 1 && dist_squared_to_origin != 0.0f) {
- return v3_scale(v, 1.0f / sqrtf(dist_squared_to_origin));
- }
- }
- return V3(0, 0, 0);
-}
-
-typedef struct {
- float x, y, z, w;
-} v4;
-
-static v4 const v4_zero = {0, 0, 0, 0};
-
-static v4 V4(float x, float y, float z, float w) {
- v4 v;
- v.x = x;
- v.y = y;
- v.z = z;
- v.w = w;
- return v;
-}
-
-static v4 v4_add(v4 a, v4 b) {
- return V4(a.x + b.x, a.y + b.y, a.z + b.z, a.w + b.w);
-}
-
-static v4 v4_sub(v4 a, v4 b) {
- return V4(a.x - b.x, a.y - b.y, a.z - b.z, a.w - b.w);
-}
-
-static v4 v4_scale(v4 v, float s) {
- return V4(v.x * s, v.y * s, v.z * s, v.w * s);
-}
-
-static v4 v4_scale_xyz(v4 v, float s) {
- return V4(v.x * s, v.y * s, v.z * s, v.w);
-}
-
-static v4 v4_lerp(float x, v4 a, v4 b) {
- return V4(lerpf(x, a.x, b.x), lerpf(x, a.y, b.y), lerpf(x, a.z, b.z), lerpf(x, a.w, b.w));
-}
-
-static float v4_dot(v4 u, v4 v) {
- return u.x*v.x + u.y*v.y + u.z*v.z + u.w*v.w;
-}
-
-// create a new vector by multiplying the respective components of u and v
-static v4 v4_mul(v4 u, v4 v) {
- return V4(u.x * v.x, u.y * v.y, u.z * v.z, u.w * v.w);
-}
-
-static float v4_len(v4 v) {
- return sqrtf(v4_dot(v, v));
-}
-
-static v4 v4_normalize(v4 v) {
- float len = v4_len(v);
- float mul = len == 0.0f ? 1.0f : 1.0f/len;
- return v4_scale(v, mul);
-}
-
-static v3 v4_xyz(v4 v) {
- return V3(v.x, v.y, v.z);
-}
-
-static v4 v4_rand(void) {
- return V4(randf(), randf(), randf(), randf());
-}
-
-static void v4_print(v4 v) {
- printf("(%f, %f, %f, %f)\n", v.x, v.y, v.z, v.w);
-}
-
-typedef struct {
- double x, y;
-} v2d;
-
-static v2d V2D(double x, double y) {
- v2d v;
- v.x = x;
- v.y = y;
- return v;
-}
-
-// matrices are column-major, because that's what they are in OpenGL
-typedef struct {
- float e[16];
-} m4;
-
-static m4 const m4_identity = {{
- 1, 0, 0, 0,
- 0, 1, 0, 0,
- 0, 0, 1, 0,
- 0, 0, 0, 1
-}};
-
-static void m4_print(m4 m) {
- int i;
- for (i = 0; i < 4; ++i)
- printf("[ %f %f %f %f ]\n", m.e[i], m.e[i+4], m.e[i+8], m.e[i+12]);
- printf("\n");
-}
-
-static m4 M4(
- float a, float b, float c, float d,
- float e, float f, float g, float h,
- float i, float j, float k, float l,
- float m, float n, float o, float p) {
- m4 ret;
- float *x = ret.e;
- x[0] = a; x[4] = b; x[ 8] = c; x[12] = d;
- x[1] = e; x[5] = f; x[ 9] = g; x[13] = h;
- x[2] = i; x[6] = j; x[10] = k; x[14] = l;
- x[3] = m; x[7] = n; x[11] = o; x[15] = p;
- return ret;
-}
-
-// see https://en.wikipedia.org/wiki/Rotation_matrix#General_rotations
-static m4 m4_yaw(float yaw) {
- float c = cosf(yaw), s = sinf(yaw);
- return M4(
- c, 0, -s, 0,
- 0, 1, 0, 0,
- s, 0, c, 0,
- 0, 0, 0, 1
- );
-}
-
-static m4 m4_pitch(float pitch) {
- float c = cosf(pitch), s = sinf(pitch);
- return M4(
- 1, 0, 0, 0,
- 0, c, -s, 0,
- 0, s, c, 0,
- 0, 0, 0, 1
- );
-}
-
-// https://en.wikipedia.org/wiki/Translation_(geometry)
-static m4 m4_translate(v3 t) {
- return M4(
- 1, 0, 0, t.x,
- 0, 1, 0, t.y,
- 0, 0, 1, t.z,
- 0, 0, 0, 1
- );
-}
-
-// multiply m by [v.x, v.y, v.z, 1]
-static v3 m4_mul_v3(m4 m, v3 v) {
- return v3_add(v3_scale(V3(m.e[0], m.e[1], m.e[2]), v.x), v3_add(v3_scale(V3(m.e[4], m.e[5], m.e[6]), v.y),
- v3_add(v3_scale(V3(m.e[8], m.e[9], m.e[10]), v.z), V3(m.e[12], m.e[13], m.e[14]))));
-}
-
-/*
-4x4 perspective matrix.
-fov - field of view in radians, aspect - width:height aspect ratio, z_near/z_far - clipping planes
-math stolen from gluPerspective (https://www.khronos.org/registry/OpenGL-Refpages/gl2.1/xhtml/gluPerspective.xml)
-*/
-static m4 m4_perspective(float fov, float aspect, float z_near, float z_far) {
- float f = 1.0f / tanf(fov / 2.0f);
- return M4(
- f/aspect, 0, 0, 0,
- 0, f, 0, 0,
- 0, 0, (z_far+z_near) / (z_near-z_far), (2.0f*z_far*z_near) / (z_near-z_far),
- 0, 0, -1, 0
- );
-}
-
-// windows.h defines near and far, so let's not use those
-static m4 m4_ortho(float left, float right, float bottom, float top, float near_, float far_) {
- float tx = -(right + left)/(right - left);
- float ty = -(top + bottom)/(top - bottom);
- float tz = -(far_ + near_)/(far_ - near_);
- return M4(
- 2.0f / (right - left), 0, 0, tx,
- 0, 2.0f / (top - bottom), 0, ty,
- 0, 0, -2.0f / (far_ - near_), tz,
- 0, 0, 0, 1
- );
-}
-
-
-static m4 m4_mul(m4 a, m4 b) {
- m4 prod = {0};
- int i, j;
- float *x = prod.e;
- for (i = 0; i < 4; ++i) {
- for (j = 0; j < 4; ++j, ++x) {
- float *as = &a.e[j];
- float *bs = &b.e[4*i];
- *x = as[0]*bs[0] + as[4]*bs[1] + as[8]*bs[2] + as[12]*bs[3];
- }
- }
- return prod;
-}
-
-static m4 m4_inv(m4 mat) {
- m4 ret;
- float *inv = ret.e;
- float *m = mat.e;
-
- inv[0] = m[5] * m[10] * m[15] - m[5] * m[11] * m[14] - m[9] * m[6] * m[15] + m[9] * m[7] * m[14] + m[13] * m[6] * m[11] - m[13] * m[7] * m[10];
- inv[4] = -m[4] * m[10] * m[15] + m[4] * m[11] * m[14] + m[8] * m[6] * m[15] - m[8] * m[7] * m[14] - m[12] * m[6] * m[11] + m[12] * m[7] * m[10];
- inv[8] = m[4] * m[9] * m[15] - m[4] * m[11] * m[13] - m[8] * m[5] * m[15] + m[8] * m[7] * m[13] + m[12] * m[5] * m[11] - m[12] * m[7] * m[9];
- inv[12] = -m[4] * m[9] * m[14] + m[4] * m[10] * m[13] + m[8] * m[5] * m[14] - m[8] * m[6] * m[13] - m[12] * m[5] * m[10] + m[12] * m[6] * m[9];
- inv[1] = -m[1] * m[10] * m[15] + m[1] * m[11] * m[14] + m[9] * m[2] * m[15] - m[9] * m[3] * m[14] - m[13] * m[2] * m[11] + m[13] * m[3] * m[10];
- inv[5] = m[0] * m[10] * m[15] - m[0] * m[11] * m[14] - m[8] * m[2] * m[15] + m[8] * m[3] * m[14] + m[12] * m[2] * m[11] - m[12] * m[3] * m[10];
- inv[9] = -m[0] * m[9] * m[15] + m[0] * m[11] * m[13] + m[8] * m[1] * m[15] - m[8] * m[3] * m[13] - m[12] * m[1] * m[11] + m[12] * m[3] * m[9];
- inv[13] = m[0] * m[9] * m[14] - m[0] * m[10] * m[13] - m[8] * m[1] * m[14] + m[8] * m[2] * m[13] + m[12] * m[1] * m[10] - m[12] * m[2] * m[9];
- inv[2] = m[1] * m[6] * m[15] - m[1] * m[7] * m[14] - m[5] * m[2] * m[15] + m[5] * m[3] * m[14] + m[13] * m[2] * m[7] - m[13] * m[3] * m[6];
- inv[6] = -m[0] * m[6] * m[15] + m[0] * m[7] * m[14] + m[4] * m[2] * m[15] - m[4] * m[3] * m[14] - m[12] * m[2] * m[7] + m[12] * m[3] * m[6];
- inv[10] = m[0] * m[5] * m[15] - m[0] * m[7] * m[13] - m[4] * m[1] * m[15] + m[4] * m[3] * m[13] + m[12] * m[1] * m[7] - m[12] * m[3] * m[5];
- inv[14] = -m[0] * m[5] * m[14] + m[0] * m[6] * m[13] + m[4] * m[1] * m[14] - m[4] * m[2] * m[13] - m[12] * m[1] * m[6] + m[12] * m[2] * m[5];
- inv[3] = -m[1] * m[6] * m[11] + m[1] * m[7] * m[10] + m[5] * m[2] * m[11] - m[5] * m[3] * m[10] - m[9] * m[2] * m[7] + m[9] * m[3] * m[6];
- inv[7] = m[0] * m[6] * m[11] - m[0] * m[7] * m[10] - m[4] * m[2] * m[11] + m[4] * m[3] * m[10] + m[8] * m[2] * m[7] - m[8] * m[3] * m[6];
- inv[11] = -m[0] * m[5] * m[11] + m[0] * m[7] * m[9] + m[4] * m[1] * m[11] - m[4] * m[3] * m[9] - m[8] * m[1] * m[7] + m[8] * m[3] * m[5];
- inv[15] = m[0] * m[5] * m[10] - m[0] * m[6] * m[9] - m[4] * m[1] * m[10] + m[4] * m[2] * m[9] + m[8] * m[1] * m[6] - m[8] * m[2] * m[5];
-
- float det = m[0] * inv[0] + m[1] * inv[4] + m[2] * inv[8] + m[3] * inv[12];
-
- if (det == 0) {
- memset(inv, 0, sizeof *inv);
- } else {
- det = 1 / det;
-
- for (int i = 0; i < 16; i++)
- inv[i] *= det;
- }
-
- return ret;
-}
-typedef struct {
- int x, y;
-} v2i;
-
-static v2i V2I(int x, int y) {
- v2i v;
- v.x = x;
- v.y = y;
- return v;
-}
-
-static void rgba_u32_to_floats(u32 rgba, float floats[4]) {
- floats[0] = (float)((rgba >> 24) & 0xFF) / 255.f;
- floats[1] = (float)((rgba >> 16) & 0xFF) / 255.f;
- floats[2] = (float)((rgba >> 8) & 0xFF) / 255.f;
- floats[3] = (float)((rgba >> 0) & 0xFF) / 255.f;
-}
-
-static v4 rgba_u32_to_v4(u32 rgba) {
- float c[4];
- rgba_u32_to_floats(rgba, c);
- return V4(c[0], c[1], c[2], c[3]);
-}
-
-static u32 rgba_v4_to_u32(v4 color) {
- return (u32)(color.x * 255) << 24
- | (u32)(color.y * 255) << 16
- | (u32)(color.z * 255) << 8
- | (u32)(color.w * 255);
-}
-
-// returns average of red green and blue components of color
-static float rgba_brightness(u32 color) {
- u8 r = (u8)(color >> 24), g = (u8)(color >> 16), b = (u8)(color >> 8);
- return ((float)r+(float)g+(float)b) * (1.0f / 3);
-}
-
-static bool rect_contains_point_v2(v2 pos, v2 size, v2 point) {
- float x1 = pos.x, y1 = pos.y, x2 = pos.x + size.x, y2 = pos.y + size.y,
- x = point.x, y = point.y;
- return x >= x1 && x < x2 && y >= y1 && y < y2;
-}
-
-static bool centered_rect_contains_point(v2 center, v2 size, v2 point) {
- return rect_contains_point_v2(v2_sub(center, v2_scale(size, 0.5f)), size, point);
-}
-
-typedef struct {
- v2 pos, size;
-} Rect;
-
-static Rect rect(v2 pos, v2 size) {
- Rect r;
- r.pos = pos;
- r.size = size;
- return r;
-}
-
-static Rect rect_endpoints(v2 e1, v2 e2) {
- Rect r;
- r.pos = e1;
- r.size = v2_sub(e2, e1);
- return r;
-}
-
-static Rect rect4(float x1, float y1, float x2, float y2) {
- assert(x2 >= x1);
- assert(y2 >= y1);
- return rect(V2(x1,y1), V2(x2-x1, y2-y1));
-}
-
-static Rect rect_xywh(float x, float y, float w, float h) {
- assert(w >= 0);
- assert(h >= 0);
- return rect(V2(x, y), V2(w, h));
-}
-
-static Rect rect_centered(v2 center, v2 size) {
- Rect r;
- r.pos = v2_sub(center, v2_scale(size, 0.5f));
- r.size = size;
- return r;
-}
-
-static v2 rect_center(Rect r) {
- return v2_add(r.pos, v2_scale(r.size, 0.5f));
-}
-
-static bool rect_contains_point(Rect r, v2 point) {
- return rect_contains_point_v2(r.pos, r.size, point);
-}
-
-static Rect rect_translate(Rect r, v2 by) {
- return rect(v2_add(r.pos, by), r.size);
-}
-
-static float rect_x1(Rect r) { return r.pos.x; }
-static float rect_y1(Rect r) { return r.pos.y; }
-static float rect_x2(Rect r) { return r.pos.x + r.size.x; }
-static float rect_y2(Rect r) { return r.pos.y + r.size.y; }
-static float rect_xmid(Rect r) { return r.pos.x + r.size.x * 0.5f; }
-static float rect_ymid(Rect r) { return r.pos.y + r.size.y * 0.5f; }
-
-static void rect_coords(Rect r, float *x1, float *y1, float *x2, float *y2) {
- *x1 = r.pos.x;
- *y1 = r.pos.y;
- *x2 = r.pos.x + r.size.x;
- *y2 = r.pos.y + r.size.y;
-}
-
-static void rect_print(Rect r) {
- printf("Position: (%f, %f), Size: (%f, %f)\n", r.pos.x, r.pos.y, r.size.x, r.size.y);
-}
-
-
-static float rects_intersect(Rect r1, Rect r2) {
- if (r1.pos.x >= r2.pos.x + r2.size.x) return false; // r1 is to the right of r2
- if (r2.pos.x >= r1.pos.x + r1.size.x) return false; // r2 is to the right of r1
- if (r1.pos.y >= r2.pos.y + r2.size.y) return false; // r1 is above r2
- if (r2.pos.y >= r1.pos.y + r1.size.y) return false; // r2 is above r1
- return true;
-}
-
-// returns whether or not there is any of the clipped rectangle left
-static bool rect_clip_to_rect(Rect *clipped, Rect clipper) {
- v2 start_pos = clipped->pos;
- clipped->pos.x = maxf(clipped->pos.x, clipper.pos.x);
- clipped->pos.y = maxf(clipped->pos.y, clipper.pos.y);
- clipped->size = v2_add(clipped->size, v2_sub(start_pos, clipped->pos));
-
- clipped->size.x = clampf(clipped->size.x, 0, clipper.pos.x + clipper.size.x - clipped->pos.x);
- clipped->size.y = clampf(clipped->size.y, 0, clipper.pos.y + clipper.size.y - clipped->pos.y);
- return clipped->size.x > 0 && clipped->size.y > 0;
-}
-
-// removes `amount` from all sides of r
-static Rect rect_shrink(Rect r, float amount) {
- r.pos.x += amount;
- r.pos.y += amount;
- r.size.x -= 2 * amount;
- r.size.y -= 2 * amount;
- r.size.x = maxf(r.size.x, 0);
- r.size.y = maxf(r.size.y, 0);
- return r;
-}
-
-// adds `amount` to all sides of r
-static Rect rect_grow(Rect r, float amount) {
- r.pos.x -= amount;
- r.pos.y -= amount;
- r.size.x += 2 * amount;
- r.size.y += 2 * amount;
- return r;
-}
-
-static v4 color_rgba_to_hsva(v4 rgba) {
- float R = rgba.x, G = rgba.y, B = rgba.z, A = rgba.w;
- float M = maxf(R, maxf(G, B));
- float m = minf(R, minf(G, B));
- float C = M - m;
- float H = 0;
- if (C == 0)
- H = 0;
- else if (M == R)
- H = fmodf((G - B) / C, 6);
- else if (M == G)
- H = (B - R) / C + 2;
- else if (M == B)
- H = (R - G) / C + 4;
- H *= 60;
- float V = M;
- float S = V == 0 ? 0 : C / V;
- return V4(H, S, V, A);
-}
-
-static v4 color_hsva_to_rgba(v4 hsva) {
- float H = hsva.x, S = hsva.y, V = hsva.z, A = hsva.w;
- H /= 60;
- float C = S * V;
- float X = C * (1 - fabsf(fmodf(H, 2) - 1));
- float R, G, B;
- if (H <= 1)
- R=C, G=X, B=0;
- else if (H <= 2)
- R=X, G=C, B=0;
- else if (H <= 3)
- R=0, G=C, B=X;
- else if (H <= 4)
- R=0, G=X, B=C;
- else if (H <= 5)
- R=X, G=0, B=C;
- else
- R=C, G=0, B=X;
-
- float m = V-C;
- R += m;
- G += m;
- B += m;
- return V4(R, G, B, A);
-}
-
-static u32 color_interpolate(float x, u32 color1, u32 color2) {
- x = x * x * (3 - 2*x); // hermite interpolation
-
- v4 c1 = rgba_u32_to_v4(color1), c2 = rgba_u32_to_v4(color2);
- // to make it interpolate more nicely, convert to hsv, interpolate in that space, then convert back
- c1 = color_rgba_to_hsva(c1);
- c2 = color_rgba_to_hsva(c2);
- // v_1/2 named differently to avoid shadowing
- float h1 = c1.x, s1 = c1.y, v_1 = c1.z, a1 = c1.w;
- float h2 = c2.x, s2 = c2.y, v_2 = c2.z, a2 = c2.w;
-
- float s_out = lerpf(x, s1, s2);
- float v_out = lerpf(x, v_1, v_2);
- float a_out = lerpf(x, a1, a2);
-
- float h_out;
- // because hue is on a circle, we need to make sure we take the shorter route around the circle
- if (fabsf(h1 - h2) < 180) {
- h_out = lerpf(x, h1, h2);
- } else if (h1 > h2) {
- h_out = lerpf(x, h1, h2 + 360);
- } else {
- h_out = lerpf(x, h1 + 360, h2);
- }
- h_out = fmodf(h_out, 360);
-
- v4 c_out = V4(h_out, s_out, v_out, a_out);
- c_out = color_hsva_to_rgba(c_out);
- return rgba_v4_to_u32(c_out);
-}