Reworking the build system

1 files changed, 780 insertions, 0 deletions

diff --git a/kit/math.h b/kit/math.h
new file mode 100644
index 0000000..d7044f1
--- /dev/null
+++ b/kit/math.h
@@ -0,0 +1,780 @@
+#ifndef KIT_MATH_H
+#define KIT_MATH_H
+
+#include "types.h"
+
+#ifndef _USE_MATH_DEFINES
+#  define _USE_MATH_DEFINES
+#endif
+#include <math.h>
+
+#include <assert.h>
+#include <string.h>
+
+//  TODO
+//  - Gamma correction.
+//  - Custom prefixes
+//  - Unnecesary XYZ scaling.
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#ifndef M_PI
+#  define M_PI 3.14159265358979323846
+#endif
+
+#ifndef M_E
+#  define M_E 2.71828182845904523536
+#endif
+
+#ifndef KIT_VEC_TYPE
+#  define KIT_VEC_TYPE f32
+#endif
+
+#define KIT_EPSILON .00001
+
+#define KIT_COLOR_REF_X 95.047f
+#define KIT_COLOR_REF_Y 100.0f
+#define KIT_COLOR_REF_Z 108.883f
+
+typedef KIT_VEC_TYPE vec_t;
+
+typedef struct {
+  vec_t v[2];
+} vec2_t;
+
+typedef struct {
+  vec_t v[3];
+} vec3_t;
+
+typedef struct {
+  vec_t v[4];
+} vec4_t, quat_t;
+
+typedef struct {
+  vec_t v[9];
+} mat3_t;
+
+typedef struct {
+  vec_t v[16];
+} mat4_t;
+
+#ifdef __GNUC__
+#  pragma GCC diagnostic push
+#  pragma GCC diagnostic ignored "-Wunused-function"
+#  pragma GCC diagnostic ignored "-Wunknown-pragmas"
+#  pragma GCC            push_options
+#  pragma GCC            optimize("O3")
+#endif
+
+static vec2_t vec2(vec_t x, vec_t y) {
+  vec2_t v = { { x, y } };
+  return v;
+}
+
+static vec3_t vec3(vec_t x, vec_t y, vec_t z) {
+  vec3_t v = { { x, y, z } };
+  return v;
+}
+
+static vec4_t vec4(vec_t x, vec_t y, vec_t z, vec_t w) {
+  vec4_t v = { { x, y, z, w } };
+  return v;
+}
+
+static mat3_t mat3(                  //
+    vec_t _00, vec_t _01, vec_t _02, //
+    vec_t _10, vec_t _11, vec_t _12, //
+    vec_t _20, vec_t _21, vec_t _22) {
+  mat3_t m = { {
+      _00, _01, _02, //
+      _10, _11, _12, //
+      _20, _21, _22  //
+  } };
+  return m;
+}
+
+static mat4_t mat4(vec_t _00, vec_t _01, vec_t _02, vec_t _03,
+                   vec_t _10, vec_t _11, vec_t _12, vec_t _13,
+                   vec_t _20, vec_t _21, vec_t _22, vec_t _23,
+                   vec_t _30, vec_t _31, vec_t _32, vec_t _33) {
+  mat4_t m = { {
+      _00, _01, _02, _03, //
+      _10, _11, _12, _13, //
+      _20, _21, _22, _23, //
+      _30, _31, _32, _33  //
+  } };
+  return m;
+}
+
+static mat3_t vec3_to_mat3(vec3_t a, vec3_t b, vec3_t c) {
+  return mat3(a.v[0], a.v[1], a.v[2], //
+              b.v[0], b.v[1], b.v[2], //
+              c.v[0], c.v[1], c.v[2]);
+}
+
+static mat4_t vec4_to_mat4(vec4_t a, vec4_t b, vec4_t c, vec4_t d) {
+  return mat4(a.v[0], a.v[1], a.v[2], a.v[3], //
+              b.v[0], b.v[1], b.v[2], b.v[3], //
+              c.v[0], c.v[1], c.v[2], c.v[3], //
+              d.v[0], d.v[1], d.v[2], d.v[3]);
+}
+
+static mat4_t mat3_to_mat4(mat3_t a) {
+  return mat4(a.v[0], a.v[1], a.v[2], 0.f, //
+              a.v[3], a.v[4], a.v[5], 0.f, //
+              a.v[6], a.v[7], a.v[8], 0.f, //
+              0.f, 0.f, 0.f, 1.f);
+}
+
+static mat3_t mat4_to_mat3(mat4_t a) {
+  return mat3(a.v[0], a.v[1], a.v[2], //
+              a.v[4], a.v[5], a.v[6], //
+              a.v[8], a.v[9], a.v[10]);
+}
+
+static quat_t quat(vec_t x, vec_t y, vec_t z, vec_t w) {
+  return vec4(x, y, z, w);
+}
+
+static quat_t vec3_to_quat(vec3_t v, vec_t w) {
+  return quat(v.v[0], v.v[1], v.v[2], w);
+}
+
+static vec3_t vec4_to_vec3(vec4_t a) {
+  return vec3(a.v[0], a.v[1], a.v[2]);
+}
+
+static vec3_t quat_to_vec3(quat_t a) {
+  return vec4_to_vec3(a);
+}
+
+static vec_t vec_abs(vec_t x) {
+  return fabs(x);
+}
+
+static vec_t vec_min(vec_t x, vec_t y) {
+  return fmin(x, y);
+}
+
+static vec_t vec_max(vec_t x, vec_t y) {
+  return fmax(x, y);
+}
+
+static vec_t vec_clamp(vec_t x, vec_t x0, vec_t x1) {
+  return fmax(x0, fmin(x, x1));
+}
+
+static vec3_t vec3_clamp(vec3_t x, vec3_t x0, vec3_t x1) {
+  return vec3(vec_clamp(x.v[0], x0.v[0], x1.v[0]),
+              vec_clamp(x.v[1], x0.v[1], x1.v[1]),
+              vec_clamp(x.v[2], x0.v[2], x1.v[2]));
+}
+
+static vec_t vec_sqrt(vec_t x) {
+  return sqrtf(x);
+}
+
+static vec_t vec_pow(vec_t x, vec_t y) {
+  return powf(x, y);
+}
+
+static vec_t vec_sin(vec_t x) {
+  return sinf(x);
+}
+
+static vec_t vec_cos(vec_t x) {
+  return cosf(x);
+}
+
+static vec_t vec_asin(vec_t x) {
+  return asinf(x);
+}
+
+static vec_t vec_acos(vec_t x) {
+  return acosf(x);
+}
+
+static vec_t vec_atan(vec_t x) {
+  return atanf(x);
+}
+
+static vec_t vec_atan2(vec_t y, vec_t x) {
+  return atan2f(y, x);
+}
+
+static vec3_t vec3_neg(vec3_t a) {
+  return vec3(-a.v[0], -a.v[1], -a.v[2]);
+}
+
+static vec3_t vec3_add(vec3_t a, vec3_t b) {
+  return vec3(a.v[0] + b.v[0], a.v[1] + b.v[1], a.v[2] + b.v[2]);
+}
+
+static vec3_t vec3_add3(vec3_t a, vec3_t b, vec3_t c) {
+  return vec3(a.v[0] + b.v[0] + c.v[0], a.v[1] + b.v[1] + c.v[1],
+              a.v[2] + b.v[2] + c.v[2]);
+}
+
+static vec3_t vec3_sub(vec3_t a, vec3_t b) {
+  return vec3(a.v[0] - b.v[0], a.v[1] - b.v[1], a.v[2] - b.v[2]);
+}
+
+static vec3_t vec3_mul(vec3_t v, vec_t x) {
+  return vec3(v.v[0] * x, v.v[1] * x, v.v[2] * x);
+}
+
+static vec_t vec3_dot(vec3_t a, vec3_t b) {
+  return a.v[0] * b.v[0] + a.v[1] * b.v[1] + a.v[2] * b.v[2];
+}
+
+static vec3_t vec3_cross(vec3_t a, vec3_t b) {
+  return vec3(a.v[1] * b.v[2] - a.v[2] * b.v[1], //
+              a.v[2] * b.v[0] - a.v[0] * b.v[2], //
+              a.v[0] * b.v[1] - a.v[1] * b.v[0]);
+}
+
+static vec3_t vec3_normal(vec3_t v) {
+  vec_t x = v.v[0];
+  vec_t y = v.v[1];
+  vec_t z = v.v[2];
+
+  vec_t length_squared = x * x + y * y + z * z;
+  assert(length_squared >= KIT_EPSILON);
+
+  if (length_squared < KIT_EPSILON) {
+    vec3_t n = { { 0.f, 0.f, 1.f } };
+    return n;
+  }
+
+  vec_t  length = vec_sqrt(length_squared);
+  vec3_t n      = { { x / length, y / length, z / length } };
+
+  return n;
+}
+
+static vec4_t vec4_neg(vec4_t a) {
+  return vec4(-a.v[0], -a.v[1], -a.v[2], -a.v[3]);
+}
+
+static vec4_t vec4_add(vec4_t a, vec4_t b) {
+  return vec4(a.v[0] + b.v[0], a.v[1] + b.v[1], a.v[2] + b.v[2],
+              a.v[3] + b.v[3]);
+}
+
+static vec4_t vec4_mul(vec4_t v, vec_t x) {
+  return vec4(v.v[0] * x, v.v[1] * x, v.v[2] * x, v.v[3] * x);
+}
+
+static vec4_t vec4_div(vec4_t v, vec_t x) {
+  assert(x < KIT_EPSILON);
+  assert(x > -KIT_EPSILON);
+
+  if (x >= -KIT_EPSILON || x < KIT_EPSILON)
+    return vec4(0.f, 0.f, 0.f, 0.f);
+
+  return vec4(v.v[0] / x, v.v[1] / x, v.v[2] / x, v.v[3] / x);
+}
+
+static vec_t vec4_dot(vec4_t a, vec4_t b) {
+  return a.v[0] * b.v[0] + a.v[1] * b.v[1] + a.v[2] * b.v[2] +
+         a.v[3] * b.v[3];
+}
+
+static vec4_t vec4_normal(vec4_t v) {
+  vec_t x = v.v[0];
+  vec_t y = v.v[1];
+  vec_t z = v.v[2];
+  vec_t w = v.v[3];
+
+  vec_t length_squared = x * x + y * y + z * z + w * w;
+  assert(length_squared >= KIT_EPSILON);
+
+  if (length_squared < KIT_EPSILON) {
+    vec4_t n = { { 0.f, 0.f, 0.f, 1.f } };
+    return n;
+  }
+
+  vec_t  length = vec_sqrt(length_squared);
+  vec4_t n = { { x / length, y / length, z / length, w / length } };
+
+  return n;
+}
+
+static quat_t quat_conjugate(quat_t q) {
+  return quat(-q.v[0], -q.v[1], -q.v[2], q.v[3]);
+}
+
+static quat_t quat_normal(quat_t q) {
+  return vec4_normal(q);
+}
+
+static quat_t quat_mul(quat_t left, quat_t right) {
+  vec_t i0 = left.v[0];
+  vec_t j0 = left.v[1];
+  vec_t k0 = left.v[2];
+  vec_t r0 = left.v[3];
+
+  vec_t i1 = right.v[0];
+  vec_t j1 = right.v[1];
+  vec_t k1 = right.v[2];
+  vec_t r1 = right.v[3];
+
+  quat_t q = { { r0 * i1 + i0 * r1 + j0 * k1 - k0 * j1, //
+                 r0 * j1 - i0 * k1 + j0 * r1 + k0 * i1, //
+                 r0 * k1 + i0 * j1 - j0 * i1 + k0 * r1, //
+                 r0 * r1 - i0 * i1 - j0 * j1 - k0 * k1 } };
+
+  return q;
+}
+
+static quat_t quat_rotation(vec_t angle, vec3_t axis) {
+  vec_t s = vec_sin(angle / 2.f);
+  vec_t c = vec_cos(angle / 2.f);
+
+  vec_t x = axis.v[0];
+  vec_t y = axis.v[1];
+  vec_t z = axis.v[2];
+
+  quat_t q = { { x * s, y * s, z * s, c } };
+
+  return q;
+}
+
+static mat3_t mat3_transpose(mat3_t m) {
+  return mat3(m.v[0], m.v[3], m.v[6], //
+              m.v[1], m.v[4], m.v[7], //
+              m.v[2], m.v[5], m.v[8]);
+}
+
+static mat4_t mat4_transpose(mat4_t m) {
+  return mat4(m.v[0], m.v[4], m.v[8], m.v[12],  //
+              m.v[1], m.v[5], m.v[9], m.v[13],  //
+              m.v[2], m.v[6], m.v[10], m.v[14], //
+              m.v[3], m.v[7], m.v[11], m.v[15]);
+}
+
+static mat3_t mat3_look_at(vec3_t direction, vec3_t up) {
+  vec3_t f = vec3_normal(vec3_neg(direction));
+  vec3_t s = vec3_normal(vec3_cross(f, up));
+  vec3_t u = vec3_cross(s, f);
+
+  return mat3(s.v[0], u.v[0], f.v[0], //
+              s.v[1], u.v[1], f.v[1], //
+              s.v[2], u.v[2], f.v[2]);
+}
+
+static vec3_t vec3_rotate(vec3_t v, quat_t rotation) {
+  vec3_t u = vec3(rotation.v[0], rotation.v[1], rotation.v[2]);
+  vec_t  s = rotation.v[3];
+
+  return vec3_add3(vec3_mul(u, 2.f * vec3_dot(u, v)),
+                   vec3_mul(v, s * s - vec3_dot(u, u)),
+                   vec3_mul(vec3_cross(u, v), 2.f * s));
+}
+
+static mat3_t quat_to_mat3(quat_t q) {
+  vec_t xx = q.v[0] * q.v[0];
+  vec_t yy = q.v[1] * q.v[1];
+  vec_t zz = q.v[2] * q.v[2];
+  vec_t xz = q.v[0] * q.v[2];
+  vec_t xy = q.v[0] * q.v[1];
+  vec_t yz = q.v[1] * q.v[2];
+  vec_t wx = q.v[3] * q.v[0];
+  vec_t wy = q.v[3] * q.v[1];
+  vec_t wz = q.v[3] * q.v[2];
+
+  return mat3(1.f - 2.f * (yy + zz), //
+              2.f * (xy + wz),       //
+              2.f * (xz - wy),       //
+              2.f * (xy - wz),       //
+              1.f - 2.f * (xx + zz), //
+              2.f * (yz + wx),       //
+              2.f * (xz + wy),       //
+              2.f * (yz - wx),       //
+              1.f - 2.f * (xx + yy));
+}
+
+static quat_t mat3_to_quat(mat3_t m) {
+  vec_t a = m.v[0] - m.v[4] - m.v[8];
+  vec_t b = m.v[4] - m.v[0] - m.v[8];
+  vec_t c = m.v[8] - m.v[0] - m.v[4];
+  vec_t d = m.v[0] + m.v[4] + m.v[8];
+
+  int   n = 0;
+  vec_t h = d;
+  if (a > h) {
+    h = a;
+    n = 1;
+  }
+  if (b > h) {
+    h = b;
+    n = 2;
+  }
+  if (c > h) {
+    h = c;
+    n = 3;
+  }
+
+  vec_t s = vec_sqrt(h + 1.f) * .5f;
+  vec_t k = .25f / s;
+
+  switch (n) {
+    case 0:
+      return quat((m.v[5] - m.v[7]) * k, //
+                  (m.v[6] - m.v[2]) * k, //
+                  (m.v[1] - m.v[3]) * k, //
+                  s);
+    case 1:
+      return quat(s,                     //
+                  (m.v[1] + m.v[3]) * k, //
+                  (m.v[6] + m.v[2]) * k, //
+                  (m.v[5] - m.v[7]) * k);
+    case 2:
+      return quat((m.v[1] + m.v[3]) * k, //
+                  s,                     //
+                  (m.v[5] + m.v[7]) * k, //
+                  (m.v[6] - m.v[2]) * k);
+    case 3:
+      return quat((m.v[6] + m.v[2]) * k, //
+                  (m.v[5] + m.v[7]) * k, //
+                  s,                     //
+                  (m.v[1] - m.v[3]) * k);
+    default:;
+  }
+
+  assert(0);
+  return quat(0.f, 0.f, 0.f, 1.f);
+}
+
+static quat_t quat_look_at(vec3_t direction, vec3_t up) {
+  vec3_t z     = vec3_normal(vec3_neg(direction));
+  vec3_t right = vec3_cross(up, z);
+  vec3_t x     = vec3_mul(
+          right,
+          1.f / vec_sqrt(vec_max(KIT_EPSILON, vec3_dot(right, right))));
+  vec3_t y = vec3_cross(z, x);
+
+  return mat3_to_quat(vec3_to_mat3(x, y, z));
+}
+
+static mat3_t mat3_mul(mat3_t left, mat3_t right) {
+  mat3_t m;
+  memset(&m, 0, sizeof m);
+
+  for (int j = 0; j < 3; j++)
+    for (int i = 0; i < 3; i++)
+      for (int k = 0; k < 3; k++)
+        m.v[j * 3 + i] += left.v[k * 3 + i] * right.v[j * 3 + k];
+
+  return m;
+}
+
+static mat4_t mat4_mul(mat4_t left, mat4_t right) {
+  mat4_t m;
+  memset(&m, 0, sizeof m);
+
+  for (int j = 0; j < 4; j++)
+    for (int i = 0; i < 4; i++)
+      for (int k = 0; k < 4; k++)
+        m.v[j * 4 + i] += left.v[k * 4 + i] * right.v[j * 4 + k];
+
+  return m;
+}
+
+static mat4_t mat4_move(vec3_t offset) {
+  vec_t x = offset.v[0];
+  vec_t y = offset.v[1];
+  vec_t z = offset.v[2];
+
+  return mat4(1.f, 0.f, 0.f, 0.f, //
+              0.f, 1.f, 0.f, 0.f, //
+              0.f, 0.f, 1.f, 0.f, //
+              x, y, z, 1.f);
+}
+
+static mat4_t mat4_scale(vec3_t scale) {
+  vec_t x = scale.v[0];
+  vec_t y = scale.v[1];
+  vec_t z = scale.v[2];
+
+  return mat4(x, 0.f, 0.f, 0.f, //
+              0.f, y, 0.f, 0.f, //
+              0.f, 0.f, z, 0.f, //
+              0.f, 0.f, 0.f, 1.f);
+}
+
+static mat4_t mat4_frustum(vec_t left, vec_t right, vec_t bottom,
+                           vec_t top, vec_t znear, vec_t zfar) {
+  vec_t t0 = 2.f * znear;
+  vec_t t1 = right - left;
+  vec_t t2 = top - bottom;
+  vec_t t3 = zfar - znear;
+
+  return mat4(t0 / t1,              //
+              0.f,                  //
+              0.f,                  //
+              0.f,                  //
+              0.f,                  //
+              t0 / t2,              //
+              0.f,                  //
+              0.f,                  //
+              (right + left) / t1,  //
+              (top + bottom) / t2,  //
+              (-zfar - znear) / t3, //
+              -1.f,                 //
+              0.f,                  //
+              0.f,                  //
+              (-t0 * zfar) / t3,    //
+              0.f);
+}
+
+static mat4_t mat4_perspective(vec_t fovy, vec_t aspect_ratio_,
+                               vec_t znear, vec_t zfar) {
+  vec_t ymax = znear * tanf(fovy);
+  vec_t xmax = ymax * aspect_ratio_;
+
+  return mat4_frustum(-xmax, xmax, -ymax, ymax, znear, zfar);
+}
+
+static vec_t vec_lerp(vec_t x0, vec_t x1, vec_t t) {
+  assert(t >= 0.f);
+  assert(t <= 1.f);
+
+  if (t <= 0.f)
+    return x0;
+  if (t >= 1.f)
+    return x1;
+
+  return x0 + (x1 - x0) * t;
+}
+
+static vec3_t vec3_lerp(vec3_t x0, vec3_t x1, vec_t t) {
+  return vec3(vec_lerp(x0.v[0], x1.v[0], t),
+              vec_lerp(x0.v[1], x1.v[1], t),
+              vec_lerp(x0.v[2], x1.v[2], t));
+}
+
+static vec4_t vec4_lerp(vec4_t x0, vec4_t x1, vec_t t) {
+  return vec4(
+      vec_lerp(x0.v[0], x1.v[0], t), vec_lerp(x0.v[1], x1.v[1], t),
+      vec_lerp(x0.v[2], x1.v[2], t), vec_lerp(x0.v[3], x1.v[3], t));
+}
+
+static quat_t quat_slerp(quat_t q0, quat_t q1, vec_t t) {
+  assert(t >= 0.f);
+  assert(t <= 1.f);
+
+  if (t <= 0.f)
+    return q0;
+  if (t >= 1.f)
+    return q1;
+
+  quat_t q = q1;
+
+  vec_t cos_theta = vec4_dot(q0, q1);
+
+  if (cos_theta < 0.f) {
+    q         = vec4_neg(q1);
+    cos_theta = -cos_theta;
+  }
+
+  if (cos_theta > 1.f - KIT_EPSILON)
+    return vec4_lerp(q0, q1, t);
+
+  vec_t angle = vec_acos(cos_theta);
+
+  return vec4_div(vec4_add(vec4_mul(q0, vec_sin((1.f - t) * angle)),
+                           vec4_mul(q, vec_sin(t * angle))),
+                  vec_sin(angle));
+}
+
+static vec3_t rgb_to_xyz(vec3_t rgb) {
+  //  FIXME
+  //  Remove unnecessary x100 scaling.
+  //
+
+  vec_t red   = rgb.v[0];
+  vec_t green = rgb.v[1];
+  vec_t blue  = rgb.v[2];
+
+  //  Gamma correction
+  //
+  /*
+  if (red > 0.04045f)
+    red = vec_pow(((red + 0.055f) / 1.055f), 2.4f);
+  else
+    red = red / 12.92f;
+
+  if (green > 0.04045f)
+    green = vec_pow(((green + 0.055f) / 1.055f), 2.4f);
+  else
+    green = green / 12.92f;
+
+  if (blue > 0.04045f)
+    blue = vec_pow(((blue + 0.055f) / 1.055f), 2.4f);
+  else
+    blue = blue / 12.92f;
+  */
+
+  red   = red * 100.f;
+  green = green * 100.f;
+  blue  = blue * 100.f;
+
+  return vec3(red * 0.4124f + green * 0.3576f + blue * 0.1805f,
+              red * 0.2126f + green * 0.7152f + blue * 0.0722f,
+              red * 0.0193f + green * 0.1192f + blue * 0.9505f);
+}
+
+static vec3_t xyz_to_rgb(vec3_t xyz) {
+  //  FIXME
+  //  Remove unnecessary x100 scaling.
+  //
+
+  vec_t x = xyz.v[0] / 100.f;
+  vec_t y = xyz.v[1] / 100.f;
+  vec_t z = xyz.v[2] / 100.f;
+
+  vec_t red   = x * 3.2406f + (y * -1.5372f) + z * (-0.4986f);
+  vec_t green = x * (-0.9689f) + y * 1.8758f + z * 0.0415f;
+  vec_t blue  = x * 0.0557f + y * (-0.2040f) + z * 1.0570f;
+
+  //  Gamma correction
+  //
+  /*
+  if (red > 0.0031308f)
+    red = 1.055f * vec_pow(red, (1.0f / 2.4)) - 0.055f;
+  else
+    red = 12.92f * red;
+
+  if (green > 0.0031308f)
+    green = 1.055f * vec_pow(green, (1.0f / 2.4)) - 0.055f;
+  else
+    green = 12.92f * green;
+
+  if (blue > 0.0031308f)
+    blue = 1.055f * vec_pow(blue, (1.0f / 2.4)) - 0.055f;
+  else
+    blue = 12.92f * blue;
+  */
+
+  return vec3_clamp(vec3(red, green, blue), vec3(0.f, 0.f, 0.f),
+                    vec3(1.f, 1.f, 1.f));
+}
+
+static vec3_t lab_to_xyz(vec3_t lab) {
+  //  FIXME
+  //  Remove unnecessary x100 scaling.
+  //
+
+  vec_t lightness = lab.v[0];
+  vec_t a         = lab.v[1];
+  vec_t b         = lab.v[2];
+
+  vec_t y = (lightness + 16.f) / 116.f;
+  vec_t x = (a / 500.f) + y;
+  vec_t z = y - (b / 200.f);
+
+  if (vec_pow(y, 3.f) > 0.008856)
+    y = vec_pow(y, 3.f);
+  else
+    y = (y - (16.f / 116.f)) / 7.787f;
+
+  if (vec_pow(x, 3.f) > 0.008856f)
+    x = vec_pow(x, 3.f);
+  else
+    x = (x - (16.f / 116.f)) / 7.787f;
+
+  if (vec_pow(z, 3.f) > 0.008856f)
+    z = vec_pow(z, 3.f);
+  else
+    z = (z - (16.f / 116.f)) / 7.787f;
+
+  return vec3(KIT_COLOR_REF_X * x, KIT_COLOR_REF_Y * y,
+              KIT_COLOR_REF_Z * z);
+}
+
+static vec3_t xyz_to_lab(vec3_t xyz) {
+  //  FIXME
+  //  Remove unnecessary x100 scaling.
+  //
+
+  vec_t x = (xyz.v[0] / KIT_COLOR_REF_X);
+  vec_t y = (xyz.v[1] / KIT_COLOR_REF_Y);
+  vec_t z = (xyz.v[2] / KIT_COLOR_REF_Z);
+
+  if (x > 0.008856f)
+    x = vec_pow(x, (1.f / 3.f));
+  else
+    x = (7.787f * x) + (16.f / 116.f);
+
+  if (y > 0.008856f)
+    y = vec_pow(y, (1.f / 3.f));
+  else
+    y = (7.787f * y) + (16.f / 116.f);
+
+  if (z > 0.008856f)
+    z = vec_pow(z, (1.f / 3.f));
+  else
+    z = (7.787f * z) + (16.f / 116.f);
+
+  vec_t lightness = (116.f * y) - 16.f;
+  vec_t a         = 500.f * (x - y);
+  vec_t b         = 200.f * (y - z);
+
+  return vec3(lightness, a, b);
+}
+
+static vec3_t lab_to_lch(vec3_t lab) {
+  vec_t lightness = lab.v[0];
+  vec_t a         = lab.v[1];
+  vec_t b         = lab.v[2];
+
+  vec_t chroma = sqrtf(a * a + b * b);
+  vec_t hue    = a == 0.f ? 0.f : vec_atan(b / a);
+
+  return vec3(lightness, chroma, hue);
+}
+
+static vec3_t lch_to_lab(vec3_t lch) {
+  vec_t lightness = lch.v[0];
+  vec_t chroma    = lch.v[1];
+  vec_t hue       = lch.v[2];
+
+  vec_t a = chroma * cos(hue);
+  vec_t b = chroma * sin(hue);
+
+  return vec3(lightness, a, b);
+}
+
+static vec3_t rgb_to_lch(vec3_t rgb) {
+  vec3_t xyz = rgb_to_xyz(rgb);
+  vec3_t lab = xyz_to_lab(xyz);
+
+  return lab_to_lch(lab);
+}
+
+static vec3_t lch_to_rgb(vec3_t lch) {
+  vec3_t lab = lch_to_lab(lch);
+  vec3_t xyz = lab_to_xyz(lab);
+
+  return xyz_to_rgb(xyz);
+}
+
+static vec4_t rgba_from_lcha(vec_t lightness, vec_t chroma, vec_t hue,
+                             vec_t alpha) {
+  vec3_t rgb = lch_to_rgb(vec3(lightness, chroma, hue));
+  return vec4(rgb.v[0], rgb.v[1], rgb.v[2], alpha);
+}
+
+#ifdef __GNUC__
+#  pragma GCC            pop_options
+#  pragma GCC diagnostic pop
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif