From 9c62a21ba91d347617856dee9ccdf6da1175a83f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Camilla=20L=C3=B6wy?= <elmindreda@elmindreda.org>
Date: Wed, 10 Mar 2021 21:39:16 +0100
Subject: [PATCH] Update linmath.h

---
 deps/linmath.h | 151 +++++++++++++++++++++----------------------------
 1 file changed, 66 insertions(+), 85 deletions(-)

diff --git a/deps/linmath.h b/deps/linmath.h
index 8b2a8c2e..be37caf6 100644
--- a/deps/linmath.h
+++ b/deps/linmath.h
@@ -1,13 +1,8 @@
 #ifndef LINMATH_H
 #define LINMATH_H
 
+#include <string.h>
 #include <math.h>
-
-/* 2020-03-02 Camilla Löwy <elmindreda@elmindreda.org>
- * - Added inclusion of string.h for memcpy
- * - Replaced tan and acos with tanf and acosf
- * - Replaced double constants with float equivalents
- */
 #include <string.h>
 
 #ifdef LINMATH_NO_INLINE
@@ -38,7 +33,7 @@ LINMATH_H_FUNC void vec##n##_scale(vec##n r, vec##n const v, float const s) \
 } \
 LINMATH_H_FUNC float vec##n##_mul_inner(vec##n const a, vec##n const b) \
 { \
-	float p = 0.; \
+	float p = 0.f; \
 	int i; \
 	for(i=0; i<n; ++i) \
 		p += b[i]*a[i]; \
@@ -64,6 +59,12 @@ LINMATH_H_FUNC void vec##n##_max(vec##n r, vec##n const a, vec##n const b) \
 	int i; \
 	for(i=0; i<n; ++i) \
 		r[i] = a[i]>b[i] ? a[i] : b[i]; \
+} \
+LINMATH_H_FUNC void vec##n##_dup(vec##n r, vec##n const src) \
+{ \
+	int i; \
+	for(i=0; i<n; ++i) \
+		r[i] = src[i]; \
 }
 
 LINMATH_H_DEFINE_VEC(2)
@@ -79,13 +80,13 @@ LINMATH_H_FUNC void vec3_mul_cross(vec3 r, vec3 const a, vec3 const b)
 
 LINMATH_H_FUNC void vec3_reflect(vec3 r, vec3 const v, vec3 const n)
 {
-	float p  = 2.f*vec3_mul_inner(v, n);
+	float p = 2.f * vec3_mul_inner(v, n);
 	int i;
 	for(i=0;i<3;++i)
 		r[i] = v[i] - p*n[i];
 }
 
-LINMATH_H_FUNC void vec4_mul_cross(vec4 r, vec4 a, vec4 b)
+LINMATH_H_FUNC void vec4_mul_cross(vec4 r, vec4 const a, vec4 const b)
 {
 	r[0] = a[1]*b[2] - a[2]*b[1];
 	r[1] = a[2]*b[0] - a[0]*b[2];
@@ -93,7 +94,7 @@ LINMATH_H_FUNC void vec4_mul_cross(vec4 r, vec4 a, vec4 b)
 	r[3] = 1.f;
 }
 
-LINMATH_H_FUNC void vec4_reflect(vec4 r, vec4 v, vec4 n)
+LINMATH_H_FUNC void vec4_reflect(vec4 r, vec4 const v, vec4 const n)
 {
 	float p  = 2.f*vec4_mul_inner(v, n);
 	int i;
@@ -109,61 +110,59 @@ LINMATH_H_FUNC void mat4x4_identity(mat4x4 M)
 		for(j=0; j<4; ++j)
 			M[i][j] = i==j ? 1.f : 0.f;
 }
-LINMATH_H_FUNC void mat4x4_dup(mat4x4 M, mat4x4 N)
+LINMATH_H_FUNC void mat4x4_dup(mat4x4 M, mat4x4 const N)
 {
-	int i, j;
+	int i;
 	for(i=0; i<4; ++i)
-		for(j=0; j<4; ++j)
-			M[i][j] = N[i][j];
+		vec4_dup(M[i], N[i]);
 }
-LINMATH_H_FUNC void mat4x4_row(vec4 r, mat4x4 M, int i)
+LINMATH_H_FUNC void mat4x4_row(vec4 r, mat4x4 const M, int i)
 {
 	int k;
 	for(k=0; k<4; ++k)
 		r[k] = M[k][i];
 }
-LINMATH_H_FUNC void mat4x4_col(vec4 r, mat4x4 M, int i)
+LINMATH_H_FUNC void mat4x4_col(vec4 r, mat4x4 const M, int i)
 {
 	int k;
 	for(k=0; k<4; ++k)
 		r[k] = M[i][k];
 }
-LINMATH_H_FUNC void mat4x4_transpose(mat4x4 M, mat4x4 N)
+LINMATH_H_FUNC void mat4x4_transpose(mat4x4 M, mat4x4 const N)
 {
+    // Note: if M and N are the same, the user has to
+    // explicitly make a copy of M and set it to N.
 	int i, j;
 	for(j=0; j<4; ++j)
 		for(i=0; i<4; ++i)
 			M[i][j] = N[j][i];
 }
-LINMATH_H_FUNC void mat4x4_add(mat4x4 M, mat4x4 a, mat4x4 b)
+LINMATH_H_FUNC void mat4x4_add(mat4x4 M, mat4x4 const a, mat4x4 const b)
 {
 	int i;
 	for(i=0; i<4; ++i)
 		vec4_add(M[i], a[i], b[i]);
 }
-LINMATH_H_FUNC void mat4x4_sub(mat4x4 M, mat4x4 a, mat4x4 b)
+LINMATH_H_FUNC void mat4x4_sub(mat4x4 M, mat4x4 const a, mat4x4 const b)
 {
 	int i;
 	for(i=0; i<4; ++i)
 		vec4_sub(M[i], a[i], b[i]);
 }
-LINMATH_H_FUNC void mat4x4_scale(mat4x4 M, mat4x4 a, float k)
+LINMATH_H_FUNC void mat4x4_scale(mat4x4 M, mat4x4 const a, float k)
 {
 	int i;
 	for(i=0; i<4; ++i)
 		vec4_scale(M[i], a[i], k);
 }
-LINMATH_H_FUNC void mat4x4_scale_aniso(mat4x4 M, mat4x4 a, float x, float y, float z)
+LINMATH_H_FUNC void mat4x4_scale_aniso(mat4x4 M, mat4x4 const a, float x, float y, float z)
 {
-	int i;
 	vec4_scale(M[0], a[0], x);
 	vec4_scale(M[1], a[1], y);
 	vec4_scale(M[2], a[2], z);
-	for(i = 0; i < 4; ++i) {
-		M[3][i] = a[3][i];
-	}
+	vec4_dup(M[3], a[3]);
 }
-LINMATH_H_FUNC void mat4x4_mul(mat4x4 M, mat4x4 a, mat4x4 b)
+LINMATH_H_FUNC void mat4x4_mul(mat4x4 M, mat4x4 const a, mat4x4 const b)
 {
 	mat4x4 temp;
 	int k, r, c;
@@ -174,7 +173,7 @@ LINMATH_H_FUNC void mat4x4_mul(mat4x4 M, mat4x4 a, mat4x4 b)
 	}
 	mat4x4_dup(M, temp);
 }
-LINMATH_H_FUNC void mat4x4_mul_vec4(vec4 r, mat4x4 M, vec4 v)
+LINMATH_H_FUNC void mat4x4_mul_vec4(vec4 r, mat4x4 const M, vec4 const v)
 {
 	int i, j;
 	for(j=0; j<4; ++j) {
@@ -200,13 +199,13 @@ LINMATH_H_FUNC void mat4x4_translate_in_place(mat4x4 M, float x, float y, float
 		M[3][i] += vec4_mul_inner(r, t);
 	}
 }
-LINMATH_H_FUNC void mat4x4_from_vec3_mul_outer(mat4x4 M, vec3 a, vec3 b)
+LINMATH_H_FUNC void mat4x4_from_vec3_mul_outer(mat4x4 M, vec3 const a, vec3 const b)
 {
 	int i, j;
 	for(i=0; i<4; ++i) for(j=0; j<4; ++j)
 		M[i][j] = i<3 && j<3 ? a[i] * b[j] : 0.f;
 }
-LINMATH_H_FUNC void mat4x4_rotate(mat4x4 R, mat4x4 M, float x, float y, float z, float angle)
+LINMATH_H_FUNC void mat4x4_rotate(mat4x4 R, mat4x4 const M, float x, float y, float z, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -234,13 +233,13 @@ LINMATH_H_FUNC void mat4x4_rotate(mat4x4 R, mat4x4 M, float x, float y, float z,
 		mat4x4_add(T, T, C);
 		mat4x4_add(T, T, S);
 
-		T[3][3] = 1.;		
+		T[3][3] = 1.f;
 		mat4x4_mul(R, M, T);
 	} else {
 		mat4x4_dup(R, M);
 	}
 }
-LINMATH_H_FUNC void mat4x4_rotate_X(mat4x4 Q, mat4x4 M, float angle)
+LINMATH_H_FUNC void mat4x4_rotate_X(mat4x4 Q, mat4x4 const M, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -252,7 +251,7 @@ LINMATH_H_FUNC void mat4x4_rotate_X(mat4x4 Q, mat4x4 M, float angle)
 	};
 	mat4x4_mul(Q, M, R);
 }
-LINMATH_H_FUNC void mat4x4_rotate_Y(mat4x4 Q, mat4x4 M, float angle)
+LINMATH_H_FUNC void mat4x4_rotate_Y(mat4x4 Q, mat4x4 const M, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -264,7 +263,7 @@ LINMATH_H_FUNC void mat4x4_rotate_Y(mat4x4 Q, mat4x4 M, float angle)
 	};
 	mat4x4_mul(Q, M, R);
 }
-LINMATH_H_FUNC void mat4x4_rotate_Z(mat4x4 Q, mat4x4 M, float angle)
+LINMATH_H_FUNC void mat4x4_rotate_Z(mat4x4 Q, mat4x4 const M, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -276,7 +275,7 @@ LINMATH_H_FUNC void mat4x4_rotate_Z(mat4x4 Q, mat4x4 M, float angle)
 	};
 	mat4x4_mul(Q, M, R);
 }
-LINMATH_H_FUNC void mat4x4_invert(mat4x4 T, mat4x4 M)
+LINMATH_H_FUNC void mat4x4_invert(mat4x4 T, mat4x4 const M)
 {
 	float s[6];
 	float c[6];
@@ -317,10 +316,10 @@ LINMATH_H_FUNC void mat4x4_invert(mat4x4 T, mat4x4 M)
 	T[3][2] = (-M[3][0] * s[3] + M[3][1] * s[1] - M[3][2] * s[0]) * idet;
 	T[3][3] = ( M[2][0] * s[3] - M[2][1] * s[1] + M[2][2] * s[0]) * idet;
 }
-LINMATH_H_FUNC void mat4x4_orthonormalize(mat4x4 R, mat4x4 M)
+LINMATH_H_FUNC void mat4x4_orthonormalize(mat4x4 R, mat4x4 const M)
 {
 	mat4x4_dup(R, M);
-	float s = 1.;
+	float s = 1.f;
 	vec3 h;
 
 	vec3_norm(R[2], R[2]);
@@ -398,7 +397,7 @@ LINMATH_H_FUNC void mat4x4_perspective(mat4x4 m, float y_fov, float aspect, floa
 	m[3][2] = -((2.f * f * n) / (f - n));
 	m[3][3] = 0.f;
 }
-LINMATH_H_FUNC void mat4x4_look_at(mat4x4 m, vec3 eye, vec3 center, vec3 up)
+LINMATH_H_FUNC void mat4x4_look_at(mat4x4 m, vec3 const eye, vec3 const center, vec3 const up)
 {
 	/* Adapted from Android's OpenGL Matrix.java.                        */
 	/* See the OpenGL GLUT documentation for gluLookAt for a description */
@@ -441,24 +440,18 @@ LINMATH_H_FUNC void mat4x4_look_at(mat4x4 m, vec3 eye, vec3 center, vec3 up)
 }
 
 typedef float quat[4];
+#define quat_add vec4_add
+#define quat_sub vec4_sub
+#define quat_norm vec4_norm
+#define quat_scale vec4_scale
+#define quat_mul_inner vec4_mul_inner
+
 LINMATH_H_FUNC void quat_identity(quat q)
 {
 	q[0] = q[1] = q[2] = 0.f;
 	q[3] = 1.f;
 }
-LINMATH_H_FUNC void quat_add(quat r, quat a, quat b)
-{
-	int i;
-	for(i=0; i<4; ++i)
-		r[i] = a[i] + b[i];
-}
-LINMATH_H_FUNC void quat_sub(quat r, quat a, quat b)
-{
-	int i;
-	for(i=0; i<4; ++i)
-		r[i] = a[i] - b[i];
-}
-LINMATH_H_FUNC void quat_mul(quat r, quat p, quat q)
+LINMATH_H_FUNC void quat_mul(quat r, quat const p, quat const q)
 {
 	vec3 w;
 	vec3_mul_cross(r, p, q);
@@ -468,37 +461,22 @@ LINMATH_H_FUNC void quat_mul(quat r, quat p, quat q)
 	vec3_add(r, r, w);
 	r[3] = p[3]*q[3] - vec3_mul_inner(p, q);
 }
-LINMATH_H_FUNC void quat_scale(quat r, quat v, float s)
-{
-	int i;
-	for(i=0; i<4; ++i)
-		r[i] = v[i] * s;
-}
-LINMATH_H_FUNC float quat_inner_product(quat a, quat b)
-{
-	float p = 0.f;
-	int i;
-	for(i=0; i<4; ++i)
-		p += b[i]*a[i];
-	return p;
-}
-LINMATH_H_FUNC void quat_conj(quat r, quat q)
+LINMATH_H_FUNC void quat_conj(quat r, quat const q)
 {
 	int i;
 	for(i=0; i<3; ++i)
 		r[i] = -q[i];
 	r[3] = q[3];
 }
-LINMATH_H_FUNC void quat_rotate(quat r, float angle, vec3 axis) {
-	vec3 v;
-	vec3_scale(v, axis, sinf(angle / 2));
-	int i;
-	for(i=0; i<3; ++i)
-		r[i] = v[i];
-	r[3] = cosf(angle / 2);
+LINMATH_H_FUNC void quat_rotate(quat r, float angle, vec3 const axis) {
+    vec3 axis_norm;
+    vec3_norm(axis_norm, axis);
+    float s = sinf(angle / 2);
+    float c = cosf(angle / 2);
+    vec3_scale(r, axis_norm, s);
+    r[3] = c;
 }
-#define quat_norm vec4_norm
-LINMATH_H_FUNC void quat_mul_vec3(vec3 r, quat q, vec3 v)
+LINMATH_H_FUNC void quat_mul_vec3(vec3 r, quat const q, vec3 const v)
 {
 /*
  * Method by Fabian 'ryg' Giessen (of Farbrausch)
@@ -518,7 +496,7 @@ v' = v + q.w * t + cross(q.xyz, t)
 	vec3_add(r, v, t);
 	vec3_add(r, r, u);
 }
-LINMATH_H_FUNC void mat4x4_from_quat(mat4x4 M, quat q)
+LINMATH_H_FUNC void mat4x4_from_quat(mat4x4 M, quat const q)
 {
 	float a = q[3];
 	float b = q[0];
@@ -548,18 +526,21 @@ LINMATH_H_FUNC void mat4x4_from_quat(mat4x4 M, quat q)
 	M[3][3] = 1.f;
 }
 
-LINMATH_H_FUNC void mat4x4o_mul_quat(mat4x4 R, mat4x4 M, quat q)
+LINMATH_H_FUNC void mat4x4o_mul_quat(mat4x4 R, mat4x4 const M, quat const q)
 {
-/*  XXX: The way this is written only works for othogonal matrices. */
+/*  XXX: The way this is written only works for orthogonal matrices. */
 /* TODO: Take care of non-orthogonal case. */
 	quat_mul_vec3(R[0], q, M[0]);
 	quat_mul_vec3(R[1], q, M[1]);
 	quat_mul_vec3(R[2], q, M[2]);
 
 	R[3][0] = R[3][1] = R[3][2] = 0.f;
-	R[3][3] = 1.f;
+	R[0][3] = M[0][3];
+	R[1][3] = M[1][3];
+	R[2][3] = M[2][3];
+	R[3][3] = M[3][3];  // typically 1.0, but here we make it general
 }
-LINMATH_H_FUNC void quat_from_mat4x4(quat q, mat4x4 M)
+LINMATH_H_FUNC void quat_from_mat4x4(quat q, mat4x4 const M)
 {
 	float r=0.f;
 	int i;
@@ -589,7 +570,7 @@ LINMATH_H_FUNC void quat_from_mat4x4(quat q, mat4x4 M)
 	q[3] = (M[p[2]][p[1]] - M[p[1]][p[2]])/(2.f*r);
 }
 
-LINMATH_H_FUNC void mat4x4_arcball(mat4x4 R, mat4x4 M, vec2 _a, vec2 _b, float s)
+LINMATH_H_FUNC void mat4x4_arcball(mat4x4 R, mat4x4 const M, vec2 const _a, vec2 const _b, float s)
 {
 	vec2 a; memcpy(a, _a, sizeof(a));
 	vec2 b; memcpy(b, _b, sizeof(b));
@@ -597,14 +578,14 @@ LINMATH_H_FUNC void mat4x4_arcball(mat4x4 R, mat4x4 M, vec2 _a, vec2 _b, float s
 	float z_a = 0.;
 	float z_b = 0.;
 
-	if(vec2_len(a) < 1.f) {
-		z_a = sqrtf(1.f - vec2_mul_inner(a, a));
+	if(vec2_len(a) < 1.) {
+		z_a = sqrtf(1. - vec2_mul_inner(a, a));
 	} else {
 		vec2_norm(a, a);
 	}
 
-	if(vec2_len(b) < 1.f) {
-		z_b = sqrtf(1.f - vec2_mul_inner(b, b));
+	if(vec2_len(b) < 1.) {
+		z_b = sqrtf(1. - vec2_mul_inner(b, b));
 	} else {
 		vec2_norm(b, b);
 	}
@@ -615,7 +596,7 @@ LINMATH_H_FUNC void mat4x4_arcball(mat4x4 R, mat4x4 M, vec2 _a, vec2 _b, float s
 	vec3 c_;
 	vec3_mul_cross(c_, a_, b_);
 
-	float const angle = acosf(vec3_mul_inner(a_, b_)) * s;
+	float const angle = acos(vec3_mul_inner(a_, b_)) * s;
 	mat4x4_rotate(R, M, c_[0], c_[1], c_[2], angle);
 }
 #endif