Update linmath.h

This updates our linmath.h to the latest version plus minor local fixes
for MSVC and Clang.

Fixes #1653.

deps/linmath.h

@@ -3,31 +3,40 @@
 
 #include <math.h>
 
-#ifdef _MSC_VER
+/* 2020-03-02 Camilla Löwy <elmindreda@elmindreda.org>
-#define inline __inline
+ * - 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
+#define LINMATH_H_FUNC static
+#else
+#define LINMATH_H_FUNC static inline
 #endif
 
 #define LINMATH_H_DEFINE_VEC(n) \
 typedef float vec##n[n]; \
-static inline void vec##n##_add(vec##n r, vec##n const a, vec##n const b) \
+LINMATH_H_FUNC void vec##n##_add(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]; \
 } \
-static inline void vec##n##_sub(vec##n r, vec##n const a, vec##n const b) \
+LINMATH_H_FUNC void vec##n##_sub(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]; \
 } \
-static inline void vec##n##_scale(vec##n r, vec##n const v, float const s) \
+LINMATH_H_FUNC void vec##n##_scale(vec##n r, vec##n const v, float const s) \
 { \
 	int i; \
 	for(i=0; i<n; ++i) \
 		r[i] = v[i] * s; \
 } \
-static inline float vec##n##_mul_inner(vec##n const a, vec##n const b) \
+LINMATH_H_FUNC float vec##n##_mul_inner(vec##n const a, vec##n const b) \
 { \
 	float p = 0.; \
 	int i; \
@@ -35,28 +44,40 @@ static inline float vec##n##_mul_inner(vec##n const a, vec##n const b) \
 		p += b[i]*a[i]; \
 	return p; \
 } \
-static inline float vec##n##_len(vec##n const v) \
+LINMATH_H_FUNC float vec##n##_len(vec##n const v) \
 { \
-	return (float) sqrt(vec##n##_mul_inner(v,v)); \
+	return sqrtf(vec##n##_mul_inner(v,v)); \
 } \
-static inline void vec##n##_norm(vec##n r, vec##n const v) \
+LINMATH_H_FUNC void vec##n##_norm(vec##n r, vec##n const v) \
 { \
 	float k = 1.f / vec##n##_len(v); \
 	vec##n##_scale(r, v, k); \
+} \
+LINMATH_H_FUNC void vec##n##_min(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##_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_DEFINE_VEC(2)
 LINMATH_H_DEFINE_VEC(3)
 LINMATH_H_DEFINE_VEC(4)
 
-static inline void vec3_mul_cross(vec3 r, vec3 const a, vec3 const b)
+LINMATH_H_FUNC void vec3_mul_cross(vec3 r, vec3 const a, vec3 const b)
 {
 	r[0] = a[1]*b[2] - a[2]*b[1];
 	r[1] = a[2]*b[0] - a[0]*b[2];
 	r[2] = a[0]*b[1] - a[1]*b[0];
 }
 
-static inline void vec3_reflect(vec3 r, vec3 const v, vec3 const n)
+LINMATH_H_FUNC void vec3_reflect(vec3 r, vec3 const v, vec3 const n)
 {
 	float p  = 2.f*vec3_mul_inner(v, n);
 	int i;
@@ -64,7 +85,7 @@ static inline void vec3_reflect(vec3 r, vec3 const v, vec3 const n)
 		r[i] = v[i] - p*n[i];
 }
 
-static inline void vec4_mul_cross(vec4 r, vec4 a, vec4 b)
+LINMATH_H_FUNC void vec4_mul_cross(vec4 r, vec4 a, vec4 b)
 {
 	r[0] = a[1]*b[2] - a[2]*b[1];
 	r[1] = a[2]*b[0] - a[0]*b[2];
@@ -72,7 +93,7 @@ static inline void vec4_mul_cross(vec4 r, vec4 a, vec4 b)
 	r[3] = 1.f;
 }
 
-static inline void vec4_reflect(vec4 r, vec4 v, vec4 n)
+LINMATH_H_FUNC void vec4_reflect(vec4 r, vec4 v, vec4 n)
 {
 	float p  = 2.f*vec4_mul_inner(v, n);
 	int i;
@@ -81,58 +102,58 @@ static inline void vec4_reflect(vec4 r, vec4 v, vec4 n)
 }
 
 typedef vec4 mat4x4[4];
-static inline void mat4x4_identity(mat4x4 M)
+LINMATH_H_FUNC void mat4x4_identity(mat4x4 M)
 {
 	int i, j;
 	for(i=0; i<4; ++i)
 		for(j=0; j<4; ++j)
 			M[i][j] = i==j ? 1.f : 0.f;
 }
-static inline void mat4x4_dup(mat4x4 M, mat4x4 N)
+LINMATH_H_FUNC void mat4x4_dup(mat4x4 M, mat4x4 N)
 {
 	int i, j;
 	for(i=0; i<4; ++i)
 		for(j=0; j<4; ++j)
 			M[i][j] = N[i][j];
 }
-static inline void mat4x4_row(vec4 r, mat4x4 M, int i)
+LINMATH_H_FUNC void mat4x4_row(vec4 r, mat4x4 M, int i)
 {
 	int k;
 	for(k=0; k<4; ++k)
 		r[k] = M[k][i];
 }
-static inline void mat4x4_col(vec4 r, mat4x4 M, int i)
+LINMATH_H_FUNC void mat4x4_col(vec4 r, mat4x4 M, int i)
 {
 	int k;
 	for(k=0; k<4; ++k)
 		r[k] = M[i][k];
 }
-static inline void mat4x4_transpose(mat4x4 M, mat4x4 N)
+LINMATH_H_FUNC void mat4x4_transpose(mat4x4 M, mat4x4 N)
 {
 	int i, j;
 	for(j=0; j<4; ++j)
 		for(i=0; i<4; ++i)
 			M[i][j] = N[j][i];
 }
-static inline void mat4x4_add(mat4x4 M, mat4x4 a, mat4x4 b)
+LINMATH_H_FUNC void mat4x4_add(mat4x4 M, mat4x4 a, mat4x4 b)
 {
 	int i;
 	for(i=0; i<4; ++i)
 		vec4_add(M[i], a[i], b[i]);
 }
-static inline void mat4x4_sub(mat4x4 M, mat4x4 a, mat4x4 b)
+LINMATH_H_FUNC void mat4x4_sub(mat4x4 M, mat4x4 a, mat4x4 b)
 {
 	int i;
 	for(i=0; i<4; ++i)
 		vec4_sub(M[i], a[i], b[i]);
 }
-static inline void mat4x4_scale(mat4x4 M, mat4x4 a, float k)
+LINMATH_H_FUNC void mat4x4_scale(mat4x4 M, mat4x4 a, float k)
 {
 	int i;
 	for(i=0; i<4; ++i)
 		vec4_scale(M[i], a[i], k);
 }
-static inline void mat4x4_scale_aniso(mat4x4 M, mat4x4 a, float x, float y, float z)
+LINMATH_H_FUNC void mat4x4_scale_aniso(mat4x4 M, mat4x4 a, float x, float y, float z)
 {
 	int i;
 	vec4_scale(M[0], a[0], x);
@@ -142,7 +163,7 @@ static inline void mat4x4_scale_aniso(mat4x4 M, mat4x4 a, float x, float y, floa
 		M[3][i] = a[3][i];
 	}
 }
-static inline void mat4x4_mul(mat4x4 M, mat4x4 a, mat4x4 b)
+LINMATH_H_FUNC void mat4x4_mul(mat4x4 M, mat4x4 a, mat4x4 b)
 {
 	mat4x4 temp;
 	int k, r, c;
@@ -153,7 +174,7 @@ static inline void mat4x4_mul(mat4x4 M, mat4x4 a, mat4x4 b)
 	}
 	mat4x4_dup(M, temp);
 }
-static inline void mat4x4_mul_vec4(vec4 r, mat4x4 M, vec4 v)
+LINMATH_H_FUNC void mat4x4_mul_vec4(vec4 r, mat4x4 M, vec4 v)
 {
 	int i, j;
 	for(j=0; j<4; ++j) {
@@ -162,14 +183,14 @@ static inline void mat4x4_mul_vec4(vec4 r, mat4x4 M, vec4 v)
 			r[j] += M[i][j] * v[i];
 	}
 }
-static inline void mat4x4_translate(mat4x4 T, float x, float y, float z)
+LINMATH_H_FUNC void mat4x4_translate(mat4x4 T, float x, float y, float z)
 {
 	mat4x4_identity(T);
 	T[3][0] = x;
 	T[3][1] = y;
 	T[3][2] = z;
 }
-static inline void mat4x4_translate_in_place(mat4x4 M, float x, float y, float z)
+LINMATH_H_FUNC void mat4x4_translate_in_place(mat4x4 M, float x, float y, float z)
 {
 	vec4 t = {x, y, z, 0};
 	vec4 r;
@@ -179,33 +200,32 @@ static inline void mat4x4_translate_in_place(mat4x4 M, float x, float y, float z
 		M[3][i] += vec4_mul_inner(r, t);
 	}
 }
-static inline void mat4x4_from_vec3_mul_outer(mat4x4 M, vec3 a, vec3 b)
+LINMATH_H_FUNC void mat4x4_from_vec3_mul_outer(mat4x4 M, vec3 a, vec3 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;
 }
-static inline void mat4x4_rotate(mat4x4 R, mat4x4 M, float x, float y, float z, float angle)
+LINMATH_H_FUNC void mat4x4_rotate(mat4x4 R, mat4x4 M, float x, float y, float z, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
 	vec3 u = {x, y, z};
 
 	if(vec3_len(u) > 1e-4) {
-		mat4x4 T, C, S = {{0}};
-
 		vec3_norm(u, u);
+		mat4x4 T;
 		mat4x4_from_vec3_mul_outer(T, u, u);
 
-		S[1][2] =  u[0];
+		mat4x4 S = {
-		S[2][1] = -u[0];
+			{    0,  u[2], -u[1], 0},
-		S[2][0] =  u[1];
+			{-u[2],     0,  u[0], 0},
-		S[0][2] = -u[1];
+			{ u[1], -u[0],     0, 0},
-		S[0][1] =  u[2];
+			{    0,     0,     0, 0}
-		S[1][0] = -u[2];
+		};
-
 		mat4x4_scale(S, S, s);
 
+		mat4x4 C;
 		mat4x4_identity(C);
 		mat4x4_sub(C, C, T);
 
@@ -214,13 +234,13 @@ static inline 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.;		
 		mat4x4_mul(R, M, T);
 	} else {
 		mat4x4_dup(R, M);
 	}
 }
-static inline void mat4x4_rotate_X(mat4x4 Q, mat4x4 M, float angle)
+LINMATH_H_FUNC void mat4x4_rotate_X(mat4x4 Q, mat4x4 M, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -232,7 +252,7 @@ static inline void mat4x4_rotate_X(mat4x4 Q, mat4x4 M, float angle)
 	};
 	mat4x4_mul(Q, M, R);
 }
-static inline void mat4x4_rotate_Y(mat4x4 Q, mat4x4 M, float angle)
+LINMATH_H_FUNC void mat4x4_rotate_Y(mat4x4 Q, mat4x4 M, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -244,7 +264,7 @@ static inline void mat4x4_rotate_Y(mat4x4 Q, mat4x4 M, float angle)
 	};
 	mat4x4_mul(Q, M, R);
 }
-static inline void mat4x4_rotate_Z(mat4x4 Q, mat4x4 M, float angle)
+LINMATH_H_FUNC void mat4x4_rotate_Z(mat4x4 Q, mat4x4 M, float angle)
 {
 	float s = sinf(angle);
 	float c = cosf(angle);
@@ -256,9 +276,8 @@ static inline void mat4x4_rotate_Z(mat4x4 Q, mat4x4 M, float angle)
 	};
 	mat4x4_mul(Q, M, R);
 }
-static inline void mat4x4_invert(mat4x4 T, mat4x4 M)
+LINMATH_H_FUNC void mat4x4_invert(mat4x4 T, mat4x4 M)
 {
-	float idet;
 	float s[6];
 	float c[6];
 	s[0] = M[0][0]*M[1][1] - M[1][0]*M[0][1];
@@ -274,10 +293,10 @@ static inline void mat4x4_invert(mat4x4 T, mat4x4 M)
 	c[3] = M[2][1]*M[3][2] - M[3][1]*M[2][2];
 	c[4] = M[2][1]*M[3][3] - M[3][1]*M[2][3];
 	c[5] = M[2][2]*M[3][3] - M[3][2]*M[2][3];
-
+	
 	/* Assumes it is invertible */
-	idet = 1.0f/( s[0]*c[5]-s[1]*c[4]+s[2]*c[3]+s[3]*c[2]-s[4]*c[1]+s[5]*c[0] );
+	float idet = 1.0f/( s[0]*c[5]-s[1]*c[4]+s[2]*c[3]+s[3]*c[2]-s[4]*c[1]+s[5]*c[0] );
-
+	
 	T[0][0] = ( M[1][1] * c[5] - M[1][2] * c[4] + M[1][3] * c[3]) * idet;
 	T[0][1] = (-M[0][1] * c[5] + M[0][2] * c[4] - M[0][3] * c[3]) * idet;
 	T[0][2] = ( M[3][1] * s[5] - M[3][2] * s[4] + M[3][3] * s[3]) * idet;
@@ -298,35 +317,34 @@ static inline 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;
 }
-static inline void mat4x4_orthonormalize(mat4x4 R, mat4x4 M)
+LINMATH_H_FUNC void mat4x4_orthonormalize(mat4x4 R, mat4x4 M)
 {
+	mat4x4_dup(R, M);
 	float s = 1.;
 	vec3 h;
 
-	mat4x4_dup(R, M);
 	vec3_norm(R[2], R[2]);
-
+	
-	s = vec3_mul_inner(R[1], R[2]);
-	vec3_scale(h, R[2], s);
-	vec3_sub(R[1], R[1], h);
-	vec3_norm(R[2], R[2]);
-
 	s = vec3_mul_inner(R[1], R[2]);
 	vec3_scale(h, R[2], s);
 	vec3_sub(R[1], R[1], h);
 	vec3_norm(R[1], R[1]);
 
+	s = vec3_mul_inner(R[0], R[2]);
+	vec3_scale(h, R[2], s);
+	vec3_sub(R[0], R[0], h);
+
 	s = vec3_mul_inner(R[0], R[1]);
 	vec3_scale(h, R[1], s);
 	vec3_sub(R[0], R[0], h);
 	vec3_norm(R[0], R[0]);
 }
 
-static inline void mat4x4_frustum(mat4x4 M, float l, float r, float b, float t, float n, float f)
+LINMATH_H_FUNC void mat4x4_frustum(mat4x4 M, float l, float r, float b, float t, float n, float f)
 {
 	M[0][0] = 2.f*n/(r-l);
 	M[0][1] = M[0][2] = M[0][3] = 0.f;
-
+	
 	M[1][1] = 2.f*n/(t-b);
 	M[1][0] = M[1][2] = M[1][3] = 0.f;
 
@@ -334,11 +352,11 @@ static inline void mat4x4_frustum(mat4x4 M, float l, float r, float b, float t,
 	M[2][1] = (t+b)/(t-b);
 	M[2][2] = -(f+n)/(f-n);
 	M[2][3] = -1.f;
-
+	
 	M[3][2] = -2.f*(f*n)/(f-n);
 	M[3][0] = M[3][1] = M[3][3] = 0.f;
 }
-static inline void mat4x4_ortho(mat4x4 M, float l, float r, float b, float t, float n, float f)
+LINMATH_H_FUNC void mat4x4_ortho(mat4x4 M, float l, float r, float b, float t, float n, float f)
 {
 	M[0][0] = 2.f/(r-l);
 	M[0][1] = M[0][2] = M[0][3] = 0.f;
@@ -348,17 +366,17 @@ static inline void mat4x4_ortho(mat4x4 M, float l, float r, float b, float t, fl
 
 	M[2][2] = -2.f/(f-n);
 	M[2][0] = M[2][1] = M[2][3] = 0.f;
-
+	
 	M[3][0] = -(r+l)/(r-l);
 	M[3][1] = -(t+b)/(t-b);
 	M[3][2] = -(f+n)/(f-n);
 	M[3][3] = 1.f;
 }
-static inline void mat4x4_perspective(mat4x4 m, float y_fov, float aspect, float n, float f)
+LINMATH_H_FUNC void mat4x4_perspective(mat4x4 m, float y_fov, float aspect, float n, float f)
 {
 	/* NOTE: Degrees are an unhandy unit to work with.
 	 * linmath.h uses radians for everything! */
-	float const a = 1.f / (float) tan(y_fov / 2.f);
+	float const a = 1.f / tanf(y_fov / 2.f);
 
 	m[0][0] = a / aspect;
 	m[0][1] = 0.f;
@@ -380,7 +398,7 @@ static inline void mat4x4_perspective(mat4x4 m, float y_fov, float aspect, float
 	m[3][2] = -((2.f * f * n) / (f - n));
 	m[3][3] = 0.f;
 }
-static inline void mat4x4_look_at(mat4x4 m, vec3 eye, vec3 center, vec3 up)
+LINMATH_H_FUNC void mat4x4_look_at(mat4x4 m, vec3 eye, vec3 center, vec3 up)
 {
 	/* Adapted from Android's OpenGL Matrix.java.                        */
 	/* See the OpenGL GLUT documentation for gluLookAt for a description */
@@ -389,15 +407,14 @@ static inline void mat4x4_look_at(mat4x4 m, vec3 eye, vec3 center, vec3 up)
 	/* TODO: The negation of of can be spared by swapping the order of
 	 *       operands in the following cross products in the right way. */
 	vec3 f;
+	vec3_sub(f, center, eye);	
+	vec3_norm(f, f);	
+	
 	vec3 s;
-	vec3 t;
-
-	vec3_sub(f, center, eye);
-	vec3_norm(f, f);
-
 	vec3_mul_cross(s, f, up);
 	vec3_norm(s, s);
 
+	vec3 t;
 	vec3_mul_cross(t, s, f);
 
 	m[0][0] =  s[0];
@@ -424,24 +441,24 @@ static inline void mat4x4_look_at(mat4x4 m, vec3 eye, vec3 center, vec3 up)
 }
 
 typedef float quat[4];
-static inline void quat_identity(quat q)
+LINMATH_H_FUNC void quat_identity(quat q)
 {
 	q[0] = q[1] = q[2] = 0.f;
 	q[3] = 1.f;
 }
-static inline void quat_add(quat r, quat a, quat b)
+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];
 }
-static inline void quat_sub(quat r, quat a, quat b)
+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];
 }
-static inline void quat_mul(quat r, quat p, quat q)
+LINMATH_H_FUNC void quat_mul(quat r, quat p, quat q)
 {
 	vec3 w;
 	vec3_mul_cross(r, p, q);
@@ -451,13 +468,13 @@ static inline 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);
 }
-static inline void quat_scale(quat r, quat v, float s)
+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;
 }
-static inline float quat_inner_product(quat a, quat b)
+LINMATH_H_FUNC float quat_inner_product(quat a, quat b)
 {
 	float p = 0.f;
 	int i;
@@ -465,42 +482,43 @@ static inline float quat_inner_product(quat a, quat b)
 		p += b[i]*a[i];
 	return p;
 }
-static inline void quat_conj(quat r, quat q)
+LINMATH_H_FUNC void quat_conj(quat r, quat q)
 {
 	int i;
 	for(i=0; i<3; ++i)
 		r[i] = -q[i];
 	r[3] = q[3];
 }
-static inline void quat_rotate(quat r, float angle, vec3 axis) {
+LINMATH_H_FUNC void quat_rotate(quat r, float angle, vec3 axis) {
-	int i;
 	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);
 }
 #define quat_norm vec4_norm
-static inline void quat_mul_vec3(vec3 r, quat q, vec3 v)
+LINMATH_H_FUNC void quat_mul_vec3(vec3 r, quat q, vec3 v)
 {
 /*
  * Method by Fabian 'ryg' Giessen (of Farbrausch)
 t = 2 * cross(q.xyz, v)
 v' = v + q.w * t + cross(q.xyz, t)
  */
-	vec3 t = {q[0], q[1], q[2]};
+	vec3 t;
+	vec3 q_xyz = {q[0], q[1], q[2]};
 	vec3 u = {q[0], q[1], q[2]};
 
-	vec3_mul_cross(t, t, v);
+	vec3_mul_cross(t, q_xyz, v);
 	vec3_scale(t, t, 2);
 
-	vec3_mul_cross(u, u, t);
+	vec3_mul_cross(u, q_xyz, t);
 	vec3_scale(t, t, q[3]);
 
 	vec3_add(r, v, t);
 	vec3_add(r, r, u);
 }
-static inline void mat4x4_from_quat(mat4x4 M, quat q)
+LINMATH_H_FUNC void mat4x4_from_quat(mat4x4 M, quat q)
 {
 	float a = q[3];
 	float b = q[0];
@@ -510,7 +528,7 @@ static inline void mat4x4_from_quat(mat4x4 M, quat q)
 	float b2 = b*b;
 	float c2 = c*c;
 	float d2 = d*d;
-
+	
 	M[0][0] = a2 + b2 - c2 - d2;
 	M[0][1] = 2.f*(b*c + a*d);
 	M[0][2] = 2.f*(b*d - a*c);
@@ -530,7 +548,7 @@ static inline void mat4x4_from_quat(mat4x4 M, quat q)
 	M[3][3] = 1.f;
 }
 
-static inline void mat4x4o_mul_quat(mat4x4 R, mat4x4 M, quat q)
+LINMATH_H_FUNC void mat4x4o_mul_quat(mat4x4 R, mat4x4 M, quat q)
 {
 /*  XXX: The way this is written only works for othogonal matrices. */
 /* TODO: Take care of non-orthogonal case. */
@@ -541,7 +559,7 @@ static inline void mat4x4o_mul_quat(mat4x4 R, mat4x4 M, quat q)
 	R[3][0] = R[3][1] = R[3][2] = 0.f;
 	R[3][3] = 1.f;
 }
-static inline void quat_from_mat4x4(quat q, mat4x4 M)
+LINMATH_H_FUNC void quat_from_mat4x4(quat q, mat4x4 M)
 {
 	float r=0.f;
 	int i;
@@ -557,7 +575,7 @@ static inline void quat_from_mat4x4(quat q, mat4x4 M)
 		p = &perm[i];
 	}
 
-	r = (float) sqrt(1.f + M[p[0]][p[0]] - M[p[1]][p[1]] - M[p[2]][p[2]] );
+	r = sqrtf(1.f + M[p[0]][p[0]] - M[p[1]][p[1]] - M[p[2]][p[2]] );
 
 	if(r < 1e-6) {
 		q[0] = 1.f;
@@ -571,4 +589,33 @@ static inline 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)
+{
+	vec2 a; memcpy(a, _a, sizeof(a));
+	vec2 b; memcpy(b, _b, sizeof(b));
+	
+	float z_a = 0.;
+	float z_b = 0.;
+
+	if(vec2_len(a) < 1.f) {
+		z_a = sqrtf(1.f - 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));
+	} else {
+		vec2_norm(b, b);
+	}
+	
+	vec3 a_ = {a[0], a[1], z_a};
+	vec3 b_ = {b[0], b[1], z_b};
+
+	vec3 c_;
+	vec3_mul_cross(c_, a_, b_);
+
+	float const angle = acosf(vec3_mul_inner(a_, b_)) * s;
+	mat4x4_rotate(R, M, c_[0], c_[1], c_[2], angle);
+}
 #endif