use std::iter;
use std::mem;
use std::ops::*;
use rand::distributions::{Distribution, Standard};
use rand::Rng;
use num_traits::{cast, NumCast};
use structure::*;
use angle::Rad;
use approx;
use euler::Euler;
use matrix::{Matrix3, Matrix4};
use num::BaseFloat;
use point::Point3;
use rotation::{Basis3, Rotation, Rotation3};
use vector::Vector3;
#[cfg(feature = "simd")]
use simd::f32x4 as Simdf32x4;
#[cfg(feature = "mint")]
use mint;
#[repr(C)]
#[derive(Copy, Clone, Debug, PartialEq)]
#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
pub struct Quaternion<S> {
pub s: S,
pub v: Vector3<S>,
}
#[cfg(feature = "simd")]
impl From<Simdf32x4> for Quaternion<f32> {
#[inline]
fn from(f: Simdf32x4) -> Self {
unsafe {
let mut ret: Self = mem::uninitialized();
{
let ret_mut: &mut [f32; 4] = ret.as_mut();
f.store(ret_mut.as_mut(), 0 as usize);
}
ret
}
}
}
#[cfg(feature = "simd")]
impl Into<Simdf32x4> for Quaternion<f32> {
#[inline]
fn into(self) -> Simdf32x4 {
let self_ref: &[f32; 4] = self.as_ref();
Simdf32x4::load(self_ref.as_ref(), 0 as usize)
}
}
impl<S> Quaternion<S> {
#[inline]
pub const fn new(w: S, xi: S, yj: S, zk: S) -> Quaternion<S> {
Quaternion::from_sv(w, Vector3::new(xi, yj, zk))
}
#[inline]
pub const fn from_sv(s: S, v: Vector3<S>) -> Quaternion<S> {
Quaternion { s: s, v: v }
}
}
impl<S: BaseFloat> Quaternion<S> {
pub fn from_arc(
src: Vector3<S>,
dst: Vector3<S>,
fallback: Option<Vector3<S>>,
) -> Quaternion<S> {
let mag_avg = (src.magnitude2() * dst.magnitude2()).sqrt();
let dot = src.dot(dst);
if ulps_eq!(dot, &mag_avg) {
Quaternion::<S>::one()
} else if ulps_eq!(dot, &-mag_avg) {
let axis = fallback.unwrap_or_else(|| {
let mut v = Vector3::unit_x().cross(src);
if ulps_eq!(v, &Zero::zero()) {
v = Vector3::unit_y().cross(src);
}
v.normalize()
});
Quaternion::from_axis_angle(axis, Rad::turn_div_2())
} else {
Quaternion::from_sv(mag_avg + dot, src.cross(dst)).normalize()
}
}
#[inline]
pub fn conjugate(self) -> Quaternion<S> {
Quaternion::from_sv(self.s, -self.v)
}
pub fn nlerp(self, other: Quaternion<S>, amount: S) -> Quaternion<S> {
(self * (S::one() - amount) + other * amount).normalize()
}
pub fn slerp(self, other: Quaternion<S>, amount: S) -> Quaternion<S> {
let dot = self.dot(other);
let dot_threshold = cast(0.9995f64).unwrap();
if dot > dot_threshold {
self.nlerp(other, amount)
} else {
let robust_dot = if dot > S::one() {
S::one()
} else if dot < -S::one() {
-S::one()
} else {
dot
};
let theta = Rad::acos(robust_dot.clone());
let scale1 = Rad::sin(theta * (S::one() - amount));
let scale2 = Rad::sin(theta * amount);
(self * scale1 + other * scale2) * Rad::sin(theta).recip()
}
}
pub fn is_finite(&self) -> bool {
self.s.is_finite() && self.v.is_finite()
}
}
impl<S: BaseFloat> Zero for Quaternion<S> {
#[inline]
fn zero() -> Quaternion<S> {
Quaternion::from_sv(S::zero(), Vector3::zero())
}
#[inline]
fn is_zero(&self) -> bool {
ulps_eq!(self, &Quaternion::<S>::zero())
}
}
impl<S: BaseFloat> One for Quaternion<S> {
#[inline]
fn one() -> Quaternion<S> {
Quaternion::from_sv(S::one(), Vector3::zero())
}
}
impl<S: BaseFloat> iter::Sum<Quaternion<S>> for Quaternion<S> {
#[inline]
fn sum<I: Iterator<Item = Quaternion<S>>>(iter: I) -> Quaternion<S> {
iter.fold(Quaternion::<S>::zero(), Add::add)
}
}
impl<'a, S: 'a + BaseFloat> iter::Sum<&'a Quaternion<S>> for Quaternion<S> {
#[inline]
fn sum<I: Iterator<Item = &'a Quaternion<S>>>(iter: I) -> Quaternion<S> {
iter.fold(Quaternion::<S>::zero(), Add::add)
}
}
impl<S: BaseFloat> iter::Product<Quaternion<S>> for Quaternion<S> {
#[inline]
fn product<I: Iterator<Item = Quaternion<S>>>(iter: I) -> Quaternion<S> {
iter.fold(Quaternion::<S>::one(), Mul::mul)
}
}
impl<'a, S: 'a + BaseFloat> iter::Product<&'a Quaternion<S>> for Quaternion<S> {
#[inline]
fn product<I: Iterator<Item = &'a Quaternion<S>>>(iter: I) -> Quaternion<S> {
iter.fold(Quaternion::<S>::one(), Mul::mul)
}
}
impl<S: BaseFloat> VectorSpace for Quaternion<S> {
type Scalar = S;
}
impl<S: BaseFloat> MetricSpace for Quaternion<S> {
type Metric = S;
#[inline]
fn distance2(self, other: Self) -> S {
(other - self).magnitude2()
}
}
impl<S: NumCast + Copy> Quaternion<S> {
pub fn cast<T: BaseFloat>(&self) -> Option<Quaternion<T>> {
let s = match NumCast::from(self.s) {
Some(s) => s,
None => return None,
};
let v = match self.v.cast() {
Some(v) => v,
None => return None,
};
Some(Quaternion::from_sv(s, v))
}
}
#[cfg(not(feature = "simd"))]
impl<S: BaseFloat> InnerSpace for Quaternion<S> {
#[inline]
fn dot(self, other: Quaternion<S>) -> S {
self.s * other.s + self.v.dot(other.v)
}
}
#[cfg(feature = "simd")]
impl<S: BaseFloat> InnerSpace for Quaternion<S> {
#[inline]
default fn dot(self, other: Quaternion<S>) -> S {
self.s * other.s + self.v.dot(other.v)
}
}
#[cfg(feature = "simd")]
impl InnerSpace for Quaternion<f32> {
#[inline]
fn dot(self, other: Quaternion<f32>) -> f32 {
let lhs: Simdf32x4 = self.into();
let rhs: Simdf32x4 = other.into();
let r = lhs * rhs;
r.extract(0) + r.extract(1) + r.extract(2) + r.extract(3)
}
}
impl<A> From<Euler<A>> for Quaternion<A::Unitless>
where
A: Angle + Into<Rad<<A as Angle>::Unitless>>,
{
fn from(src: Euler<A>) -> Quaternion<A::Unitless> {
let half = cast(0.5f64).unwrap();
let (s_x, c_x) = Rad::sin_cos(src.x.into() * half);
let (s_y, c_y) = Rad::sin_cos(src.y.into() * half);
let (s_z, c_z) = Rad::sin_cos(src.z.into() * half);
Quaternion::new(
-s_x * s_y * s_z + c_x * c_y * c_z,
s_x * c_y * c_z + s_y * s_z * c_x,
-s_x * s_z * c_y + s_y * c_x * c_z,
s_x * s_y * c_z + s_z * c_x * c_y,
)
}
}
#[cfg(not(feature = "simd"))]
impl_operator!(<S: BaseFloat> Neg for Quaternion<S> {
fn neg(quat) -> Quaternion<S> {
Quaternion::from_sv(-quat.s, -quat.v)
}
});
#[cfg(feature = "simd")]
impl_operator_default!(<S: BaseFloat> Neg for Quaternion<S> {
fn neg(quat) -> Quaternion<S> {
Quaternion::from_sv(-quat.s, -quat.v)
}
});
#[cfg(feature = "simd")]
impl_operator_simd!{
[Simdf32x4]; Neg for Quaternion<f32> {
fn neg(lhs) -> Quaternion<f32> {
(-lhs).into()
}
}
}
#[cfg(not(feature = "simd"))]
impl_operator!(<S: BaseFloat> Mul<S> for Quaternion<S> {
fn mul(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s * rhs, lhs.v * rhs)
}
});
#[cfg(feature = "simd")]
impl_operator_default!(<S: BaseFloat> Mul<S> for Quaternion<S> {
fn mul(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s * rhs, lhs.v * rhs)
}
});
#[cfg(feature = "simd")]
impl_operator_simd!{@rs
[Simdf32x4]; Mul<f32> for Quaternion<f32> {
fn mul(lhs, rhs) -> Quaternion<f32> {
(lhs * rhs).into()
}
}
}
#[cfg(not(feature = "simd"))]
impl_assignment_operator!(<S: BaseFloat> MulAssign<S> for Quaternion<S> {
fn mul_assign(&mut self, scalar) { self.s *= scalar; self.v *= scalar; }
});
#[cfg(feature = "simd")]
impl_assignment_operator_default!(<S: BaseFloat> MulAssign<S> for Quaternion<S> {
fn mul_assign(&mut self, scalar) { self.s *= scalar; self.v *= scalar; }
});
#[cfg(feature = "simd")]
impl MulAssign<f32> for Quaternion<f32> {
fn mul_assign(&mut self, other: f32) {
let s: Simdf32x4 = (*self).into();
let other = Simdf32x4::splat(other);
*self = (s * other).into();
}
}
#[cfg(not(feature = "simd"))]
impl_operator!(<S: BaseFloat> Div<S> for Quaternion<S> {
fn div(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s / rhs, lhs.v / rhs)
}
});
#[cfg(feature = "simd")]
impl_operator_default!(<S: BaseFloat> Div<S> for Quaternion<S> {
fn div(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s / rhs, lhs.v / rhs)
}
});
#[cfg(feature = "simd")]
impl_operator_simd!{@rs
[Simdf32x4]; Div<f32> for Quaternion<f32> {
fn div(lhs, rhs) -> Quaternion<f32> {
(lhs / rhs).into()
}
}
}
#[cfg(not(feature = "simd"))]
impl_assignment_operator!(<S: BaseFloat> DivAssign<S> for Quaternion<S> {
fn div_assign(&mut self, scalar) { self.s /= scalar; self.v /= scalar; }
});
#[cfg(feature = "simd")]
impl_assignment_operator_default!(<S: BaseFloat> DivAssign<S> for Quaternion<S> {
fn div_assign(&mut self, scalar) { self.s /= scalar; self.v /= scalar; }
});
#[cfg(feature = "simd")]
impl DivAssign<f32> for Quaternion<f32> {
fn div_assign(&mut self, other: f32) {
let s: Simdf32x4 = (*self).into();
let other = Simdf32x4::splat(other);
*self = (s / other).into();
}
}
impl_operator!(<S: BaseFloat> Rem<S> for Quaternion<S> {
fn rem(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s % rhs, lhs.v % rhs)
}
});
impl_assignment_operator!(<S: BaseFloat> RemAssign<S> for Quaternion<S> {
fn rem_assign(&mut self, scalar) { self.s %= scalar; self.v %= scalar; }
});
impl_operator!(<S: BaseFloat> Mul<Vector3<S> > for Quaternion<S> {
fn mul(lhs, rhs) -> Vector3<S> {{
let rhs = rhs.clone();
let two: S = cast(2i8).unwrap();
let tmp = lhs.v.cross(rhs) + (rhs * lhs.s);
(lhs.v.cross(tmp) * two) + rhs
}}
});
#[cfg(not(feature = "simd"))]
impl_operator!(<S: BaseFloat> Add<Quaternion<S> > for Quaternion<S> {
fn add(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s + rhs.s, lhs.v + rhs.v)
}
});
#[cfg(feature = "simd")]
impl_operator_default!(<S: BaseFloat> Add<Quaternion<S> > for Quaternion<S> {
fn add(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s + rhs.s, lhs.v + rhs.v)
}
});
#[cfg(feature = "simd")]
impl_operator_simd!{
[Simdf32x4]; Add<Quaternion<f32>> for Quaternion<f32> {
fn add(lhs, rhs) -> Quaternion<f32> {
(lhs + rhs).into()
}
}
}
#[cfg(not(feature = "simd"))]
impl_assignment_operator!(<S: BaseFloat> AddAssign<Quaternion<S> > for Quaternion<S> {
fn add_assign(&mut self, other) { self.s += other.s; self.v += other.v; }
});
#[cfg(feature = "simd")]
impl_assignment_operator_default!(<S: BaseFloat> AddAssign<Quaternion<S> > for Quaternion<S> {
fn add_assign(&mut self, other) { self.s += other.s; self.v += other.v; }
});
#[cfg(feature = "simd")]
impl AddAssign for Quaternion<f32> {
#[inline]
fn add_assign(&mut self, rhs: Self) {
let s: Simdf32x4 = (*self).into();
let rhs: Simdf32x4 = rhs.into();
*self = (s + rhs).into();
}
}
#[cfg(not(feature = "simd"))]
impl_operator!(<S: BaseFloat> Sub<Quaternion<S> > for Quaternion<S> {
fn sub(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s - rhs.s, lhs.v - rhs.v)
}
});
#[cfg(feature = "simd")]
impl_operator_default!(<S: BaseFloat> Sub<Quaternion<S> > for Quaternion<S> {
fn sub(lhs, rhs) -> Quaternion<S> {
Quaternion::from_sv(lhs.s - rhs.s, lhs.v - rhs.v)
}
});
#[cfg(feature = "simd")]
impl_operator_simd!{
[Simdf32x4]; Sub<Quaternion<f32>> for Quaternion<f32> {
fn sub(lhs, rhs) -> Quaternion<f32> {
(lhs - rhs).into()
}
}
}
#[cfg(not(feature = "simd"))]
impl_assignment_operator!(<S: BaseFloat> SubAssign<Quaternion<S> > for Quaternion<S> {
fn sub_assign(&mut self, other) { self.s -= other.s; self.v -= other.v; }
});
#[cfg(feature = "simd")]
impl_assignment_operator_default!(<S: BaseFloat> SubAssign<Quaternion<S> > for Quaternion<S> {
fn sub_assign(&mut self, other) { self.s -= other.s; self.v -= other.v; }
});
#[cfg(feature = "simd")]
impl SubAssign for Quaternion<f32> {
#[inline]
fn sub_assign(&mut self, rhs: Self) {
let s: Simdf32x4 = (*self).into();
let rhs: Simdf32x4 = rhs.into();
*self = (s - rhs).into();
}
}
#[cfg(not(feature = "simd"))]
impl_operator!(<S: BaseFloat> Mul<Quaternion<S> > for Quaternion<S> {
fn mul(lhs, rhs) -> Quaternion<S> {
Quaternion::new(
lhs.s * rhs.s - lhs.v.x * rhs.v.x - lhs.v.y * rhs.v.y - lhs.v.z * rhs.v.z,
lhs.s * rhs.v.x + lhs.v.x * rhs.s + lhs.v.y * rhs.v.z - lhs.v.z * rhs.v.y,
lhs.s * rhs.v.y + lhs.v.y * rhs.s + lhs.v.z * rhs.v.x - lhs.v.x * rhs.v.z,
lhs.s * rhs.v.z + lhs.v.z * rhs.s + lhs.v.x * rhs.v.y - lhs.v.y * rhs.v.x,
)
}
});
#[cfg(feature = "simd")]
impl_operator_default!(<S: BaseFloat> Mul<Quaternion<S> > for Quaternion<S> {
fn mul(lhs, rhs) -> Quaternion<S> {
Quaternion::new(
lhs.s * rhs.s - lhs.v.x * rhs.v.x - lhs.v.y * rhs.v.y - lhs.v.z * rhs.v.z,
lhs.s * rhs.v.x + lhs.v.x * rhs.s + lhs.v.y * rhs.v.z - lhs.v.z * rhs.v.y,
lhs.s * rhs.v.y + lhs.v.y * rhs.s + lhs.v.z * rhs.v.x - lhs.v.x * rhs.v.z,
lhs.s * rhs.v.z + lhs.v.z * rhs.s + lhs.v.x * rhs.v.y - lhs.v.y * rhs.v.x,
)
}
});
#[cfg(feature = "simd")]
impl_operator_simd!{
[Simdf32x4]; Mul<Quaternion<f32>> for Quaternion<f32> {
fn mul(lhs, rhs) -> Quaternion<f32> {
{
let p0 = Simdf32x4::splat(lhs.extract(0)) * rhs;
let p1 = Simdf32x4::splat(lhs.extract(1)) * Simdf32x4::new(
-rhs.extract(1), rhs.extract(0), -rhs.extract(3), rhs.extract(2)
);
let p2 = Simdf32x4::splat(lhs.extract(2)) * Simdf32x4::new(
-rhs.extract(2), rhs.extract(3), rhs.extract(0), -rhs.extract(1)
);
let p3 = Simdf32x4::splat(lhs.extract(3)) * Simdf32x4::new(
-rhs.extract(3), -rhs.extract(2), rhs.extract(1), rhs.extract(0)
);
(p0 + p1 + p2 + p3).into()
}
}
}
}
macro_rules! impl_scalar_mul {
($S:ident) => {
impl_operator!(Mul<Quaternion<$S>> for $S {
fn mul(scalar, quat) -> Quaternion<$S> {
Quaternion::from_sv(scalar * quat.s, scalar * quat.v)
}
});
};
}
macro_rules! impl_scalar_div {
($S:ident) => {
impl_operator!(Div<Quaternion<$S>> for $S {
fn div(scalar, quat) -> Quaternion<$S> {
Quaternion::from_sv(scalar / quat.s, scalar / quat.v)
}
});
};
}
impl_scalar_mul!(f32);
impl_scalar_mul!(f64);
impl_scalar_div!(f32);
impl_scalar_div!(f64);
impl<S: BaseFloat> approx::AbsDiffEq for Quaternion<S> {
type Epsilon = S::Epsilon;
#[inline]
fn default_epsilon() -> S::Epsilon {
S::default_epsilon()
}
#[inline]
fn abs_diff_eq(&self, other: &Self, epsilon: S::Epsilon) -> bool {
S::abs_diff_eq(&self.s, &other.s, epsilon)
&& Vector3::abs_diff_eq(&self.v, &other.v, epsilon)
}
}
impl<S: BaseFloat> approx::RelativeEq for Quaternion<S> {
#[inline]
fn default_max_relative() -> S::Epsilon {
S::default_max_relative()
}
#[inline]
fn relative_eq(&self, other: &Self, epsilon: S::Epsilon, max_relative: S::Epsilon) -> bool {
S::relative_eq(&self.s, &other.s, epsilon, max_relative)
&& Vector3::relative_eq(&self.v, &other.v, epsilon, max_relative)
}
}
impl<S: BaseFloat> approx::UlpsEq for Quaternion<S> {
#[inline]
fn default_max_ulps() -> u32 {
S::default_max_ulps()
}
#[inline]
fn ulps_eq(&self, other: &Self, epsilon: S::Epsilon, max_ulps: u32) -> bool {
S::ulps_eq(&self.s, &other.s, epsilon, max_ulps)
&& Vector3::ulps_eq(&self.v, &other.v, epsilon, max_ulps)
}
}
impl<S: BaseFloat> From<Quaternion<S>> for Matrix3<S> {
fn from(quat: Quaternion<S>) -> Matrix3<S> {
let x2 = quat.v.x + quat.v.x;
let y2 = quat.v.y + quat.v.y;
let z2 = quat.v.z + quat.v.z;
let xx2 = x2 * quat.v.x;
let xy2 = x2 * quat.v.y;
let xz2 = x2 * quat.v.z;
let yy2 = y2 * quat.v.y;
let yz2 = y2 * quat.v.z;
let zz2 = z2 * quat.v.z;
let sy2 = y2 * quat.s;
let sz2 = z2 * quat.s;
let sx2 = x2 * quat.s;
#[cfg_attr(rustfmt, rustfmt_skip)]
Matrix3::new(
S::one() - yy2 - zz2, xy2 + sz2, xz2 - sy2,
xy2 - sz2, S::one() - xx2 - zz2, yz2 + sx2,
xz2 + sy2, yz2 - sx2, S::one() - xx2 - yy2,
)
}
}
impl<S: BaseFloat> From<Quaternion<S>> for Matrix4<S> {
fn from(quat: Quaternion<S>) -> Matrix4<S> {
let x2 = quat.v.x + quat.v.x;
let y2 = quat.v.y + quat.v.y;
let z2 = quat.v.z + quat.v.z;
let xx2 = x2 * quat.v.x;
let xy2 = x2 * quat.v.y;
let xz2 = x2 * quat.v.z;
let yy2 = y2 * quat.v.y;
let yz2 = y2 * quat.v.z;
let zz2 = z2 * quat.v.z;
let sy2 = y2 * quat.s;
let sz2 = z2 * quat.s;
let sx2 = x2 * quat.s;
#[cfg_attr(rustfmt, rustfmt_skip)]
Matrix4::new(
S::one() - yy2 - zz2, xy2 + sz2, xz2 - sy2, S::zero(),
xy2 - sz2, S::one() - xx2 - zz2, yz2 + sx2, S::zero(),
xz2 + sy2, yz2 - sx2, S::one() - xx2 - yy2, S::zero(),
S::zero(), S::zero(), S::zero(), S::one(),
)
}
}
impl<S: BaseFloat> From<Quaternion<S>> for Basis3<S> {
#[inline]
fn from(quat: Quaternion<S>) -> Basis3<S> {
Basis3::from_quaternion(&quat)
}
}
impl<S: BaseFloat> Rotation<Point3<S>> for Quaternion<S> {
#[inline]
fn look_at(dir: Vector3<S>, up: Vector3<S>) -> Quaternion<S> {
Matrix3::look_at(dir, up).into()
}
#[inline]
fn between_vectors(a: Vector3<S>, b: Vector3<S>) -> Quaternion<S> {
let k_cos_theta = a.dot(b);
if ulps_eq!(k_cos_theta, S::one()) {
return Quaternion::<S>::one();
}
let k = (a.magnitude2() * b.magnitude2()).sqrt();
if ulps_eq!(k_cos_theta / k, -S::one()) {
let mut orthogonal = a.cross(Vector3::unit_x());
if ulps_eq!(orthogonal.magnitude2(), S::zero()) {
orthogonal = a.cross(Vector3::unit_y());
}
return Quaternion::from_sv(S::zero(), orthogonal.normalize());
}
Quaternion::from_sv(k + k_cos_theta, a.cross(b)).normalize()
}
#[inline]
fn rotate_vector(&self, vec: Vector3<S>) -> Vector3<S> {
self * vec
}
#[inline]
fn invert(&self) -> Quaternion<S> {
self.conjugate() / self.magnitude2()
}
}
impl<S: BaseFloat> Rotation3<S> for Quaternion<S> {
#[inline]
fn from_axis_angle<A: Into<Rad<S>>>(axis: Vector3<S>, angle: A) -> Quaternion<S> {
let (s, c) = Rad::sin_cos(angle.into() * cast(0.5f64).unwrap());
Quaternion::from_sv(c, axis * s)
}
}
impl<S: BaseFloat> Into<[S; 4]> for Quaternion<S> {
#[inline]
fn into(self) -> [S; 4] {
match self.into() {
(w, xi, yj, zk) => [w, xi, yj, zk],
}
}
}
impl<S: BaseFloat> AsRef<[S; 4]> for Quaternion<S> {
#[inline]
fn as_ref(&self) -> &[S; 4] {
unsafe { mem::transmute(self) }
}
}
impl<S: BaseFloat> AsMut<[S; 4]> for Quaternion<S> {
#[inline]
fn as_mut(&mut self) -> &mut [S; 4] {
unsafe { mem::transmute(self) }
}
}
impl<S: BaseFloat> From<[S; 4]> for Quaternion<S> {
#[inline]
fn from(v: [S; 4]) -> Quaternion<S> {
Quaternion::new(v[0], v[1], v[2], v[3])
}
}
impl<'a, S: BaseFloat> From<&'a [S; 4]> for &'a Quaternion<S> {
#[inline]
fn from(v: &'a [S; 4]) -> &'a Quaternion<S> {
unsafe { mem::transmute(v) }
}
}
impl<'a, S: BaseFloat> From<&'a mut [S; 4]> for &'a mut Quaternion<S> {
#[inline]
fn from(v: &'a mut [S; 4]) -> &'a mut Quaternion<S> {
unsafe { mem::transmute(v) }
}
}
impl<S: BaseFloat> Into<(S, S, S, S)> for Quaternion<S> {
#[inline]
fn into(self) -> (S, S, S, S) {
match self {
Quaternion {
s,
v: Vector3 { x, y, z },
} => (s, x, y, z),
}
}
}
impl<S: BaseFloat> AsRef<(S, S, S, S)> for Quaternion<S> {
#[inline]
fn as_ref(&self) -> &(S, S, S, S) {
unsafe { mem::transmute(self) }
}
}
impl<S: BaseFloat> AsMut<(S, S, S, S)> for Quaternion<S> {
#[inline]
fn as_mut(&mut self) -> &mut (S, S, S, S) {
unsafe { mem::transmute(self) }
}
}
impl<S: BaseFloat> From<(S, S, S, S)> for Quaternion<S> {
#[inline]
fn from(v: (S, S, S, S)) -> Quaternion<S> {
match v {
(w, xi, yj, zk) => Quaternion::new(w, xi, yj, zk),
}
}
}
impl<'a, S: BaseFloat> From<&'a (S, S, S, S)> for &'a Quaternion<S> {
#[inline]
fn from(v: &'a (S, S, S, S)) -> &'a Quaternion<S> {
unsafe { mem::transmute(v) }
}
}
impl<'a, S: BaseFloat> From<&'a mut (S, S, S, S)> for &'a mut Quaternion<S> {
#[inline]
fn from(v: &'a mut (S, S, S, S)) -> &'a mut Quaternion<S> {
unsafe { mem::transmute(v) }
}
}
macro_rules! index_operators {
($S:ident, $Output:ty, $I:ty) => {
impl<$S: BaseFloat> Index<$I> for Quaternion<$S> {
type Output = $Output;
#[inline]
fn index<'a>(&'a self, i: $I) -> &'a $Output {
let v: &[$S; 4] = self.as_ref(); &v[i]
}
}
impl<$S: BaseFloat> IndexMut<$I> for Quaternion<$S> {
#[inline]
fn index_mut<'a>(&'a mut self, i: $I) -> &'a mut $Output {
let v: &mut [$S; 4] = self.as_mut(); &mut v[i]
}
}
}
}
index_operators!(S, S, usize);
index_operators!(S, [S], Range<usize>);
index_operators!(S, [S], RangeTo<usize>);
index_operators!(S, [S], RangeFrom<usize>);
index_operators!(S, [S], RangeFull);
impl<S> Distribution<Quaternion<S>> for Standard
where Standard: Distribution<S>,
Standard: Distribution<Vector3<S>>,
S: BaseFloat {
#[inline]
fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> Quaternion<S> {
Quaternion::from_sv(rng.gen(), rng.gen())
}
}
#[cfg(feature = "mint")]
impl<S> From<mint::Quaternion<S>> for Quaternion<S> {
fn from(q: mint::Quaternion<S>) -> Self {
Quaternion {
s: q.s,
v: q.v.into(),
}
}
}
#[cfg(feature = "mint")]
impl<S: Clone> Into<mint::Quaternion<S>> for Quaternion<S> {
fn into(self) -> mint::Quaternion<S> {
mint::Quaternion {
s: self.s,
v: self.v.into(),
}
}
}
#[cfg(test)]
mod tests {
use quaternion::*;
use vector::*;
const QUATERNION: Quaternion<f32> = Quaternion {
s: 1.0,
v: Vector3 {
x: 2.0,
y: 3.0,
z: 4.0,
},
};
#[test]
fn test_into() {
let v = QUATERNION;
{
let v: [f32; 4] = v.into();
assert_eq!(v, [1.0, 2.0, 3.0, 4.0]);
}
{
let v: (f32, f32, f32, f32) = v.into();
assert_eq!(v, (1.0, 2.0, 3.0, 4.0));
}
}
#[test]
fn test_as_ref() {
let v = QUATERNION;
{
let v: &[f32; 4] = v.as_ref();
assert_eq!(v, &[1.0, 2.0, 3.0, 4.0]);
}
{
let v: &(f32, f32, f32, f32) = v.as_ref();
assert_eq!(v, &(1.0, 2.0, 3.0, 4.0));
}
}
#[test]
fn test_as_mut() {
let mut v = QUATERNION;
{
let v: &mut [f32; 4] = v.as_mut();
assert_eq!(v, &mut [1.0, 2.0, 3.0, 4.0]);
}
{
let v: &mut (f32, f32, f32, f32) = v.as_mut();
assert_eq!(v, &mut (1.0, 2.0, 3.0, 4.0));
}
}
#[test]
fn test_from() {
assert_eq!(Quaternion::from([1.0, 2.0, 3.0, 4.0]), QUATERNION);
{
let v = &[1.0, 2.0, 3.0, 4.0];
let v: &Quaternion<_> = From::from(v);
assert_eq!(v, &QUATERNION);
}
{
let v = &mut [1.0, 2.0, 3.0, 4.0];
let v: &mut Quaternion<_> = From::from(v);
assert_eq!(v, &QUATERNION);
}
assert_eq!(Quaternion::from((1.0, 2.0, 3.0, 4.0)), QUATERNION);
{
let v = &(1.0, 2.0, 3.0, 4.0);
let v: &Quaternion<_> = From::from(v);
assert_eq!(v, &QUATERNION);
}
{
let v = &mut (1.0, 2.0, 3.0, 4.0);
let v: &mut Quaternion<_> = From::from(v);
assert_eq!(v, &QUATERNION);
}
}
}