/*************************************************************************/ /* vector3.h */ /*************************************************************************/ /* This file is part of: */ /* GODOT ENGINE */ /* https://godotengine.org */ /*************************************************************************/ /* Copyright (c) 2007-2022 Juan Linietsky, Ariel Manzur. */ /* Copyright (c) 2014-2022 Godot Engine contributors (cf. AUTHORS.md). */ /* */ /* Permission is hereby granted, free of charge, to any person obtaining */ /* a copy of this software and associated documentation files (the */ /* "Software"), to deal in the Software without restriction, including */ /* without limitation the rights to use, copy, modify, merge, publish, */ /* distribute, sublicense, and/or sell copies of the Software, and to */ /* permit persons to whom the Software is furnished to do so, subject to */ /* the following conditions: */ /* */ /* The above copyright notice and this permission notice shall be */ /* included in all copies or substantial portions of the Software. */ /* */ /* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */ /* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */ /* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.*/ /* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */ /* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */ /* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */ /* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ /*************************************************************************/ #ifndef VECTOR3_H #define VECTOR3_H #include "core/error/error_macros.h" #include "core/math/math_funcs.h" class String; struct Basis; struct Vector2; struct Vector3i; struct _NO_DISCARD_ Vector3 { static const int AXIS_COUNT = 3; enum Axis { AXIS_X, AXIS_Y, AXIS_Z, }; union { struct { real_t x; real_t y; real_t z; }; real_t coord[3] = { 0 }; }; _FORCE_INLINE_ const real_t &operator[](const int p_axis) const { DEV_ASSERT((unsigned int)p_axis < 3); return coord[p_axis]; } _FORCE_INLINE_ real_t &operator[](const int p_axis) { DEV_ASSERT((unsigned int)p_axis < 3); return coord[p_axis]; } _FORCE_INLINE_ Vector3::Axis min_axis_index() const { return x < y ? (x < z ? Vector3::AXIS_X : Vector3::AXIS_Z) : (y < z ? Vector3::AXIS_Y : Vector3::AXIS_Z); } _FORCE_INLINE_ Vector3::Axis max_axis_index() const { return x < y ? (y < z ? Vector3::AXIS_Z : Vector3::AXIS_Y) : (x < z ? Vector3::AXIS_Z : Vector3::AXIS_X); } _FORCE_INLINE_ real_t length() const; _FORCE_INLINE_ real_t length_squared() const; _FORCE_INLINE_ void normalize(); _FORCE_INLINE_ Vector3 normalized() const; _FORCE_INLINE_ bool is_normalized() const; _FORCE_INLINE_ Vector3 inverse() const; Vector3 limit_length(const real_t p_len = 1.0) const; _FORCE_INLINE_ void zero(); void snap(const Vector3 p_val); Vector3 snapped(const Vector3 p_val) const; void rotate(const Vector3 &p_axis, const real_t p_angle); Vector3 rotated(const Vector3 &p_axis, const real_t p_angle) const; /* Static Methods between 2 vector3s */ _FORCE_INLINE_ Vector3 lerp(const Vector3 &p_to, const real_t p_weight) const; _FORCE_INLINE_ Vector3 slerp(const Vector3 &p_to, const real_t p_weight) const; _FORCE_INLINE_ Vector3 cubic_interpolate(const Vector3 &p_b, const Vector3 &p_pre_a, const Vector3 &p_post_b, const real_t p_weight) const; _FORCE_INLINE_ Vector3 cubic_interpolate_in_time(const Vector3 &p_b, const Vector3 &p_pre_a, const Vector3 &p_post_b, const real_t p_weight, const real_t &p_b_t, const real_t &p_pre_a_t, const real_t &p_post_b_t) const; _FORCE_INLINE_ Vector3 bezier_interpolate(const Vector3 &p_control_1, const Vector3 &p_control_2, const Vector3 &p_end, const real_t p_t) const; _FORCE_INLINE_ Vector3 bezier_derivative(const Vector3 &p_control_1, const Vector3 &p_control_2, const Vector3 &p_end, const real_t p_t) const; Vector3 move_toward(const Vector3 &p_to, const real_t p_delta) const; Vector2 octahedron_encode() const; static Vector3 octahedron_decode(const Vector2 &p_oct); Vector2 octahedron_tangent_encode(const float sign) const; static Vector3 octahedron_tangent_decode(const Vector2 &p_oct, float *sign); _FORCE_INLINE_ Vector3 cross(const Vector3 &p_with) const; _FORCE_INLINE_ real_t dot(const Vector3 &p_with) const; Basis outer(const Vector3 &p_with) const; _FORCE_INLINE_ Vector3 abs() const; _FORCE_INLINE_ Vector3 floor() const; _FORCE_INLINE_ Vector3 sign() const; _FORCE_INLINE_ Vector3 ceil() const; _FORCE_INLINE_ Vector3 round() const; Vector3 clamp(const Vector3 &p_min, const Vector3 &p_max) const; _FORCE_INLINE_ real_t distance_to(const Vector3 &p_to) const; _FORCE_INLINE_ real_t distance_squared_to(const Vector3 &p_to) const; _FORCE_INLINE_ Vector3 posmod(const real_t p_mod) const; _FORCE_INLINE_ Vector3 posmodv(const Vector3 &p_modv) const; _FORCE_INLINE_ Vector3 project(const Vector3 &p_to) const; _FORCE_INLINE_ real_t angle_to(const Vector3 &p_to) const; _FORCE_INLINE_ real_t signed_angle_to(const Vector3 &p_to, const Vector3 &p_axis) const; _FORCE_INLINE_ Vector3 direction_to(const Vector3 &p_to) const; _FORCE_INLINE_ Vector3 slide(const Vector3 &p_normal) const; _FORCE_INLINE_ Vector3 bounce(const Vector3 &p_normal) const; _FORCE_INLINE_ Vector3 reflect(const Vector3 &p_normal) const; bool is_equal_approx(const Vector3 &p_v) const; bool is_zero_approx() const; bool is_finite() const; /* Operators */ _FORCE_INLINE_ Vector3 &operator+=(const Vector3 &p_v); _FORCE_INLINE_ Vector3 operator+(const Vector3 &p_v) const; _FORCE_INLINE_ Vector3 &operator-=(const Vector3 &p_v); _FORCE_INLINE_ Vector3 operator-(const Vector3 &p_v) const; _FORCE_INLINE_ Vector3 &operator*=(const Vector3 &p_v); _FORCE_INLINE_ Vector3 operator*(const Vector3 &p_v) const; _FORCE_INLINE_ Vector3 &operator/=(const Vector3 &p_v); _FORCE_INLINE_ Vector3 operator/(const Vector3 &p_v) const; _FORCE_INLINE_ Vector3 &operator*=(const real_t p_scalar); _FORCE_INLINE_ Vector3 operator*(const real_t p_scalar) const; _FORCE_INLINE_ Vector3 &operator/=(const real_t p_scalar); _FORCE_INLINE_ Vector3 operator/(const real_t p_scalar) const; _FORCE_INLINE_ Vector3 operator-() const; _FORCE_INLINE_ bool operator==(const Vector3 &p_v) const; _FORCE_INLINE_ bool operator!=(const Vector3 &p_v) const; _FORCE_INLINE_ bool operator<(const Vector3 &p_v) const; _FORCE_INLINE_ bool operator<=(const Vector3 &p_v) const; _FORCE_INLINE_ bool operator>(const Vector3 &p_v) const; _FORCE_INLINE_ bool operator>=(const Vector3 &p_v) const; operator String() const; operator Vector3i() const; _FORCE_INLINE_ Vector3() {} _FORCE_INLINE_ Vector3(const real_t p_x, const real_t p_y, const real_t p_z) { x = p_x; y = p_y; z = p_z; } }; Vector3 Vector3::cross(const Vector3 &p_with) const { Vector3 ret( (y * p_with.z) - (z * p_with.y), (z * p_with.x) - (x * p_with.z), (x * p_with.y) - (y * p_with.x)); return ret; } real_t Vector3::dot(const Vector3 &p_with) const { return x * p_with.x + y * p_with.y + z * p_with.z; } Vector3 Vector3::abs() const { return Vector3(Math::abs(x), Math::abs(y), Math::abs(z)); } Vector3 Vector3::sign() const { return Vector3(SIGN(x), SIGN(y), SIGN(z)); } Vector3 Vector3::floor() const { return Vector3(Math::floor(x), Math::floor(y), Math::floor(z)); } Vector3 Vector3::ceil() const { return Vector3(Math::ceil(x), Math::ceil(y), Math::ceil(z)); } Vector3 Vector3::round() const { return Vector3(Math::round(x), Math::round(y), Math::round(z)); } Vector3 Vector3::lerp(const Vector3 &p_to, const real_t p_weight) const { Vector3 res = *this; res.x = Math::lerp(res.x, p_to.x, p_weight); res.y = Math::lerp(res.y, p_to.y, p_weight); res.z = Math::lerp(res.z, p_to.z, p_weight); return res; } Vector3 Vector3::slerp(const Vector3 &p_to, const real_t p_weight) const { // This method seems more complicated than it really is, since we write out // the internals of some methods for efficiency (mainly, checking length). real_t start_length_sq = length_squared(); real_t end_length_sq = p_to.length_squared(); if (unlikely(start_length_sq == 0.0f || end_length_sq == 0.0f)) { // Zero length vectors have no angle, so the best we can do is either lerp or throw an error. return lerp(p_to, p_weight); } Vector3 axis = cross(p_to); real_t axis_length_sq = axis.length_squared(); if (unlikely(axis_length_sq == 0.0f)) { // Colinear vectors have no rotation axis or angle between them, so the best we can do is lerp. return lerp(p_to, p_weight); } axis /= Math::sqrt(axis_length_sq); real_t start_length = Math::sqrt(start_length_sq); real_t result_length = Math::lerp(start_length, Math::sqrt(end_length_sq), p_weight); real_t angle = angle_to(p_to); return rotated(axis, angle * p_weight) * (result_length / start_length); } Vector3 Vector3::cubic_interpolate(const Vector3 &p_b, const Vector3 &p_pre_a, const Vector3 &p_post_b, const real_t p_weight) const { Vector3 res = *this; res.x = Math::cubic_interpolate(res.x, p_b.x, p_pre_a.x, p_post_b.x, p_weight); res.y = Math::cubic_interpolate(res.y, p_b.y, p_pre_a.y, p_post_b.y, p_weight); res.z = Math::cubic_interpolate(res.z, p_b.z, p_pre_a.z, p_post_b.z, p_weight); return res; } Vector3 Vector3::cubic_interpolate_in_time(const Vector3 &p_b, const Vector3 &p_pre_a, const Vector3 &p_post_b, const real_t p_weight, const real_t &p_b_t, const real_t &p_pre_a_t, const real_t &p_post_b_t) const { Vector3 res = *this; res.x = Math::cubic_interpolate_in_time(res.x, p_b.x, p_pre_a.x, p_post_b.x, p_weight, p_b_t, p_pre_a_t, p_post_b_t); res.y = Math::cubic_interpolate_in_time(res.y, p_b.y, p_pre_a.y, p_post_b.y, p_weight, p_b_t, p_pre_a_t, p_post_b_t); res.z = Math::cubic_interpolate_in_time(res.z, p_b.z, p_pre_a.z, p_post_b.z, p_weight, p_b_t, p_pre_a_t, p_post_b_t); return res; } Vector3 Vector3::bezier_interpolate(const Vector3 &p_control_1, const Vector3 &p_control_2, const Vector3 &p_end, const real_t p_t) const { Vector3 res = *this; res.x = Math::bezier_interpolate(res.x, p_control_1.x, p_control_2.x, p_end.x, p_t); res.y = Math::bezier_interpolate(res.y, p_control_1.y, p_control_2.y, p_end.y, p_t); res.z = Math::bezier_interpolate(res.z, p_control_1.z, p_control_2.z, p_end.z, p_t); return res; } Vector3 Vector3::bezier_derivative(const Vector3 &p_control_1, const Vector3 &p_control_2, const Vector3 &p_end, const real_t p_t) const { Vector3 res = *this; res.x = Math::bezier_derivative(res.x, p_control_1.x, p_control_2.x, p_end.x, p_t); res.y = Math::bezier_derivative(res.y, p_control_1.y, p_control_2.y, p_end.y, p_t); res.z = Math::bezier_derivative(res.z, p_control_1.z, p_control_2.z, p_end.z, p_t); return res; } real_t Vector3::distance_to(const Vector3 &p_to) const { return (p_to - *this).length(); } real_t Vector3::distance_squared_to(const Vector3 &p_to) const { return (p_to - *this).length_squared(); } Vector3 Vector3::posmod(const real_t p_mod) const { return Vector3(Math::fposmod(x, p_mod), Math::fposmod(y, p_mod), Math::fposmod(z, p_mod)); } Vector3 Vector3::posmodv(const Vector3 &p_modv) const { return Vector3(Math::fposmod(x, p_modv.x), Math::fposmod(y, p_modv.y), Math::fposmod(z, p_modv.z)); } Vector3 Vector3::project(const Vector3 &p_to) const { return p_to * (dot(p_to) / p_to.length_squared()); } real_t Vector3::angle_to(const Vector3 &p_to) const { return Math::atan2(cross(p_to).length(), dot(p_to)); } real_t Vector3::signed_angle_to(const Vector3 &p_to, const Vector3 &p_axis) const { Vector3 cross_to = cross(p_to); real_t unsigned_angle = Math::atan2(cross_to.length(), dot(p_to)); real_t sign = cross_to.dot(p_axis); return (sign < 0) ? -unsigned_angle : unsigned_angle; } Vector3 Vector3::direction_to(const Vector3 &p_to) const { Vector3 ret(p_to.x - x, p_to.y - y, p_to.z - z); ret.normalize(); return ret; } /* Operators */ Vector3 &Vector3::operator+=(const Vector3 &p_v) { x += p_v.x; y += p_v.y; z += p_v.z; return *this; } Vector3 Vector3::operator+(const Vector3 &p_v) const { return Vector3(x + p_v.x, y + p_v.y, z + p_v.z); } Vector3 &Vector3::operator-=(const Vector3 &p_v) { x -= p_v.x; y -= p_v.y; z -= p_v.z; return *this; } Vector3 Vector3::operator-(const Vector3 &p_v) const { return Vector3(x - p_v.x, y - p_v.y, z - p_v.z); } Vector3 &Vector3::operator*=(const Vector3 &p_v) { x *= p_v.x; y *= p_v.y; z *= p_v.z; return *this; } Vector3 Vector3::operator*(const Vector3 &p_v) const { return Vector3(x * p_v.x, y * p_v.y, z * p_v.z); } Vector3 &Vector3::operator/=(const Vector3 &p_v) { x /= p_v.x; y /= p_v.y; z /= p_v.z; return *this; } Vector3 Vector3::operator/(const Vector3 &p_v) const { return Vector3(x / p_v.x, y / p_v.y, z / p_v.z); } Vector3 &Vector3::operator*=(const real_t p_scalar) { x *= p_scalar; y *= p_scalar; z *= p_scalar; return *this; } // Multiplication operators required to workaround issues with LLVM using implicit conversion // to Vector3i instead for integers where it should not. _FORCE_INLINE_ Vector3 operator*(const float p_scalar, const Vector3 &p_vec) { return p_vec * p_scalar; } _FORCE_INLINE_ Vector3 operator*(const double p_scalar, const Vector3 &p_vec) { return p_vec * p_scalar; } _FORCE_INLINE_ Vector3 operator*(const int32_t p_scalar, const Vector3 &p_vec) { return p_vec * p_scalar; } _FORCE_INLINE_ Vector3 operator*(const int64_t p_scalar, const Vector3 &p_vec) { return p_vec * p_scalar; } Vector3 Vector3::operator*(const real_t p_scalar) const { return Vector3(x * p_scalar, y * p_scalar, z * p_scalar); } Vector3 &Vector3::operator/=(const real_t p_scalar) { x /= p_scalar; y /= p_scalar; z /= p_scalar; return *this; } Vector3 Vector3::operator/(const real_t p_scalar) const { return Vector3(x / p_scalar, y / p_scalar, z / p_scalar); } Vector3 Vector3::operator-() const { return Vector3(-x, -y, -z); } bool Vector3::operator==(const Vector3 &p_v) const { return x == p_v.x && y == p_v.y && z == p_v.z; } bool Vector3::operator!=(const Vector3 &p_v) const { return x != p_v.x || y != p_v.y || z != p_v.z; } bool Vector3::operator<(const Vector3 &p_v) const { if (x == p_v.x) { if (y == p_v.y) { return z < p_v.z; } return y < p_v.y; } return x < p_v.x; } bool Vector3::operator>(const Vector3 &p_v) const { if (x == p_v.x) { if (y == p_v.y) { return z > p_v.z; } return y > p_v.y; } return x > p_v.x; } bool Vector3::operator<=(const Vector3 &p_v) const { if (x == p_v.x) { if (y == p_v.y) { return z <= p_v.z; } return y < p_v.y; } return x < p_v.x; } bool Vector3::operator>=(const Vector3 &p_v) const { if (x == p_v.x) { if (y == p_v.y) { return z >= p_v.z; } return y > p_v.y; } return x > p_v.x; } _FORCE_INLINE_ Vector3 vec3_cross(const Vector3 &p_a, const Vector3 &p_b) { return p_a.cross(p_b); } _FORCE_INLINE_ real_t vec3_dot(const Vector3 &p_a, const Vector3 &p_b) { return p_a.dot(p_b); } real_t Vector3::length() const { real_t x2 = x * x; real_t y2 = y * y; real_t z2 = z * z; return Math::sqrt(x2 + y2 + z2); } real_t Vector3::length_squared() const { real_t x2 = x * x; real_t y2 = y * y; real_t z2 = z * z; return x2 + y2 + z2; } void Vector3::normalize() { real_t lengthsq = length_squared(); if (lengthsq == 0) { x = y = z = 0; } else { real_t length = Math::sqrt(lengthsq); x /= length; y /= length; z /= length; } } Vector3 Vector3::normalized() const { Vector3 v = *this; v.normalize(); return v; } bool Vector3::is_normalized() const { // use length_squared() instead of length() to avoid sqrt(), makes it more stringent. return Math::is_equal_approx(length_squared(), 1, (real_t)UNIT_EPSILON); } Vector3 Vector3::inverse() const { return Vector3(1.0f / x, 1.0f / y, 1.0f / z); } void Vector3::zero() { x = y = z = 0; } // slide returns the component of the vector along the given plane, specified by its normal vector. Vector3 Vector3::slide(const Vector3 &p_normal) const { #ifdef MATH_CHECKS ERR_FAIL_COND_V_MSG(!p_normal.is_normalized(), Vector3(), "The normal Vector3 must be normalized."); #endif return *this - p_normal * this->dot(p_normal); } Vector3 Vector3::bounce(const Vector3 &p_normal) const { return -reflect(p_normal); } Vector3 Vector3::reflect(const Vector3 &p_normal) const { #ifdef MATH_CHECKS ERR_FAIL_COND_V_MSG(!p_normal.is_normalized(), Vector3(), "The normal Vector3 must be normalized."); #endif return 2.0f * p_normal * this->dot(p_normal) - *this; } #endif // VECTOR3_H