lionkor / vector2

A templated header-only 2D Vector class

Vector2

No longer in development. Use glm instead.

_{Contact: [email protected]}

• Overview
• Features
• Development Progress
• Notes
• How to use
  • Construct a vector
  • Vector operations

Overview

This is a templated header-only class for a 2D Vector, written in C++17, with all common vector operations and some useful additional methods.

Features

• Header-only, MIT licensed, templated.

• All equality- and relational operators

• Supported operations:

  • + - * / with another vector
  • + - * / with any scalar (templated)
  • rotation (deg & rad)
  • reflection (using surface normal)
  • distance & squared distance
  • magnitude (length) & squared magnitude
  • normalization
  • dot product
  • linear interpolation (clamped & unclamped)
  • square root

• Typedefs for all primitive types (Vector2f, Vector2i, Vector2d, etc.)

Development Progress

• Common operators
• Templated operators
• License (MIT)
• Full in-source documentation
• Common operations like normalizing, dot/cross/scalar product, distance, magnitude, etc.
• Extra fancy operations like linear interpolation, abs, rotation, reflection

Notes

• Floating point equality comparisons have been implemented roughly following this article. Feel free to override the functions used by #define-ing VECTOR2_FLOAT_COMPARE, VECTOR2_DOUBLE_COMPARE and VECTOR2_LONG_DOUBLE_COMPARE for float, double and long double comparisons, respectively. They're used like FUNCTION(first_float, second_float).
• Vector-rotation is implemented to rotate CLOCKWISE.
• rotated_rad is possibly faster than rotated_deg, since the latter converts from degrees to radians first.
• The typedefs which specify size (for example Vector2i8) use int_fast8_t and similar (fast variant). Feel free to open an issue if this causes problems

How to use

It should be intuitive, but here's how intuitive:

Construct a vector

With typedefs...

Vector2f vec {};                // (0, 0)

Vector2u vec (6);               // (6, 6)

Vector2d vec { 2.0, 5.0 };      // (2.0, 5.0)

Vector2i vec (6, -15);          // (6, -15)

...or the typey way:

Vector2<long double> vec (2.5L, -18.5L);

Vector Operations

rotate (degrees)

auto result_vec = vec.rotated_deg (45.0);

rotate (radians)

auto result_vec = vec.rotated_rad (PI / 3.0);

linearly interpolate with other_vec

auto result_vec = vec.lerp (other_vec, 0.75);

linearly interpolate with other_vec unclamped

^{unclamped means that the value is not restricted between 0.0 and 1.0.}

auto result_vec = vec.lerp_unclamped (other_vec, 1.5);

reflect off surface with normal_vec

auto result_vec = vec.reflected (normal_vec);

distance to other_vec

auto result_vec = vec.distance (other_vec);

squared distance to other_vec

auto result_vec = vec.sqr_distance (other_vec);

magnitude

double mag = vec.magnitude ();

squared magnitude

double sqr_mag = vec.sqr_magnitude ();

dot product with other_vec

auto result_vec = vec.dot (other_vec);

square root

auto result_vec = vec.square_root ();

normalize

result_vec = vec.normalized ();

vector2's Issues

rotate_deg does not work correctly

When rotating a vector, rotate_deg does not calculate correctly.

Example:
Vector2 (2, 5).rotated_deg (90) should rotate vector (2, 5) by 90 degrees, which should return a vector (-5, 2) (correct example result).

It instead returns a vector (2, -5).

Using the latest version, you can reproduce it with this code:

std::cout << Vector2f (2, 5).rotated_deg (90) << std::endl;

Which will output x: 2 y: -5, a wrong result.

This is reproducible with many different vectors and degrees.

Vector2/Vector2.hpp

Lines 364 to 371 in 17cdcf4

	constexpr Vector2 rotated_deg (double angle_degrees) const
	{
	// convert to radians
	long double radians = angle_degrees * (VECTOR2_MATH_PI<long double> / 180.0);
	return Vector2 (
	x * sin (radians) + x * cos (radians),
	x * cos (radians) - y * sin (radians));
	}