/* SPDX-License-Identifier: GPL-2.0-or-later * Copyright 2001-2002 NaN Holding BV. All rights reserved. */ #pragma once /** \file * \ingroup bli */ #if defined(_MSC_VER) && !defined(_USE_MATH_DEFINES) # define _USE_MATH_DEFINES #endif #include "BLI_assert.h" #include "BLI_math_inline.h" #include "BLI_sys_types.h" #include #ifndef M_PI # define M_PI 3.14159265358979323846 /* pi */ #endif #ifndef M_PI_2 # define M_PI_2 1.57079632679489661923 /* pi/2 */ #endif #ifndef M_PI_4 # define M_PI_4 0.78539816339744830962 /* pi/4 */ #endif #ifndef M_SQRT2 # define M_SQRT2 1.41421356237309504880 /* sqrt(2) */ #endif #ifndef M_SQRT1_2 # define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */ #endif #ifndef M_SQRT3 # define M_SQRT3 1.73205080756887729352 /* sqrt(3) */ #endif #ifndef M_SQRT1_3 # define M_SQRT1_3 0.57735026918962576450 /* 1/sqrt(3) */ #endif #ifndef M_1_PI # define M_1_PI 0.318309886183790671538 /* 1/pi */ #endif #ifndef M_E # define M_E 2.7182818284590452354 /* e */ #endif #ifndef M_LOG2E # define M_LOG2E 1.4426950408889634074 /* log_2 e */ #endif #ifndef M_LOG10E # define M_LOG10E 0.43429448190325182765 /* log_10 e */ #endif #ifndef M_LN2 # define M_LN2 0.69314718055994530942 /* log_e 2 */ #endif #ifndef M_LN10 # define M_LN10 2.30258509299404568402 /* log_e 10 */ #endif #if defined(__GNUC__) # define NAN_FLT __builtin_nanf("") #else /* evil quiet NaN definition */ static const int NAN_INT = 0x7FC00000; # define NAN_FLT (*((float *)(&NAN_INT))) #endif #if BLI_MATH_DO_INLINE # include "intern/math_base_inline.c" #endif #ifdef BLI_MATH_GCC_WARN_PRAGMA # pragma GCC diagnostic push # pragma GCC diagnostic ignored "-Wredundant-decls" #endif #ifdef __cplusplus extern "C" { #endif /******************************* Float ******************************/ /* `powf` is really slow for raising to integer powers. */ MINLINE float pow2f(float x); MINLINE float pow3f(float x); MINLINE float pow4f(float x); MINLINE float pow7f(float x); MINLINE float sqrt3f(float f); MINLINE double sqrt3d(double d); MINLINE float sqrtf_signed(float f); MINLINE float saacosf(float f); MINLINE float saasinf(float f); MINLINE float sasqrtf(float f); MINLINE float saacos(float fac); MINLINE float saasin(float fac); MINLINE float sasqrt(float fac); MINLINE float interpf(float a, float b, float t); MINLINE double interpd(double a, double b, double t); MINLINE float ratiof(float min, float max, float pos); MINLINE double ratiod(double min, double max, double pos); /** * Map a normalized value, i.e. from interval [0, 1] to interval [a, b]. */ MINLINE float scalenorm(float a, float b, float x); /** * Map a normalized value, i.e. from interval [0, 1] to interval [a, b]. */ MINLINE double scalenormd(double a, double b, double x); /* NOTE: Compilers will upcast all types smaller than int to int when performing arithmetic * operation. */ MINLINE int square_s(short a); MINLINE int square_uchar(unsigned char a); MINLINE int cube_s(short a); MINLINE int cube_uchar(unsigned char a); MINLINE int square_i(int a); MINLINE unsigned int square_uint(unsigned int a); MINLINE float square_f(float a); MINLINE double square_d(double a); MINLINE int cube_i(int a); MINLINE unsigned int cube_uint(unsigned int a); MINLINE float cube_f(float a); MINLINE double cube_d(double a); MINLINE float min_ff(float a, float b); MINLINE float max_ff(float a, float b); MINLINE float min_fff(float a, float b, float c); MINLINE float max_fff(float a, float b, float c); MINLINE float min_ffff(float a, float b, float c, float d); MINLINE float max_ffff(float a, float b, float c, float d); MINLINE double min_dd(double a, double b); MINLINE double max_dd(double a, double b); MINLINE int min_ii(int a, int b); MINLINE int max_ii(int a, int b); MINLINE int min_iii(int a, int b, int c); MINLINE int max_iii(int a, int b, int c); MINLINE int min_iiii(int a, int b, int c, int d); MINLINE int max_iiii(int a, int b, int c, int d); MINLINE uint min_uu(uint a, uint b); MINLINE uint max_uu(uint a, uint b); MINLINE size_t min_zz(size_t a, size_t b); MINLINE size_t max_zz(size_t a, size_t b); MINLINE char min_cc(char a, char b); MINLINE char max_cc(char a, char b); MINLINE int clamp_i(int value, int min, int max); MINLINE float clamp_f(float value, float min, float max); MINLINE size_t clamp_z(size_t value, size_t min, size_t max); /** * Almost-equal for IEEE floats, using absolute difference method. * * \param max_diff: the maximum absolute difference. */ MINLINE int compare_ff(float a, float b, float max_diff); /** * Almost-equal for IEEE floats, using their integer representation * (mixing ULP and absolute difference methods). * * \param max_diff: is the maximum absolute difference (allows to take care of the near-zero area, * where relative difference methods cannot really work). * \param max_ulps: is the 'maximum number of floats + 1' * allowed between \a a and \a b to consider them equal. * * \see https://randomascii.wordpress.com/2012/02/25/comparing-floating-point-numbers-2012-edition/ */ MINLINE int compare_ff_relative(float a, float b, float max_diff, int max_ulps); MINLINE bool compare_threshold_relative(float value1, float value2, float thresh); MINLINE float signf(float f); MINLINE int signum_i_ex(float a, float eps); MINLINE int signum_i(float a); /** * Used for zoom values. */ MINLINE float power_of_2(float f); /** * Returns number of (base ten) *significant* digits of integer part of given float * (negative in case of decimal-only floats, 0.01 returns -1 e.g.). */ MINLINE int integer_digits_f(float f); /** * Returns number of (base ten) *significant* digits of integer part of given double * (negative in case of decimal-only floats, 0.01 returns -1 e.g.). */ MINLINE int integer_digits_d(double d); MINLINE int integer_digits_i(int i); /* These don't really fit anywhere but were being copied about a lot. */ MINLINE int is_power_of_2_i(int n); MINLINE unsigned int log2_floor_u(unsigned int x); MINLINE unsigned int log2_ceil_u(unsigned int x); /** * Returns next (or previous) power of 2 or the input number if it is already a power of 2. */ MINLINE int power_of_2_max_i(int n); MINLINE int power_of_2_min_i(int n); MINLINE unsigned int power_of_2_max_u(unsigned int x); MINLINE unsigned int power_of_2_min_u(unsigned int x); /** * Integer division that rounds 0.5 up, particularly useful for color blending * with integers, to avoid gradual darkening when rounding down. */ MINLINE int divide_round_i(int a, int b); /** * Integer division that returns the ceiling, instead of flooring like normal C division. */ MINLINE uint divide_ceil_u(uint a, uint b); MINLINE uint64_t divide_ceil_ul(uint64_t a, uint64_t b); /** * Returns \a a if it is a multiple of \a b or the next multiple or \a b after \b a . */ MINLINE uint ceil_to_multiple_u(uint a, uint b); MINLINE uint64_t ceil_to_multiple_ul(uint64_t a, uint64_t b); /** * modulo that handles negative numbers, works the same as Python's. */ MINLINE int mod_i(int i, int n); /** * Round to closest even number, halfway cases are rounded away from zero. */ MINLINE float round_to_even(float f); MINLINE signed char round_fl_to_char(float a); MINLINE unsigned char round_fl_to_uchar(float a); MINLINE short round_fl_to_short(float a); MINLINE unsigned short round_fl_to_ushort(float a); MINLINE int round_fl_to_int(float a); MINLINE unsigned int round_fl_to_uint(float a); MINLINE signed char round_db_to_char(double a); MINLINE unsigned char round_db_to_uchar(double a); MINLINE short round_db_to_short(double a); MINLINE unsigned short round_db_to_ushort(double a); MINLINE int round_db_to_int(double a); MINLINE unsigned int round_db_to_uint(double a); MINLINE signed char round_fl_to_char_clamp(float a); MINLINE unsigned char round_fl_to_uchar_clamp(float a); MINLINE short round_fl_to_short_clamp(float a); MINLINE unsigned short round_fl_to_ushort_clamp(float a); MINLINE int round_fl_to_int_clamp(float a); MINLINE unsigned int round_fl_to_uint_clamp(float a); MINLINE signed char round_db_to_char_clamp(double a); MINLINE unsigned char round_db_to_uchar_clamp(double a); MINLINE short round_db_to_short_clamp(double a); MINLINE unsigned short round_db_to_ushort_clamp(double a); MINLINE int round_db_to_int_clamp(double a); MINLINE unsigned int round_db_to_uint_clamp(double a); int pow_i(int base, int exp); /** * \param ndigits: must be between 0 and 21. */ double double_round(double x, int ndigits); /** * Floor to the nearest power of 10, e.g.: * - 15.0 -> 10.0 * - 0.015 -> 0.01 * - 1.0 -> 1.0 * * \param f: Value to floor, must be over 0.0. * \note If we wanted to support signed values we could if this becomes necessary. */ float floor_power_of_10(float f); /** * Ceiling to the nearest power of 10, e.g.: * - 15.0 -> 100.0 * - 0.015 -> 0.1 * - 1.0 -> 1.0 * * \param f: Value to ceiling, must be over 0.0. * \note If we wanted to support signed values we could if this becomes necessary. */ float ceil_power_of_10(float f); #ifdef BLI_MATH_GCC_WARN_PRAGMA # pragma GCC diagnostic pop #endif /* Asserts, some math functions expect normalized inputs * check the vector is unit length, or zero length (which can't be helped in some cases). */ #ifndef NDEBUG /** \note 0.0001 is too small because normals may be converted from short's: see T34322. */ # define BLI_ASSERT_UNIT_EPSILON 0.0002f # define BLI_ASSERT_UNIT_EPSILON_DB 0.0002 /** * \note Checks are flipped so NAN doesn't assert. * This is done because we're making sure the value was normalized and in the case we * don't want NAN to be raising asserts since there is nothing to be done in that case. */ # define BLI_ASSERT_UNIT_V3(v) \ { \ const float _test_unit = len_squared_v3(v); \ BLI_assert(!(fabsf(_test_unit - 1.0f) >= BLI_ASSERT_UNIT_EPSILON) || \ !(fabsf(_test_unit) >= BLI_ASSERT_UNIT_EPSILON)); \ } \ (void)0 # define BLI_ASSERT_UNIT_V3_DB(v) \ { \ const double _test_unit = len_squared_v3_db(v); \ BLI_assert(!(fabs(_test_unit - 1.0) >= BLI_ASSERT_UNIT_EPSILON_DB) || \ !(fabs(_test_unit) >= BLI_ASSERT_UNIT_EPSILON_DB)); \ } \ (void)0 # define BLI_ASSERT_UNIT_V2(v) \ { \ const float _test_unit = len_squared_v2(v); \ BLI_assert(!(fabsf(_test_unit - 1.0f) >= BLI_ASSERT_UNIT_EPSILON) || \ !(fabsf(_test_unit) >= BLI_ASSERT_UNIT_EPSILON)); \ } \ (void)0 # define BLI_ASSERT_UNIT_QUAT(q) \ { \ const float _test_unit = dot_qtqt(q, q); \ BLI_assert(!(fabsf(_test_unit - 1.0f) >= BLI_ASSERT_UNIT_EPSILON * 10) || \ !(fabsf(_test_unit) >= BLI_ASSERT_UNIT_EPSILON * 10)); \ } \ (void)0 # define BLI_ASSERT_ZERO_M3(m) \ { \ BLI_assert(dot_vn_vn((const float *)m, (const float *)m, 9) != 0.0); \ } \ (void)0 # define BLI_ASSERT_ZERO_M4(m) \ { \ BLI_assert(dot_vn_vn((const float *)m, (const float *)m, 16) != 0.0); \ } \ (void)0 # define BLI_ASSERT_UNIT_M3(m) \ { \ BLI_ASSERT_UNIT_V3((m)[0]); \ BLI_ASSERT_UNIT_V3((m)[1]); \ BLI_ASSERT_UNIT_V3((m)[2]); \ } \ (void)0 #else # define BLI_ASSERT_UNIT_V2(v) (void)(v) # define BLI_ASSERT_UNIT_V3(v) (void)(v) # define BLI_ASSERT_UNIT_V3_DB(v) (void)(v) # define BLI_ASSERT_UNIT_QUAT(v) (void)(v) # define BLI_ASSERT_ZERO_M3(m) (void)(m) # define BLI_ASSERT_ZERO_M4(m) (void)(m) # define BLI_ASSERT_UNIT_M3(m) (void)(m) #endif #ifdef __cplusplus } #endif