#ifndef COSMOPOLITAN_LIBC_MATH_H_ #define COSMOPOLITAN_LIBC_MATH_H_ /*─────────────────────────────────────────────────────────────────────────────╗ │ cosmopolitan § mathematics │ ╚─────────────────────────────────────────────────────────────────────────────*/ #define M_E 2.7182818284590452354 /* 𝑒 */ #define M_LOG2E 1.4426950408889634074 /* log₂𝑒 */ #define M_LOG10E 0.43429448190325182765 /* log₁₀𝑒 */ #define M_LN2 0.69314718055994530942 /* logₑ2 */ #define M_LN10 2.30258509299404568402 /* logₑ10 */ #define M_PI 3.14159265358979323846 /* 𝜋 */ #define M_PI_2 1.57079632679489661923 /* 𝜋/2 */ #define M_PI_4 0.78539816339744830962 /* 𝜋/4 */ #define M_1_PI 0.31830988618379067154 /* 1/𝜋 */ #define M_2_PI 0.63661977236758134308 /* 2/𝜋 */ #define M_2_SQRTPI 1.12837916709551257390 /* 2/sqrt(𝜋) */ #define M_SQRT2 1.41421356237309504880 /* sqrt(2) */ #define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */ #if defined(_GNU_SOURCE) || defined(_COSMO_SOURCE) #define M_Ef 2.7182818284590452354f /* 𝑒 */ #define M_LOG2Ef 1.4426950408889634074f /* log₂𝑒 */ #define M_LOG10Ef 0.43429448190325182765f /* log₁₀𝑒 */ #define M_LN2f 0.69314718055994530942f /* logₑ2 */ #define M_LN10f 2.30258509299404568402f /* logₑ10 */ #define M_PIf 3.14159265358979323846f /* 𝜋 */ #define M_PI_2f 1.57079632679489661923f /* 𝜋/2 */ #define M_PI_4f 0.78539816339744830962f /* 𝜋/4 */ #define M_1_PIf 0.31830988618379067154f /* 1/𝜋 */ #define M_2_PIf 0.63661977236758134308f /* 2/𝜋 */ #define M_2_SQRTPIf 1.12837916709551257390f /* 2/sqrt(𝜋) */ #define M_SQRT2f 1.41421356237309504880f /* sqrt(2) */ #define M_SQRT1_2f 0.70710678118654752440f /* 1/sqrt(2) */ #endif #if defined(_GNU_SOURCE) || defined(_COSMO_SOURCE) #define M_El 2.718281828459045235360287471352662498L /* 𝑒 */ #define M_LOG2El 1.442695040888963407359924681001892137L /* log₂𝑒 */ #define M_LOG10El 0.434294481903251827651128918916605082L /* log₁₀𝑒 */ #define M_LN2l 0.693147180559945309417232121458176568L /* logₑ2 */ #define M_LN10l 2.302585092994045684017991454684364208L /* logₑ10 */ #define M_PIl 3.141592653589793238462643383279502884L /* 𝜋 */ #define M_PI_2l 1.570796326794896619231321691639751442L /* 𝜋/2 */ #define M_PI_4l 0.785398163397448309615660845819875721L /* 𝜋/4 */ #define M_1_PIl 0.318309886183790671537767526745028724L /* 1/𝜋 */ #define M_2_PIl 0.636619772367581343075535053490057448L /* 2/𝜋 */ #define M_2_SQRTPIl 1.128379167095512573896158903121545172L /* 2/sqrt(𝜋) */ #define M_SQRT2l 1.414213562373095048801688724209698079L /* sqrt(2) */ #define M_SQRT1_2l 0.707106781186547524400844362104849039L /* 1/sqrt(2) */ #endif #define DBL_DECIMAL_DIG __DBL_DECIMAL_DIG__ #define DBL_DIG __DBL_DIG__ #define DBL_EPSILON __DBL_EPSILON__ #define DBL_HAS_SUBNORM __DBL_HAS_DENORM__ #define DBL_IS_IEC_60559 __DBL_IS_IEC_60559__ #define DBL_MANT_DIG __DBL_MANT_DIG__ #define DBL_MANT_DIG __DBL_MANT_DIG__ #define DBL_MAX __DBL_MAX__ #define DBL_MAX_10_EXP __DBL_MAX_10_EXP__ #define DBL_MAX_EXP __DBL_MAX_EXP__ #define DBL_MIN __DBL_MIN__ /* 2.23e–308 ↔ 1.79e308 */ #define DBL_MIN_10_EXP __DBL_MIN_10_EXP__ #define DBL_MIN_EXP __DBL_MIN_EXP__ #define DBL_NORM_MAX __DBL_NORM_MAX__ #define DBL_TRUE_MIN __DBL_DENORM_MIN__ #define DECIMAL_DIG __LDBL_DECIMAL_DIG__ #define FLT_DECIMAL_DIG __FLT_DECIMAL_DIG__ #define FLT_DIG __FLT_DIG__ #define FLT_EPSILON __FLT_EPSILON__ #define FLT_HAS_SUBNORM __FLT_HAS_DENORM__ #define FLT_IS_IEC_60559 __FLT_IS_IEC_60559__ #define FLT_MANT_DIG __FLT_MANT_DIG__ #define FLT_MANT_DIG __FLT_MANT_DIG__ #define FLT_MAX __FLT_MAX__ #define FLT_MAX_10_EXP __FLT_MAX_10_EXP__ #define FLT_MAX_EXP __FLT_MAX_EXP__ #define FLT_MIN __FLT_MIN__ /* 1.18e–38 ↔ 3.40e38 */ #define FLT_MIN_10_EXP __FLT_MIN_10_EXP__ #define FLT_MIN_EXP __FLT_MIN_EXP__ #define FLT_NORM_MAX __FLT_NORM_MAX__ #define FLT_RADIX __FLT_RADIX__ #define FLT_TRUE_MIN __FLT_DENORM_MIN__ #define HLF_MAX 6.50e4f #define HLF_MIN 3.10e-5f #define LDBL_DECIMAL_DIG __LDBL_DECIMAL_DIG__ #define LDBL_DIG __LDBL_DIG__ #define LDBL_EPSILON __LDBL_EPSILON__ #define LDBL_HAS_SUBNORM __LDBL_HAS_DENORM__ #define LDBL_IS_IEC_60559 __LDBL_IS_IEC_60559__ #define LDBL_MANT_DIG __LDBL_MANT_DIG__ #define LDBL_MANT_DIG __LDBL_MANT_DIG__ #define LDBL_MAX __LDBL_MAX__ #define LDBL_MAX_10_EXP __LDBL_MAX_10_EXP__ #define LDBL_MAX_EXP __LDBL_MAX_EXP__ #define LDBL_MIN __LDBL_MIN__ /* 3.37e–4932 ↔ 1.18e4932 */ #define LDBL_MIN_10_EXP __LDBL_MIN_10_EXP__ #define LDBL_MIN_EXP __LDBL_MIN_EXP__ #define LDBL_NORM_MAX __LDBL_NORM_MAX__ #define LDBL_TRUE_MIN __LDBL_DENORM_MIN__ #define FP_NAN 0 #define FP_INFINITE 1 #define FP_ZERO 2 #define FP_SUBNORMAL 3 #define FP_NORMAL 4 #define FP_ILOGB0 (-2147483647 - 1) #define FP_ILOGBNAN (-2147483647 - 1) #define MATH_ERRNO 1 #define MATH_ERREXCEPT 2 #ifdef __FAST_MATH__ #define math_errhandling 0 #elif defined(__NO_MATH_ERRNO__) #define math_errhandling (MATH_ERREXCEPT) #else #define math_errhandling (MATH_ERRNO | MATH_ERREXCEPT) #endif #ifdef __FP_FAST_FMA #define FP_FAST_FMA 1 #endif #ifdef __FP_FAST_FMAF #define FP_FAST_FMAF 1 #endif #ifdef __FP_FAST_FMAL #define FP_FAST_FMAL 1 #endif COSMOPOLITAN_C_START_ #define NAN __builtin_nanf("") #define INFINITY __builtin_inff() #define HUGE_VAL __builtin_inf() #define HUGE_VALF __builtin_inff() #define HUGE_VALL __builtin_infl() #if __FLT_EVAL_METHOD__ + 0 == 2 typedef long double float_t; typedef long double double_t; #else typedef float float_t; typedef double double_t; #endif #define isinf(x) __builtin_isinf(x) #define isnan(x) __builtin_isnan(x) #define isfinite(x) __builtin_isfinite(x) #define isnormal(x) __builtin_isnormal(x) #define isgreater(x, y) __builtin_isgreater(x, y) #define isgreaterequal(x, y) __builtin_isgreaterequal(x, y) #define isless(x, y) __builtin_isless(x, y) #define islessequal(x, y) __builtin_islessequal(x, y) #define islessgreater(x, y) __builtin_islessgreater(x, y) #define isunordered(x, y) __builtin_isunordered(x, y) #define fpclassify(x) \ __builtin_fpclassify(FP_NAN, FP_INFINITE, FP_NORMAL, FP_SUBNORMAL, FP_ZERO, x) #define signbit(x) \ (sizeof(x) == sizeof(long double) ? __builtin_signbitl(x) \ : sizeof(x) == sizeof(float) ? __builtin_signbitf(x) \ : __builtin_signbit(x)) extern int signgam; double acos(double) libcesque; double acosh(double) libcesque; double asin(double) libcesque; double asinh(double) libcesque; double atan(double) libcesque; double atan2(double, double) libcesque; double atanh(double) libcesque; double cbrt(double) libcesque; double ceil(double) libcesque; double copysign(double, double) libcesque; double cos(double) libcesque; double cosh(double) libcesque; double drem(double, double) libcesque; double erf(double) libcesque; double erfc(double) libcesque; double exp(double) libcesque; double exp10(double) libcesque; double exp2(double) libcesque; double expm1(double) libcesque; double fabs(double) libcesque; double fdim(double, double) libcesque; double floor(double) libcesque; double fma(double, double, double) libcesque; double fmax(double, double) libcesque; double fmin(double, double) libcesque; double fmod(double, double) libcesque; double hypot(double, double) libcesque; double ldexp(double, int) libcesque; double log(double) libcesque; double log10(double) libcesque; double log1p(double) libcesque; double log2(double) libcesque; double logb(double) libcesque; double nearbyint(double) libcesque; double nextafter(double, double) libcesque; double nexttoward(double, long double) libcesque; double pow(double, double) libcesque; double pow10(double) libcesque; double powi(double, int) libcesque; double remainder(double, double) libcesque; double rint(double) libcesque; double round(double) libcesque; double scalb(double, double) libcesque; double scalbln(double, long int) libcesque; double scalbn(double, int) libcesque; double significand(double) libcesque; double sin(double) libcesque; double sinh(double) libcesque; double sqrt(double) libcesque; double tan(double) libcesque; double tanh(double) libcesque; double trunc(double) libcesque; double tgamma(double) libcesque; double lgamma(double) libcesque; double lgamma_r(double, int *) libcesque; int finite(double) libcesque; float acosf(float) libcesque; float acoshf(float) libcesque; float asinf(float) libcesque; float asinhf(float) libcesque; float atan2f(float, float) libcesque; float atanf(float) libcesque; float atanhf(float) libcesque; float cbrtf(float) libcesque; float ceilf(float) libcesque; float copysignf(float, float) libcesque; float cosf(float) libcesque; float coshf(float) libcesque; float dremf(float, float) libcesque; float erfcf(float) libcesque; float erff(float) libcesque; float exp10f(float) libcesque; float exp2f(float) libcesque; float expf(float) libcesque; float expm1f(float) libcesque; float fabsf(float) libcesque; float fdimf(float, float) libcesque; float floorf(float) libcesque; float fmaf(float, float, float) libcesque; float fmaxf(float, float) libcesque; float fminf(float, float) libcesque; float fmodf(float, float) libcesque; float hypotf(float, float) libcesque; float ldexpf(float, int) libcesque; float lgammaf(float) libcesque; float lgammaf_r(float, int *) libcesque; float log10f(float) libcesque; float log1pf(float) libcesque; float log2f(float) libcesque; float logbf(float) libcesque; float logf(float) libcesque; float nearbyintf(float) libcesque; float nextafterf(float, float) libcesque; float nexttowardf(float, long double) libcesque; float pow10f(float) libcesque; float powf(float, float) libcesque; float powif(float, int) libcesque; float remainderf(float, float) libcesque; float rintf(float) libcesque; float roundf(float) libcesque; float scalbf(float, float) libcesque; float scalblnf(float, long int) libcesque; float scalbnf(float, int) libcesque; float significandf(float) libcesque; float sinf(float) libcesque; float sinhf(float) libcesque; float sqrtf(float) libcesque; float tanf(float) libcesque; float tanhf(float) libcesque; float tgammaf(float) libcesque; float truncf(float) libcesque; int finitef(float) libcesque; int finitel(long double) libcesque; long double acoshl(long double) libcesque; long double acosl(long double) libcesque; long double asinhl(long double) libcesque; long double asinl(long double) libcesque; long double atan2l(long double, long double) libcesque; long double atanhl(long double) libcesque; long double atanl(long double) libcesque; long double cbrtl(long double) libcesque; long double ceill(long double) libcesque; long double copysignl(long double, long double) libcesque; long double coshl(long double) libcesque; long double cosl(long double) libcesque; long double dreml(long double, long double) libcesque; long double erfcl(long double) libcesque; long double erfl(long double) libcesque; long double exp10l(long double) libcesque; long double exp2l(long double) libcesque; long double expl(long double) libcesque; long double expm1l(long double) libcesque; long double fabsl(long double) libcesque; long double fdiml(long double, long double) libcesque; long double floorl(long double) libcesque; long double fmal(long double, long double, long double) libcesque; long double fmaxl(long double, long double) libcesque; long double fminl(long double, long double) libcesque; long double fmodl(long double, long double) libcesque; long double hypotl(long double, long double) libcesque; long double ldexpl(long double, int) libcesque; long double lgammal(long double) libcesque; long double lgammal_r(long double, int *) libcesque; long double log10l(long double) libcesque; long double log1pl(long double) libcesque; long double log2l(long double) libcesque; long double logbl(long double) libcesque; long double logl(long double) libcesque; long double nearbyintl(long double) libcesque; long double nextafterl(long double, long double) libcesque; long double nexttowardl(long double, long double) libcesque; long double pow10l(long double) libcesque; long double powl(long double, long double) libcesque; long double remainderl(long double, long double) libcesque; long double rintl(long double) libcesque; long double roundl(long double) libcesque; long double scalbl(long double, long double) libcesque; long double scalblnl(long double, long int) libcesque; long double scalbnl(long double, int) libcesque; long double significandl(long double) libcesque; long double sinhl(long double) libcesque; long double sinl(long double) libcesque; long double sqrtl(long double) libcesque; long double tanhl(long double) libcesque; long double tanl(long double) libcesque; long double tgammal(long double) libcesque; long double truncl(long double) libcesque; long lrint(double) libcesque; long lrintf(float) libcesque; long lrintl(long double) libcesque; long lround(double) libcesque; long lroundf(float) libcesque; long lroundl(long double) libcesque; int ilogbf(float) libcesque; int ilogb(double) libcesque; int ilogbl(long double) libcesque; long long llrint(double) libcesque; long long llrintf(float) libcesque; long long llrintl(long double) libcesque; long long llround(double) libcesque; long long llroundf(float) libcesque; long long llroundl(long double) libcesque; double frexp(double, int *) libcesque; double modf(double, double *) libcesque; double nan(const char *) libcesque; double remquo(double, double, int *) libcesque; float frexpf(float, int *) libcesque; float modff(float, float *) libcesque; float nanf(const char *) libcesque; float remquof(float, float, int *) libcesque; long double frexpl(long double, int *) libcesque; long double modfl(long double, long double *) libcesque; long double nanl(const char *) libcesque; long double remquol(long double, long double, int *) libcesque; void sincos(double, double *, double *) libcesque; void sincosf(float, float *, float *) libcesque; void sincosl(long double, long double *, long double *) libcesque; double fsumf(const float *, size_t) libcesque; double fsum(const double *, size_t) libcesque; double j0(double) libcesque; double j1(double) libcesque; double jn(int, double) libcesque; float j0f(float) libcesque; float j1f(float) libcesque; float jnf(int, float) libcesque; double y0(double) libcesque; double y1(double) libcesque; double yn(int, double) libcesque; float y0f(float) libcesque; float y1f(float) libcesque; float ynf(int, float) libcesque; COSMOPOLITAN_C_END_ #endif /* COSMOPOLITAN_LIBC_MATH_H_ */