diff options
author | Andrew Hsieh <andrewhsieh@google.com> | 2013-11-11 16:01:46 +0800 |
---|---|---|
committer | Andrew Hsieh <andrewhsieh@google.com> | 2013-11-12 15:05:12 +0800 |
commit | 7b0de5ae35083554f258e68dcfb0c9ada702e70a (patch) | |
tree | 6db6ed333d1275a724874c1f9d9d28d200397706 /9/platforms/android-15/arch-x86/usr/include/math.h | |
parent | e9808d310fa92e2aa27a9208fe59144a43a8d4b6 (diff) | |
download | prebuilts_ndk-7b0de5ae35083554f258e68dcfb0c9ada702e70a.zip prebuilts_ndk-7b0de5ae35083554f258e68dcfb0c9ada702e70a.tar.gz prebuilts_ndk-7b0de5ae35083554f258e68dcfb0c9ada702e70a.tar.bz2 |
Refresh prebuilts/ndk/9/platforms with r9b+, part 2/2
Part 2: add new API level 12, 13, 15, 16, 17 and 19
Change-Id: I305970f7766b3ed3e5faf2532952eaa03398fcbd
Diffstat (limited to '9/platforms/android-15/arch-x86/usr/include/math.h')
-rw-r--r-- | 9/platforms/android-15/arch-x86/usr/include/math.h | 498 |
1 files changed, 498 insertions, 0 deletions
diff --git a/9/platforms/android-15/arch-x86/usr/include/math.h b/9/platforms/android-15/arch-x86/usr/include/math.h new file mode 100644 index 0000000..e46cf46 --- /dev/null +++ b/9/platforms/android-15/arch-x86/usr/include/math.h @@ -0,0 +1,498 @@ +/* + * ==================================================== + * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved. + * + * Developed at SunPro, a Sun Microsystems, Inc. business. + * Permission to use, copy, modify, and distribute this + * software is freely granted, provided that this notice + * is preserved. + * ==================================================== + */ + +/* + * from: @(#)fdlibm.h 5.1 93/09/24 + * $FreeBSD: src/lib/msun/src/math.h,v 1.61 2005/04/16 21:12:47 das Exp $ + */ + +#ifndef _MATH_H_ +#define _MATH_H_ + +#include <sys/cdefs.h> +#include <sys/types.h> +#include <limits.h> + +#define __pure2 + +/* + * ANSI/POSIX + */ +extern const union __infinity_un { + unsigned char __uc[8]; + double __ud; +} __infinity; + +extern const union __nan_un { + unsigned char __uc[sizeof(float)]; + float __uf; +} __nan; + +/* #if __GNUC_PREREQ__(3, 3) || (defined(__INTEL_COMPILER) && __INTEL_COMPILER >= 800) */ +#if 1 +#define __MATH_BUILTIN_CONSTANTS +#endif + +/* #if __GNUC_PREREQ__(3, 0) && !defined(__INTEL_COMPILER) */ +#if 1 +#define __MATH_BUILTIN_RELOPS +#endif + +/* #ifdef __MATH_BUILTIN_CONSTANTS */ +#if 1 +#define HUGE_VAL __builtin_huge_val() +#else +#define HUGE_VAL (__infinity.__ud) +#endif + +/* #if __ISO_C_VISIBLE >= 1999 */ +#if 0 +#define FP_ILOGB0 (-__INT_MAX) +#define FP_ILOGBNAN __INT_MAX +#else +#define FP_ILOGB0 (-INT_MAX) +#define FP_ILOGBNAN INT_MAX +#endif + +#ifdef __MATH_BUILTIN_CONSTANTS +#define HUGE_VALF __builtin_huge_valf() +#define HUGE_VALL __builtin_huge_vall() +#define INFINITY __builtin_inf() +#define NAN __builtin_nan("") +#else +#define HUGE_VALF (float)HUGE_VAL +#define HUGE_VALL (long double)HUGE_VAL +#define INFINITY HUGE_VALF +#define NAN (__nan.__uf) +#endif /* __MATH_BUILTIN_CONSTANTS */ + +#define MATH_ERRNO 1 +#define MATH_ERREXCEPT 2 +#define math_errhandling MATH_ERREXCEPT + +/* XXX We need a <machine/math.h>. */ +#if defined(__ia64__) || defined(__sparc64__) +#define FP_FAST_FMA +#endif +#ifdef __ia64__ +#define FP_FAST_FMAL +#endif +#define FP_FAST_FMAF + +/* Symbolic constants to classify floating point numbers. */ +#define FP_INFINITE 0x01 +#define FP_NAN 0x02 +#define FP_NORMAL 0x04 +#define FP_SUBNORMAL 0x08 +#define FP_ZERO 0x10 +#define fpclassify(x) \ + ((sizeof (x) == sizeof (float)) ? __fpclassifyf(x) \ + : (sizeof (x) == sizeof (double)) ? __fpclassifyd(x) \ + : __fpclassifyl(x)) + +#define isfinite(x) \ + ((sizeof (x) == sizeof (float)) ? __isfinitef(x) \ + : (sizeof (x) == sizeof (double)) ? __isfinite(x) \ + : __isfinitel(x)) +#define isinf(x) \ + ((sizeof (x) == sizeof (float)) ? __isinff(x) \ + : (sizeof (x) == sizeof (double)) ? __isinf(x) \ + : __isinfl(x)) +#define isnan(x) \ + ((sizeof (x) == sizeof (float)) ? isnanf(x) \ + : (sizeof (x) == sizeof (double)) ? isnan(x) \ + : __isnanl(x)) +#define isnormal(x) \ + ((sizeof (x) == sizeof (float)) ? __isnormalf(x) \ + : (sizeof (x) == sizeof (double)) ? __isnormal(x) \ + : __isnormall(x)) + +#ifdef __MATH_BUILTIN_RELOPS +#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)) +#else +#define isgreater(x, y) (!isunordered((x), (y)) && (x) > (y)) +#define isgreaterequal(x, y) (!isunordered((x), (y)) && (x) >= (y)) +#define isless(x, y) (!isunordered((x), (y)) && (x) < (y)) +#define islessequal(x, y) (!isunordered((x), (y)) && (x) <= (y)) +#define islessgreater(x, y) (!isunordered((x), (y)) && \ + ((x) > (y) || (y) > (x))) +#define isunordered(x, y) (isnan(x) || isnan(y)) +#endif /* __MATH_BUILTIN_RELOPS */ + +#define signbit(x) \ + ((sizeof (x) == sizeof (float)) ? __signbitf(x) \ + : (sizeof (x) == sizeof (double)) ? __signbit(x) \ + : __signbitl(x)) + +#if 0 +typedef __double_t double_t; +typedef __float_t float_t; +#endif +/* #endif */ /* __ISO_C_VISIBLE >= 1999 */ + +/* + * XOPEN/SVID + */ +/* #if __BSD_VISIBLE || __XSI_VISIBLE */ +#define M_E 2.7182818284590452354 /* e */ +#define M_LOG2E 1.4426950408889634074 /* log 2e */ +#define M_LOG10E 0.43429448190325182765 /* log 10e */ +#define M_LN2 0.69314718055994530942 /* log e2 */ +#define M_LN10 2.30258509299404568402 /* log e10 */ +#define M_PI 3.14159265358979323846 /* pi */ +#define M_PI_2 1.57079632679489661923 /* pi/2 */ +#define M_PI_4 0.78539816339744830962 /* pi/4 */ +#define M_1_PI 0.31830988618379067154 /* 1/pi */ +#define M_2_PI 0.63661977236758134308 /* 2/pi */ +#define M_2_SQRTPI 1.12837916709551257390 /* 2/sqrt(pi) */ +#define M_SQRT2 1.41421356237309504880 /* sqrt(2) */ +#define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */ + +#define MAXFLOAT ((float)3.40282346638528860e+38) +extern int signgam; +/* #endif */ /* __BSD_VISIBLE || __XSI_VISIBLE */ + +#if __BSD_VISIBLE +#if 0 +/* Old value from 4.4BSD-Lite math.h; this is probably better. */ +#define HUGE HUGE_VAL +#else +#define HUGE MAXFLOAT +#endif +#endif /* __BSD_VISIBLE */ + +/* + * Most of these functions depend on the rounding mode and have the side + * effect of raising floating-point exceptions, so they are not declared + * as __pure2. In C99, FENV_ACCESS affects the purity of these functions. + */ +__BEGIN_DECLS +/* + * ANSI/POSIX + */ +int __fpclassifyd(double) __NDK_FPABI_MATH__ __pure2; +int __fpclassifyf(float) __NDK_FPABI_MATH__ __pure2; +int __fpclassifyl(long double) __NDK_FPABI_MATH__ __pure2; +int __isfinitef(float) __NDK_FPABI_MATH__ __pure2; +int __isfinite(double) __NDK_FPABI_MATH__ __pure2; +int __isfinitel(long double) __NDK_FPABI_MATH__ __pure2; +int __isinff(float) __NDK_FPABI_MATH__ __pure2; +int __isinf(double) __NDK_FPABI_MATH__ __pure2; +int __isinfl(long double) __NDK_FPABI_MATH__ __pure2; +int __isnanl(long double) __NDK_FPABI_MATH__ __pure2; +int __isnormalf(float) __NDK_FPABI_MATH__ __pure2; +int __isnormal(double) __NDK_FPABI_MATH__ __pure2; +int __isnormall(long double) __NDK_FPABI_MATH__ __pure2; +int __signbit(double) __NDK_FPABI_MATH__ __pure2; +int __signbitf(float) __NDK_FPABI_MATH__ __pure2; +int __signbitl(long double) __NDK_FPABI_MATH__ __pure2; + +double acos(double) __NDK_FPABI_MATH__; +double asin(double) __NDK_FPABI_MATH__; +double atan(double) __NDK_FPABI_MATH__; +double atan2(double, double) __NDK_FPABI_MATH__; +double cos(double) __NDK_FPABI_MATH__; +double sin(double) __NDK_FPABI_MATH__; +double tan(double) __NDK_FPABI_MATH__; + +double cosh(double) __NDK_FPABI_MATH__; +double sinh(double) __NDK_FPABI_MATH__; +double tanh(double) __NDK_FPABI_MATH__; + +double exp(double) __NDK_FPABI_MATH__; +double frexp(double, int *) __NDK_FPABI_MATH__; /* fundamentally !__pure2 */ +double ldexp(double, int) __NDK_FPABI_MATH__; +double log(double) __NDK_FPABI_MATH__; +double log10(double) __NDK_FPABI_MATH__; +double modf(double, double *) __NDK_FPABI_MATH__; /* fundamentally !__pure2 */ + +double pow(double, double) __NDK_FPABI_MATH__; +double sqrt(double) __NDK_FPABI_MATH__; + +double ceil(double) __NDK_FPABI_MATH__; +double fabs(double) __NDK_FPABI_MATH__ __pure2; +double floor(double) __NDK_FPABI_MATH__; +double fmod(double, double) __NDK_FPABI_MATH__; + +/* + * These functions are not in C90. + */ +/* #if __BSD_VISIBLE || __ISO_C_VISIBLE >= 1999 || __XSI_VISIBLE */ +double acosh(double) __NDK_FPABI_MATH__; +double asinh(double) __NDK_FPABI_MATH__; +double atanh(double) __NDK_FPABI_MATH__; +double cbrt(double) __NDK_FPABI_MATH__; +double erf(double) __NDK_FPABI_MATH__; +double erfc(double) __NDK_FPABI_MATH__; +double exp2(double) __NDK_FPABI_MATH__; +double expm1(double) __NDK_FPABI_MATH__; +double fma(double, double, double) __NDK_FPABI_MATH__; +double hypot(double, double) __NDK_FPABI_MATH__; +int ilogb(double) __NDK_FPABI_MATH__ __pure2; +/* int (isinf)(double) __NDK_FPABI_MATH__ __pure2; */ +int (isnan)(double) __NDK_FPABI_MATH__ __pure2; +double lgamma(double) __NDK_FPABI_MATH__; +long long llrint(double) __NDK_FPABI_MATH__; +long long llround(double) __NDK_FPABI_MATH__; +double log1p(double) __NDK_FPABI_MATH__; +double logb(double) __NDK_FPABI_MATH__; +long lrint(double) __NDK_FPABI_MATH__; +long lround(double) __NDK_FPABI_MATH__; +double nan(const char *) __NDK_FPABI_MATH__ __pure2; +double nextafter(double, double) __NDK_FPABI_MATH__; +double remainder(double, double) __NDK_FPABI_MATH__; +double remquo(double, double, int *) __NDK_FPABI_MATH__; +double rint(double) __NDK_FPABI_MATH__; +/* #endif */ /* __BSD_VISIBLE || __ISO_C_VISIBLE >= 1999 || __XSI_VISIBLE */ + +/* #if __BSD_VISIBLE || __XSI_VISIBLE */ +double j0(double) __NDK_FPABI_MATH__; +double j1(double) __NDK_FPABI_MATH__; +double jn(int, double) __NDK_FPABI_MATH__; +double scalb(double, double) __NDK_FPABI_MATH__; +double y0(double) __NDK_FPABI_MATH__; +double y1(double) __NDK_FPABI_MATH__; +double yn(int, double) __NDK_FPABI_MATH__; + +/* #if __XSI_VISIBLE <= 500 || __BSD_VISIBLE */ +double gamma(double) __NDK_FPABI_MATH__; +/* #endif */ +/* #endif */ /* __BSD_VISIBLE || __XSI_VISIBLE */ + +/* #if __BSD_VISIBLE || __ISO_C_VISIBLE >= 1999 */ +double copysign(double, double) __NDK_FPABI_MATH__ __pure2; +double fdim(double, double) __NDK_FPABI_MATH__; +double fmax(double, double) __NDK_FPABI_MATH__ __pure2; +double fmin(double, double) __NDK_FPABI_MATH__ __pure2; +double nearbyint(double) __NDK_FPABI_MATH__; +double round(double) __NDK_FPABI_MATH__; +double scalbln(double, long) __NDK_FPABI_MATH__; +double scalbn(double, int) __NDK_FPABI_MATH__; +double tgamma(double) __NDK_FPABI_MATH__; +double trunc(double) __NDK_FPABI_MATH__; +/* #endif */ + +/* + * BSD math library entry points + */ +/* #if __BSD_VISIBLE */ +double drem(double, double) __NDK_FPABI_MATH__; +int finite(double) __NDK_FPABI_MATH__ __pure2; +int isnanf(float) __NDK_FPABI_MATH__ __pure2; + +/* + * Reentrant version of gamma & lgamma; passes signgam back by reference + * as the second argument; user must allocate space for signgam. + */ +double gamma_r(double, int *) __NDK_FPABI_MATH__; +double lgamma_r(double, int *) __NDK_FPABI_MATH__; + +/* + * IEEE Test Vector + */ +double significand(double) __NDK_FPABI_MATH__; +/* #endif */ /* __BSD_VISIBLE */ + +/* float versions of ANSI/POSIX functions */ +/*#if __ISO_C_VISIBLE >= 1999 */ +float acosf(float) __NDK_FPABI_MATH__; +float asinf(float) __NDK_FPABI_MATH__; +float atanf(float) __NDK_FPABI_MATH__; +float atan2f(float, float) __NDK_FPABI_MATH__; +float cosf(float) __NDK_FPABI_MATH__; +float sinf(float) __NDK_FPABI_MATH__; +float tanf(float) __NDK_FPABI_MATH__; + +float coshf(float) __NDK_FPABI_MATH__; +float sinhf(float) __NDK_FPABI_MATH__; +float tanhf(float) __NDK_FPABI_MATH__; + +float exp2f(float) __NDK_FPABI_MATH__; +float expf(float) __NDK_FPABI_MATH__; +float expm1f(float) __NDK_FPABI_MATH__; +float frexpf(float, int *) __NDK_FPABI_MATH__; /* fundamentally !__pure2 */ +int ilogbf(float) __NDK_FPABI_MATH__ __pure2; +float ldexpf(float, int) __NDK_FPABI_MATH__; +float log10f(float) __NDK_FPABI_MATH__; +float log1pf(float) __NDK_FPABI_MATH__; +float logf(float) __NDK_FPABI_MATH__; +float modff(float, float *) __NDK_FPABI_MATH__; /* fundamentally !__pure2 */ + +float powf(float, float) __NDK_FPABI_MATH__; +float sqrtf(float) __NDK_FPABI_MATH__; + +float ceilf(float) __NDK_FPABI_MATH__; +float fabsf(float) __NDK_FPABI_MATH__ __pure2; +float floorf(float) __NDK_FPABI_MATH__; +float fmodf(float, float) __NDK_FPABI_MATH__; +float roundf(float) __NDK_FPABI_MATH__; + +float erff(float) __NDK_FPABI_MATH__; +float erfcf(float) __NDK_FPABI_MATH__; +float hypotf(float, float) __NDK_FPABI_MATH__; +float lgammaf(float) __NDK_FPABI_MATH__; +float tgammaf(float) __NDK_FPABI_MATH__; + +float acoshf(float) __NDK_FPABI_MATH__; +float asinhf(float) __NDK_FPABI_MATH__; +float atanhf(float) __NDK_FPABI_MATH__; +float cbrtf(float) __NDK_FPABI_MATH__; +float logbf(float) __NDK_FPABI_MATH__; +float copysignf(float, float) __NDK_FPABI_MATH__ __pure2; +long long llrintf(float) __NDK_FPABI_MATH__; +long long llroundf(float) __NDK_FPABI_MATH__; +long lrintf(float) __NDK_FPABI_MATH__; +long lroundf(float) __NDK_FPABI_MATH__; +float nanf(const char *) __NDK_FPABI_MATH__ __pure2; +float nearbyintf(float) __NDK_FPABI_MATH__; +float nextafterf(float, float) __NDK_FPABI_MATH__; +float remainderf(float, float) __NDK_FPABI_MATH__; +float remquof(float, float, int *) __NDK_FPABI_MATH__; +float rintf(float) __NDK_FPABI_MATH__; +float scalblnf(float, long) __NDK_FPABI_MATH__; +float scalbnf(float, int) __NDK_FPABI_MATH__; +float truncf(float) __NDK_FPABI_MATH__; + +float fdimf(float, float) __NDK_FPABI_MATH__; +float fmaf(float, float, float) __NDK_FPABI_MATH__; +float fmaxf(float, float) __NDK_FPABI_MATH__ __pure2; +float fminf(float, float) __NDK_FPABI_MATH__ __pure2; +/* #endif */ + +/* + * float versions of BSD math library entry points + */ +/* #if __BSD_VISIBLE */ +float dremf(float, float) __NDK_FPABI_MATH__; +int finitef(float) __NDK_FPABI_MATH__ __pure2; +float gammaf(float) __NDK_FPABI_MATH__; +float j0f(float) __NDK_FPABI_MATH__; +float j1f(float) __NDK_FPABI_MATH__; +float jnf(int, float) __NDK_FPABI_MATH__; +float scalbf(float, float) __NDK_FPABI_MATH__; +float y0f(float) __NDK_FPABI_MATH__; +float y1f(float) __NDK_FPABI_MATH__; +float ynf(int, float) __NDK_FPABI_MATH__; + +/* + * Float versions of reentrant version of gamma & lgamma; passes + * signgam back by reference as the second argument; user must + * allocate space for signgam. + */ +float gammaf_r(float, int *) __NDK_FPABI_MATH__; +float lgammaf_r(float, int *) __NDK_FPABI_MATH__; + +/* + * float version of IEEE Test Vector + */ +float significandf(float) __NDK_FPABI_MATH__; +/* #endif */ /* __BSD_VISIBLE */ + +/* + * long double versions of ISO/POSIX math functions + */ +/* #if __ISO_C_VISIBLE >= 1999 */ +#if 0 +long double acoshl(long double) __NDK_FPABI_MATH__; +long double acosl(long double) __NDK_FPABI_MATH__; +long double asinhl(long double) __NDK_FPABI_MATH__; +long double asinl(long double) __NDK_FPABI_MATH__; +long double atan2l(long double, long double) __NDK_FPABI_MATH__; +long double atanhl(long double) __NDK_FPABI_MATH__; +long double atanl(long double) __NDK_FPABI_MATH__; +long double cbrtl(long double) __NDK_FPABI_MATH__; +#endif +long double ceill(long double) __NDK_FPABI_MATH__; +long double copysignl(long double, long double) __NDK_FPABI_MATH__ __pure2; +#if 0 +long double coshl(long double) __NDK_FPABI_MATH__; +long double cosl(long double) __NDK_FPABI_MATH__; +long double erfcl(long double) __NDK_FPABI_MATH__; +long double erfl(long double) __NDK_FPABI_MATH__; +long double exp2l(long double) __NDK_FPABI_MATH__; +long double expl(long double) __NDK_FPABI_MATH__; +long double expm1l(long double) __NDK_FPABI_MATH__; +#endif +long double fabsl(long double) __NDK_FPABI_MATH__ __pure2; +long double fdiml(long double, long double) __NDK_FPABI_MATH__; +long double floorl(long double) __NDK_FPABI_MATH__; +long double fmal(long double, long double, long double) __NDK_FPABI_MATH__; +long double fmaxl(long double, long double) __NDK_FPABI_MATH__ __pure2; +long double fminl(long double, long double) __NDK_FPABI_MATH__ __pure2; +#if 0 +long double fmodl(long double, long double) __NDK_FPABI_MATH__; +#endif +long double frexpl(long double value, int *) __NDK_FPABI_MATH__; /* fundamentally !__pure2 */ +#if 0 +long double hypotl(long double, long double) __NDK_FPABI_MATH__; +#endif +int ilogbl(long double) __NDK_FPABI_MATH__ __pure2; +long double ldexpl(long double, int) __NDK_FPABI_MATH__; +#if 0 +long double lgammal(long double) __NDK_FPABI_MATH__; +long long llrintl(long double) __NDK_FPABI_MATH__; +#endif +long long llroundl(long double) __NDK_FPABI_MATH__; +#if 0 +long double log10l(long double) __NDK_FPABI_MATH__; +long double log1pl(long double) __NDK_FPABI_MATH__; +long double log2l(long double) __NDK_FPABI_MATH__; +long double logbl(long double) __NDK_FPABI_MATH__; +long double logl(long double) __NDK_FPABI_MATH__; +long lrintl(long double) __NDK_FPABI_MATH__; +#endif +long lroundl(long double) __NDK_FPABI_MATH__; +#if 0 +long double modfl(long double, long double *) __NDK_FPABI_MATH__; /* fundamentally !__pure2 */ +#endif +long double nanl(const char *) __NDK_FPABI_MATH__ __pure2; +#if 0 +long double nearbyintl(long double) __NDK_FPABI_MATH__; +#endif +long double nextafterl(long double, long double) __NDK_FPABI_MATH__; +double nexttoward(double, long double) __NDK_FPABI_MATH__; +float nexttowardf(float, long double) __NDK_FPABI_MATH__; +long double nexttowardl(long double, long double) __NDK_FPABI_MATH__; +#if 0 +long double powl(long double, long double) __NDK_FPABI_MATH__; +long double remainderl(long double, long double) __NDK_FPABI_MATH__; +long double remquol(long double, long double, int *) __NDK_FPABI_MATH__; +long double rintl(long double) __NDK_FPABI_MATH__; +#endif +long double roundl(long double) __NDK_FPABI_MATH__; +long double scalblnl(long double, long) __NDK_FPABI_MATH__; +long double scalbnl(long double, int) __NDK_FPABI_MATH__; +#if 0 +long double sinhl(long double) __NDK_FPABI_MATH__; +long double sinl(long double) __NDK_FPABI_MATH__; +long double sqrtl(long double) __NDK_FPABI_MATH__; +long double tanhl(long double) __NDK_FPABI_MATH__; +long double tanl(long double) __NDK_FPABI_MATH__; +long double tgammal(long double) __NDK_FPABI_MATH__; +#endif +long double truncl(long double) __NDK_FPABI_MATH__; + +/* BIONIC: GLibc compatibility - required by the ARM toolchain */ +#ifdef _GNU_SOURCE +void sincos(double x, double *sin, double *cos) __NDK_FPABI_MATH__; +void sincosf(float x, float *sin, float *cos) __NDK_FPABI_MATH__; +void sincosl(long double x, long double *sin, long double *cos) __NDK_FPABI_MATH__; +#endif + +/* #endif */ /* __ISO_C_VISIBLE >= 1999 */ +__END_DECLS + +#endif /* !_MATH_H_ */ |