aboutsummaryrefslogtreecommitdiff
path: root/win32/include/math.h
diff options
context:
space:
mode:
authorbellard <bellard>2005-04-17 13:10:37 +0000
committerbellard <bellard>2005-04-17 13:10:37 +0000
commit5556cf15656ad1e433ec20502411411e30a4cdc7 (patch)
tree2105947a5bdaefd49ef9db9af5080498ef04ef42 /win32/include/math.h
parent6144d433211b4c5ceb68d6ffa500d6d54c3a2d67 (diff)
downloadtinycc-5556cf15656ad1e433ec20502411411e30a4cdc7.tar.gz
tinycc-5556cf15656ad1e433ec20502411411e30a4cdc7.tar.bz2
dos2unix
Diffstat (limited to 'win32/include/math.h')
-rw-r--r--win32/include/math.h876
1 files changed, 438 insertions, 438 deletions
diff --git a/win32/include/math.h b/win32/include/math.h
index 72210e9..ffb133a 100644
--- a/win32/include/math.h
+++ b/win32/include/math.h
@@ -1,438 +1,438 @@
-/*
- * math.h
- *
- * Mathematical functions.
- *
- * This file is part of the Mingw32 package.
- *
- * Contributors:
- * Created by Colin Peters <colin@bird.fu.is.saga-u.ac.jp>
- *
- * THIS SOFTWARE IS NOT COPYRIGHTED
- *
- * This source code is offered for use in the public domain. You may
- * use, modify or distribute it freely.
- *
- * This code is distributed in the hope that it will be useful but
- * WITHOUT ANY WARRANTY. ALL WARRANTIES, EXPRESS OR IMPLIED ARE HEREBY
- * DISCLAIMED. This includes but is not limited to warranties of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * $Revision: 1.1 $
- * $Author: bellard $
- * $Date: 2005/04/14 23:49:21 $
- *
- */
-
-#ifndef _MATH_H_
-#define _MATH_H_
-
-/* All the headers include this file. */
-#include <_mingw.h>
-
-/*
- * Types for the _exception structure.
- */
-
-#define _DOMAIN 1 /* domain error in argument */
-#define _SING 2 /* singularity */
-#define _OVERFLOW 3 /* range overflow */
-#define _UNDERFLOW 4 /* range underflow */
-#define _TLOSS 5 /* total loss of precision */
-#define _PLOSS 6 /* partial loss of precision */
-
-/*
- * Exception types with non-ANSI names for compatibility.
- */
-
-#ifndef __STRICT_ANSI__
-#ifndef _NO_OLDNAMES
-
-#define DOMAIN _DOMAIN
-#define SING _SING
-#define OVERFLOW _OVERFLOW
-#define UNDERFLOW _UNDERFLOW
-#define TLOSS _TLOSS
-#define PLOSS _PLOSS
-
-#endif /* Not _NO_OLDNAMES */
-#endif /* Not __STRICT_ANSI__ */
-
-
-/* These are also defined in Mingw float.h; needed here as well to work
- around GCC build issues. */
-#ifndef __STRICT_ANSI__
-#ifndef __MINGW_FPCLASS_DEFINED
-#define __MINGW_FPCLASS_DEFINED 1
-/* IEEE 754 classication */
-#define _FPCLASS_SNAN 0x0001 /* Signaling "Not a Number" */
-#define _FPCLASS_QNAN 0x0002 /* Quiet "Not a Number" */
-#define _FPCLASS_NINF 0x0004 /* Negative Infinity */
-#define _FPCLASS_NN 0x0008 /* Negative Normal */
-#define _FPCLASS_ND 0x0010 /* Negative Denormal */
-#define _FPCLASS_NZ 0x0020 /* Negative Zero */
-#define _FPCLASS_PZ 0x0040 /* Positive Zero */
-#define _FPCLASS_PD 0x0080 /* Positive Denormal */
-#define _FPCLASS_PN 0x0100 /* Positive Normal */
-#define _FPCLASS_PINF 0x0200 /* Positive Infinity */
-#endif /* __MINGW_FPCLASS_DEFINED */
-#endif /* Not __STRICT_ANSI__ */
-
-#ifndef RC_INVOKED
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-/*
- * HUGE_VAL is returned by strtod when the value would overflow the
- * representation of 'double'. There are other uses as well.
- *
- * __imp__HUGE is a pointer to the actual variable _HUGE in
- * MSVCRT.DLL. If we used _HUGE directly we would get a pointer
- * to a thunk function.
- *
- * NOTE: The CRTDLL version uses _HUGE_dll instead.
- */
-
-#ifndef __DECLSPEC_SUPPORTED
-
-#ifdef __MSVCRT__
-extern double* __imp__HUGE;
-#define HUGE_VAL (*__imp__HUGE)
-#else
-/* CRTDLL */
-extern double* __imp__HUGE_dll;
-#define HUGE_VAL (*__imp__HUGE_dll)
-#endif
-
-#else /* __DECLSPEC_SUPPORTED */
-
-#ifdef __MSVCRT__
-__MINGW_IMPORT double _HUGE;
-#define HUGE_VAL _HUGE
-#else
-/* CRTDLL */
-__MINGW_IMPORT double _HUGE_dll;
-#define HUGE_VAL _HUGE_dll
-#endif
-
-#endif /* __DECLSPEC_SUPPORTED */
-
-struct _exception
-{
- int type;
- char *name;
- double arg1;
- double arg2;
- double retval;
-};
-
-
-double sin (double);
-double cos (double);
-double tan (double);
-double sinh (double);
-double cosh (double);
-double tanh (double);
-double asin (double);
-double acos (double);
-double atan (double);
-double atan2 (double, double);
-double exp (double);
-double log (double);
-double log10 (double);
-double pow (double, double);
-double sqrt (double);
-double ceil (double);
-double floor (double);
-double fabs (double);
-double ldexp (double, int);
-double frexp (double, int*);
-double modf (double, double*);
-double fmod (double, double);
-
-
-#ifndef __STRICT_ANSI__
-
-/* Complex number (for cabs) */
-struct _complex
-{
- double x; /* Real part */
- double y; /* Imaginary part */
-};
-
-double _cabs (struct _complex);
-double _hypot (double, double);
-double _j0 (double);
-double _j1 (double);
-double _jn (int, double);
-double _y0 (double);
-double _y1 (double);
-double _yn (int, double);
-int _matherr (struct _exception *);
-
-/* These are also declared in Mingw float.h; needed here as well to work
- around GCC build issues. */
-/* BEGIN FLOAT.H COPY */
-/*
- * IEEE recommended functions
- */
-
-double _chgsign (double);
-double _copysign (double, double);
-double _logb (double);
-double _nextafter (double, double);
-double _scalb (double, long);
-
-int _finite (double);
-int _fpclass (double);
-int _isnan (double);
-
-/* END FLOAT.H COPY */
-
-#if !defined (_NO_OLDNAMES) \
- || (defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L )
-
-/*
- * Non-underscored versions of non-ANSI functions. These reside in
- * liboldnames.a. They are now also ISO C99 standand names.
- * Provided for extra portability.
- */
-
-double cabs (struct _complex);
-double hypot (double, double);
-double j0 (double);
-double j1 (double);
-double jn (int, double);
-double y0 (double);
-double y1 (double);
-double yn (int, double);
-
-#endif /* Not _NO_OLDNAMES */
-
-#endif /* Not __STRICT_ANSI__ */
-
-#ifdef __cplusplus
-}
-#endif
-#endif /* Not RC_INVOKED */
-
-
-#ifndef __NO_ISOCEXT
-
-#define INFINITY HUGE_VAL
-#define NAN (0.0F/0.0F)
-
-/*
- Return values for fpclassify.
- These are based on Intel x87 fpu condition codes
- in the high byte of status word and differ from
- the return values for MS IEEE 754 extension _fpclass()
-*/
-#define FP_NAN 0x0100
-#define FP_NORMAL 0x0400
-#define FP_INFINITE (FP_NAN | FP_NORMAL)
-#define FP_ZERO 0x4000
-#define FP_SUBNORMAL (FP_NORMAL | FP_ZERO)
-/* 0x0200 is signbit mask */
-
-#ifndef RC_INVOKED
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-double nan(const char *tagp);
-float nanf(const char *tagp);
-
-#ifndef __STRICT_ANSI__
-#define nan() nan("")
-#define nanf() nanf("")
-#endif
-
-
-/*
- We can't inline float, because we want to ensure truncation
- to semantic type before classification. If we extend to long
- double, we will also need to make double extern only.
- (A normal long double value might become subnormal when
- converted to double, and zero when converted to float.)
-*/
-extern __inline__ int __fpclassify (double x){
- unsigned short sw;
- __asm__ ("fxam; fstsw %%ax;" : "=a" (sw): "t" (x));
- return sw & (FP_NAN | FP_NORMAL | FP_ZERO );
-}
-
-extern int __fpclassifyf (float);
-
-#define fpclassify(x) ((sizeof(x) == sizeof(float)) ? __fpclassifyf(x) \
- : __fpclassify(x))
-
-/* We don't need to worry about trucation here:
- A NaN stays a NaN. */
-
-extern __inline__ int __isnan (double _x)
-{
- unsigned short sw;
- __asm__ ("fxam;"
- "fstsw %%ax": "=a" (sw) : "t" (_x));
- return (sw & (FP_NAN | FP_NORMAL | FP_INFINITE | FP_ZERO | FP_SUBNORMAL))
- == FP_NAN;
-}
-
-extern __inline__ int __isnanf (float _x)
-{
- unsigned short sw;
- __asm__ ("fxam;"
- "fstsw %%ax": "=a" (sw) : "t" (_x));
- return (sw & (FP_NAN | FP_NORMAL | FP_INFINITE | FP_ZERO | FP_SUBNORMAL))
- == FP_NAN;
-}
-
-#define isnan(x) ((sizeof(x) == sizeof(float)) ? __isnanf(x) \
- : __isnan(x))
-
-
-#define isfinite(x) ((fpclassify(x) & FP_NAN) == 0)
-#define isinf(x) (fpclassify(x) == FP_INFINITE)
-#define isnormal(x) (fpclassify(x) == FP_NORMAL)
-
-
-extern __inline__ int __signbit (double x) {
- unsigned short stw;
- __asm__ ( "fxam; fstsw %%ax;": "=a" (stw) : "t" (x));
- return stw & 0x0200;
-}
-
-extern __inline__ int __signbitf (float x) {
- unsigned short stw;
- __asm__ ("fxam; fstsw %%ax;": "=a" (stw) : "t" (x));
- return stw & 0x0200;
-}
-
-#define signbit(x) ((sizeof(x) == sizeof(float)) ? __signbitf(x) \
- : __signbit(x))
-/*
- * With these functions, comparisons involving quiet NaNs set the FP
- * condition code to "unordered". The IEEE floating-point spec
- * dictates that the result of floating-point comparisons should be
- * false whenever a NaN is involved, with the exception of the !=,
- * which always returns true.
- */
-
-#if __GNUC__ >= 3
-
-#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
-/* helper */
-extern __inline__ int __fp_unordered_compare (double x, double y){
- unsigned short retval;
- __asm__ ("fucom %%st(1);"
- "fnstsw;": "=a" (retval) : "t" (x), "u" (y));
- return retval;
-}
-
-#define isgreater(x, y) ((__fp_unordered_compare(x, y) \
- & 0x4500) == 0)
-#define isless(x, y) ((__fp_unordered_compare (y, x) \
- & 0x4500) == 0)
-#define isgreaterequal(x, y) ((__fp_unordered_compare (x, y) \
- & FP_INFINITE) == 0)
-#define islessequal(x, y) ((__fp_unordered_compare(y, x) \
- & FP_INFINITE) == 0)
-#define islessgreater(x, y) ((__fp_unordered_compare(x, y) \
- & FP_SUBNORMAL) == 0)
-#define isunordered(x, y) ((__fp_unordered_compare(x, y) \
- & 0x4500) == 0x4500)
-
-#endif
-
-/* round, using fpu control word settings */
-extern __inline__ double rint (double x)
-{
- double retval;
- __asm__ ("frndint;": "=t" (retval) : "0" (x));
- return retval;
-}
-
-extern __inline__ float rintf (float x)
-{
- float retval;
- __asm__ ("frndint;" : "=t" (retval) : "0" (x) );
- return retval;
-}
-
-/* round away from zero, regardless of fpu control word settings */
-extern double round (double);
-extern float roundf (float);
-
-/* round towards zero, regardless of fpu control word settings */
-extern double trunc (double);
-extern float truncf (float);
-
-
-/* fmax and fmin.
- NaN arguments are treated as missing data: if one argument is a NaN and the other numeric, then the
- these functions choose the numeric value.
-*/
-
-extern double fmax (double, double);
-extern double fmin (double, double);
-extern float fmaxf (float, float);
-float fminf (float, float);
-
-/* return x * y + z as a ternary op */
-extern double fma (double, double, double);
-extern float fmaf (float, float, float);
-
-/* one lonely transcendental */
-extern double log2 (double _x);
-extern float log2f (float _x);
-
-/* The underscored versions are in MSVCRT.dll.
- The stubs for these are in libmingwex.a */
-
-double copysign (double, double);
-float copysignf (float, float);
-double logb (double);
-float logbf (float);
-double nextafter (double, double);
-float nextafterf (float, float);
-double scalb (double, long);
-float scalbf (float, long);
-
-#if !defined (__STRICT_ANSI__) /* inline using non-ANSI functions */
-extern __inline__ double copysign (double x, double y)
- { return _copysign(x, y); }
-extern __inline__ float copysignf (float x, float y)
- { return _copysign(x, y); }
-extern __inline__ double logb (double x)
- { return _logb(x); }
-extern __inline__ float logbf (float x)
- { return _logb(x); }
-extern __inline__ double nextafter(double x, double y)
- { return _nextafter(x, y); }
-extern __inline__ float nextafterf(float x, float y)
- { return _nextafter(x, y); }
-extern __inline__ double scalb (double x, long i)
- { return _scalb (x, i); }
-extern __inline__ float scalbf (float x, long i)
- { return _scalb(x, i); }
-#endif /* (__STRICT_ANSI__) */
-
-#ifdef __cplusplus
-}
-#endif
-#endif /* Not RC_INVOKED */
-
-#endif /* __NO_ISOCEXT */
-
-#endif /* Not _MATH_H_ */
-
+/*
+ * math.h
+ *
+ * Mathematical functions.
+ *
+ * This file is part of the Mingw32 package.
+ *
+ * Contributors:
+ * Created by Colin Peters <colin@bird.fu.is.saga-u.ac.jp>
+ *
+ * THIS SOFTWARE IS NOT COPYRIGHTED
+ *
+ * This source code is offered for use in the public domain. You may
+ * use, modify or distribute it freely.
+ *
+ * This code is distributed in the hope that it will be useful but
+ * WITHOUT ANY WARRANTY. ALL WARRANTIES, EXPRESS OR IMPLIED ARE HEREBY
+ * DISCLAIMED. This includes but is not limited to warranties of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * $Revision: 1.2 $
+ * $Author: bellard $
+ * $Date: 2005/04/17 13:14:29 $
+ *
+ */
+
+#ifndef _MATH_H_
+#define _MATH_H_
+
+/* All the headers include this file. */
+#include <_mingw.h>
+
+/*
+ * Types for the _exception structure.
+ */
+
+#define _DOMAIN 1 /* domain error in argument */
+#define _SING 2 /* singularity */
+#define _OVERFLOW 3 /* range overflow */
+#define _UNDERFLOW 4 /* range underflow */
+#define _TLOSS 5 /* total loss of precision */
+#define _PLOSS 6 /* partial loss of precision */
+
+/*
+ * Exception types with non-ANSI names for compatibility.
+ */
+
+#ifndef __STRICT_ANSI__
+#ifndef _NO_OLDNAMES
+
+#define DOMAIN _DOMAIN
+#define SING _SING
+#define OVERFLOW _OVERFLOW
+#define UNDERFLOW _UNDERFLOW
+#define TLOSS _TLOSS
+#define PLOSS _PLOSS
+
+#endif /* Not _NO_OLDNAMES */
+#endif /* Not __STRICT_ANSI__ */
+
+
+/* These are also defined in Mingw float.h; needed here as well to work
+ around GCC build issues. */
+#ifndef __STRICT_ANSI__
+#ifndef __MINGW_FPCLASS_DEFINED
+#define __MINGW_FPCLASS_DEFINED 1
+/* IEEE 754 classication */
+#define _FPCLASS_SNAN 0x0001 /* Signaling "Not a Number" */
+#define _FPCLASS_QNAN 0x0002 /* Quiet "Not a Number" */
+#define _FPCLASS_NINF 0x0004 /* Negative Infinity */
+#define _FPCLASS_NN 0x0008 /* Negative Normal */
+#define _FPCLASS_ND 0x0010 /* Negative Denormal */
+#define _FPCLASS_NZ 0x0020 /* Negative Zero */
+#define _FPCLASS_PZ 0x0040 /* Positive Zero */
+#define _FPCLASS_PD 0x0080 /* Positive Denormal */
+#define _FPCLASS_PN 0x0100 /* Positive Normal */
+#define _FPCLASS_PINF 0x0200 /* Positive Infinity */
+#endif /* __MINGW_FPCLASS_DEFINED */
+#endif /* Not __STRICT_ANSI__ */
+
+#ifndef RC_INVOKED
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * HUGE_VAL is returned by strtod when the value would overflow the
+ * representation of 'double'. There are other uses as well.
+ *
+ * __imp__HUGE is a pointer to the actual variable _HUGE in
+ * MSVCRT.DLL. If we used _HUGE directly we would get a pointer
+ * to a thunk function.
+ *
+ * NOTE: The CRTDLL version uses _HUGE_dll instead.
+ */
+
+#ifndef __DECLSPEC_SUPPORTED
+
+#ifdef __MSVCRT__
+extern double* __imp__HUGE;
+#define HUGE_VAL (*__imp__HUGE)
+#else
+/* CRTDLL */
+extern double* __imp__HUGE_dll;
+#define HUGE_VAL (*__imp__HUGE_dll)
+#endif
+
+#else /* __DECLSPEC_SUPPORTED */
+
+#ifdef __MSVCRT__
+__MINGW_IMPORT double _HUGE;
+#define HUGE_VAL _HUGE
+#else
+/* CRTDLL */
+__MINGW_IMPORT double _HUGE_dll;
+#define HUGE_VAL _HUGE_dll
+#endif
+
+#endif /* __DECLSPEC_SUPPORTED */
+
+struct _exception
+{
+ int type;
+ char *name;
+ double arg1;
+ double arg2;
+ double retval;
+};
+
+
+double sin (double);
+double cos (double);
+double tan (double);
+double sinh (double);
+double cosh (double);
+double tanh (double);
+double asin (double);
+double acos (double);
+double atan (double);
+double atan2 (double, double);
+double exp (double);
+double log (double);
+double log10 (double);
+double pow (double, double);
+double sqrt (double);
+double ceil (double);
+double floor (double);
+double fabs (double);
+double ldexp (double, int);
+double frexp (double, int*);
+double modf (double, double*);
+double fmod (double, double);
+
+
+#ifndef __STRICT_ANSI__
+
+/* Complex number (for cabs) */
+struct _complex
+{
+ double x; /* Real part */
+ double y; /* Imaginary part */
+};
+
+double _cabs (struct _complex);
+double _hypot (double, double);
+double _j0 (double);
+double _j1 (double);
+double _jn (int, double);
+double _y0 (double);
+double _y1 (double);
+double _yn (int, double);
+int _matherr (struct _exception *);
+
+/* These are also declared in Mingw float.h; needed here as well to work
+ around GCC build issues. */
+/* BEGIN FLOAT.H COPY */
+/*
+ * IEEE recommended functions
+ */
+
+double _chgsign (double);
+double _copysign (double, double);
+double _logb (double);
+double _nextafter (double, double);
+double _scalb (double, long);
+
+int _finite (double);
+int _fpclass (double);
+int _isnan (double);
+
+/* END FLOAT.H COPY */
+
+#if !defined (_NO_OLDNAMES) \
+ || (defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L )
+
+/*
+ * Non-underscored versions of non-ANSI functions. These reside in
+ * liboldnames.a. They are now also ISO C99 standand names.
+ * Provided for extra portability.
+ */
+
+double cabs (struct _complex);
+double hypot (double, double);
+double j0 (double);
+double j1 (double);
+double jn (int, double);
+double y0 (double);
+double y1 (double);
+double yn (int, double);
+
+#endif /* Not _NO_OLDNAMES */
+
+#endif /* Not __STRICT_ANSI__ */
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* Not RC_INVOKED */
+
+
+#ifndef __NO_ISOCEXT
+
+#define INFINITY HUGE_VAL
+#define NAN (0.0F/0.0F)
+
+/*
+ Return values for fpclassify.
+ These are based on Intel x87 fpu condition codes
+ in the high byte of status word and differ from
+ the return values for MS IEEE 754 extension _fpclass()
+*/
+#define FP_NAN 0x0100
+#define FP_NORMAL 0x0400
+#define FP_INFINITE (FP_NAN | FP_NORMAL)
+#define FP_ZERO 0x4000
+#define FP_SUBNORMAL (FP_NORMAL | FP_ZERO)
+/* 0x0200 is signbit mask */
+
+#ifndef RC_INVOKED
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+double nan(const char *tagp);
+float nanf(const char *tagp);
+
+#ifndef __STRICT_ANSI__
+#define nan() nan("")
+#define nanf() nanf("")
+#endif
+
+
+/*
+ We can't inline float, because we want to ensure truncation
+ to semantic type before classification. If we extend to long
+ double, we will also need to make double extern only.
+ (A normal long double value might become subnormal when
+ converted to double, and zero when converted to float.)
+*/
+extern __inline__ int __fpclassify (double x){
+ unsigned short sw;
+ __asm__ ("fxam; fstsw %%ax;" : "=a" (sw): "t" (x));
+ return sw & (FP_NAN | FP_NORMAL | FP_ZERO );
+}
+
+extern int __fpclassifyf (float);
+
+#define fpclassify(x) ((sizeof(x) == sizeof(float)) ? __fpclassifyf(x) \
+ : __fpclassify(x))
+
+/* We don't need to worry about trucation here:
+ A NaN stays a NaN. */
+
+extern __inline__ int __isnan (double _x)
+{
+ unsigned short sw;
+ __asm__ ("fxam;"
+ "fstsw %%ax": "=a" (sw) : "t" (_x));
+ return (sw & (FP_NAN | FP_NORMAL | FP_INFINITE | FP_ZERO | FP_SUBNORMAL))
+ == FP_NAN;
+}
+
+extern __inline__ int __isnanf (float _x)
+{
+ unsigned short sw;
+ __asm__ ("fxam;"
+ "fstsw %%ax": "=a" (sw) : "t" (_x));
+ return (sw & (FP_NAN | FP_NORMAL | FP_INFINITE | FP_ZERO | FP_SUBNORMAL))
+ == FP_NAN;
+}
+
+#define isnan(x) ((sizeof(x) == sizeof(float)) ? __isnanf(x) \
+ : __isnan(x))
+
+
+#define isfinite(x) ((fpclassify(x) & FP_NAN) == 0)
+#define isinf(x) (fpclassify(x) == FP_INFINITE)
+#define isnormal(x) (fpclassify(x) == FP_NORMAL)
+
+
+extern __inline__ int __signbit (double x) {
+ unsigned short stw;
+ __asm__ ( "fxam; fstsw %%ax;": "=a" (stw) : "t" (x));
+ return stw & 0x0200;
+}
+
+extern __inline__ int __signbitf (float x) {
+ unsigned short stw;
+ __asm__ ("fxam; fstsw %%ax;": "=a" (stw) : "t" (x));
+ return stw & 0x0200;
+}
+
+#define signbit(x) ((sizeof(x) == sizeof(float)) ? __signbitf(x) \
+ : __signbit(x))
+/*
+ * With these functions, comparisons involving quiet NaNs set the FP
+ * condition code to "unordered". The IEEE floating-point spec
+ * dictates that the result of floating-point comparisons should be
+ * false whenever a NaN is involved, with the exception of the !=,
+ * which always returns true.
+ */
+
+#if __GNUC__ >= 3
+
+#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
+/* helper */
+extern __inline__ int __fp_unordered_compare (double x, double y){
+ unsigned short retval;
+ __asm__ ("fucom %%st(1);"
+ "fnstsw;": "=a" (retval) : "t" (x), "u" (y));
+ return retval;
+}
+
+#define isgreater(x, y) ((__fp_unordered_compare(x, y) \
+ & 0x4500) == 0)
+#define isless(x, y) ((__fp_unordered_compare (y, x) \
+ & 0x4500) == 0)
+#define isgreaterequal(x, y) ((__fp_unordered_compare (x, y) \
+ & FP_INFINITE) == 0)
+#define islessequal(x, y) ((__fp_unordered_compare(y, x) \
+ & FP_INFINITE) == 0)
+#define islessgreater(x, y) ((__fp_unordered_compare(x, y) \
+ & FP_SUBNORMAL) == 0)
+#define isunordered(x, y) ((__fp_unordered_compare(x, y) \
+ & 0x4500) == 0x4500)
+
+#endif
+
+/* round, using fpu control word settings */
+extern __inline__ double rint (double x)
+{
+ double retval;
+ __asm__ ("frndint;": "=t" (retval) : "0" (x));
+ return retval;
+}
+
+extern __inline__ float rintf (float x)
+{
+ float retval;
+ __asm__ ("frndint;" : "=t" (retval) : "0" (x) );
+ return retval;
+}
+
+/* round away from zero, regardless of fpu control word settings */
+extern double round (double);
+extern float roundf (float);
+
+/* round towards zero, regardless of fpu control word settings */
+extern double trunc (double);
+extern float truncf (float);
+
+
+/* fmax and fmin.
+ NaN arguments are treated as missing data: if one argument is a NaN and the other numeric, then the
+ these functions choose the numeric value.
+*/
+
+extern double fmax (double, double);
+extern double fmin (double, double);
+extern float fmaxf (float, float);
+float fminf (float, float);
+
+/* return x * y + z as a ternary op */
+extern double fma (double, double, double);
+extern float fmaf (float, float, float);
+
+/* one lonely transcendental */
+extern double log2 (double _x);
+extern float log2f (float _x);
+
+/* The underscored versions are in MSVCRT.dll.
+ The stubs for these are in libmingwex.a */
+
+double copysign (double, double);
+float copysignf (float, float);
+double logb (double);
+float logbf (float);
+double nextafter (double, double);
+float nextafterf (float, float);
+double scalb (double, long);
+float scalbf (float, long);
+
+#if !defined (__STRICT_ANSI__) /* inline using non-ANSI functions */
+extern __inline__ double copysign (double x, double y)
+ { return _copysign(x, y); }
+extern __inline__ float copysignf (float x, float y)
+ { return _copysign(x, y); }
+extern __inline__ double logb (double x)
+ { return _logb(x); }
+extern __inline__ float logbf (float x)
+ { return _logb(x); }
+extern __inline__ double nextafter(double x, double y)
+ { return _nextafter(x, y); }
+extern __inline__ float nextafterf(float x, float y)
+ { return _nextafter(x, y); }
+extern __inline__ double scalb (double x, long i)
+ { return _scalb (x, i); }
+extern __inline__ float scalbf (float x, long i)
+ { return _scalb(x, i); }
+#endif /* (__STRICT_ANSI__) */
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* Not RC_INVOKED */
+
+#endif /* __NO_ISOCEXT */
+
+#endif /* Not _MATH_H_ */
+