
MSVC-specific, MSVCRT-specific, or Windows-specific. Because Clang can also define _MSC_VER, and MSVCRT is not necessarily the only C runtime, these macros should not be used interchangeably. This patch divides all Windows-related bits into the aforementioned categories. Two new macros are introduced: - _LIBCPP_MSVC: Defined when compiling with MSVC. Detected using _MSC_VER, excluding Clang. - _LIBCPP_MSVCRT: Defined when using the Microsoft CRT. This is the default when _WIN32 is defined. This leaves _WIN32 for code using the Windows API. This also corrects the spelling of _LIBCP_HAS_IS_BASE_OF to _LIBCPP_HAS_IS_BASE_OF. Nico, please prepare a patch for CREDITS.TXT, thanks. git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@187593 91177308-0d34-0410-b5e6-96231b3b80d8
114 lines
2.9 KiB
C++
114 lines
2.9 KiB
C++
// -*- C++ -*-
|
|
//===---------------------- support/win32/math_win32.h --------------------===//
|
|
//
|
|
// The LLVM Compiler Infrastructure
|
|
//
|
|
// This file is dual licensed under the MIT and the University of Illinois Open
|
|
// Source Licenses. See LICENSE.TXT for details.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#ifndef _LIBCPP_SUPPORT_WIN32_MATH_WIN32_H
|
|
#define _LIBCPP_SUPPORT_WIN32_MATH_WIN32_H
|
|
|
|
#if !defined(_LIBCPP_MSVCRT)
|
|
#error "This header complements Microsoft's C Runtime library, and should not be included otherwise."
|
|
#else
|
|
|
|
#include <math.h>
|
|
|
|
typedef float float_t;
|
|
typedef double double_t;
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool isfinite( double num )
|
|
{
|
|
return _finite(num) != 0;
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE bool isinf( double num )
|
|
{
|
|
return !isfinite(num) && !_isnan(num);
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE bool isnan( double num )
|
|
{
|
|
return _isnan(num) != 0;
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE bool isnormal( double num )
|
|
{
|
|
int class_ = _fpclass(num);
|
|
return class_ == _FPCLASS_NN || class_ == _FPCLASS_PN;
|
|
}
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool isgreater( double x, double y )
|
|
{
|
|
if(_fpclass(x) == _FPCLASS_SNAN || _fpclass(y) == _FPCLASS_SNAN) return false;
|
|
else return x > y;
|
|
}
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool isgreaterequal( double x, double y )
|
|
{
|
|
if(_fpclass(x) == _FPCLASS_SNAN || _fpclass(y) == _FPCLASS_SNAN) return false;
|
|
else return x >= y;
|
|
}
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool isless( double x, double y )
|
|
{
|
|
if(_fpclass(x) == _FPCLASS_SNAN || _fpclass(y) == _FPCLASS_SNAN) return false;
|
|
else return x < y;
|
|
}
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool islessequal( double x, double y )
|
|
{
|
|
if(::_fpclass(x) == _FPCLASS_SNAN || ::_fpclass(y) == _FPCLASS_SNAN) return false;
|
|
else return x <= y;
|
|
}
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool islessgreater( double x, double y )
|
|
{
|
|
if(::_fpclass(x) == _FPCLASS_SNAN || ::_fpclass(y) == _FPCLASS_SNAN) return false;
|
|
else return x < y || x > y;
|
|
}
|
|
|
|
_LIBCPP_ALWAYS_INLINE bool isunordered( double x, double y )
|
|
{
|
|
return isnan(x) || isnan(y);
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE bool signbit( double num )
|
|
{
|
|
switch(_fpclass(num))
|
|
{
|
|
case _FPCLASS_SNAN:
|
|
case _FPCLASS_QNAN:
|
|
case _FPCLASS_NINF:
|
|
case _FPCLASS_NN:
|
|
case _FPCLASS_ND:
|
|
case _FPCLASS_NZ:
|
|
return true;
|
|
case _FPCLASS_PZ:
|
|
case _FPCLASS_PD:
|
|
case _FPCLASS_PN:
|
|
case _FPCLASS_PINF:
|
|
return false;
|
|
}
|
|
return false;
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE float copysignf( float x, float y )
|
|
{
|
|
return (signbit (x) != signbit (y) ? - x : x);
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE double copysign( double x, double y )
|
|
{
|
|
return ::_copysign(x,y);
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE double copysignl( long double x, long double y )
|
|
{
|
|
return ::_copysignl(x,y);
|
|
}
|
|
_LIBCPP_ALWAYS_INLINE int fpclassify( double num )
|
|
{
|
|
return _fpclass(num);
|
|
}
|
|
|
|
#endif // _LIBCPP_MSVCRT
|
|
|
|
#endif // _LIBCPP_SUPPORT_WIN32_MATH_WIN32_H
|