mirror of
https://git.FreeBSD.org/src.git
synced 2024-12-23 11:18:54 +00:00
b58aba6b66
I've looked at the GCC sources and I now understand what's going wrong. THe C11 keywords are simply nonexistent when using C++ mode. They are marked as C-only in the parser. This is absolutely impractical for multiple reasons: - The C11 keywords do not conflict with C++ naming rules. They all start with _[A-Z]. There is no reason to make them C-only. - It makes it practically impossible for people to use these keywords in C header files and expect them to work from within C++ sources. As I said in my previous commit message: GCC is by far the weirdest compiler that I've ever used.
211 lines
8.8 KiB
C
211 lines
8.8 KiB
C
/*-
|
|
* Copyright (c) 2004 Stefan Farfeleder.
|
|
* All rights reserved.
|
|
*
|
|
* Copyright (c) 2012 Ed Schouten <ed@FreeBSD.org>
|
|
* All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions
|
|
* are met:
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer.
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
* documentation and/or other materials provided with the distribution.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
* SUCH DAMAGE.
|
|
*
|
|
* $FreeBSD$
|
|
*/
|
|
|
|
#ifndef _TGMATH_H_
|
|
#define _TGMATH_H_
|
|
|
|
#include <complex.h>
|
|
#include <math.h>
|
|
|
|
/*
|
|
* This implementation of <tgmath.h> uses the two following macros,
|
|
* which are based on the macros described in C11 proposal N1404:
|
|
* __tg_impl_simple(x, y, z, fnl, fn, fnf, ...)
|
|
* Invokes fnl() if the corresponding real type of x, y or z is long
|
|
* double, fn() if it is double or any has an integer type, and fnf()
|
|
* otherwise.
|
|
* __tg_impl_full(x, y, cfnl, cfn, cfnf, fnl, fn, fnf, ...)
|
|
* Invokes [c]fnl() if the corresponding real type of x or y is long
|
|
* double, [c]fn() if it is double or any has an integer type, and
|
|
* [c]fnf() otherwise. The function with the 'c' prefix is called if
|
|
* any of x or y is a complex number.
|
|
* Both macros call the chosen function with all additional arguments passed
|
|
* to them, as given by __VA_ARGS__.
|
|
*
|
|
* Note that these macros cannot be implemented with C's ?: operator,
|
|
* because the return type of the whole expression would incorrectly be long
|
|
* double complex regardless of the argument types.
|
|
*
|
|
* The structure of the C11 implementation of these macros can in
|
|
* principle be reused for non-C11 compilers, but due to an integer
|
|
* promotion bug for complex types in GCC 4.2, simply let non-C11
|
|
* compilers use an inefficient yet reliable version.
|
|
*/
|
|
|
|
#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 201112L) || \
|
|
__has_extension(c_generic_selections)
|
|
#define __tg_generic(x, cfnl, cfn, cfnf, fnl, fn, fnf) \
|
|
_Generic(x, \
|
|
long double _Complex: cfnl, \
|
|
double _Complex: cfn, \
|
|
float _Complex: cfnf, \
|
|
long double: fnl, \
|
|
default: fn, \
|
|
float: fnf \
|
|
)
|
|
#define __tg_type(x) \
|
|
__tg_generic(x, (long double _Complex)0, (double _Complex)0, \
|
|
(float _Complex)0, (long double)0, (double)0, (float)0)
|
|
#define __tg_impl_simple(x, y, z, fnl, fn, fnf, ...) \
|
|
__tg_generic( \
|
|
__tg_type(x) + __tg_type(y) + __tg_type(z), \
|
|
fnl, fn, fnf, fnl, fn, fnf)(__VA_ARGS__)
|
|
#define __tg_impl_full(x, y, cfnl, cfn, cfnf, fnl, fn, fnf, ...) \
|
|
__tg_generic( \
|
|
__tg_type(x) + __tg_type(y), \
|
|
cfnl, cfn, cfnf, fnl, fn, fnf)(__VA_ARGS__)
|
|
#elif defined(__generic)
|
|
#define __tg_generic_simple(x, fnl, fn, fnf) \
|
|
__generic(x, long double _Complex, fnl, \
|
|
__generic(x, double _Complex, fn, \
|
|
__generic(x, float _Complex, fnf, \
|
|
__generic(x, long double, fnl, \
|
|
__generic(x, float, fnf, fn)))))
|
|
#define __tg_impl_simple(x, y, z, fnl, fn, fnf, ...) \
|
|
__tg_generic_simple(x, \
|
|
__tg_generic_simple(y, \
|
|
__tg_generic_simple(z, fnl, fnl, fnl), \
|
|
__tg_generic_simple(z, fnl, fnl, fnl), \
|
|
__tg_generic_simple(z, fnl, fnl, fnl)), \
|
|
__tg_generic_simple(y, \
|
|
__tg_generic_simple(z, fnl, fnl, fnl), \
|
|
__tg_generic_simple(z, fnl, fn , fn ), \
|
|
__tg_generic_simple(z, fnl, fn , fn )), \
|
|
__tg_generic_simple(y, \
|
|
__tg_generic_simple(z, fnl, fnl, fnl), \
|
|
__tg_generic_simple(z, fnl, fn , fn ), \
|
|
__tg_generic_simple(z, fnl, fn , fnf)))(__VA_ARGS__)
|
|
#define __tg_generic_full(x, cfnl, cfn, cfnf, fnl, fn, fnf) \
|
|
__generic(x, long double _Complex, cfnl, \
|
|
__generic(x, double _Complex, cfn, \
|
|
__generic(x, float _Complex, cfnf, \
|
|
__generic(x, long double, fnl, \
|
|
__generic(x, float, fnf, fn)))))
|
|
#define __tg_impl_full(x, y, cfnl, cfn, cfnf, fnl, fn, fnf, ...) \
|
|
__tg_generic_full(x, \
|
|
__tg_generic_full(y, cfnl, cfnl, cfnl, cfnl, cfnl, cfnl), \
|
|
__tg_generic_full(y, cfnl, cfn , cfn , cfnl, cfn , cfn ), \
|
|
__tg_generic_full(y, cfnl, cfn , cfnf, cfnl, cfn , cfnf), \
|
|
__tg_generic_full(y, cfnl, cfnl, cfnl, fnl , fnl , fnl ), \
|
|
__tg_generic_full(y, cfnl, cfn , cfn , fnl , fn , fn ), \
|
|
__tg_generic_full(y, cfnl, cfn , cfnf, fnl , fn , fnf )) \
|
|
(__VA_ARGS__)
|
|
#else
|
|
#error "<tgmath.h> not implemented for this compiler"
|
|
#endif
|
|
|
|
/* Macros to save lots of repetition below */
|
|
#define __tg_simple(x, fn) \
|
|
__tg_impl_simple(x, x, x, fn##l, fn, fn##f, x)
|
|
#define __tg_simple2(x, y, fn) \
|
|
__tg_impl_simple(x, x, y, fn##l, fn, fn##f, x, y)
|
|
#define __tg_simple3(x, y, z, fn) \
|
|
__tg_impl_simple(x, y, z, fn##l, fn, fn##f, x, y, z)
|
|
#define __tg_simplev(x, fn, ...) \
|
|
__tg_impl_simple(x, x, x, fn##l, fn, fn##f, __VA_ARGS__)
|
|
#define __tg_full(x, fn) \
|
|
__tg_impl_full(x, x, c##fn##l, c##fn, c##fn##f, fn##l, fn, fn##f, x)
|
|
#define __tg_full2(x, y, fn) \
|
|
__tg_impl_full(x, y, c##fn##l, c##fn, c##fn##f, fn##l, fn, fn##f, x, y)
|
|
|
|
/* 7.22#4 -- These macros expand to real or complex functions, depending on
|
|
* the type of their arguments. */
|
|
#define acos(x) __tg_full(x, acos)
|
|
#define asin(x) __tg_full(x, asin)
|
|
#define atan(x) __tg_full(x, atan)
|
|
#define acosh(x) __tg_full(x, acosh)
|
|
#define asinh(x) __tg_full(x, asinh)
|
|
#define atanh(x) __tg_full(x, atanh)
|
|
#define cos(x) __tg_full(x, cos)
|
|
#define sin(x) __tg_full(x, sin)
|
|
#define tan(x) __tg_full(x, tan)
|
|
#define cosh(x) __tg_full(x, cosh)
|
|
#define sinh(x) __tg_full(x, sinh)
|
|
#define tanh(x) __tg_full(x, tanh)
|
|
#define exp(x) __tg_full(x, exp)
|
|
#define log(x) __tg_full(x, log)
|
|
#define pow(x, y) __tg_full2(x, y, pow)
|
|
#define sqrt(x) __tg_full(x, sqrt)
|
|
|
|
/* "The corresponding type-generic macro for fabs and cabs is fabs." */
|
|
#define fabs(x) __tg_impl_full(x, x, cabsl, cabs, cabsf, \
|
|
fabsl, fabs, fabsf, x)
|
|
|
|
/* 7.22#5 -- These macros are only defined for arguments with real type. */
|
|
#define atan2(x, y) __tg_simple2(x, y, atan2)
|
|
#define cbrt(x) __tg_simple(x, cbrt)
|
|
#define ceil(x) __tg_simple(x, ceil)
|
|
#define copysign(x, y) __tg_simple2(x, y, copysign)
|
|
#define erf(x) __tg_simple(x, erf)
|
|
#define erfc(x) __tg_simple(x, erfc)
|
|
#define exp2(x) __tg_simple(x, exp2)
|
|
#define expm1(x) __tg_simple(x, expm1)
|
|
#define fdim(x, y) __tg_simple2(x, y, fdim)
|
|
#define floor(x) __tg_simple(x, floor)
|
|
#define fma(x, y, z) __tg_simple3(x, y, z, fma)
|
|
#define fmax(x, y) __tg_simple2(x, y, fmax)
|
|
#define fmin(x, y) __tg_simple2(x, y, fmin)
|
|
#define fmod(x, y) __tg_simple2(x, y, fmod)
|
|
#define frexp(x, y) __tg_simplev(x, frexp, x, y)
|
|
#define hypot(x, y) __tg_simple2(x, y, hypot)
|
|
#define ilogb(x) __tg_simple(x, ilogb)
|
|
#define ldexp(x, y) __tg_simplev(x, ldexp, x, y)
|
|
#define lgamma(x) __tg_simple(x, lgamma)
|
|
#define llrint(x) __tg_simple(x, llrint)
|
|
#define llround(x) __tg_simple(x, llround)
|
|
#define log10(x) __tg_simple(x, log10)
|
|
#define log1p(x) __tg_simple(x, log1p)
|
|
#define log2(x) __tg_simple(x, log2)
|
|
#define logb(x) __tg_simple(x, logb)
|
|
#define lrint(x) __tg_simple(x, lrint)
|
|
#define lround(x) __tg_simple(x, lround)
|
|
#define nearbyint(x) __tg_simple(x, nearbyint)
|
|
#define nextafter(x, y) __tg_simple2(x, y, nextafter)
|
|
#define nexttoward(x, y) __tg_simplev(x, nexttoward, x, y)
|
|
#define remainder(x, y) __tg_simple2(x, y, remainder)
|
|
#define remquo(x, y, z) __tg_impl_simple(x, x, y, remquol, remquo, \
|
|
remquof, x, y, z)
|
|
#define rint(x) __tg_simple(x, rint)
|
|
#define round(x) __tg_simple(x, round)
|
|
#define scalbn(x, y) __tg_simplev(x, scalbn, x, y)
|
|
#define scalbln(x, y) __tg_simplev(x, scalbln, x, y)
|
|
#define tgamma(x) __tg_simple(x, tgamma)
|
|
#define trunc(x) __tg_simple(x, trunc)
|
|
|
|
/* 7.22#6 -- These macros always expand to complex functions. */
|
|
#define carg(x) __tg_simple(x, carg)
|
|
#define cimag(x) __tg_simple(x, cimag)
|
|
#define conj(x) __tg_simple(x, conj)
|
|
#define cproj(x) __tg_simple(x, cproj)
|
|
#define creal(x) __tg_simple(x, creal)
|
|
|
|
#endif /* !_TGMATH_H_ */
|