.\" Copyright (c) 2004 Stefan Farfeleder
.\" 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 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 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.
.\"
.\" $FreeBSD$
.\"
.Dd January 4, 2012
.Dt TGMATH 3
.Os
.Sh NAME
.Nm tgmath
.Nd "type-generic macros"
.Sh SYNOPSIS
.In tgmath.h
.Sh DESCRIPTION
The header
.In tgmath.h
provides type-generic macros
for
.In math.h
and
.In complex.h
functions that have
.Vt float
(suffixed with
.Sy f ) ,
.Vt double
and
.Vt "long double"
(suffixed with
.Sy l )
versions.
The arguments that vary across the three functions and have type
.Vt float , double
and
.Vt "long double" ,
respectively, are called
.Em "generic arguments" .
.Pp
The following rules describe which function is actually called if a
type-generic macro is invoked.
If any generic argument has type
.Vt "long double"
or
.Vt "long double complex" ,
the
.Vt "long double"
function is called.
Else, if any generic argument has type
.Vt double , "double complex"
or an integer type, the
.Vt double
version is invoked.
Otherwise, the macro expands to the
.Vt float
implementation.
.Pp
For the macros in the following table, both real and complex functions
exist.
The real functions are prototyped in
.In math.h
and the complex equivalents in
.In complex.h .
The complex function is called if any of the generic arguments is a
complex value.
Otherwise, the real equivalent is called.
.Bl -column -offset indent ".Fn acosh" "Sy real function" ".Sy complex function"
.It Sy Macro Ta Sy real function Ta Sy complex function
.It Fn acos Ta Fn acos Ta Fn cacos
.It Fn asin Ta Fn asin Ta Fn casin
.It Fn atan Ta Fn atan Ta Fn catan
.It Fn acosh Ta Fn acosh Ta Fn cacosh
.It Fn asinh Ta Fn asinh Ta Fn casinh
.It Fn atanh Ta Fn atanh Ta Fn catanh
.It Fn cos Ta Fn cos Ta Fn ccos
.It Fn sin Ta Fn sin Ta Fn csin
.It Fn tan  Ta Fn tan Ta Fn ctan
.It Fn cosh Ta Fn cosh Ta Fn ccosh
.It Fn sinh Ta Fn sinh Ta Fn csinh
.It Fn tanh Ta Fn tanh Ta Fn ctanh
.It Fn exp Ta Fn exp Ta Fn cexp
.It Fn log Ta Fn log Ta Fn clog
.It Fn pow Ta Fn pow Ta Fn cpow
.It Fn sqrt Ta Fn sqrt Ta Fn csqrt
.It Fn fabs Ta Fn fabs Ta Fn cabs
.El
.Pp
No complex functions exist for the following macros, so passing a
complex value to a generic argument invokes undefined behaviour:
.Bl -column -offset indent ".Fn nexttoward" ".Fn nexttoward" ".Fn nexttoward" ".Fn nexttoward"
.It Fn atan2 Ta Fn fma Ta Fn llround Ta Fn remainder
.It Fn cbrt Ta Fn fmax Ta Fn log10 Ta Fn remquo
.It Fn ceil Ta Fn fmin Ta Fn log1p Ta Fn rint
.It Fn copysign Ta Fn fmod Ta Fn log2 Ta Fn round
.It Fn erf Ta Fn frexp Ta Fn logb Ta Fn scalbn
.It Fn erfc Ta Fn hypot Ta Fn lrint Ta Fn scalbln
.It Fn exp2 Ta Fn ilogb Ta Fn lround Ta Fn tgamma
.It Fn expm1 Ta Fn ldexp Ta Fn nextbyint Ta Fn trunc
.It Fn fdim Ta Fn lgamma Ta Fn nextafter Ta \&
.It Fn floor Ta Fn llrint Ta Fn nexttoward Ta \&
.El
.Pp
The following macros always expand to a complex function:
.Bl -column -offset indent ".Fn cimag" ".Fn cimag" ".Fn cimag" ".Fn cimag" ".Fn cimag"
.It Fn carg Ta Fn cimag Ta Fn conj Ta Fn cproj Ta Fn creal
.El
.Pp
This header includes
.In complex.h
and
.In math.h .
.Sh STANDARDS
The header
.In tgmath.h
conforms to
.St -isoC-99 .
.Sh HISTORY
The header
.In tgmath.h
first appeared in
.Fx 5.3 .
.Sh COMPILER SUPPORT
Before
.St -isoC-2011 ,
the header
.In tgmath.h
could not be implemented with strictly conforming C code and needed
special compiler support.
As of
.St -isoC-2011 ,
this header file can be implemented using the
.Fn _Generic
language keyword.
In addition to compilers that support this keyword, this header file
works with GCC.
.Sh BUGS
Many of the functions mentioned here are not prototyped in
.In math.h
or
.In complex.h
as they are not yet implemented.
This prevents the corresponding type-generic macro from working at all.