tgmath.h

Type-generic macros for invoking functions declared in math.h and complex.h.

The math.h and complex.h header files declare many functions in variants to accept arguments and return values in each of the floating-point types, double, float, and long double. A function typically uses the double type. A function of the same name with a suffix of "f" uses the float type and a function with a suffix of "l" uses the long double type. For example, math.h declares the cos(), cosf(), and cosl() functions. These functions all behave similarly but accept and return values of type double, float, and long double, respectively.

The macros defined in the tgmath.h header file choose and invoke a corresponding function in math.h or complex.h based on the types of the arguments passed to the macro.

The header file follows these rules, in this order, to choose a function to invoke:
if any of the macro's arguments has a complex type, use the complex variant of the macro's corresponding function
if any of the macro's arguments has a type of long double then use a long double variant
if any of the macro arguments has type double or an integer type then use a double variant
otherwise, use the float variant

The table below table lists the macros in tgmath.h and the corresponding function variants in math.h and complex.h.

Table 1. Macros in tgmath.h
Macro	Math.h Variant	Complex.h Variant
acos()	acos	cacos
asin()	asin	casin
atan()	atan	catan
acosh()	acosh	cacosh
asinh()	asinh	casinh
atanh()	atanh	catanh
`cos()`	`cos`	`ccos`
`sin()`	`sin`	`csin`
`tan()`	`tan`	`ctan`
`cosh()`	`cosh`	`ccosh`
`sinh()`	`sinh`	`csinh`
`tanh()`	`tanh`	`ctanh`
`exp()`	`exp`	`cexp`
`pow()`	`pow`	`cpow`
`sqrt()`	`sqrt`	`csqrt`
`fabs()`	`fabs`	`cabs`
`log()`	`log`	`clog`
`atan2()`	`atan2`
`cbrt()`	`cbrt`
`ceil()`	`ceil`
`copysign()`	`copysign`
`erf()`	`erf`
`erfc()`	`erfc`
`exp2()`	`exp2`
`expm1()`	`expml`
`fdim()`	`fdim`
`floor()`	`floor`
`fma()`	`fma`
`fmax()`	`fmax`
`fmin()`	`fmin`
`fmod()`	`fmod`
`frexp()`	`frexp`
`hypot()`	`hypot`
`ilogb()`	`ilogb`
`ldexp()`	`ldexp`
`lgamma()`	`lgamma`
`llrint()`	`llrint`
`llround()`	`llround`
`log10()`	`log10`
`log1p()`	`log1p`
`log2()`	`log2`
`logb()`	`logb`
`lround()`	`lround`
`lrint()`	`lrint`
`nearbyint()`	`nearbyint`
`nextafter()`	`nextafter`
`nexttoward()`	`nexttoward`
`remainder()`	`remainder`
`remquo()`	`remquo`
`rint()`	`rint`
`round()`	`round`
`scalbn()`	`scalbn`
`scalbln()`	`scalbln`
`tgamma()`	`tgamma`
`trunc()`	`trunc`
`carg()`		`carg`
`cimag()`		`cimag`
`conj()`		`conj`
`cproj()`		`cproj`
`creal()`		`creal`