This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 1/4] Macroize inclusion of math-finite.h (part 1)
On Mon, 13 Mar 2017, Gabriel F. T. Gomes wrote:
> +/* Only declare float and long double versions of these functions if
> + __USE_ISOC99 is defined. Other restrictions may apply for each
> + specific function and are dealt with individually. */
> +#if __MATH_DECLARING_DOUBLE || defined __USE_ISOC99
> +# define __MATH_REDIRCALL_X(from, args, to) \
> + extern _Mdouble_ __REDIRECT_NTH (from, args, to)
> +# define __MATH_REDIRCALL(function, reentrant, args) \
> + __MATH_REDIRCALL_X \
> + (__REDIRFROM (function, reentrant, _MSUF_), args, \
> + __REDIRTO (function, reentrant, _MSUF_))
> +# define __MATH_REDIRCALL_2(from, reentrant, args, to) \
> + __MATH_REDIRCALL_X \
> + (__REDIRFROM (from, reentrant, _MSUF_), args, \
> + __REDIRTO (to, reentrant, _MSUF_))
> +#else
> +# define __MATH_REDIRCALL(function)
> #endif
My inclination is that here it would be cleaner to make the include of
math-finite.h from math.h for float and double conditional on
__USE_ISOC99, just as the includes of mathcalls.h are conditional.
(Then, for float128, the includes of mathcalls.h and math-finite.h would
be conditional on the relevant conditional for "declare float128
interfaces", but the include of mathcalls-helper-functions.h would be
conditional on "float128 is supported by glibc and ABI-distinct from long
double".)
--
Joseph S. Myers
joseph@codesourcery.com