[PATCH] Add nextup and nextdown math functions

[Joseph Myers]

On Fri, 20 May 2016, Rajalakshmi Srinivasaraghavan wrote:

> diff --git a/math/Makefile b/math/Makefile
> index 57c3ec1..03c54ed 100644
> --- a/math/Makefile
> +++ b/math/Makefile
> @@ -63,7 +63,7 @@ libm-calls = e_acos e_acosh e_asin e_atan2 e_atanh e_cosh e_exp e_fmod	\
>  	     s_fma s_lrint s_llrint s_lround s_llround e_exp10 w_log2	\
>  	     s_issignaling $(calls:s_%=m_%) x2y2m1 k_casinh	\
>  	     gamma_product k_standard lgamma_neg lgamma_product		\
> -	     w_lgamma_compat
> +	     w_lgamma_compat s_nextdown s_nextup

This is the standard way of adding a function to be built for each 
floating-point type.  It assumes that, when double and long double have 
the same ABI, the double version will define the long double name as an 
alias.

But I don't see anything in your double versions actually to define such 
aliases if NO_LONG_DOUBLE, so I'd expect you to be missing those functions 
for configurations without a long double format distinct from double (e.g. 
ARM).

-- 
Joseph S. Myers
joseph@codesourcery.com