strict ansi C++ with isnan() et al

[Jeff Johnston]

While I question why the C++ cmath header file introduces the loop in 
the first place, I have no problem with the patch and I have checked it 
in (reversed btw with a ChangeLog entry :)).

-- Jeff J.

Lev Bishop wrote:
> An issue was reported on cygwin here:
> http://cygwin.com/ml/cygwin/2006-11/msg00786.html
> 
> I know little about how system headers are supposed to work, but would
> this be an appropriate patch? It fixes this specific problem, for me.
> Or is the root problem elsewhere than newlib?
> 
> --- /usr/include/math.h    2006-11-30 17:45:31.875000000 -0500
> +++ /usr/include/math.h.old    2006-11-30 17:42:21.468750000 -0500
> @@ -102,7 +102,7 @@
> #endif /* ! defined (__math_68881) */
> #endif /* ! defined (_REENT_ONLY) */
> 
> -#if !defined(__STRICT_ANSI__) || defined(__cplusplus)
> +#ifndef __STRICT_ANSI__
> 
> /* ISO C99 types and macros. */
> 
> @@ -233,7 +233,7 @@
> 
> #endif /* ! defined (_REENT_ONLY) */
> 
> -#endif /* ! defined (__STRICT_ANSI__) || defined (__cplusplus) */
> +#endif /* ! defined (__STRICT_ANSI__) */
> 
> #if !defined(__STRICT_ANSI__) || defined(__cplusplus)
>