[fortran] Fix Fortran configure for aarch64

A while back I checked in a Fortran configure patch that allowed me to build
Fortran for the mips-mti-elf target.  I used the same mechanism (checking
with_newlib) that libstdc++ and libjava use.  This change broke the aarch64
Fortran build because that platform's newlib does not have strtold.  I would
like to change configure to not define HAVE_STRTOLD.  I could just remove that
define but instead I would again like to use the same code that the libstdc++
uses, by checking long_double_math_on_this_cpu.  This variable is actually
never set, and because of this HAVE_STRTOLD is never set but that allows us to
build Fortran on mips and aarch64.

Mips would not use strtold even if HAVE_STRTOLD was defined because mips does
not define REAL*10 and the REAL*16 code checks GFC_REAL_16_IS_FLOAT128
(set for mips) before checking for HAVE_STRTOLD.

Marcus Shawcroft said it works for him on aarch64 and I tested it on mips,
OK for checkin?

Steve Ellcey
sellcey@mips.com


2013-11-15  Steve Ellcey  <sellcey@mips.com>

	* configure.ac: Do not define HAVE_STRTOLD.
	* configure: Regenerate.

Hi Steve,

It certainly looks good to me and if it works on the intended targets,
it cannot be bad!  OK for trunk.

Thanks for the patch

Paul

On 15 November 2013 18:33, Steve Ellcey <sellcey@mips.com> wrote:
>
> A while back I checked in a Fortran configure patch that allowed me to build
> Fortran for the mips-mti-elf target.  I used the same mechanism (checking
> with_newlib) that libstdc++ and libjava use.  This change broke the aarch64
> Fortran build because that platform's newlib does not have strtold.  I would
> like to change configure to not define HAVE_STRTOLD.  I could just remove that
> define but instead I would again like to use the same code that the libstdc++
> uses, by checking long_double_math_on_this_cpu.  This variable is actually
> never set, and because of this HAVE_STRTOLD is never set but that allows us to
> build Fortran on mips and aarch64.
>
> Mips would not use strtold even if HAVE_STRTOLD was defined because mips does
> not define REAL*10 and the REAL*16 code checks GFC_REAL_16_IS_FLOAT128
> (set for mips) before checking for HAVE_STRTOLD.
>
> Marcus Shawcroft said it works for him on aarch64 and I tested it on mips,
> OK for checkin?
>
> Steve Ellcey
> sellcey@mips.com
>
>
> 2013-11-15  Steve Ellcey  <sellcey@mips.com>
>
>         * configure.ac: Do not define HAVE_STRTOLD.
>         * configure: Regenerate.
>
>
>
> diff --git a/libgfortran/configure.ac b/libgfortran/configure.ac
> index 6417373..6acc0f2 100644
> --- a/libgfortran/configure.ac
> +++ b/libgfortran/configure.ac
> @@ -267,12 +267,18 @@ if test "x${with_newlib}" = "xyes"; then
>     # link executables.
>     AC_DEFINE(HAVE_MKSTEMP, 1, [Define if you have mkstemp.])
>     AC_DEFINE(HAVE_STRTOF, 1, [Define if you have strtof.])
> -   AC_DEFINE(HAVE_STRTOLD, 1, [Define if you have strtold.])
>     AC_DEFINE(HAVE_SNPRINTF, 1, [Define if you have snprintf.])
>     AC_DEFINE(HAVE_STRCASESTR, 1, [Define if you have strcasestr.])
>     AC_DEFINE(HAVE_VSNPRINTF, 1, [Define if you have vsnprintf.])
>     AC_DEFINE(HAVE_LOCALTIME_R, 1, [Define if you have localtime_r.])
>     AC_DEFINE(HAVE_GMTIME_R, 1, [Define if you have gmtime_r.])
> +
> +   # At some point, we should differentiate between architectures
> +   # like x86, which have long double versions, and alpha/powerpc/etc.,
> +   # which don't. For the time being, punt.
> +   if test x"long_double_math_on_this_cpu" = x"yes"; then
> +     AC_DEFINE(HAVE_STRTOLD, 1, [Define if you have strtold.])
> +   fi
>  else
>     AC_CHECK_FUNCS_ONCE(getrusage times mkstemp strtof strtold snprintf \
>     ftruncate chsize chdir getlogin gethostname kill link symlink sleep ttyname \
>