Re: [PATCH] PR 16917: Add DFLOAT to gfortran

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Wed, Aug 11, 2004 at 05:16:17PM +0100, Paul Brook wrote:
> 
> Steven B pointed out that it would be simpler to just add dfloat as an alias 
> for dble. This means we accept a wider range of arguments to dfloat, but I 
> don't imagine anyone's going to be bothered about that.
> 

(snip)

>  
> +  make_alias ("dfloat");
> +
>    make_generic ("dble", GFC_ISYM_DBLE);
>  

I originally did the above patch, but changed my mind
because you are technically no longer compatible with
g77, f2c, and f77 on HP-UX.  There may be other vendor
compilers that support DFLOAT.  In all cases, the input 
to DFLOAT is restricted to a scalar integer argument.
Thus, the following will fail with g77 et al

      program d
      real :: x(3) = 1.2e0
      print *, dfloat(x)
      end

but gfortran with your patch will now compile the above.

Additionally, I plan to add a warning in gfc_check_dfloat
if -std is set.  I, however, need to determine how
gfc_notify_std () works.  I think we should issue something
like "Extension:  DFLOAT() is a nonstandard intrinsic.  Consider
using DBLE()."  'make_alias ("dfloat");' does not permit this
type of checking.

-- 
Steve