[PATCH] Fix the -d8 options in gfortran

[Andrew Pinski]

On Mar 9, 2005, at 10:38 AM, Steve Kargl wrote:

> On Wed, Mar 09, 2005 at 03:06:30PM +0000, Paul Brook wrote:
>>>> http://gcc.gnu.org/ml/fortran/2005-01/msg00137.html
>>>
>>> Doesn't seem relevant.  Toon talks about documenting
>>> the broken behaviour.  I'm actually trying to fix the
>>> problem.
>>
>> I meant the second part of the email. ie. -d8 should set
>> "DOUBLE PRECISION" to  REAL*8, not be an alias for -i8 -r8.
>>
>> Paul
>
> I've implemented 1. from msg00137.html.  Toon prefers 2.
> (presumably for a Cray).  If you actually read the code,
> all comments suggest -d8 is equivalent to -i8 and -r8.
>
> Does gfortran even work on an OS with 128-bit floating
> point?

Yes AIX and ppc-darwin8.0 both which have 128bit fp (not the normal
IEEE format though).

-- Pinski