Hi Greg
The 2003 standard helps my F90 and F95 compilers ;)
What does the PathScale compiler do? I suspected it wasn't standard (I
would never write such an expression in my code, it was sent to me as a
bug in my fortran function parser/scripting language) but was quite amazed
to see that 1/2 the compilers I have access to demonstrated a problem.
My opteron box has been taken away, so I wasn't able to check the
PathScale compilers.
Stu.
--
<--------------------------------------------------------------------->
Dr Stuart Midgley | [log in to unmask]
Supercomputer Facility | [log in to unmask]
Leonard Huxley Building 56 | +61 (0)2 6125 5988 Work
Australian National University | +61 (0)2 6125 8199 Fax
CANBERRA ACT 0200 | +61 (0)4 1125 2488 Mob
> On Sat, Nov 20, 2004 at 01:57:06PM +1100, Stuart Midgley wrote:
>
>> program test
>> write(*,*) 1d-1/-2d-1
>> end program test
>
> Stuart,
>
> In the 2003 draft standard, table 7.7 shows that the precedence of / is
> higher than that of unary -. So this isn't a valid expression, you need
> ().
>
> -- greg
>
>
|