Re: writing a double precission varibale with format f3.0 is not rounded

[Tim Prince <tprince at myrealbox dot com>]

Hugo Mahr wrote:
> Hello,
> I realized that there is a rounding probelm:
>
> Code:
>   -> cat f0999.f
>        program f0999
>        double precision rdum
>        rdum = 9.99999D-1
>        write(*,'(F3.0)') rdum
>        write(*,'(d24.6)') rdum
>        end
>
> compilation:
>   -> gfortran f0999.f
>
> execution:
>   -> ./a.out
>  0.
>             0.999999D+00
> I expect 1 as outcome instead of 0.999999d0.
>
> With f77 that is working as expected - and this
> is also according to the book fortran 95/2003 explained.
> Can you please help me to clarify this.
>
>
> fortran --version
> GNU Fortran 95 (GCC) 4.1.0 (SUSE Linux)
>
>   
If you want to discuss potential bugs with compilers other than 
gfortran, that is off topic.  What does your gfortran do?  Is it a 
current version?  In Metcalf/Reid/Cohen, I see one remark about 
rounding, but not enough to convince me that your example is taken from 
their book.  I have no clue about which compiler you mean by "F77."

Here is the echo obtained with my gfortran:
$ ./a
1.
           0.999999D+00