bug report : gfortran E edit descriptor

[Dominique d'Humières]

> I have an trouble in E edit descriptor of gfortran.
> This is a example.
> (source file: test.f95)
>        program test
>        implicit none
>        real a,b
>        a=135.0
>        b=1737.5
>        write(*,*)a,b
>        write(*,'(e9.3,a,f7.1)')a,' ',b
>        write(*,'(1pe9.3,a,f7.1)')a,' ',b
>        end
> (compile)
> gfortran test.f95
> (execute)
> a.exe
> (result)
>    135.000000       1737.50000    
> 0.135E+03  1737.5
> 1.350E+02 17375.0  ( <--- wrong value )
>
> I suppose there are some bugs in the E edit descriptor.
IIRC 
(1) the P ‘modifier’ applies to all the subsequent numeric descriptors,
(2) when rP is applied to an F descriptor, the value is ‘multiplied’ by 10**r.
write(*,'(e9.3,a,1pf7.1)')a,' ‘,b
gives
0.135E+03 17375.0
and
write(*,'(1pe9.3,a,e9.3)')a,' ‘,b
gives
1.350E+02 1.738E+03
So I think gfortran is right.

> I like Fortran and I wish gfortran of GCC to be better compiler.
> best regards.
If you want to contribute, discussion about gfortran should go to fortran@gcc.gnu.org and bug should be filed in bugzilla.

Cheers,

Dominique