[gfortran] Patch for PR 14901 Internal writes do not pad with blanks

[Paul Brook]

> Index:
> gcc/gcc/testsuite/gfortran.fortran-torture/execute/internal_write.f90
> =================================================================== RCS
> file: gcc/gcc/testsuite/gfortran.fortran-torture/execute/internal_write.f90
> diff -N
> gcc/gcc/testsuite/gfortran.fortran-torture/execute/internal_write.f90 ***
> /dev/null   1 Jan 1970 00:00:00 -0000
> --- gcc/gcc/testsuite/gfortran.fortran-torture/execute/internal_write.f90  
>     10 Apr 2004 02:18:54 -0000 ***************
> *** 0 ****
> --- 1,11 ----
> + ! PR 14901
> + ! Internal writes were appending CR after the last char
> + ! written by the format statement.
> +        CHARACTER*10 A
> +        WRITE(A,'(3HGCC)')
> +        IF (A.NE.'GCC       ') THEN
> +           PRINT*,'A was not filled correctly by internal write'
> +           PRINT*,' A = ',A

This sort of debugging information in the testsuite is IMO entirely redundant, 
and just slows the testsuite down.

If the purpose or previously observed failure mode of a test is not obvious it 
should be documented with source comments, not with verbose test output.

> +           CALL ABORT
> +        ENDIF
> +        END
> Index: gcc/libgfortran/ChangeLog
> ===================================================================
> RCS file: /cvs/gcc/gcc/libgfortran/Attic/ChangeLog,v
> retrieving revision 1.1.2.41
> diff -c -3 -p -r1.1.2.41 ChangeLog
> *** gcc/libgfortran/ChangeLog   4 Apr 2004 03:06:56 -0000       1.1.2.41
> --- gcc/libgfortran/ChangeLog   10 Apr 2004 02:18:55 -0000
> ***************
> *** 1,3 ****
> --- 1,9 ----
> + 2004-04-09  Bud Davis  <bdavis9659@comcast.net>
> +
> +       PR gfortran/14901
> +       * io/transfer.c (next_record_w) : No '\n' if internal.
> +       * io/unix.c (empty_internal_buffer) : Init to spaces, not '\n'.

Ok.

Paul