Re: [Patch, fortran] PR28585: Add Fortran 2003 NEW_LINE intrinsic function

[Brooks Moses <bmoses at stanford dot edu>]

FX Coudert wrote:
> > I also need to go read the F2003 standard to see
> > what this intrinsic does because its name suggests that gfortran
> > can constant-fold the intrinsics away.
>
> I'm a bit worried that not all systems supported by gfortran use \n  
> as newline. E.g., i386-pc-mingw32 uses \n\r.

This is irrelevant.  The Fortran 2003 Standard states that "If [the 
argument] is of the default character type and the character in position 
10 of the ASCII collating sequence is representable in the default 
character set, then the result is ACHAR(10).  ...  If [the argument] is 
of the ASCII character type or the ISO 10646 character type, then the 
result is CHAR(10,KIND(A))."

Thus, the appropriate result is \n, regardless of what the system uses 
as a newline in files.

Unfortunately, this _does_ mean that in a standard-compliant 
implementation, the following commands will produce different files on a 
system that uses \n\r as a newline -- but that's an issue with the 
standard, not with the implementation:

    WRITE(*,'A') this // NEW_LINE(this) // that
    WRITE(*,'A') this, that

- Brooks