This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [Patch, Fortran, OOP] PR 46971: [4.6 Regression] ICE on long class names
Il giorno gio, 30/12/2010 alle 14.47 -0800, Steve Kargl ha scritto:
>
> Just out of curosity, why does a make utility need to
> know the name mangling scheme of a runtime library?
>
First, I had misunderstood that the user-defined subroutine names were
affected.
Second, even in that case I think cmake is overdoing it. Most people out
there are still using canned Fortran libraries which go back years, and
they use the old tricks needed to call F77 from C, instead of writing
wrappers with ISO_C_BINDING; if you are writing code with modules, you
might as well write a proper wrapper (which ForTrilinos *is* doing), but
not everyone does it apparently. But I don't control cmake (or
ForTrilinos for that matter)
Salvatore
> gfortran deliberately uses _gfortran_ as a prefix for
> most routines. A Fortran standard conforming application
> cannot directly call these routines, and no user written
> application (in some other language) should directly call
> those routines. And yes, I am aware of projects that grub
> around inside libgfortran.
>
>