Re: g95 as a system fortran compiler?

From: Anton Shterenlikht <mexas_at_bristol.ac.uk>
Date: Mon, 21 Dec 2009 10:58:26 +0000
On Mon, Dec 21, 2009 at 11:54:29AM +0100, Gerald Pfeifer wrote:
> On Sun, 20 Dec 2009, b. f. wrote:
> > The decision to remove the Fortran compiler from the base system was
> > made long ago, and will probably not be reversed now.  There is no
> > Fortran code in the base system, and Fortran is needed only by a
> > minority of users.
> 
> With even popular scripting languages residing in ports, not the
> base system, it really would be hard to argue for Fortran being
> part of the system compiler.  That, and FreeBSD has been notorious
> for letting the system compiler and toolchain rot (sorry, it's hard
> to find a friendlier term), so users would not be served too well
> by such a move, especially seeing how actively and quickly GCC
> Fortran develops in contrast.  This level of agility is really
> where our Ports Collection shines.

yes, I get it now, it makes sense

> > Yes, it's unfortunate that the gcc maintainers discontinued support
> > for a number of architectures.  But maybe someone will step forward
> > and fix it?  Or llvm?  In the meantime, why don't you ask gerald_at_ to
> > make the default Fortran compiler on ia64 the latest version of
> > gfortran that will still work on that architecture?  You can do it
> > yourself by making some small local patches to ports/Mk/bsd.gcc.mk,
> > and to the relevant lang/gcc4X port, while you are waiting for him ...
> 
> Anton has been working with me and really has been trying to get 
> (upstream) attention.  With FreeBSD being a niche OS and Itanium
> going the way of the Alpha and the Dodo, this is not a healthy 
> intersection, sadly, and nobody has stepped up yet to fix the
> issue for real though some advice has been given. :-(
> 
> ( http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40959 )
> 
> Using different compiler versions on different architectures is
> going to make other ports maintainers pretty unhappy campers and
> would not see a lot of testing, so I would not recommend going
> down that route, nor would I want to make ports more complicated
> when that only benefits one or two users globally.

I agree


-- 
Anton Shterenlikht
Room 2.6, Queen's Building
Mech Eng Dept
Bristol University
University Walk, Bristol BS8 1TR, UK
Tel: +44 (0)117 331 5944
Fax: +44 (0)117 929 4423
Received on Mon Dec 21 2009 - 09:58:33 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:59 UTC