Re: [G95] planned obsolescence?

[Steven Bosscher]

Op di 14-01-2003, om 20:16 schreef Steven G. Johnson:
---- 8< ----
> Wouldn't it be advisable to use a
> generic name like "gfc" instead of a version-dependent one like "g95?"

While I don't think the decision has been made yet, it seems that
everybody involved here feels that it would indeed be the Right Thing to
change the name of the compiler from g95 to gfc (that is, at least the
driver program should be called gfc).

> 
> This is not entirely an abstract question for me, since I have done a lot
> of the maintainance work on Fortran support in GNU Autoconf.  Recently,
> I've proposed a patch to add support for newer Fortran dialects
> simultaneously with F77:
> 
>       http://mail.gnu.org/archive/html/autoconf-patches/2002-11/msg00009.html
> 
> The basic viewpoint of this patch is that future Fortran support in
> Autoconf should not have a new set of macros for every revision of the
> standard; everything should be handled with $(FC), $(FCFLAGS), etcetera,
> just like e.g. C and C++, with the exception of F77 (which keeps a
> separate $F77 etc. for legacy code).
> 
> Since you guys are at the forefront of Fortran development, I'm curious to
> hear your perspective.

Suppose we do change the name, would this also mean that your patches
would be accepted to go into autoconf?

One of the things we're having trouble with all the time, is
autoconf/automake.  Right now, Paul Brook has to maintain is own set of
patches to autoconf[1] and automake to allow us to build our runtime
library  (which is partially written in Fortran 90).
This is obviously very inconvenient.  The sooner the support in the
official releases is there, the better[2].

It would certainly help if autoconf and automake would start supporting
$(FC).  The bulk of the work already exists, even for automake.  It's
just a matter of accepting it.

Greetz
Steven

[1] Available from http://sourceforge.net/projects/gcc-g95/
[2] Now if only GCC would move to a more recent autoconf soon ;-)