patch for -Wno-long-long and early GNAT compilers

[Alexandre Oliva]

On Mar 19, 2002, dewar@gnat.com (Robert Dewar) wrote:

> The recommended procedure is to use the pair of binaries that we
> package, that have been verified to work together, and in particular
> to successfully bootstrap.

I can see why ACT would not want to support GNAT builds using
arbitrary bootstrap compilers, just like Red Hat-supplied GNU Pro
toolchains are only supported when built by a particular version of
GNU Pro native tools.  However, we're not talking about ACT here.
We're talking about the GNU Ada compiler, that is part of the GNU
Compiler Collection.  GCC has always been quite conservative in
requirements imposed to the bootstrap compiler, which makes it
possible for us to build GCC starting from lots of different
compilers.  Of course, sometimes you'll run into problems, if you use
a sufficiently broken compiler, but odds are that it will just work.

Now, I get the impression that Ada has come into GCC with a totally
different mindset.  It appears that, instead of even making attempts
to widen the set of bootstrap compilers that could be used, the
maintainers of the GNU Ada Compiler seem to imply that the only way
they care about is a particular version of GNAT released by ACT.
Contrast this with the position that has typically been taken in GCC,
of often attempting to cope with bootstrap compilers that don't have
problems that are too severe.  I, for one, have been bootstrapping
weekly GCC snapshots with old and relatively broken compilers on a
number of platforms, for example, on sparc-solaris machines using a
relatively broken SPARCompiler released about a decade ago.  Using
such bootstrap compilers has actually helped revealing portability
bugs in GCC, and fixing them made GCC's code more robust.

I can understand ACT's position of recommending its own builds for
bootstrapping, and I can see how it may be justifiable with the
scarcity of human resources available to GNAT working with multiple
bootstrap compilers.

But from that to claiming that using any other compiler is asking for
trouble, there's a long way.  `Be lenient in what you accept and
strict in what you generate.' -now, who wrote this, and where?  Was it
in the sendmail book that I read this?

I fear that this position may lead us to reject patches for GNAT from
random contributors just because they're not necessary for builds with
the recommended version.

I'd much rather treat changes that cause releases of GNAT that used to
be usable for bootstrapping to stop working as regressions that should
be addressed in a timely manner, instead of promoting support for the
use of a single bootstrap compiler or actively breaking builds that
fail to use such compiler without a very good reason to do so.  I
don't mean to imply that we have the latter now, but I fear the
current trend may get us there, and I don't want us to get there.

That said, my interest in Ada is extremely small, and I only have it
on my own build machines because I have the disk space for it.  If it
stops building Ada at some point because configure is adjusted so as
to start requiring a newer version of GNAT that the one I have, I
probably won't notice.  My comments are just meant as food for thought
and, while at that, probably not worth more than $0.02 :-)

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist                Professional serial bug killer