This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH INSTALLED]: const typedefs part 19/N


On 8/27/07, Kaveh R. GHAZI <ghazi@caip.rutgers.edu> wrote:
> On Mon, 27 Aug 2007, Richard Guenther wrote:
> > That said, I vote for forcing stage1 to be build with a C++ compiler for next
> > stage1.
> > Richard.
>
> Saying "forcing stage1 to be build with a C++ compiler" could be
> interpreted two ways:
>
> It could mean making it *possible* to build with a C++ compiler, i.e. use
> -Wc++-compat for the whole compiler but still allow using plain C to
> bootstrap.  (Has anyone tried this recently?  Gaby?)
>
> Or using a more radical interpretation, it could mean *only* allow
> building stage1 with C++, i.e. using C++ features in the compiler that
> make it impossible to build stage1 with a plain C compiler.
>
> Which did you vote for?

I did vote for requiring a host C++ compiler and build stage1 with it (that is,
libbackend and the C frontend).  stage2/3 would be continued to be built
with a C compiler.  The effect would be that libbackend and the C frontend
would be buildable with both C or C++.  The benefits are 1) we don't regress
in this ability, 2) it's easy to experiment using some C++ features if you
just disable bootstrap.  Especially I would like to have 2) as at the moment
the benefits of using C++ are only theoretical and it is hard to try out some
things as building with C++ breaks all the time.

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]