Re: new concept checks and the 3.0 ABI

[Gabriel Dos Reis <Gabriel dot Dos-Reis at cmla dot ens-cachan dot fr>]

Joe Buck <jbuck@racerx.synopsys.com> writes:

| > What we need to tell users is what exactly is meant by "C++ ABI won't
| > change".  As discussed, it is the fact that what is described in
| > 
| > 	http://www.codesourcery.com/cxx-abi/
| > 
| > will be implemented in GCC/g++ and that interface won't change.
| > 
| > I don't recall any discussion to the effect that libstdc++ ABI was
| > meant.  There are so many things not in their final shape that I can't
| > even imagine that the whole libstdc++ was implied.
| 
| Sigh.  This should not be news.

Then it is, because libstdc++ developpers were never told that the 
"C++ ABI" issue was about libstdc++ and not what has been implemented
as the "new C++ ABI" or "C++ ABI v3".

| Gaby, we were talking about this way back when we were negotiating
| the egcs/gcc remerge with Richard Stallman two years ago.  The raison
| d'etre of 3.0, the reason for the bump in the major version number, was
| that it was going to have a stable C++ ABI.  Not what the CodeSourcery
| document was promising (after all, this was before Mark signed up to be
| RM) -- the whole thing, with all that it implies.

Please note that not just because the document appears to be at
CodeSourcery web site means it is a "CodeSourcery document".  It used
to be hosted on another web site.

| Note that http://www.codesourcery.com/cxx-abi/ talks about the multivendor
| ABI -- we never signed up, nor should we sign up, to having our library
| interoperable with other vendors' libraries.

If you don't want GCC/g++ comply to the multivendor ABI, that is not an
issue I'll dispute with you.  It *is* the fact that ABI is implemented
in the compiler.

| Now, it may be that this extremely ambitious goal is simply not
| achievable, but we cannot honestly pretend that it was not the goal.  But
| if (or rather, when) we back off from this goal, then I think we need to
| see if a subset is achievable (e.g. iostreams)

IOstreams are not even in their final shape.  We still need away to
have an implementation not built on top of stdio.h.

-- Gaby