[RFC] C++1x breaking the ABI in one more place :(

[Gabriel Dos Reis]

On Thu, May 20, 2010 at 11:57 AM, Mark Mitchell <mark@codesourcery.com> wrote:

> I do think the earlier suggestion of "add a few spare vtable slots" is a
> good bit of future-proofing, in that it improves the odds that any given
> change will be ABI-compatible.

How many "spare vtable slots" is future-proof? 12? 24? 1024? 1G?

Also, I have to note that the LWG does not take ABI breaking changes lightly.
I reported a defect in the IO library with respect to the French locale.
A simple, easy to understand, straightforward solution would have
required addition of a new virtual function to the facets.  Rather, LWG
worked hard to come up with a very clever (albeit brittle in my view) solution
to avoid the addition of a new virtual function.

Personally, I believe that the risk posed by the original patch is being
inflated beyond the reasonable.  That does not deny the general more
abstract problem though.

-- Gaby