C++ ABI: Proposed key function tweak

[Jason Merrill]

>>>>> "Mark" == Mark Mitchell <mark at codesourcery.com> writes:

>>>>> "Jason" == Jason Merrill <jason at redhat.com> writes:
Jason> Since a pure virtual destructor must still be defined, we
Jason> can use it as the key function even though it's pure.
Jason> Making this change will allow users to avoid unnecessary
Jason> duplicate vtables for abstract classes.

> I think that's a sensible change -- but I think we're past the point
> where we can make these kind of changes.

> HP's compiler is in its final stabilization phase, and so is Intel's.

And so is ours.  But this is a one-line change to the compiler; I don't
think it's unreasonable to make it at this point.

Have folks at HP and Intel actually implemented the full ABI?  If so, I'm
surprised we haven't heard about issues that came up during implementation.

> I don't think we should be doing anything but clarifying at this
> point.

Perhaps.  But this seems like a clarification to me.

Other opinions?

Jason