New port: CRX

[Hans-Peter Nilsson]

On Thu, 8 Jul 2004, Zack Weinberg wrote:
> Daniel Jacobowitz <drow@false.org> writes:
>
> > On Wed, Jul 07, 2004 at 06:48:58PM +0100, Nick Clifton wrote:
> >>   * Support for the VTABLE_INHERIT and VTABLE_ENTRY relocs used by G++.
> >>  (Lack of this support is causing 2 unexpected failures in the GAS
> >> testsuite).
> >
> > Is this still true?  I don't think g++ 3.x uses these relocations any
> > more.
>
> Yup, it doesn't.  Didn't do anything useful with the new ABI, if I
> recall correctly.

It (-fvtable-gc semantics) wasn't implemented and (prove me
wrong) can probably not be implemented in general, but maybe
it's possible for special cases like a chain of single
inheritance.  The option -fvtable-gc has been dropped from GCC,
and the pseudos are not emitted by GCC in CVS (unless that
changed in the last month, but probably not judging from Zack's
comment).

I suggest new binutils ports should not (have to) implement
support for those relocs (and support in old ports be removed
some time).

brgds, H-P