Invalid macro names in tests

[Jonathan Wakely]

On Sat, 14 Dec 2024 at 13:16, François Dumont <frs.dumont@gmail.com> wrote:
>
> Ok, I see, I've adapted _Rb_tree the same way.
>
> Just a small remark on this patch. As the new code is only used in C++11
> and later several std::__addressof could be std::addressof, no ?

Yes but that makes an additional function call.

Maybe we should make std::addressof use __builtin_addressof directly,
to avoid that. Or just make them both always_inline.


>
> François
>
> On 11/12/2024 20:36, Jonathan Wakely wrote:
> > On Wed, 11 Dec 2024 at 18:47, Jonathan Wakely <jwakely@redhat.com> wrote:
> >> On Wed, 11 Dec 2024 at 18:32, François Dumont <frs.dumont@gmail.com> wrote:
> >>> Hi
> >>>
> >>> I don't know if you will come to the same conclusion but I think
> >>> alloc_ptr_ignored.cc test cannot PASS, at best it could XFAIL.
> >> It passes for me with the patch I'm testing, attached. The relevant
> >> parts that were needed to make alloc_ptr_ignored.cc work are the
> >> changes to _List_base::_M_destroy_node and list::_M_create_node.
> > Ha, I made a last-minute change to us 'if constexpr' in _M_put_node
> > and broke the test again.
> >
> > The version attached here makes alloc_ptr_ignored.cc PASS again.
> >
> >
> >>> Unless when _GLIBCXX_USE_ALLOC_PTR_FOR_XXX is 0 you simply want to
> >>> ignore the _Alloc template parameter entirely and fallback on
> >>> std::allocator ?
> >>>
> >>> Otherwise you have problems with node_handle definition or with the
> >>> result of CustomPointerAlloc::allocate that cannot be converted to a _Node*.
> >> Then the container needs to be fixed :-)