Also as a bonus, once compilers start supporting "Deducing this", the
pointers to deducing this member functions are free function pointers.


On Fri, Jan 5, 2024 at 11:21 AM Jarrad Waterloo <descender76_at_[hidden]>
wrote:

> That example will scare people away for 2 reasons. Besides the one
> presented, there is the awkward ergometric of having to specify the
> instance TWICE in memfunc_to_func and invoke_func_as_memfunc which then
> leads to the presented problem.
>
> At compile time for KNOWN instance `a` you can call both Base::Go and
> Derived1::Go.
> At compile time for KNOWN instance `b` you can call both Base::Go and
> Derived2::Go.
>
> function_ref allows you to do these same 4 tasks and you don't have to
> make Base, Derived1 or Derived2 virtual at all.
>
> function_ref, Dyno and the others are really a better form of polymorphism
> that you ?should? defer to 85/90/95/99% of the time.
>
>
> https://youtube.com/clip/Ugkxo9iqdK7DOY8JCS6hvw9x-P5FqO849zSQ?si=J85Eh-aLJdpgBOqX
>
>
>
>
> On Fri, Jan 5, 2024 at 10:54 AM Frederick Virchanza Gotham via
> Std-Proposals <std-proposals_at_[hidden]> wrote:
>
>>
>> On Friday, January 5, 2024, Frederick Virchanza Gotham wrote:
>>
>>>
>>> The above code will invoke the 'Derived2::Go' member function on an
>>> object of type 'Derived1'.
>>>
>>
>>
>> Typographical error:
>> Swap 'Derived2' with 'Derived1'.
>> --
>> Std-Proposals mailing list
>> Std-Proposals_at_[hidden]
>> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals
>>
>