Date: Fri, 21 Nov 2025 23:08:20 +0000
On Friday, November 21st, 2025 at 10:05 AM, Frederick Virchanza Gotham via Std-Proposals <std-proposals_at_[hidden]> wrote:
>
> https://quuxplusone.github.io/blog/2022/12/01/chimeric-ptr/
>
> Now . . . three years on . . . I pay homage to my absorbed twin . . .
> I've written 'std::chimeric_ptr' into the GNU g++ compiler. Here's my
> compiler patch:
Looks like intersection type to me, retrofitted
to subtyping polymorphism based on pointers.
Java has it in the language, where you can
declare a type refinement like this
<T extends Colorful & Text>
And I think being a library is the part making
the post less cool, because chimeric_ptr would
waive away all the possible lifetime management
and type computation behind those types.
Why can't std::unique_ptr<Colorful & Text> and
std::polymorphic<Colorful & Text> "work?"
>
> https://quuxplusone.github.io/blog/2022/12/01/chimeric-ptr/
>
> Now . . . three years on . . . I pay homage to my absorbed twin . . .
> I've written 'std::chimeric_ptr' into the GNU g++ compiler. Here's my
> compiler patch:
Looks like intersection type to me, retrofitted
to subtyping polymorphism based on pointers.
Java has it in the language, where you can
declare a type refinement like this
<T extends Colorful & Text>
And I think being a library is the part making
the post less cool, because chimeric_ptr would
waive away all the possible lifetime management
and type computation behind those types.
Why can't std::unique_ptr<Colorful & Text> and
std::polymorphic<Colorful & Text> "work?"
-- Zhihao Yuan, ID lichray The best way to predict the future is to invent it. _______________________________________________
Received on 2025-11-21 23:08:31
