On Wednesday, 26 November 2025 15:33:42 Pacific Standard Time Frederick
Virchanza Gotham via Std-Proposals wrote:
> I mean I think my GodBolt link
> tells ya nearly everything you need to know -- you can edit the code a
> little, see what it does, add another base class and come back to me
> with "I notice it does X, but shouldn't it do Y instead?" -- that
> would open up discussion.

Your Godbolt links prove something *can* be done (whether it inherently has to
go through UB is different story). We're not asking for that.

We're asking for an explanation of *why* something must be done, why it must
be done this way and not in other ways, why it must be in the Standard, etc.
It should include an analysis of future compatibility and cost too.

> I'm preparing a paper on "std::chimeric_ptr" but I want more
> discussion on it first -- particularly I want to disassemble Thiago's
> argument that it's only use is to compensate for bad programming.

Which wouldn't have been necessary if you had prepared a paper with the use-
cases (who benefits, why they need it, etc.).

-- 
Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
  Principal Engineer - Intel Data Center - Platform & Sys. Eng.