On Thursday, October 9, 2025, Sebastian Wittmeier via Std-Proposals <
std-proposals_at_[hidden]> wrote:

> What is the difference to static_cast<unsigned int>(size())
>
> (or some suitably longer <size_t> type)
>
> Or you want to definitely have functions differentiated (overloaded) by
> return type?
>



I was purely going for the ergonomics of it.

You're right though that it would sort of be like allowing function
overloading by return type, but we could stipulate that the unsigned and
signed forms of the function shall have the same address in memory (because
the machine code will be the same anyway). So then the trigraph, +|- ,
would really do nothing more than convert a function pointer to another
function pointer type.

What I mean, is that if you had the following function pointer:

    uint64_t (*p)(void)

and if you were to place +|- before it, then the following expression:

    +|- p

would be the same as:

    ( (int64_t (*)(void)) p)

So you're not really providing an alternative implementation in the true
sense of "function overloading", but more so you're just casting a function
pointer before you invoke it. This will work fine on every CPU instruction
set I can think of . . . arm 32 & 64, x86 32 & 64, hppa, motorolla,
PowerPC, RISC, Sparc.

Again it's just a half-baked idea.