> The part about "Constant-checked preconditions" is orthogonal to the
> feature in question. Would it make sense to split it in its own paper?
> (I guess the only reason it's in *this* paper is because no-one has
> needed it before.)

Perhaps it would make sense. Perhaps it could just be done in the repo
or as an LWG issue. I've already opened a repo issue related to this
change here: https://github.com/cplusplus/draft/issues/7904