[unique.ptr.general] Intro wording is misleadingly restrictive

https://eel.is/c++draft/unique.ptr.general#1 (and remaining) paragraphs make overly restrictive claims:

• a unique pointer owns an object
• a unique pointer stores a pointer

Neither of these claims are necessarily true. The following program is valid, and compiles with libc++:

#include <memory>

struct deleter {
    using pointer = int(*)();
    void operator()(pointer) const {}
};

int main() {
    std::unique_ptr<int&, deleter> ptr;
    ptr.get();
}

• What is the owned object here? Neither int& nor int() are objects types.
• While pointer is actually a pointer in this example, it is generally any pointer-like type that satisfies NullablePointer. It's not required to be a pointer.

I think it's all arguably not a big problem, and this is simply questionable wording in the semi-normative intro paragraphs. However, it's also not nice that the wording is making claims that get contradicted/superseded/corrected by later wording. It would be nice if this wasn't the case.

Start with the linked unique.ptr.general section and read the introductory and remaining paragraphs together, then compare their descriptions with the NullablePointer requirements and the example in the issue. Done means the introductory wording no longer makes claims contradicted by the later specification; the issue does not name a repository file or test to run.