Proposed resolution:

This wording is relative to N4988.

1. Modify [meta.type.synop], header <type_traits> synopsis, as indicated:

   […]
   // [meta.const.eval], constant evaluation context
   constexpr bool is_constant_evaluated() noexcept;
   template<class T>
   consteval bool is_within_lifetime(const Tauto*) noexcept;
   

2. Modify [meta.const.eval] as indicated:

   template<class T>
   consteval bool is_within_lifetime(const Tauto* p) noexcept;
   

   -?- Mandates: is_object_v<T> is true.

   -3- Returns: true if p is a pointer to an object that is within its lifetime ([basic.life]); otherwise, false.

   -4- Remarks: During the evaluation of an expression E as a core constant expression, a call to this function is ill-formed unless p points to an object that is usable in constant expressions or whose complete object's lifetime began within E.

[ 2025-10-14; Reflector poll ]

Set priority to 3 after reflector poll.

"What about `void*`, should that be allowed? The compiler should know which union member the `void*` came from, as this is only using during constant evaluation."

"NAD, a function could have its lifetime constrained by a dynamic loader, so it seems wrong to assume that taling of function lifetimes makes no sense."

int f();
std::is_within_lifetime<int()>(f);

This is currently well-formed, and only fails when evaluated because is_within_lifetime is not constrained. However talking of lifetime of a non-object does not make sense, and the lack of constraint makes the implementation and use of that function and underlying built-in more convoluted than necessary.