Proposed resolution:

This wording is relative to N3936.

1. Introduce the following new definition to the existing list in [definitions]: [Drafting note: If LWG 2234 is accepted before this issue, the accepted wording for the new definition should be used instead — end drafting note]

   constant subexpression [defns.const.subexpr]

   an expression whose evaluation as a subexpression of a conditional-expression CE ([expr.cond]) would not prevent CE from being a core constant expression ([expr.const]).

2. Change header <memory> synopsis, [memory.syn] as indicated:

   namespace std {
     […]
     // [specialized.algorithms], specialized algorithms:
     template <class T> constexpr T* addressof(T& r) noexcept;
     […]
   }
   

3. Change [specialized.addressof] as indicated:

   template <class T> constexpr T* addressof(T& r) noexcept;
   

   -1- Returns: The actual address of the object or function referenced by r, even in the presence of an overloaded operator&.

   -?- Remarks: An expression std::addressof(E) is a constant subexpression ([defns.const.subexpr]), if E is an lvalue constant subexpression.

[ 2015-05, Lenexa ]

STL: the intent of this change is good; I think the wording is good
- I'm a bit worried about asking for a compiler hook
- if every implementer says: yes they can do it we should be good
EB: there is missing the word "a" before "subexpression" (in multiple places)
MC: the editor should do - we rely on our editors
MC: move to Review with a note stating that we wait for implementation experience first
- in favor: 13, opposed: 0, abstain: 2 HB: N4430 will bring something which is addressing this issue
MC: good we didn't go to ready then

[ 2014-06-09, further improvements ]

A new wording form is now used similar to the approach used by LWG 2234, which is a stricter way to impose the necessary implementation requirements.

[ 2014-06-08, Daniel improves wording ]

It has been ensured that the wording is in sync with the recent working paper and the usage of "any" has been improved to say "every" instead (the fix is similar to that applied by LWG 2150).

Previous resolution from Daniel [SUPERSEDED]:

1. Change header <memory> synopsis, [memory.syn] as indicated:

   namespace std {
     […]
     // [specialized.algorithms], specialized algorithms:
     template <class T> constexpr T* addressof(T& r) noexcept;
     […]
   }
   

2. Change [specialized.addressof] as indicated:

   template <class T> constexpr T* addressof(T& r) noexcept;
   

   -1- Returns: The actual address of the object or function referenced by r, even in the presence of an overloaded operator&.

   -?- Remarks: For every lvalue core constant expression e ([expr.const]), the expression std::addressof(e) is a core constant expression.

[ 2013-09 Chicago ]

I'm writing a function that needs to be constexpr and I wanted to take the address of its input. I was thinking of using std::addressof to be safe, but it isn't currently constexpr. A sample implementation couldn't be constexpr under the C++11 rules, though.

Daniel Krügler:

Indeed the core language clarified by CWG 1312 and by CWG 1384, that such emulations of std::addressof implementations are not valid in constant expressions, therefore it seems more like a defect than a feature request to ask for the guarantee that std::addressof is a constexpr function. It should be added that a similar requirement already exists for offsetof indirectly via the C99 standard as of 7.17 p3:

The macros are […]

offsetof(type, member-designator)

which expands to an integer constant expression that has type size_t […]

combined with the noted property in C++11 that:

"offsetof is required to work as specified even if unary operator& is overloaded for any of the types involved"

Therefore implementations should already be able without heroic efforts to realize this functionality by some intrinsic. The wording needs at least to ensure that for any lvalue core constant expression e the expression std::addressof(e) is a core constant expression.