Non-static data members

< cpp‎ | language

Non-static data members are declared in a member specification of a class.

class S
    int n;                // non-static data member
    int& r;               // non-static data member of reference type
    int a[10] = {1, 2};   // non-static data member with initializer (C++11)
    std::string s, *ps;   // two non-static data members
    struct NestedS {
        std::string s;
    } d5, *d6;            // two non-static data members of nested type
    char bit : 2;         // two-bit bitfield

Any simple declarations are allowed, except

  • extern and register storage class specifiers are not allowed;
  • thread_local storage class specifier is not allowed (but it is allowed for static data members);
  • incomplete types are not allowed: in particular, a class C cannot have a non-static data member of type C, although it can have a non-static data member of type C& (reference to C) or C* (pointer to C);
  • a non-static data member cannot have the same name as the name of the class if at least one user-declared constructor is present;
  • the auto specifier cannot be used in a non-static data member declaration (although it is allowed for static data members that are initialized in the class definition).

In addition, bit field declarations are allowed.


[edit] Layout

When an object of some class C is created, each non-static data member of non-reference type is allocated in some part of the object representation of C. Whether reference members occupy any storage is implementation-defined, but their storage duration is the same as that of the object in which they are members.

For non-union class types, members with the same member access are always allocated so that the members declared later have higher addresses within a class object. Members with different access control are allocated in unspecified order (the compiler may group them together). Alignment requirements may necessitate padding between members, or after the last member of a class.

[edit] Standard layout

A class where all non-static data members have the same access control and certain other conditions are satisfied is known as standard layout type (see StandardLayoutType for the list of requirements).

Two standard-layout non-union class types may have a common initial sequence of non-static data members and bit-fields (since C++14), for a sequence of one or more initial members (in order of declaration), if the members have layout-compatible types and if they are bit-fields, they have the same width (since C++14).

struct A { int a; char b; };
struct B { const int b1; volatile char b2; }; 
// A and B's common initial sequence is A.a, A.b and B.b1, B.b2
struct C { int c; unsigned : 0; char b; };
// A and C's common initial sequence is A.a and C.c
struct D { int d; char b : 4; };
// A and D's common initial sequence is A.a and D.d
struct E { unsigned int e; char b; };
// A and E's common initial sequence is empty

Two standard-layout non-union class types are called layout-compatible if they are the same type ignoring cv-qualifiers, if any (since C++14), are layout-compatible enumerations, or if their common initial sequence consists of every non-static data member and bit field (since C++14) (in the example above, A and B are layout-compatible)

Two standard-layout unions are called layout-compatible if they have the same number of non-static data members and corresponding non-static data members (in any order) have layout-compatible types.

Standard layout types have the following special properties:

  • If a standard-layout union holds two (or more) standard-layout structs as members, and these structs have a common initial sequence of data members, it is well-defined to examine any member of that common initial sequence regardless of which member of the union is active.
(until C++14)
  • In a standard-layout union with an active member of non-union class type T1, it is permitted to read a non-static data member m of another union member of non-union class type T2 provided m is part of the common initial sequence of T1 and T2 (except that reading a volatile member through non-volatile glvalue is undefined).
(since C++14)
  • A pointer to an object of standard-layout struct type can be reinterpret_cast to pointer to its first non-static data member (if it has non-static data members) or otherwise its first base class subobject (if it has any), and vice versa. (padding is not allowed before the first data member). Note that strict aliasing rules still apply to the result of such cast.
  • The macro offsetof may be used to determine the offset of any member from the beginning of a standard-layout struct.

[edit] Member initialization

Non-static data members may be initialized in one of two ways:

1) In the member initializer list of the constructor.
struct S
    int n;
    std::string s;
    S() : n(7) // direct-initializes n, default-initializes s
    { }
2) Through a default member initializer, which is simply a brace or equals initializer included in the member declaration, which is used if the member is omitted in the member initializer list
struct S
    int n = 7;
    std::string s{'a', 'b', 'c'};
    S() // copy-initializes n, list-initializes s
    { }

If a member has a default member initializer and also appears in the member initialization list in a constructor, the default member initializer is ignored.

#include <iostream>
int x = 0;
struct S
    int n = ++x;
    S() { }                 // uses default member initializer
    S(int arg) : n(arg) { } // uses member initializer list
int main()
    std::cout << x << '\n'; // prints 0
    S s1;
    std::cout << x << '\n'; // prints 1 (default initializer ran)
    S s2(7);
    std::cout << x << '\n'; // prints 1 (default initializer did not run)



Default member initializers are not allowed for bit field members.

Members of array type cannot deduce their size from member initializers:

struct X {
   int a[] = {1,2,3}; // error
   int b[3] = {1,2,3}; // OK

Default member initializers are not allowed to cause the implicit definition of a defaulted default constructor for the enclosing class or the exception specification of that constructor:

struct node {
    node* p = new node; // error: use of implicit or defaulted node::node() 
(since C++11)

Reference members cannot be bound to temporaries in a default member initializer (note; same rule exists for member initializer lists)

struct A
    A() = default;          // OK
    A(int v) : v(v) { }     // OK
    const int& v = 42;      // OK
A a1;    // error: ill-formed binding of temporary to reference
A a2(1); // OK (default member initializer ignored because v appears in a constructor)
         // however a2.v is a dangling reference

Reference members cannot be initialized with a default member initializer if it has a subexpression that would execute aggregate initialization which would use the same initializer:

struct A;
extern A a;
struct A
    const A& a1{ A{a, a} }; // OK
    const A& a2{ A{} };     // error
A a{a, a};                  // OK
(since C++14)

[edit] Usage

The name of a non-static data member or a non-static member function can only appear in the following three situations:

1) As a part of class member access expression, in which the class either has this member or is derived from a class that has this member, including the implicit this-> member access expressions that appear when a non-static member name is used in any of the contexts where this is allowed (inside member function bodies, in member initializer lists, in the in-class default member initializers).
struct S
    int m;
    int n;
    int x = m;            // OK: implicit this-> allowed in default initializers (C++11)
    S(int i) : m(i), n(m) // OK: implicit this-> allowed in member initializer lists
        this->f();        // explicit member access expression
        f();              // implicit this-> allowed in member function bodies
    void f();
2) To form a pointer to non-static member.
struct S
   int m;
   void f();
int S::*p = &S::m;       // OK: use of m to make a pointer to member
void (S::*fp)() = &S::f; // OK: use of f to make a pointer to member
3) (for data members only, not member functions) When used in unevaluated operands.
struct S
   int m;
   static const std::size_t sz = sizeof m; // OK: m in unevaluated operand
std::size_t j = sizeof(S::m + 42); // OK: even though there is no "this" object for m
(since C++03)

[edit] Defect reports

The following behavior-changing defect reports were applied retroactively to previously published C++ standards.

DR Applied to Behavior as published Correct behavior
CWG 613 C++03 unevaluated uses of non-static data members not allowed such uses are allowed
CWG 1696 C++14 reference members could be initialized to temporaries (whose lifetime would end at the end of ctor) such init is ill-formed
CWG 1397 C++11 class was regarded as complete in the default member initializers default member init cannot trigger definition of default ctor
CWG 1719 C++14 differently cv-qualified types weren't layout-compatible cv-quals ignored, spec improved

[edit] See also