______________________________________________________________________

  13   Overloading                                      [over]

  ______________________________________________________________________

1 When two or more different declarations are  specified  for  a  single
  name in the same scope, that name is said to be overloaded.  By exten­
  sion, two declarations in the same scope that declare  the  same  name
  but  with  different  types  are called overloaded declarations.  Only
  function declarations can be overloaded; object and type  declarations
  cannot be overloaded.

2 When  an  overloaded  function name is used, which overloaded function
  declaration is being referenced is determined by comparing  the  types
  of  the arguments at the point of use with the types of the parameters
  in the overloaded declarations that are visible at the point  of  use.
  This  function  selection process is called overload resolution and is
  defined in _over.match_.  For example,
          double abs(double);
          int abs(int);
          abs(1);       // call abs(int);
          abs(1.0);     // call abs(double);

  13.1  Overloadable declarations                            [over.load]

1 Not all function declarations can be overloaded.  Those that cannot be
  overloaded are specified here.  A program is ill-formed if it contains
  two such non-overloadable declarations in the same scope.

2 Certain function declarations that cannot be distinguished by overload
  resolution cannot be overloaded:

  --Since  for any type T, a parameter of type T and a parameter of type
    ``reference to T accept the same set of initializer values, function
    declarations  with  parameter  types  differing only in this respect
    cannot be overloaded.

  +-------                 BEGIN BOX 1                -------+
  This restriction is hard to check across translation units.  Moreover,
  ambiguities can be detected just fine at call time.  Perhaps we should
  remove it.
  +-------                  END BOX 1                 -------+

  For example,
          int f(int i)
          {
              // ...
          }

          int f(int& r)  // error: function types
                         // not sufficiently different
          {
              // ...
          }
    It is, however, possible to distinguish between reference  to  const
    T,  reference  to  volatile  T, and plain reference to T so function
    declarations that differ only in this  respect  can  be  overloaded.
    Similarly, it is possible to distinguish between pointer to const T,
    pointer to volatile T, and plain reference to T so function declara­
    tions that differ only in this respect can be overloaded.

  --Function  declarations that differ only in the return type cannot be
    overloaded.

  --Member function declarations with the same name and the same parame­
    ter  types  cannot  be  overloaded if any of them is a static member
    function declaration (_class.static_).  The types  of  the  implicit
    object  parameters constructed for the member functions for the pur­
    pose of overload resolution (_over.match.funcs_) are not  considered
    when  comparing  parameter  types  for enforcement of this rule.  In
    contrast, if there is no static member function declaration among  a
    set  of member function declarations with the same name and the same
    parameter types, then these  member  function  declarations  can  be
    overloaded  if  they  differ  in  the  type of their implicit object
    parameter.  The following example illustrates this distinction:
              class X {
                  static void f();
                  void f();                  // ill-formed
                  void f() const;            // ill-formed
                  void f() const volatile;   // ill-formed
                  void g();
                  void g() const;            // Ok: no static g
                  void g() const volatile;   // Ok: no static g
              };

3 Function declarations  that  have  equivalent  parameter  declarations
  declare the same function and therefore cannot be overloaded:

  --Parameter  declarations  that  differ  only in the use of equivalent
    typedef types are equivalent.  A typedef is not a separate type, but
    only a synonym for another type (_dcl.typedef_).  For example,
              typedef int Int;

              void f(int i);
              void f(Int i);                  // OK: redeclaration of f(int)
              void f(int i) { /* ... */ }
              void f(Int i) { /* ... */ }     // error: redefinition of f(int)

    Enumerations,  on the other hand, are distinct types and can be used
    to distinguish overloaded function declarations.  For example,

              enum E { a };

              void f(int i) { /* ... */ }
              void f(E i)   { /* ... */ }

  --Parameter declarations that differ only in a  pointer  *  versus  an
    array [] are equivalent.  That is, the array declaration is adjusted
    to become a pointer declaration (_dcl.fct_).   Note  that  only  the
    second  and subsequent array dimensions are significant in parameter
    types (_dcl.array_).
              f(char*);
              f(char[]);      // same as f(char*);
              f(char[7]);     // same as f(char*);
              f(char[9]);     // same as f(char*);
              g(char(*)[10]);
              g(char[5][10]);  // same as g(char(*)[10]);
              g(char[7][10]);  // same as g(char(*)[10]);
              g(char(*)[20]);  // different from g(char(*)[10]);

  --Parameter declarations that differ only in the presence  or  absence
    of  const  and/or  volatile  are equivalent.  That is, the const and
    volatile type-specifiers for each parameter type  are  ignored  when
    determining  which  function  is being declared, defined, or called.
    For example,
              typedef const int cInt;

              int f (int);
              int f (const int);      // redeclaration of f (int);
              int f (int) { ... }     // definition of f (int)
              int f (cInt) { ... }    // error: redefinition of f (int)

    Only the const and volatile type-specifiers at the  outermost  level
    of  the  parameter  type  specification are ignored in this fashion;
    const and volatile type-specifiers buried within  a  parameter  type
    specification  are  significant and can be used to distinguish over­
    loaded function  declarations.   In  particular,  for  any  type  T,
    pointer to T, pointer to const T, and pointer to volatile T are con­
    sidered distinct parameter types, as are reference to  T,  reference
    to const T, and reference to volatile T.

  --Two  parameter  declarations  that differ only in their default ini­
    tialization are equivalent.  Consider the following example
              void f (int i, int j);
              void f (int i, int j = 99);         // Ok: redeclaration of f (int, int)
              void f (int i = 88, int j = 99);    // Ok: redeclaration of f (int, int)
              void f ();                          // Ok: overloaded declaration of f

              void prog ()
              {
                  f (1, 2);  // Ok: call f (int, int)
                  f (1);     // Ok: call f (int, int)
                  f ();      // Error: f (int, int) or f ()?
              }

  13.1.1  Declaration matching                                [over.dcl]

1 Two function declarations of the same name refer to the same  function
  if  they  are in the same scope and have equivalent parameter declara­
  tions (_over.load_).  A function member of a derived class is  not  in
  the  same scope as a function member of the same name in a base class.
  For example,
          class B {
          public:
              int f(int);
          };

          class D : public B {
          public:
              int f(char*);
          };
  Here D::f(char*) hides B::f(int) rather than overloading it.
          void h(D* pd)
          {
              pd->f(1);       // error:
                              // D::f(char*) hides B::f(int)
              pd->B::f(1);    // ok
              pd->f("Ben");   // ok, calls D::f
          }
  A locally declared function is not in the same scope as a function  in
  a containing scope.
          int f(char*);
          void g()
          {
              extern f(int);
              f("asdf");  // error: f(int) hides f(char*)
                          // so there is no f(char*) in this scope
          }
          void caller ()
          {
              void callee (int, int);
              {
                  void callee (int);  // hides callee (int, int)
                  callee (88, 99);    // error: only callee (int) in scope
              }
          )

2 Different  versions of an overloaded member function can be given dif­
  ferent access rules.  For example,
          class buffer {
          private:
              char* p;
              int size;
          protected:
              buffer(int s, char* store) { size = s; p = store; }
              // ...

          public:
              buffer(int s) { p = new char[size = s]; }
              // ...
          };

  13.2  Overload resolution                                 [over.match]

1 Overload resolution is a mechanism for selecting the best function  to
  call  given  a list of expressions that are to be the arguments of the
  call and a set of candidate functions that can be called based on  the
  context of the call.  The selection criteria for the best function are
  the number of arguments, how well the arguments match the types of the
  parameters  of the candidate function, and certain other properties of
  the candidate function.  The function selected by overload  resolution
  is  not  guaranteed to be appropriate for the context.  Other restric­
  tions, such as the accessibility of the function, can make its use  in
  the calling context ill-formed.

2 Overload resolution selects the function to call in five distinct con­
  texts within the language:

  --Invocation  of  a  function  named  in  the  function  call   syntax
    (_expr.call_)

  --Invocation  of  a function call operator, a pointer-to-function con­
    version function, or a reference-to-function conversion function  of
    a class object named in the function call syntax (_over.match.call_)

  --Invocation of the operator referenced in an expression (_expr_)

  --Invocation of a constructor during initialization of a class  object
    via a parenthesized expression list (_class.expl.init_)

  --Invocation  of  a user-defined conversion during initialization from
    an expression (_dcl.init_, _dcl.init.ref_)

3 Each of these contexts defines the set of candidate functions and  the
  list  of  arguments  in  its  own unique way.  But, once the candidate
  functions and argument lists have been identified,  the  selection  of
  the best function is the same in all cases:

  --First,  a  subset  of  the  candidate functions--those that have the
    proper number of arguments and meet  certain  other  conditions---is
    selected to form a set of viable functions.

  --Then the best viable function is selected based on the implicit con­
    version sequences (_over.best.ics_) needed to match each argument to
    the corresponding parameter of each viable function.

4 If  a  best  viable function exists and is unique, overload resolution
  succeeds and produces it as the result.  Otherwise overload resolution
  fails and the invocation is ill-formed.

  13.2.1  Candidate functions and argument lists      [over.match.funcs]

1 The following subclauses describe the set of candidate  functions  and
  the argument list submitted to overload resolution in each of the five
  contexts in which overload resolution is used.  The source transforma­
  tions  and  constructions defined in these subclauses are only for the
  purpose of describing the overload resolution process.  An implementa­
  tion is not required to use such transformations and constructions.

2 The  set of candidate functions can contain both member and non-member
  functions to be resolved against the  same  argument  list.   So  that
  argument  and parameter lists are comparable within this heterogeneous
  set, a member function is  considered  to  have  an  extra  parameter,
  called  the implicit object parameter, which represents the object for
  which the member function has been called.  For the purposes of  over­
  load  resolution,  both static and non-static member functions have an
  implicit object parameter, but constructors do not.

3 Similarly, when appropriate, the context  can  construct  an  argument
  list  that contains an implied object argument to denote the object to
  be operated on.  Since arguments  and  parameters  are  associated  by
  position  within  their  respective  lists, the convention is that the
  implicit object parameter, if present, is always the  first  parameter
  and the implied object argument, if present, is always the first argu­
  ment.

4 For non-static member functions,  the  type  of  the  implicit  object
  parameter  is  reference to cv X where X is the class that defines the
  member function and cv is the cv-qualification on the member  function
  declaration.  For example, for a const member function of class X, the
  extra parameter is assumed to have type  reference  to  const X.   For
  static  member  functions, the implicit object parameter is considered
  to match any object (since if the function is selected, the object  is
  discarded).

5 During  overload  resolution, the implied object argument is indistin­
  guishable from other arguments.  The implicit object  parameter,  how­
  ever,  retains  its  identity  since  conversions on the corresponding
  argument shall obey these additional rules:

  --no temporary object can be introduced to hold the argument  for  the
    implicit object parameter

  --no  user-defined  conversions can be applied to achieve a type match
    with it

  --even if the implicit object parameter  is  not  const-qualified,  an
    rvalue  temporary  can  be  bound to the parameter as long as in all
    other respects the temporary can be converted to  the  type  of  the
    implicit object parameter.

  13.2.1.1  Function call syntax                       [over.match.call]

1 Recall from _expr.call_, that a function call is a postfix-expression,
  possibly nested  arbitrarily  deep  in  parentheses,  followed  by  an
  optional expression-list enclosed in parentheses:
          (...(opt postfix-expression )...)opt (expression-listopt)
  Overload  resolution  is required if the postfix-expression yields the
  name of a function, an object of class type, or a set of  pointers-to-
  function.

2 Subclauses   _over.call.func_  and  _over.call.object_,  respectively,
  describe how overload resolution is used in the  first  two  cases  to
  determine the function to call.

3 The  third case arises from a postfix-expression of the form &F, where
  F names a set of overloaded functions.  In the context of  a  function
  call,  the  set  of functions named by F shall contain only non-member
  functions and static member functions1).  And in this context using &F
  behaves  the  same  as  using   the   name   F   by   itself.    Thus,
  (&F)(expression-listopt)  is  simply (F)(expression-listopt), which is
  discussed in _over.call.func_.  (The resolution of &F  in  other  con­
  texts is described in _over.over_.)

  13.2.1.1.1  Call to named function                    [over.call.func]

1 Of  interest  in this subclause are only those function calls in which
  the postfix-expression ultimately contains a name that denotes one  or
  more  functions that might be called.  Such a postfix-expression, per­
  haps nested arbitrarily deep in parentheses, has one of the  following
  forms:
          postfix-expression:
                  postfix-expression . id-expression
                  postfix-expression -> id-expression
                  primary-expression
  These  represent two syntactic subcategories of function calls: quali­
  fied function calls and unqualified function calls.

2 In qualified function calls,  the  name  to  be  resolved  is  an  id-
  expression  and  is  preceded by an -> or .  operator.  Since the con­
  struct A->B is generally equivalent to (*A).B, the rest of this clause
  assumes,  without  loss  of generality, that all member function calls
  have been normalized to the form that uses an object and the .  opera­
  tor.   Furthermore,  this  clause  assumes that the postfix-expression
  that is the left operand of the .  operator has  type  cv  T  where  T
  denotes a class2).  Under this assumption, the  id-expression  in  the
  call  is  looked  up as a member function of T following the rules for
  looking up names in classes (_class.derived_).  If a  member  function
  _________________________
  1) If F names a non-static member function, &F is a pointer-to-member,
  which cannot be used with the function call syntax.
  2)  Note  that cv-qualifiers on the type of objects are significant in
  overload resolution for both lvalue and rvalue objects.

  is found, that function and its overloaded declarations constitute the
  set of candidate functions.  Because of the usual name  hiding  rules,
  these  will  all  be declared in T or they will all be declared in the
  same base class of T.  The argument list is the expression-list in the
  call  augmented by the addition of the left operand of the .  operator
  in the normalized member function call as the implied object argument.

3 In unqualified function calls, the name is not qualified by an -> or .
  operator and has the more general form of a  primary-expression.   The
  name  is  looked  up in the context of the function call following the
  normal rules for name lookup.  If the name resolves  to  a  non-member
  function  declaration,  that  function and its overloaded declarations
  constitute the set of candidate functions.  Because of the usual  name
  hiding  rules,  these will all be declared in the same block or names­
  pace.  The argument list is the same as  the  expression-list  in  the
  call.   If  the  name resolves to a member function, then the function
  call is actually a member function call.  If the keyword  this  is  in
  scope  and refers to the class of that member function, then the func­
  tion call is transformed into a  normalized  qualified  function  call
  using  (*this)  as the postfix-expression to the left of the .  opera­
  tor.  The candidate functions and argument list are as  described  for
  qualified  function  calls above.  If the keyword this is not in scope
  or refers to another class, then name resolution found a static member
  of  some  class  T.   In this case, all overloaded declarations of the
  function name in T become candidate functions and a  contrived  object
  of type T becomes the implied object argument3).   The  call  is  ill-
  formed,  however, if overload resolution selects one of the non-static
  member functions of T in this case.

  13.2.1.1.2  Call to object of class type            [over.call.object]

1 If the primary-expression E in the function call syntax evaluates to a
  class  object  of  type  cv  T,  then  the  set of candidate functions
  includes at least the function call operators of T.  The function call
  operators  of T are obtained by ordinary lookup of the name operator()
  in the context of (E).operator().  Because of the  usual  name  hiding
  rules, these will all be declared in T or they will all be declared in
  the same base class of T.

2 In addition, for each conversion function declared in T of the form
          operator conversion-type-id () cv-qualifier;
  where conversion-type-id denotes the type  pointer  to  function  with
  parameters  of  type  P1,...,Pn  and  returning R or type reference to
  function with parameters of type P1,...,Pn and returning R,  a  surro­
  gate  call  function with the unique name call-function and having the
  _________________________
  3)  An  implied object argument must be contrived to correspond to the
  implicit object parameter attributed to member functions during  over­
  load resolution.  It is not used in the call to the selected function.
  Since the member functions all have the same implicit  object  parame­
  ter,  the contrived object will not be the cause to select or reject a
  function.

  form
          R call-function (conversion-type-id F, P1 a1,...,Pn an) { return F (a1,...,an); }
  is also considered as a candidate function.  Similarly, surrogate call
  functions are added to the set of candidate functions for each conver­
  sion function declared in an accessible base class provided the  func­
  tion is not hidden within T by another intervening declaration4).

3 If such a surrogate call function is selected by overload  resolution,
  its  body,  as  defined  above,  will  be executed to convert E to the
  appropriate function and then to invoke that function with  the  argu­
  ments of the call.

4 The  argument  list  submitted  to overload resolution consists of the
  argument expressions present in the function call syntax  preceded  by
  the  implied object argument (E).  When comparing the call against the
  function call operators,  the  implied  object  argument  is  compared
  against  the  implicit object parameter of the function call operator.
  When comparing the call against a surrogate call funtion, the  implied
  object  argument is compared against the first parameter of the surro­
  gate call function.  The conversion function from which the  surrogate
  call  function was derived will be used in the conversion sequence for
  that parameter since it converts the implied object  argument  to  the
  appropriate  function  pointer  or  reference  required  by that first
  parameter.

  13.2.1.2  Operators in expressions                   [over.match.oper]

1 If no operand of the operator has a type that is a class or an enumer­
  ation,  the  operator  is assumed to be a built-in operator and inter­
  preted according to clause _expr_.  For example,

  _________________________
  4) Note that this construction can yield candidate call functions that
  cannot be differentiated one from the other by overload resolution be­
  cause  they have identical declarations or differ only in their return
  type.  The call will be ambiguous if overload resolution cannot select
  a match to the call that is uniquely better than such undifferentiable
  functions.

          class String {
          public:
             String (const String&);
             String (char*);
                  operator char* ();
          };
          String operator + (const String&, const String&);

          void f(void)
          {
             char* p= "one" + "two"; // ill-formed because neither
                                     // operand has user defined type
             int I = 1 + 1;          // Always evaluates to 2 even if
                                     // user defined types exist which
                                     // would perform the operation.
          }

2 If either operand has a type that is a  class  or  an  enumeration,  a
  user-defined  operator function might be declared that implements this
  operator or a user-defined conversion can be necessary to convert  the
  operand  to  a  type  that is appropriate for a built-in operator.  In
  this case, overload resolution is used  to  determine  which  operator
  function  is  to be invoked to implement the operator.  Therefore, the
  operator notation is first transformed to the equivalent function-call
  notation  as  summarized in Table 1 (where @ denotes one of the opera­
  tors covered in the specified subclause).

    Table 1--relationship between operator and function call notation

  +--------------+------------+--------------------+------------------------+
  |Subclause     | Expression | As member function | As non-member function |
  +--------------+------------+--------------------+------------------------+
  |_over.unary_  | @a         | (a).operator@ ()   | operator@ (a)          |
  |_over.binary_ | a@b        | (a).operator@ (b)  | operator@ (a, b)       |
  |_over.ass_    | a=b        | (a).operator= (b)  |                        |
  |_over.sub_    | a[b]       | (a).operator[](b)  |                        |
  |_over.ref_    | a->        | (a).operator-> ()  |                        |
  |_over.inc_    | a@         | (a).operator@ (0)  | operator@ (a, 0)       |
  +--------------+------------+--------------------+------------------------+

3 Three sets of candidate functions are constructed as follows:

  --If the first operand of the operator is an object or reference to an
    object  of  class  X,  the operator could be implemented by a member
    operator function of X.  The expression is transformed to  a  quali­
    fied  function  call  per column 3 of Table 1 and a set of candidate
    functions is constructed for the transformed call according  to  the
    rules in _over.call.func_.  This set is designated the member candi­
    dates .

  --If the operator is either a unary or binary operator  (_over.unary_,

    _over.binary_,  or _over.inc_), the operator could be implemented by
    a non-member operator function.  The expression is transformed to an
    unqualified  function  call  per  column 4 of Table 1.  The operator
    name is looked up in the context of  the  expression  following  the
    usual  rules  for  name  lookup except that all member functions are
    ignored.  Thus, if the operator name resolves to any declaration, it
    will be to a non-member function declaration.  That function and its
    overloaded declarations constitute the set  of  candidate  functions
    designated  the  non-member  candidates.  Because of the name hiding
    rules, these will all be declared in the same block or  namespace5).

  +-------                 BEGIN BOX 2                -------+
  A  motion  is  expected  in Valley Forge that would eliminate all name
  hiding when resolving non-member operator names so that the non-member
  candidates  would include all operators of the same name with a decla­
  ration in any enclosing block or namespace.
  +-------                  END BOX 2                 -------+

  --In any case, a set of candidate functions, called the built-in  can­
    didates,  is  constructed.   For  the binary operator , or the unary
    operator &, the built-in candidates set is  empty.   For  all  other
    operators, the built-in candidates include all of the built-in oper­
    ators defined in _over.built_ that, compared to the given operator,

    --have the same operator name, and

    --accept the same number of operands, and

    --accept operand types to which the given operand or operands can be
      converted according to _over.best.ics_.

4 For the built-in assignment operators, conversions of the left operand
  are restricted as follows:
  _________________________
  5)  Note  that the look up rules for operators in expressions are dif­
  ferent than the lookup rules for operator function names in a function
  call as shown in the following example:
  struct A { };
  void operator + (A, A);

  struct B {
    void operator + (B);
    void f ();
  };

  A a;

  void B::f() {
    operator+ (a,a);      // ERROR - global operator hidden by member
    a + a;                // OK - calls global operator+
  }

  --no temporaries are introduced to hold the left operand

  --no user-defined conversions are applied to achieve a type match with
    it

5 For all other operators, no such restrictions apply.

6 If  a built-in candidate is selected by overload resolution, any class
  operands are first converted to the appropriate type for the operator.
  Then  the  operator  is treated as the corresponding built-in operator
  and interpreted according to clause  _expr_.   The  set  of  candidate
  functions  for  overload  resolution is the union of the member candi­
  dates, the non-member candidates, and the  built-in  candidates.   The
  argument list contains all of the operands of the operator.

7 If  the  operator  is the binary operator ,or the unary operator & and
  overload resolution is unsuccessful, then the operator is  assumed  to
  be the built-in operator and interpreted according to clause _expr_.

  13.2.1.3  Initialization by user-defined             [over.match.user]
       conversions

1 Under the conditions specified in  _dcl.init_  and  _dcl.init.ref_,  a
  user-defined  conversion  can  be  invoked  to convert the assignment-
  expression of an initializer-clause to the type of  the  object  being
  initialized (which might be a temporary in the reference case).  Over­
  load resolution is used to select the user-defined  conversion  to  be
  invoked.  Assuming that cv1 T is the type of the object being initial­
  ized, the candidate functions are selected as follows:

  --When T is a class type, the constructors of T  are  candidate  func­
    tions

  --When the type of the assignment-expression is a class type cv S, the
    conversion functions of S  and  its  base  classes  are  considered.
    Those  that  are  not hidden within S and yield type cv2 T or a type
    that can be converted to type cv2 T, for any cv2 that  is  the  same
    cv-qualification  as,  or  lesser  cv-qualification than, cv1, via a
    standard conversion sequence (_over.ics.scs_)  are  candidate  func­
    tions

2 In  both  cases,  the  argument  list  has  one argument, which is the
  assignment-expression of the initializer-clause.  This  argument  will
  be  compared  against  the  first  parameter  of  the constructors and
  against the implicit object parameter of the conversion functions.

3 Because only one user-defined conversion is  allowed  in  an  implicit
  conversion sequence, special rules apply when selecting the best user-
  defined conversion (_over.match.best_, _over.best.ics_).

  13.2.1.4  Initialization by constructor              [over.match.ctor]

1 When objects of classes  with  constructors  are  initialized  with  a
  parenthesized expression-list (_class.expl.init_), overload resolution
  selects the constructor.  The candidate functions  are  all  the  con­
  structors  of the class of the object being initialized.  The argument
  list is the expression-list within the parentheses of the initializer.

  13.2.2  Viable functions                           [over.match.viable]

1 From  the  set  of candidate functions constructed for a given context
  (_over.match.funcs_), a set of viable functions is chosen, from  which
  the  best  function  will be selected by comparing argument conversion
  sequences for the best  fit  (_over.match.best_).   The  selection  of
  viable  functions  considers relationships between arguments and func­
  tion parameters other than the ranking of conversion sequences.

2 First, to be a viable function, a candidate function shall have enough
  parameters to agree in number with the arguments in the list.

  --If there are m arguments in the list, all candidate functions having
    exactly m parameters are viable.

  --A candidate function having fewer than m parameters is  viable  only
    if  it  has  an ellipsis in its parameter list (_dcl.fct_).  For the
    purposes of overload resolution, its parameter list is  extended  to
    the right with ellipses so that there are exactly m parameters.

  --A candidate function having more than m parameters is viable only if
    the    (m+1)-st    parameter    has    a     default     initializer
    (_dcl.fct.default_).   For  the purposes of overload resolution, the
    parameter list is truncated on the right, so that there are  exactly
    m parameters.

3 Second,  for  F  to  be  a viable function, there shall exist for each
  argument an implicit conversion sequence (_over.best.ics_)  that  con­
  verts  that  argument  to  the  corresponding  parameter of F.  If the
  parameter  has  reference  type,  the  implicit  conversion   sequence
  includes  the  operation of binding the reference, and the fact that a
  reference to non-const cannot be bound to an  rvalue  can  affect  the
  viability of the function (see _over.ics.ref_).

  13.2.3  Best Viable Function                         [over.match.best]

1 Let  ICSi(F) denote the implicit conversion sequence that converts the
  i-th argument in the list to the type of the i-th parameter of  viable
  function F.  Subclause _over.best.ics_ defines the implicit conversion
  sequences and subclause _over.ics.rank_ defines what it means for  one
  implicit  conversion  sequence  to  be a better conversion sequence or
  worse conversion sequence than another.  Given  these  definitions,  a
  viable  function  F1  is  defined to be a better function than another
  viable function F2 if for all arguments i, ICSi(F1)  is  not  a  worse
  conversion sequence than ICSi(F2), and then

  --for  some  argument j, ICSj(F1) is a better conversion sequence than
    ICSj(F2), or, if not that,

  --F1 is a non-template function and F2 is a template function, or,  if
    not that,

  --the  context  is  an  initialization by user-defined conversion (see
    _dcl.init_  and  _over.match.user_)  and  the  standard   conversion
    sequence  from  the return type of F1 to the destination type (i.e.,
    the type of the entity being initialized)  is  a  better  conversion
    sequence  than the standard conversion sequence from the return type
    of F2 to the destination type.  For example,
              struct A {
                  A();
                  operator int();
                  operator double();
              } a;
              int i = a;     // a.operator int() followed by no conversion is better
                             // than a.operator double() followed by a conversion
                             // to int
              float x = a;   // ambiguous: both possibilities require conversions,
                             // and neither is better than the other

2 If there is exactly one viable function that is a better function than
  all  other  viable  functions, then it is the one selected by overload
  resolution; otherwise the call is ill-formed6).

3 Examples:

  _________________________
  6) The algorithm for selecting the best viable function is  linear  in
  the  number  of  viable  functions.  Run a simple tournament to find a
  function W that is not worse than any opponent it faced.  Although an­
  other  function F that W did not face might be better than W, F cannot
  be the best function because at some point in the tournament F encoun­
  tered  another function G such that F was not better than G.  Hence, W
  is either the best function or there is no best function.  So, make  a
  second  pass over the viable functions to verify that W is better than
  all other functions.

          void Fcn(const int*,  short);
          void Fcn(int*, int);

          int i;
          short s = 0;

          Fcn(&i, s);     // is ambiguous because
                          // &i -> int* is better than &i -> const int*
                          // but s -> short is also better than s -> int

          Fcn(&i, 1L);    // calls Fcn(int*, int), because
                          // &i -> int* is better than &i -> const int*
                          // and 1L -> short and 1L -> int are indistinguishable

          Fcn(&i,'c');    // calls Fcn(int*, int), because
                          // &i -> int* is better than &i -> const int*
                          // and 'c' -> int is better than 'c' -> short

  13.2.3.1  Implicit conversion sequences                [over.best.ics]

1 An implicit conversion sequence is a sequence of conversions  used  to
  convert  an argument in a function call to the type of the correspond­
  ing parameter of the function being called.  The sequence  of  conver­
  sions is governed by the rules for initialization of an object or ref­
  erence by a single expression (_dcl.init_ and _dcl.init.ref_).

2 Implicit conversion sequences are concerned only with  the  type,  cv-
  qualification,  and lvalue-ness of the argument and how these are con­
  verted to match the corresponding properties of the parameter.   Other
  properties,  such as the lifetime, storage class, alignment, or acces­
  sibility of the argument and whether or not the  argument  is  a  bit-
  field  are  ignored.  So, although an implicit conversion sequence can
  be defined for a given argument-parameter pair,  the  conversion  from
  the  argument  to the parameter might still be ill-formed in the final
  analysis.

3 Except in the context of an initialization by user-defined  conversion
  (_over.match.user_), a well-formed implicit conversion sequence is one
  of the following forms:

  --a standard conversion sequence (_over.ics.scs_),

  --a user-defined conversion sequence (_over.ics.user_), or

  --an ellipsis conversion sequence (_over.ics.ellipsis_).

4 In the context of an initialization by user-defined conversion  (i.e.,
  when  considering  the argument of a user-defined conversion function;
  see _over.match.user_), only standard conversion sequences and  ellip­
  sis conversion sequences are allowed.

5 When  initializing a reference, the operation of binding the reference
  to an object or temporary occurs after any  conversion.   The  binding

  operation  is  not  a conversion, but it is considered to be part of a
  standard conversion sequence, and it can affect the rank of  the  con­
  version sequence.  See _over.ics.ref_.

6 In  all  contexts, when converting to the implicit object parameter or
  when converting to the left operand of an  assignment  operation  only
  standard  conversion sequences that create no temporary object for the
  result are allowed.

7 If no conversions are required to match an  argument  to  a  parameter
  type,  the  implicit  conversion  sequence  is the standard conversion
  sequence consisting of the identity conversion (_over.ics.scs_).

8 If no sequence of conversions can be found to convert an argument to a
  parameter  type or the conversion is otherwise ill-formed, an implicit
  conversion sequence cannot be formed.

9 If several different sequences of conversions exist that each  convert
  the  argument  to the parameter type, the implicit conversion sequence
  is a sequence among these that is not worse than all the rest  accord­
  ing to _over.ics.rank_7).  If that conversion sequence in  not  better
  than all the rest and a function that uses such an implicit conversion
  sequence is selected as the best viable function, then the  call  will
  be  ill-formed  because  the conversion of one of the arguments in the
  call is ambiguous.

10The three forms of implicit conversion sequences mentioned  above  are
  defined in the following subclauses.

  _________________________
  7)  This  rule prevents a function from becoming non-viable because of
  an ambiguous conversion sequence for one of its parameters.   Consider
  this example,
          class B;
          class A { A (B&); };
          class B { operator A (); };
          class C { C (B&); };
          f(A) { }
          f(C) { }
          B b;
          f(b);   // ambiguous since b -> C via constructor and
                  // b -> A via constructor or conversion function.
  If  it  were  not  for this rule, f(A) would be eliminated as a viable
  function for the call f(b) causing overload resolution to select  f(C)
  as the function to call even though it is not clearly the best choice.
  On the other hand, if an f(B) were to be declared then f(b) would  re­
  solved  to  that f(B) because the exact match with f(B) is better than
  any of the sequences required to match f(A).

  13.2.3.1.1  Standard conversion sequences               [over.ics.scs]

1 Table 2 summarizes the conversions defined in clause _conv_ and parti­
  tions them into four disjoint categories: Lvalue Transformation, Qual­
  ification Adjustment, Promotion, and Conversion.  Note that these cat­
  egories are orthogonal with respect to lvalue-ness,  cv-qualification,
  and  data representation: the Lvalue Transformations do not change the
  cv-qualification or data representation of the type; the Qualification
  Adjustments  do  not  change the lvalue-ness or data representation of
  the type; and the Promotions and Conversions do not change the lvalue-
  ness or cv-qualification of the type.

2 A  standard  conversion  sequence is either the Identity conversion by
  itself or consists of one to four conversions from the other four cat­
  egories.   At  most  one conversion from each category is allowed in a
  single standard conversion sequence.  If there are two or more conver­
  sions  in  the  sequence, the conversions are applied in the canonical
  order: Lvalue  Transformation,  Promotion,  Conversion,  Qualification
  Adjustment.

3 Each  conversion  in Table 2 also has an associated rank (Exact Match,
  Promotion, or Conversion).  These are used to rank standard conversion
  sequences  (_over.ics.rank_).   The  rank  of a conversion sequence is
  determined by considering the rank of each conversion in the  sequence
  and  the  rank  of  any reference binding (_over.ics.ref_).  If any of
  those has Conversion rank, the sequence has  Conversion  rank;  other­
  wise,  if  any of those has Promotion rank, the sequence has Promotion
  rank; otherwise, the sequence has Exact Match rank.

                           Table 2--conversions

  +-------------------------------+--------------------------+-------------+-----------------+
  |Conversion                     |         Category         |    Rank     |    Subclause    |
  +-------------------------------+--------------------------+-------------+-----------------+
  +-------------------------------+--------------------------+-------------+-----------------+
  |No conversions required        |         Identity         |             |                 |
  +-------------------------------+--------------------------+             +-----------------+
  |Lvalue-to-rvalue conversion    |                          |             |   _conv.lval_   |
  +-------------------------------+                          |             +-----------------+
  |Array-to-pointer conversion    |  Lvalue Transformation   | Exact Match |  _conv.array_   |
  +-------------------------------+                          |             +-----------------+
  |Function-to-pointer conversion |                          |             |   _conv.func_   |
  +-------------------------------+--------------------------+             +-----------------+
  |Qualification conversions      | Qualification Adjustment |             |   _conv.qual_   |
  +-------------------------------+--------------------------+-------------+-----------------+
  |Integral promotions            |                          |             |   _conv.prom_   |
  +-------------------------------+        Promotion         |  Promotion  +-----------------+
  |Floating point promotion       |                          |             |  _conv.fpprom_  |
  +-------------------------------+--------------------------+-------------+-----------------+
  |Integral conversions           |                          |             | _conv.integral_ |
  +-------------------------------+                          |             +-----------------+
  |Floating point conversions     |                          |             |  _conv.double_  |
  +-------------------------------+                          |             +-----------------+
  |Floating-integral conversions  |                          |             |  _conv.fpint_   |
  +-------------------------------+                          |             +-----------------+
  |Pointer conversions            |        Conversion        | Conversion  |   _conv.ptr_    |
  +-------------------------------+                          |             +-----------------+
  |Pointer to member conversions  |                          |             |   _conv.mem_    |
  +-------------------------------+                          |             +-----------------+
  |Base class conversion          |                          |             |  _conv.class_   |
  +-------------------------------+                          |             +-----------------+
  |Boolean conversions            |                          |             |   _conv.bool_   |
  +-------------------------------+--------------------------+-------------+-----------------+

  13.2.3.1.2  User-defined conversion sequences          [over.ics.user]

1 A user-defined conversion sequence consists  of  an  initial  standard
  conversion    sequence   followed   by   a   user-defined   conversion
  (_class.conv_) followed by a second standard conversion sequence.   If
  the   user-defined   conversion   is   specified   by   a  constructor
  (_class.conv.ctor_), the initial standard conversion sequence converts
  the  source type to the type required by the argument of the construc­
  tor.  If the user-defined conversion  is  specified  by  a  conversion
  function  (_class.conv.fct_), the initial standard conversion sequence
  converts the source type to the implicit object parameter of the  con­
  version function.

2 The  second  standard  conversion  sequence converts the result of the
  user-defined conversion to the target type for the sequence.  Since an

  implicit  conversion  sequence is an initialization, the special rules
  for initialization by user-defined conversion apply when selecting the
  best  user-defined  conversion  for a user-defined conversion sequence
  (see _over.match.best_ and _over.best.ics_)

3 It should be noted that a conversion of an expression of class type to
  the same class type or to a base class of that type is a standard con­
  version rather than a user-defined conversion in  spite  of  the  fact
  that  a copy constructor (i.e., a user-defined conversion function) is
  called.

  13.2.3.1.3  Ellipsis conversion sequences          [over.ics.ellipsis]

1 An ellipsis conversion sequence occurs when an argument in a  function
  call is matched with the ellipsis parameter specification of the func­
  tion called.

  13.2.3.1.4  Reference binding                           [over.ics.ref]

1 The operation of binding a reference is not a conversion, but for  the
  purposes of overload resolution it is considered to be part of a stan­
  dard conversion sequence (specifically, it is the last step in such  a
  sequence).

2 A standard conversion sequence cannot be formed if it requires binding
  a reference to non-const to an rvalue (except when binding an implicit
  object   parameter;   see   the   special   rules  for  that  case  in
  _over.match.funcs_).  This means, for example, that a candidate  func­
  tion  cannot  be  a  viable  function  if it has a non-const reference
  parameter (other than the implicit object parameter)  and  the  corre­
  sponding argument is a temporary or would require one to be created to
  initialize the reference (see _dcl.init.ref_).

3 Other restrictions on binding a reference to a particular argument  do
  not  affect  the formation of a standard conversion sequence, however.
  For example, a function with a reference to int  parameter  can  be  a
  viable  candidate  even  if  the corresponding argument is an int bit-
  field.  The formation of implicit conversion sequences treats the  int
  bit-field  as  an int lvalue and finds an exact match with the parame­
  ter.  If the function is selected by  overload  resolution,  the  call
  will nonetheless be ill-formed because of the prohibition on binding a
  non-const reference to a bit-field (_dcl.init.ref_).

4 A reference binding in general has no effect on the rank of a standard
  conversion sequence, but there is one exception: the binding of a ref­
  erence to a (possibly cv-qualified) class to an expression of a  (pos­
  sibly  cv-qualified)  class  derived from that class gives the overall
  standard conversion sequence Conversion rank.

  13.2.3.2  Ranking implicit conversion sequences        [over.ics.rank]

1 This  clause  defines  a  partial  ordering  of  implicit   conversion
  sequences  based  on  the relationships better conversion sequence and
  better conversion.  If an implicit conversion sequence S1  is  defined
  by  these rules to be a better conversion sequence than S2, then it is
  also the case that S2 is a worse conversion sequence than S1.  If con­
  version  sequence  S1 is neither better than nor worse than conversion
  sequence S2, S1 and S2 are said  to  be  indistinguishable  conversion
  sequences.

2 When  comparing  the  basic forms of implicit conversion sequences (as
  defined in _over.best.ics_)

  --A standard conversion sequence (_over.ics.scs_) is a better  conver­
    sion sequence than a user-defined conversion sequence or an ellipsis
    conversion sequence

  --A user-defined conversion sequence  (_over.ics.user_)  is  a  better
    conversion   sequence   than   an   ellipsis   conversion   sequence
    (_over.ics.ellipsis_)

3 Two implicit conversion sequences of the same form are  indistinguish­
  able conversion sequences unless one of the following rules apply:

  --Standard conversion sequence S1 is a better conversion sequence than
    standard conversion sequence S2 if

    --S1 is a proper subsequence of S2, or, if not that,

    --the dominant conversion of S1 is better than the dominant  conver­
      sion of S2 (by the rules defined below), or, if not that,

    --S1  and  S2 differ only in their qualification conversion and they
      yield types identical except for cv-qualifiers and S2 adds all the
      qualifiers  that  S1 adds (and in the same places) and S2 adds yet
      more cv-qualifiers than S1, or the  similar  case  with  reference
      binding  (see  the  definition  of reference-compatible with added
      qualification in _dcl.init.ref_).

  --User-defined conversion sequence U1 is a better conversion  sequence
    than another user-defined conversion sequence U2 if they contain the
    same user-defined conversion operator or constructor and if the sec­
    ond  standard  conversion  sequence  of U1 is better than the second
    standard conversion sequence of U2.

4 Standard conversions are ordered by their ranks: an Exact Match  is  a
  better  conversion than a Promotion, which is a better conversion than
  a Conversion.  Two conversions with the same rank  are  indistinguish­
  able unless one of the following rules applies:

  --If  class  B is derived directly or indirectly from class A, conver­
    sion of B* to A* is better than conversion of B* to void*.

  --If class B is derived directly or indirectly from class A and  class
    C is derived directly or indirectly from B,

    --conversion of C* to B* is better than conversion of C* to A*

    --Binding  of  an  expression of type C to a reference of type B& is
      better than binding an expression of type C to a reference of type
      A&

    --conversion  of  A::*  to B::* is better than conversion of A::* to
      C::*

  13.3  Address of overloaded function                       [over.over]

1 A use of a function name without arguments selects,  among  all  func­
  tions of that name that are in scope, the (only) function that exactly
  matches the target.  The target can be

  --an object being initialized (_dcl.init_)

  --the left side of an assignment (_expr.ass_)

  --a parameter of a function (_expr.call_)

  --a parameter of a user-defined operator (_over.oper_)

  --the return value of a function,  operator  function,  or  conversion
    (_stmt.return_)

  --an explicit type conversion (_expr.type.conv_, _expr.cast_)

2 Non-member functions match targets of type pointer-to-function; member
  functions match targets of type pointer-to-member-function.

3 Note that if f() and g() are  both  overloaded  functions,  the  cross
  product  of  possibilities must be considered to resolve f(&g), or the
  equivalent expression f(g).

4 For example,
          int f(double);
          int f(int);
          (int (*)(int))&f;           // cast expression as selector
          int (*pfd)(double) = &f;    // selects f(double)
          int (*pfi)(int) = &f;       // selects f (int)
          int (*pfe)(...) = &f;       // error: type mismatch
  The last  initialization  is  ill-formed  because  no  f()  with  type
  int(...)  has been defined, and not because of any ambiguity.

5 Note  also  that  there  are  no  standard conversions (_conv_) of one
  pointer-to-function type or  pointer-to-member-function  into  another
  (_conv.ptr_).  In particular, even if B is a public base of D we have

          D* f();
          B* (*p1)() = &f;       // error
          void g(D*);
          void (*p2)(B*) = &g;   // error

6 Note  that  if  the  target  type is a pointer to member function, the
  function type of the pointer to member is used to  select  the  member
  function from a set of overloaded member functions. For example:
          struct X {
              int f(int);
              static int f(long);
          };
          int (X::*p1)(int)  = &X::f;   // OK
          int    (*p2)(int)  = &X::f;   // error: mismatch
          int    (*p3)(long) = &X::f;   // OK
          int (X::*p4)(long) = &X::f;   // error: mismatch
          int (X::*p5)(int)  = &(X::f); // error: wrong syntax for
                                        // pointer to member
          int    (*p6)(long) = &(X::f); // OK

  13.4  Overloaded operators                                 [over.oper]

1 A  function declaration having one of the following operator-function-
  ids as its name declares an operator function.  An  operator  function
  is said to implement the operator named in its operator-function-id.
          operator-function-id:
                  operator operator
          operator: one of
                  new  delete    new[]     delete[]
                  +    -    *    /    %    ^    &    |    ~
                  !    =    <    >    +=   -=   *=   /=   %=
                  ^=   &=   |=   <<   >>   >>=  <<=  ==   !=
                  <=   >=   &&   ||   ++   --   ,    ->*  ->
                  ()   []
  The  last two operators are function call (_expr.call_) and subscript­
  ing (_expr.sub_).

2 Both the unary and binary forms of
                  +    -    *     &
  can be overloaded.

3 The following operators cannot be overloaded:
                  .    .*   ::    ?:
  nor can the preprocessing symbols # and ## (_cpp_).

4 Operator functions are usually not called directly; instead  they  are
  invoked  to  evaluate  the  operators  they  implement (_over.unary_ -
  _over.inc_).  They can be explicitly called, though.  For example,
          complex z = a.operator+(b);  // complex z = a+b;
          void* p = operator new(sizeof(int)*n);

5 The allocation and  deallocation  functions,  operator  new,  operator
  new[], operator delete and operator delete[], are described completely

  in _class.free_.  The attributes and restrictions found in the rest of
  this  section  do  not  apply  to  them  unless  explicitly  stated in
  _class.free_.

6 An operator function shall either be a non-static member function  or,
  be a non-member function and have at least one parameter whose type is
  a class, a reference to a class, an enumeration, or a reference to  an
  enumeration.   It  is not possible to change the precedence, grouping,
  or number of operands of operators.  The meaning of the  operators  =,
  (unary) &, and , (comma), predefined for each type, can be changed for
  specific types by defining operator  functions  that  implement  these
  operators.   Operator  functions are inherited the same as other func­
  tions, but because an instance  of  operator=  is  automatically  con­
  structed for each class (_class.copy_, _over.ass_), operator= is never
  inherited by a class from its bases.

7 The identities among certain predefined  operators  applied  to  basic
  types (for example, ++a == a+=1) need not hold for operator functions.
  Some predefined operators, such as +=, require an  operand  to  be  an
  lvalue  when  applied to basic types; this is not required by operator
  functions.

8 An    operator    function    cannot    have     default     arguments
  (_dcl.fct.default_).

9 Operators  not  mentioned explicitly below in _over.ass_ to _over.inc_
  act as ordinary unary and binary operators obeying the rules  of  sec­
  tion _over.unary_ or _over.binary_.

  13.4.1  Unary operators                                   [over.unary]

1 A  prefix  unary  operator  can  be implemented by a non-static member
  function (_class.mfct_) with no parameters or  a  non-member  function
  with  one parameter.  Thus, for any prefix unary operator @, @x can be
  interpreted as either x.operator@() or operator@(x).  If both forms of
  the   operator   function   have   been   declared,   the   rules   in
  _over.match.oper_ determine which, if  any,  interpretation  is  used.
  See  _over.inc_  for  an explanation of the postfix unary operators ++
  and --.

2 The unary and binary forms of the same operator are considered to have
  the same name.  Consequently, a unary operator can hide a binary oper­
  ator from an enclosing scope, and vice versa.

  13.4.2  Binary operators                                 [over.binary]

1 A binary operator can be implemented either  by  a  non-static  member
  function (_class.mfct_) with one parameter or by a non-member function
  with two parameters.  Thus, for any binary  operator  @,  x@y  can  be
  interpreted as either x.operator@(y) or operator@(x,y).  If both forms
  of  the  operator  function  have  been   declared,   the   rules   in
  _over.match.oper_ determines which, if any, interpretation is used.

  13.4.3  Assignment                                          [over.ass]

1 An  overloaded  assignment operator shall be a non-static member func­
  tion with exactly one parameter.  Because an instance of operator=  is
  constructed  for each class (_class.copy_), it is never inherited by a
  derived class.

2 A copy assignment operator operator= is a non-static  member  function
  of  class  X  with  exactly  one  parameter  of  type  X& or const X&.
  _class.copy_ describes the copy assignment operator.

  13.4.4  Function call                                      [over.call]

1 operator() shall be a non-static member function.  It  implements  the
  function call syntax
          postfix-expression ( expression-listopt )
  where  the postfix-expression evaluates to a class object and the pos­
  sibly empty expression-list matches the parameter list  of  an  opera­
  tor() member function of the class.  Thus, a call x(arg1,arg2,arg3) is
  interpreted as x.operator()(arg1,arg2,arg3) for a class  object  x  of
  type  T  if  T::operator()(T1,  T2,  T3) exists and if the operator is
  selected as the best match function by the overload resolution  mecha­
  nism (_over.match.best_).

  13.4.5  Subscripting                                        [over.sub]

1 operator[]  shall  be a non-static member function.  It implements the
  subscripting syntax
          postfix-expression [ expression ]
  Thus, a subscripting expression x[y] is interpreted as x.operator[](y)
  for  a class object x of type T if T::operator()(T1) exists and if the
  operator is selected as the best match function by the overload  reso­
  lution mechanism (_over.match.best_).

  13.4.6  Class member access                                 [over.ref]

1 operator-> shall be a non-static member function taking no parameters.
  It implements class member access using ->
          postfix-expression -> primary-expression
  An expression x->m is interpreted as (x.operator->())->m for  a  class
  object  x  of  type T if T::operator->() exists and if the operator is
  selected as the best match function by the overload resolution  mecha­
  nism  (_over.match_).  It follows that operator-> must return either a
  pointer to a class that has a member m or an object of or a  reference
  to a class for which operator-> is defined.

  13.4.7  Increment and decrement                             [over.inc]

1 The  prefix  and  postfix  increment operators can be implemented by a
  function called operator++.  If this function  is  a  member  function
  with no parameters, or a non-member function with one class parameter,
  it defines the prefix increment operator ++ for objects of that class.
  If  the  function is a member function with one parameter (which shall
  be of type int) or a non-member  function  with  two  parameters  (the

  second  shall be of type int), it defines the postfix increment opera­
  tor ++ for objects of that  class.   When  the  postfix  increment  is
  called, the int argument will have value zero.  For example,
          class X {
          public:
              const X&   operator++();     // prefix ++a
              const X&   operator++(int);  // postfix a++
          };
          class Y {
          public:
          };
          const Y&   operator++(Y&);       // prefix ++b
          const Y&   operator++(Y&, int);  // postfix b++
          void f(X a, Y b)
          {
              ++a;        // a.operator++();
              a++;        // a.operator++(0);
              ++b;        // operator++(b);
              b++;        // operator++(b, 0);
              a.operator++();    // explicit call: like ++a;
              a.operator++(0);   // explicit call: like a++;
              operator++(b);     // explicit call: like ++b;
              operator++(b, 0);  // explicit call: like b++;
          }

2 The prefix and postfix decrement operators -- are handled similarly.

  13.5  Built-in operators                                  [over.built]

1 The  built-in  operators  (_expr_)  participate in overload resolution
  (_over.match.oper_) as though declared as specified in  this  section.
  For  operator,  and  unary  operator&, a built-in operator is selected
  only if there are no user-defined operator candidates.  For all  other
  built-in operators, since they take only operands with non-class type,
  and operator overload resolution occurs only when an  operand  expres­
  sion  originally  has  class  type,  operator  overload resolution can
  resolve to a built-in operator only when an operand has a  class  type
  which  has  a  user-defined conversion to a non-class type appropriate
  for the operator.

2 In this section, the term promoted integral type is used to  refer  to
  those  integral  types  which  are  preserved  by  integral  promotion
  (including e.g.  int but excluding e.g.  char).  Similarly,  the  term
  promoted arithmetic type refers to promoted integral types plus float­
  ing types.

3 For every pair T, VQ), where T is an arithmetic type, and VQ is either
  volatile or empty, there exist
          VQ T&   operator++(VQ T&);
          VQ T&   operator--(VQ T&);
          T       operator++(VQ T&, int);
          T       operator--(VQ T&, int);

4 For  every  pair T, VQ), where T is a cv-qualified or unqualified com­
  plete object type, and VQ is either volatile or empty, there exist
          T*VQ&   operator++(T*VQ&);
          T*VQ&   operator--(T*VQ&);
          T*      operator++(T*VQ&, int);
          T*      operator--(T*VQ&, int);

5 For every cv-qualified or unqualified complete object  type  T,  there
  exists
          T&      operator*(T*);

6 For every function type T, there exists
          T&      operator*(T*);

7 For every type T, there exist
          T*      operator&(T&);
          T*      operator+(T*);

8 For every promoted arithmetic type T, there exist
          T       operator+(T);
          T       operator-(T);

9 For every promoted integral type T, there exists
          T       operator~(T);

10For  every  quadruple C, T, CV1, CV2), where C is a class type, T is a
  complete object type or a function type,  and  CV1  and  CV2  are  cv-
  qualifier-seqs, there exists
          CV12 T& operator->*(CV1 C*, CV2 T C::*);
  where CV12 is the union of CV1 and CV2.

11For every pair of promoted arithmetic types L and R, there exist
          LR      operator*(L, R);
          LR      operator/(L, R);
          LR      operator+(L, R);
          LR      operator-(L, R);
          bool    operator<(L, R);
          bool    operator>(L, R);
          bool    operator<=(L, R);
          bool    operator>=(L, R);
          bool    operator==(L, R);
          bool    operator!=(L, R);
  where  LR  is  the  result of the usual arithmetic conversions between
  types L and R.

12For every pair of types T and I, where T is a cv-qualified or unquali­
  fied  complete  object  type  and I is a promoted integral type, there
  exist
          T*      operator+(T*, I);
          T&      operator[](T*, I);
          T*      operator-(T*, I);
          T*      operator+(I, T*);
          T&      operator[](I, T*);

13For every triple T, CV1, CV2), where T is a complete object type,  and
  CV1 and CV2 are cv-qualifier-seqs, there exists
          ptrdiff_t operator-(CV1 T*, CV2 T*);

14For  every  triple  T, CV1, CV2), where T is any type, and CV1 and CV2
  are cv-qualifier-seqs, there exist
          bool    operator<(CV1 T*, CV2 T*);
          bool    operator>(CV1 T*, CV2 T*);
          bool    operator<=(CV1 T*, CV2 T*);
          bool    operator>=(CV1 T*, CV2 T*);
          bool    operator==(CV1 T*, CV2 T*);
          bool    operator!=(CV1 T*, CV2 T*);

15For every quadruple C, T, CV1, CV2), where C is a class type, T is any
  type, and CV1 and CV2 are cv-qualifier-seqs, there exist
          bool    operator==(CV1 T C::*, CV2 T C::*);
          bool    operator!=(CV1 T C::*, CV2 T C::*);

16For every pair of promoted integral types L and R, there exist
          LR      operator%(L, R);
          LR      operator&(L, R);
          LR      operator^(L, R);
          LR      operator|(L, R);
          L       operator<<(L, R);
          L       operator>>(L, R);
  where  LR  is  the  result of the usual arithmetic conversions between
  types L and R.

17For every triple L, VQ, R), where L  is  an  arithmetic  type,  VQ  is
  either  volatile  or empty, and R is a promoted arithmetic type, there
  exist
          VQ L&   operator=(VQ L&, R);
          VQ L&   operator*=(VQ L&, R);
          VQ L&   operator/=(VQ L&, R);
          VQ L&   operator+=(VQ L&, R);
          VQ L&   operator-=(VQ L&, R);

18For every pair T, VQ), where T is any type and VQ is  either  volatile
  or empty, there exists
          T*VQ&   operator=(T*VQ&, T*);

19For  every  triple T, VQ, I), where T is a cv-qualified or unqualified
  complete object type, VQ is either volatile or empty, and I is a  pro­
  moted integral type, there exist
          T*VQ&   operator+=(T*VQ&, I);
          T*VQ&   operator-=(T*VQ&, I);

20For  every triple L, VQ, R), where L is an integral type, VQ is either
  volatile or empty, and R is a promoted integral type, there exist

          VQ L&   operator%=(VQ L&, R);
          VQ L&   operator<<=(VQ L&, R);
          VQ L&   operator>>=(VQ L&, R);
          VQ L&   operator&=(VQ L&, R);
          VQ L&   operator^=(VQ L&, R);
          VQ L&   operator|=(VQ L&, R);

21For every pair of types L and R, there exists
          R       operator,(L, R);

22There also exist
          bool    operator!(bool);
          bool    operator&&(bool, bool);
          bool    operator||(bool, bool);