Document Number:N4920
Date:
Revises:N4908
Editor: Thomas Köppe
Google DeepMind
tkoeppe@google.com

Working Draft, C++ Extensions for Library Fundamentals, Version 3

Note: this is an early draft. It’s known to be incomplet and incorrekt, and it has lots of bad formatting.

1

General

[general]
1.1

Scope

[general.scope]

This technical specification describes extensions to the C++ Standard Library (1.2). These extensions are classes and functions that are likely to be used widely within a program and/or on the interface boundaries between libraries written by different organizations.

This technical specification is non-normative. Some of the library components in this technical specification may be considered for standardization in a future version of C++, but they are not currently part of any C++ standard. Some of the components in this technical specification may never be standardized, and others may be standardized in a substantially changed form.

The goal of this technical specification is to build more widespread existing practice for an expanded C++ standard library. It gives advice on extensions to those vendors who wish to provide them.

1.2

Normative references

[general.references]

The following referenced document is indispensable for the application of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

  • ISO/IEC 14882:2020, Programming Languages — C++

ISO/IEC 14882:2020 is herein called the C++ Standard. References to clauses within the C++ Standard are written as "C++20 §3.2". The library described in ISO/IEC 14882:2020 clauses 16–32 is herein called the C++ Standard Library.

Unless otherwise specified, the whole of the C++ Standard's Library introduction (C++20 §16) is included into this Technical Specification by reference.

1.3

Namespaces, headers, and modifications to standard classes

[general.namespaces]

Since the extensions described in this technical specification are experimental and not part of the C++ standard library, they should not be declared directly within namespace std. Unless otherwise specified, all components described in this technical specification either:

  • modify an existing interface in the C++ Standard Library in-place,
  • are declared in a namespace whose name appends ::experimental::fundamentals_v3 to a namespace defined in the C++ Standard Library, such as std or std::chrono, or
  • are declared in a subnamespace of a namespace described in the previous bullet, whose name is not the same as an existing subnamespace of namespace std.
[ Example: This TS does not define std::experimental::fundamentals_v3::pmr because the C++ Standard Library defines std::pmr. end example ]

Each header described in this technical specification shall import the contents of std::experimental::fundamentals_v3 into std::experimental as if by

namespace std::experimental::inline fundamentals_v3 {}

This technical specification also describes some experimental modifications to existing interfaces in the C++ Standard Library. These modifications are described by quoting the affected parts of the standard and using underlining to represent added text and strike-through to represent deleted text.

Unless otherwise specified, references to other entities described in this technical specification are assumed to be qualified with std::experimental::fundamentals_v3::, and references to entities described in the standard are assumed to be qualified with std::.

Extensions that are expected to eventually be added to an existing header <meow> are provided inside the <experimental/meow> header, which shall include the standard contents of <meow> as if by

#include <meow>

New headers are also provided in the <experimental/> directory, but without such an #include.

1.4

Future plans (Informative)

[general.plans]

This section describes tentative plans for future versions of this technical specification and plans for moving content into future versions of the C++ Standard.

The C++ committee intends to release new versions of this technical specification periodically, containing the library extensions we hope to add to a near-future version of the C++ Standard. Future versions will define their contents in std::experimental::fundamentals_v4, std::experimental::fundamentals_v5, etc., with the most recent implemented version inlined into std::experimental.

When an extension defined in this or a future version of this technical specification represents enough existing practice, it will be moved into the next version of the C++ Standard by removing the experimental::fundamentals_vN segment of its namespace and by removing the experimental/ prefix from its header's path.

1.5

Feature-testing recommendations (Informative)

[general.feature.test]

For the sake of improved portability between partial implementations of various C++ standards, WG21 (the ISO technical committee for the C++ programming language) recommends that implementers and programmers follow the guidelines in this section concerning feature-test macros. [ Note: WG21's SD-6 makes similar recommendations for the C++ Standard itself. end note ]

Implementers who provide a new standard feature should define a macro with the recommended name, in the same circumstances under which the feature is available (for example, taking into account relevant command-line options), to indicate the presence of support for that feature. Implementers should define that macro with the value specified in the most recent version of this technical specification that they have implemented. The recommended macro name is "__cpp_lib_experimental_" followed by the string in the "Macro Name Suffix" column.

Programmers who wish to determine whether a feature is available in an implementation should base that determination on the presence of the header (determined with __has_include(<header/name>)) and the state of the macro with the recommended name. (The absence of a tested feature may result in a program with decreased functionality, or the relevant functionality may be provided in a different way. A program that strictly depends on support for a feature can just try to use the feature unconditionally; presumably, on an implementation lacking necessary support, translation will fail.)

Table 2 — Significant features in this technical specification
Doc. No. Title Primary Section Macro Name Suffix Value Header
N4388 A Proposal to Add a Const-Propagating Wrapper to the Standard Library 3.2 propagate_const 201505 <experimental/propagate_const>
P0052R10 Generic Scope Guard and RAII Wrapper for the Standard Library 3.3 scope 201902 <experimental/scope>
N3866 Invocation type traits 3.4.2 invocation_type 201406 <experimental/type_traits>
N4502 The C++ Detection Idiom 3.4.3 detect 201505 <experimental/type_traits>
N3916 Type-erased allocator for std::function 4.2 function_erased_allocator 201406 <experimental/functional>, <experimental/utility>
N3916 Polymorphic Memory Resources 5.4 memory_resources 201803 <experimental/memory_resouce>
N4282 The World’s Dumbest Smart Pointer 5.2 observer_ptr 201411 <experimental/memory>
N4257 Delimited iterators 6.2 ostream_joiner 201411 <experimental/iterator>
N3916 Type-erased allocator for std::promise 7.2 promise_erased_allocator 201406 <experimental/future>, <experimental/utility>
N3916 Type-erased allocator for std::packaged_task 7.3 packaged_task_erased_allocator 201406 <experimental/future>, <experimental/utility>
N3925 A sample Proposal 8.2 sample 201402 <experimental/algorithm>
N4531 std::rand replacement 9.1.2 randint 201511 <experimental/random>
2

Modifications to the C++ Standard Library

[mods]

Implementations that conform to this technical specification shall behave as if the modifications contained in this section are made to the C++ Standard.

2.1

Exception Requirements

[mods.exception.requirements]

The following changes to the library introduction allow the destructor of scope_success to throw exceptions.

16.5.4.8 Other functions [res.on.functions]

In certain cases […]

In particular, the effects are undefined in the following cases:

  • […]
  • if any replacement function or handler function or destructor operation exits via an exception, unless specifically allowed in the applicable Required behavior: or Throws: paragraph.
  • if an incomplete type (6.9) is used as a template argument when instantiating a template component, unless specifically allowed for that component.

16.5.5.13 Restrictions on exception handling [res.on.exception.handling]

[…]

Functions from the C standard library shall not throw exceptions181 except when such a function calls a program-supplied function that throws an exception.182

Unless otherwise specified, destructorDestructor operations defined in the C++ standard library shall not throw exceptions. Every destructor without an exception specification in the C++ standard library shall behave as if it had a non-throwing exception specification.

Functions defined in the C++ standard library […]

2.2

Uses-allocator construction

[mods.allocator.uses]

The following changes to the uses_allocator trait and to the description of uses-allocator construction allow a memory_resource pointer act as an allocator in many circumstances. [ Note: Existing programs that use standard allocators would be unaffected by this change. end note ]

20.10.8 uses_allocator [allocator.uses]

20.10.8.1 uses_allocator trait [allocator.uses.trait]

template <class T, class Alloc> struct uses_allocator;
Remarks:
Automatically detects whether T has a nested allocator_type that is convertible from Alloc. Meets the Cpp17BinaryTypeTrait requirements (C++20 §20.15.1). The implementation shall provide a definition that is derived from true_type if a type T::allocator_type exists and either is_convertible_v<Alloc, T::allocator_type> != false or T::allocator_type is an alias for std::experimental::erased_type (3.1.2), otherwise it shall be derived from false_type. A program may specialize this template to derive from true_type for a user-defined type T that does not have a nested allocator_type but nonetheless can be constructed with an allocator where either:
  • the first argument of a constructor has type allocator_arg_t and the second argument has type Alloc or
  • the last argument of a constructor has type Alloc.

20.10.8.2 uses-allocator construction [allocator.uses.construction]

Uses-allocator construction with allocator Alloc refers to the construction of an object obj of type T, using constructor arguments v1, v2, ..., vN of types V1, V2, ..., VN, respectively, and an allocator alloc of type Alloc, where Alloc either (1) meets the requirements of an allocator (C++20 §16.5.3.5), or (2) is a pointer type convertible to std::pmr::memory_resource* (C++20 §20.12), according to the following rules:

3

General utilities library

[utilities]
3.1

Utility components

[utility]
3.1.1

Header <experimental/utility> synopsis

[utility.syn]
#include <utility>

namespace std::experimental::inline fundamentals_v3 {

  // 3.1.2, Class erased_type
  struct erased_type { };

} // namespace std::experimental::inline fundamentals_v3
3.1.2

Class erased_type

[utility.erased.type]
struct erased_type { };

The erased_type struct is an empty struct that serves as a placeholder for a type T in situations where the actual type T is determined at runtime. For example, the nested type, allocator_type, is an alias for erased_type in classes that use type-erased allocators (see 5.3).

3.2

Constness propagation

[propagate_const]
3.2.1

Header <experimental/propagate_const> synopsis

[propagate_const.syn]
namespace std {
  namespace experimental::inline fundamentals_v3 {

    // 3.2.2.1, Class template propagate_const overview
    template <class T> class propagate_const;

    // 3.2.2.9, propagate_const relational operators
    template <class T>
      constexpr bool operator==(const propagate_const<T>& pt, nullptr_t);
    template <class T>
      constexpr bool operator==(nullptr_t, const propagate_const<T>& pu);

    template <class T>
      constexpr bool operator!=(const propagate_const<T>& pt, nullptr_t);
    template <class T>
      constexpr bool operator!=(nullptr_t, const propagate_const<T>& pu);

    template <class T, class U>
      constexpr bool operator==(const propagate_const<T>& pt, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator!=(const propagate_const<T>& pt, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator<(const propagate_const<T>& pt, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator>(const propagate_const<T>& pt, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator<=(const propagate_const<T>& pt, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator>=(const propagate_const<T>& pt, const propagate_const<U>& pu);

    template <class T, class U>
      constexpr bool operator==(const propagate_const<T>& pt, const U& u);
    template <class T, class U>
      constexpr bool operator!=(const propagate_const<T>& pt, const U& u);
    template <class T, class U>
      constexpr bool operator<(const propagate_const<T>& pt, const U& u);
    template <class T, class U>
      constexpr bool operator>(const propagate_const<T>& pt, const U& u);
    template <class T, class U>
      constexpr bool operator<=(const propagate_const<T>& pt, const U& u);
    template <class T, class U>
      constexpr bool operator>=(const propagate_const<T>& pt, const U& u);

    template <class T, class U>
      constexpr bool operator==(const T& t, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator!=(const T& t, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator<(const T& t, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator>(const T& t, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator<=(const T& t, const propagate_const<U>& pu);
    template <class T, class U>
      constexpr bool operator>=(const T& t, const propagate_const<U>& pu);

    // 3.2.2.10, propagate_const specialized algorithms
    template <class T>
      constexpr void swap(propagate_const<T>& pt, propagate_const<T>& pt2) noexcept(see below);

    // 3.2.2.11, propagate_const underlying pointer access
    template <class T>
      constexpr const T& get_underlying(const propagate_const<T>& pt) noexcept;
    template <class T>
      constexpr T& get_underlying(propagate_const<T>& pt) noexcept;

  } // namespace experimental::inline fundamentals_v3

  // 3.2.2.12, propagate_const hash support
  template <class T> struct hash;
  template <class T>
    struct hash<experimental::fundamentals_v3::propagate_const<T>>;

  // 3.2.2.13, propagate_const comparison function objects
  template <class T> struct equal_to;
  template <class T>
    struct equal_to<experimental::fundamentals_v3::propagate_const<T>>;
  template <class T> struct not_equal_to;
  template <class T>
    struct not_equal_to<experimental::fundamentals_v3::propagate_const<T>>;
  template <class T> struct less;
  template <class T>
    struct less<experimental::fundamentals_v3::propagate_const<T>>;
  template <class T> struct greater;
  template <class T>
    struct greater<experimental::fundamentals_v3::propagate_const<T>>;
  template <class T> struct less_equal;
  template <class T>
    struct less_equal<experimental::fundamentals_v3::propagate_const<T>>;
  template <class T> struct greater_equal;
  template <class T>
    struct greater_equal<experimental::fundamentals_v3::propagate_const<T>>;

} // namespace std
3.2.2

Class template propagate_const

[propagate_const]
3.2.2.1

Class template propagate_const overview

[propagate_const.overview]
namespace std::experimental::inline fundamentals_v3 {

  template <class T> class propagate_const {
  public:
    using element_type = remove_reference_t<decltype(*declval<T&>())>;

    // 3.2.2.4, propagate_const constructors
    constexpr propagate_const() = default;
    propagate_const(const propagate_const& p) = delete;
    constexpr propagate_const(propagate_const&& p) = default;
    template <class U>
      explicit(!is_convertible_v<U, T>) constexpr propagate_const(propagate_const<U>&& pu);
    template <class U>
      explicit(!is_convertible_v<U, T>) constexpr propagate_const(U&& u);

    // 3.2.2.5, propagate_const assignment
    propagate_const& operator=(const propagate_const& p) = delete;
    constexpr propagate_const& operator=(propagate_const&& p) = default;
    template <class U>
      constexpr propagate_const& operator=(propagate_const<U>&& pu);
    template <class U>
      constexpr propagate_const& operator=(U&& u);

    // 3.2.2.6, propagate_const const observers
    explicit constexpr operator bool() const;
    constexpr const element_type* operator->() const;
    constexpr operator const element_type*() const; // Not always defined
    constexpr const element_type& operator*() const;
    constexpr const element_type* get() const;

    // 3.2.2.7, propagate_const non-const observers
    constexpr element_type* operator->();
    constexpr operator element_type*(); // Not always defined
    constexpr element_type& operator*();
    constexpr element_type* get();

    // 3.2.2.8, propagate_const modifiers
    constexpr void swap(propagate_const& pt) noexcept(is_nothrow_swappable<T>);

  private:
    T t_; //exposition only
  };

} // namespace std::experimental::inline fundamentals_v3

propagate_const is a wrapper around a pointer-like object type T which treats the wrapped pointer as a pointer to const when the wrapper is accessed through a const access path.

3.2.2.2

propagate_const general requirements on T

[propagate_const.requirements]

T shall be an object pointer type or a class type for which decltype(*declval<T&>()) is an lvalue reference; otherwise the program is ill-formed.

If T is an array type, reference type, pointer to function type or pointer to (possibly cv-qualified) void, then the program is ill-formed.

[ Note: propagate_const<const int*> is well-formed end note ]

3.2.2.3

propagate_const requirements on class type T

[propagate_const.class_type_requirements]

If T is class type then it shall satisfy the following requirements. In this subclause t denotes a non-const lvalue of type T, ct is a const T& bound to t, element_type denotes an object type.

T and const T shall be contextually convertible to bool.

If T is implicitly convertible to element_type*, (element_type*)t == t.get() shall be true.

If const T is implicitly convertible to const element_type*, (const element_type*)ct == ct.get() shall be true.

Table 3 — Requirements on class types T
Expression Return type Pre-conditions Operational semantics
t.get() element_type*
ct.get() const element_type* or element_type* t.get() == ct.get().
*t element_type& t.get() != nullptr *t refers to the same object as *(t.get())
*ct const element_type& or element_type& ct.get() != nullptr *ct refers to the same object as *(ct.get())
t.operator->() element_type* t.get() != nullptr t.operator->() == t.get()
ct.operator->() const element_type* or element_type* ct.get() != nullptr ct.operator->() == ct.get()
(bool)t bool (bool)t is equivalent to t.get() != nullptr
(bool)ct bool (bool)ct is equivalent to ct.get() != nullptr
3.2.2.4

propagate_const constructors

[propagate_const.ctor]
template <class U>
explicit(!is_convertible_v<U, T>) constexpr propagate_const(propagate_const<U>&& pu);
Constraints:
is_constructible_v<T, U> is true.
Effects:
Initializes t_ as if direct-non-list-initializing an object of type T with the expression std::move(pu.t_).
template <class U>
explicit(!is_convertible_v<U, T>) constexpr propagate_const(U&& u);
Constraints:
is_constructible_v<T, U> is true and decay_t<U> is not a specialization of propagate_const.
Effects:
Initializes t_ as if direct-non-list-initializing an object of type T with the expression std::forward<U>(u).
3.2.2.5

propagate_const assignment

[propagate_const.assignment]
template <class U>
constexpr propagate_const& operator=(propagate_const<U>&& pu);
Constraints:
U is implicitly convertible to T.
Effects:
t_ = std::move(pu.t_).
Returns:
*this.
template <class U>
constexpr propagate_const& operator=(U&& u);
Constraints:
U is implicitly convertible to T and decay_t<U> is not a specialization of propagate_const.
Effects:
t_ = std::forward<U>(u).
Returns:
*this.
3.2.2.6

propagate_const const observers

[propagate_const.const_observers]
explicit constexpr operator bool() const;
Returns:
(bool)t_.
constexpr const element_type* operator->() const;
Preconditions:
get() != nullptr.
Returns:
get().
constexpr operator const element_type*() const;
Constraints:
T is an object pointer type or has an implicit conversion to const element_type*.
Returns:
get().
constexpr const element_type& operator*() const;
Preconditions:
get() != nullptr.
Returns:
*get().
constexpr const element_type* get() const;
Returns:
t_ if T is an object pointer type, otherwise t_.get().
3.2.2.7

propagate_const non-const observers

[propagate_const.non_const_observers]
constexpr element_type* operator->();
Preconditions:
get() != nullptr.
Returns:
get().
constexpr operator element_type*();
Constraints:
T is an object pointer type or has an implicit conversion to element_type*.
Returns:
get().
constexpr element_type& operator*();
Preconditions:
get() != nullptr.
Returns:
*get().
constexpr element_type* get();
Returns:
t_ if T is an object pointer type, otherwise t_.get().
3.2.2.8

propagate_const modifiers

[propagate_const.modifiers]
constexpr void swap(propagate_const& pt) noexcept(is_nothrow_swappable<T>);
Preconditions:
Lvalues of type T are swappable (C++20 §16.5.3.2).
Effects:
swap(t_, pt.t_).
3.2.2.9

propagate_const relational operators

[propagate_const.relational]
template <class T>
constexpr bool operator==(const propagate_const<T>& pt, nullptr_t);
Returns:
pt.t_ == nullptr.
template <class T>
constexpr bool operator==(nullptr_t, const propagate_const<T>& pt);
Returns:
nullptr == pt.t_.
template <class T>
constexpr bool operator!=(const propagate_const<T>& pt, nullptr_t);
Returns:
pt.t_ != nullptr.
template <class T>
constexpr bool operator!=(nullptr_t, const propagate_const<T>& pt);
Returns:
nullptr != pt.t_.
template <class T, class U>
constexpr bool operator==(const propagate_const<T>& pt, const propagate_const<U>& pu);
Returns:
pt.t_ == pu.t_.
template <class T, class U>
constexpr bool operator!=(const propagate_const<T>& pt, const propagate_const<U>& pu);
Returns:
pt.t_ != pu.t_.
template <class T, class U>
constexpr bool operator<(const propagate_const<T>& pt, const propagate_const<U>& pu);
Returns:
pt.t_ < pu.t_.
template <class T, class U>
constexpr bool operator>(const propagate_const<T>& pt, const propagate_const<U>& pu);
Returns:
pt.t_ > pu.t_.
template <class T, class U>
constexpr bool operator<=(const propagate_const<T>& pt, const propagate_const<U>& pu);
Returns:
pt.t_ <= pu.t_.
template <class T, class U>
constexpr bool operator>=(const propagate_const<T>& pt, const propagate_const<U>& pu);
Returns:
pt.t_ >= pu.t_.
template <class T, class U>
constexpr bool operator==(const propagate_const<T>& pt, const U& u);
Returns:
pt.t_ == u.
template <class T, class U>
constexpr bool operator!=(const propagate_const<T>& pt, const U& u);
Returns:
pt.t_ != u.
template <class T, class U>
constexpr bool operator<(const propagate_const<T>& pt, const U& u);
Returns:
pt.t_ < u.
template <class T, class U>
constexpr bool operator>(const propagate_const<T>& pt, const U& u);
Returns:
pt.t_ > u.
template <class T, class U>
constexpr bool operator<=(const propagate_const<T>& pt, const U& u);
Returns:
pt.t_ <= u.
template <class T, class U>
constexpr bool operator>=(const propagate_const<T>& pt, const U& u);
Returns:
pt.t_ >= u.
template <class T, class U>
constexpr bool operator==(const T& t, const propagate_const<U>& pu);
Returns:
t == pu.t_.
template <class T, class U>
constexpr bool operator!=(const T& t, const propagate_const<U>& pu);
Returns:
t != pu.t_.
template <class T, class U>
constexpr bool operator<(const T& t, const propagate_const<U>& pu);
Returns:
t < pu.t_.
template <class T, class U>
constexpr bool operator>(const T& t, const propagate_const<U>& pu);
Returns:
t > pu.t_.
template <class T, class U>
constexpr bool operator<=(const T& t, const propagate_const<U>& pu);
Returns:
t <= pu.t_.
template <class T, class U>
constexpr bool operator>=(const T& t, const propagate_const<U>& pu);
Returns:
t >= pu.t_.
3.2.2.10

propagate_const specialized algorithms

[propagate_const.algorithms]
template <class T>
constexpr void swap(propagate_const<T>& pt1, propagate_const<T>& pt2) noexcept(see below);
Constraints:
is_swappable_v<T> is true.
Effects:
Equivalent to: pt1.swap(pt2).
Remarks:
The expression inside noexcept is equivalent to:
noexcept(pt1.swap(pt2))
3.2.2.11

propagate_const underlying pointer access

[propagate_const.underlying]

Access to the underlying object pointer type is through free functions rather than member functions. These functions are intended to resemble cast operations to encourage caution when using them.

template <class T>
constexpr const T& get_underlying(const propagate_const<T>& pt) noexcept;
Returns:
a reference to the underlying object pointer type.
template <class T>
constexpr T& get_underlying(propagate_const<T>& pt) noexcept;
Returns:
a reference to the underlying object pointer type.
3.2.2.12

propagate_const hash support

[propagate_const.hash]
template <class T>
struct hash<experimental::fundamentals_v3::propagate_const<T>>;

The specialization hash<experimental::fundamentals_v3::propagate_const<T>> is enabled (C++20 §20.14.18) if and only if hash<T> is enabled. When enabled, for an object p of type propagate_const<T>, hash<experimental::fundamentals_v3::propagate_const<T>>()(p) evaluates to the same value as hash<T>()(p.t_).

3.2.2.13

propagate_const comparison function objects

[propagate_const.comparison_function_objects]
template <class T>
struct equal_to<experimental::fundamentals_v3::propagate_const<T>>;

For objects p, q of type propagate_const<T>, equal_to<experimental::fundamentals_v3::propagate_const<T>>()(p, q) shall evaluate to the same value as equal_to<T>()(p.t_, q.t_).

Mandates:
The specialization equal_to<T> is well-formed
Preconditions:
The specialization equal_to<T> is well-defined.
template <class T>
struct not_equal_to<experimental::fundamentals_v3::propagate_const<T>>;

For objects p, q of type propagate_const<T>, not_equal_to<experimental::fundamentals_v3::propagate_const<T>>()(p, q) shall evaluate to the same value as not_equal_to<T>()(p.t_, q.t_).

Mandates:
The specialization not_equal_to<T> is well-formed
Preconditions:
The specialization not_equal_to<T> is well-defined.
template <class T>
struct less<experimental::fundamentals_v3::propagate_const<T>>;

For objects p, q of type propagate_const<T>, less<experimental::fundamentals_v3::propagate_const<T>>()(p, q) shall evaluate to the same value as less<T>()(p.t_, q.t_).

Mandates:
The specialization less<T> is well-formed
Preconditions:
The specialization less<T> is well-defined.
template <class T>
struct greater<experimental::fundamentals_v3::propagate_const<T>>;

For objects p, q of type propagate_const<T>, greater<experimental::fundamentals_v3::propagate_const<T>>()(p, q) shall evaluate to the same value as greater<T>()(p.t_, q.t_).

Mandates:
The specialization greater<T> is well-formed
Preconditions:
The specialization greater<T> is well-defined.
template <class T>
struct less_equal<experimental::fundamentals_v3::propagate_const<T>>;

For objects p, q of type propagate_const<T>, less_equal<experimental::fundamentals_v3::propagate_const<T>>()(p, q) shall evaluate to the same value as less_equal<T>()(p.t_, q.t_).

Mandates:
The specialization less_equal<T> is well-formed
Preconditions:
The specialization less_equal<T> is well-defined.
template <class T>
struct greater_equal<experimental::fundamentals_v3::propagate_const<T>>;

For objects p, q of type propagate_const<T>, greater_equal<experimental::fundamentals_v3::propagate_const<T>>()(p, q) shall evaluate to the same value as greater_equal<T>()(p.t_, q.t_).

Mandates:
The specialization greater_equal<T> is well-formed
Preconditions:
The specialization greater_equal<T> is well-defined.
3.3

Scope guard support

[scopeguard]
3.3.1

Header <experimental/scope> synopsis

[scope.syn]
namespace std::experimental::inline fundamentals_v3 {

  // 3.3.2, Class templates scope_exit, scope_fail, and scope_success
  template <class EF>
    class scope_exit;
  template <class EF>
    class scope_fail;
  template <class EF>
    class scope_success;

  // 3.3.3, Class template unique_resource
  template <class R, class D>
    class unique_resource;

  // 3.3.3.6, unique_resource creation
  template <class R, class D, class S=decay_t<R>>
    unique_resource<decay_t<R>, decay_t<D>>
      make_unique_resource_checked(R&& r, const S& invalid, D&& d) noexcept(see below);

} // namespace std::experimental::inline fundamentals_v3
3.3.2

Class templates scope_exit, scope_fail, and scope_success

[scopeguard.exit]

The class templates scope_exit, scope_fail, and scope_success define scope guards that wrap a function object to be called on their destruction.

In this subclause, the placeholder scope-guard denotes each of these class templates. In descriptions of the class members, scope-guard refers to the enclosing class.

namespace std::experimental::inline fundamentals_v3 {

  template <class EF> class scope-guard {
  public:
    template <class EFP>
      explicit scope-guard(EFP&& f) noexcept(see below);
    scope-guard(scope-guard&& rhs) noexcept(see below);

    scope-guard(const scope-guard&) = delete;
    scope-guard& operator=(const scope-guard&) = delete;
    scope-guard& operator=(scope-guard&&) = delete;

    ~scope-guard () noexcept(see below);

    void release() noexcept;

  private:
    EF exit_function;                                 // exposition only
    bool execute_on_destruction{true};                // exposition only
    int uncaught_on_creation{uncaught_exceptions()};  // exposition only
  };

  template <class EF>
    scope-guard(EF) -> scope-guard<EF>;

}  // namespace std::experimental::inline fundamentals_v3

The class template scope_exit is a general-purpose scope guard that calls its exit function when a scope is exited. The class templates scope_fail and scope_success share the scope_exit interface, only the situation when the exit function is called differs.

[ Example:
void grow(vector<int>& v) {
  scope_success guard([]{ cout << "Good!" << endl; });
  v.resize(1024);
}
end example ]

[ Note: If the exit function object of a scope_success or scope_exit object refers to a local variable of the function where it is defined, e.g., as a lambda capturing the variable by reference, and that variable is used as a return operand in that function, that variable might have already been returned when the scope-guard’s destructor executes, calling the exit function. This can lead to surprising behavior. end note ]

Template argument EF shall be a function object type (C++20 §20.14), lvalue reference to function, or lvalue reference to function object type. If EF is an object type, it shall meet the Cpp17Destructible requirements (C++20 Table 30). Given an lvalue g of type remove_reference_t<EF>, the expression g() shall be well-formed.

The constructor parameter f in the following constructors shall be a reference to a function or a reference to a function object (C++20 §20.14).

template <class EFP>
explicit scope-guard(EFP&& f) noexcept(
    is_nothrow_constructible_v<EF, EFP> ||
    is_nothrow_constructible_v<EF, EFP&>);
Constraints:
is_same_v<remove_cvref_t<EFP>, scope-guard> is false and is_constructible_v<EF, EFP> is true.
Mandates:
The expression f() is well-formed.
Preconditions:
Calling f() has well-defined behavior. For scope_exit and scope_fail, calling f() does not throw an exception.
Effects:
If EFP is not an lvalue reference type and is_nothrow_constructible_v<EF, EFP> is true, initialize exit_function with std::forward<EFP>(f); otherwise initialize exit_function with f. For scope_exit and scope_fail, if the initialization of exit_function throws an exception, calls f(). [ Note: For scope_success, f() will not be called if the initialization fails. end note ]
Throws:
Any exception thrown during the initialization of exit_function.
scope-guard(scope-guard&& rhs) noexcept(see below)
Constraints:
(is_nothrow_move_constructible_v<EF> || is_copy_constructible_v<EF>) is true.
Preconditions:
If EF is an object type:
  • if is_nothrow_move_constructible_v<EF> is true, EF meets the Cpp17MoveConstructible requirements (C++20 Table 26),
  • otherwise EF meets the Cpp17CopyConstructible requirements (C++20 Table 27).
Effects:
If is_nothrow_move_constructible_v<EF> is true, initializes exit_function with std::forward<EF>(rhs.exit_function), otherwise initializes exit_function with rhs.exit_function. Initializes execute_on_destruction from rhs.execute_on_destruction and uncaught_on_creation from rhs.uncaught_on_creation. If construction succeeds, call rhs.release(). [ Note: Copying instead of moving provides the strong exception guarantee. end note ]
Postconditions:
execute_on_destruction yields the value rhs.execute_on_destruction yielded before the construction. uncaught_on_creation yields the value rhs.uncaught_on_creation yielded before the construction.
Throws:
Any exception thrown during the initialization of exit_function.
Remarks:
The expression inside noexcept is equivalent to:
is_nothrow_move_constructible_v<EF> || is_nothrow_copy_constructible_v<EF>
~scope_exit() noexcept(true);
Effects:
Equivalent to:
if (execute_on_destruction)
  exit_function();
~scope_fail() noexcept(true);
Effects:
Equivalent to:
if (execute_on_destruction && uncaught_exceptions() > uncaught_on_creation)
  exit_function();
~scope_success() noexcept(noexcept(exit_function()));
Effects:
Equivalent to:
if (execute_on_destruction && uncaught_exceptions() <= uncaught_on_creation)
  exit_function();
[ Note: If noexcept(exit_function()) is false, exit_function() may throw an exception, notwithstanding the restrictions of C++20 §16.5.5.13. end note ]
Throws:
Any exception thrown by exit_function().
void release() noexcept;
Effects:
Equivalent to execute_on_destruction = false.
3.3.3

Class template unique_resource

[scopeguard.uniqueres]
3.3.3.1

Overview

[scopeguard.uniqueres.overview]
namespace std::experimental::inline fundamentals_v3 {

  template <class R, class D> class unique_resource {
  public:
    // 3.3.3.2, Constructors
    unique_resource();
    template <class RR, class DD>
      unique_resource(RR&& r, DD&& d) noexcept(see below);
    unique_resource(unique_resource&& rhs) noexcept(see below);

    // 3.3.3.3, Destructor
    ~unique_resource();

    // 3.3.3.4, Assignment
    unique_resource& operator=(unique_resource&& rhs) noexcept(see below);

    // 3.3.3.5, Other member functions
    void reset() noexcept;
    template <class RR>
      void reset(RR&& r);
    void release() noexcept;
    const R& get() const noexcept;
    see below operator*() const noexcept;
    R operator->() const noexcept;
    const D& get_deleter() const noexcept;

  private:
    using R1 = conditional_t<is_reference_v<R>, reference_wrapper<remove_reference_t<R>>, R>;  // exposition only
    R1 resource;                  // exposition only
    D deleter;                    // exposition only
    bool execute_on_reset{true};  // exposition only
  };

  template<class R, class D>
    unique_resource(R, D) -> unique_resource<R, D>;

}  // namespace std::experimental::inline fundamentals_v3

[ Note: unique_resource is a universal RAII wrapper for resource handles. Typically, such resource handles are of trivial type and come with a factory function and a clean-up or deleter function that do not throw exceptions. The clean-up function together with the result of the creation function is used to create a unique_resource variable, that on destruction will call the clean-up function. Access to the underlying resource handle is achieved through get() and in case of a pointer type resource through a set of convenience pointer operator functions. end note ]

The template argument D shall meet the requirements of a Cpp17Destructible (C++20 Table 30) function object type (C++20 §20.14), for which, given a lvalue d of type D and a lvalue r of type R, the expression d(r) shall be well-formed. D shall either meet the Cpp17CopyConstructible requirements (C++20 Table 27), or D shall meet the Cpp17MoveConstructible requirements (C++20 Table 26) and is_nothrow_move_constructible_v<D> shall be true.

For the purpose of this subclause, a resource type T is an object type that meets the requirements of Cpp17CopyConstructible (C++20 Table 27), or is an object type that meets the requirements of Cpp17MoveConstructible (C++20 Table 26) and is_nothrow_move_constructible_v<T> is true, or is an lvalue reference to a resource type. R shall be a resource type.

For the scope of the adjacent subclauses, let RESOURCE be defined as follows:

  • resource.get() if is_reference_v<R> is true,
  • resource otherwise.

3.3.3.2

Constructors

[scopeguard.uniqueres.ctor]
unique_resource()
Constraints:
is_default_constructible_v<R> && is_default_constructible_v<D> is true.
Effects:
Value-initializes resource and deleter; execute_on_reset is initialized with false.
template <class RR, class DD>
unique_resource(RR&& r, DD&& d) noexcept(see below)
Constraints:
is_constructible_v<R1, RR> &&
is_constructible_v<D , DD> &&
(is_nothrow_constructible_v<R1, RR> || is_constructible_v<R1,RR&>) &&
(is_nothrow_constructible_v<D , DD> || is_constructible_v<D ,DD&>)
is true. [ Note: The first two conditions prohibit initialization from an rvalue reference when either R1 or D is a specialization of reference_wrapper. end note ]
Mandates:
The expressions d(r), d(RESOURCE) and deleter(RESOURCE) are well-formed.
Preconditions:
Calling d(r), d(RESOURCE) or deleter(RESOURCE) has well-defined behavior and does not throw an exception.
Effects:
If is_nothrow_constructible_v<R1, RR> is true, initializes resource with std::forward<RR>(r), otherwise initializes resource with r. Then, if is_nothrow_constructible_v<D, DD> is true, initializes deleter with std::forward<DD>(d), otherwise initializes deleter with d. If initialization of resource throws an exception, calls d(r). If initialization of deleter throws an exception, calls d(RESOURCE). [ Note: The explained mechanism ensures no leaking of resources. end note ]
Throws:
Any exception thrown during initialization of resource or deleter.
Remarks:
The expression inside noexcept is equivalent to:
(is_nothrow_constructible_v<R1, RR> || is_nothrow_constructible_v<R1, RR&>) &&
(is_nothrow_constructible_v<D , DD> || is_nothrow_constructible_v<D , DD&>)
unique_resource(unique_resource&& rhs) noexcept(see below);
Effects:
First, initialize resource as follows:
  • If is_nothrow_move_constructible_v<R1> is true, from std::move(rhs.resource);
  • otherwise, from rhs.resource.
[ Note: If initialization of resource throws an exception, rhs is left owning the resource and will free it in due time. end note ] Then, initialize deleter as follows:
  • If is_nothrow_move_constructible_v<D> is true, from std::move(rhs.deleter);
  • otherwise, from rhs.deleter.
If initialization of deleter throws an exception and is_nothrow_move_constructible_v<R1> is true and rhs.execute_on_reset is true:
rhs.deleter(RESOURCE);
rhs.release();
Finally, execute_on_reset is initialized with exchange(rhs.execute_on_reset, false). [ Note: The explained mechanism ensures no leaking and no double release of resources. end note ]
Remarks:
The expression inside noexcept is equivalent to:
is_nothrow_move_constructible_v<R1> && is_nothrow_move_constructible_v<D>
3.3.3.3

Destructor

[scopeguard.uniqueres.dtor]
~unique_resource();
Effects:
Equivalent to reset().
3.3.3.4

Assignment

[scopeguard.uniqueres.assign]
unique_resource& operator=(unique_resource&& rhs) noexcept(see below);
Preconditions:
If is_nothrow_move_assignable_v<R1> is true, R1 meets the Cpp17MoveAssignable (C++20 Table 28) requirements; otherwise R1 meets the Cpp17CopyAssignable (C++20 Table 29) requirements. If is_nothrow_move_assignable_v<D> is true, D meets the Cpp17MoveAssignable (C++20 Table 28) requirements; otherwise D meets the Cpp17CopyAssignable (C++20 Table 29) requirements.
Effects:
Equivalent to:
reset();
if constexpr (is_nothrow_move_assignable_v<R1>) {
  if constexpr (is_nothrow_move_assignable_v<D>) {
    resource = std::move(rhs.resource);
    deleter = std::move(rhs.deleter);
  } else {
    deleter = rhs.deleter;
    resource = std::move(rhs.resource);
  }
} else {
  if constexpr (is_nothrow_move_assignable_v<D>) {
    resource = rhs.resource;
    deleter = std::move(rhs.deleter);
  } else {
    resource = rhs.resource;
    deleter = rhs.deleter;
  }
}
execute_on_reset = exchange(rhs.execute_on_reset, false);
[ Note: If a copy of a member throws an exception, this mechanism leaves rhs intact and *this in the released state. end note ]
Returns:
*this.
Throws:
Any exception thrown during a copy-assignment of a member that cannot be moved without an exception.
Remarks:
The expression inside noexcept is equivalent to:
is_nothrow_move_assignable_v<R1> && is_nothrow_move_assignable_v<D>
3.3.3.5

Other member functions

[scopeguard.uniqueres.members]
void reset() noexcept;
Effects:
Equivalent to:
if (execute_on_reset) {
  execute_on_reset = false;
  deleter(RESOURCE);
}
template <class RR> void reset(RR&& r);
Constraints:
the selected assignment expression statement assigning resource is well-formed.
Mandates:
The expression deleter(r) is well-formed.
Preconditions:
Calling deleter(r) has well-defined behavior and does not throw an exception.
Effects:
Equivalent to:
reset();
if constexpr (is_nothrow_assignable_v<R1&, RR>) {
  resource = std::forward<RR>(r);
} else {
  resource = as_const(r);
}
execute_on_reset = true;
If copy-assignment of resource throws an exception, calls deleter(r).
void release() noexcept;
Effects:
Equivalent to execute_on_reset = false.
const R& get() const noexcept;
Returns:
resource.
see below operator*() const noexcept;
Constraints:
is_pointer_v<R> is true and is_void_v<remove_pointer_t<R>> is false.
Effects:
Equivalent to: return *get();
Remarks:
The return type is add_lvalue_reference_t<remove_pointer_t<R>>.
R operator->() const noexcept;
Constraints:
is_pointer_v<R> is true.
Returns:
get().
const D& get_deleter() const noexcept;
Returns:
deleter.
3.3.3.6

unique_resource creation

[scopeguard.uniqueres.create]
template <class R, class D, class S=decay_t<R>>
unique_resource<decay_t<R>, decay_t<D>>
  make_unique_resource_checked(R&& resource, const S& invalid, D&& d)
  noexcept(is_nothrow_constructible_v<decay_t<R>, R> &&
           is_nothrow_constructible_v<decay_t<D>, D>);
Mandates:
The expression (resource == invalid ? true : false) is well-formed.
Preconditions:
Evaluation of the expression (resource == invalid ? true : false) has well-defined behavior and does not throw an exception.
Effects:
Returns an object constructed with members initialized from std::forward<R>(resource), std::forward<D>(d), and !bool(resource == invalid). Any failure during construction of the return value will not call d(resource) if bool(resource == invalid) is true.

[ Note: This creation function exists to avoid calling a deleter function with an invalid argument. end note ]

[ Example: The following example shows its use to avoid calling fclose when fopen fails.
auto file = make_unique_resource_checked(
    ::fopen("potentially_nonexistent_file.txt", "r"),
    nullptr,
    [](auto fptr){ ::fclose(fptr); });
end example ]
3.4

Metaprogramming and type traits

[meta]
3.4.1

Header <experimental/type_traits> synopsis

[meta.type.syn]
#include <type_traits>

namespace std::experimental::inline fundamentals_v3 {

  // 3.4.2, Other type transformations
  template <class> class invocation_type; // not defined
  template <class F, class... ArgTypes> class invocation_type<F(ArgTypes...)>;
  template <class> class raw_invocation_type; // not defined
  template <class F, class... ArgTypes> class raw_invocation_type<F(ArgTypes...)>;

  template <class T>
    using invocation_type_t = typename invocation_type<T>::type;
  template <class T>
    using raw_invocation_type_t = typename raw_invocation_type<T>::type;

  // 3.4.3, Detection idiom
  struct nonesuch;

  template <template<class...> class Op, class... Args>
    using is_detected = see below;
  template <template<class...> class Op, class... Args>
    inline constexpr bool is_detected_v
      = is_detected<Op, Args...>::value;
  template <template<class...> class Op, class... Args>
    using detected_t = see below;
  template <class Default, template<class...> class Op, class... Args>
    using detected_or = see below;
  template <class Default, template<class...> class Op, class... Args>
    using detected_or_t = typename detected_or<Default, Op, Args...>::type;
  template <class Expected, template<class...> class Op, class... Args>
    using is_detected_exact = is_same<Expected, detected_t<Op, Args...>>;
  template <class Expected, template<class...> class Op, class... Args>
    inline constexpr bool is_detected_exact_v
      = is_detected_exact<Expected, Op, Args...>::value;
  template <class To, template<class...> class Op, class... Args>
    using is_detected_convertible = is_convertible<detected_t<Op, Args...>, To>;
  template <class To, template<class...> class Op, class... Args>
    inline constexpr bool is_detected_convertible_v
      = is_detected_convertible<To, Op, Args...>::value;

} // namespace std::experimental::inline fundamentals_v3
3.4.2

Other type transformations

[meta.trans.other]

This subclause contains templates that may be used to transform one type to another following some predefined rule.

Each of the templates in this subclause shall be a TransformationTrait (C++20 §20.15.1).

Within this section, define the invocation parameters of INVOKE(f, t1, t2, ..., tN) as follows, in which T1 is the possibly cv-qualified type of t1 and U1 denotes T1& if t1 is an lvalue or T1&& if t1 is an rvalue:

  • When f is a pointer to a member function of a class T the invocation parameters are U1 followed by the parameters of f matched by t2, ..., tN.
  • When N == 1 and f is a pointer to member data of a class T the invocation parameter is U1.
  • If f is a class object, the invocation parameters are the parameters matching t1, ..., tN of the best viable function (C++20 §12.4.3) for the arguments t1, ..., tN among the function call operators and surrogate call functions of f.
  • In all other cases, the invocation parameters are the parameters of f matching t1, ... tN.

In all of the above cases, if an argument tI matches the ellipsis in the function's parameter-declaration-clause, the corresponding invocation parameter is defined to be the result of applying the default argument promotions (C++20 §7.6.1.2) to tI.

[ Example: Assume S is defined as
struct S {
  int f(double const &) const;
  void operator()(int, int);
  void operator()(char const *, int i = 2, int j = 3);
  void operator()(...);
};
  • The invocation parameters of INVOKE(&S::f, S(), 3.5) are (S &&, double const &).
  • The invocation parameters of INVOKE(S(), 1, 2) are (int, int).
  • The invocation parameters of INVOKE(S(), "abc", 5) are (const char *, int). The defaulted parameter j does not correspond to an argument.
  • The invocation parameters of INVOKE(S(), locale(), 5) are (locale, int). Arguments corresponding to ellipsis maintain their types.
end example ]
Table 4 — Other type transformations
TemplateConditionComments
template <class Fn, class... ArgTypes>
struct raw_invocation_type<Fn(ArgTypes...)>;
Fn and all types in the parameter pack ArgTypes shall be complete types, (possibly cv-qualified) void, or arrays of unknown bound. see below
template <class Fn, class... ArgTypes>
struct invocation_type<Fn(ArgTypes...)>;
Fn and all types in the parameter pack ArgTypes shall be complete types, (possibly cv-qualified) void, or arrays of unknown bound. see below

Access checking is performed as if in a context unrelated to Fn and ArgTypes. Only the validity of the immediate context of the expression is considered. [ Note: The compilation of the expression can result in side effects such as the instantiation of class template specializations and function template specializations, the generation of implicitly-defined functions, and so on. Such side effects are not in the "immediate context" and can result in the program being ill-formed. end note ]

The member raw_invocation_type<Fn(ArgTypes...)>::type shall be defined as follows. If the expression INVOKE(declval<Fn>(), declval<ArgTypes>()...) is ill-formed when treated as an unevaluated operand (C++20 §7), there shall be no member type. Otherwise:

  • Let R denote result_of_t<Fn(ArgTypes...)>.
  • Let the types Ti be the invocation parameters of INVOKE(declval<Fn>(), declval<ArgTypes>()...).
  • Then the member type shall name the function type R(T1, T2, ...).

The member invocation_type<Fn(ArgTypes...)>::type shall be defined as follows. If raw_invocation_type<Fn(ArgTypes...)>::type does not exist, there shall be no member type. Otherwise:

  • Let A1, A2, … denote ArgTypes...
  • Let R(T1, T2, …) denote raw_invocation_type_t<Fn(ArgTypes...)>
  • Then the member type shall name the function type R(U1, U2, …) where Ui is decay_t<Ai> if declval<Ai>() is an rvalue otherwise Ti.
3.4.3

Detection idiom

[meta.detect]
struct nonesuch {
  ~nonesuch() = delete;
  nonesuch(nonesuch const&) = delete;
  void operator=(nonesuch const&) = delete;
};

nonesuch has no default constructor (C++20 §11.4.4) or initializer-list constructor (C++20 §9.4.4), and is not an aggregate (C++20 §9.4.1).

template <class Default, class AlwaysVoid,
          template<class...> class Op, class... Args>
struct DETECTOR { // exposition only
  using value_t = false_type;
  using type = Default;
};

template <class Default, template<class...> class Op, class... Args>
struct DETECTOR<Default, void_t<Op<Args...>>, Op, Args...> { // exposition only
  using value_t = true_type;
  using type = Op<Args...>;
};

template <template<class...> class Op, class... Args>
  using is_detected = typename DETECTOR<nonesuch, void, Op, Args...>::value_t;

template <template<class...> class Op, class... Args>
  using detected_t = typename DETECTOR<nonesuch, void, Op, Args...>::type;

template <class Default, template<class...> class Op, class... Args>
  using detected_or = DETECTOR<Default, void, Op, Args...>;
[ Example:
// archetypal helper alias for a copy assignment operation:
template <class T>
  using copy_assign_t = decltype(declval<T&>() = declval<T const &>());

// plausible implementation for the is_assignable type trait:
template <class T>
  using is_copy_assignable = is_detected<copy_assign_t, T>;

// plausible implementation for an augmented is_assignable type trait
// that also checks the return type:
template <class T>
  using is_canonical_copy_assignable = is_detected_exact<T&, copy_assign_t, T>;
end example ]
[ Example:
// archetypal helper alias for a particular type member:
template <class T>
  using diff_t = typename T::difference_type;

// alias the type member, if it exists, otherwise alias ptrdiff_t:
template <class Ptr>
  using difference_type = detected_or_t<ptrdiff_t, diff_t, Ptr>;
end example ]
4

Function objects

[func]
4.1

Header <experimental/functional> synopsis

[functional.syn]
#include <functional>

namespace std {
  namespace experimental::inline fundamentals_v3 {

    // 4.2, Class template function
    template<class> class function; // undefined
    template<class R, class... ArgTypes> class function<R(ArgTypes...)>;

    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&);

    template<class R, class... ArgTypes>
    bool operator==(const function<R(ArgTypes...)>&, nullptr_t) noexcept;
    template<class R, class... ArgTypes>
    bool operator==(nullptr_t, const function<R(ArgTypes...)>&) noexcept;
    template<class R, class... ArgTypes>
    bool operator!=(const function<R(ArgTypes...)>&, nullptr_t) noexcept;
    template<class R, class... ArgTypes>
    bool operator!=(nullptr_t, const function<R(ArgTypes...)>&) noexcept;

  } // namespace experimental::inline fundamentals_v3

  template<class R, class... ArgTypes, class Alloc>
  struct uses_allocator<experimental::function<R(ArgTypes...)>, Alloc>;

} // namespace std
4.2

Class template function

[func.wrap.func]
4.2.1

Overview

[func.wrap.func.overview]

The specification of all declarations within subclause 4.2 are the same as the corresponding declarations, as specified in C++20 §20.14.16.2, unless explicitly specified otherwise. [ Note: std::experimental::function uses std::bad_function_call, there is no additional type std::experimental::bad_function_call end note ] .

namespace std {
  namespace experimental::inline fundamentals_v3 {

    template<class> class function; // undefined

    template<class R, class... ArgTypes>
    class function<R(ArgTypes...)> {
    public:
      using result_type = R;
      using argument_type = T1;
      using first_argument_type T1;
      using second_argument_type = T2;

      using allocator_type = erased_type;

      function() noexcept;
      function(nullptr_t) noexcept;
      function(const function&);
      function(function&&);
      template<class F> function(F);
      template<class A> function(allocator_arg_t, const A&) noexcept;
      template<class A> function(allocator_arg_t, const A&,
        nullptr_t) noexcept;
      template<class A> function(allocator_arg_t, const A&,
        const function&);
      template<class A> function(allocator_arg_t, const A&,
        function&&);
      template<class F, class A> function(allocator_arg_t, const A&, F);

      function& operator=(const function&);
      function& operator=(function&&);
      function& operator=(nullptr_t) noexcept;
      template<class F> function& operator=(F&&);
      template<class F> function& operator=(reference_wrapper<F>);

      ~function();

      void swap(function&);

      explicit operator bool() const noexcept;

      R operator()(ArgTypes...) const;

      const type_info& target_type() const noexcept;
      template<class T> T* target() noexcept;
      template<class T> const T* target() const noexcept;

      pmr::memory_resource* get_memory_resource() const noexcept;
    };

    template <class R, class... ArgTypes>
    bool operator==(const function<R(ArgTypes...)>&, nullptr_t) noexcept;
    template <class R, class... ArgTypes>
    bool operator==(nullptr_t, const function<R(ArgTypes...)>&) noexcept;

    template <class R, class... ArgTypes>
    bool operator!=(const function<R(ArgTypes...)>&, nullptr_t) noexcept;
    template <class R, class... ArgTypes>
    bool operator!=(nullptr_t, const function<R(ArgTypes...)>&) noexcept;

    template <class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&);

  } // namespace experimental::inline fundamentals_v3

  template <class R, class... ArgTypes, class Alloc>
  struct uses_allocator<experimental::function<R(ArgTypes...)>, Alloc>
    : true_type { };

} // namespace std
4.2.2

function construct/copy/destroy

[func.wrap.func.con]

When a function constructor that takes a first argument of type allocator_arg_t is invoked, the second argument is treated as a type-erased allocator (5.3). If the constructor moves or makes a copy of a function object (C++20 §20.14), including an instance of the experimental::function class template, then that move or copy is performed by using-allocator construction with allocator get_memory_resource().

In the following descriptions, let ALLOCATOR_OF(f) be the allocator specified in the construction of function f, or the value of experimental::pmr::get_default_resource() at the time of the construction of f if no allocator was specified.

function& operator=(const function& f);
Effects:
function(allocator_arg, ALLOCATOR_OF(*this), f).swap(*this);
Returns:
*this.
function& operator=(function&& f);
Effects:
function(allocator_arg, ALLOCATOR_OF(*this), std::move(f)).swap(*this);
Returns:
*this.
function& operator=(nullptr_t) noexcept;
Effects:
If *this != nullptr, destroys the target of this.
Postconditions:
!(*this). The memory resource returned by get_memory_resource() after the assignment is equivalent to the memory resource before the assignment. [ Note: the address returned by get_memory_resource() might change end note ]
Returns:
*this.
template<class F> function& operator=(F&& f);
Constraints:
declval<decay_t<F>&>() is Lvalue-Callable (C++20 §20.14.16.2) for argument types ArgTypes... and return type R.
Effects:
function(allocator_arg, ALLOCATOR_OF(*this), std::forward<F>(f)).swap(*this);
Returns:
*this.
template<class F> function& operator=(reference_wrapper<F> f);
Effects:
function(allocator_arg, ALLOCATOR_OF(*this), f).swap(*this);
Returns:
*this.
4.2.3

function modifiers

[func.wrap.func.mod]
void swap(function& other);
Preconditions:
*this->get_memory_resource() == *other.get_memory_resource().
Effects:
Interchanges the targets of *this and other.
Remarks:
The allocators of *this and other are not interchanged.
5

Memory

[memory]
5.1

Header <experimental/memory> synopsis

[memory.syn]
#include <memory>

namespace std {
  namespace experimental::inline fundamentals_v3 {

    // 5.2, Non-owning (observer) pointers
    template <class W> class observer_ptr;

    // 5.2.6, observer_ptr specialized algorithms
    template <class W>
    void swap(observer_ptr<W>&, observer_ptr<W>&) noexcept;
    template <class W>
    observer_ptr<W> make_observer(W*) noexcept;
    // (in)equality operators
    template <class W1, class W2>
    bool operator==(observer_ptr<W1>, observer_ptr<W2>);

    template <class W1, class W2>
    bool operator!=(observer_ptr<W1>, observer_ptr<W2>);
    template <class W>
    bool operator==(observer_ptr<W>, nullptr_t) noexcept;
    template <class W>
    bool operator!=(observer_ptr<W>, nullptr_t) noexcept;
    template <class W>
    bool operator==(nullptr_t, observer_ptr<W>) noexcept;
    template <class W>
    bool operator!=(nullptr_t, observer_ptr<W>) noexcept;
    // ordering operators
    template <class W1, class W2>
    bool operator<(observer_ptr<W1>, observer_ptr<W2>);
    template <class W1, class W2>
    bool operator>(observer_ptr<W1>, observer_ptr<W2>);
    template <class W1, class W2>
    bool operator<=(observer_ptr<W1>, observer_ptr<W2>);
    template <class W1, class W2>
    bool operator>=(observer_ptr<W1>, observer_ptr<W2>);

  } // namespace experimental::inline fundamentals_v3

  // 5.2.7, observer_ptr hash support
  template <class T> struct hash;
  template <class T> struct hash<experimental::observer_ptr<T>>;

} // namespace std
5.2

Non-owning (observer) pointers

[memory.observer.ptr]
5.2.1

Class template observer_ptr overview

[memory.observer.ptr.overview]
namespace std::experimental::inline fundamentals_v3 {

  template <class W> class observer_ptr {
    using pointer = add_pointer_t<W>;            // exposition-only
    using reference = add_lvalue_reference_t<W>; // exposition-only
  public:
    // publish our template parameter and variations thereof
    using element_type = W;

    // 5.2.2, observer_ptr constructors
    // default constructor
    constexpr observer_ptr() noexcept;

    // pointer-accepting constructors
    constexpr observer_ptr(nullptr_t) noexcept;
    constexpr explicit observer_ptr(pointer) noexcept;

    // copying constructors (in addition to the implicit copy constructor)
    template <class W2> constexpr observer_ptr(observer_ptr<W2>) noexcept;

    // 5.2.3, observer_ptr observers
    constexpr pointer get() const noexcept;
    constexpr reference operator*() const;
    constexpr pointer operator->() const noexcept;
    constexpr explicit operator bool() const noexcept;

    // 5.2.4, observer_ptr conversions
    constexpr explicit operator pointer() const noexcept;

    // 5.2.5, observer_ptr modifiers
    constexpr pointer release() noexcept;
    constexpr void reset(pointer = nullptr) noexcept;
    constexpr void swap(observer_ptr&) noexcept;
  }; // observer_ptr<>

} // namespace std::experimental::inline fundamentals_v3

A non-owning pointer, known as an observer, is an object o that stores a pointer to a second object, w. In this context, w is known as a watched object. [ Note: There is no watched object when the stored pointer is nullptr. end note ] An observer takes no responsibility or ownership of any kind for its watched object, if any; in particular, there is no inherent relationship between the lifetimes of o and w.

Specializations of observer_ptr shall meet the requirements of a Cpp17CopyConstructible and Cpp17CopyAssignable type. The template parameter W of an observer_ptr shall not be a reference type, but may be an incomplete type.

[ Note: The uses of observer_ptr include clarity of interface specification in new code, and interoperability with pointer-based legacy code. end note ]

5.2.2

observer_ptr constructors

[memory.observer.ptr.ctor]
constexpr observer_ptr() noexcept;constexpr observer_ptr(nullptr_t) noexcept;
Effects:
Constructs an observer_ptr object that has no corresponding watched object.
Postconditions:
get() == nullptr.
constexpr explicit observer_ptr(pointer other) noexcept;
Postconditions:
get() == other.
template <class W2> constexpr observer_ptr(observer_ptr<W2> other) noexcept;
Constraints:
W2* is convertible to W*.
Postconditions:
get() == other.get().
5.2.3

observer_ptr observers

[memory.observer.ptr.obs]
constexpr pointer get() const noexcept;
Returns:
The stored pointer.
constexpr reference operator*() const;
Preconditions:
get() != nullptr is true.
Returns:
*get().
Throws:
Nothing.
constexpr pointer operator->() const noexcept;
Returns:
get().
constexpr explicit operator bool() const noexcept;
Returns:
get() != nullptr.
5.2.4

observer_ptr conversions

[memory.observer.ptr.conv]
constexpr explicit operator pointer() const noexcept;
Returns:
get().
5.2.5

observer_ptr modifiers

[memory.observer.ptr.mod]
constexpr pointer release() noexcept;
Postconditions:
get() == nullptr.
Returns:
The value get() had at the start of the call to release.
constexpr void reset(pointer p = nullptr) noexcept;
Postconditions:
get() == p.
constexpr void swap(observer_ptr& other) noexcept;
Effects:
Invokes swap on the stored pointers of *this and other.
5.2.6

observer_ptr specialized algorithms

[memory.observer.ptr.special]
template <class W>
void swap(observer_ptr<W>& p1, observer_ptr<W>& p2) noexcept;
Effects:
p1.swap(p2).
template <class W> observer_ptr<W> make_observer(W* p) noexcept;
Returns:
observer_ptr<W>{p}.
template <class W1, class W2>
bool operator==(observer_ptr<W1> p1, observer_ptr<W2> p2);
Returns:
p1.get() == p2.get().
template <class W1, class W2>
bool operator!=(observer_ptr<W1> p1, observer_ptr<W2> p2);
Returns:
not (p1 == p2).
template <class W>
bool operator==(observer_ptr<W> p, nullptr_t) noexcept;template <class W>
bool operator==(nullptr_t, observer_ptr<W> p) noexcept;
Returns:
not p.
template <class W>
bool operator!=(observer_ptr<W> p, nullptr_t) noexcept;template <class W>
bool operator!=(nullptr_t, observer_ptr<W> p) noexcept;
Returns:
(bool)p.
template <class W1, class W2>
bool operator<(observer_ptr<W1> p1, observer_ptr<W2> p2);
Returns:
less<W3>()(p1.get(), p2.get()), where W3 is the composite pointer type (C++20 §7) of W1* and W2*.
template <class W1, class W2>
bool operator>(observer_ptr<W1> p1, observer_ptr<W2> p2);
Returns:
p2 < p1.
template <class W1, class W2>
bool operator<=(observer_ptr<W1> p1, observer_ptr<W2> p2);
Returns:
not (p2 < p1).
template <class W1, class W2>
bool operator>=(observer_ptr<W1> p1, observer_ptr<W2> p2);
Returns:
not (p1 < p2).
5.2.7

observer_ptr hash support

[memory.observer.ptr.hash]
template <class T> struct hash<experimental::observer_ptr<T>>;

The specialization is enabled (C++20 §20.14.18). For an object p of type observer_ptr<T>, hash<observer_ptr<T>>()(p) evaluates to the same value as hash<T*>()(p.get()).

5.3

Type-erased allocator

[memory.type.erased.allocator]

A type-erased allocator is an allocator or memory resource, alloc, used to allocate internal data structures for an object X of type C, but where C is not dependent on the type of alloc. Once alloc has been supplied to X (typically as a constructor argument), alloc can be retrieved from X only as a pointer rptr of static type std::pmr::memory_resource* (C++20 §20.12.2). The process by which rptr is computed from alloc depends on the type of alloc as described in Table 5:

Table 5 — Computed memory_resource for type-erased allocator
If the type of alloc is then the value of rptr is
non-existent — no alloc specified The value of pmr::get_default_resource() at the time of construction.
nullptr_t The value of pmr::get_default_resource() at the time of construction.
a pointer type convertible to pmr::memory_resource* static_cast<pmr::memory_resource*>(alloc)
pmr::polymorphic_allocator<U> alloc.resource()
any other type meeting the Cpp17Allocator requirements (C++20 §16.5.3.5) a pointer to a value of type pmr::resource_adaptor<A> where A is the type of alloc. rptr remains valid only for the lifetime of X.
None of the above The program is ill-formed.

Additionally, class C meets the following requirements:

  • C::allocator_type denotes std::experimental::erased_type.
  • X.get_memory_resource() returns rptr.
5.4

Header <experimental/memory_resource> synopsis

[memory.resource.syn]
namespace std::experimental::inline fundamentals_v3::pmr {

  // The name resource_adaptor_imp is for exposition only.
  template <class Allocator> class resource_adaptor_imp;

  template <class Allocator>
    using resource_adaptor = resource_adaptor_imp<
      typename allocator_traits<Allocator>::template rebind_alloc<char>>;

} // namespace std::experimental::inline fundamentals_v3::pmr
5.5

Alias template resource_adaptor

[memory.resource.adaptor]
5.5.1

resource_adaptor

[memory.resource.adaptor.overview]

An instance of resource_adaptor<Allocator> is an adaptor that wraps a memory_resource interface around Allocator. In order that resource_adaptor<X<T>> and resource_adaptor<X<U>> are the same type for any allocator template X and types T and U, resource_adaptor<Allocator> is rendered as an alias to a class template such that Allocator is rebound to a char value type in every specialization of the class template. The requirements on this class template are defined below. The name resource_adaptor_imp is for exposition only and is not normative, but the definitions of the members of that class, whatever its name, are normative. In addition to the Cpp17Allocator requirements (C++20 §16.5.3.5), the parameter to resource_adaptor shall meet the following additional requirements:

  • typename allocator_traits<Allocator>::pointer shall be identical to typename allocator_traits<Allocator>::value_type*.
  • typename allocator_traits<Allocator>::const_pointer shall be identical to typename allocator_traits<Allocator>::value_type const*.
  • typename allocator_traits<Allocator>::void_pointer shall be identical to void*.
  • typename allocator_traits<Allocator>::const_void_pointer shall be identical to void const*.

// The name resource_adaptor_imp is for exposition only.
template <class Allocator>
class resource_adaptor_imp : public memory_resource {
  // for exposition only
  Allocator m_alloc;

public:
  using allocator_type = Allocator;

  resource_adaptor_imp() = default;
  resource_adaptor_imp(const resource_adaptor_imp&) = default;
  resource_adaptor_imp(resource_adaptor_imp&&) = default;

  explicit resource_adaptor_imp(const Allocator& a2);
  explicit resource_adaptor_imp(Allocator&& a2);

  resource_adaptor_imp& operator=(const resource_adaptor_imp&) = default;

  allocator_type get_allocator() const { return m_alloc; }

protected:
  virtual void* do_allocate(size_t bytes, size_t alignment);
  virtual void do_deallocate(void* p, size_t bytes, size_t alignment);

  virtual bool do_is_equal(const memory_resource& other) const noexcept;
};

template <class Allocator>
  using resource_adaptor = typename resource_adaptor_imp<
    typename allocator_traits<Allocator>::template rebind_alloc<char>>;
5.5.2

resource_adaptor_imp constructors

[memory.resource.adaptor.ctor]
explicit resource_adaptor_imp(const Allocator& a2);
Effects:
Initializes m_alloc with a2.
explicit resource_adaptor_imp(Allocator&& a2);
Effects:
Initializes m_alloc with std::move(a2).
5.5.3

resource_adaptor_imp member functions

[memory.resource.adaptor.mem]
void* do_allocate(size_t bytes, size_t alignment);
Returns:
Allocated memory obtained by calling m_alloc.allocate. The size and alignment of the allocated memory shall meet the requirements for a class derived from memory_resource (C++20 §20.12.2).
void do_deallocate(void* p, size_t bytes, size_t alignment);
Preconditions:
p was previously allocated using A.allocate, where A == m_alloc, and not subsequently deallocated.
Effects:
Returns memory to the allocator using m_alloc.deallocate().
bool do_is_equal(const memory_resource& other) const noexcept;

Let p be dynamic_cast<const resource_adaptor_imp*>(&other).

Returns:
false if p is null, otherwise the value of m_alloc == p->m_alloc.
6

Iterators library

[iterator]
6.1

Header <experimental/iterator> synopsis

[iterator.syn]
#include <iterator>

namespace std::experimental::inline fundamentals_v3 {

  // 6.2, Class template ostream_joiner
  template <class DelimT, class charT = char, class traits = char_traits<charT> >
      class ostream_joiner;
  template <class charT, class traits, class DelimT>
    ostream_joiner<decay_t<DelimT>, charT, traits>
    make_ostream_joiner(basic_ostream<charT, traits>& os, DelimT&& delimiter);

} // namespace std::experimental::inline fundamentals_v3
6.2

Class template ostream_joiner

[iterator.ostream.joiner]
6.2.1

Overview

[iterator.ostream.joiner.overview]

ostream_joiner writes (using operator<<) successive elements onto the output stream from which it was constructed. The delimiter that it was constructed with is written to the stream between every two Ts that are written. It is not possible to get a value out of the output iterator. Its only use is as an output iterator in situations like

while (first != last)
  *result++ = *first++;

ostream_joiner is defined as

namespace std::experimental::inline fundamentals_v3 {

  template <class DelimT, class charT = char, class traits = char_traits<charT> >
  class ostream_joiner {
  public:
    using char_type = charT;
    using traits_type = traits;
    using ostream_type = basic_ostream<charT, traits>;
    using iterator_category = output_iterator_tag;
    using value_type = void;
    using difference_type = void;
    using pointer = void;
    using reference = void;

    ostream_joiner(ostream_type& s, const DelimT& delimiter);
    ostream_joiner(ostream_type& s, DelimT&& delimiter);
    template<typename T>
    ostream_joiner& operator=(const T& value);
    ostream_joiner& operator*() noexcept;
    ostream_joiner& operator++() noexcept;
    ostream_joiner& operator++(int) noexcept;

  private:
    ostream_type* out_stream; // exposition only
    DelimT delim;             // exposition only
    bool first_element;       // exposition only
  };

} // namespace std::experimental::inline fundamentals_v3
6.2.2

ostream_joiner constructor

[iterator.ostream.joiner.cons]
ostream_joiner(ostream_type& s, const DelimT& delimiter);
Effects:
Initializes out_stream with std::addressof(s), delim with delimiter, and first_element with true.
ostream_joiner(ostream_type& s, DelimT&& delimiter);
Effects:
Initializes out_stream with std::addressof(s), delim with move(delimiter), and first_element with true.
6.2.3

ostream_joiner operations

[iterator.ostream.joiner.ops]
template<typename T>
ostream_joiner& operator=(const T& value);
Effects:
if (!first_element)
  *out_stream << delim;
first_element = false;
*out_stream << value;
return *this;
ostream_joiner& operator*() noexcept;
Returns:
*this.
ostream_joiner& operator++() noexcept;ostream_joiner& operator++(int) noexcept;
Returns:
*this.
6.2.4

ostream_joiner creation function

[iterator.ostream.joiner.creation]
template <class charT, class traits, class DelimT>
ostream_joiner<decay_t<DelimT>, charT, traits>
make_ostream_joiner(basic_ostream<charT, traits>& os, DelimT&& delimiter);
Returns:
ostream_joiner<decay_t<DelimT>, charT, traits>(os, forward<DelimT>(delimiter));
7

Futures

[futures]
7.1

Header <experimental/future> synopsis

[future.syn]
#include <future>

namespace std {
  namespace experimental::inline fundamentals_v3 {

    template <class R> class promise;
    template <class R> class promise<R&>;
    template <> class promise<void>;

    template <class R>
    void swap(promise<R>& x, promise<R>& y) noexcept;

    template <class> class packaged_task; // undefined
    template <class R, class... ArgTypes>
    class packaged_task<R(ArgTypes...)>;

    template <class R, class... ArgTypes>
    void swap(packaged_task<R(ArgTypes...)>&, packaged_task<R(ArgTypes...)>&) noexcept;

  } // namespace experimental::inline fundamentals_v3

  template <class R, class Alloc>
  struct uses_allocator<experimental::promise<R>, Alloc>;

  template <class R, class Alloc>
  struct uses_allocator<experimental::packaged_task<R>, Alloc>;

} // namespace std
7.2

Class template promise

[futures.promise]

The specification of all declarations within this subclause 7.2 are the same as the corresponding declarations, as specified in C++20 §32.9.6, unless explicitly specified otherwise.

namespace std {
  namespace experimental::inline fundamentals_v3 {

    template <class R>
    class promise {
    public:
      using allocator_type = erased_type;

      promise();
      template <class Allocator>
      promise(allocator_arg_t, const Allocator& a);
      promise(promise&& rhs) noexcept;
      promise(const promise& rhs) = delete;
      ~promise();

      promise& operator=(promise&& rhs) noexcept;
      promise& operator=(const promise& rhs) = delete;
      void swap(promise& other) noexcept;

      future<R> get_future();

      void set_value(see below);
      void set_exception(exception_ptr p);

      void set_value_at_thread_exit(const R& r);
      void set_value_at_thread_exit(see below);
      void set_exception_at_thread_exit(exception_ptr p);

      pmr::memory_resource* get_memory_resource() const noexcept;
    };

    template <class R>
    void swap(promise<R>& x, promise<R>& y) noexcept;

  } // namespace experimental::inline fundamentals_v3

  template <class R, class Alloc>
  struct uses_allocator<experimental::promise<R>, Alloc>;

} // namespace std

When a promise constructor that takes a first argument of type allocator_arg_t is invoked, the second argument is treated as a type-erased allocator (5.3).

7.3

Class template packaged_task

[futures.task]

The specification of all declarations within this subclause 7.3 are the same as the corresponding declarations, as specified in C++20 §32.9.10, unless explicitly specified otherwise.

namespace std {
  namespace experimental::inline fundamentals_v3 {

    template <class R, class... ArgTypes>
    class packaged_task<R(ArgTypes...)> {
    public:
      using allocator_type = erased_type;

      packaged_task() noexcept;
      template <class F>
      explicit packaged_task(F&& f);
      template <class F, class Allocator>
      explicit packaged_task(allocator_arg_t, const Allocator& a, F&& f);
      ~packaged_task();

      packaged_task(const packaged_task&) = delete;
      packaged_task& operator=(const packaged_task&) = delete;

      packaged_task(packaged_task&& rhs) noexcept;
      packaged_task& operator=(packaged_task&& rhs) noexcept;
      void swap(packaged_task& other) noexcept;

      bool valid() const noexcept;

      future<R> get_future();

      void operator()(ArgTypes... );
      void make_ready_at_thread_exit(ArgTypes...);

      void reset();

      pmr::memory_resource* get_memory_resource() const noexcept;
    };

    template <class R, class... ArgTypes>
    void swap(packaged_task<R(ArgTypes...)>&, packaged_task<R(ArgTypes...)>&) noexcept;

  } // namespace experimental::inline fundamentals_v3

  template <class R, class Alloc>
  struct uses_allocator<experimental::packaged_task<R>, Alloc>;

} // namespace std

When a packaged_task constructor that takes a first argument of type allocator_arg_t is invoked, the second argument is treated as a type-erased allocator (5.3).

8

Algorithms library

[algorithms]
8.1

Header <experimental/algorithm> synopsis

[algorithm.syn]
#include <algorithm>

namespace std::experimental::inline fundamentals_v3 {

  // 8.2, Sampling
  template<class PopulationIterator, class SampleIterator, class Distance>
  SampleIterator sample(PopulationIterator first, PopulationIterator last,
                        SampleIterator out, Distance n);

  // 8.3, Shuffle
  template<class RandomAccessIterator>
  void shuffle(RandomAccessIterator first, RandomAccessIterator last);

} // namespace std::experimental::inline fundamentals_v3
8.2

Sampling

[alg.random.sample]
template<class PopulationIterator, class SampleIterator, class Distance>
SampleIterator sample(PopulationIterator first, PopulationIterator last,
                      SampleIterator out, Distance n);
Effects:
Equivalent to:
return ::std::sample(first, last, out, n, g);
where g denotes the per-thread engine (). To the extent that the implementation of this function makes use of random numbers, the object g serves as the implementation’s source of randomness.
8.3

Shuffle

[alg.random.shuffle]
template<class RandomAccessIterator>
  void shuffle(RandomAccessIterator first, RandomAccessIterator last);
Preconditions:
RandomAccessIterator meets the Cpp17ValueSwappable requirements (C++20 §16.5.3.2).
Effects:
Permutes the elements in the range [first,last) such that each possible permutation of those elements has equal probability of appearance.
Complexity:
Exactly (last - first) - 1 swaps.
Remarks:
To the extent that the implementation of this function makes use of random numbers, the per-thread engine (9.1.2) serves as the implementation's source of randomness.
9

Numerics library

[numeric]
9.1

Random number generation

[rand]
9.1.1

Header <experimental/random> synopsis

[rand.syn]
#include <random>

namespace std::experimental::inline fundamentals_v3 {

  // 9.1.2, Function template randint
  template <class IntType>
  IntType randint(IntType a, IntType b);
  void reseed();
  void reseed(default_random_engine::result_type value);

} // namespace std::experimental::inline fundamentals_v3
9.1.2

Function template randint

[rand.randint]

A separate per-thread engine of type default_random_engine (C++20 §26.6.5), initialized to an unpredictable state, shall be maintained for each thread.

template<class IntType>
IntType randint(IntType a, IntType b);
Mandates:
The template argument meets the requirements for a template parameter named IntType in C++20 §26.6.2.1.
Preconditions:
ab.
Returns:
A random integer i, aib, produced from a thread-local instance of uniform_int_distribution<IntType> (C++20 §26.6.8.2.1) invoked with the per-thread engine.
void reseed();void reseed(default_random_engine::result_type value);
Effects:
Let g be the per-thread engine. The first form sets g to an unpredictable state. The second form invokes g.seed(value).
Postconditions:
Subsequent calls to randint do not depend on values produced by g before calling reseed. [ Note: reseed also resets any instances of uniform_int_distribution used by randint. end note ]