This page is a snapshot from the LWG issues list, see the Library Active Issues List for more information and the meaning of C++20 status.

3133. Modernizing numeric type requirements

Section: 29.2 [numeric.requirements] Status: C++20 Submitter: Tim Song Opened: 2018-07-05 Last modified: 2021-02-25

Priority: 0

View all other issues in [numeric.requirements].

View all issues with C++20 status.

Discussion:

29.2 [numeric.requirements] contains some very old wording that hasn't been changed since C++98 except for issue 2699(i). As a result, it is at once over- and under-restrictive. For example:

We can significantly clean up this wording by using the existing named requirements. For ease of review, the following table provides a side-by-side comparison of the current and proposed wording.

Before After
A C++ program shall instantiate these components only with a type T that satisfies the following requirements: [Footnote … ] A C++ program shall instantiate these components only with a cv-unqualified object type T that satisfies the Cpp17DefaultConstructible, Cpp17CopyConstructible, Cpp17CopyAssignable, and Cpp17Destructiblefollowing requirements (16.4.4.2 [utility.arg.requirements]).:
(1.1) — T is not an abstract class (it has no pure virtual member functions); Cpp17DefaultConstructible
(1.2) — T is not a reference type;
(1.3) — T is not cv-qualified;
Implied by "cv-unqualified object type"
(1.4) — If T is a class, it has a public default constructor; Cpp17DefaultConstructible
(1.5) — If T is a class, it has a public copy constructor with the signature T::T(const T&); Cpp17CopyConstructible
(1.6) — If T is a class, it has a public destructor; Cpp17Destructible
(1.7) — If T is a class, it has a public copy assignment operator whose signature is either T& T::operator=(const T&) or T& T::operator=(T); Cpp17CopyAssignable
(1.8) — If T is a class, its assignment operator, copy and default constructors, and destructor shall correspond to each other in the following sense:
(1.8.1) — Initialization of raw storage using the copy constructor on the value of T(), however obtained, is semantically equivalent to value-initialization of the same raw storage.
(1.8.2) — Initialization of raw storage using the default constructor, followed by assignment, is semantically equivalent to initialization of raw storage using the copy constructor.
(1.8.3) — Destruction of an object, followed by initialization of its raw storage using the copy constructor, is semantically equivalent to assignment to the original object.
[Note: […] — end note]
These requirements are implied by Cpp17CopyConstructible and Cpp17CopyAssignable's requirement that the value of the copy is equivalent to the source.
(1.9) — If T is a class, it does not overload unary operator&. omitted now that we have std::addressof

[2019-01-20 Reflector prioritization]

Set Priority to 0 and status to Tentatively Ready

Proposed resolution:

This wording is relative to the post-Rapperswil 2018 working draft.

  1. Edit 29.2 [numeric.requirements] p1 as indicated, striking the entire bulleted list:

    -1- The complex and valarray components are parameterized by the type of information they contain and manipulate. A C++ program shall instantiate these components only with a cv-unqualified object type T that satisfies the Cpp17DefaultConstructible, Cpp17CopyConstructible, Cpp17CopyAssignable, and Cpp17Destructiblefollowing requirements (16.4.4.2 [utility.arg.requirements]). :[Footnote: … ]

    (1.1) — T is not an abstract class (it has no pure virtual member functions);

    […]

    (1.9) — If T is a class, it does not overload unary operator&.

  2. Edit 29.6.2.4 [valarray.access] p3-4 as indicated:

    const T&  operator[](size_t n) const;
    T& operator[](size_t n);
    

    -1- Requires: n < size().

    -2- Returns: […]

    -3- Remarks: The expression &addressof(a[i+j]) == &addressof(a[i]) + j evaluates to true for all size_t i and size_t j such that i+j < a.size().

    -4- The expression &addressof(a[i]) != &addressof(b[j]) evaluates to true for any two arrays a and b and for any size_t i and size_t j such that i < a.size() and j < b.size(). [Note: […] — end note ]