Movatterモバイル変換

This is an unofficial snapshot of the ISO/IEC JTC1 SC22 WG21 Core Issues List revision 119a. See http://www.open-std.org/jtc1/sc22/wg21/ for the official list.

2025-12-20

2448. Cv-qualification of arithmetic types and deprecation of volatile

[Accepted as a DR at the June, 2021 meeting.]

According to the definitions in 6.9.2 [basic.fundamental],the arithmetic types include only the non-cv-qualified versions.In the taxonomy of fundamental types, the first mention of“cv-qualified versions of these types” is forscalar types (6.9 [basic.types] paragraph 9). However,7.6.1.6 [expr.post.incr] paragraph 1 and7.6.2.3 [expr.pre.incr] paragraph 1 both say:

The type of the operand shall be an arithmetic type otherthancvbool, or...

which is a contradiction, since cv-qualifiedboolis not an arithmetic type. Similarly, 7.6.19 [expr.assign] paragraph 6requires an arithmetic type for+=and-=. D.4 [depr.volatile.type] deprecatesthe increment and decrement operators when applied tovolatile-qualified arithmetic types, but the wording alreadymade those ill-formed (since the normative wording requiresan arithmetic type and not a possibly cv-qualified versionthereof).

A related question is whether 12.5 [over.built],which explicitly allows for cv-qualified arithmetic types,should also note the deprecation.

See alsoissue 2185.

Notes from the July, 2020 teleconference:

CWG felt that no changes should be made to12.5 [over.built].

Proposed resolution (April, 2021):

1. Change 6.9.2 [basic.fundamental] paragraphs 11 and 12 as follows,splitting paragraph 12 as indicated:

Typesbool,char,wchar_t,char8_t,char16_t,char32_t, andthe signed and unsigned integer types, and cv-qualifiedversions (6.9.5 [basic.type.qualifier]) thereof, arecollectivelycalledtermedintegral types. A synonym for integral typeisinteger type. [Note 8: Enumerations(9.8.1 [dcl.enum]) are not integral; however,unscoped enumerations can be promoted to integral types asspecified in 7.3.7 [conv.prom]. —endnote]

There arethree floating-point types:The three distinct typesfloat,double, andlong doublecan representfloating-point numbers. The typedoubleprovides at least as much precision asfloat, and the typelong double provides at least as much precision asdouble. The set of values of thetypefloat is a subset of the set of values of thetypedouble; the set of values of thetypedouble is a subset of the set of values of thetypelong double.The typesfloat,double, andlong double, and cv-qualifiedversions (6.9.5 [basic.type.qualifier]) thereof, arecollectively termedfloating-point types. Thevalue representation of floating-point types isimplementation-defined.[Note 9: This documentimposes no requirements on the accuracy of floating-pointoperations; see also17.3 [support.limits]. —end note]

Integral and floating-point types arecollectivelycalledtermedarithmetic types. Specializations of the standardlibrary templatestd::numeric_limits(17.3 [support.limits]) shall specify the maximum andminimum values of each arithmetic type for animplementation.

2. Change 6.9.5 [basic.type.qualifier] paragraph 1 as follows,splitting the paragraph as indicated:

A type mentioned in 6.9.2 [basic.fundamental] and6.9.4 [basic.compound] is acv-unqualifiedtype. Each typewhich is a cv-unqualified object typeor isvoid (6.9 [basic.types]) has threecorresponding cv-qualified versions of its typeother than a function or reference type is part of agroup of four distinct, but related,types:acv-unqualified version,aconst-qualified version,avolatile-qualified version, andaconst-volatile-qualified version.The type of an object (6.8.2 [intro.object])includes thecv-qualifiers specified inthedecl-specifier-seq (9.2 [dcl.spec]),declarator (9.3 [dcl.decl]),type-id(9.3.2 [dcl.name]), ornew-type-id(7.6.2.8 [expr.new]) when the object iscreated.The types in each such group shall havethe same representation and alignment requirements(6.8.3 [basic.align]). [Footnote: The samerepresentation and alignment requirements are meant to implyinterchangeability as arguments to functions, return valuesfrom functions, and non-static data members ofunions. —end footnote] A function or referencetype is always cv-unqualified.

• Aconst object is an object oftypeconst T or a non-mutable subobject of a constobject.

• Avolatile object is an object oftypevolatile T or a subobject of a volatileobject.

• Aconst volatile object is an object oftypeconst volatile T, a non-mutable subobject of aconst volatile object, a const subobject of a volatileobject, or a non-mutable volatile subobject of a constobject.

The cv-qualified or cv-unqualified versions of atype are distinct types; however, they shall have the samerepresentation and alignment requirements(6.8.3 [basic.align]).⁴⁰[Note:The type of an object (6.8.2 [intro.object]) includesthecv-qualifiers specified in thedecl-specifier-seq (9.2 [dcl.spec]),declarator (9.3 [dcl.decl]),type-id(9.3.2 [dcl.name]), ornew-type-id(7.6.2.8 [expr.new]) when the object is created.—end note]

3. Change 12.5 [over.built] paragraphs 2-10 as follows:

In this subclause, the termpromoted integral typeis used to refer to thosecv-unqualified integraltypes which are preserved by integral promotion(7.3.7 [conv.prom]) (including e.g.intandlong but excluding e.g.char). [Note 2: In all cases where a promoted integraltype is required, an operand of unscoped enumeration typewill be acceptable by way of the integralpromotions. —end note]

In the remainder of this subclause,vqrepresents eithervolatile or no cv-qualifier.

For every pair (T, vq), whereT isana cv-unqualified arithmetic type otherthanboolor a cv-unqualified pointer to(possibly cv-qualified) object type, there existcandidate operator functions of the form

vq T& operator++(vq T&);
Toperator++(vq T&, int);

For every pair (T, vq), whereT is anarithmetic type other thanbool, there existcandidate operator functions of the form

vq T& operator--(vq T&);
Toperator--(vq T&, int);

For every pair (T, vq), whereT is acv-qualified or cv-unqualified object type, there existcandidate operator functions of the form

T*vq& operator++(T*vq&);
T*vq& operator--(T*vq&);
T*    operator++(T*vq&, int);
T*    operator--(T*vq&, int);

For everycv-qualified or cv-unqualified(possibly cv-qualified) object typeTand for every function typeT that has neithercv-qualifiers nor aref-qualifier,there exist candidate operator functions of the form

T& operator*(T*);

For every function typeT that does not havecv-qualifiers or a ref-qualifier, there exist candidateoperator functions of the form

T& operator*(T*);

For every typeT there exist candidate operatorfunctions of the form

T* operator+(T*);

For everycv-unqualified floating-point orpromoted integral typeT, there exist candidateoperator functions of the form

T operator+(T);
Tl operator-(T);

[Drafting note: Clause 21 [meta] regardingtype traits appropriately handles cv-qualified andcv-unqualified types and does not require revision.]