Operators in C and C++: Difference between revisions

Revision as of 14:49, 31 January 2025

Latest revision as of 02:44, 23 April 2025

(8 intermediate revisions by 3 users not shown)

Line 9:

Most of the operators available in C and C++ are also available in other [[C-family]] languages such as [[C Sharp (programming language)|C#]], [[D (programming language)|D]], [[Java (programming language)|Java]], [[Perl]], and [[PHP]] with the same precedence, associativity, and semantics.

Many ofoperators thespecified operatorsby containinga multi-charactersequence sequencesof symbols are givencommonly "names"referred builtto fromby thea operatorname that consists of the name of each charactersymbol. For example, <code>+=</code> and <code>-=</code> are often called ''"plus equal(s)''" and ''"minus equal(s)''", instead of the more verbose "assignment by addition" and "assignment by subtraction".

==Operators==

Line 67 ⟶ 69:

| {{rh}} colspan="2" | Postfix increment

| align="center" | <code>a'''++'''</code>

| {{cpp|1=R K::operator ++(int);}} {{efn|name=dummy-int|The {{cpp|int}} is a dummy parameter to differentiate between prefix and postfix.}}

| {{cpp|1=R operator ++(K& a, int);}} {{efn|name=dummy-int}}

|-

| {{rh}} colspan="2" | Prefix [[Increment and decrement operators|decrement]]

Line 77 ⟶ 79:

| {{rh}} colspan="2" | Postfix decrement

| align="center" | <code>a'''--'''</code>

| {{cpp|1=R K::operator --(int);}} {{efn|name=dummy-int}}

| {{cpp|1=R operator --(K& a, int);}} {{efn|name=dummy-int}}

|}

Line 126 ⟶ 128:

|-

| {{rh}} colspan="2" | [[Three-way comparison]]<ref name="threewaycomparison" group="lower-alpha"/>{{efn|Possible return types: <code>std::weak_ordering</code>, <code>std::strong_ordering</code> and <code>std::partial_ordering</code> to which they all are convertible to.}}

| rowspan="2" style="text-align:center;" | <code>a '''&lt;=&gt;''' b</code> || rowspan="2"{{no}}

| {{cpp|1=auto K::operator <=>(const S &b);}}

| {{cpp|1=auto operator <=>(const K &a, const S &b);}}

Line 173 ⟶ 175:

! outside class

|-

| {{rh}} colspan="2" | [[Bitwise operation#NOT|Bitwise NOT]]

| align="center" | <code>'''~'''a</code><br/>

| {{cpp|1=R K::operator ~();}}

| {{cpp|1=R operator ~(K a);}}

|-

| {{rh}} colspan="2" | [[Bitwise operation#AND|Bitwise AND]]

| style="text-align:center;" | <code>a '''&''' b</code>

| {{cpp|1=R K::operator &(S b);}}

| {{cpp|1=R operator &(K a, S b);}}

|-

| {{rh}} colspan="2" | [[Bitwise operation#OR|Bitwise OR]]

| style="text-align:center;" | <code>a '''<nowiki>|</nowiki>''' b</code>

| {{cpp|1=R K::operator {{!}}(S b);|lang=cpp}}

| {{cpp|1=R operator {{!}}(K a, S b);|lang=cpp}}

|-

| {{rh}} colspan="2" | [[Bitwise operation#XOR|Bitwise XOR]]

| style="text-align:center;" | <code>a '''^''' b</code>

| {{cpp|1=R K::operator ^(S b);}}

| {{cpp|1=R operator ^(K a, S b);}}

|-

| {{rh}} colspan="2" | [[Bitwise shift|BitwiseShift]] left shift]]<ref name="bitshift" group="lower-alpha" />

| style="text-align:center;" | <code>a '''<<''' b</code>

| {{cpp|1=R K::operator <<(S b);}}

| {{cpp|1=R operator <<(K a, S b);}}

|-

| {{rh}} colspan="2" | [[Bitwise shift|BitwiseShift right shift]]<ref name="bitshift" group="lower-alpha" />{{Refn | Operation="rightbitshift" | group="lower-alpha" | According to the C99 standard, the right shift of a negative number is implementation defined. Most implementations, e.g., the GCC,<ref name="Integers">{{Citation | contribution = Integers implementation | url = //gcc.gnu.org/onlinedocs/gcc-4.3.3/gcc/Integers-implementation.html#Integers-implementation | title = GCC 4.3.3 | publisher = GNU}}.</ref> use an [[arithmetic shift]] (i.e., sign extension), but a [[logical shift]] is possible.}}

| style="text-align:center;" | <code>a '''>>''' b</code>

| {{cpp|1=R K::operator >>(S b);}}

Line 208 ⟶ 210:

C and C++ have the same assignment operators and all can be overloaded in C++.

For the combination operators, <code>a ⊚= b</code> (where <code>⊚</code> represnetsrepresents an operation) is equivalent to <code>a = a ⊚ b</code>, except that <code>a</code> is evaluated only once.

{| class="wikitable" style="width:100%"

Line 289 ⟶ 291:

| {{yes}}

| {{yes}}

| {{cpp|1=R& K::operator [](S b);}}<br/>{{cpp|1=R& K::operator [](S b, ...); // since C++23}}<brref name="sinceCXX23" group="lower-alpha" />

| {{n/a}}

|-

Line 352 ⟶ 354:

| colspan="2" {{n/a}}

|-

| {{rh}} colspan="2" | User-defined literals<ref name="ud-literal" group="lower-alpha" /><brref name="sinceCXX11" group="lower-alpha" />{{small|(since C++11)}}

| style="text-align: center;" | <code><nowiki>"a"_b</nowiki></code> || {{yes}} || {{no}}

| {{n/a}}

Line 361 ⟶ 363:

| colspan="2" {{n/a}}

|-

| {{rh}} colspan="2" | Size of [[variadic template|parameter pack]]<br>{{small|(sinceref C++11)}}name="sinceCXX11" group="lower-alpha" />

| style="text-align:center;" | <code>'''sizeof...'''(Args)</code> || {{no}} || {{no}}

| colspan="2" {{n/a}}

|-

| {{rh}} colspan="2" | Alignof<br>{{small|(sinceref C++11)}}name="sinceCXX11" group="lower-alpha" />

| style="text-align:center;" | <code>'''alignof'''(R)</code> <br> or <code>'''_Alignof'''(R)</code><ref name="alignof" group="lower-alpha"/> || {{no}} || {{yes}}

| colspan="2" {{n/a}}

|-

| {{rh}} colspan="2" | [[Decltype]]<brref name="sinceCXX11" group="lower-alpha" />{{small|(since C++11)}}

| style="text-align:center;" | <code>'''decltype''' (a)</code><br/><code>'''decltype''' (R)</code> || {{no}} || {{no}}

| colspan="2" {{n/a}}

Line 382 ⟶ 384:

| {{n/a}}

|-

| {{rh}} colspan="2" | [[type conversion|Conversion]] {{efn|Behaves like const_cast/static_cast/reinterpret_cast. In the last two cases, the <code>auto</code> specifier is replaced with the type of the invented variable x declared with <code>auto x(a);</code> (which is never interpreted as a function declaration) or <code>auto x{a};</code>, respectively.}}<ref>[https://en.cppreference.com/w/cpp/language/explicit_cast Explicit type conversion] in C++</ref>

| style="text-align:center;" | <code>R(a)</code><br><code>R{a}</code>^{sinceref name="sinceCXX11" group="lower-alpha" C++11}<br><code>auto(a)</code>^{sinceref name="sinceCXX23" group="lower-alpha" C++23}<br><code>auto{a}</code>^{sinceref name="sinceCXX23" group="lower-alpha" C++23} || {{no}} || {{no}}

| {{rh}} colspan="2" {{n/a}}

|-

| {{rh}} colspan="2" | [[static_cast]] conversion {{efn|For user-defined conversions, the return type implicitly and necessarily matches the operator name unless the type is inferred (e.g. {{cpp|1=operator auto()}}, {{cpp|1=operator decltype(auto)()}} etc.).}}

| style="text-align:center;" | <code>'''static_cast'''<R>(a)</code> || {{yes}} || {{no}}

| {{cpp|1=K::operator R();}}<br>{{cpp|1=explicit K::operator R();}}<ref {{small|(sincename="sinceCXX11" C++11)}}group="lower-alpha" />

| {{n/a}}

|-

Line 423 ⟶ 425:

| {{cpp|1=void operator delete[](void* a);}}

|-

| {{rh}} colspan="2" | Exception check<br>{{small|(sinceref C++11)}}name="sinceCXX11" group="lower-alpha" />

| style="text-align:center;" | <code>'''noexcept'''(a)</code> || {{no}} || {{no}}

| colspan="2" {{n/a}}

|}

=== Synonyms ===

Notes:

C++ defines keywords to act as aliases for a number of operators:<ref name="Committee">{{cite book | title = ISO/IEC 14882:1998(E) Programming Language C++ | date = 1 September 1998 | publisher = open-std.org – The C++ Standards Committee | pages = 40–41}}</ref>

{{reflist |group="lower-alpha"|refs=

{| class="wikitable" style="width:30%; text-align:center;"

<ref name="modulo" group="lower-alpha">The modulus operator only supports integer operands; for floating point, a function such as [[math.h|{{code|fmod|cpp}}]] can be used.</ref>

! Keyword || Operator

<ref name="bitshift" group="lower-alpha">In the context of [[iostream]]s in C++, writers often will refer to {{cpp|<<}} and {{cpp|>>}} as the "put-to" or "stream insertion" and "get-from" or "stream extraction" operators, respectively.</ref>

|-

<ref name="addressof2" group="lower-alpha">The actual address of an object with an overloaded <code>operator &</code> can be obtained with [https://en.cppreference.com/w/cpp/memory/addressof <code>std::addressof</code>]</ref>

| {{code|and}} || <code>&&</code>

<ref name="arrowptr" group="lower-alpha">The return type of {{cpp|operator->()}} must be a type for which the {{cpp |->}} operation can be applied, such as a pointer type. If {{cpp|x}} is of type {{cpp |C}} where {{cpp |C}} overloads {{cpp|operator->()}}, {{cpp|x->y}} gets expanded to {{cpp|x.operator->()->y}}.</ref>

|-

<ref name="arrowstar" group= "lower-alpha">{{Citation | publisher = Aristeia | url = http://aristeia.com/Papers/DDJ_Oct_1999.pdf | title = Implementing operator-&gt;* for Smart Pointers | first = Scott | last = Meyers | journal = Dr. Dobb's Journal |date=Oct 1999}}.</ref>

| {{code|and_eq}} || <code>&=</code>

<ref name="scopal" group="lower-alpha">Although a <code>::</code> punctuator exists in C as of C23, it is not used as a scope resolution operator.</ref>

|-

<ref name="ud-literal" group="lower-alpha">About [http://en.cppreference.com/w/cpp/language/user_literal C++11 User-defined literals]</ref>

| {{code|bitand}} || <code>&</code>

<ref name="sizeof" group="lower-alpha">The parentheses are not necessary when taking the size of a value, only when taking the size of a type. However, they are usually used regardless.{{Citation needed|date=July 2024}}</ref>

|-

<ref name="alignof" group = "lower-alpha">C++ defines <code>alignof</code> operator, whereas C defines <code>_Alignof</code> (C23 defines both). Both operators have the same semantics.</ref>

| {{code|bitor}} || <code>&#x7C;</code>

<ref name="threewaycomparison" group="lower-alpha">About [https://en.cppreference.com/w/cpp/language/operator_comparison#Three-way_comparison C++20 three-way comparison]</ref>

|-

| {{code|compl}} || <code>~</code>

|-

| {{code|not}} || <code>!</code>

|-

| {{code|not_eq}} || <code>!=</code>

|-

| {{code|or}} || <code>&#x7C;&#x7C;</code>

|-

| {{code|or_eq}} || <code>&#x7C;=</code>

|-

| {{code|xor}} || <code>^</code>

|-

| {{code|xor_eq}} || <code>^=</code>

|}

Each keyword is a different way to specify an operator and as such can be used instead of the corresponding symbolic variation. For example, {{code|1=(a > 0 and not flag)}} and {{code|1=(a > 0 && !flag)}} specify the same behavior. As another example, the <code>bitand</code> keyword may be used to replace not only the ''bitwise-and'' operator but also the ''address-of'' operator, and it can be used to specify reference types (e.g., {{code|1=int bitand ref = n}}).

<ref name="infer" group="lower-alpha">The type name can also be inferred (e.g <code>new auto</code>) if an initializer is provided.</ref>

<ref name="infer2" group="lower-alpha">The array size can also be inferred if an initializer is provided.</ref>

The ISO C specification makes allowance for these keywords as preprocessor macros in the header file [[iso646.h|{{code|iso646.h}}]]. For compatibility with C, C++ also provides the header {{code|iso646.h}}, the inclusion of which has no effect. Until C++20, it also provided the corresponding header [[ciso646|{{code|ciso646}}]] which had no effect as well.

}}

==Expression evaluation order==

During expression evaluation, the order in which sub-expressions are evaluated is determined by [[order of operations|precedence]] and [[operator associativity|associativity]]. An operator with higher precedence is evaluated before a operator of lower precedence and the operands of an operator are evaluated based on associativity. The following table describes the precedence and associativity of the C and C++ operators. Operators are shown in groups of equal precedence with groups ordered in descending precedence from top to bottom (lower order is higher precedence).<ref>{{cite book | title = ISO/IEC 9899:201x Programming Languages - C | date = 19 December 2011 | publisher = open-std.org – The C Standards Committee | pages = 465}}</ref><ref>{{cite tech report |title=the ISO C 1999 standard, section 6.5.6 note 71 |institution=ISO |year=1999 }}</ref><ref>{{cite web |title=C++ Built-in Operators, Precedence and Associativity |url=https://docs.microsoft.com/en-US/cpp/cpp/cpp-built-in-operators-precedence-and-associativity |website=docs.microsoft.com |access-date=11 May 2020 |language=en-us}}</ref>

Operator precedence is not affected by overloading.

Line 452 ⟶ 469:

{| class="wikitable"

|-

! style="text-align: left" | PrecedenceOrder

! style="text-align: left" | Operator

! style="text-align: left" | Description

Line 681 ⟶ 698:

|}

===NotesDetails===

Although this table is adequate for describing most evaluation order, it does not describe a few details. The [[ternary operator]] allows any arbitrary expression as its middle operand, despite being listed as having higher precedence than the assignment and comma operators. Thus <code>a ? b, c : d</code> is interpreted as <code>a ? (b, c) : d</code>, and not as the meaningless <code>(a ? b), (c : d)</code>. So, the expression in the middle of the conditional operator (between <code>'''?'''</code> and <code>''':'''</code>) is parsed as if parenthesized. Also, the immediate, un-parenthesized result of a C cast expression cannot be the operand of <code>sizeof</code>. Therefore, <code>sizeof (int) * x</code> is interpreted as <code>(sizeof(int)) * x</code> and not <code>sizeof ((int) * x)</code>.

===Chained expressions===

The precedence table determines the order of binding in chained expressions, when it is not expressly specified by parentheses.

* For example, <code>++x*3</code> is ambiguous without some precedence rule(s). The precedence table tells us that: {{mono|x}} is 'bound' more tightly to {{mono|++}} than to {{mono|*}}, so that whatever {{mono|++}} does (now or later—see below), it does it ONLY to {{mono|x}} (and not to <code>x*3</code>); it is equivalent to (<code>++x</code>, <code>x*3</code>).

Line 690 ⟶ 708:

* Abstracting the issue of precedence or binding, consider the diagram above for the expression 3+2*y[i]++. The compiler's job is to resolve the diagram into an expression, one in which several unary operators (call them 3+( . ), 2*( . ), ( . )++ and ( . )[ i ]) are competing to bind to y. The order of precedence table resolves the final sub-expression they each act upon: ( . )[ i ] acts only on y, ( . )++ acts only on y[i], 2*( . ) acts only on y[i]++ and 3+( . ) acts 'only' on 2*((y[i])++). It is important to note that WHAT sub-expression gets acted on by each operator is clear from the precedence table but WHEN each operator acts is not resolved by the precedence table; in this example, the ( . )++ operator acts only on y[i] by the precedence rules but binding levels alone do not indicate the timing of the postfix ++ (the ( . )++ operator acts only after y[i] is evaluated in the expression).

===Binding===

Many of the operators containing multi-character sequences are given "names" built from the operator name of each character. For example, <code>+=</code> and <code>-=</code> are often called ''plus equal(s)'' and ''minus equal(s)'', instead of the more verbose "assignment by addition" and "assignment by subtraction".

The binding of operators in C and C++ is specified (in the corresponding Standards) by a factored language grammar, rather than a precedence table. This creates some subtle conflicts. For example, in C, the syntax for a conditional expression is:

<syntaxhighlight lang="c">logical-OR-expression ? expression : conditional-expression</syntaxhighlight>

while in C++ it is:

Line 703 ⟶ 721:

which is a valid expression.<ref>{{cite web |title=C Operator Precedence - cppreference.com |url=https://en.cppreference.com/w/c/language/operator_precedence |website=en.cppreference.com |access-date=10 April 2020}}</ref><ref>{{Cite web|url=https://stackoverflow.com/questions/13515434/does-the-c-c-ternary-operator-actually-have-the-same-precedence-as-assignment/13515505|title=Does the C/C++ ternary operator actually have the same precedence as assignment operators?|website=Stack Overflow|access-date=2019-09-22}}</ref>

IfTo youuse wantthe tocomma use comma-as-operator withinin a single function call argument expression, variable assignment, or othera comma-separated list, youuse needof toparentheses useis parentheses,required.<ref>{{cite web |title=Other operators - cppreference.com |url=https://en.cppreference.com/w/c/language/operator_other |website=en.cppreference.com |access-date=10 April 2020}}</ref><ref>{{cite web |title=c++ - How does the Comma Operator work |url=https://stackoverflow.com/questions/54142/how-does-the-comma-operator-work/ |website=Stack Overflow |access-date=1 April 2020}}</ref> e.g.:For example,

<syntaxhighlight lang="cpp">
int a = 1, b = 2, weirdVariable = (++a, b), d = 4;
</syntaxhighlight>

===Criticism of bitwise and equality operators precedence===

Line 714 ⟶ 736:

Moreover, in C++ (and later versions of C) equality operations, with the exception of the three-way comparison operator, yield [[Boolean data type|bool]] type values which are conceptually a single bit (1 or 0) and as such do not properly belong in "bitwise" operations.

== Synonyms Notes==

{{reflist |group="lower-alpha"|refs=

C++ defines keywords to act as aliases for a number of operators:<ref name="Committee">{{cite book | title = ISO/IEC 14882:1998(E) Programming Language C++ | date = 1 September 1998 | publisher = open-std.org – The C++ Standards Committee | pages = 40–41}}</ref>

<ref name="sinceCXX11" group="lower-alpha">since C++11</ref>

{| class="wikitable" style="width:30%; text-align:center;"

<ref name="sinceCXX23" group="lower-alpha">since C++23</ref>

! Keyword || Operator

<ref name="modulo" group="lower-alpha">The modulus operator only supports integer operands; for floating point, a function such as [[math.h|{{code|fmod|cpp}}]] can be used.</ref>

|-

<ref name="bitshift" group="lower-alpha">In the context of [[iostream]]s in C++, writers often will refer to {{cpp|<<}} and {{cpp|>>}} as the "put-to" or "stream insertion" and "get-from" or "stream extraction" operators, respectively.</ref>

| {{code|and}} || <code>&&</code>

<ref name="addressof2" group="lower-alpha">The actual address of an object with an overloaded <code>operator &</code> can be obtained with [https://en.cppreference.com/w/cpp/memory/addressof <code>std::addressof</code>]</ref>

|-

<ref name="arrowptr" group="lower-alpha">The return type of {{cpp|operator->()}} must be a type for which the {{cpp |->}} operation can be applied, such as a pointer type. If {{cpp|x}} is of type {{cpp |C}} where {{cpp |C}} overloads {{cpp|operator->()}}, {{cpp|x->y}} gets expanded to {{cpp|x.operator->()->y}}.</ref>

| {{code|and_eq}} || <code>&=</code>

<ref name="arrowstar" group= "lower-alpha">{{Citation | publisher = Aristeia | url = http://aristeia.com/Papers/DDJ_Oct_1999.pdf | title = Implementing operator-&gt;* for Smart Pointers | first = Scott | last = Meyers | journal = Dr. Dobb's Journal |date=Oct 1999}}.</ref>

|-

<ref name="scopal" group="lower-alpha">Although a <code>::</code> punctuator exists in C as of C23, it is not used as a scope resolution operator.</ref>

| {{code|bitand}} || <code>&</code>

<ref name="ud-literal" group="lower-alpha">About [http://en.cppreference.com/w/cpp/language/user_literal C++11 User-defined literals]</ref>

|-

<ref name="sizeof" group="lower-alpha">The parentheses are not necessary when taking the size of a value, only when taking the size of a type. However, they are usually used regardless.{{Citation needed|date=July 2024}}</ref>

| {{code|bitor}} || <code>&#x7C;</code>

<ref name="alignof" group = "lower-alpha">C++ defines <code>alignof</code> operator, whereas C defines <code>_Alignof</code> (C23 defines both). Both operators have the same semantics.</ref>

|-

<ref name="threewaycomparison" group="lower-alpha">About [https://en.cppreference.com/w/cpp/language/operator_comparison#Three-way_comparison C++20 three-way comparison]</ref>

| {{code|compl}} || <code>~</code>

|-

| {{code|not}} || <code>!</code>

|-

| {{code|not_eq}} || <code>!=</code>

|-

| {{code|or}} || <code>&#x7C;&#x7C;</code>

|-

| {{code|or_eq}} || <code>&#x7C;=</code>

|-

| {{code|xor}} || <code>^</code>

|-

| {{code|xor_eq}} || <code>^=</code>

|}

<ref name="infer" group="lower-alpha">The type name can also be inferred (e.g <code>new auto</code>) if an initializer is provided.</ref>

Each keyword is a different way to specify an operator and as such can be used instead of the corresponding symbolic variation. For example, {{code|1=(a > 0 and not flag)}} and {{code|1=(a > 0 && !flag)}} specify the same behavior. As another example, the <code>bitand</code> keyword may be used to replace not only the ''bitwise-and'' operator but also the ''address-of'' operator, and it can be used to specify reference types (e.g., {{code|1=int bitand ref = n}}).

<ref name="infer2" group="lower-alpha">The array size can also be inferred if an initializer is provided.</ref>

}}

The ISO C specification makes allowance for these keywords as preprocessor macros in the header file [[iso646.h|{{code|iso646.h}}]]. For compatibility with C, C++ also provides the header {{code|iso646.h}}, the inclusion of which has no effect. Until C++20, it also provided the corresponding header [[ciso646|{{code|ciso646}}]] which had no effect as well.

==See also==