InC andC++, asequence point defines any point in acomputer program'sexecution at which it is guaranteed that allside effects of previous evaluations will have been performed, and no side effects from subsequent evaluations have yet been performed. They are a core concept for determining the validity of and, if valid, the possible results of expressions. Adding more sequence points is sometimes necessary to make an expression defined and to ensure a single valid order of evaluation.
Documentation forC11 andC++11 stopped using the term "sequence point" and now uses alternative terms:[1][2][3]
The execution of unsequenced evaluations can overlap, leading to potentially catastrophicundefined behavior if they sharestate. This situation can arise inparallel computing, causingrace conditions, but undefined behavior can also result in single-threaded situations. For example,a[i] = i++; (wherea is an array andi is an integer) has undefined behavior.
Consider twofunctionsf() andg(). In C and C++, the+ operator is not associated with a sequence point, and therefore in theexpressionf()+g() it is possible that eitherf() org() will be executed first. Thecomma operator introduces a sequence point, and therefore in the codef(),g() the order of evaluation is defined: firstf() is called, and theng() is called.
Sequence points also come into play when the same variable is modified more than once within a single expression. An often-cited example is theC expressioni=i++, which apparently both assignsi its previous value and incrementsi. The final value ofi is ambiguous, because, depending on the order of expression evaluation, the increment may occur before, after, or interleaved with the assignment. The definition of a particular language might specify one of the possible behaviors or simply say the behavior isundefined. In C and C++, evaluating such an expression yields undefined behavior.[4] Other languages, such asC#, define theprecedence of theassignment operator and theincrement operator in such a way that the result of the expressioni=i++ is guaranteed.
In C[5] and C++,[6] sequence points occur in the following places. (In C++,overloaded operators act like functions, and thus operators that have been overloaded introduce sequence points in the same way as function calls.)
&& (logical AND),|| (logical OR) (as part ofshort-circuit evaluation), andcomma operators. For example, in the expression*p++!=0&&*q++!=0, all side effects of the sub-expression*p++!=0 are completed before any attempt to accessq.a=(*p++)?(*p++):0 there is a sequence point after the first*p++, meaning it has already been incremented by the time the second instance is executed.a=b;),return statements, the controlling expressions ofif,switch,while, ordo-while statements, and each of the three expressions in afor statement.f(i++)+g(j++)+h(k++),f is called with a parameter of the original value ofi, buti is incremented before entering the body off. Similarly,j andk are updated before enteringg andh respectively. However, it is not specified in which orderf(),g(),h() are executed, nor in which orderi,j,k are incremented. If the body off accesses the variablesj andk, it might find both, neither, or just one of them to have been incremented. (The function callf(a,b,c) isnot a use of the comma operator; the order of evaluation fora,b, andc is unspecified.)5 in the declarationinta=5;.a++ inintx=a++,y=a++.[8] (This isnot an example of the comma operator.)printf("foo %n %d",&a,42), there is a sequence point after the%n is evaluated and before printing42.This sectionneeds expansion. You can help byadding missing information.(April 2023) |
C++11 brought diverse changes, including features that improve the language's performance, usability, and multithreading.[9] Partially because of the introduction of language support for threads, C11 and C++11 introduced new terminology for evaluation order. An operation may be “sequenced before” another, the two can be “indeterminately sequenced” (one must complete before the other), or the two can be “unsequenced” (the operations in each expression may be interleaved).
C++17 restricted several aspects of evaluation order. Thenew expression will always perform the memory allocation before evaluating the constructor arguments. The operators<<,>>,.,.*,->*, and the subscript and function call operator are guaranteed to be evaluated left to right (whether they are overloaded or not). For example, the code
std::cout<<a()<<b()<<c();// parsed as (((std::cout << a()) << b()) << c());
is newly guaranteed to calla,b, andc in that order. The right-hand side of any assignment-like operator is evaluated before the left-hand side, so thatb() *= a(); is guaranteed to evaluatea first. Finally, although the order in which function parameters are evaluated remains implementation-defined, the compiler is no longer allowed to interleavesub-expressions across multiple parameters.[10]