|
|
Member functions | ||||
(C++11) | ||||
Non-member functions | ||||
(until C++20)(until C++20)(until C++20)(until C++20)(until C++20)(C++20) | ||||
(C++11) | ||||
(C++11) | ||||
Helper classes | ||||
(C++11) | ||||
(C++11) | ||||
(C++23) | ||||
(C++23) | ||||
piecewise_construct_t (C++11) | ||||
Deduction guides(C++17) |
Defined in header <utility> | ||
struct piecewise_construct_t{explicit piecewise_construct_t()=default;}; | (1) | (since C++11) |
constexprstd::piecewise_construct_t piecewise_construct{}; | (2) | (since C++11) (inline since C++17) |
std::piecewise_construct_t
is an empty class tag type used to disambiguate between different functions that take two tuple arguments.std::piecewise_construct
is an instance of(1).The overloads that do not usestd::piecewise_construct_t
assume that each tuple argument becomes the element of a pair. The overloads that usestd::piecewise_construct_t
assume that each tuple argument is used to construct, piecewise, a new object of specified type, which will become the element of the pair.
Contents |
The following standard library types and functions use it as a disambiguation tag:
implements binary tuple, i.e. a pair of values (class template)[edit] | |
prepares the argument list matching the flavor of uses-allocator construction required by the given type (function template)[edit] | |
aview consisting of a generated sequence by repeatedly producing the same value(class template)(customization point object)[edit] |
#include <iostream>#include <tuple>#include <utility> struct Foo{ Foo(std::tuple<int,float>){std::cout<<"Constructed a Foo from a tuple\n";} Foo(int,float){std::cout<<"Constructed a Foo from an int and a float\n";}}; int main(){std::tuple<int,float> t(1,3.14); std::cout<<"Creating p1...\n";std::pair<Foo, Foo> p1(t, t); std::cout<<"Creating p2...\n";std::pair<Foo, Foo> p2(std::piecewise_construct, t, t);}
Output:
Creating p1...Constructed a Foo from a tupleConstructed a Foo from a tupleCreating p2...Constructed a Foo from an int and a floatConstructed a Foo from an int and a float
The following behavior-changing defect reports were applied retroactively to previously published C++ standards.
DR | Applied to | Behavior as published | Correct behavior |
---|---|---|---|
LWG 2510 | C++11 | the default constructor was non-explicit, which could lead to ambiguity | made explicit |
constructs new pair (public member function of std::pair<T1,T2> ) |