Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Should all/most setter functions in C++11 be written as function templates accepting universal references?

Consider a class X with N member variables, each of some copiable and movable type, and N corresponding setter functions. In C++98, the definition of X would likely look something like this:

class X { public:     void set_a(A const& a) { _a = a; }     void set_b(B const& b) { _b = b; }     ... private:     A _a;     B _b;     ... }; 

Setter functions of class X above can bind both to lvalue and to rvalue arguments. Depending on the actual argument, this might result in the creation of a temporary and will eventually result in a copy assignment; due to this, non-copiable types are not supported by this design.

With C++11 we have move semantics, perfect forwarding, and universal references (Scott Meyers's terminology), which allow for a more efficient and generalized use of setter functions by rewriting them this way:

class X { public:     template<typename T>     void set_a(T&& a) { _a = std::forward<T>(a); }      template<typename T>     void set_b(T&& b) { _b = std::forward<T>(b); }     ... private:     A _a;     B _b;     ... }; 

Universal references can bind to const/non-const, volatile/non-volatile, and to any convertible type in general, avoiding the creation of temporaries and passing values straight to operator =. Non-copiable, movable types are now supported. Possibly undesired bindings can be eliminated either through static_assert or through std::enable_if.

So my question is: as a design guideline, should all (let's say, most) setter functions in C++11 be written as function templates accepting universal references?

Apart from the more cumbersome syntax and the impossibility of using Intellisense-like helper tools when writing code in those setter functions, are there any relevant disadvantages with the hypothetical principle "write setter functions as function templates accepting universal references whenever possible"?

like image 821
Andy Prowl Avatar asked Jan 07 '13 14:01

Andy Prowl


1 Answers

You know the classes A and B, so you know if they are movable or not and if this design is ultimately necessary. For something like std::string, it's a waste of time changing the existing code unless you know you have a performance problem here. If you're dealing with auto_ptr, then it's time to rip it out and use unique_ptr.

It's usually preferred now to take arguments by value if you don't know anything more specific- such as

void set_a(A a) { _a = std::move(a); } 

This permits the use of any of the constructors of A without requiring anything except movability and offers a relatively intuitive interface.

like image 138
Puppy Avatar answered Sep 18 '22 01:09

Puppy