C++ std::move and

std::forward
C++ std::move does not move and std::forward does not forward. This article dives deep into a
long list of rules on lvalues, rvalues, references, overloads and templates to be able to explain a
few deceivingly simple lines of code using std::move and std::forward.

Motivation
N4543 suggests a solution for when the content of a  std::function  is not copyable. But if first
starts with the code below (and I’m going to ignore the rest of N4543 here). It has
a  commands  variable that maps strings to functions, and an utility function to insert a new
command.

std::map<std::string, std::function<void()>> commands;

template<typename ftor>
void install_command(std::string name, ftor && handler)
{
commands.insert( {std::move(name),std::forward<ftor>(handler)} );
}

The code above is easy to read, but has subtleties. The goal of the article is to provide enough
background information to be able to understand in detail each line, in particular why does it
use  std::move  in one place (at line 7) and  std::forward  in another (at line 8).

There is a lot of background information required. We’ll start with the almost mathematical
theory, but then dive quickly into C++ specifics and historical artefacts.

Value categories (lvalue, prvalue and xvalue)

A C++ expression has, in addition to a type, a value category. Traditionally the main value
categories were  lvalue  and  rvalue  with a rough meaning that it if could stand on the left side of
an assignment it’s an  lvalue , otherwise it’s an  rvalue .

With the advent of C++ 11, additional value categories have been identified and organized in a
systematic way based on the observation that there are two important properties of an
expresion:
has identity (i.e. ‘can we get its address’) and can be moved from.

The naming of the main value categories is illustrated using Venn diagrams below.

 If if has identity, but cannot be moved it’s an  lvalue ; otherwise it’s an  rvalue . A
typical  lvalue  is a variable name  a .
 If it can be moved, but has no identity is a  prvalue  (pure right value); otherwise it’s
a  glvalue  (generalized left value). A typical  prvalue  is a temporary resulting from a
function call/operator (with a non-reference return type) like  s.substr(1, 2)  or  a + b  or
integral constant like  42 .
 If it has an identity and can be moved it’s an  xvalue  (because that was considered
strange, and  x  is a good prefix for weird things). A typical  xvalue  is  std::move(a) .
The above categories are the main ones. There are additional ones (e.g.  void  has a category with
no identity and that can’t be moved from), but I’m going to skip over them in this article.

References as function parameters (are lvalues)

References as function parameters are relevant here because they allow us to bind to arguments
depending on their value category.

// i is a function parameter
void fn(int i) { }

int main()
{
int j = 42;
// j is a function argument
fn(j);
}

There are two types of reference declarations in C++. The pre-C++ 11 is called now  lvalue
reference  (and uses one  & ), and the new C++ 11 called  rvalue reference  (that looks like  && ).

If a function has  lvalue reference  parameter, then it can be called with an  lvalue  argument,
but not an  rvalue  argument.

// parameter is lvalue reference

void fn(X &) { std::cout<< "X &\n"; }

int main()
{
X a;
fn(a); // works, argument is an lvalue

fn(X()); // compiler error, argument is an rvalue

}

Similarly if a function has a  rvalue reference  parameter, then it can be called with
an  rvalue  argument, but not an  lvalue  argument.

// parameter is rvalue reference

void fn(X &&) { std::cout<< "X &&\n"; }

int main()
{
X a;
fn(a); // compiler error, argument is an lvalue

fn(X()); // works, argument is an rvalue

}

But when used inside the function body, a parameter, whether  lvalue reference  or  rvalue
reference , is an  lvalue  itself: it has a name like any other variable.
 // parameter is rvalue reference
void fn(X && x)
{
// but here expression x has an lvalue value category
// can use std::move to convert it to an xvalue
}

From here, things get even more complicated when one notices that  const  matters. In the
example below,  fn  can be called with both an  lvalue  and an  rvalue  argument. This is pre-C++
11 behaviour that is unchanged.

// parameter is const rvalue reference

void fn(const X &) { std::cout<< "const X &\n"; }

int main()
{
X a;
fn(a); // works, argument is an lvalue

fn(X()); // also works, argument is an rvalue

}

References and function overloads

We could provide overloads for  fn , and we end up with three main overload options. If for an
expression the preferred overload is not available, there is a fallback mechanism until all options
are exhausted, and then we get a compiler error.

struct X {};

// overloads
void fn(X &) { std::cout<< "X &\n"; }
void fn(const X &) { std::cout<< "const X &\n"; }
void fn(X &&) { std::cout<< "X &&\n"; }

int main()
{
X a;
fn(a);
// lvalue selects fn(X &)
// fallbacks on fn(const X &)

const X b;
fn(b);
// const lvalue requires fn(const X &)

fn(X());
 // rvalue selects fn(X &&)
// and then on fn(const X &)
}

In addition to the three overloads above there is of course the option of the overload with
a  const X &&  argument, but I’m going to skip over it in this article.

A typical usage of this rules is for the typical copy/move constructor/assignment quadruple. For
a user defined class  X  we expect two overloads requiring:

 const X &  for copy constructor or assignment

 X &&  for the move constructor or assignment

Template argument deduction and reference collapsing rules

If a templated function takes an  rvalue reference  template argument, special template
argument deduction rules kick in. Despite the syntactic similarities with the  rvalue
reference  above, the rules for this case where specifically designed to support argument
forwarding and are called forwarding references.

template<typename T>
void foo(T &&); // forwarding reference here

template<typename T>
void bar(std::vector<T> &&); // but not here

The rules allow the function  foo  above to be called with either an  lvalue  or an  rvalue :

 When called with an  lvalue  of type  X , then  T  resolves to  X &
 When called with and  rvalue  of type  X , then  T  resolves to  X
When applying these rules we end up with an argument being  X & && . So there are even more
rules to collapse the outcome:

 X & &  collapses to  X &

 X & &&  collapses to  X &

 X && &  collapses to  X &

 X && &&  collapses to  X &&

Combining the two rules we can have:

template<typename T>
void fn(T &&) { std::cout<< "template\n"; }
 int main()
{
X a;
fn(a);
// argument expression is lvalue of type X
// resolves to T being X &
// X & && collapses to X &

fn(X());
// argument expression is rvalue of type X
// resolves to T being X
// X && stays X &&
}

static_cast<X &&>
Once we have an expression of a value category, we can convert it to an expression of a different
value category. If we have a  rvalue  we can assign it to a variable, or take a reference, hence
becoming a  lvalue . If we have a  lvalue  we can return it from a function, so we get a  rvalue .

But one important rule is that: one can covert from a  lvalue  to a  rvalue  (to an  xvalue  more
precisely) by using static_cast<X &&> without creating temporaries. And this is the last
piece of the puzzle to understand  std::move  and  std::forward .

std::move
The idiomatic use of  std::move  is to ensure the argument passed to a function is an  rvalue  so
that you can move from it (choose move semantics). By function I mean an actual function or a
constructor or an operator (e.g. assignment operator).

Here is an example where  std::move  is used twice in an idiomatic way:

struct X
{
X(){}

X(const X & other) : s_{ other.s_ } {}

X(X && other) : s_{ std::move(other.s_) } {}

// other is an lvalue and so it's other.s_
// use std::move to force using the move constructor for s_
// don't use other.s_ after std::move (other than to destruct)
 std::string s_;
};

int main()
{
X a;

X b = std::move(a);
// a is an lvalue
// use std::move to convert to a rvalue,
// xvalue to be precise,
// so that the move constructor for X is used
// don't use a after std::move (other than to destruct)
}

Here is a possible implementations for  std::move .

template<typename T> struct remove_reference { typedef T type; };

template<typename T> struct remove_reference<T&> { typedef T type; };
template<typename T> struct remove_reference<T&&> { typedef T type; };

template<typename T>
constexpr typename remove_reference<T>::type && move(T && arg) noexcept
{
return static_cast<typename remove_reference<T>::type &&>(arg);
}

First of all  std::move  is a template taking a  rvalue reference  argument which means that it can
be called with either a  lvalue  or an  rvalue , and the reference collapsing rules apply.

Because the type  T  is deduced, we did not have to specify when using  std::move .

Then all it does is a  static_cast .

The  remove_reference  template specializations are used to get the underlying type for  T  without
any references, and that type is decorated with  &&  for the  static_cast  and return type.

In conclusion  std::move  does not move, all it does is to return a  rvalue  so that the function that
actually moves, eventually receiving a  rvalue reference , is selected by the compiler.

std::forward
The idiomatic use of  std::forward  is inside a templated function with an  rvalue argument ,
where the argument is now  lvalue , used to retrieve the original value category, that it was called
with, and pass it on further down the call chain (perfect forwarding).
Here is an example where  std::forward  is used twice in an idiomatic way:

struct Y
{
Y(){}
Y(const Y &){ std::cout << "Copy constructor\n"; }
Y(Y &&){ std::cout << "Move constructor\n"; }
};

struct X
{
template<typename A, typename B>
X(A && a, B && b) :
// retrieve the original value category from constructor call
// and pass on to member variables
a_{ std::forward<A>(a) },
b_{ std::forward<B>(b) }
{
}

Y a_;
Y b_;
};

template<typename A, typename B>

X factory(A && a, B && b)
{
// retrieve the original value category from the factory call
// and pass on to X constructor
return X(std::forward<A>(a), std::forward<B>(b));
}

int main()
{
Y y;
X two = factory(y, Y());
// the first argument is a lvalue, eventually a_ will have the
// copy constructor called
// the second argument is an rvalue, eventually b_ will have the
// move constructor called
}

// prints:
// Copy constructor
// Move constructor

Here is a possible implementations for  std::forward .

 template<typename T> struct is_lvalue_reference { static constexpr bool value
= false; };
template<typename T> struct is_lvalue_reference<T&> { static constexpr bool
value = true; };

template<typename T>
constexpr T&& forward(typename remove_reference<T>::type & arg) noexcept
{
return static_cast<T&&>(arg);
}

template<typename T>
constexpr T&& forward(typename remove_reference<T>::type && arg) noexcept
{
static_assert(!is_lvalue_reference<T>::value, "invalid rvalue to lvalue
conversion");
return static_cast<T&&>(arg);
}

First of all  std::forward  is more complex than  std::move . This version is the result of several
iterations.

The type  T  is not deduced, therefore we had to specify it when using  std::forward .

Then all it does is a  static_cast .

The  static_assert  is there to stop at compile time attempts to convert from an  rvalue  to
an  lvalue  (that would have the dangling reference problem: a reference pointing to a temporary
long gone). This is explained in more details in N2835, but the gist is:

forward<const Y&>(Y()); // does not compile

// static assert in forward triggers compilation failure for line above
// with "invalid rvalue to lvalue conversion"

Some non-obvious properties of  std::forward  are that the return value can be more cv-qualified
(i.e. can add a  const ). Also it allows for the case where the argument and return are different e.g.
to forward expressions from derived type to it’s base type (even some scenarios where the base is
derived from as  private ).

Conclusion
Going back to the code we started with:

std::map<std::string, std::function<void()>> commands;

template<typename ftor>
 void install_command(std::string name, ftor && handler)
{
commands.insert({
std::move(name),
std::forward<ftor>(handler)
});
}

The first parameter,  name , for the function  install_command  is passed by value. That is realy a
temporary, but has a name, hence it’s an  lvalue  expression inside  install_command . The second
parameter  handler  is a  rvalue reference . Because it has a name, it’s an  lvalue  expression as
well inside  install_command .

The  std::map  has an  insert  overload that accepts an templated  rvalue reference  for the
key/value pair to insert. For the key we can provide an  rvalue  using  std::move  because really
we don’t need  name  any more. If we did not use  std::move  we would do a silly copy. For the
value we provide whatever we the  install_command  was called with for the  handler . We
use  std::forward  to retrieve the original value category. If for the  handler  we provided
an  rvalue  then  insert  will move from it. If for the  handler  we provided
an  lvalue  then  insert  will copy it.

There are a lot of rules that come into play for the initial deceivingly simple code. They are the
result of maintaining backward compatibility and plumbing move semantics and perfect
forwarding support on top of that, while making it so that most common scenarios are easy to
write and read.

References
 Stroustrup on C++ 11 value category
classification: http://www.stroustrup.com/terminology.pdf
 Cppreference with details on value
category: http://en.cppreference.com/w/cpp/language/value_category
 Thomas Becker on rvalue
references: http://thbecker.net/articles/rvalue_references/section_08.html
 MSDN on rvalue references: https://msdn.microsoft.com/en-us/library/dd293668.aspx
 Final C++ 11 versions of std::forward and std::move: http://www.open-
std.org/jtc1/sc22/wg21/docs/papers/2010/n3143.html
 Use cases for std::forward: http://www.open-
std.org/jtc1/sc22/wg21/docs/papers/2009/n2951.html
 Explaining the static_assert in std::forward: http://www.open-
std.org/jtc1/sc22/wg21/docs/papers/2009/n2835.html
 On reference binding rules: http://www.open-
std.org/JTC1/SC22/WG21/docs/papers/2008/n2812.html
 Howard Hinnant on Stack Overflow: http://stackoverflow.com/a/9672202/5495780