question on r value reference

[chrisben]

Hi,
Could someone explain?

For the following statement, why void foo(X const &) can be called on both r and l value, but void foo(X&) cannot. I faintly remember it is about reference binding on temporary, but not clear on the concept now.

Thanks


"
If you implement
void foo(X&);
but not
void foo(X&&);
then of course the behavior is unchanged: foo can be called on l-values, but not on r-values.

If you implement

void foo(X const &);

but not

void foo(X&&);

then again, the behavior is unchanged: foo can be called on l-values and r-values, but it is not possible to make it distinguish between l-values and r-values.
"

[helios]

For example, suppose X is int and you define void foo(int &). Now suppose it's possible to bind int & to an rvalue. What should happen if you call foo(0)?

[chrisben]

that will be a compiling error for temp variable. So what magic does void foo(int const &) do? Keep the lifetime of the temp through the lifetime of the function?

[helios]

Well, yes, of course. It can be trivially shown that temporaries passed as function parameters live until their callee ends. There's nothing magic about that.

struct A{
    A(){ std::cout << "A::A()\n"; }
    ~A(){ std::cout << "A::~A()\n"; }
};

void foo(const A &a){
    std::cout << "foo()\n";
}

foo(A());

[chrisben]

thanks