I've been studying rvalue references lately and came to a conclusion that it's quite advantageous to use pass-by-value everywhere where complete copy of an object will be made (for complete justification see e.g. How to reduce redundant code when adding rvalue reference operator overloads? and Want speed? Pass by value!), because the compiler can automatically optimize a copy away in cases such as f(std::move(a));
, where f
is defined as void f(A a);
.
One negative consequence of pass-by-value-everywhere is that all the code becomes littered with std::move
even in simple cases such as:
void Object::value(A a)
{
value_ = std::move(a);
}
Obviously, if I wrote only the following:
void Object::value(A a)
{
value_ = a;
}
it shouldn't be hard for the compiler to recognize that a
is near the end of its lifetime even without the hint and not to penalize me with additional copy. In fact, the compiler should be able to recognize this even in complex functions.
The questions:
Is this optimization allowed by the C++0x Standard?
Do the compilers employ it? Even in complex cases, i.e. the function consists from more than one line?
How reliable is this optimization, i.e. can I expect the compiler to utilize it as much as I expect the compiler to apply Return Value Optimization?