Temporary objects - when are they created, how do

Question

In Eckel, Vol 1, pg:367

//: C08:ConstReturnValues.cpp
// Constant return by value
// Result cannot be used as an lvalue
class X {
   int i;
public:
   X(int ii = 0);
   void modify();
};

X::X(int ii) { i = ii; }

void X::modify() { i++; }

X f5() {
   return X();
}

const X f6() {
   return X();
}

void f7(X& x) { // Pass by non-const reference
   x.modify();
}

int main() {
   f5() = X(1); // OK -- non-const return value
   f5().modify(); // OK
// Causes compile-time errors:
//! f7(f5());
//! f6() = X(1);
//! f6().modify();
//! f7(f6());
} ///:~

Why does f5() = X(1) succed? What is going on here???

Q1. When he does X(1) - what is going on here? Is this a constructor call - shouldn't this then read X::X(1); Is it class instantiation - isn't class instantiation something like: X a(1); How does the compiler determine what X(1) is?? I mean.. name decoration takes place so.. X(1) the constructor call would translate to something like: globalScope_X_int as the function name.. ???

Q2. Surely a temporary object is used to store the resulting object that X(1) creates and then would't that be then assigned to the object f5() returns (which would also be a temporary object)? Given that f5() returns a temporary object that will be soon be discarded, how can he assign one constant temporary to another constant temporary??? Could someone explain clearly why: f7(f5()); should reult in a constant temporary and not plain old f5();

Answer 1

1. This is indeed a constructor call, an expression evaluating to a temporary object of type X. Expressions of the form X([...]) with X being the name of a type are constructor calls that create temporary objects of type X (though I don't know how to explain that in proper standardese, and there are special cases where the parser can behave differently). This is the same construct you use in your f5 and f6 functions, just omitting the optional ii argument.

2. The temporary created by X(1) lives (doesn't get destructed/invalid) until the end of the full expression containing it, which usually means (like in this case with the assignment expression) until the semicolon. Likewise does f5 create a temporary X and return it to the call site (inside main), thus copying it. So in main the f5 call also returns a temporary X. This temporary X is then assigned the temporary X created by X(1). After that is done (and the semicolon reached, if you want), both temporaries get destroyed. This assignment works because those functions return ordinary non-constant objects, no matter if they are just temprorary and destroyed after the expression is fully evaluated (thus making the assignment more or less senseless, even though perfectly valid).

   It doesn't work with f6 since that returns a const X onto which you cannot assign. Likewise does f7(f5()) not work, since f5 creates a temporary and temporary objects don't bind to non-const lvalue references X& (C++11 introduced rvalue references X&& for this purpose, but that's a different story). It would work if f7 took a const reference const X&, as constant lvalue references bind to temporaries (but then f7 itself wouldn't work anymore, of course).

Answer 2

All your questions boil down to a rule in C++ which says that a temporary object (one that has no name) cannot be bound to a non-const reference. (Because Stroustrup felt it could provoke logical errors...)

The one catch, is that you can invoke a method on a temporary: so X(1).modify() is fine but f7(X(1)) is not.

As for where the temporary is created, this is the compiler job. The rules of the language precise that the temporary should only survive until the end of the current full-expression (and no longer) which is important for temporary instances of classes whose destructor has a side-effect.

Therefore, the following statement X(1).modify(); can be fully translated to:

{
    X __0(1);
    __0.modify();
} // automatic cleanup of __0

With that in mind, we can attack f5() = X(1);. We have two temporaries here, and an assignment. Both arguments of the assignment must be fully evaluated before the assignment is called, but the order is not precise. One possible translation is:

{
    X __0(f5());
    X __1(1);
    __0.operator=(__1);
}

(the other translation is swapping the order in which __0 and __1 are initialized)

And the key to it working is that __0.operator=(__1) is a method invocation, and methods can be invoked on temporaries :)

Answer 3

Here is an example what actually happens when you execute your code. I've made some modifications to clarify the processes behind the scene:

#include <iostream>

struct Object
{
    Object( int x = 0 ) {std::cout << this << ": " << __PRETTY_FUNCTION__ << std::endl;}
    ~Object() {std::cout << this << ": " << __PRETTY_FUNCTION__ << std::endl;}
    Object( const Object& rhs ){std::cout << this << ": " << __PRETTY_FUNCTION__ << " rhs = " << &rhs << std::endl;}
    Object& operator=( const Object& rhs )
    {
        std::cout << this << ": " << __PRETTY_FUNCTION__ << " rhs = " << &rhs << std::endl;
        return *this;
    }
    static Object getObject()
    {
        return Object();
    }
};

void TestTemporary()
{
    // Output on my machine
    //0x22fe0e: Object::Object(int) -> The Object from the right side of = is created Object();
    //0x22fdbf: Object::Object(int) -> In getObject method the Temporary Unnamed object is created
    //0x22fe0f: Object::Object(const Object&) rhs = 0x22fdbf -> Temporary is copy-constructed from the previous line object
    //0x22fdbf: Object::~Object() -> Temporary Unnamed is no longer needed and it is destroyed
    //0x22fe0f: Object& Object::operator=(const Object&) rhs = 0x22fe0e -> assignment operator of the returned object from getObject is called to assigne the right object
    //0x22fe0f: Object::~Object() - The return object from getObject is destroyed
    //0x22fe0e: Object::~Object() -> The Object from the right side of = is destroyed Object();

    Object::getObject() = Object();
}

You have to know that on most modern compilers the copy construction will be avoided. This is because the optimization which is made (Return Value Optimization) by the compiler. In my output I have removed explicitly the optimization to show what actually happens according to the standard. If you want to remove this optimization too use the following option:

-fno-elide-constructors

Answer 4

I wasn't entirely satisfied by the answers, so I took a look at:

"More Effective C++", Scott Meyers. Item 19: "Understand the origin of temporary Objects"

. Regarding Bruce Eckel's coverage of "Temporaries", well, as I suspect and as Christian Rau directly points out, it's plain wrong! Grrr! He's (Eckel's) using us as guinea pigs!! (it would be a good book for newbies like me once he corrects all his mistakes)

Meyer: "True temporary objects in C++ are invisible - they don't appear in your source code. They arise whenever a non-heap object is created but not named. Such unnamed objects usually arise in one of two situations: when implicit type conversions are applied to make function calls succeed and when functions return objects."

"Consider first the case in which temporary objects are created to make function calls succeed. This happens when the type of object passed to a function is not the same as the type of the parameter to which it is being bound."

"These conversions occur only when passing objects by value or when passing to a reference-to-const parameter. They do not occur when passing an object to a reference-to-non-const parameter."

"The second set of circumstances under which temporary objects are created is when a function returns an object."

"Anytime you see a reference-to-const parameter, the possibility exists that a temporary will be created to bind to that parameter. Anytime you see a function returning an object, a temporary will be created (and later destroyed)."

The other part of the answer is found in: "Meyer: Effective C++", in the "Introduction":

"a copy constructor is used to initialize an object with a different object of the same type:"

String s1;       // call default constructor
String s2(s1);   // call copy constructor
String s3 = s2;  // call copy constructor

"Probably the most important use of the copy constructor is to define what it means to pass and return objects by value."

Regarding my questions:

f5() = X(1) //what is happening?

Here a new object isn't being initialized, ergo this is not initialization(copy constructor): it's an assignment (as Matthieu M pointed out).

The temporaries are created because as per Meyer (top paragraphs), both functions return values, so temporary objects are being created. As Matthieu pointed out using pseudo-code, it becomes: __0.operator=(__1) and a bitwise copy takes place(done by the compiler).

Regarding:

void f7(X& x);
f7(f5);

ergo, a temporary cannot be created (Meyer: top paragraphs). If it had been declared: void f7(const X& x); then a temporary would have been created.

Regarding a temporary object being a constant:

Meyer says it (and Matthieu): "a temporary will be created to bind to that parameter."

So a temporary is only bound to a constant reference and is itself not a "const" object.

Regarding: what is X(1)?

Meyer, Item27, Effective C++ - 3e, he says:

"C-style casts look like this: (T)expression //cast expression to be of type T

Function-style casts use this syntax: T(expression) //cast expression to be of type T"

So X(1) is a function-style cast. 1 the expression is being cast to type X.

And Meyer says it again:

"About the only time I use an old-style cast is when I want to call an explicit constructor to pass an object to a function. For example:

class Widget {
  public:
    explicit Widget(int size);
    ...
};

void doSomeWork(const Widget& w);
doSomeWork(Widget(15)); //create Widget from int
                        //with function-style cast

doSomeWork(static_cast<Widget>(15));

Somehow, deliberate object creation doesn't "feel" like a cast, so I'd probably use the function-style cast instead of the static_cast in this case."