Explicit conversion

[Stub]

Docs says that "The compiler does not use an explicit constructor to
implement an implied conversion of types. It's purpose is reserved
explicitly for construction."

I put up code of three cases at the bottom. Hope you can help me understand
the "explicit" keyword and its usage. Specifically,

Is "explicit" keyword only associated with constructor in C++?

What's "implied conversion of types"?

Is "ExClass Ex=5;" so-called implied conversion since it converts an int
into ExClass type? If so then why "ExClass Ex(5);" is not since it calls a
conversion constructor?


Thank you for your help!


Case 1:
---------------
struct ExClass{
explicit ExClass(int) { cout << "explicit conversion"; }
ExClass(long) { cout << "implicit conversion"; }
};

int main (){
ExClass Ex=5;
return 0;
}

OUTPUT:
implicit conversion


Case 2:
----------------
struct ExClass{
ExClass(int) { cout << "explicit conversion"; }
ExClass(long) { cout << "implicit conversion"; }
};

int main (){
ExClass Ex=5;
return 0;
}

OUTPUT:
explicit conversion


Case 3:
-------------------
struct ExClass{
ExClass(int) { cout << "explicit conversion"; }
ExClass(long) { cout << "implicit conversion"; }
};

int main (){
ExClass Ex(5);
return 0;
}

OUTPUT:
explicit conversion

 

[Michael Kochetkov]

Docs says that "The compiler does not use an explicit constructor to
implement an implied conversion of types. It's purpose is reserved
explicitly for construction."

I put up code of three cases at the bottom. Hope you can help me understand
the "explicit" keyword and its usage. Specifically,

Is "explicit" keyword only associated with constructor in C++?

I would read it as "... with a constructor in C++?" if you do not mind. Then
no, the "explicit" keyword deals with constructors declarations only withing
class declarations.

What's "implied conversion of types"?

Let us say, an object "a" of type A may be implicitly converted to a type B
if and only if the B b = a; declaration exists.

Is "ExClass Ex=5;" so-called implied conversion since it converts an int
into ExClass type?

If you do not mind the definition above then it is so by definition.

If so then why "ExClass Ex(5);" is not since it calls a
conversion constructor?

The expression above is the direct initialization. Direct initializations
are a kind of explicit initialization.

Thank you for your help!


Case 1:
---------------
struct ExClass{
explicit ExClass(int) { cout << "explicit conversion"; }
ExClass(long) { cout << "implicit conversion"; }
};

int main (){
ExClass Ex=5;
return 0;
}

OUTPUT:
implicit conversion

Yes, it is. It is so by definition. I see, that your compiler is not an
Intel-made one. Though it looks like other EDG-base compilers can handle it
right.

Case 2:
----------------
struct ExClass{
ExClass(int) { cout << "explicit conversion"; }
ExClass(long) { cout << "implicit conversion"; }
};

int main (){
ExClass Ex=5;
return 0;
}

OUTPUT:
explicit conversion

ExClass is the best conversion as far as 5 is of type int.

Case 3:
-------------------
struct ExClass{
ExClass(int) { cout << "explicit conversion"; }
ExClass(long) { cout << "implicit conversion"; }
};

int main (){
ExClass Ex(5);
return 0;
}

OUTPUT:
explicit conversion

Direct initialization is the explicit conversion indeed.
Other explicit conversions are:
explicit cast conversion in cast notation: ExClass Ex = (ExClass)5;
static_cast expression: ExClass Ex = static_cast<ExClass>(5);

I cannot think out the explicit cast conversion in functional notation -- it
constantly leads my to direct initialization. But it might be applicable
too.