CODEWITHSUNDEEP
c++
q0987
q0987

Reputation: 35982

which return comply with C++ standard

class A {};

class B 
{
public:
    B(const A& a, int i = 10) : m_a(a), m_i(i) {}
private:
    int m_i;
    A m_a;
};

B getB(void)
{
    //return B(A());  // Method one
    //return A();     // Method two
}

Both method one and method two pass the compilation of VS2010.

Question 1> Which one is better?

Question 2> Is it true that an implicit constructor supports more than one pass-in parameters if all except the first parameters have default values?

Thank you

Upvotes: 2

Views: 82

Answers (3)

Kerrek SB
Kerrek SB

Reputation: 477188

I would decorate the constructor as explicit, and then use method one as the only available method:

class B {
public:
    explicit B(A const & a, int b = 10) : m_a(a), m_i(b) { }
    // ...
};

B foo()  { return B(A()); }

That way you can never accidentally construct a B from an A. If the constructor of B is expensive or may throw, then having this extra level of deliberateness in your code may well help make it more readable and less error-prone.

Upvotes: 1

user405725
user405725

Reputation:

Both are correct. First one explicitly creates B instance, and the second one implicitly creates B from A. I like neither of those. They just increase confusion level and nothing more.

Upvotes: 1

Billy ONeal
Billy ONeal

Reputation: 106569

Both are valid and both do the same thing. I'd use the first one because I find the second one confusing, but different strokes for different folks.

EDIT: Actually, I'd do even more than that. I'd mark B's constructor as explicit in order to prevent just the sort of thing that makes the second one compile.

Note: I would avoid use of (void) in method signatures in C++.

Upvotes: 2

Related Questions