C++ Programmer

This is a frequent c++ interview question. Copy constructor is a special constructor of a class which is used to create copy of an object. Compiler will give a default copy constructor if you don't define one. This implicit constructor will copy all the members of source object to target object.

Implicit copy constructors are not recommended, because if the source object contains pointers they will be copied to target object, and it may cause heap corruption when both the objects with pointers referring to the same location does an update to the memory location. In this case its better to define a custom copy constructor and do a deep copy of the object.

class SampleClass{

public:

int* ptr;

SampleClass();

// Copy constructor declaration

SampleClass(SampleClass &obj);

};

SampleClass::SampleClass(){

ptr = new int();

*ptr = 5;

}

// Copy constructor definition

SampleClass::SampleClass(SampleClass &obj){

//create a new object for the pointer

ptr = new int();

// Now manually assign the value

*ptr = *(obj.ptr);

cout<<"Copy constructor...n";

}

C++ inline functions are special functions, for which the compiler replaces the function call with body/definition of function. Inline functions makes the program execute faster than the normal functions, since the overhead involved in saving current state to stack on the function call is avoided. By giving developer the control of making a function as inline, he can further optimize the code based on application logic. But actually, it's the compiler that decides whether to make a function inline or not regardless of it's declaration. Compiler may choose to make a non inline function inline and vice versa. Declaring a function as inline is in effect a request to the compiler to make it inline, which compiler may ignore. So, please note this point for the interview that, it is upto the compiler to make a function inline or not.

inline int min(int a, int b)

{

return (a < b)? a : b;

}

int main( )

{

cout << "min (20,10): " << min(20,10) << endl;

cout << "min (0,200): " << min(0,200) << endl;

cout << "min (100,1010): " << min(100,1010) << endl;

return 0;

}

If the complier decides to make the function min as inline, then the above code will internally look as if it was written like

int main( )

{

cout << "min (20,10): " << ((20 < 10)? 20 : 10) << endl;

cout << "min (0,200): " << ((0 < 200)? 0 : 200) << endl;

cout << "min (100,1010): " << ((100 < 1010)? 100 : 1010) << endl;

return 0;

}

A const member function is one which isn’t allowed to modify the members of the object for which it is called. A static member function is one which can’t be called for a specific object.

Thus, the const modifier for a static member function is meaningless, because there is no object associated with the call.

A more detailed explanation of this reason comes from the C programming language. In C, there were no classes and member functions, so all functions were global. A member function call is translated to a global function call. Consider a member function like this:

void foo(int i);

A call like this:

obj.foo(10);

…is translated to this:

foo(&obj, 10);

This means that the member function foo has a hidden first argument of type T*:

void foo(T* const this, int i);

If a member function is const, this is of type const T* const this:

void foo(const T* const this, int i);

Static member functions don’t have such a hidden argument, so there is no this pointer to be const or not.

When you do a shallow copy, all the fields of the source object is copied to target object as it is. That means, if there is a dynamically created field in the source object, shallow copy will copy the same pointer to target object. So you will have two objects with fields that are pointing to same memory location which is not what you usually want.

In case of deep copy, instead of copying the pointer, the object itself is copied to target. In this case if you modify the target object, it will not affect the source. By default copy constructors and assignment operators do shallow copy. To make it as deep copy, you need to create a custom copy constructor and override assignment operator.

While it’s ideal to avoid virtual inheritance altogether (you should know how your class is going to be used) having a solid understanding of how virtual inheritance works is still important:

So when you have a class (class A) which inherits from 2 parents (B and C), both of which share a parent (class D), as demonstrated below:

#include <iostream>

class D {

public:

void foo() {

std::cout << "Foooooo" << std::endl;

}

};

class C: public D {

};

class B: public D {

};

class A: public B, public C {

};

int main(int argc, const char * argv[]) {

A a;

a.foo();

}

If you don’t use virtual inheritance in this case, you will get two copies of D in class A: one from B and one from C. To fix this you need to change the declarations of classes C and B to be virtual, as follows:

class C: virtual public D {

};

class B: virtual public D {

};

We organize our C++ programs into different source files (.cpp, .cxx etc). When you consider a source file, at the preprocessing stage, some extra content may get added to the source code ( for example, the contents of header files included) and some content may get removed ( for example, the part of the code in the #ifdef of #ifndef block which resolve to false/0 based on the symbols defined). This effective content is called a translation unit. In other words, a translation unit consists of

☛ Contents of source file

☛ Plus contents of files included directly or indirectly

☛ Minus source code lines ignored by any conditional pre processing directives ( the lines ignored by #ifdef,#ifndef etc)