Question

What is the Difference in between 2 statements

int main() 
{ 
    A a = new A();
    A a;
}

Please explain this two object creation statements.

Answer 1

The first command allocates a variable on the stack (A a), and initializes on the heap (new A()).

The second one only allocates the variable on the stack. It is not initialized, and therefore cannot be used until you assign it, either by a return value from a function or calling the class constructor.

---

Side Note: When your program is compiled and being run, it doesn’t even remotely resemble the code you wrote. Variables are loaded right before you need them. The code you have above would roughly look like this:

.method private hidebysig static void  Main(string[] args) cil managed
{
  .entrypoint
  // Code size       8 (0x8)
  .maxstack  1
  .locals init ([0] class DERP.Program/A a,   // This code here declares a local
                                              // variable: A a;
           [1] class DERP.Program/A b)        // another local variable: A b;
   nop
   newobj     instance void DERP.Program/A::.ctor()  // This is: new A()
   stloc.1                           // this loads the new A() we created
                                     // into A b;  (stloc.1 means to store
                                     // the last item we created into the
                                     // local variable at index [1]
   ret
} // end of method Program::Main

It’s okay if you don’t completely understand what each of those commands are, but I’ve commented it to try and make it as straight forwards as possibly. Since we never assign A a to anything, it just sits on the local stack, hanging out, doing nothing. We can’t use it because it doesn’t point to any object.

---

There are many different reason in code you might see the declaration separated from the actual assignment.

For example, sometimes you need to declare a variable outside of a try {} catch {} clause. Lets say your class takes a value in its constructor. You have a function that gets that data, say from a database. However, since its a DB call, you want to catch the exception, and if its thrown, initialize the class with a default value, instead of the returned value from the DB call.

Due to the way scoping works in C#, variables declared inside a try {} catch {} are not accessible outside of it, hence you would need to declare the variable earlier in code before you initialize it.