How new differentiates between PODs and user defined data types?

Question

First:

int *p = new int;

Second:

class A{};

A *pa = new A;

How does new and compiler determines when to call constructor? In first case compiler does not generate code to call constructor of p and in second case it generates code to call constructor of A. Which mechanism is used to make such choice?

Your second code snippet *does not* call any constructor of A (5.3.4/15, we're in the case where "the object created has indeterminate value"), although since the compiler-generated default constructor does nothing, that's a bit of a moot point in this example. See http://stackoverflow.com/questions/620137/do-the-parentheses-after-the-type-name-make-a-difference-with-new — Steve Jessop, Dec 17 '10 at 12:36

score 7 · Accepted Answer · answered Dec 17 '10 at 11:19

7

The compiler knows that A is a class, because it has seen the class declaration, so it uses the synthesised default constructor. It knows an int is an int, because the language grammar says it is.

answered Dec 17 '10 at 11:19

unquiet mind

1,082
6
11

score 6 · Answer 2 · edited May 23 '17 at 11:47

In your case A is a POD as well. To learn the correct definition of POD's take a look at this.

As far as your code is concerned, the compiler knows that int is a built-in type and doesn't have a constructor whatsoever.

Edit: Your question is rather strange. The compiler knows which type is a pod, and which isn't, also it knows which are built-in and not-built in because it is the compiler that compiles your code :) If the compiler didn't know that information, who would?

How new differentiates between PODs and user defined data types?

2 Answers2