why a struct?

I've just started reading some material on sockets programming. These definitions in include/netinet/in.h caught my attention

Code:

---

typedef uint32_t in_addr_t;
struct in_addr
  {
    in_addr_t s_addr;
  };

---

and I'm wondering why they didn't simply define

Code:

---

typedef uint32_t in_addr;

---

?

What's the point of a struct consisting of just a single member?

Type safety.

It prevents an in_addr from being assigned any old number. Combined with a special creation function for in_addr objects, the programmer using the librairy does not even need to know that a in_addr is an 32 bit integer under the hood.

I don't get it. What prevents a programmer from assigning

Code:

---

in_addr.s_addr = 15;

---

Quote:

---

Originally Posted by R.Stiltskin

What prevents a programmer from assigning

---

The intelligence to use the interface instead of relying on implementation detail?

Ouch! :eek:

OK, first I'll learn the interface, then I'll ask the questions.

Quote:

---

Originally Posted by R.Stiltskin

Ouch!

---

Hehe, I see how it can be read now. Nah, I was not trying to insult your intelligence :p

Quote:

---

Originally Posted by R.Stiltskin

I don't get it. What prevents a programmer from assigning

Code:

---

in_addr.s_addr = 15;

---

---

Nothing. But it won't happen implicitly, and is less likely to be done accidentally.

For example, if you have a function that takes an in_addr, as the first arguement, and an int the second, if a programmer were to try to pass the arguments in the wrong order, it would lead to a compile time error.

It also allows a library developer to later change the member(s) of the struct, while being sure that the client code will still work, because the client does not directly need to access the members of the struct.