Question

looking at the source code of GMPlib (the Gnu library for multy-precision calculations) I found this kind of code for building its mp*_t structures. And I have replicated it in various other works I have done, but I don’t fully understand it.

typedef struct
{
  int _mp_alloc;           /* Number of *limbs* allocated and pointed
                              to by the _mp_d field.  */
  int _mp_size;            /* abs(_mp_size) is the number of limbs the
                              last field points to.  If _mp_size is
                              negative this is a negative number.  */
  mp_limb_t *_mp_d;        /* Pointer to the limbs.  */
} __mpz_struct;

I understand that this defines the ‘shape’ of a structure with two integers and a mp_limb_t and typedefs it to __mpz_struct

Then comes this line:

typedef __mpz_struct mpz_t[1];

And after a while, this other one:

typedef __mpz_struct *mpz_ptr;

I understand that the second one is typedefining __mpz_struct * to mpz_ptr (Which is used in the function prototypes)

But I don’t understand what the first one does and why it works so I can declare a mpz_t. Can anyone explain why it works?

Thanks!

Answer 1

The simplest way to understand a typedef is the rule Define it like you use it. In other words, when you declare a typedef you’re making a new name, but should you remove the typedef keyword you’ll get an object of the aliased type.

For example:

typedef __mpz_struct mpz_t[1];

This makes a new type name, called “mpz_t”. But if we rewrite:

__mpz_struct some_obj[1];

we immediately understand it makes an array of whatever __mpz_struct is. So now we know what the new type mpz_t a does: it makes an array of __mpz_struct with one element.