Sizeof

Application to arrays

When sizeof is applied to the name of an array, the result is the number of bytes required to store the entire array. This is one of the few exceptions to the rule that the name of an array is converted to a pointer to the first element of the array, and is possible just because the actual array size is fixed and known at compile time, when the sizeof operator is evaluated. The following program uses sizeof to determine the size of a declared array, avoiding a buffer overflow when copying characters:

#include <stdio.h>

int main(int argc, char **argv)
{
  char buffer[10]; /* Array of 10 chars */

  /* Copy at most 9 characters from argv[1] into buffer,
   * null-terminate the buffer. */
  snprintf(buffer, sizeof buffer, "%s", argv[1]);

  return 0;
}

Here, sizeof buffer is equivalent to 10 * sizeof buffer [0], which evaluates to 10, because the size of the type char is defined as 1.

C99 adds support for flexible array members to structures. This form of array declaration is allowed as the last element in structures only, and differs from normal arrays in that no length is specified to the compiler. For a structure named s containing a flexible array member named a, sizeof s is therefore equivalent to offsetof (s, a):

#include <stdio.h>

struct flexarray {
    char val;
    int array[];  /* Flexible array member; must be last element of struct */
};

int main(int argc, char **argv)
{
    printf("sizeof (struct flexarray) == %zu\n", sizeof (struct flexarray));
    return 0;
}

In this case the sizeof operator returns the size of the structure, including any padding, but without any storage allowed for the array. Most platforms produce the following output:

sizeof (struct flexarray) == 4

C99 also allows variable length arrays that have the length specified at runtime,^[2] although the feature is considered an optional implementation in later versions of the C standard. In such cases, the sizeof operator is evaluated in part at runtime to determine the storage occupied by the array.

#include <stddef.h>

size_t flexsize(int n)
{
   char b[n + 3];   /* Variable length array */
   return sizeof b;  /* Execution time sizeof */
}

int main(void)
{
  size_t size = flexsize(10); /* flexsize returns 13 */
  return 0;
}

sizeof can be used to determine the number of elements in an array, by dividing the size of the entire array by the size of a single element. This should be used with caution; When passing an array to another function, it will "decay" to a pointer type. At this point, sizeof will return the size of the pointer, not the total size of the array. As an example with a proper array:

int main (void)
{
   int tab [10];
   printf ("Number of elements in the array: %zu\n", sizeof tab / sizeof tab [0]); /* yields 10 */
   return 0;
}

Incomplete types

sizeof can only be applied to "completely" defined types. With arrays, this means that the dimensions of the array must be present in its declaration, and that the type of the elements must be completely defined. For structs and unions, this means that there must be a member list of completely defined types. For example, consider the following two source files:

/* file1.c */
int arr [10];
struct x {int one; int two;};
/* more code */

/* file2.c */
extern int arr [];
struct x;
/* more code */

Both files are perfectly legal C, and code in file1.c can apply sizeof to arr and struct x. However, it is illegal for code in file2.c to do this, because the definitions in file2.c are not complete. In the case of arr, the code does not specify the dimension of the array; without this information, the compiler has no way of knowing how many elements are in the array, and cannot calculate the array's overall size. Likewise, the compiler cannot calculate the size of struct x because it does not know what members it is made up of, and therefore cannot calculate the sum of the sizes of the structure's members (and padding). If the programmer provided the size of the array in its declaration in file2.c, or completed the definition of struct x by supplying a member list, this would allow the application of sizeof to arr or struct x in that source file.

Object members

C++11 introduced the possibility to apply the sizeof parameter to specific members of a class without the necessity to instantiate the object to achieve this.^[3] The following example for instance yields 4 and 8 on most platforms.

#include <iostream>

struct foo {
  int a;
  int b;
};

int main ()
{
  std::cout << sizeof foo::a << "\n" << sizeof (foo) << "\n";
}

Variadic template packs

C++11 introduced variadic templates; the keyword sizeof followed by ellipsis returns the number of elements in a parameter pack.

template <typename... Args>
void print_size (Args... args)
{
  std::cout << sizeof... (args) << "\n";
}

int main ()
{
  print_size (); // outputs 0
  print_size ("Is the answer", 42, true); // outputs 3
}

Structure padding

When calculating the size of any object type, the compiler must take into account any required data structure alignment to meet efficiency or architectural constraints. Many computer architectures do not support multiple-byte access starting at any byte address that is not a multiple of the word size, and even when the architecture allows it, usually the processor can fetch a word-aligned object faster than it can fetch an object that straddles multiple words in memory.^[4] Therefore, compilers usually align data structures to at least a word boundary, and also align individual members to their respective boundaries. In the following example, the structure student is likely to be aligned on a word boundary, which is also where the member grade begins, and the member age is likely to start at the next word address. The compiler accomplishes the latter by inserting padding bytes between members as needed to satisfy the alignment requirements. There may also be padding at the end of a structure to ensure proper alignment in case the structure is used as an element of an array.

Thus, the aggregate size of a structure in C can be greater than the sum of the sizes of its individual members. For example, on many systems the following code prints 8:

struct student {
  char grade; /* char is 1 byte long */
  int age; /* int is 4 bytes long */
};

printf ("%zu", sizeof (struct student));