Section VI: Extended Integer Types

As described in Chapter 3, "Data and C," the C99 header file inttypes.h provides a systematic set of alternative names for the various integer types. These names describe the properties of the type more clearly than do the standard names. For example, type int might be 16 bits, 32 bits, or 64 bits, but the int32_t type always is 32 bits.

More precisely, the inttypes.h header file defines macros that can be used with scanf() and printf() to read and write integers of these types. This header file includes the stdlib.h header file, which provides the actual type definitions. The formatting macros are strings that can be concatenated with other strings to produce the proper formatting directions.

The types are defined using typedef. For example, a system with a 32-bit int might use this definition:

typedef int int32_t;

The format specifiers are defined using the #define directive. For example, a system using the previous definition for int32_t might have this definition:

#define PRId32 "d"   // output specifier

#define SCNd32 "d"   // input specifier

Using these definitions, you could declare an extended integer variable, input a value, and display it as follows:

int32_t cd_sales;  // 32-bit integer

scanf("%" SCNd32, &cd_sales);

printf("CD sales = %10" PRId32 " units\n", cd_sales);

String concatenation then combines strings, if needed, to get the final control string. Thus, the previous code gets converted to the following:

int cd_sales;  // 32-bit integer

scanf("%d", &cd_sales);

printf("CD sales = %10d units\n", cd_sales);

If you moved the original code to a system with a 16-bit int, that system might define int32_t as long, PRId32 as "ld", and SCNd32 as "ld". But you could use the same code, knowing that it uses a 32-bit integer.

The rest of this reference section lists the extended types along with the format specifiers and macros representing the type limits.

Exact-Width Types

One set of typedefs identify types with precise sizes. The general form is intN_t for signed types and uintN_t for unsigned types, with N indicating the number of bits. Note, however, that not all systems can support all the types. For example, there could be a system for which the smallest usable memory size is 16 bits; such a system would not support the int8_t and uint8_t types. The format macros can use either d or i for the signed types, so PRIi8 and SCNi8 also work. For the unsigned types, you can substitute o, x, or X for u to obtain the %o , %x, or %X specifier instead of %u. For example, you can use PRIX32 to print a uint32_t type value in hexadecimal format. Table RS.VI.1 lists the exact-width types, format specifiers, and value limits.

Table RS.VI.1. Exact-Width Types

Type Name	printf() Specifier	scanf() Specifier	Minimum Value	Maximum Value
int8_t	PRId8	SCNd8	INT8_MIN	INT8_MAX
int16_t	PRId16	SCNd16	INT16_MIN	INT16_MAX
int32_t	PRId32	SCNd32	INT32_MIN	INT32_MAX
int64_t	PRId64	SCNd64	INT64_MIN	INT64_MAX
uint8_t	PRIu8	SCNu8	0	UINT8_MAX
uint16_t	PRIu16	SCNu16	0	UINT16_MAX
uint32_t	PRIu32	SCNu32	0	UINT32_MAX
uint64_t	PRIu64	SCNu64	0	UINT64_MAX

Minimum-Width Types

The minimum-width types guarantee a type that is at least a certain number of bits in size. These types always exist. For example, a system that does not support 8-bit units could define int_least_8 as a 16-bit type. Table RS.VI.2 lists minimum-width types, format specifiers, and value limits.

Table RS.VI.2. Minimum-Width Types

TypeName	printf() Specifier	scanf() Specifier	Minimum Value	Maximum Value
int_least8_t	PRILEASTd8	SCNLEASTd8	INT_LEAST8_MIN	INT_LEAST8_MAX
int_least16_t	PRILEASTd16	SCNLEASTd16	INT_LEAST16_MIN	INT_LEAST16_MAX
int_least32_t	PRILEASTd32	SCNLEASTd32	INT_LEAST32_MIN	INT_LEAST32_MAX
int_least 64_t	PRILEASTd64	SCNLEASTd64	INT_LEAST64_MIN	INT_LEAST64_MAX
uint_least 8_t	PRILEASTu8	SCNLEASTu8	0	UINT_LEAST8_MAX
uint_least 16_t	PRILEASTu16	SCNLEASTu16	0	UINT_LEAST16_MAX
uint_least 32_t	PRILEASTu32	SCNLEASTu32	0	UINT_LEAST32_MAX
uint_least 64_t	PRILEASTu64	SCNLEASTu64	0	UINT_LEAST64_MAX

Fastest Minimum-Width Types

For a particular system, some integer representations can be faster than others. For example, int_least16_t might be implemented as short, but the system might do arithmetic faster using type int. So inttypes.h also defines the fastest type for representing at least a certain number of bits. These types always exist. In some cases, there might be no clear-cut choice for fastest; in that case, the system simply specifies one of the choices. Table RS.VI.3 lists fastest minimum-width types, format specifiers, and value limits.

Table RS.VI.3. Fastest Minimum-Width Types

Type Name	printf() Specifier	scanf() Specifier	Minimum Value	Maximum Value
int_fast8_t	PRIFASTd8	SCNFASTd8	INT_FAST8_MIN	INT_FAST8_MAX
int_fast16_t	PRIFASTd16	SCNFASTd16	INT_FAST16_MIN	INT_FAST16_MAX
int_fast32_t	PRIFASTd32	SCNFASTd32	INT_FAST32_MIN	INT_FAST32_MAX
int_fast 64_t	PRIFASTd64	SCNFASTd64	INT_FAST64_MIN	INT_FAST64_MAX
uint_fast 8_t	PRIFASTu8	SCNFASTu8	0	UINT_FAST8_MAX
uint_fast 16_t	PRIFASTu16	SCNFASTu16	0	UINT_FAST16_MAX
uint_fast 32_t	PRIFASTu32	SCNFASTu32	0	UINT_FAST32_MAX
uint_fast 64_t	PRIFASTu64	SCNFASTu64	0	UINT_FAST64_MAX

Maximum-Width Types

Sometimes you may want the largest integer type available. Table RS.VI.4 lists these types. They may, in fact, be wider than long long or unsigned long long, because a system may provide additional types wider than the required types.

Table RS.VI.4. Maximum-Width Types

Type Name	printf() Specifier	scanf() Specifier	Minimum Value	Maximum Value
intmax_t	PRIdMAX	SCNdMAX	INTMAX_MIN	INTMAX_MAX
uintmax_t	PRIuMAX	SCBuMAX	0	UINTMAX_MAX

Integers That Can Hold Pointer Values

The inttypes.h header file (via the included stdint.h header file) defines two integer types, listed in Table RS.VI.5, that can hold pointer values accurately. That is, if you assign a type void * value to one of these types, and then assign the integer type back to the pointer, no information is lost. Either or both types might not exist.

Table RS.VI.5. Integer Types for Holding Pointer Values

Type Name	printf() Specifier	scanf() Specifier	Minimum Value	Maximum Value
intptr_t	PRIdPTR	SCNdPTR	INTPTR_MIN	INTPTR_MAX
uintptr_t	PRIuPTR	SCBuPTR	0	UINTPTR_MAX

Extended Integer Constants

You can indicate a long constant with the L suffix, as in 445566L. How do you indicate that a constant is type int32_t? Use macros defined in inttypes.h. For example, the expression INT32_C(445566) expands to a type int32_t constant. Essentially, the macro is a type cast to the underlying type—that is, to the fundamental type that represents int32_t in a particular implementation.

The macro names are formed by taking the type name, replacing the _t with _C, and making all the letters uppercase. For example, to make 1000 a type uint_least64_t constant, use the expression UINT_LEAST64_C(1000).