c++11 - Fixed-width integer literals in C++?

Question

C++11 first introduced support for defining new literals into C++ by means of user-defined literals. Does C++11 or later also predefine suffixes for fixed-width integer literals for types in <cstdint>?

Answer 1

No. As of C++14 the only literal suffixes defined by the standard are provided by <chrono>, <complex> and <string> headers in the standard library. The <chrono> header defines the h, min, s, ms, us, ns suffixes for time durations, <complex> defines the i, il and if suffixes for imaginary numbers, and <string> defines the s suffix for basic_string literals.

However, one can easily define their own fixed-width literals like this:

#include <cstdint>

constexpr std::int8_t operator "" _int8(unsigned long long v)
{ return static_cast<std::int8_t>(v); }

constexpr std::uint8_t operator "" _uint8(unsigned long long v)
{ return static_cast<std::uint8_t>(v); }

constexpr std::int16_t operator "" _int16(unsigned long long v)
{ return static_cast<std::int16_t>(v); }

constexpr std::uint16_t operator "" _uint16(unsigned long long v)
{ return static_cast<std::uint16_t>(v); }

constexpr std::int32_t operator "" _int32(unsigned long long v)
{ return static_cast<std::int32_t>(v); }

constexpr std::uint32_t operator "" _uint32(unsigned long long v)
{ return static_cast<std::uint32_t>(v); }

constexpr std::int64_t operator "" _int64(unsigned long long v)
{ return static_cast<std::int64_t>(v); }

constexpr std::uint64_t operator "" _uint64(unsigned long long v)
{ return static_cast<std::uint64_t>(v); }

constexpr std::int_fast8_t operator "" _int_fast8(unsigned long long v)
{ return static_cast<std::int_fast8_t>(v); }

constexpr std::uint_fast8_t operator "" _uint_fast8(unsigned long long v)
{ return static_cast<std::uint_fast8_t>(v); }

constexpr std::int_fast16_t operator "" _int_fast16(unsigned long long v)
{ return static_cast<std::int_fast16_t>(v); }

constexpr std::uint_fast16_t operator "" _uint_fast16(unsigned long long v)
{ return static_cast<std::uint_fast16_t>(v); }

constexpr std::int_fast32_t operator "" _int_fast32(unsigned long long v)
{ return static_cast<std::int_fast32_t>(v); }

constexpr std::uint_fast32_t operator "" _uint_fast32(unsigned long long v)
{ return static_cast<std::uint_fast32_t>(v); }

constexpr std::int_fast64_t operator "" _int_fast64(unsigned long long v)
{ return static_cast<std::int_fast64_t>(v); }

constexpr std::uint_fast64_t operator "" _uint_fast64(unsigned long long v)
{ return static_cast<std::uint_fast64_t>(v); }

constexpr std::int_least8_t operator "" _int_least8(unsigned long long v)
{ return static_cast<std::int_least8_t>(v); }

constexpr std::uint_least8_t operator "" _uint_least8(unsigned long long v)
{ return static_cast<std::uint_least8_t>(v); }

constexpr std::int_least16_t operator "" _int_least16(unsigned long long v)
{ return static_cast<std::int_least16_t>(v); }

constexpr std::uint_least16_t operator "" _uint_least16(unsigned long long v)
{ return static_cast<std::uint_least16_t>(v); }

constexpr std::int_least32_t operator "" _int_least32(unsigned long long v)
{ return static_cast<std::int_least32_t>(v); }

constexpr std::uint_least32_t operator "" _uint_least32(unsigned long long v)
{ return static_cast<std::uint_least32_t>(v); }

constexpr std::int_least64_t operator "" _int_least64(unsigned long long v)
{ return static_cast<std::int_least64_t>(v); }

constexpr std::uint_least64_t operator "" _uint_least64(unsigned long long v)
{ return static_cast<std::uint_least64_t>(v); }

constexpr std::intmax_t operator "" _intmax(unsigned long long v)
{ return static_cast<std::intmax_t>(v); }

constexpr std::uintmax_t operator "" _uintmax(unsigned long long v)
{ return static_cast<std::uintmax_t>(v); }

constexpr std::intptr_t operator "" _intptr(unsigned long long v)
{ return static_cast<std::intptr_t>(v); }

constexpr std::uintptr_t operator "" _uintptr(unsigned long long v)
{ return static_cast<std::uintptr_t>(v); }

Warning: The above code will silently give the wrong result if used on literals which don't fit into unsigned long long, as well as overflow if the literal value, doesn't fit into the requested type, e.g. 999_int8. A better implementation (GPL-3 licensed) would probably have to parse the literal character-by-character and static_assert on overflow, like this.

The downside of using user defined literals is that one needs to prefix the suffixes with an underscore _, because suffixes without the underscore are reserved for future standardization according to §17.6.4.3.4.