A nonnegative C value
`unsigned x = 240u; unsigned y = x + 5u;` declares `x` and `y` with type `unsigned` (short for `unsigned int`). The `+` operator combines the values of its operands; evaluating `x` supplies the value 240. section 4.1 section 5.7.1
Unsigned integer types represent values from 0 through a type-specific maximum. For `unsigned`, that bound is `UINT_MAX` from `<limits.h>`; `unsigned long` and `unsigned long long` have their own bounds. Their precision is implementation-defined, so do not assume a particular maximum. The semantic type `size_t` is also unsigned and has a bound named `SIZE_MAX`. section 5.2 section 5.7.1
An unsigned type with precision `p` represents exactly 0 through `2^p − 1`. Its value is the sum of its set bits, each weighted by a power of two. For example, in a 16-bit unsigned type, 240 has bits 4, 5, 6, and 7 set. Unsigned arithmetic is closed over its range: mathematically, its results are described modulo `2^p`. section 5.7.1
TAKEAWAY
- Check the relevant `_MAX` macro rather than assuming an unsigned type’s range. section 5.2 section 5.7.1
- A narrow unsigned integer may be promoted before arithmetic; on the realistic platforms described here, that promotion produces `signed int`. section 5.2
typedef union unsignedInspect unsignedInspect;
union unsignedInspect {
unsigned val;
unsigned char bytes[sizeof(unsigned)];
};
unsignedInspect twofold = { .val = 0xAABBCCDD, };
This union overlays an `unsigned` value with an array of `unsigned char` so its individual bytes can be inspected. Byte order is implementation-defined; the order observed on one machine is not a portable rule. section 12.2
