A sign representation encodes signed values with a sign bit and the remaining representation bits. A zero sign bit denotes a positive value; a one sign bit denotes a negative value. Positive signed values use the same representation as corresponding unsigned values. section 5.7.5
C permits three sign representations:
- Sign and magnitude keeps the positive magnitude and uses the sign bit as a minus sign. section 5.7.5
- Ones’ complement complements every bit of the corresponding positive value. section 5.7.5
- Two’s complement performs the same bit arithmetic as unsigned values, while representations with a high-order one are interpreted as negative. section 5.7.5
Negating 240 in 16 bits
For unsigned arithmetic, negate A by computing ~A + 1. Starting with 240, or 0000000011110000, complementing gives 65295, or 1111111100001111; adding one gives 65296, or 1111111100010000. Two’s complement assigns that resulting bit pattern its negative signed interpretation. section 5.7.5
bool is_negative(unsigned a) {
unsigned const int_max = UINT_MAX/2;
return a > int_max;
}
bool is_signed_less(unsigned a, unsigned b) {
if (is_negative(b) && !is_negative(a)) return false;
else return a < b;
}
!@%STYLE%@!
{"css":"{\"css\": \"font-weight: bold;\"}","target":"[[{\"line\":0,\"ch\":17},{\"line\":0,\"ch\":25}],[{\"line\":1,\"ch\":2},{\"line\":1,\"ch\":10}],[{\"line\":4,\"ch\":20},{\"line\":4,\"ch\":28}],[{\"line\":1,\"ch\":2},{\"line\":1,\"ch\":16}],[{\"line\":1,\"ch\":27},{\"line\":1,\"ch\":35}],[{\"line\":2,\"ch\":2},{\"line\":2,\"ch\":8}],[{\"line\":5,\"ch\":41},{\"line\":5,\"ch\":47}],[{\"line\":6,\"ch\":7},{\"line\":6,\"ch\":13}],[{\"line\":0,\"ch\":17},{\"line\":0,\"ch\":25}],[{\"line\":1,\"ch\":2},{\"line\":1,\"ch\":10}],[{\"line\":4,\"ch\":20},{\"line\":4,\"ch\":28}],[{\"line\":0,\"ch\":17},{\"line\":0,\"ch\":25}],[{\"line\":1,\"ch\":2},{\"line\":1,\"ch\":10}],[{\"line\":4,\"ch\":20},{\"line\":4,\"ch\":28}],[{\"line\":5,\"ch\":2},{\"line\":5,\"ch\":4}],[{\"line\":2,\"ch\":2},{\"line\":2,\"ch\":8}],[{\"line\":5,\"ch\":41},{\"line\":5,\"ch\":47}],[{\"line\":6,\"ch\":7},{\"line\":6,\"ch\":13}],[{\"line\":5,\"ch\":48},{\"line\":5,\"ch\":53}],[{\"line\":6,\"ch\":2},{\"line\":6,\"ch\":13}]]"}
!@%STYLE%@!
This implementation compares unsigned bit patterns as signed values by treating values above UINT_MAX/2 as negative. section 5.7.5
Sign representation is part of a type’s binary representation: the standard leaves some representation details platform-dependent, but programs can observe the resulting model through values, constants, operators, and library interfaces. section 5.1.3
