bits and bytes

definitions

the signed values use the high-most bit for a sign, so numbers are limited to 127 to -127 for signed and to 255 for unsigned

signed integers are not always reproducible straight from 010101 by two's conversion, but may also be stored in (as) twos complement, ones complement, sign-magnitude. This can lead to portability problems as on different systems the unsigned integers are represented differently.

it is usually best therefore to use the unsigned type to perform integer conversions

negative numbers representation

sign-magnitude

easiest to understand, -16 and 16 are same in binary, but the first bit is a "1" when it's -16

ones complement

simply reverses the bits to represent negative numbers (can tell diff between +16 and -16 by looking at the first bit)

twos complement

the most widely used

is not a simple oring of values: reverses all the bits and then adds one

if the programmer performs bitwise operations on numbers, results may not be as expected when signed types are used

representations

how do these look in declarations?

unsigned msb

in this example, represents the most significant byte in a network address

get this with msb = * (unsigned char *) &adr_inet.sin_addr.s_addr

unsigned char

operators

you know of course the bitwise operators are different from the logical operators...

bit arrays are bit arrays, not of predefined types, on which bit operations will be performed

memset operates on bit arrays, zeroing out all the bits