Module varint 

Available on crate feature alloc only.

Simple and efficient encoding of variably-sized integers

This encoding is little-endian, appending the seven least-significant bits to a byte vector, until the remaining bits are zero. The eight bit is reserved to mark whether or not the encoding continues.

VARINT_VALUE_MASK = (1 << 7) - 1
fn encode_varint(value) {
  if value == 0 {
    return [0]
  }

  result = []
  while value != 0 {
    next_byte = value & VARINT_VALUE_MASK
    value >>= 7
    next_byte |= (value != 0) << 7
    result.push(next_byte)
  }
  return result
}

fn decode_varint(bytestream) {
  result = 0
  i = 0
  while true {
    next_byte = bytestream.next_byte()
    result |= (next_byte & VARINT_VALUE_MASK) << i
    if (next_byte >> 7) == 0 {
      # Check this was canonical, without unnecessary extra bytes
      assert (i == 0) || (next_byte != 0)
      return result
    }
    i += 7
  }
}

Implementations MUST ensure that (next_byte & mask) << i DOES NOT shift any set bits past the boundary of the container. Implementations SHOULD terminate early if i exceeds the bit-length of the container (as either a bit will overflow the container, or the final byte will be zero and therefore the number is non-canonically encoded, either case being invalid).

Constants

VARINT_VALUE_MASK 🔒

Functions

decode_varint 🔒

This function runs in time variable to the input.

encode_varint 🔒

This function runs in time variable to the input.