1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
use std::io;
use std::io::Result;
use std::mem::size_of;
/// Most-significant byte, == 0x80
pub const MSB: u8 = 0b1000_0000;
/// All bits except for the most significant. Can be used as bitmask to drop the most-signficant
/// bit using `&` (binary-and).
const DROP_MSB: u8 = 0b0111_1111;
/// How many bytes an integer uses when being encoded as a VarInt.
#[inline]
fn required_encoded_space_unsigned(mut v: u64) -> usize {
if v == 0 {
return 1;
}
let mut logcounter = 0;
while v > 0 {
logcounter += 1;
v >>= 7;
}
logcounter
}
/// How many bytes an integer uses when being encoded as a VarInt.
#[inline]
fn required_encoded_space_signed(v: i64) -> usize {
required_encoded_space_unsigned(zigzag_encode(v))
}
/// Varint (variable length integer) encoding, as described in
/// https://developers.google.com/protocol-buffers/docs/encoding.
///
/// Uses zigzag encoding (also described there) for signed integer representation.
pub trait VarInt: Sized + Copy {
/// Returns the number of bytes this number needs in its encoded form. Note: This varies
/// depending on the actual number you want to encode.
fn required_space(self) -> usize;
/// Decode a value from the slice. Returns the value and the number of bytes read from the
/// slice (can be used to read several consecutive values from a big slice)
/// return None if all bytes has MSB set.
fn decode_var(src: &[u8]) -> Option<(Self, usize)>;
/// Encode a value into the slice. The slice must be at least `required_space()` bytes long.
/// The number of bytes taken by the encoded integer is returned.
fn encode_var(self, src: &mut [u8]) -> usize;
/// Helper: Encode a value and return the encoded form as Vec. The Vec must be at least
/// `required_space()` bytes long.
fn encode_var_vec(self) -> Vec<u8> {
let mut v = Vec::new();
v.resize(self.required_space(), 0);
self.encode_var(&mut v);
v
}
}
#[inline]
fn zigzag_encode(from: i64) -> u64 {
((from << 1) ^ (from >> 63)) as u64
}
// see: http://stackoverflow.com/a/2211086/56332
// casting required because operations like unary negation
// cannot be performed on unsigned integers
#[inline]
fn zigzag_decode(from: u64) -> i64 {
((from >> 1) ^ (-((from & 1) as i64)) as u64) as i64
}
pub(crate) trait VarIntMaxSize {
fn varint_max_size() -> usize;
}
impl<VI: VarInt> VarIntMaxSize for VI {
fn varint_max_size() -> usize {
(size_of::<VI>() * 8 + 7) / 7
}
}
macro_rules! impl_varint {
($t:ty, unsigned) => {
impl VarInt for $t {
fn required_space(self) -> usize {
required_encoded_space_unsigned(self as u64)
}
fn decode_var(src: &[u8]) -> Option<(Self, usize)> {
let (n, s) = u64::decode_var(src)?;
Some((n as Self, s))
}
fn encode_var(self, dst: &mut [u8]) -> usize {
(self as u64).encode_var(dst)
}
}
};
($t:ty, signed) => {
impl VarInt for $t {
fn required_space(self) -> usize {
required_encoded_space_signed(self as i64)
}
fn decode_var(src: &[u8]) -> Option<(Self, usize)> {
let (n, s) = i64::decode_var(src)?;
Some((n as Self, s))
}
fn encode_var(self, dst: &mut [u8]) -> usize {
(self as i64).encode_var(dst)
}
}
};
}
impl_varint!(u32, unsigned);
impl_varint!(u16, unsigned);
impl_varint!(u8, unsigned);
impl_varint!(i32, signed);
impl_varint!(i16, signed);
impl_varint!(i8, signed);
// Below are the "base implementations" doing the actual encodings; all other integer types are
// first cast to these biggest types before being encoded.
impl VarInt for u64 {
fn required_space(self) -> usize {
required_encoded_space_unsigned(self)
}
#[inline]
fn decode_var(src: &[u8]) -> Option<(Self, usize)> {
let mut result: u64 = 0;
let mut shift = 0;
let mut success = false;
for b in src.iter() {
let msb_dropped = b & DROP_MSB;
result |= (msb_dropped as u64) << shift;
shift += 7;
if b & MSB == 0 || shift > (9 * 7) {
success = b & MSB == 0;
break;
}
}
if success {
Some((result, shift / 7))
} else {
None
}
}
#[inline]
fn encode_var(self, dst: &mut [u8]) -> usize {
assert!(dst.len() >= self.required_space());
let mut n = self;
let mut i = 0;
while n >= 0x80 {
dst[i] = MSB | (n as u8);
i += 1;
n >>= 7;
}
dst[i] = n as u8;
i + 1
}
}
impl VarInt for i64 {
fn required_space(self) -> usize {
required_encoded_space_signed(self)
}
#[inline]
fn decode_var(src: &[u8]) -> Option<(Self, usize)> {
if let Some((result, size)) = u64::decode_var(src) {
Some((zigzag_decode(result) as Self, size))
} else {
None
}
}
#[inline]
fn encode_var(self, dst: &mut [u8]) -> usize {
assert!(dst.len() >= self.required_space());
let mut n: u64 = zigzag_encode(self as i64);
let mut i = 0;
while n >= 0x80 {
dst[i] = MSB | (n as u8);
i += 1;
n >>= 7;
}
dst[i] = n as u8;
i + 1
}
}
/// A trait for reading VarInts from any other `Reader`.
///
/// It's recommended to use a buffered reader, as many small reads will happen.
pub trait VarIntReader {
/// Returns either the decoded integer, or an error.
///
/// In general, this always reads a whole varint. If the encoded varint's value is bigger
/// than the valid value range of `VI`, then the value is truncated.
///
/// On EOF, an io::Error with io::ErrorKind::UnexpectedEof is returned.
fn read_varint<VI: VarInt>(&mut self) -> Result<VI>;
}
/// VarIntProcessor encapsulates the logic for decoding a VarInt byte-by-byte.
#[derive(Default)]
pub(crate) struct VarIntProcessor {
buf: [u8; 10],
maxsize: usize,
pub i: usize,
}
impl VarIntProcessor {
pub fn new<VI: VarIntMaxSize>() -> VarIntProcessor {
VarIntProcessor {
maxsize: VI::varint_max_size(),
..VarIntProcessor::default()
}
}
pub fn push(&mut self, b: u8) -> Result<()> {
if self.i >= self.maxsize {
return Err(io::Error::new(
io::ErrorKind::InvalidData,
"Unterminated varint",
));
}
self.buf[self.i] = b;
self.i += 1;
Ok(())
}
pub fn finished(&self) -> bool {
self.i > 0 && (self.buf[self.i - 1] & MSB == 0)
}
pub fn decode<VI: VarInt>(&self) -> Option<VI> {
Some(VI::decode_var(&self.buf[0..self.i])?.0)
}
}
impl<R: io::Read> VarIntReader for R {
fn read_varint<VI: VarInt>(&mut self) -> Result<VI> {
let mut buf = [0_u8; 1];
let mut p = VarIntProcessor::new::<VI>();
while !p.finished() {
let read = self.read(&mut buf)?;
// EOF
if read == 0 && p.i == 0 {
return Err(io::Error::new(io::ErrorKind::UnexpectedEof, "Reached EOF"));
}
if read == 0 {
break;
}
p.push(buf[0])?;
}
p.decode()
.ok_or_else(|| io::Error::new(io::ErrorKind::UnexpectedEof, "Reached EOF"))
}
}