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
use crate::codec::decoder::Decoder;
use crate::codec::encoder::Encoder;
use bytes::{Buf, BufMut, Bytes, BytesMut};
use std::{cmp, fmt, io, str};
const DEFAULT_SEEK_DELIMITERS: &[u8] = b",;\n\r";
const DEFAULT_SEQUENCE_WRITER: &[u8] = b",";
/// A simple [`Decoder`] and [`Encoder`] implementation that splits up data into chunks based on any character in the given delimiter string.
///
/// [`Decoder`]: crate::codec::Decoder
/// [`Encoder`]: crate::codec::Encoder
///
/// # Example
/// Decode string of bytes containing various different delimiters.
///
/// [`BytesMut`]: bytes::BytesMut
/// [`Error`]: std::io::Error
///
/// ```
/// use tokio_util::codec::{AnyDelimiterCodec, Decoder};
/// use bytes::{BufMut, BytesMut};
///
/// #
/// # #[tokio::main(flavor = "current_thread")]
/// # async fn main() -> Result<(), std::io::Error> {
/// let mut codec = AnyDelimiterCodec::new(b",;\r\n".to_vec(),b";".to_vec());
/// let buf = &mut BytesMut::new();
/// buf.reserve(200);
/// buf.put_slice(b"chunk 1,chunk 2;chunk 3\n\r");
/// assert_eq!("chunk 1", codec.decode(buf).unwrap().unwrap());
/// assert_eq!("chunk 2", codec.decode(buf).unwrap().unwrap());
/// assert_eq!("chunk 3", codec.decode(buf).unwrap().unwrap());
/// assert_eq!("", codec.decode(buf).unwrap().unwrap());
/// assert_eq!(None, codec.decode(buf).unwrap());
/// # Ok(())
/// # }
/// ```
///
#[derive(Clone, Debug, Eq, PartialEq, Ord, PartialOrd, Hash)]
pub struct AnyDelimiterCodec {
// Stored index of the next index to examine for the delimiter character.
// This is used to optimize searching.
// For example, if `decode` was called with `abc` and the delimiter is '{}', it would hold `3`,
// because that is the next index to examine.
// The next time `decode` is called with `abcde}`, the method will
// only look at `de}` before returning.
next_index: usize,
/// The maximum length for a given chunk. If `usize::MAX`, chunks will be
/// read until a delimiter character is reached.
max_length: usize,
/// Are we currently discarding the remainder of a chunk which was over
/// the length limit?
is_discarding: bool,
/// The bytes that are using for search during decode
seek_delimiters: Vec<u8>,
/// The bytes that are using for encoding
sequence_writer: Vec<u8>,
}
impl AnyDelimiterCodec {
/// Returns a `AnyDelimiterCodec` for splitting up data into chunks.
///
/// # Note
///
/// The returned `AnyDelimiterCodec` will not have an upper bound on the length
/// of a buffered chunk. See the documentation for [`new_with_max_length`]
/// for information on why this could be a potential security risk.
///
/// [`new_with_max_length`]: crate::codec::AnyDelimiterCodec::new_with_max_length()
pub fn new(seek_delimiters: Vec<u8>, sequence_writer: Vec<u8>) -> AnyDelimiterCodec {
AnyDelimiterCodec {
next_index: 0,
max_length: usize::MAX,
is_discarding: false,
seek_delimiters,
sequence_writer,
}
}
/// Returns a `AnyDelimiterCodec` with a maximum chunk length limit.
///
/// If this is set, calls to `AnyDelimiterCodec::decode` will return a
/// [`AnyDelimiterCodecError`] when a chunk exceeds the length limit. Subsequent calls
/// will discard up to `limit` bytes from that chunk until a delimiter
/// character is reached, returning `None` until the delimiter over the limit
/// has been fully discarded. After that point, calls to `decode` will
/// function as normal.
///
/// # Note
///
/// Setting a length limit is highly recommended for any `AnyDelimiterCodec` which
/// will be exposed to untrusted input. Otherwise, the size of the buffer
/// that holds the chunk currently being read is unbounded. An attacker could
/// exploit this unbounded buffer by sending an unbounded amount of input
/// without any delimiter characters, causing unbounded memory consumption.
///
/// [`AnyDelimiterCodecError`]: crate::codec::AnyDelimiterCodecError
pub fn new_with_max_length(
seek_delimiters: Vec<u8>,
sequence_writer: Vec<u8>,
max_length: usize,
) -> Self {
AnyDelimiterCodec {
max_length,
..AnyDelimiterCodec::new(seek_delimiters, sequence_writer)
}
}
/// Returns the maximum chunk length when decoding.
///
/// ```
/// use std::usize;
/// use tokio_util::codec::AnyDelimiterCodec;
///
/// let codec = AnyDelimiterCodec::new(b",;\n".to_vec(), b";".to_vec());
/// assert_eq!(codec.max_length(), usize::MAX);
/// ```
/// ```
/// use tokio_util::codec::AnyDelimiterCodec;
///
/// let codec = AnyDelimiterCodec::new_with_max_length(b",;\n".to_vec(), b";".to_vec(), 256);
/// assert_eq!(codec.max_length(), 256);
/// ```
pub fn max_length(&self) -> usize {
self.max_length
}
}
impl Decoder for AnyDelimiterCodec {
type Item = Bytes;
type Error = AnyDelimiterCodecError;
fn decode(&mut self, buf: &mut BytesMut) -> Result<Option<Bytes>, AnyDelimiterCodecError> {
loop {
// Determine how far into the buffer we'll search for a delimiter. If
// there's no max_length set, we'll read to the end of the buffer.
let read_to = cmp::min(self.max_length.saturating_add(1), buf.len());
let new_chunk_offset = buf[self.next_index..read_to].iter().position(|b| {
self.seek_delimiters
.iter()
.any(|delimiter| *b == *delimiter)
});
match (self.is_discarding, new_chunk_offset) {
(true, Some(offset)) => {
// If we found a new chunk, discard up to that offset and
// then stop discarding. On the next iteration, we'll try
// to read a chunk normally.
buf.advance(offset + self.next_index + 1);
self.is_discarding = false;
self.next_index = 0;
}
(true, None) => {
// Otherwise, we didn't find a new chunk, so we'll discard
// everything we read. On the next iteration, we'll continue
// discarding up to max_len bytes unless we find a new chunk.
buf.advance(read_to);
self.next_index = 0;
if buf.is_empty() {
return Ok(None);
}
}
(false, Some(offset)) => {
// Found a chunk!
let new_chunk_index = offset + self.next_index;
self.next_index = 0;
let mut chunk = buf.split_to(new_chunk_index + 1);
chunk.truncate(chunk.len() - 1);
let chunk = chunk.freeze();
return Ok(Some(chunk));
}
(false, None) if buf.len() > self.max_length => {
// Reached the maximum length without finding a
// new chunk, return an error and start discarding on the
// next call.
self.is_discarding = true;
return Err(AnyDelimiterCodecError::MaxChunkLengthExceeded);
}
(false, None) => {
// We didn't find a chunk or reach the length limit, so the next
// call will resume searching at the current offset.
self.next_index = read_to;
return Ok(None);
}
}
}
}
fn decode_eof(&mut self, buf: &mut BytesMut) -> Result<Option<Bytes>, AnyDelimiterCodecError> {
Ok(match self.decode(buf)? {
Some(frame) => Some(frame),
None => {
// return remaining data, if any
if buf.is_empty() {
None
} else {
let chunk = buf.split_to(buf.len());
self.next_index = 0;
Some(chunk.freeze())
}
}
})
}
}
impl<T> Encoder<T> for AnyDelimiterCodec
where
T: AsRef<str>,
{
type Error = AnyDelimiterCodecError;
fn encode(&mut self, chunk: T, buf: &mut BytesMut) -> Result<(), AnyDelimiterCodecError> {
let chunk = chunk.as_ref();
buf.reserve(chunk.len() + 1);
buf.put(chunk.as_bytes());
buf.put(self.sequence_writer.as_ref());
Ok(())
}
}
impl Default for AnyDelimiterCodec {
fn default() -> Self {
Self::new(
DEFAULT_SEEK_DELIMITERS.to_vec(),
DEFAULT_SEQUENCE_WRITER.to_vec(),
)
}
}
/// An error occurred while encoding or decoding a chunk.
#[derive(Debug)]
pub enum AnyDelimiterCodecError {
/// The maximum chunk length was exceeded.
MaxChunkLengthExceeded,
/// An IO error occurred.
Io(io::Error),
}
impl fmt::Display for AnyDelimiterCodecError {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
match self {
AnyDelimiterCodecError::MaxChunkLengthExceeded => {
write!(f, "max chunk length exceeded")
}
AnyDelimiterCodecError::Io(e) => write!(f, "{e}"),
}
}
}
impl From<io::Error> for AnyDelimiterCodecError {
fn from(e: io::Error) -> AnyDelimiterCodecError {
AnyDelimiterCodecError::Io(e)
}
}
impl std::error::Error for AnyDelimiterCodecError {}