Revision control

Copy as Markdown

Other Tools

// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
use crate::{
huffman_decode_helper::{huffman_decoder_root, HuffmanDecoderNode},
huffman_table::HUFFMAN_TABLE,
Error, Res,
};
struct BitReader<'a> {
input: &'a [u8],
offset: usize,
current_bit: u8,
}
impl<'a> BitReader<'a> {
pub const fn new(input: &'a [u8]) -> Self {
BitReader {
input,
offset: 0,
current_bit: 8,
}
}
pub fn read_bit(&mut self) -> Res<u8> {
if self.input.len() == self.offset {
return Err(Error::NeedMoreData);
}
if self.current_bit == 0 {
self.offset += 1;
if self.offset == self.input.len() {
return Err(Error::NeedMoreData);
}
self.current_bit = 8;
}
self.current_bit -= 1;
Ok((self.input[self.offset] >> self.current_bit) & 0x01)
}
pub fn verify_ending(&mut self, i: u8) -> Res<()> {
if (i + self.current_bit) > 7 {
return Err(Error::HuffmanDecompressionFailed);
}
if self.input.is_empty() {
Ok(())
} else if self.offset != self.input.len() {
Err(Error::HuffmanDecompressionFailed)
} else if self.input[self.input.len() - 1] & ((0x1 << (i + self.current_bit)) - 1)
== ((0x1 << (i + self.current_bit)) - 1)
{
self.current_bit = 0;
Ok(())
} else {
Err(Error::HuffmanDecompressionFailed)
}
}
pub const fn has_more_data(&self) -> bool {
!self.input.is_empty() && (self.offset != self.input.len() || (self.current_bit != 0))
}
}
/// Decodes huffman encoded input.
///
/// # Errors
///
/// This function may return `HuffmanDecompressionFailed` if `input` is not a correct
/// huffman-encoded array of bits.
///
/// # Panics
///
/// Never, but rust can't know that.
pub fn decode_huffman(input: &[u8]) -> Res<Vec<u8>> {
let mut reader = BitReader::new(input);
let mut output = Vec::new();
while reader.has_more_data() {
if let Some(c) = decode_character(&mut reader)? {
if c == 256 {
return Err(Error::HuffmanDecompressionFailed);
}
output.push(u8::try_from(c).unwrap());
}
}
Ok(output)
}
fn decode_character(reader: &mut BitReader) -> Res<Option<u16>> {
let mut node: &HuffmanDecoderNode = huffman_decoder_root();
let mut i = 0;
while node.value.is_none() {
match reader.read_bit() {
Err(_) => {
reader.verify_ending(i)?;
return Ok(None);
}
Ok(b) => {
i += 1;
if let Some(next) = &node.next[usize::from(b)] {
node = next;
} else {
reader.verify_ending(i)?;
return Ok(None);
}
}
}
}
debug_assert!(node.value.is_some());
Ok(node.value)
}
/// # Panics
///
/// Never, but rust doesn't know that.
#[must_use]
pub fn encode_huffman(input: &[u8]) -> Vec<u8> {
let mut output: Vec<u8> = Vec::new();
let mut left: u8 = 8;
let mut saved: u8 = 0;
for c in input {
let mut e = HUFFMAN_TABLE[*c as usize];
// Fill the previous byte
if e.len < left {
let b = u8::try_from(e.val & 0xFF).unwrap();
saved |= b << (left - e.len);
left -= e.len;
e.len = 0;
} else {
let v: u8 = u8::try_from(e.val >> (e.len - left)).unwrap();
saved |= v;
output.push(saved);
e.len -= left;
left = 8;
saved = 0;
}
// Write full bytes
while e.len >= 8 {
let v: u8 = u8::try_from((e.val >> (e.len - 8)) & 0xFF).unwrap();
output.push(v);
e.len -= 8;
}
// Write the rest into saved.
if e.len > 0 {
saved = u8::try_from(e.val & ((1 << e.len) - 1)).unwrap() << (8 - e.len);
left = 8 - e.len;
}
}
if left < 8 {
let v: u8 = (1 << left) - 1;
saved |= v;
output.push(saved);
}
output
}
#[cfg(test)]
mod tests {
use super::{decode_huffman, encode_huffman, Error};
struct TestElement {
pub val: &'static [u8],
pub res: &'static [u8],
}
const TEST_CASES: &[TestElement] = &[
TestElement {
val: b"www.example.com",
res: &[
0xf1, 0xe3, 0xc2, 0xe5, 0xf2, 0x3a, 0x6b, 0xa0, 0xab, 0x90, 0xf4, 0xff,
],
},
TestElement {
val: b"no-cache",
res: &[0xa8, 0xeb, 0x10, 0x64, 0x9c, 0xbf],
},
TestElement {
val: b"custom-key",
res: &[0x25, 0xa8, 0x49, 0xe9, 0x5b, 0xa9, 0x7d, 0x7f],
},
TestElement {
val: b"custom-value",
res: &[0x25, 0xa8, 0x49, 0xe9, 0x5b, 0xb8, 0xe8, 0xb4, 0xbf],
},
TestElement {
val: b"private",
res: &[0xae, 0xc3, 0x77, 0x1a, 0x4b],
},
TestElement {
val: b"Mon, 21 Oct 2013 20:13:21 GMT",
res: &[
0xd0, 0x7a, 0xbe, 0x94, 0x10, 0x54, 0xd4, 0x44, 0xa8, 0x20, 0x05, 0x95, 0x04, 0x0b,
0x81, 0x66, 0xe0, 0x82, 0xa6, 0x2d, 0x1b, 0xff,
],
},
TestElement {
res: &[
0x9d, 0x29, 0xad, 0x17, 0x18, 0x63, 0xc7, 0x8f, 0x0b, 0x97, 0xc8, 0xe9, 0xae, 0x82,
0xae, 0x43, 0xd3,
],
},
TestElement {
val: b"Mon, 21 Oct 2013 20:13:22 GMT",
res: &[
0xd0, 0x7a, 0xbe, 0x94, 0x10, 0x54, 0xd4, 0x44, 0xa8, 0x20, 0x05, 0x95, 0x04, 0x0b,
0x81, 0x66, 0xe0, 0x84, 0xa6, 0x2d, 0x1b, 0xff,
],
},
TestElement {
val: b"gzip",
res: &[0x9b, 0xd9, 0xab],
},
TestElement {
val: b"foo=ASDJKHQKBZXOQWEOPIUAXQWEOIU; max-age=3600; version=1",
res: &[
0x94, 0xe7, 0x82, 0x1d, 0xd7, 0xf2, 0xe6, 0xc7, 0xb3, 0x35, 0xdf, 0xdf, 0xcd, 0x5b,
0x39, 0x60, 0xd5, 0xaf, 0x27, 0x08, 0x7f, 0x36, 0x72, 0xc1, 0xab, 0x27, 0x0f, 0xb5,
0x29, 0x1f, 0x95, 0x87, 0x31, 0x60, 0x65, 0xc0, 0x03, 0xed, 0x4e, 0xe5, 0xb1, 0x06,
0x3d, 0x50, 0x07,
],
},
TestElement {
val: b"<?\\ >",
res: &[0xff, 0xf9, 0xfe, 0x7f, 0xff, 0x05, 0x3f, 0xef],
},
];
const WRONG_END: &[u8] = &[0xa8, 0xeb, 0x10, 0x64, 0x9c, 0xaf];
#[test]
fn encoder() {
for e in TEST_CASES {
let out = encode_huffman(e.val);
assert_eq!(out[..], *e.res);
}
}
#[test]
fn decoder() {
for e in TEST_CASES {
let res = decode_huffman(e.res);
assert!(res.is_ok());
assert_eq!(res.unwrap()[..], *e.val);
}
}
#[test]
fn decoder_error_wrong_ending() {
assert_eq!(
decode_huffman(WRONG_END),
Err(Error::HuffmanDecompressionFailed)
);
}
}