ruzstd/blocks/literals_section.rs
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
//! Utilities and representations for the first half of a block, the literals section.
//! It contains data that is then copied from by the sequences section.
use super::super::decoding::bit_reader::{BitReader, GetBitsError};
/// A compressed block consists of two sections, a literals section, and a sequences section.
///
/// This is the first of those two sections. A literal is just any arbitrary data, and it is copied by the sequences section
pub struct LiteralsSection {
/// - If this block is of type [LiteralsSectionType::Raw], then the data is `regenerated_bytes`
/// bytes long, and it contains the raw literals data to be used during the second section,
/// the sequences section.
/// - If this block is of type [LiteralsSectionType::RLE],
/// then the literal consists of a single byte repeated `regenerated_size` times.
/// - For types [LiteralsSectionType::Compressed] or [LiteralsSectionType::Treeless],
/// then this is the size of the decompressed data.
pub regenerated_size: u32,
/// - For types [LiteralsSectionType::Raw] and [LiteralsSectionType::RLE], this value is not present.
/// - For types [LiteralsSectionType::Compressed] and [LiteralsSectionType::Treeless], this value will
/// be set to the size of the compressed data.
pub compressed_size: Option<u32>,
/// This value will be either 1 stream or 4 streams if the literal is of type
/// [LiteralsSectionType::Compressed] or [LiteralsSectionType::Treeless], and it
/// is not used for RLE or uncompressed literals.
pub num_streams: Option<u8>,
/// The type of the literal section.
pub ls_type: LiteralsSectionType,
}
/// The way which a literal section is encoded.
pub enum LiteralsSectionType {
/// Literals are stored uncompressed.
Raw,
/// Literals consist of a single byte value repeated [LiteralsSection::regenerated_size] times.
RLE,
/// This is a standard Huffman-compressed block, starting with a Huffman tree description.
/// In this mode, there are at least *2* different literals represented in the Huffman tree
/// description.
Compressed,
/// This is a Huffman-compressed block,
/// using the Huffman tree from the previous [LiteralsSectionType::Compressed] block
/// in the sequence. If this mode is triggered without any previous Huffman-tables in the
/// frame (or dictionary), it should be treated as data corruption.
Treeless,
}
#[derive(Debug)]
#[non_exhaustive]
pub enum LiteralsSectionParseError {
IllegalLiteralSectionType { got: u8 },
GetBitsError(GetBitsError),
NotEnoughBytes { have: usize, need: u8 },
}
#[cfg(feature = "std")]
impl std::error::Error for LiteralsSectionParseError {
fn source(&self) -> Option<&(dyn std::error::Error + 'static)> {
match self {
LiteralsSectionParseError::GetBitsError(source) => Some(source),
_ => None,
}
}
}
impl core::fmt::Display for LiteralsSectionParseError {
fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
match self {
LiteralsSectionParseError::IllegalLiteralSectionType { got } => {
write!(
f,
"Illegal literalssectiontype. Is: {}, must be in: 0, 1, 2, 3",
got
)
}
LiteralsSectionParseError::GetBitsError(e) => write!(f, "{:?}", e),
LiteralsSectionParseError::NotEnoughBytes { have, need } => {
write!(
f,
"Not enough byte to parse the literals section header. Have: {}, Need: {}",
have, need,
)
}
}
}
}
impl From<GetBitsError> for LiteralsSectionParseError {
fn from(val: GetBitsError) -> Self {
Self::GetBitsError(val)
}
}
impl core::fmt::Display for LiteralsSectionType {
fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> Result<(), core::fmt::Error> {
match self {
LiteralsSectionType::Compressed => write!(f, "Compressed"),
LiteralsSectionType::Raw => write!(f, "Raw"),
LiteralsSectionType::RLE => write!(f, "RLE"),
LiteralsSectionType::Treeless => write!(f, "Treeless"),
}
}
}
impl Default for LiteralsSection {
fn default() -> Self {
Self::new()
}
}
impl LiteralsSection {
/// Create a new [LiteralsSection].
pub fn new() -> LiteralsSection {
LiteralsSection {
regenerated_size: 0,
compressed_size: None,
num_streams: None,
ls_type: LiteralsSectionType::Raw,
}
}
/// Given the first byte of a header, determine the size of the whole header, from 1 to 5 bytes.
pub fn header_bytes_needed(&self, first_byte: u8) -> Result<u8, LiteralsSectionParseError> {
let ls_type: LiteralsSectionType = Self::section_type(first_byte)?;
let size_format = (first_byte >> 2) & 0x3;
match ls_type {
LiteralsSectionType::RLE | LiteralsSectionType::Raw => {
match size_format {
0 | 2 => {
// size_format actually only uses one bit
// regenerated_size uses 5 bits
Ok(1)
}
1 => {
// size_format uses 2 bit
// regenerated_size uses 12 bits
Ok(2)
}
3 => {
// size_format uses 2 bit
// regenerated_size uses 20 bits
Ok(3)
}
_ => panic!(
"This is a bug in the program. There should only be values between 0..3"
),
}
}
LiteralsSectionType::Compressed | LiteralsSectionType::Treeless => {
match size_format {
0 | 1 => {
// Only differ in num_streams
// both regenerated and compressed sizes use 10 bit
Ok(3)
}
2 => {
// both regenerated and compressed sizes use 14 bit
Ok(4)
}
3 => {
// both regenerated and compressed sizes use 18 bit
Ok(5)
}
_ => panic!(
"This is a bug in the program. There should only be values between 0..3"
),
}
}
}
}
/// Parse the header into `self`, and returns the number of bytes read.
pub fn parse_from_header(&mut self, raw: &[u8]) -> Result<u8, LiteralsSectionParseError> {
let mut br: BitReader<'_> = BitReader::new(raw);
let block_type = br.get_bits(2)? as u8;
self.ls_type = Self::section_type(block_type)?;
let size_format = br.get_bits(2)? as u8;
let byte_needed = self.header_bytes_needed(raw[0])?;
if raw.len() < byte_needed as usize {
return Err(LiteralsSectionParseError::NotEnoughBytes {
have: raw.len(),
need: byte_needed,
});
}
match self.ls_type {
LiteralsSectionType::RLE | LiteralsSectionType::Raw => {
self.compressed_size = None;
match size_format {
0 | 2 => {
// size_format actually only uses one bit
// regenerated_size uses 5 bits
self.regenerated_size = u32::from(raw[0]) >> 3;
Ok(1)
}
1 => {
// size_format uses 2 bit
// regenerated_size uses 12 bits
self.regenerated_size = (u32::from(raw[0]) >> 4) + (u32::from(raw[1]) << 4);
Ok(2)
}
3 => {
// size_format uses 2 bit
// regenerated_size uses 20 bits
self.regenerated_size = (u32::from(raw[0]) >> 4)
+ (u32::from(raw[1]) << 4)
+ (u32::from(raw[2]) << 12);
Ok(3)
}
_ => panic!(
"This is a bug in the program. There should only be values between 0..3"
),
}
}
LiteralsSectionType::Compressed | LiteralsSectionType::Treeless => {
match size_format {
0 => {
self.num_streams = Some(1);
}
1..=3 => {
self.num_streams = Some(4);
}
_ => panic!(
"This is a bug in the program. There should only be values between 0..3"
),
};
match size_format {
0 | 1 => {
// Differ in num_streams see above
// both regenerated and compressed sizes use 10 bit
// 4 from the first, six from the second byte
self.regenerated_size =
(u32::from(raw[0]) >> 4) + ((u32::from(raw[1]) & 0x3f) << 4);
// 2 from the second, full last byte
self.compressed_size =
Some(u32::from(raw[1] >> 6) + (u32::from(raw[2]) << 2));
Ok(3)
}
2 => {
// both regenerated and compressed sizes use 14 bit
// 4 from first, full second, 2 from the third byte
self.regenerated_size = (u32::from(raw[0]) >> 4)
+ (u32::from(raw[1]) << 4)
+ ((u32::from(raw[2]) & 0x3) << 12);
// 6 from the third, full last byte
self.compressed_size =
Some((u32::from(raw[2]) >> 2) + (u32::from(raw[3]) << 6));
Ok(4)
}
3 => {
// both regenerated and compressed sizes use 18 bit
// 4 from first, full second, six from third byte
self.regenerated_size = (u32::from(raw[0]) >> 4)
+ (u32::from(raw[1]) << 4)
+ ((u32::from(raw[2]) & 0x3F) << 12);
// 2 from third, full fourth, full fifth byte
self.compressed_size = Some(
(u32::from(raw[2]) >> 6)
+ (u32::from(raw[3]) << 2)
+ (u32::from(raw[4]) << 10),
);
Ok(5)
}
_ => panic!(
"This is a bug in the program. There should only be values between 0..3"
),
}
}
}
}
/// Given the first two bits of a header, determine the type of a header.
fn section_type(raw: u8) -> Result<LiteralsSectionType, LiteralsSectionParseError> {
let t = raw & 0x3;
match t {
0 => Ok(LiteralsSectionType::Raw),
1 => Ok(LiteralsSectionType::RLE),
2 => Ok(LiteralsSectionType::Compressed),
3 => Ok(LiteralsSectionType::Treeless),
other => Err(LiteralsSectionParseError::IllegalLiteralSectionType { got: other }),
}
}
}