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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
#![no_std] //! # Rut //! //! Rut is a small UTF-8 parsing library for applications that need to parse individual `char`s.\ //! It provides a byte-wise parsing mechanism, and functions for processing byte slices. //! //! It is completely `#[no_std]` and should produce very small binaries.<sup>[[*citation needed*](index.html)]</sup> //! //! # Conformance //! //! Rut aims to be fully conformant to the specifications and restrictions of the [Unicode standard].\ //! Due to the nature of byte-wise parsing, some [extra caution] might be required when using Rut. //! //! The [`parse_one`] and [`parse`] functions take care of this. //! //! # Testing //! //! A few tests validating the expected behavior are already in place, but it is not comprehensive by any means yet. //! More tests will be added. //! //! I have thrown a fuzzer at it for several minutes, and it passes this [stress test for UTF-8 decoders](https://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt). //! //! # Examples //! //! ``` //! use rut::Utf8Parser; //! //! // UTF-8 encoding of '€' //! let bytes = [0xE2, 0x82, 0xAC]; //! //! let mut p = Utf8Parser::new(); //! //! assert_eq!(p.parse_byte(bytes[0]), Ok(None)); //! assert_eq!(p.parse_byte(bytes[1]), Ok(None)); //! assert_eq!(p.parse_byte(bytes[2]), Ok(Some('€'))); //! ``` //! //! [Unicode standard]: http://www.unicode.org/standard/standard.html //! [extra caution]: struct.Utf8Parser.html#conformance //! [`parse_one`]: fn.parse_one.html //! [`parse`]: fn.parse.html /// A UTF-8 parser error. #[derive(Copy, Clone, Debug, Eq, PartialEq)] pub enum Utf8ParseError { /// A byte value of `C0`, `C1`, or `F5..=FF` was encountered. InvalidByte, /// A continuation was encountered outside of a sequence. UnexpectedContinuation, /// A sequence was terminated by a byte that isn't a valid continuation. BrokenSequence, /// An [overlong encoding](https://en.wikipedia.org/wiki/UTF-8#Overlong_encodings) was encountered. OverlongEncoding, /// An encoding which would result in an invalid unicode scalar value (U+D800-U+DFFF or > U+10FFF) was encountered. InvalidCodePoint, /// The end of input was reached before the sequence could be fully parsed. TruncatedSequence, } /// A byte-wise UTF-8 parser. #[derive(Copy, Clone, Debug, Default)] pub struct Utf8Parser { // Code point parsed up to this point. value: u32, // Number of continuation bytes in this sequence. length: u8, // Leading byte of this sequence. lead: u8, } /// Bitmask used for leading bytes. /// 110x xxxx for 2 byte, /// 1110 xxxx for 3 byte, and /// 1111 0xxx for 4 byte sequences. const LEAD_MASK: [u8; 3] = [0x1F, 0x0F, 0x07]; /// Bitmask used for continuation bytes. /// Always 10xx xxxx. const CONT_MASK: u8 = 0x3F; impl Utf8Parser { /// Creates a new parser. /// This is exactly equivalent to `Utf8Parser::default()`. #[inline] pub fn new() -> Utf8Parser { Utf8Parser::default() } /// Resets the parser to the base state. /// This is generally not necessary, as a reset happens whenever an error is encountered during parsing. #[inline] pub fn reset(&mut self) { // NOTE: `self.value` and `self.lead` are not reset here, // because this implicitly happens in the // "new character, leading byte" branch of `parse_byte`. self.length = 0; } /// Checks if the parser is currently inside a sequence. /// Since it has no concept of an 'end of input' condition, /// this can be used to check if a sequence is unterminated so far. #[inline] pub fn is_mid_parse(&self) -> bool { self.length != 0 } /// Parses a single UTF-8 byte. /// /// Returns `Ok(Some(char))` if a full character was parsed, /// `Ok(None)` if the byte was parsed but no full character yet, /// and `Err(Utf8ParseError)` otherwise. /// /// # Conformance /// /// Due to the nature of this parser, incorrect use may lead to behavior not conformant to the Unicode standard. /// Namely, Unicode forbids interpreting well-formed sequences as part of ill-formed ones. /// /// For example, the byte sequence `C2 41 42` would yield the result `<None> <Error> B`. /// Unicode expects this sequence to instead result in `<Error> A B`. /// /// Implementing this would prove very difficult, so instead, if `parse_byte` returns `Err(BrokenSequence)`, /// it should be called with the same byte value again, not with the next one. /// /// The following example uses the Unicode replacement character (�) on errors. /// /// # Example /// /// ``` /// use std::fmt::Write; /// use rut::{Utf8Parser, Utf8ParseError::*}; /// /// let mut s = String::new(); /// let mut p = Utf8Parser::new(); /// /// // An ill-formed UTF-8 sequence as described above. /// let bytes = [0xC2, 0x41, 0x42]; /// /// for &byte in &bytes { /// let mut result = p.parse_byte(byte); /// /// // If we get an invalid continuation byte, try again. /// // This makes sure that if `byte` is the start of a well-formed sequence, /// // it isn't "swallowed" by the parser. /// if let Err(BrokenSequence) = result { /// write!(&mut s, "�"); /// result = p.parse_byte(byte); /// } /// /// match result { /// Ok(Some(c)) => write!(&mut s, "{}", c).unwrap(), /// Err(_) => write!(&mut s, "�").unwrap(), /// _ => () /// } /// } /// /// assert_eq!(&s, "�AB") /// ``` pub fn parse_byte(&mut self, byte: u8) -> Result<Option<char>, Utf8ParseError> { use core::convert::TryFrom; use self::Utf8ParseError::*; if !is_valid_utf8_byte(byte) { self.reset(); return Err(InvalidByte); } // Start parsing a new sequence. if self.length == 0 { if byte < 0x80 { return Ok(Some(char::from(byte))); } else if is_utf8_continuation_byte(byte) { // We do not call self.reset() here because we are already in the base state. return Err(UnexpectedContinuation); } // We subtract 1 and treat it as the number of bytes following this one. self.lead = byte; self.length = utf8_length(byte) - 1; self.value = (byte & LEAD_MASK[(self.length - 1) as usize]) as u32; // Parsing is Ok, but we don't have a full char yet. return Ok(None); } else { // Continue parsing the current sequence // Implies byte is in 0x80..=0xBF if !is_utf8_continuation_byte(byte) { self.reset(); return Err(BrokenSequence); } // This branch is only hit on the first continuation byte, and checks for the sequence's validity. if self.lead != 0 { // E0 XX for XX < A0 and F0 XX for XX < 90 represent a non-shortest encoding. if self.lead == 0xE0 && byte < 0xA0 || self.lead == 0xF0 && byte < 0x90 { self.reset(); return Err(OverlongEncoding); } // ED XX for XX > 9F and F4 XX for XX > 8F represent illegal unicode scalar values if self.lead == 0xED && byte > 0x9F || self.lead == 0xF4 && byte > 0x8F { self.reset(); return Err(InvalidCodePoint); } self.lead = 0; } self.value = (self.value << 6) | (byte & CONT_MASK) as u32; self.length -= 1; // We're done if self.length == 0 { return Ok(Some( char::try_from(self.value).expect("Uncought Invalid Code Point"), )); } else { // Parsing is Ok, but we don't have a full char yet. return Ok(None); } } } } /// Parses one UTF-8 sequence from a byte slice, returning a `Result` and the unparsed remainder of the slice. /// /// # Panics /// /// Panics if `bytes` is empty. /// /// # Examples /// /// ``` /// // Valid UTF-8 encoding of '€' /// let bytes = [0xE2, 0x82, 0xAC]; /// /// let (result, rest) = rut::parse_one(&bytes); /// /// assert_eq!(result, Ok('€')); /// assert_eq!(rest, &[]); /// ``` /// /// ``` /// use rut::Utf8ParseError::*; /// /// // Ill-formed sequence followed by 2 valid characters /// let bytes = [0xC2, 0x41, 0x42]; /// /// let (result1, rest1) = rut::parse_one(&bytes); /// let (result2, rest2) = rut::parse_one(rest1); /// let (result3, rest3) = rut::parse_one(rest2); /// /// assert_eq!(result1, Err(BrokenSequence)); /// assert_eq!(result2, Ok('A')); /// assert_eq!(result3, Ok('B')); /// assert_eq!(rest1, &[0x41, 0x42]); /// assert_eq!(rest2, &[0x42]); /// assert_eq!(rest3, &[]); /// ``` pub fn parse_one<'a>(bytes: &'a [u8]) -> (Result<char, Utf8ParseError>, &'a [u8]) { use self::Utf8ParseError::*; assert!(!bytes.is_empty()); let mut idx = 0; let mut p = Utf8Parser::new(); loop { if let Some(&byte) = bytes.get(idx) { match p.parse_byte(byte) { // If we didn't get anything yet, continue the loop Ok(None) => (), // If we got a char, return it and start from the next byte Ok(Some(c)) => return (Ok(c), &bytes[idx+1..]), // If we got a broken sequence, we need to start parsing from the current byte again. Err(BrokenSequence) => return (Err(BrokenSequence), &bytes[idx..]), // On all other errors, continue to the next byte. Err(e) => return (Err(e), &bytes[idx+1..]) } } else { return (Err(TruncatedSequence), &[]) } idx += 1; } } /// An iterator for parsing UTF-8 sequences from a byte slice. /// /// See the [`parse`](fn.parse.html) function for more information. #[must_use = "iterators are lazy and do nothing unless consumed"] #[derive(Clone)] pub struct Parse<'a> { bytes: &'a [u8] } /// Creates an iterator for parsing UTF-8 sequences from a byte slice. /// /// # Examples /// /// ``` /// // Valid UTF-8 encoding of '€' /// let bytes = [0xE2, 0x82, 0xAC]; /// /// let mut it = rut::parse(&bytes); /// /// assert_eq!(it.next(), Some(Ok('€'))); /// assert_eq!(it.next(), None); /// ``` /// /// ``` /// use rut::Utf8ParseError::*; /// /// // Ill-formed sequence followed by 2 valid characters /// let bytes = [0xC2, 0x41, 0x42]; /// /// let mut it = rut::parse(&bytes); /// /// assert_eq!(it.next(), Some(Err(BrokenSequence))); /// assert_eq!(it.next(), Some(Ok('A'))); /// assert_eq!(it.next(), Some(Ok('B'))); /// assert_eq!(it.next(). None); /// ``` #[inline] pub fn parse<'a>(bytes: &'a [u8]) -> Parse<'a> { Parse { bytes } } impl ::core::iter::Iterator for Parse<'_> { type Item = Result<char, Utf8ParseError>; fn next(&mut self) -> Option<Self::Item> { if self.bytes.is_empty() { None } else { let (result, rest) = parse_one(self.bytes); self.bytes = rest; Some(result) } } } impl ::core::iter::FusedIterator for Parse<'_> {} // Helper functions #[inline] fn is_valid_utf8_byte(byte: u8) -> bool { match byte { 0xC0 | 0xC1 | 0xF5..=0xFF => false, _ => true, } } #[inline] fn is_utf8_continuation_byte(byte: u8) -> bool { // Continuation bytes look like 10xx xxxx, // so we look at the top 2 bits and see if they match. (byte >> 6) == 0b10 } #[inline] fn utf8_length(byte: u8) -> u8 { // This function assumes that `byte` is both a valid UTF-8 byte and *not* a continuation byte. debug_assert!(is_valid_utf8_byte(byte) && !is_utf8_continuation_byte(byte)); if byte < 0x80 { 1 } else { (!byte).leading_zeros() as u8 } }