| 1 | #![cfg_attr (not(feature = "std" ), no_std)] |
| 2 | #![deny (missing_docs)] |
| 3 | #![cfg_attr (test, deny(warnings))] |
| 4 | |
| 5 | //! # httparse |
| 6 | //! |
| 7 | //! A push library for parsing HTTP/1.x requests and responses. |
| 8 | //! |
| 9 | //! The focus is on speed and safety. Unsafe code is used to keep parsing fast, |
| 10 | //! but unsafety is contained in a submodule, with invariants enforced. The |
| 11 | //! parsing internals use an `Iterator` instead of direct indexing, while |
| 12 | //! skipping bounds checks. |
| 13 | //! |
| 14 | //! With Rust 1.27.0 or later, support for SIMD is enabled automatically. |
| 15 | //! If building an executable to be run on multiple platforms, and thus |
| 16 | //! not passing `target_feature` or `target_cpu` flags to the compiler, |
| 17 | //! runtime detection can still detect SSE4.2 or AVX2 support to provide |
| 18 | //! massive wins. |
| 19 | //! |
| 20 | //! If compiling for a specific target, remembering to include |
| 21 | //! `-C target_cpu=native` allows the detection to become compile time checks, |
| 22 | //! making it *even* faster. |
| 23 | |
| 24 | use core::{fmt, result, str}; |
| 25 | use core::mem::{self, MaybeUninit}; |
| 26 | |
| 27 | use crate::iter::Bytes; |
| 28 | |
| 29 | mod iter; |
| 30 | #[macro_use ] mod macros; |
| 31 | mod simd; |
| 32 | |
| 33 | /// Determines if byte is a token char. |
| 34 | /// |
| 35 | /// > ```notrust |
| 36 | /// > token = 1*tchar |
| 37 | /// > |
| 38 | /// > tchar = "!" / "#" / "$" / "%" / "&" / "'" / "*" |
| 39 | /// > / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~" |
| 40 | /// > / DIGIT / ALPHA |
| 41 | /// > ; any VCHAR, except delimiters |
| 42 | /// > ``` |
| 43 | #[inline ] |
| 44 | fn is_token(b: u8) -> bool { |
| 45 | b > 0x1F && b < 0x7F |
| 46 | } |
| 47 | |
| 48 | // ASCII codes to accept URI string. |
| 49 | // i.e. A-Z a-z 0-9 !#$%&'*+-._();:@=,/?[]~^ |
| 50 | // TODO: Make a stricter checking for URI string? |
| 51 | static URI_MAP: [bool; 256] = byte_map![ |
| 52 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 53 | // \0 \n |
| 54 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 55 | // commands |
| 56 | 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 57 | // \w ! " # $ % & ' ( ) * + , - . / |
| 58 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, |
| 59 | // 0 1 2 3 4 5 6 7 8 9 : ; < = > ? |
| 60 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 61 | // @ A B C D E F G H I J K L M N O |
| 62 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 63 | // P Q R S T U V W X Y Z [ \ ] ^ _ |
| 64 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 65 | // ` a b c d e f g h i j k l m n o |
| 66 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, |
| 67 | // p q r s t u v w x y z { | } ~ del |
| 68 | // ====== Extended ASCII (aka. obs-text) ====== |
| 69 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 70 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 71 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 72 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 73 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 74 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 75 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 76 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 77 | ]; |
| 78 | |
| 79 | #[inline ] |
| 80 | fn is_uri_token(b: u8) -> bool { |
| 81 | URI_MAP[b as usize] |
| 82 | } |
| 83 | |
| 84 | static HEADER_NAME_MAP: [bool; 256] = byte_map![ |
| 85 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 86 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 87 | 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, |
| 88 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, |
| 89 | 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 90 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, |
| 91 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 92 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, |
| 93 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 94 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 95 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 96 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 97 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 98 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 99 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 100 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 101 | ]; |
| 102 | |
| 103 | #[inline ] |
| 104 | fn is_header_name_token(b: u8) -> bool { |
| 105 | HEADER_NAME_MAP[b as usize] |
| 106 | } |
| 107 | |
| 108 | static HEADER_VALUE_MAP: [bool; 256] = byte_map![ |
| 109 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, |
| 110 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 111 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 112 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 113 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 114 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 115 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 116 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, |
| 117 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 118 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 119 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 120 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 121 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 122 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 123 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 124 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 125 | ]; |
| 126 | |
| 127 | |
| 128 | #[inline ] |
| 129 | fn is_header_value_token(b: u8) -> bool { |
| 130 | HEADER_VALUE_MAP[b as usize] |
| 131 | } |
| 132 | |
| 133 | /// An error in parsing. |
| 134 | #[derive(Copy, Clone, PartialEq, Eq, Debug)] |
| 135 | pub enum Error { |
| 136 | /// Invalid byte in header name. |
| 137 | HeaderName, |
| 138 | /// Invalid byte in header value. |
| 139 | HeaderValue, |
| 140 | /// Invalid byte in new line. |
| 141 | NewLine, |
| 142 | /// Invalid byte in Response status. |
| 143 | Status, |
| 144 | /// Invalid byte where token is required. |
| 145 | Token, |
| 146 | /// Parsed more headers than provided buffer can contain. |
| 147 | TooManyHeaders, |
| 148 | /// Invalid byte in HTTP version. |
| 149 | Version, |
| 150 | } |
| 151 | |
| 152 | impl Error { |
| 153 | #[inline ] |
| 154 | fn description_str(&self) -> &'static str { |
| 155 | match *self { |
| 156 | Error::HeaderName => "invalid header name" , |
| 157 | Error::HeaderValue => "invalid header value" , |
| 158 | Error::NewLine => "invalid new line" , |
| 159 | Error::Status => "invalid response status" , |
| 160 | Error::Token => "invalid token" , |
| 161 | Error::TooManyHeaders => "too many headers" , |
| 162 | Error::Version => "invalid HTTP version" , |
| 163 | } |
| 164 | } |
| 165 | } |
| 166 | |
| 167 | impl fmt::Display for Error { |
| 168 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
| 169 | f.write_str(self.description_str()) |
| 170 | } |
| 171 | } |
| 172 | |
| 173 | #[cfg (feature = "std" )] |
| 174 | impl std::error::Error for Error { |
| 175 | fn description(&self) -> &str { |
| 176 | self.description_str() |
| 177 | } |
| 178 | } |
| 179 | |
| 180 | /// An error in parsing a chunk size. |
| 181 | // Note: Move this into the error enum once v2.0 is released. |
| 182 | #[derive(Debug, PartialEq, Eq)] |
| 183 | pub struct InvalidChunkSize; |
| 184 | |
| 185 | impl fmt::Display for InvalidChunkSize { |
| 186 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
| 187 | f.write_str("invalid chunk size" ) |
| 188 | } |
| 189 | } |
| 190 | |
| 191 | /// A Result of any parsing action. |
| 192 | /// |
| 193 | /// If the input is invalid, an `Error` will be returned. Note that incomplete |
| 194 | /// data is not considered invalid, and so will not return an error, but rather |
| 195 | /// a `Ok(Status::Partial)`. |
| 196 | pub type Result<T> = result::Result<Status<T>, Error>; |
| 197 | |
| 198 | /// The result of a successful parse pass. |
| 199 | /// |
| 200 | /// `Complete` is used when the buffer contained the complete value. |
| 201 | /// `Partial` is used when parsing did not reach the end of the expected value, |
| 202 | /// but no invalid data was found. |
| 203 | #[derive(Copy, Clone, Eq, PartialEq, Debug)] |
| 204 | pub enum Status<T> { |
| 205 | /// The completed result. |
| 206 | Complete(T), |
| 207 | /// A partial result. |
| 208 | Partial |
| 209 | } |
| 210 | |
| 211 | impl<T> Status<T> { |
| 212 | /// Convenience method to check if status is complete. |
| 213 | #[inline ] |
| 214 | pub fn is_complete(&self) -> bool { |
| 215 | match *self { |
| 216 | Status::Complete(..) => true, |
| 217 | Status::Partial => false |
| 218 | } |
| 219 | } |
| 220 | |
| 221 | /// Convenience method to check if status is partial. |
| 222 | #[inline ] |
| 223 | pub fn is_partial(&self) -> bool { |
| 224 | match *self { |
| 225 | Status::Complete(..) => false, |
| 226 | Status::Partial => true |
| 227 | } |
| 228 | } |
| 229 | |
| 230 | /// Convenience method to unwrap a Complete value. Panics if the status is |
| 231 | /// `Partial`. |
| 232 | #[inline ] |
| 233 | pub fn unwrap(self) -> T { |
| 234 | match self { |
| 235 | Status::Complete(t) => t, |
| 236 | Status::Partial => panic!("Tried to unwrap Status::Partial" ) |
| 237 | } |
| 238 | } |
| 239 | } |
| 240 | |
| 241 | /// Parser configuration. |
| 242 | #[derive(Clone, Debug, Default)] |
| 243 | pub struct ParserConfig { |
| 244 | allow_spaces_after_header_name_in_responses: bool, |
| 245 | allow_obsolete_multiline_headers_in_responses: bool, |
| 246 | allow_multiple_spaces_in_request_line_delimiters: bool, |
| 247 | allow_multiple_spaces_in_response_status_delimiters: bool, |
| 248 | ignore_invalid_headers_in_responses: bool, |
| 249 | } |
| 250 | |
| 251 | impl ParserConfig { |
| 252 | /// Sets whether spaces and tabs should be allowed after header names in responses. |
| 253 | pub fn allow_spaces_after_header_name_in_responses( |
| 254 | &mut self, |
| 255 | value: bool, |
| 256 | ) -> &mut Self { |
| 257 | self.allow_spaces_after_header_name_in_responses = value; |
| 258 | self |
| 259 | } |
| 260 | |
| 261 | /// Sets whether multiple spaces are allowed as delimiters in request lines. |
| 262 | /// |
| 263 | /// # Background |
| 264 | /// |
| 265 | /// The [latest version of the HTTP/1.1 spec][spec] allows implementations to parse multiple |
| 266 | /// whitespace characters in place of the `SP` delimiters in the request line, including: |
| 267 | /// |
| 268 | /// > SP, HTAB, VT (%x0B), FF (%x0C), or bare CR |
| 269 | /// |
| 270 | /// This option relaxes the parser to allow for multiple spaces, but does *not* allow the |
| 271 | /// request line to contain the other mentioned whitespace characters. |
| 272 | /// |
| 273 | /// [spec]: https://httpwg.org/http-core/draft-ietf-httpbis-messaging-latest.html#rfc.section.3.p.3 |
| 274 | pub fn allow_multiple_spaces_in_request_line_delimiters(&mut self, value: bool) -> &mut Self { |
| 275 | self.allow_multiple_spaces_in_request_line_delimiters = value; |
| 276 | self |
| 277 | } |
| 278 | |
| 279 | /// Whether multiple spaces are allowed as delimiters in request lines. |
| 280 | pub fn multiple_spaces_in_request_line_delimiters_are_allowed(&self) -> bool { |
| 281 | self.allow_multiple_spaces_in_request_line_delimiters |
| 282 | } |
| 283 | |
| 284 | /// Sets whether multiple spaces are allowed as delimiters in response status lines. |
| 285 | /// |
| 286 | /// # Background |
| 287 | /// |
| 288 | /// The [latest version of the HTTP/1.1 spec][spec] allows implementations to parse multiple |
| 289 | /// whitespace characters in place of the `SP` delimiters in the response status line, |
| 290 | /// including: |
| 291 | /// |
| 292 | /// > SP, HTAB, VT (%x0B), FF (%x0C), or bare CR |
| 293 | /// |
| 294 | /// This option relaxes the parser to allow for multiple spaces, but does *not* allow the status |
| 295 | /// line to contain the other mentioned whitespace characters. |
| 296 | /// |
| 297 | /// [spec]: https://httpwg.org/http-core/draft-ietf-httpbis-messaging-latest.html#rfc.section.4.p.3 |
| 298 | pub fn allow_multiple_spaces_in_response_status_delimiters(&mut self, value: bool) -> &mut Self { |
| 299 | self.allow_multiple_spaces_in_response_status_delimiters = value; |
| 300 | self |
| 301 | } |
| 302 | |
| 303 | /// Whether multiple spaces are allowed as delimiters in response status lines. |
| 304 | pub fn multiple_spaces_in_response_status_delimiters_are_allowed(&self) -> bool { |
| 305 | self.allow_multiple_spaces_in_response_status_delimiters |
| 306 | } |
| 307 | |
| 308 | /// Sets whether obsolete multiline headers should be allowed. |
| 309 | /// |
| 310 | /// This is an obsolete part of HTTP/1. Use at your own risk. If you are |
| 311 | /// building an HTTP library, the newlines (`\r` and `\n`) should be |
| 312 | /// replaced by spaces before handing the header value to the user. |
| 313 | /// |
| 314 | /// # Example |
| 315 | /// |
| 316 | /// ```rust |
| 317 | /// let buf = b"HTTP/1.1 200 OK \r\nFolded-Header: hello \r\n there \r\n\r\n" ; |
| 318 | /// let mut headers = [httparse::EMPTY_HEADER; 16]; |
| 319 | /// let mut response = httparse::Response::new(&mut headers); |
| 320 | /// |
| 321 | /// let res = httparse::ParserConfig::default() |
| 322 | /// .allow_obsolete_multiline_headers_in_responses(true) |
| 323 | /// .parse_response(&mut response, buf); |
| 324 | /// |
| 325 | /// assert_eq!(res, Ok(httparse::Status::Complete(buf.len()))); |
| 326 | /// |
| 327 | /// assert_eq!(response.headers.len(), 1); |
| 328 | /// assert_eq!(response.headers[0].name, "Folded-Header" ); |
| 329 | /// assert_eq!(response.headers[0].value, b"hello \r\n there" ); |
| 330 | /// ``` |
| 331 | pub fn allow_obsolete_multiline_headers_in_responses( |
| 332 | &mut self, |
| 333 | value: bool, |
| 334 | ) -> &mut Self { |
| 335 | self.allow_obsolete_multiline_headers_in_responses = value; |
| 336 | self |
| 337 | } |
| 338 | |
| 339 | /// Whether obsolete multiline headers should be allowed. |
| 340 | pub fn obsolete_multiline_headers_in_responses_are_allowed(&self) -> bool { |
| 341 | self.allow_obsolete_multiline_headers_in_responses |
| 342 | } |
| 343 | |
| 344 | /// Parses a request with the given config. |
| 345 | pub fn parse_request<'headers, 'buf>( |
| 346 | &self, |
| 347 | request: &mut Request<'headers, 'buf>, |
| 348 | buf: &'buf [u8], |
| 349 | ) -> Result<usize> { |
| 350 | request.parse_with_config(buf, self) |
| 351 | } |
| 352 | |
| 353 | /// Parses a request with the given config and buffer for headers |
| 354 | pub fn parse_request_with_uninit_headers<'headers, 'buf>( |
| 355 | &self, |
| 356 | request: &mut Request<'headers, 'buf>, |
| 357 | buf: &'buf [u8], |
| 358 | headers: &'headers mut [MaybeUninit<Header<'buf>>], |
| 359 | ) -> Result<usize> { |
| 360 | request.parse_with_config_and_uninit_headers(buf, self, headers) |
| 361 | } |
| 362 | |
| 363 | /// Sets whether invalid header lines should be silently ignored in responses. |
| 364 | /// |
| 365 | /// This mimicks the behaviour of major browsers. You probably don't want this. |
| 366 | /// You should only want this if you are implementing a proxy whose main |
| 367 | /// purpose is to sit in front of browsers whose users access arbitrary content |
| 368 | /// which may be malformed, and they expect everything that works without |
| 369 | /// the proxy to keep working with the proxy. |
| 370 | /// |
| 371 | /// This option will prevent `ParserConfig::parse_response` from returning |
| 372 | /// an error encountered when parsing a header, except if the error was caused |
| 373 | /// by the character NUL (ASCII code 0), as Chrome specifically always reject |
| 374 | /// those, or if the error was caused by a lone character `\r`, as Firefox and |
| 375 | /// Chrome behave differently in that case. |
| 376 | /// |
| 377 | /// The ignorable errors are: |
| 378 | /// * empty header names; |
| 379 | /// * characters that are not allowed in header names, except for `\0` and `\r`; |
| 380 | /// * when `allow_spaces_after_header_name_in_responses` is not enabled, |
| 381 | /// spaces and tabs between the header name and the colon; |
| 382 | /// * missing colon between header name and value; |
| 383 | /// * when `allow_obsolete_multiline_headers_in_responses` is not enabled, |
| 384 | /// headers using obsolete line folding. |
| 385 | /// * characters that are not allowed in header values except for `\0` and `\r`. |
| 386 | /// |
| 387 | /// If an ignorable error is encountered, the parser tries to find the next |
| 388 | /// line in the input to resume parsing the rest of the headers. As lines |
| 389 | /// contributing to a header using obsolete line folding always start |
| 390 | /// with whitespace, those will be ignored too. An error will be emitted |
| 391 | /// nonetheless if it finds `\0` or a lone `\r` while looking for the |
| 392 | /// next line. |
| 393 | pub fn ignore_invalid_headers_in_responses( |
| 394 | &mut self, |
| 395 | value: bool, |
| 396 | ) -> &mut Self { |
| 397 | self.ignore_invalid_headers_in_responses = value; |
| 398 | self |
| 399 | } |
| 400 | |
| 401 | /// Parses a response with the given config. |
| 402 | pub fn parse_response<'headers, 'buf>( |
| 403 | &self, |
| 404 | response: &mut Response<'headers, 'buf>, |
| 405 | buf: &'buf [u8], |
| 406 | ) -> Result<usize> { |
| 407 | response.parse_with_config(buf, self) |
| 408 | } |
| 409 | |
| 410 | /// Parses a response with the given config and buffer for headers |
| 411 | pub fn parse_response_with_uninit_headers<'headers, 'buf>( |
| 412 | &self, |
| 413 | response: &mut Response<'headers, 'buf>, |
| 414 | buf: &'buf [u8], |
| 415 | headers: &'headers mut [MaybeUninit<Header<'buf>>], |
| 416 | ) -> Result<usize> { |
| 417 | response.parse_with_config_and_uninit_headers(buf, self, headers) |
| 418 | } |
| 419 | } |
| 420 | |
| 421 | /// A parsed Request. |
| 422 | /// |
| 423 | /// The optional values will be `None` if a parse was not complete, and did not |
| 424 | /// parse the associated property. This allows you to inspect the parts that |
| 425 | /// could be parsed, before reading more, in case you wish to exit early. |
| 426 | /// |
| 427 | /// # Example |
| 428 | /// |
| 429 | /// ```no_run |
| 430 | /// let buf = b"GET /404 HTTP/1.1 \r\nHost:" ; |
| 431 | /// let mut headers = [httparse::EMPTY_HEADER; 16]; |
| 432 | /// let mut req = httparse::Request::new(&mut headers); |
| 433 | /// let res = req.parse(buf).unwrap(); |
| 434 | /// if res.is_partial() { |
| 435 | /// match req.path { |
| 436 | /// Some(ref path) => { |
| 437 | /// // check router for path. |
| 438 | /// // /404 doesn't exist? we could stop parsing |
| 439 | /// }, |
| 440 | /// None => { |
| 441 | /// // must read more and parse again |
| 442 | /// } |
| 443 | /// } |
| 444 | /// } |
| 445 | /// ``` |
| 446 | #[derive(Debug, Eq, PartialEq)] |
| 447 | pub struct Request<'headers, 'buf> { |
| 448 | /// The request method, such as `GET`. |
| 449 | pub method: Option<&'buf str>, |
| 450 | /// The request path, such as `/about-us`. |
| 451 | pub path: Option<&'buf str>, |
| 452 | /// The request minor version, such as `1` for `HTTP/1.1`. |
| 453 | pub version: Option<u8>, |
| 454 | /// The request headers. |
| 455 | pub headers: &'headers mut [Header<'buf>] |
| 456 | } |
| 457 | |
| 458 | impl<'h, 'b> Request<'h, 'b> { |
| 459 | /// Creates a new Request, using a slice of headers you allocate. |
| 460 | #[inline ] |
| 461 | pub fn new(headers: &'h mut [Header<'b>]) -> Request<'h, 'b> { |
| 462 | Request { |
| 463 | method: None, |
| 464 | path: None, |
| 465 | version: None, |
| 466 | headers, |
| 467 | } |
| 468 | } |
| 469 | |
| 470 | fn parse_with_config_and_uninit_headers( |
| 471 | &mut self, |
| 472 | buf: &'b [u8], |
| 473 | config: &ParserConfig, |
| 474 | mut headers: &'h mut [MaybeUninit<Header<'b>>], |
| 475 | ) -> Result<usize> { |
| 476 | let orig_len = buf.len(); |
| 477 | let mut bytes = Bytes::new(buf); |
| 478 | complete!(skip_empty_lines(&mut bytes)); |
| 479 | const GET: [u8; 4] = *b"GET " ; |
| 480 | const POST: [u8; 4] = *b"POST" ; |
| 481 | let method = match bytes.peek_n::<[u8; 4]>(4) { |
| 482 | Some(GET) => { |
| 483 | unsafe { |
| 484 | bytes.advance_and_commit(4); |
| 485 | } |
| 486 | "GET" |
| 487 | } |
| 488 | Some(POST) if bytes.peek_ahead(4) == Some(b' ' ) => { |
| 489 | unsafe { |
| 490 | bytes.advance_and_commit(5); |
| 491 | } |
| 492 | "POST" |
| 493 | } |
| 494 | _ => complete!(parse_token(&mut bytes)), |
| 495 | }; |
| 496 | self.method = Some(method); |
| 497 | if config.allow_multiple_spaces_in_request_line_delimiters { |
| 498 | complete!(skip_spaces(&mut bytes)); |
| 499 | } |
| 500 | self.path = Some(complete!(parse_uri(&mut bytes))); |
| 501 | if config.allow_multiple_spaces_in_request_line_delimiters { |
| 502 | complete!(skip_spaces(&mut bytes)); |
| 503 | } |
| 504 | self.version = Some(complete!(parse_version(&mut bytes))); |
| 505 | newline!(bytes); |
| 506 | |
| 507 | let len = orig_len - bytes.len(); |
| 508 | let headers_len = complete!(parse_headers_iter_uninit( |
| 509 | &mut headers, |
| 510 | &mut bytes, |
| 511 | &ParserConfig::default(), |
| 512 | )); |
| 513 | /* SAFETY: see `parse_headers_iter_uninit` guarantees */ |
| 514 | self.headers = unsafe { assume_init_slice(headers) }; |
| 515 | |
| 516 | Ok(Status::Complete(len + headers_len)) |
| 517 | } |
| 518 | |
| 519 | /// Try to parse a buffer of bytes into the Request, |
| 520 | /// except use an uninitialized slice of `Header`s. |
| 521 | /// |
| 522 | /// For more information, see `parse` |
| 523 | pub fn parse_with_uninit_headers( |
| 524 | &mut self, |
| 525 | buf: &'b [u8], |
| 526 | headers: &'h mut [MaybeUninit<Header<'b>>], |
| 527 | ) -> Result<usize> { |
| 528 | self.parse_with_config_and_uninit_headers(buf, &Default::default(), headers) |
| 529 | } |
| 530 | |
| 531 | fn parse_with_config(&mut self, buf: &'b [u8], config: &ParserConfig) -> Result<usize> { |
| 532 | let headers = mem::replace(&mut self.headers, &mut []); |
| 533 | |
| 534 | /* SAFETY: see `parse_headers_iter_uninit` guarantees */ |
| 535 | unsafe { |
| 536 | let headers: *mut [Header<'_>] = headers; |
| 537 | let headers = headers as *mut [MaybeUninit<Header<'_>>]; |
| 538 | match self.parse_with_config_and_uninit_headers(buf, config, &mut *headers) { |
| 539 | Ok(Status::Complete(idx)) => Ok(Status::Complete(idx)), |
| 540 | other => { |
| 541 | // put the original headers back |
| 542 | self.headers = &mut *(headers as *mut [Header<'_>]); |
| 543 | other |
| 544 | }, |
| 545 | } |
| 546 | } |
| 547 | } |
| 548 | |
| 549 | /// Try to parse a buffer of bytes into the Request. |
| 550 | /// |
| 551 | /// Returns byte offset in `buf` to start of HTTP body. |
| 552 | pub fn parse(&mut self, buf: &'b [u8]) -> Result<usize> { |
| 553 | self.parse_with_config(buf, &Default::default()) |
| 554 | } |
| 555 | } |
| 556 | |
| 557 | #[inline ] |
| 558 | fn skip_empty_lines(bytes: &mut Bytes<'_>) -> Result<()> { |
| 559 | loop { |
| 560 | let b = bytes.peek(); |
| 561 | match b { |
| 562 | Some(b' \r' ) => { |
| 563 | // there's `\r`, so it's safe to bump 1 pos |
| 564 | unsafe { bytes.bump() }; |
| 565 | expect!(bytes.next() == b' \n' => Err(Error::NewLine)); |
| 566 | }, |
| 567 | Some(b' \n' ) => { |
| 568 | // there's `\n`, so it's safe to bump 1 pos |
| 569 | unsafe { bytes.bump(); } |
| 570 | }, |
| 571 | Some(..) => { |
| 572 | bytes.slice(); |
| 573 | return Ok(Status::Complete(())); |
| 574 | }, |
| 575 | None => return Ok(Status::Partial) |
| 576 | } |
| 577 | } |
| 578 | } |
| 579 | |
| 580 | #[inline ] |
| 581 | fn skip_spaces(bytes: &mut Bytes<'_>) -> Result<()> { |
| 582 | loop { |
| 583 | let b = bytes.peek(); |
| 584 | match b { |
| 585 | Some(b' ' ) => { |
| 586 | // there's ` `, so it's safe to bump 1 pos |
| 587 | unsafe { bytes.bump() }; |
| 588 | } |
| 589 | Some(..) => { |
| 590 | bytes.slice(); |
| 591 | return Ok(Status::Complete(())); |
| 592 | } |
| 593 | None => return Ok(Status::Partial), |
| 594 | } |
| 595 | } |
| 596 | } |
| 597 | |
| 598 | /// A parsed Response. |
| 599 | /// |
| 600 | /// See `Request` docs for explanation of optional values. |
| 601 | #[derive(Debug, Eq, PartialEq)] |
| 602 | pub struct Response<'headers, 'buf> { |
| 603 | /// The response minor version, such as `1` for `HTTP/1.1`. |
| 604 | pub version: Option<u8>, |
| 605 | /// The response code, such as `200`. |
| 606 | pub code: Option<u16>, |
| 607 | /// The response reason-phrase, such as `OK`. |
| 608 | /// |
| 609 | /// Contains an empty string if the reason-phrase was missing or contained invalid characters. |
| 610 | pub reason: Option<&'buf str>, |
| 611 | /// The response headers. |
| 612 | pub headers: &'headers mut [Header<'buf>] |
| 613 | } |
| 614 | |
| 615 | impl<'h, 'b> Response<'h, 'b> { |
| 616 | /// Creates a new `Response` using a slice of `Header`s you have allocated. |
| 617 | #[inline ] |
| 618 | pub fn new(headers: &'h mut [Header<'b>]) -> Response<'h, 'b> { |
| 619 | Response { |
| 620 | version: None, |
| 621 | code: None, |
| 622 | reason: None, |
| 623 | headers, |
| 624 | } |
| 625 | } |
| 626 | |
| 627 | /// Try to parse a buffer of bytes into this `Response`. |
| 628 | pub fn parse(&mut self, buf: &'b [u8]) -> Result<usize> { |
| 629 | self.parse_with_config(buf, &ParserConfig::default()) |
| 630 | } |
| 631 | |
| 632 | fn parse_with_config(&mut self, buf: &'b [u8], config: &ParserConfig) -> Result<usize> { |
| 633 | let headers = mem::replace(&mut self.headers, &mut []); |
| 634 | |
| 635 | unsafe { |
| 636 | let headers: *mut [Header<'_>] = headers; |
| 637 | let headers = headers as *mut [MaybeUninit<Header<'_>>]; |
| 638 | match self.parse_with_config_and_uninit_headers(buf, config, &mut *headers) { |
| 639 | Ok(Status::Complete(idx)) => Ok(Status::Complete(idx)), |
| 640 | other => { |
| 641 | // put the original headers back |
| 642 | self.headers = &mut *(headers as *mut [Header<'_>]); |
| 643 | other |
| 644 | }, |
| 645 | } |
| 646 | } |
| 647 | } |
| 648 | |
| 649 | fn parse_with_config_and_uninit_headers( |
| 650 | &mut self, |
| 651 | buf: &'b [u8], |
| 652 | config: &ParserConfig, |
| 653 | mut headers: &'h mut [MaybeUninit<Header<'b>>], |
| 654 | ) -> Result<usize> { |
| 655 | let orig_len = buf.len(); |
| 656 | let mut bytes = Bytes::new(buf); |
| 657 | |
| 658 | complete!(skip_empty_lines(&mut bytes)); |
| 659 | self.version = Some(complete!(parse_version(&mut bytes))); |
| 660 | space!(bytes or Error::Version); |
| 661 | if config.allow_multiple_spaces_in_response_status_delimiters { |
| 662 | complete!(skip_spaces(&mut bytes)); |
| 663 | } |
| 664 | self.code = Some(complete!(parse_code(&mut bytes))); |
| 665 | |
| 666 | // RFC7230 says there must be 'SP' and then reason-phrase, but admits |
| 667 | // its only for legacy reasons. With the reason-phrase completely |
| 668 | // optional (and preferred to be omitted) in HTTP2, we'll just |
| 669 | // handle any response that doesn't include a reason-phrase, because |
| 670 | // it's more lenient, and we don't care anyways. |
| 671 | // |
| 672 | // So, a SP means parse a reason-phrase. |
| 673 | // A newline means go to headers. |
| 674 | // Anything else we'll say is a malformed status. |
| 675 | match next!(bytes) { |
| 676 | b' ' => { |
| 677 | if config.allow_multiple_spaces_in_response_status_delimiters { |
| 678 | complete!(skip_spaces(&mut bytes)); |
| 679 | } |
| 680 | bytes.slice(); |
| 681 | self.reason = Some(complete!(parse_reason(&mut bytes))); |
| 682 | }, |
| 683 | b' \r' => { |
| 684 | expect!(bytes.next() == b' \n' => Err(Error::Status)); |
| 685 | bytes.slice(); |
| 686 | self.reason = Some("" ); |
| 687 | }, |
| 688 | b' \n' => { |
| 689 | bytes.slice(); |
| 690 | self.reason = Some("" ); |
| 691 | } |
| 692 | _ => return Err(Error::Status), |
| 693 | } |
| 694 | |
| 695 | |
| 696 | let len = orig_len - bytes.len(); |
| 697 | let headers_len = complete!(parse_headers_iter_uninit( |
| 698 | &mut headers, |
| 699 | &mut bytes, |
| 700 | config |
| 701 | )); |
| 702 | /* SAFETY: see `parse_headers_iter_uninit` guarantees */ |
| 703 | self.headers = unsafe { assume_init_slice(headers) }; |
| 704 | Ok(Status::Complete(len + headers_len)) |
| 705 | } |
| 706 | } |
| 707 | |
| 708 | /// Represents a parsed header. |
| 709 | #[derive(Copy, Clone, Eq, PartialEq)] |
| 710 | pub struct Header<'a> { |
| 711 | /// The name portion of a header. |
| 712 | /// |
| 713 | /// A header name must be valid ASCII-US, so it's safe to store as a `&str`. |
| 714 | pub name: &'a str, |
| 715 | /// The value portion of a header. |
| 716 | /// |
| 717 | /// While headers **should** be ASCII-US, the specification allows for |
| 718 | /// values that may not be, and so the value is stored as bytes. |
| 719 | pub value: &'a [u8], |
| 720 | } |
| 721 | |
| 722 | impl<'a> fmt::Debug for Header<'a> { |
| 723 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
| 724 | let mut f = f.debug_struct("Header" ); |
| 725 | f.field("name" , &self.name); |
| 726 | if let Ok(value) = str::from_utf8(self.value) { |
| 727 | f.field("value" , &value); |
| 728 | } else { |
| 729 | f.field("value" , &self.value); |
| 730 | } |
| 731 | f.finish() |
| 732 | } |
| 733 | } |
| 734 | |
| 735 | /// An empty header, useful for constructing a `Header` array to pass in for |
| 736 | /// parsing. |
| 737 | /// |
| 738 | /// # Example |
| 739 | /// |
| 740 | /// ``` |
| 741 | /// let headers = [httparse::EMPTY_HEADER; 64]; |
| 742 | /// ``` |
| 743 | pub const EMPTY_HEADER: Header<'static> = Header { name: "" , value: b"" }; |
| 744 | |
| 745 | #[inline ] |
| 746 | fn parse_version(bytes: &mut Bytes<'_>) -> Result<u8> { |
| 747 | if let Some(eight) = bytes.peek_n::<[u8; 8]>(8) { |
| 748 | unsafe { bytes.advance(8); } |
| 749 | return match &eight { |
| 750 | b"HTTP/1.0" => Ok(Status::Complete(0)), |
| 751 | b"HTTP/1.1" => Ok(Status::Complete(1)), |
| 752 | _ => Err(Error::Version), |
| 753 | } |
| 754 | } |
| 755 | |
| 756 | // else (but not in `else` because of borrow checker) |
| 757 | |
| 758 | // If there aren't at least 8 bytes, we still want to detect early |
| 759 | // if this is a valid version or not. If it is, we'll return Partial. |
| 760 | expect!(bytes.next() == b'H' => Err(Error::Version)); |
| 761 | expect!(bytes.next() == b'T' => Err(Error::Version)); |
| 762 | expect!(bytes.next() == b'T' => Err(Error::Version)); |
| 763 | expect!(bytes.next() == b'P' => Err(Error::Version)); |
| 764 | expect!(bytes.next() == b'/' => Err(Error::Version)); |
| 765 | expect!(bytes.next() == b'1' => Err(Error::Version)); |
| 766 | expect!(bytes.next() == b'.' => Err(Error::Version)); |
| 767 | Ok(Status::Partial) |
| 768 | } |
| 769 | |
| 770 | /// From [RFC 7230](https://tools.ietf.org/html/rfc7230): |
| 771 | /// |
| 772 | /// > ```notrust |
| 773 | /// > reason-phrase = *( HTAB / SP / VCHAR / obs-text ) |
| 774 | /// > HTAB = %x09 ; horizontal tab |
| 775 | /// > VCHAR = %x21-7E ; visible (printing) characters |
| 776 | /// > obs-text = %x80-FF |
| 777 | /// > ``` |
| 778 | /// |
| 779 | /// > A.2. Changes from RFC 2616 |
| 780 | /// > |
| 781 | /// > Non-US-ASCII content in header fields and the reason phrase |
| 782 | /// > has been obsoleted and made opaque (the TEXT rule was removed). |
| 783 | #[inline ] |
| 784 | fn parse_reason<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> { |
| 785 | let mut seen_obs_text = false; |
| 786 | loop { |
| 787 | let b = next!(bytes); |
| 788 | if b == b' \r' { |
| 789 | expect!(bytes.next() == b' \n' => Err(Error::Status)); |
| 790 | return Ok(Status::Complete(unsafe { |
| 791 | let bytes = bytes.slice_skip(2); |
| 792 | if !seen_obs_text { |
| 793 | // all bytes up till `i` must have been HTAB / SP / VCHAR |
| 794 | str::from_utf8_unchecked(bytes) |
| 795 | } else { |
| 796 | // obs-text characters were found, so return the fallback empty string |
| 797 | "" |
| 798 | } |
| 799 | })); |
| 800 | } else if b == b' \n' { |
| 801 | return Ok(Status::Complete(unsafe { |
| 802 | let bytes = bytes.slice_skip(1); |
| 803 | if !seen_obs_text { |
| 804 | // all bytes up till `i` must have been HTAB / SP / VCHAR |
| 805 | str::from_utf8_unchecked(bytes) |
| 806 | } else { |
| 807 | // obs-text characters were found, so return the fallback empty string |
| 808 | "" |
| 809 | } |
| 810 | })); |
| 811 | } else if !(b == 0x09 || b == b' ' || (0x21..=0x7E).contains(&b) || b >= 0x80) { |
| 812 | return Err(Error::Status); |
| 813 | } else if b >= 0x80 { |
| 814 | seen_obs_text = true; |
| 815 | } |
| 816 | } |
| 817 | } |
| 818 | |
| 819 | #[inline ] |
| 820 | fn parse_token<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> { |
| 821 | let b = next!(bytes); |
| 822 | if !is_token(b) { |
| 823 | // First char must be a token char, it can't be a space which would indicate an empty token. |
| 824 | return Err(Error::Token); |
| 825 | } |
| 826 | |
| 827 | loop { |
| 828 | let b = next!(bytes); |
| 829 | if b == b' ' { |
| 830 | return Ok(Status::Complete(unsafe { |
| 831 | // all bytes up till `i` must have been `is_token`. |
| 832 | str::from_utf8_unchecked(bytes.slice_skip(1)) |
| 833 | })); |
| 834 | } else if !is_token(b) { |
| 835 | return Err(Error::Token); |
| 836 | } |
| 837 | } |
| 838 | } |
| 839 | |
| 840 | #[inline ] |
| 841 | fn parse_uri<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> { |
| 842 | let b = next!(bytes); |
| 843 | if !is_uri_token(b) { |
| 844 | // First char must be a URI char, it can't be a space which would indicate an empty path. |
| 845 | return Err(Error::Token); |
| 846 | } |
| 847 | |
| 848 | simd::match_uri_vectored(bytes); |
| 849 | |
| 850 | loop { |
| 851 | let b = next!(bytes); |
| 852 | if b == b' ' { |
| 853 | return Ok(Status::Complete(unsafe { |
| 854 | // all bytes up till `i` must have been `is_token`. |
| 855 | str::from_utf8_unchecked(bytes.slice_skip(1)) |
| 856 | })); |
| 857 | } else if !is_uri_token(b) { |
| 858 | return Err(Error::Token); |
| 859 | } |
| 860 | } |
| 861 | } |
| 862 | |
| 863 | |
| 864 | #[inline ] |
| 865 | fn parse_code(bytes: &mut Bytes<'_>) -> Result<u16> { |
| 866 | let hundreds = expect!(bytes.next() == b'0' ..=b'9' => Err(Error::Status)); |
| 867 | let tens = expect!(bytes.next() == b'0' ..=b'9' => Err(Error::Status)); |
| 868 | let ones = expect!(bytes.next() == b'0' ..=b'9' => Err(Error::Status)); |
| 869 | |
| 870 | Ok(Status::Complete((hundreds - b'0' ) as u16 * 100 + |
| 871 | (tens - b'0' ) as u16 * 10 + |
| 872 | (ones - b'0' ) as u16)) |
| 873 | } |
| 874 | |
| 875 | /// Parse a buffer of bytes as headers. |
| 876 | /// |
| 877 | /// The return value, if complete and successful, includes the index of the |
| 878 | /// buffer that parsing stopped at, and a sliced reference to the parsed |
| 879 | /// headers. The length of the slice will be equal to the number of properly |
| 880 | /// parsed headers. |
| 881 | /// |
| 882 | /// # Example |
| 883 | /// |
| 884 | /// ``` |
| 885 | /// let buf = b"Host: foo.bar \nAccept: */* \n\nblah blah" ; |
| 886 | /// let mut headers = [httparse::EMPTY_HEADER; 4]; |
| 887 | /// assert_eq!(httparse::parse_headers(buf, &mut headers), |
| 888 | /// Ok(httparse::Status::Complete((27, &[ |
| 889 | /// httparse::Header { name: "Host" , value: b"foo.bar" }, |
| 890 | /// httparse::Header { name: "Accept" , value: b"*/*" } |
| 891 | /// ][..])))); |
| 892 | /// ``` |
| 893 | pub fn parse_headers<'b: 'h, 'h>( |
| 894 | src: &'b [u8], |
| 895 | mut dst: &'h mut [Header<'b>], |
| 896 | ) -> Result<(usize, &'h [Header<'b>])> { |
| 897 | let mut iter = Bytes::new(src); |
| 898 | let pos = complete!(parse_headers_iter(&mut dst, &mut iter, &ParserConfig::default())); |
| 899 | Ok(Status::Complete((pos, dst))) |
| 900 | } |
| 901 | |
| 902 | #[inline ] |
| 903 | fn parse_headers_iter<'a, 'b>( |
| 904 | headers: &mut &mut [Header<'a>], |
| 905 | bytes: &'b mut Bytes<'a>, |
| 906 | config: &ParserConfig, |
| 907 | ) -> Result<usize> { |
| 908 | parse_headers_iter_uninit( |
| 909 | /* SAFETY: see `parse_headers_iter_uninit` guarantees */ |
| 910 | unsafe { deinit_slice_mut(headers) }, |
| 911 | bytes, |
| 912 | config, |
| 913 | ) |
| 914 | } |
| 915 | |
| 916 | unsafe fn deinit_slice_mut<'a, 'b, T>(s: &'a mut &'b mut [T]) -> &'a mut &'b mut [MaybeUninit<T>] { |
| 917 | let s: *mut &mut [T] = s; |
| 918 | let s = s as *mut &mut [MaybeUninit<T>]; |
| 919 | &mut *s |
| 920 | } |
| 921 | unsafe fn assume_init_slice<T>(s: &mut [MaybeUninit<T>]) -> &mut [T] { |
| 922 | let s: *mut [MaybeUninit<T>] = s; |
| 923 | let s = s as *mut [T]; |
| 924 | &mut *s |
| 925 | } |
| 926 | |
| 927 | /* Function which parsers headers into uninitialized buffer. |
| 928 | * |
| 929 | * Guarantees that it doesn't write garbage, so casting |
| 930 | * &mut &mut [Header] -> &mut &mut [MaybeUninit<Header>] |
| 931 | * is safe here. |
| 932 | * |
| 933 | * Also it promises `headers` get shrunk to number of initialized headers, |
| 934 | * so casting the other way around after calling this function is safe |
| 935 | */ |
| 936 | fn parse_headers_iter_uninit<'a, 'b>( |
| 937 | headers: &mut &mut [MaybeUninit<Header<'a>>], |
| 938 | bytes: &'b mut Bytes<'a>, |
| 939 | config: &ParserConfig, |
| 940 | ) -> Result<usize> { |
| 941 | |
| 942 | /* Flow of this function is pretty complex, especially with macros, |
| 943 | * so this struct makes sure we shrink `headers` to only parsed ones. |
| 944 | * Comparing to previous code, this only may introduce some additional |
| 945 | * instructions in case of early return */ |
| 946 | struct ShrinkOnDrop<'r1, 'r2, 'a> { |
| 947 | headers: &'r1 mut &'r2 mut [MaybeUninit<Header<'a>>], |
| 948 | num_headers: usize, |
| 949 | } |
| 950 | |
| 951 | impl<'r1, 'r2, 'a> Drop for ShrinkOnDrop<'r1, 'r2, 'a> { |
| 952 | fn drop(&mut self) { |
| 953 | let headers = mem::replace(self.headers, &mut []); |
| 954 | |
| 955 | /* SAFETY: num_headers is the number of initialized headers */ |
| 956 | let headers = unsafe { headers.get_unchecked_mut(..self.num_headers) }; |
| 957 | |
| 958 | *self.headers = headers; |
| 959 | } |
| 960 | } |
| 961 | |
| 962 | let mut autoshrink = ShrinkOnDrop { |
| 963 | headers, |
| 964 | num_headers: 0, |
| 965 | }; |
| 966 | let mut count: usize = 0; |
| 967 | let mut result = Err(Error::TooManyHeaders); |
| 968 | |
| 969 | let mut iter = autoshrink.headers.iter_mut(); |
| 970 | |
| 971 | macro_rules! maybe_continue_after_obsolete_line_folding { |
| 972 | ($bytes:ident, $label:lifetime) => { |
| 973 | if config.allow_obsolete_multiline_headers_in_responses { |
| 974 | match $bytes.peek() { |
| 975 | None => { |
| 976 | // Next byte may be a space, in which case that header |
| 977 | // is using obsolete line folding, so we may have more |
| 978 | // whitespace to skip after colon. |
| 979 | return Ok(Status::Partial); |
| 980 | } |
| 981 | Some(b' ' ) | Some(b' \t' ) => { |
| 982 | // The space will be consumed next iteration. |
| 983 | continue $label; |
| 984 | } |
| 985 | _ => { |
| 986 | // There is another byte after the end of the line, |
| 987 | // but it's not whitespace, so it's probably another |
| 988 | // header or the final line return. This header is thus |
| 989 | // empty. |
| 990 | }, |
| 991 | } |
| 992 | } |
| 993 | } |
| 994 | } |
| 995 | |
| 996 | 'headers: loop { |
| 997 | // Return the error `$err` if `ignore_invalid_headers_in_responses` |
| 998 | // is false, otherwise find the end of the current line and resume |
| 999 | // parsing on the next one. |
| 1000 | macro_rules! handle_invalid_char { |
| 1001 | ($bytes:ident, $b:ident, $err:ident) => { |
| 1002 | if !config.ignore_invalid_headers_in_responses { |
| 1003 | return Err(Error::$err); |
| 1004 | } |
| 1005 | |
| 1006 | let mut b = $b; |
| 1007 | |
| 1008 | loop { |
| 1009 | if b == b' \r' { |
| 1010 | expect!(bytes.next() == b' \n' => Err(Error::$err)); |
| 1011 | break; |
| 1012 | } |
| 1013 | if b == b' \n' { |
| 1014 | break; |
| 1015 | } |
| 1016 | if b == b' \0' { |
| 1017 | return Err(Error::$err); |
| 1018 | } |
| 1019 | b = next!($bytes); |
| 1020 | } |
| 1021 | |
| 1022 | count += $bytes.pos(); |
| 1023 | $bytes.slice(); |
| 1024 | |
| 1025 | continue 'headers; |
| 1026 | }; |
| 1027 | } |
| 1028 | |
| 1029 | // a newline here means the head is over! |
| 1030 | let b = next!(bytes); |
| 1031 | if b == b' \r' { |
| 1032 | expect!(bytes.next() == b' \n' => Err(Error::NewLine)); |
| 1033 | result = Ok(Status::Complete(count + bytes.pos())); |
| 1034 | break; |
| 1035 | } |
| 1036 | if b == b' \n' { |
| 1037 | result = Ok(Status::Complete(count + bytes.pos())); |
| 1038 | break; |
| 1039 | } |
| 1040 | if !is_header_name_token(b) { |
| 1041 | handle_invalid_char!(bytes, b, HeaderName); |
| 1042 | } |
| 1043 | |
| 1044 | // parse header name until colon |
| 1045 | let header_name: &str = 'name: loop { |
| 1046 | let mut b = next!(bytes); |
| 1047 | |
| 1048 | if is_header_name_token(b) { |
| 1049 | continue 'name; |
| 1050 | } |
| 1051 | |
| 1052 | count += bytes.pos(); |
| 1053 | let name = unsafe { |
| 1054 | str::from_utf8_unchecked(bytes.slice_skip(1)) |
| 1055 | }; |
| 1056 | |
| 1057 | if b == b':' { |
| 1058 | break 'name name; |
| 1059 | } |
| 1060 | |
| 1061 | if config.allow_spaces_after_header_name_in_responses { |
| 1062 | while b == b' ' || b == b' \t' { |
| 1063 | b = next!(bytes); |
| 1064 | |
| 1065 | if b == b':' { |
| 1066 | count += bytes.pos(); |
| 1067 | bytes.slice(); |
| 1068 | break 'name name; |
| 1069 | } |
| 1070 | } |
| 1071 | } |
| 1072 | |
| 1073 | handle_invalid_char!(bytes, b, HeaderName); |
| 1074 | }; |
| 1075 | |
| 1076 | let mut b; |
| 1077 | |
| 1078 | let value_slice = 'value: loop { |
| 1079 | // eat white space between colon and value |
| 1080 | 'whitespace_after_colon: loop { |
| 1081 | b = next!(bytes); |
| 1082 | if b == b' ' || b == b' \t' { |
| 1083 | count += bytes.pos(); |
| 1084 | bytes.slice(); |
| 1085 | continue 'whitespace_after_colon; |
| 1086 | } |
| 1087 | if is_header_value_token(b) { |
| 1088 | break 'whitespace_after_colon; |
| 1089 | } |
| 1090 | |
| 1091 | if b == b' \r' { |
| 1092 | expect!(bytes.next() == b' \n' => Err(Error::HeaderValue)); |
| 1093 | } else if b != b' \n' { |
| 1094 | handle_invalid_char!(bytes, b, HeaderValue); |
| 1095 | } |
| 1096 | |
| 1097 | maybe_continue_after_obsolete_line_folding!(bytes, 'whitespace_after_colon); |
| 1098 | |
| 1099 | count += bytes.pos(); |
| 1100 | let whitespace_slice = bytes.slice(); |
| 1101 | |
| 1102 | // This produces an empty slice that points to the beginning |
| 1103 | // of the whitespace. |
| 1104 | break 'value &whitespace_slice[0..0]; |
| 1105 | } |
| 1106 | |
| 1107 | 'value_lines: loop { |
| 1108 | // parse value till EOL |
| 1109 | |
| 1110 | simd::match_header_value_vectored(bytes); |
| 1111 | |
| 1112 | 'value_line: loop { |
| 1113 | if let Some(bytes8) = bytes.peek_n::<[u8; 8]>(8) { |
| 1114 | macro_rules! check { |
| 1115 | ($bytes:ident, $i:literal) => ({ |
| 1116 | b = $bytes[$i]; |
| 1117 | if !is_header_value_token(b) { |
| 1118 | unsafe { bytes.advance($i + 1); } |
| 1119 | break 'value_line; |
| 1120 | } |
| 1121 | }); |
| 1122 | } |
| 1123 | |
| 1124 | check!(bytes8, 0); |
| 1125 | check!(bytes8, 1); |
| 1126 | check!(bytes8, 2); |
| 1127 | check!(bytes8, 3); |
| 1128 | check!(bytes8, 4); |
| 1129 | check!(bytes8, 5); |
| 1130 | check!(bytes8, 6); |
| 1131 | check!(bytes8, 7); |
| 1132 | unsafe { bytes.advance(8); } |
| 1133 | |
| 1134 | continue 'value_line; |
| 1135 | } |
| 1136 | |
| 1137 | b = next!(bytes); |
| 1138 | if !is_header_value_token(b) { |
| 1139 | break 'value_line; |
| 1140 | } |
| 1141 | } |
| 1142 | |
| 1143 | //found_ctl |
| 1144 | let skip = if b == b' \r' { |
| 1145 | expect!(bytes.next() == b' \n' => Err(Error::HeaderValue)); |
| 1146 | 2 |
| 1147 | } else if b == b' \n' { |
| 1148 | 1 |
| 1149 | } else { |
| 1150 | handle_invalid_char!(bytes, b, HeaderValue); |
| 1151 | }; |
| 1152 | |
| 1153 | maybe_continue_after_obsolete_line_folding!(bytes, 'value_lines); |
| 1154 | |
| 1155 | count += bytes.pos(); |
| 1156 | // having just checked that a newline exists, it's safe to skip it. |
| 1157 | unsafe { |
| 1158 | break 'value bytes.slice_skip(skip); |
| 1159 | } |
| 1160 | } |
| 1161 | }; |
| 1162 | |
| 1163 | let uninit_header = match iter.next() { |
| 1164 | Some(header) => header, |
| 1165 | None => break 'headers |
| 1166 | }; |
| 1167 | |
| 1168 | // trim trailing whitespace in the header |
| 1169 | let header_value = if let Some(last_visible) = value_slice |
| 1170 | .iter() |
| 1171 | .rposition(|b| *b != b' ' && *b != b' \t' && *b != b' \r' && *b != b' \n' ) |
| 1172 | { |
| 1173 | // There is at least one non-whitespace character. |
| 1174 | &value_slice[0..last_visible+1] |
| 1175 | } else { |
| 1176 | // There is no non-whitespace character. This can only happen when value_slice is |
| 1177 | // empty. |
| 1178 | value_slice |
| 1179 | }; |
| 1180 | |
| 1181 | *uninit_header = MaybeUninit::new(Header { |
| 1182 | name: header_name, |
| 1183 | value: header_value, |
| 1184 | }); |
| 1185 | autoshrink.num_headers += 1; |
| 1186 | } |
| 1187 | |
| 1188 | result |
| 1189 | } |
| 1190 | |
| 1191 | /// Parse a buffer of bytes as a chunk size. |
| 1192 | /// |
| 1193 | /// The return value, if complete and successful, includes the index of the |
| 1194 | /// buffer that parsing stopped at, and the size of the following chunk. |
| 1195 | /// |
| 1196 | /// # Example |
| 1197 | /// |
| 1198 | /// ``` |
| 1199 | /// let buf = b"4 \r\nRust \r\n0 \r\n\r\n" ; |
| 1200 | /// assert_eq!(httparse::parse_chunk_size(buf), |
| 1201 | /// Ok(httparse::Status::Complete((3, 4)))); |
| 1202 | /// ``` |
| 1203 | pub fn parse_chunk_size(buf: &[u8]) |
| 1204 | -> result::Result<Status<(usize, u64)>, InvalidChunkSize> { |
| 1205 | const RADIX: u64 = 16; |
| 1206 | let mut bytes = Bytes::new(buf); |
| 1207 | let mut size = 0; |
| 1208 | let mut in_chunk_size = true; |
| 1209 | let mut in_ext = false; |
| 1210 | let mut count = 0; |
| 1211 | loop { |
| 1212 | let b = next!(bytes); |
| 1213 | match b { |
| 1214 | b'0' ..= b'9' if in_chunk_size => { |
| 1215 | if count > 15 { |
| 1216 | return Err(InvalidChunkSize); |
| 1217 | } |
| 1218 | count += 1; |
| 1219 | size *= RADIX; |
| 1220 | size += (b - b'0' ) as u64; |
| 1221 | }, |
| 1222 | b'a' ..= b'f' if in_chunk_size => { |
| 1223 | if count > 15 { |
| 1224 | return Err(InvalidChunkSize); |
| 1225 | } |
| 1226 | count += 1; |
| 1227 | size *= RADIX; |
| 1228 | size += (b + 10 - b'a' ) as u64; |
| 1229 | } |
| 1230 | b'A' ..= b'F' if in_chunk_size => { |
| 1231 | if count > 15 { |
| 1232 | return Err(InvalidChunkSize); |
| 1233 | } |
| 1234 | count += 1; |
| 1235 | size *= RADIX; |
| 1236 | size += (b + 10 - b'A' ) as u64; |
| 1237 | } |
| 1238 | b' \r' => { |
| 1239 | match next!(bytes) { |
| 1240 | b' \n' => break, |
| 1241 | _ => return Err(InvalidChunkSize), |
| 1242 | } |
| 1243 | } |
| 1244 | // If we weren't in the extension yet, the ";" signals its start |
| 1245 | b';' if !in_ext => { |
| 1246 | in_ext = true; |
| 1247 | in_chunk_size = false; |
| 1248 | } |
| 1249 | // "Linear white space" is ignored between the chunk size and the |
| 1250 | // extension separator token (";") due to the "implied *LWS rule". |
| 1251 | b' \t' | b' ' if !in_ext && !in_chunk_size => {} |
| 1252 | // LWS can follow the chunk size, but no more digits can come |
| 1253 | b' \t' | b' ' if in_chunk_size => in_chunk_size = false, |
| 1254 | // We allow any arbitrary octet once we are in the extension, since |
| 1255 | // they all get ignored anyway. According to the HTTP spec, valid |
| 1256 | // extensions would have a more strict syntax: |
| 1257 | // (token ["=" (token | quoted-string)]) |
| 1258 | // but we gain nothing by rejecting an otherwise valid chunk size. |
| 1259 | _ if in_ext => {} |
| 1260 | // Finally, if we aren't in the extension and we're reading any |
| 1261 | // other octet, the chunk size line is invalid! |
| 1262 | _ => return Err(InvalidChunkSize), |
| 1263 | } |
| 1264 | } |
| 1265 | Ok(Status::Complete((bytes.pos(), size))) |
| 1266 | } |
| 1267 | |
| 1268 | #[cfg (test)] |
| 1269 | mod tests { |
| 1270 | use super::{Request, Response, Status, EMPTY_HEADER, parse_chunk_size}; |
| 1271 | |
| 1272 | const NUM_OF_HEADERS: usize = 4; |
| 1273 | |
| 1274 | macro_rules! req { |
| 1275 | ($name:ident, $buf:expr, |$arg:ident| $body:expr) => ( |
| 1276 | req! {$name, $buf, Ok(Status::Complete($buf.len())), |$arg| $body } |
| 1277 | ); |
| 1278 | ($name:ident, $buf:expr, $len:expr, |$arg:ident| $body:expr) => ( |
| 1279 | #[test] |
| 1280 | fn $name() { |
| 1281 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1282 | let mut req = Request::new(&mut headers[..]); |
| 1283 | let status = req.parse($buf.as_ref()); |
| 1284 | assert_eq!(status, $len); |
| 1285 | closure(req); |
| 1286 | |
| 1287 | fn closure($arg: Request) { |
| 1288 | $body |
| 1289 | } |
| 1290 | } |
| 1291 | ) |
| 1292 | } |
| 1293 | |
| 1294 | req! { |
| 1295 | test_request_simple, |
| 1296 | b"GET / HTTP/1.1 \r\n\r\n" , |
| 1297 | |req| { |
| 1298 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1299 | assert_eq!(req.path.unwrap(), "/" ); |
| 1300 | assert_eq!(req.version.unwrap(), 1); |
| 1301 | assert_eq!(req.headers.len(), 0); |
| 1302 | } |
| 1303 | } |
| 1304 | |
| 1305 | req! { |
| 1306 | test_request_simple_with_query_params, |
| 1307 | b"GET /thing?data=a HTTP/1.1 \r\n\r\n" , |
| 1308 | |req| { |
| 1309 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1310 | assert_eq!(req.path.unwrap(), "/thing?data=a" ); |
| 1311 | assert_eq!(req.version.unwrap(), 1); |
| 1312 | assert_eq!(req.headers.len(), 0); |
| 1313 | } |
| 1314 | } |
| 1315 | |
| 1316 | req! { |
| 1317 | test_request_simple_with_whatwg_query_params, |
| 1318 | b"GET /thing?data=a^ HTTP/1.1 \r\n\r\n" , |
| 1319 | |req| { |
| 1320 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1321 | assert_eq!(req.path.unwrap(), "/thing?data=a^" ); |
| 1322 | assert_eq!(req.version.unwrap(), 1); |
| 1323 | assert_eq!(req.headers.len(), 0); |
| 1324 | } |
| 1325 | } |
| 1326 | |
| 1327 | req! { |
| 1328 | test_request_headers, |
| 1329 | b"GET / HTTP/1.1 \r\nHost: foo.com \r\nCookie: \r\n\r\n" , |
| 1330 | |req| { |
| 1331 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1332 | assert_eq!(req.path.unwrap(), "/" ); |
| 1333 | assert_eq!(req.version.unwrap(), 1); |
| 1334 | assert_eq!(req.headers.len(), 2); |
| 1335 | assert_eq!(req.headers[0].name, "Host" ); |
| 1336 | assert_eq!(req.headers[0].value, b"foo.com" ); |
| 1337 | assert_eq!(req.headers[1].name, "Cookie" ); |
| 1338 | assert_eq!(req.headers[1].value, b"" ); |
| 1339 | } |
| 1340 | } |
| 1341 | |
| 1342 | req! { |
| 1343 | test_request_headers_optional_whitespace, |
| 1344 | b"GET / HTTP/1.1 \r\nHost: \tfoo.com \t \r\nCookie: \t \r\n\r\n" , |
| 1345 | |req| { |
| 1346 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1347 | assert_eq!(req.path.unwrap(), "/" ); |
| 1348 | assert_eq!(req.version.unwrap(), 1); |
| 1349 | assert_eq!(req.headers.len(), 2); |
| 1350 | assert_eq!(req.headers[0].name, "Host" ); |
| 1351 | assert_eq!(req.headers[0].value, b"foo.com" ); |
| 1352 | assert_eq!(req.headers[1].name, "Cookie" ); |
| 1353 | assert_eq!(req.headers[1].value, b"" ); |
| 1354 | } |
| 1355 | } |
| 1356 | |
| 1357 | req! { |
| 1358 | // test the scalar parsing |
| 1359 | test_request_header_value_htab_short, |
| 1360 | b"GET / HTTP/1.1 \r\nUser-Agent: some \tagent \r\n\r\n" , |
| 1361 | |req| { |
| 1362 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1363 | assert_eq!(req.path.unwrap(), "/" ); |
| 1364 | assert_eq!(req.version.unwrap(), 1); |
| 1365 | assert_eq!(req.headers.len(), 1); |
| 1366 | assert_eq!(req.headers[0].name, "User-Agent" ); |
| 1367 | assert_eq!(req.headers[0].value, b"some \tagent" ); |
| 1368 | } |
| 1369 | } |
| 1370 | |
| 1371 | req! { |
| 1372 | // test the sse42 parsing |
| 1373 | test_request_header_value_htab_med, |
| 1374 | b"GET / HTTP/1.1 \r\nUser-Agent: 1234567890some \tagent \r\n\r\n" , |
| 1375 | |req| { |
| 1376 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1377 | assert_eq!(req.path.unwrap(), "/" ); |
| 1378 | assert_eq!(req.version.unwrap(), 1); |
| 1379 | assert_eq!(req.headers.len(), 1); |
| 1380 | assert_eq!(req.headers[0].name, "User-Agent" ); |
| 1381 | assert_eq!(req.headers[0].value, b"1234567890some \tagent" ); |
| 1382 | } |
| 1383 | } |
| 1384 | |
| 1385 | req! { |
| 1386 | // test the avx2 parsing |
| 1387 | test_request_header_value_htab_long, |
| 1388 | b"GET / HTTP/1.1 \r\nUser-Agent: 1234567890some \t1234567890agent1234567890 \r\n\r\n" , |
| 1389 | |req| { |
| 1390 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1391 | assert_eq!(req.path.unwrap(), "/" ); |
| 1392 | assert_eq!(req.version.unwrap(), 1); |
| 1393 | assert_eq!(req.headers.len(), 1); |
| 1394 | assert_eq!(req.headers[0].name, "User-Agent" ); |
| 1395 | assert_eq!(req.headers[0].value, &b"1234567890some \t1234567890agent1234567890" [..]); |
| 1396 | } |
| 1397 | } |
| 1398 | |
| 1399 | req! { |
| 1400 | test_request_headers_max, |
| 1401 | b"GET / HTTP/1.1 \r\nA: A \r\nB: B \r\nC: C \r\nD: D \r\n\r\n" , |
| 1402 | |req| { |
| 1403 | assert_eq!(req.headers.len(), NUM_OF_HEADERS); |
| 1404 | } |
| 1405 | } |
| 1406 | |
| 1407 | req! { |
| 1408 | test_request_multibyte, |
| 1409 | b"GET / HTTP/1.1 \r\nHost: foo.com \r\nUser-Agent: \xe3\x81\xb2\xe3/1.0 \r\n\r\n" , |
| 1410 | |req| { |
| 1411 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1412 | assert_eq!(req.path.unwrap(), "/" ); |
| 1413 | assert_eq!(req.version.unwrap(), 1); |
| 1414 | assert_eq!(req.headers.len(), 2); |
| 1415 | assert_eq!(req.headers[0].name, "Host" ); |
| 1416 | assert_eq!(req.headers[0].value, b"foo.com" ); |
| 1417 | assert_eq!(req.headers[1].name, "User-Agent" ); |
| 1418 | assert_eq!(req.headers[1].value, b" \xe3\x81\xb2\xe3/1.0" ); |
| 1419 | } |
| 1420 | } |
| 1421 | |
| 1422 | |
| 1423 | req! { |
| 1424 | test_request_partial, |
| 1425 | b"GET / HTTP/1.1 \r\n\r" , Ok(Status::Partial), |
| 1426 | |_req| {} |
| 1427 | } |
| 1428 | |
| 1429 | req! { |
| 1430 | test_request_partial_version, |
| 1431 | b"GET / HTTP/1." , Ok(Status::Partial), |
| 1432 | |_req| {} |
| 1433 | } |
| 1434 | |
| 1435 | req! { |
| 1436 | test_request_partial_parses_headers_as_much_as_it_can, |
| 1437 | b"GET / HTTP/1.1 \r\nHost: yolo \r\n" , |
| 1438 | Ok(crate::Status::Partial), |
| 1439 | |req| { |
| 1440 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1441 | assert_eq!(req.path.unwrap(), "/" ); |
| 1442 | assert_eq!(req.version.unwrap(), 1); |
| 1443 | assert_eq!(req.headers.len(), NUM_OF_HEADERS); // doesn't slice since not Complete |
| 1444 | assert_eq!(req.headers[0].name, "Host" ); |
| 1445 | assert_eq!(req.headers[0].value, b"yolo" ); |
| 1446 | } |
| 1447 | } |
| 1448 | |
| 1449 | req! { |
| 1450 | test_request_newlines, |
| 1451 | b"GET / HTTP/1.1 \nHost: foo.bar \n\n" , |
| 1452 | |_r| {} |
| 1453 | } |
| 1454 | |
| 1455 | req! { |
| 1456 | test_request_empty_lines_prefix, |
| 1457 | b" \r\n\r\nGET / HTTP/1.1 \r\n\r\n" , |
| 1458 | |req| { |
| 1459 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1460 | assert_eq!(req.path.unwrap(), "/" ); |
| 1461 | assert_eq!(req.version.unwrap(), 1); |
| 1462 | assert_eq!(req.headers.len(), 0); |
| 1463 | } |
| 1464 | } |
| 1465 | |
| 1466 | req! { |
| 1467 | test_request_empty_lines_prefix_lf_only, |
| 1468 | b" \n\nGET / HTTP/1.1 \n\n" , |
| 1469 | |req| { |
| 1470 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1471 | assert_eq!(req.path.unwrap(), "/" ); |
| 1472 | assert_eq!(req.version.unwrap(), 1); |
| 1473 | assert_eq!(req.headers.len(), 0); |
| 1474 | } |
| 1475 | } |
| 1476 | |
| 1477 | req! { |
| 1478 | test_request_path_backslash, |
| 1479 | b" \n\nGET / \\?wayne \\=5 HTTP/1.1 \n\n" , |
| 1480 | |req| { |
| 1481 | assert_eq!(req.method.unwrap(), "GET" ); |
| 1482 | assert_eq!(req.path.unwrap(), "/ \\?wayne \\=5" ); |
| 1483 | assert_eq!(req.version.unwrap(), 1); |
| 1484 | assert_eq!(req.headers.len(), 0); |
| 1485 | } |
| 1486 | } |
| 1487 | |
| 1488 | req! { |
| 1489 | test_request_with_invalid_token_delimiter, |
| 1490 | b"GET \n/ HTTP/1.1 \r\nHost: foo.bar \r\n\r\n" , |
| 1491 | Err(crate::Error::Token), |
| 1492 | |_r| {} |
| 1493 | } |
| 1494 | |
| 1495 | |
| 1496 | req! { |
| 1497 | test_request_with_invalid_but_short_version, |
| 1498 | b"GET / HTTP/1!" , |
| 1499 | Err(crate::Error::Version), |
| 1500 | |_r| {} |
| 1501 | } |
| 1502 | |
| 1503 | req! { |
| 1504 | test_request_with_empty_method, |
| 1505 | b" / HTTP/1.1 \r\n\r\n" , |
| 1506 | Err(crate::Error::Token), |
| 1507 | |_r| {} |
| 1508 | } |
| 1509 | |
| 1510 | req! { |
| 1511 | test_request_with_empty_path, |
| 1512 | b"GET HTTP/1.1 \r\n\r\n" , |
| 1513 | Err(crate::Error::Token), |
| 1514 | |_r| {} |
| 1515 | } |
| 1516 | |
| 1517 | req! { |
| 1518 | test_request_with_empty_method_and_path, |
| 1519 | b" HTTP/1.1 \r\n\r\n" , |
| 1520 | Err(crate::Error::Token), |
| 1521 | |_r| {} |
| 1522 | } |
| 1523 | |
| 1524 | macro_rules! res { |
| 1525 | ($name:ident, $buf:expr, |$arg:ident| $body:expr) => ( |
| 1526 | res! {$name, $buf, Ok(Status::Complete($buf.len())), |$arg| $body } |
| 1527 | ); |
| 1528 | ($name:ident, $buf:expr, $len:expr, |$arg:ident| $body:expr) => ( |
| 1529 | #[test] |
| 1530 | fn $name() { |
| 1531 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1532 | let mut res = Response::new(&mut headers[..]); |
| 1533 | let status = res.parse($buf.as_ref()); |
| 1534 | assert_eq!(status, $len); |
| 1535 | closure(res); |
| 1536 | |
| 1537 | fn closure($arg: Response) { |
| 1538 | $body |
| 1539 | } |
| 1540 | } |
| 1541 | ) |
| 1542 | } |
| 1543 | |
| 1544 | res! { |
| 1545 | test_response_simple, |
| 1546 | b"HTTP/1.1 200 OK \r\n\r\n" , |
| 1547 | |res| { |
| 1548 | assert_eq!(res.version.unwrap(), 1); |
| 1549 | assert_eq!(res.code.unwrap(), 200); |
| 1550 | assert_eq!(res.reason.unwrap(), "OK" ); |
| 1551 | } |
| 1552 | } |
| 1553 | |
| 1554 | res! { |
| 1555 | test_response_newlines, |
| 1556 | b"HTTP/1.0 403 Forbidden \nServer: foo.bar \n\n" , |
| 1557 | |_r| {} |
| 1558 | } |
| 1559 | |
| 1560 | res! { |
| 1561 | test_response_reason_missing, |
| 1562 | b"HTTP/1.1 200 \r\n\r\n" , |
| 1563 | |res| { |
| 1564 | assert_eq!(res.version.unwrap(), 1); |
| 1565 | assert_eq!(res.code.unwrap(), 200); |
| 1566 | assert_eq!(res.reason.unwrap(), "" ); |
| 1567 | } |
| 1568 | } |
| 1569 | |
| 1570 | res! { |
| 1571 | test_response_reason_missing_no_space, |
| 1572 | b"HTTP/1.1 200 \r\n\r\n" , |
| 1573 | |res| { |
| 1574 | assert_eq!(res.version.unwrap(), 1); |
| 1575 | assert_eq!(res.code.unwrap(), 200); |
| 1576 | assert_eq!(res.reason.unwrap(), "" ); |
| 1577 | } |
| 1578 | } |
| 1579 | |
| 1580 | res! { |
| 1581 | test_response_reason_missing_no_space_with_headers, |
| 1582 | b"HTTP/1.1 200 \r\nFoo: bar \r\n\r\n" , |
| 1583 | |res| { |
| 1584 | assert_eq!(res.version.unwrap(), 1); |
| 1585 | assert_eq!(res.code.unwrap(), 200); |
| 1586 | assert_eq!(res.reason.unwrap(), "" ); |
| 1587 | assert_eq!(res.headers.len(), 1); |
| 1588 | assert_eq!(res.headers[0].name, "Foo" ); |
| 1589 | assert_eq!(res.headers[0].value, b"bar" ); |
| 1590 | } |
| 1591 | } |
| 1592 | |
| 1593 | res! { |
| 1594 | test_response_reason_with_space_and_tab, |
| 1595 | b"HTTP/1.1 101 Switching Protocols \t\r\n\r\n" , |
| 1596 | |res| { |
| 1597 | assert_eq!(res.version.unwrap(), 1); |
| 1598 | assert_eq!(res.code.unwrap(), 101); |
| 1599 | assert_eq!(res.reason.unwrap(), "Switching Protocols \t" ); |
| 1600 | } |
| 1601 | } |
| 1602 | |
| 1603 | static RESPONSE_REASON_WITH_OBS_TEXT_BYTE: &[u8] = b"HTTP/1.1 200 X \xFFZ \r\n\r\n" ; |
| 1604 | res! { |
| 1605 | test_response_reason_with_obsolete_text_byte, |
| 1606 | RESPONSE_REASON_WITH_OBS_TEXT_BYTE, |
| 1607 | |res| { |
| 1608 | assert_eq!(res.version.unwrap(), 1); |
| 1609 | assert_eq!(res.code.unwrap(), 200); |
| 1610 | // Empty string fallback in case of obs-text |
| 1611 | assert_eq!(res.reason.unwrap(), "" ); |
| 1612 | } |
| 1613 | } |
| 1614 | |
| 1615 | res! { |
| 1616 | test_response_reason_with_nul_byte, |
| 1617 | b"HTTP/1.1 200 \x00\r\n\r\n" , |
| 1618 | Err(crate::Error::Status), |
| 1619 | |_res| {} |
| 1620 | } |
| 1621 | |
| 1622 | res! { |
| 1623 | test_response_version_missing_space, |
| 1624 | b"HTTP/1.1" , |
| 1625 | Ok(Status::Partial), |
| 1626 | |_res| {} |
| 1627 | } |
| 1628 | |
| 1629 | res! { |
| 1630 | test_response_code_missing_space, |
| 1631 | b"HTTP/1.1 200" , |
| 1632 | Ok(Status::Partial), |
| 1633 | |_res| {} |
| 1634 | } |
| 1635 | |
| 1636 | res! { |
| 1637 | test_response_partial_parses_headers_as_much_as_it_can, |
| 1638 | b"HTTP/1.1 200 OK \r\nServer: yolo \r\n" , |
| 1639 | Ok(crate::Status::Partial), |
| 1640 | |res| { |
| 1641 | assert_eq!(res.version.unwrap(), 1); |
| 1642 | assert_eq!(res.code.unwrap(), 200); |
| 1643 | assert_eq!(res.reason.unwrap(), "OK" ); |
| 1644 | assert_eq!(res.headers.len(), NUM_OF_HEADERS); // doesn't slice since not Complete |
| 1645 | assert_eq!(res.headers[0].name, "Server" ); |
| 1646 | assert_eq!(res.headers[0].value, b"yolo" ); |
| 1647 | } |
| 1648 | } |
| 1649 | |
| 1650 | res! { |
| 1651 | test_response_empty_lines_prefix_lf_only, |
| 1652 | b" \n\nHTTP/1.1 200 OK \n\n" , |
| 1653 | |_res| {} |
| 1654 | } |
| 1655 | |
| 1656 | res! { |
| 1657 | test_response_no_cr, |
| 1658 | b"HTTP/1.0 200 \nContent-type: text/html \n\n" , |
| 1659 | |res| { |
| 1660 | assert_eq!(res.version.unwrap(), 0); |
| 1661 | assert_eq!(res.code.unwrap(), 200); |
| 1662 | assert_eq!(res.reason.unwrap(), "" ); |
| 1663 | assert_eq!(res.headers.len(), 1); |
| 1664 | assert_eq!(res.headers[0].name, "Content-type" ); |
| 1665 | assert_eq!(res.headers[0].value, b"text/html" ); |
| 1666 | } |
| 1667 | } |
| 1668 | |
| 1669 | static RESPONSE_WITH_WHITESPACE_BETWEEN_HEADER_NAME_AND_COLON: &[u8] = |
| 1670 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials : true \r\nBread: baguette \r\n\r\n" ; |
| 1671 | |
| 1672 | #[test] |
| 1673 | fn test_forbid_response_with_whitespace_between_header_name_and_colon() { |
| 1674 | let mut headers = [EMPTY_HEADER; 2]; |
| 1675 | let mut response = Response::new(&mut headers[..]); |
| 1676 | let result = response.parse(RESPONSE_WITH_WHITESPACE_BETWEEN_HEADER_NAME_AND_COLON); |
| 1677 | |
| 1678 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 1679 | } |
| 1680 | |
| 1681 | #[test] |
| 1682 | fn test_allow_response_with_whitespace_between_header_name_and_colon() { |
| 1683 | let mut headers = [EMPTY_HEADER; 2]; |
| 1684 | let mut response = Response::new(&mut headers[..]); |
| 1685 | let result = crate::ParserConfig::default() |
| 1686 | .allow_spaces_after_header_name_in_responses(true) |
| 1687 | .parse_response(&mut response, RESPONSE_WITH_WHITESPACE_BETWEEN_HEADER_NAME_AND_COLON); |
| 1688 | |
| 1689 | assert_eq!(result, Ok(Status::Complete(77))); |
| 1690 | assert_eq!(response.version.unwrap(), 1); |
| 1691 | assert_eq!(response.code.unwrap(), 200); |
| 1692 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1693 | assert_eq!(response.headers.len(), 2); |
| 1694 | assert_eq!(response.headers[0].name, "Access-Control-Allow-Credentials" ); |
| 1695 | assert_eq!(response.headers[0].value, &b"true" [..]); |
| 1696 | assert_eq!(response.headers[1].name, "Bread" ); |
| 1697 | assert_eq!(response.headers[1].value, &b"baguette" [..]); |
| 1698 | } |
| 1699 | |
| 1700 | #[test] |
| 1701 | fn test_ignore_header_line_with_whitespaces_after_header_name() { |
| 1702 | let mut headers = [EMPTY_HEADER; 2]; |
| 1703 | let mut response = Response::new(&mut headers[..]); |
| 1704 | let result = crate::ParserConfig::default() |
| 1705 | .ignore_invalid_headers_in_responses(true) |
| 1706 | .parse_response(&mut response, RESPONSE_WITH_WHITESPACE_BETWEEN_HEADER_NAME_AND_COLON); |
| 1707 | |
| 1708 | assert_eq!(result, Ok(Status::Complete(77))); |
| 1709 | assert_eq!(response.version.unwrap(), 1); |
| 1710 | assert_eq!(response.code.unwrap(), 200); |
| 1711 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1712 | assert_eq!(response.headers.len(), 1); |
| 1713 | assert_eq!(response.headers[0].name, "Bread" ); |
| 1714 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 1715 | } |
| 1716 | |
| 1717 | static REQUEST_WITH_WHITESPACE_BETWEEN_HEADER_NAME_AND_COLON: &[u8] = |
| 1718 | b"GET / HTTP/1.1 \r\nHost : localhost \r\n\r\n" ; |
| 1719 | |
| 1720 | #[test] |
| 1721 | fn test_forbid_request_with_whitespace_between_header_name_and_colon() { |
| 1722 | let mut headers = [EMPTY_HEADER; 1]; |
| 1723 | let mut request = Request::new(&mut headers[..]); |
| 1724 | let result = request.parse(REQUEST_WITH_WHITESPACE_BETWEEN_HEADER_NAME_AND_COLON); |
| 1725 | |
| 1726 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 1727 | } |
| 1728 | |
| 1729 | static RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_START: &[u8] = |
| 1730 | b"HTTP/1.1 200 OK \r\nLine-Folded-Header: \r\n \r\n hello there \r\n\r\n" ; |
| 1731 | |
| 1732 | #[test] |
| 1733 | fn test_forbid_response_with_obsolete_line_folding_at_start() { |
| 1734 | let mut headers = [EMPTY_HEADER; 1]; |
| 1735 | let mut response = Response::new(&mut headers[..]); |
| 1736 | let result = response.parse(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_START); |
| 1737 | |
| 1738 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 1739 | } |
| 1740 | |
| 1741 | #[test] |
| 1742 | fn test_allow_response_with_obsolete_line_folding_at_start() { |
| 1743 | let mut headers = [EMPTY_HEADER; 1]; |
| 1744 | let mut response = Response::new(&mut headers[..]); |
| 1745 | let result = crate::ParserConfig::default() |
| 1746 | .allow_obsolete_multiline_headers_in_responses(true) |
| 1747 | .parse_response(&mut response, RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_START); |
| 1748 | |
| 1749 | assert_eq!(result, Ok(Status::Complete(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_START.len()))); |
| 1750 | assert_eq!(response.version.unwrap(), 1); |
| 1751 | assert_eq!(response.code.unwrap(), 200); |
| 1752 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1753 | assert_eq!(response.headers.len(), 1); |
| 1754 | assert_eq!(response.headers[0].name, "Line-Folded-Header" ); |
| 1755 | assert_eq!(response.headers[0].value, &b"hello there" [..]); |
| 1756 | } |
| 1757 | |
| 1758 | static RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_END: &[u8] = |
| 1759 | b"HTTP/1.1 200 OK \r\nLine-Folded-Header: hello there \r\n \r\n \r\n\r\n" ; |
| 1760 | |
| 1761 | #[test] |
| 1762 | fn test_forbid_response_with_obsolete_line_folding_at_end() { |
| 1763 | let mut headers = [EMPTY_HEADER; 1]; |
| 1764 | let mut response = Response::new(&mut headers[..]); |
| 1765 | let result = response.parse(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_END); |
| 1766 | |
| 1767 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 1768 | } |
| 1769 | |
| 1770 | #[test] |
| 1771 | fn test_allow_response_with_obsolete_line_folding_at_end() { |
| 1772 | let mut headers = [EMPTY_HEADER; 1]; |
| 1773 | let mut response = Response::new(&mut headers[..]); |
| 1774 | let result = crate::ParserConfig::default() |
| 1775 | .allow_obsolete_multiline_headers_in_responses(true) |
| 1776 | .parse_response(&mut response, RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_END); |
| 1777 | |
| 1778 | assert_eq!(result, Ok(Status::Complete(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_AT_END.len()))); |
| 1779 | assert_eq!(response.version.unwrap(), 1); |
| 1780 | assert_eq!(response.code.unwrap(), 200); |
| 1781 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1782 | assert_eq!(response.headers.len(), 1); |
| 1783 | assert_eq!(response.headers[0].name, "Line-Folded-Header" ); |
| 1784 | assert_eq!(response.headers[0].value, &b"hello there" [..]); |
| 1785 | } |
| 1786 | |
| 1787 | static RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_MIDDLE: &[u8] = |
| 1788 | b"HTTP/1.1 200 OK \r\nLine-Folded-Header: hello \r\n \r\n there \r\n\r\n" ; |
| 1789 | |
| 1790 | #[test] |
| 1791 | fn test_forbid_response_with_obsolete_line_folding_in_middle() { |
| 1792 | let mut headers = [EMPTY_HEADER; 1]; |
| 1793 | let mut response = Response::new(&mut headers[..]); |
| 1794 | let result = response.parse(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_MIDDLE); |
| 1795 | |
| 1796 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 1797 | } |
| 1798 | |
| 1799 | #[test] |
| 1800 | fn test_allow_response_with_obsolete_line_folding_in_middle() { |
| 1801 | let mut headers = [EMPTY_HEADER; 1]; |
| 1802 | let mut response = Response::new(&mut headers[..]); |
| 1803 | let result = crate::ParserConfig::default() |
| 1804 | .allow_obsolete_multiline_headers_in_responses(true) |
| 1805 | .parse_response(&mut response, RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_MIDDLE); |
| 1806 | |
| 1807 | assert_eq!(result, Ok(Status::Complete(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_MIDDLE.len()))); |
| 1808 | assert_eq!(response.version.unwrap(), 1); |
| 1809 | assert_eq!(response.code.unwrap(), 200); |
| 1810 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1811 | assert_eq!(response.headers.len(), 1); |
| 1812 | assert_eq!(response.headers[0].name, "Line-Folded-Header" ); |
| 1813 | assert_eq!(response.headers[0].value, &b"hello \r\n \r\n there" [..]); |
| 1814 | } |
| 1815 | |
| 1816 | static RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_EMPTY_HEADER: &[u8] = |
| 1817 | b"HTTP/1.1 200 OK \r\nLine-Folded-Header: \r\n \r\n \r\n\r\n" ; |
| 1818 | |
| 1819 | #[test] |
| 1820 | fn test_forbid_response_with_obsolete_line_folding_in_empty_header() { |
| 1821 | let mut headers = [EMPTY_HEADER; 1]; |
| 1822 | let mut response = Response::new(&mut headers[..]); |
| 1823 | let result = response.parse(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_EMPTY_HEADER); |
| 1824 | |
| 1825 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 1826 | } |
| 1827 | |
| 1828 | #[test] |
| 1829 | fn test_allow_response_with_obsolete_line_folding_in_empty_header() { |
| 1830 | let mut headers = [EMPTY_HEADER; 1]; |
| 1831 | let mut response = Response::new(&mut headers[..]); |
| 1832 | let result = crate::ParserConfig::default() |
| 1833 | .allow_obsolete_multiline_headers_in_responses(true) |
| 1834 | .parse_response(&mut response, RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_EMPTY_HEADER); |
| 1835 | |
| 1836 | assert_eq!(result, Ok(Status::Complete(RESPONSE_WITH_OBSOLETE_LINE_FOLDING_IN_EMPTY_HEADER.len()))); |
| 1837 | assert_eq!(response.version.unwrap(), 1); |
| 1838 | assert_eq!(response.code.unwrap(), 200); |
| 1839 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1840 | assert_eq!(response.headers.len(), 1); |
| 1841 | assert_eq!(response.headers[0].name, "Line-Folded-Header" ); |
| 1842 | assert_eq!(response.headers[0].value, &b"" [..]); |
| 1843 | } |
| 1844 | |
| 1845 | #[test] |
| 1846 | fn test_chunk_size() { |
| 1847 | assert_eq!(parse_chunk_size(b"0 \r\n" ), Ok(Status::Complete((3, 0)))); |
| 1848 | assert_eq!(parse_chunk_size(b"12 \r\nchunk" ), Ok(Status::Complete((4, 18)))); |
| 1849 | assert_eq!(parse_chunk_size(b"3086d \r\n" ), Ok(Status::Complete((7, 198765)))); |
| 1850 | assert_eq!(parse_chunk_size(b"3735AB1;foo bar* \r\n" ), Ok(Status::Complete((18, 57891505)))); |
| 1851 | assert_eq!(parse_chunk_size(b"3735ab1 ; baz \r\n" ), Ok(Status::Complete((16, 57891505)))); |
| 1852 | assert_eq!(parse_chunk_size(b"77a65 \r" ), Ok(Status::Partial)); |
| 1853 | assert_eq!(parse_chunk_size(b"ab" ), Ok(Status::Partial)); |
| 1854 | assert_eq!(parse_chunk_size(b"567f8a \rfoo" ), Err(crate::InvalidChunkSize)); |
| 1855 | assert_eq!(parse_chunk_size(b"567f8a \rfoo" ), Err(crate::InvalidChunkSize)); |
| 1856 | assert_eq!(parse_chunk_size(b"567xf8a \r\n" ), Err(crate::InvalidChunkSize)); |
| 1857 | assert_eq!(parse_chunk_size(b"ffffffffffffffff \r\n" ), Ok(Status::Complete((18, std::u64::MAX)))); |
| 1858 | assert_eq!(parse_chunk_size(b"1ffffffffffffffff \r\n" ), Err(crate::InvalidChunkSize)); |
| 1859 | assert_eq!(parse_chunk_size(b"Affffffffffffffff \r\n" ), Err(crate::InvalidChunkSize)); |
| 1860 | assert_eq!(parse_chunk_size(b"fffffffffffffffff \r\n" ), Err(crate::InvalidChunkSize)); |
| 1861 | } |
| 1862 | |
| 1863 | static RESPONSE_WITH_MULTIPLE_SPACE_DELIMITERS: &[u8] = |
| 1864 | b"HTTP/1.1 200 OK \r\n\r\n" ; |
| 1865 | |
| 1866 | #[test] |
| 1867 | fn test_forbid_response_with_multiple_space_delimiters() { |
| 1868 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1869 | let mut response = Response::new(&mut headers[..]); |
| 1870 | let result = response.parse(RESPONSE_WITH_MULTIPLE_SPACE_DELIMITERS); |
| 1871 | |
| 1872 | assert_eq!(result, Err(crate::Error::Status)); |
| 1873 | } |
| 1874 | |
| 1875 | #[test] |
| 1876 | fn test_allow_response_with_multiple_space_delimiters() { |
| 1877 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1878 | let mut response = Response::new(&mut headers[..]); |
| 1879 | let result = crate::ParserConfig::default() |
| 1880 | .allow_multiple_spaces_in_response_status_delimiters(true) |
| 1881 | .parse_response(&mut response, RESPONSE_WITH_MULTIPLE_SPACE_DELIMITERS); |
| 1882 | |
| 1883 | assert_eq!(result, Ok(Status::Complete(RESPONSE_WITH_MULTIPLE_SPACE_DELIMITERS.len()))); |
| 1884 | assert_eq!(response.version.unwrap(), 1); |
| 1885 | assert_eq!(response.code.unwrap(), 200); |
| 1886 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 1887 | assert_eq!(response.headers.len(), 0); |
| 1888 | } |
| 1889 | |
| 1890 | /// This is technically allowed by the spec, but we only support multiple spaces as an option, |
| 1891 | /// not stray `\r`s. |
| 1892 | static RESPONSE_WITH_WEIRD_WHITESPACE_DELIMITERS: &[u8] = |
| 1893 | b"HTTP/1.1 200 \rOK \r\n\r\n" ; |
| 1894 | |
| 1895 | #[test] |
| 1896 | fn test_forbid_response_with_weird_whitespace_delimiters() { |
| 1897 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1898 | let mut response = Response::new(&mut headers[..]); |
| 1899 | let result = response.parse(RESPONSE_WITH_WEIRD_WHITESPACE_DELIMITERS); |
| 1900 | |
| 1901 | assert_eq!(result, Err(crate::Error::Status)); |
| 1902 | } |
| 1903 | |
| 1904 | #[test] |
| 1905 | fn test_still_forbid_response_with_weird_whitespace_delimiters() { |
| 1906 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1907 | let mut response = Response::new(&mut headers[..]); |
| 1908 | let result = crate::ParserConfig::default() |
| 1909 | .allow_multiple_spaces_in_response_status_delimiters(true) |
| 1910 | .parse_response(&mut response, RESPONSE_WITH_WEIRD_WHITESPACE_DELIMITERS); |
| 1911 | assert_eq!(result, Err(crate::Error::Status)); |
| 1912 | } |
| 1913 | |
| 1914 | static REQUEST_WITH_MULTIPLE_SPACE_DELIMITERS: &[u8] = |
| 1915 | b"GET / HTTP/1.1 \r\n\r\n" ; |
| 1916 | |
| 1917 | #[test] |
| 1918 | fn test_forbid_request_with_multiple_space_delimiters() { |
| 1919 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1920 | let mut request = Request::new(&mut headers[..]); |
| 1921 | let result = request.parse(REQUEST_WITH_MULTIPLE_SPACE_DELIMITERS); |
| 1922 | |
| 1923 | assert_eq!(result, Err(crate::Error::Token)); |
| 1924 | } |
| 1925 | |
| 1926 | #[test] |
| 1927 | fn test_allow_request_with_multiple_space_delimiters() { |
| 1928 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1929 | let mut request = Request::new(&mut headers[..]); |
| 1930 | let result = crate::ParserConfig::default() |
| 1931 | .allow_multiple_spaces_in_request_line_delimiters(true) |
| 1932 | .parse_request(&mut request, REQUEST_WITH_MULTIPLE_SPACE_DELIMITERS); |
| 1933 | |
| 1934 | assert_eq!(result, Ok(Status::Complete(REQUEST_WITH_MULTIPLE_SPACE_DELIMITERS.len()))); |
| 1935 | assert_eq!(request.method.unwrap(), "GET" ); |
| 1936 | assert_eq!(request.path.unwrap(), "/" ); |
| 1937 | assert_eq!(request.version.unwrap(), 1); |
| 1938 | assert_eq!(request.headers.len(), 0); |
| 1939 | } |
| 1940 | |
| 1941 | /// This is technically allowed by the spec, but we only support multiple spaces as an option, |
| 1942 | /// not stray `\r`s. |
| 1943 | static REQUEST_WITH_WEIRD_WHITESPACE_DELIMITERS: &[u8] = |
| 1944 | b"GET \r/ \rHTTP/1.1 \r\n\r\n" ; |
| 1945 | |
| 1946 | #[test] |
| 1947 | fn test_forbid_request_with_weird_whitespace_delimiters() { |
| 1948 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1949 | let mut request = Request::new(&mut headers[..]); |
| 1950 | let result = request.parse(REQUEST_WITH_WEIRD_WHITESPACE_DELIMITERS); |
| 1951 | |
| 1952 | assert_eq!(result, Err(crate::Error::Token)); |
| 1953 | } |
| 1954 | |
| 1955 | #[test] |
| 1956 | fn test_still_forbid_request_with_weird_whitespace_delimiters() { |
| 1957 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1958 | let mut request = Request::new(&mut headers[..]); |
| 1959 | let result = crate::ParserConfig::default() |
| 1960 | .allow_multiple_spaces_in_request_line_delimiters(true) |
| 1961 | .parse_request(&mut request, REQUEST_WITH_WEIRD_WHITESPACE_DELIMITERS); |
| 1962 | assert_eq!(result, Err(crate::Error::Token)); |
| 1963 | } |
| 1964 | |
| 1965 | static REQUEST_WITH_MULTIPLE_SPACES_AND_BAD_PATH: &[u8] = b"GET /foo>ohno HTTP/1.1 \r\n\r\n" ; |
| 1966 | |
| 1967 | #[test] |
| 1968 | fn test_request_with_multiple_spaces_and_bad_path() { |
| 1969 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1970 | let mut request = Request::new(&mut headers[..]); |
| 1971 | let result = crate::ParserConfig::default() |
| 1972 | .allow_multiple_spaces_in_request_line_delimiters(true) |
| 1973 | .parse_request(&mut request, REQUEST_WITH_MULTIPLE_SPACES_AND_BAD_PATH); |
| 1974 | assert_eq!(result, Err(crate::Error::Token)); |
| 1975 | } |
| 1976 | |
| 1977 | static RESPONSE_WITH_SPACES_IN_CODE: &[u8] = b"HTTP/1.1 99 200 OK \r\n\r\n" ; |
| 1978 | |
| 1979 | #[test] |
| 1980 | fn test_response_with_spaces_in_code() { |
| 1981 | let mut headers = [EMPTY_HEADER; NUM_OF_HEADERS]; |
| 1982 | let mut response = Response::new(&mut headers[..]); |
| 1983 | let result = crate::ParserConfig::default() |
| 1984 | .allow_multiple_spaces_in_response_status_delimiters(true) |
| 1985 | .parse_response(&mut response, RESPONSE_WITH_SPACES_IN_CODE); |
| 1986 | assert_eq!(result, Err(crate::Error::Status)); |
| 1987 | } |
| 1988 | |
| 1989 | #[test] |
| 1990 | fn test_response_with_empty_header_name() { |
| 1991 | const RESPONSE: &[u8] = |
| 1992 | b"HTTP/1.1 200 OK \r\n: hello \r\nBread: baguette \r\n\r\n" ; |
| 1993 | |
| 1994 | let mut headers = [EMPTY_HEADER; 2]; |
| 1995 | let mut response = Response::new(&mut headers[..]); |
| 1996 | |
| 1997 | let result = crate::ParserConfig::default() |
| 1998 | .allow_spaces_after_header_name_in_responses(true) |
| 1999 | .parse_response(&mut response, RESPONSE); |
| 2000 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2001 | |
| 2002 | let result = crate::ParserConfig::default() |
| 2003 | .ignore_invalid_headers_in_responses(true) |
| 2004 | .parse_response(&mut response, RESPONSE); |
| 2005 | assert_eq!(result, Ok(Status::Complete(45))); |
| 2006 | |
| 2007 | assert_eq!(response.version.unwrap(), 1); |
| 2008 | assert_eq!(response.code.unwrap(), 200); |
| 2009 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 2010 | assert_eq!(response.headers.len(), 1); |
| 2011 | assert_eq!(response.headers[0].name, "Bread" ); |
| 2012 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 2013 | } |
| 2014 | |
| 2015 | #[test] |
| 2016 | fn test_request_with_whitespace_between_header_name_and_colon() { |
| 2017 | const REQUEST: &[u8] = |
| 2018 | b"GET / HTTP/1.1 \r\nAccess-Control-Allow-Credentials : true \r\nBread: baguette \r\n\r\n" ; |
| 2019 | |
| 2020 | let mut headers = [EMPTY_HEADER; 2]; |
| 2021 | let mut request = Request::new(&mut headers[..]); |
| 2022 | |
| 2023 | let result = crate::ParserConfig::default() |
| 2024 | .allow_spaces_after_header_name_in_responses(true) |
| 2025 | .parse_request(&mut request, REQUEST); |
| 2026 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2027 | |
| 2028 | let result = crate::ParserConfig::default() |
| 2029 | |
| 2030 | .ignore_invalid_headers_in_responses(true) |
| 2031 | .parse_request(&mut request, REQUEST); |
| 2032 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2033 | } |
| 2034 | |
| 2035 | #[test] |
| 2036 | fn test_response_with_invalid_char_between_header_name_and_colon() { |
| 2037 | const RESPONSE: &[u8] = |
| 2038 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials \xFF : true \r\nBread: baguette \r\n\r\n" ; |
| 2039 | |
| 2040 | let mut headers = [EMPTY_HEADER; 2]; |
| 2041 | let mut response = Response::new(&mut headers[..]); |
| 2042 | |
| 2043 | let result = crate::ParserConfig::default() |
| 2044 | .allow_spaces_after_header_name_in_responses(true) |
| 2045 | .parse_response(&mut response, RESPONSE); |
| 2046 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2047 | |
| 2048 | let result = crate::ParserConfig::default() |
| 2049 | .ignore_invalid_headers_in_responses(true) |
| 2050 | .parse_response(&mut response, RESPONSE); |
| 2051 | |
| 2052 | assert_eq!(result, Ok(Status::Complete(79))); |
| 2053 | assert_eq!(response.version.unwrap(), 1); |
| 2054 | assert_eq!(response.code.unwrap(), 200); |
| 2055 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 2056 | assert_eq!(response.headers.len(), 1); |
| 2057 | assert_eq!(response.headers[0].name, "Bread" ); |
| 2058 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 2059 | } |
| 2060 | |
| 2061 | #[test] |
| 2062 | fn test_ignore_header_line_with_missing_colon() { |
| 2063 | const RESPONSE: &[u8] = |
| 2064 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials \r\nBread: baguette \r\n\r\n" ; |
| 2065 | |
| 2066 | let mut headers = [EMPTY_HEADER; 2]; |
| 2067 | let mut response = Response::new(&mut headers[..]); |
| 2068 | |
| 2069 | let result = crate::ParserConfig::default() |
| 2070 | .parse_response(&mut response, RESPONSE); |
| 2071 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2072 | |
| 2073 | let result = crate::ParserConfig::default() |
| 2074 | .ignore_invalid_headers_in_responses(true) |
| 2075 | .parse_response(&mut response, RESPONSE); |
| 2076 | assert_eq!(result, Ok(Status::Complete(70))); |
| 2077 | |
| 2078 | assert_eq!(response.version.unwrap(), 1); |
| 2079 | assert_eq!(response.code.unwrap(), 200); |
| 2080 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 2081 | assert_eq!(response.headers.len(), 1); |
| 2082 | assert_eq!(response.headers[0].name, "Bread" ); |
| 2083 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 2084 | } |
| 2085 | |
| 2086 | #[test] |
| 2087 | fn test_header_with_missing_colon_with_folding() { |
| 2088 | const RESPONSE: &[u8] = |
| 2089 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials \r\n hello \r\nBread: baguette \r\n\r\n" ; |
| 2090 | |
| 2091 | let mut headers = [EMPTY_HEADER; 2]; |
| 2092 | let mut response = Response::new(&mut headers[..]); |
| 2093 | |
| 2094 | let result = crate::ParserConfig::default() |
| 2095 | .allow_obsolete_multiline_headers_in_responses(true) |
| 2096 | .allow_spaces_after_header_name_in_responses(true) |
| 2097 | .parse_response(&mut response, RESPONSE); |
| 2098 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2099 | |
| 2100 | let result = crate::ParserConfig::default() |
| 2101 | .ignore_invalid_headers_in_responses(true) |
| 2102 | .parse_response(&mut response, RESPONSE); |
| 2103 | assert_eq!(result, Ok(Status::Complete(81))); |
| 2104 | |
| 2105 | assert_eq!(response.version.unwrap(), 1); |
| 2106 | assert_eq!(response.code.unwrap(), 200); |
| 2107 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 2108 | assert_eq!(response.headers.len(), 1); |
| 2109 | assert_eq!(response.headers[0].name, "Bread" ); |
| 2110 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 2111 | } |
| 2112 | |
| 2113 | #[test] |
| 2114 | fn test_header_with_nul_in_header_name() { |
| 2115 | const RESPONSE: &[u8] = |
| 2116 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Cred \0entials: hello \r\nBread: baguette \r\n\r\n" ; |
| 2117 | |
| 2118 | let mut headers = [EMPTY_HEADER; 2]; |
| 2119 | let mut response = Response::new(&mut headers[..]); |
| 2120 | |
| 2121 | let result = crate::ParserConfig::default() |
| 2122 | .parse_response(&mut response, RESPONSE); |
| 2123 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2124 | |
| 2125 | let result = crate::ParserConfig::default() |
| 2126 | .ignore_invalid_headers_in_responses(true) |
| 2127 | .parse_response(&mut response, RESPONSE); |
| 2128 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2129 | } |
| 2130 | |
| 2131 | #[test] |
| 2132 | fn test_header_with_cr_in_header_name() { |
| 2133 | const RESPONSE: &[u8] = |
| 2134 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Cred \rentials: hello \r\nBread: baguette \r\n\r\n" ; |
| 2135 | |
| 2136 | let mut headers = [EMPTY_HEADER; 2]; |
| 2137 | let mut response = Response::new(&mut headers[..]); |
| 2138 | |
| 2139 | let result = crate::ParserConfig::default() |
| 2140 | .parse_response(&mut response, RESPONSE); |
| 2141 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2142 | |
| 2143 | let result = crate::ParserConfig::default() |
| 2144 | .ignore_invalid_headers_in_responses(true) |
| 2145 | .parse_response(&mut response, RESPONSE); |
| 2146 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2147 | } |
| 2148 | |
| 2149 | #[test] |
| 2150 | fn test_header_with_nul_in_whitespace_before_colon() { |
| 2151 | const RESPONSE: &[u8] = |
| 2152 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials \0: hello \r\nBread: baguette \r\n\r\n" ; |
| 2153 | |
| 2154 | let mut headers = [EMPTY_HEADER; 2]; |
| 2155 | let mut response = Response::new(&mut headers[..]); |
| 2156 | |
| 2157 | let result = crate::ParserConfig::default() |
| 2158 | .allow_spaces_after_header_name_in_responses(true) |
| 2159 | .parse_response(&mut response, RESPONSE); |
| 2160 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2161 | |
| 2162 | let result = crate::ParserConfig::default() |
| 2163 | .allow_spaces_after_header_name_in_responses(true) |
| 2164 | .ignore_invalid_headers_in_responses(true) |
| 2165 | .parse_response(&mut response, RESPONSE); |
| 2166 | assert_eq!(result, Err(crate::Error::HeaderName)); |
| 2167 | } |
| 2168 | |
| 2169 | #[test] |
| 2170 | fn test_header_with_nul_in_value() { |
| 2171 | const RESPONSE: &[u8] = |
| 2172 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials: hell \0o \r\nBread: baguette \r\n\r\n" ; |
| 2173 | |
| 2174 | let mut headers = [EMPTY_HEADER; 2]; |
| 2175 | let mut response = Response::new(&mut headers[..]); |
| 2176 | |
| 2177 | let result = crate::ParserConfig::default() |
| 2178 | .parse_response(&mut response, RESPONSE); |
| 2179 | assert_eq!(result, Err(crate::Error::HeaderValue)); |
| 2180 | |
| 2181 | let result = crate::ParserConfig::default() |
| 2182 | .ignore_invalid_headers_in_responses(true) |
| 2183 | .parse_response(&mut response, RESPONSE); |
| 2184 | assert_eq!(result, Err(crate::Error::HeaderValue)); |
| 2185 | } |
| 2186 | |
| 2187 | #[test] |
| 2188 | fn test_header_with_invalid_char_in_value() { |
| 2189 | const RESPONSE: &[u8] = |
| 2190 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials: hell \x01o \r\nBread: baguette \r\n\r\n" ; |
| 2191 | |
| 2192 | let mut headers = [EMPTY_HEADER; 2]; |
| 2193 | let mut response = Response::new(&mut headers[..]); |
| 2194 | |
| 2195 | let result = crate::ParserConfig::default() |
| 2196 | .parse_response(&mut response, RESPONSE); |
| 2197 | assert_eq!(result, Err(crate::Error::HeaderValue)); |
| 2198 | |
| 2199 | let result = crate::ParserConfig::default() |
| 2200 | .ignore_invalid_headers_in_responses(true) |
| 2201 | .parse_response(&mut response, RESPONSE); |
| 2202 | assert_eq!(result, Ok(Status::Complete(78))); |
| 2203 | |
| 2204 | assert_eq!(response.version.unwrap(), 1); |
| 2205 | assert_eq!(response.code.unwrap(), 200); |
| 2206 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 2207 | assert_eq!(response.headers.len(), 1); |
| 2208 | assert_eq!(response.headers[0].name, "Bread" ); |
| 2209 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 2210 | } |
| 2211 | |
| 2212 | #[test] |
| 2213 | fn test_header_with_invalid_char_in_value_with_folding() { |
| 2214 | const RESPONSE: &[u8] = |
| 2215 | b"HTTP/1.1 200 OK \r\nAccess-Control-Allow-Credentials: hell \x01o \n world! \r\nBread: baguette \r\n\r\n" ; |
| 2216 | |
| 2217 | let mut headers = [EMPTY_HEADER; 2]; |
| 2218 | let mut response = Response::new(&mut headers[..]); |
| 2219 | |
| 2220 | let result = crate::ParserConfig::default() |
| 2221 | .parse_response(&mut response, RESPONSE); |
| 2222 | assert_eq!(result, Err(crate::Error::HeaderValue)); |
| 2223 | |
| 2224 | let result = crate::ParserConfig::default() |
| 2225 | .ignore_invalid_headers_in_responses(true) |
| 2226 | .parse_response(&mut response, RESPONSE); |
| 2227 | assert_eq!(result, Ok(Status::Complete(88))); |
| 2228 | |
| 2229 | assert_eq!(response.version.unwrap(), 1); |
| 2230 | assert_eq!(response.code.unwrap(), 200); |
| 2231 | assert_eq!(response.reason.unwrap(), "OK" ); |
| 2232 | assert_eq!(response.headers.len(), 1); |
| 2233 | assert_eq!(response.headers[0].name, "Bread" ); |
| 2234 | assert_eq!(response.headers[0].value, &b"baguette" [..]); |
| 2235 | } |
| 2236 | } |
| 2237 | |