| OLD | NEW |
| 1 // Copyright (c) 2012, the Dart project authors. Please see the AUTHORS file | 1 // Copyright (c) 2012, the Dart project authors. Please see the AUTHORS file |
| 2 // for details. All rights reserved. Use of this source code is governed by a | 2 // for details. All rights reserved. Use of this source code is governed by a |
| 3 // BSD-style license that can be found in the LICENSE file. | 3 // BSD-style license that can be found in the LICENSE file. |
| 4 | |
| 5 library mime.multipart_transformer; | 4 library mime.multipart_transformer; |
| 6 | 5 |
| 7 import 'dart:async'; | 6 import 'dart:async'; |
| 8 import 'dart:convert'; | |
| 9 import 'dart:typed_data'; | 7 import 'dart:typed_data'; |
| 10 | 8 |
| 11 /** | 9 import 'bound_multipart_stream.dart'; |
| 12 * A Mime Multipart class representing each part parsed by | 10 import 'mime_shared.dart'; |
| 13 * [MimeMultipartTransformer]. The data is streamed in as it become available. | 11 import 'char_code.dart'; |
| 14 */ | |
| 15 abstract class MimeMultipart extends Stream<List<int>> { | |
| 16 Map<String, String> get headers; | |
| 17 } | |
| 18 | 12 |
| 19 class _MimeMultipart extends MimeMultipart { | |
| 20 final Map<String, String> headers; | |
| 21 final Stream<List<int>> _stream; | |
| 22 | 13 |
| 23 _MimeMultipart(this.headers, this._stream); | 14 Uint8List _getBoundary(String boundary) { |
| 15 var charCodes = boundary.codeUnits; |
| 24 | 16 |
| 25 StreamSubscription<List<int>> listen(void onData(List<int> data), | 17 var boundaryList = new Uint8List(4 + charCodes.length); |
| 26 {void onDone(), | 18 // Set-up the matching boundary preceding it with CRLF and two |
| 27 Function onError, | 19 // dashes. |
| 28 bool cancelOnError}) { | 20 boundaryList[0] = CharCode.CR; |
| 29 return _stream.listen(onData, | 21 boundaryList[1] = CharCode.LF; |
| 30 onDone: onDone, | 22 boundaryList[2] = CharCode.DASH; |
| 31 onError: onError, | 23 boundaryList[3] = CharCode.DASH; |
| 32 cancelOnError: cancelOnError); | 24 boundaryList.setRange(4, 4 + charCodes.length, charCodes); |
| 33 } | 25 return boundaryList; |
| 34 } | |
| 35 | |
| 36 class _CharCode { | |
| 37 static const int HT = 9; | |
| 38 static const int LF = 10; | |
| 39 static const int CR = 13; | |
| 40 static const int SP = 32; | |
| 41 static const int DASH = 45; | |
| 42 static const int COLON = 58; | |
| 43 } | 26 } |
| 44 | 27 |
| 45 /** | 28 /** |
| 46 * Parser for MIME multipart types of data as described in RFC 2046 | 29 * Parser for MIME multipart types of data as described in RFC 2046 |
| 47 * section 5.1.1. The data is transformed into [MimeMultipart] objects, each | 30 * section 5.1.1. The data is transformed into [MimeMultipart] objects, each |
| 48 * of them streaming the multipart data. | 31 * of them streaming the multipart data. |
| 49 */ | 32 */ |
| 50 class MimeMultipartTransformer | 33 class MimeMultipartTransformer |
| 51 implements StreamTransformer<List<int>, MimeMultipart> { | 34 implements StreamTransformer<List<int>, MimeMultipart> { |
| 52 static const int _START = 0; | |
| 53 static const int _FIRST_BOUNDARY_ENDING = 111; | |
| 54 static const int _FIRST_BOUNDARY_END = 112; | |
| 55 static const int _BOUNDARY_ENDING = 1; | |
| 56 static const int _BOUNDARY_END = 2; | |
| 57 static const int _HEADER_START = 3; | |
| 58 static const int _HEADER_FIELD = 4; | |
| 59 static const int _HEADER_VALUE_START = 5; | |
| 60 static const int _HEADER_VALUE = 6; | |
| 61 static const int _HEADER_VALUE_FOLDING_OR_ENDING = 7; | |
| 62 static const int _HEADER_VALUE_FOLD_OR_END = 8; | |
| 63 static const int _HEADER_ENDING = 9; | |
| 64 static const int _CONTENT = 10; | |
| 65 static const int _LAST_BOUNDARY_DASH2 = 11; | |
| 66 static const int _LAST_BOUNDARY_ENDING = 12; | |
| 67 static const int _LAST_BOUNDARY_END = 13; | |
| 68 static const int _DONE = 14; | |
| 69 static const int _FAILURE = 15; | |
| 70 | 35 |
| 71 // Bytes for '()<>@,;:\\"/[]?={} \t'. | 36 final List<int> _boundary; |
| 72 static const _SEPARATORS = const [40, 41, 60, 62, 64, 44, 59, 58, 92, 34, 47, | |
| 73 91, 93, 63, 61, 123, 125, 32, 9]; | |
| 74 | |
| 75 StreamController _controller; | |
| 76 StreamSubscription _subscription; | |
| 77 | |
| 78 StreamController _multipartController; | |
| 79 Map<String, String> _headers; | |
| 80 | |
| 81 List<int> _boundary; | |
| 82 int _state = _START; | |
| 83 int _boundaryIndex = 2; | |
| 84 | |
| 85 // Current index in the data buffer. If index is negative then it | |
| 86 // is the index into the artificial prefix of the boundary string. | |
| 87 int _index; | |
| 88 List<int> _buffer; | |
| 89 | |
| 90 List<int> _headerField = []; | |
| 91 List<int> _headerValue = []; | |
| 92 | 37 |
| 93 /** | 38 /** |
| 94 * Construct a new MIME multipart parser with the boundary | 39 * Construct a new MIME multipart parser with the boundary |
| 95 * [boundary]. The boundary should be as specified in the content | 40 * [boundary]. The boundary should be as specified in the content |
| 96 * type parameter, that is without the -- prefix. | 41 * type parameter, that is without the -- prefix. |
| 97 */ | 42 */ |
| 98 MimeMultipartTransformer(String boundary) { | 43 MimeMultipartTransformer(String boundary) |
| 99 List<int> charCodes = boundary.codeUnits; | 44 : _boundary = _getBoundary(boundary); |
| 100 _boundary = new Uint8List(4 + charCodes.length); | |
| 101 // Set-up the matching boundary preceding it with CRLF and two | |
| 102 // dashes. | |
| 103 _boundary[0] = _CharCode.CR; | |
| 104 _boundary[1] = _CharCode.LF; | |
| 105 _boundary[2] = _CharCode.DASH; | |
| 106 _boundary[3] = _CharCode.DASH; | |
| 107 _boundary.setRange(4, 4 + charCodes.length, charCodes); | |
| 108 } | |
| 109 | |
| 110 void _resumeStream() { | |
| 111 _subscription.resume(); | |
| 112 } | |
| 113 | |
| 114 void _pauseStream() { | |
| 115 _subscription.pause(); | |
| 116 } | |
| 117 | 45 |
| 118 Stream<MimeMultipart> bind(Stream<List<int>> stream) { | 46 Stream<MimeMultipart> bind(Stream<List<int>> stream) { |
| 119 _controller = new StreamController( | 47 return new BoundMultipartStream(_boundary, stream).stream; |
| 120 sync: true, | |
| 121 onPause: _pauseStream, | |
| 122 onResume:_resumeStream, | |
| 123 onCancel: () { | |
| 124 _subscription.cancel(); | |
| 125 }, | |
| 126 onListen: () { | |
| 127 _subscription = stream.listen( | |
| 128 (data) { | |
| 129 assert(_buffer == null); | |
| 130 _pauseStream(); | |
| 131 _buffer = data; | |
| 132 _index = 0; | |
| 133 _parse(); | |
| 134 }, | |
| 135 onDone: () { | |
| 136 if (_state != _DONE) { | |
| 137 _controller.addError( | |
| 138 new MimeMultipartException("Bad multipart ending")); | |
| 139 } | |
| 140 _controller.close(); | |
| 141 }, | |
| 142 onError: _controller.addError); | |
| 143 }); | |
| 144 return _controller.stream; | |
| 145 } | |
| 146 | |
| 147 void _parse() { | |
| 148 // Number of boundary bytes to artificially place before the supplied data. | |
| 149 int boundaryPrefix = 0; | |
| 150 // Position where content starts. Will be null if no known content | |
| 151 // start exists. Will be negative of the content starts in the | |
| 152 // boundary prefix. Will be zero or position if the content starts | |
| 153 // in the current buffer. | |
| 154 int contentStartIndex; | |
| 155 | |
| 156 // Function to report content data for the current part. The data | |
| 157 // reported is from the current content start index up til the | |
| 158 // current index. As the data can be artificially prefixed with a | |
| 159 // prefix of the boundary both the content start index and index | |
| 160 // can be negative. | |
| 161 void reportData() { | |
| 162 if (contentStartIndex < 0) { | |
| 163 var contentLength = boundaryPrefix + _index - _boundaryIndex; | |
| 164 if (contentLength <= boundaryPrefix) { | |
| 165 _multipartController.add( | |
| 166 _boundary.sublist(0, contentLength)); | |
| 167 } else { | |
| 168 _multipartController.add( | |
| 169 _boundary.sublist(0, boundaryPrefix)); | |
| 170 _multipartController.add( | |
| 171 _buffer.sublist(0, contentLength - boundaryPrefix)); | |
| 172 } | |
| 173 } else { | |
| 174 var contentEndIndex = _index - _boundaryIndex; | |
| 175 _multipartController.add( | |
| 176 _buffer.sublist(contentStartIndex, contentEndIndex)); | |
| 177 } | |
| 178 } | |
| 179 | |
| 180 if (_state == _CONTENT && _boundaryIndex == 0) { | |
| 181 contentStartIndex = 0; | |
| 182 } else { | |
| 183 contentStartIndex = null; | |
| 184 } | |
| 185 // The data to parse might be "artificially" prefixed with a | |
| 186 // partial match of the boundary. | |
| 187 boundaryPrefix = _boundaryIndex; | |
| 188 | |
| 189 while ((_index < _buffer.length) && _state != _FAILURE && _state != _DONE) { | |
| 190 if (_multipartController != null && _multipartController.isPaused) { | |
| 191 return; | |
| 192 } | |
| 193 int byte; | |
| 194 if (_index < 0) { | |
| 195 byte = _boundary[boundaryPrefix + _index]; | |
| 196 } else { | |
| 197 byte = _buffer[_index]; | |
| 198 } | |
| 199 switch (_state) { | |
| 200 case _START: | |
| 201 if (byte == _boundary[_boundaryIndex]) { | |
| 202 _boundaryIndex++; | |
| 203 if (_boundaryIndex == _boundary.length) { | |
| 204 _state = _FIRST_BOUNDARY_ENDING; | |
| 205 _boundaryIndex = 0; | |
| 206 } | |
| 207 } else { | |
| 208 // Restart matching of the boundary. | |
| 209 _index = _index - _boundaryIndex; | |
| 210 _boundaryIndex = 0; | |
| 211 } | |
| 212 break; | |
| 213 | |
| 214 case _FIRST_BOUNDARY_ENDING: | |
| 215 if (byte == _CharCode.CR) { | |
| 216 _state = _FIRST_BOUNDARY_END; | |
| 217 } else { | |
| 218 _expectWS(byte); | |
| 219 } | |
| 220 break; | |
| 221 | |
| 222 case _FIRST_BOUNDARY_END: | |
| 223 _expect(byte, _CharCode.LF); | |
| 224 _state = _HEADER_START; | |
| 225 break; | |
| 226 | |
| 227 case _BOUNDARY_ENDING: | |
| 228 if (byte == _CharCode.CR) { | |
| 229 _state = _BOUNDARY_END; | |
| 230 } else if (byte == _CharCode.DASH) { | |
| 231 _state = _LAST_BOUNDARY_DASH2; | |
| 232 } else { | |
| 233 _expectWS(byte); | |
| 234 } | |
| 235 break; | |
| 236 | |
| 237 case _BOUNDARY_END: | |
| 238 _expect(byte, _CharCode.LF); | |
| 239 _multipartController.close(); | |
| 240 _multipartController = null; | |
| 241 _state = _HEADER_START; | |
| 242 break; | |
| 243 | |
| 244 case _HEADER_START: | |
| 245 _headers = new Map<String, String>(); | |
| 246 if (byte == _CharCode.CR) { | |
| 247 _state = _HEADER_ENDING; | |
| 248 } else { | |
| 249 // Start of new header field. | |
| 250 _headerField.add(_toLowerCase(byte)); | |
| 251 _state = _HEADER_FIELD; | |
| 252 } | |
| 253 break; | |
| 254 | |
| 255 case _HEADER_FIELD: | |
| 256 if (byte == _CharCode.COLON) { | |
| 257 _state = _HEADER_VALUE_START; | |
| 258 } else { | |
| 259 if (!_isTokenChar(byte)) { | |
| 260 throw new MimeMultipartException("Invalid header field name"); | |
| 261 } | |
| 262 _headerField.add(_toLowerCase(byte)); | |
| 263 } | |
| 264 break; | |
| 265 | |
| 266 case _HEADER_VALUE_START: | |
| 267 if (byte == _CharCode.CR) { | |
| 268 _state = _HEADER_VALUE_FOLDING_OR_ENDING; | |
| 269 } else if (byte != _CharCode.SP && byte != _CharCode.HT) { | |
| 270 // Start of new header value. | |
| 271 _headerValue.add(byte); | |
| 272 _state = _HEADER_VALUE; | |
| 273 } | |
| 274 break; | |
| 275 | |
| 276 case _HEADER_VALUE: | |
| 277 if (byte == _CharCode.CR) { | |
| 278 _state = _HEADER_VALUE_FOLDING_OR_ENDING; | |
| 279 } else { | |
| 280 _headerValue.add(byte); | |
| 281 } | |
| 282 break; | |
| 283 | |
| 284 case _HEADER_VALUE_FOLDING_OR_ENDING: | |
| 285 _expect(byte, _CharCode.LF); | |
| 286 _state = _HEADER_VALUE_FOLD_OR_END; | |
| 287 break; | |
| 288 | |
| 289 case _HEADER_VALUE_FOLD_OR_END: | |
| 290 if (byte == _CharCode.SP || byte == _CharCode.HT) { | |
| 291 _state = _HEADER_VALUE_START; | |
| 292 } else { | |
| 293 String headerField = UTF8.decode(_headerField); | |
| 294 String headerValue = UTF8.decode(_headerValue); | |
| 295 _headers[headerField.toLowerCase()] = headerValue; | |
| 296 _headerField = []; | |
| 297 _headerValue = []; | |
| 298 if (byte == _CharCode.CR) { | |
| 299 _state = _HEADER_ENDING; | |
| 300 } else { | |
| 301 // Start of new header field. | |
| 302 _headerField.add(_toLowerCase(byte)); | |
| 303 _state = _HEADER_FIELD; | |
| 304 } | |
| 305 } | |
| 306 break; | |
| 307 | |
| 308 case _HEADER_ENDING: | |
| 309 _expect(byte, _CharCode.LF); | |
| 310 _multipartController = new StreamController( | |
| 311 sync: true, | |
| 312 onPause: () { | |
| 313 _pauseStream(); | |
| 314 }, | |
| 315 onResume: () { | |
| 316 _resumeStream(); | |
| 317 _parse(); | |
| 318 }); | |
| 319 _controller.add( | |
| 320 new _MimeMultipart(_headers, _multipartController.stream)); | |
| 321 _headers = null; | |
| 322 _state = _CONTENT; | |
| 323 contentStartIndex = _index + 1; | |
| 324 break; | |
| 325 | |
| 326 case _CONTENT: | |
| 327 if (byte == _boundary[_boundaryIndex]) { | |
| 328 _boundaryIndex++; | |
| 329 if (_boundaryIndex == _boundary.length) { | |
| 330 if (contentStartIndex != null) { | |
| 331 _index++; | |
| 332 reportData(); | |
| 333 _index--; | |
| 334 } | |
| 335 _multipartController.close(); | |
| 336 _boundaryIndex = 0; | |
| 337 _state = _BOUNDARY_ENDING; | |
| 338 } | |
| 339 } else { | |
| 340 // Restart matching of the boundary. | |
| 341 _index = _index - _boundaryIndex; | |
| 342 if (contentStartIndex == null) contentStartIndex = _index; | |
| 343 _boundaryIndex = 0; | |
| 344 } | |
| 345 break; | |
| 346 | |
| 347 case _LAST_BOUNDARY_DASH2: | |
| 348 _expect(byte, _CharCode.DASH); | |
| 349 _state = _LAST_BOUNDARY_ENDING; | |
| 350 break; | |
| 351 | |
| 352 case _LAST_BOUNDARY_ENDING: | |
| 353 if (byte == _CharCode.CR) { | |
| 354 _state = _LAST_BOUNDARY_END; | |
| 355 } else { | |
| 356 _expectWS(byte); | |
| 357 } | |
| 358 break; | |
| 359 | |
| 360 case _LAST_BOUNDARY_END: | |
| 361 _expect(byte, _CharCode.LF); | |
| 362 _multipartController.close(); | |
| 363 _multipartController = null; | |
| 364 _state = _DONE; | |
| 365 break; | |
| 366 | |
| 367 default: | |
| 368 // Should be unreachable. | |
| 369 assert(false); | |
| 370 break; | |
| 371 } | |
| 372 | |
| 373 // Move to the next byte. | |
| 374 _index++; | |
| 375 } | |
| 376 | |
| 377 // Report any known content. | |
| 378 if (_state == _CONTENT && contentStartIndex != null) { | |
| 379 reportData(); | |
| 380 } | |
| 381 | |
| 382 // Resume if at end. | |
| 383 if (_index == _buffer.length) { | |
| 384 _buffer = null; | |
| 385 _index = null; | |
| 386 _resumeStream(); | |
| 387 } | |
| 388 } | |
| 389 | |
| 390 bool _isTokenChar(int byte) { | |
| 391 return byte > 31 && byte < 128 && _SEPARATORS.indexOf(byte) == -1; | |
| 392 } | |
| 393 | |
| 394 int _toLowerCase(int byte) { | |
| 395 final int aCode = "A".codeUnitAt(0); | |
| 396 final int zCode = "Z".codeUnitAt(0); | |
| 397 final int delta = "a".codeUnitAt(0) - aCode; | |
| 398 return (aCode <= byte && byte <= zCode) ? byte + delta : byte; | |
| 399 } | |
| 400 | |
| 401 void _expect(int val1, int val2) { | |
| 402 if (val1 != val2) { | |
| 403 throw new MimeMultipartException("Failed to parse multipart mime 1"); | |
| 404 } | |
| 405 } | |
| 406 | |
| 407 void _expectWS(int byte) { | |
| 408 if (byte != _CharCode.SP && byte != _CharCode.HT) { | |
| 409 throw new MimeMultipartException("Failed to parse multipart mime 2"); | |
| 410 } | |
| 411 } | 48 } |
| 412 } | 49 } |
| 413 | |
| 414 | |
| 415 class MimeMultipartException implements Exception { | |
| 416 final String message; | |
| 417 | |
| 418 const MimeMultipartException([String this.message = ""]); | |
| 419 | |
| 420 String toString() => "MimeMultipartException: $message"; | |
| 421 } | |
| OLD | NEW |