CharacterReference.php
6 months ago
DOMTreeBuilder.php
6 months ago
EventHandler.php
6 months ago
FileInputStream.php
6 months ago
InputStream.php
6 months ago
ParseError.php
6 months ago
README.md
1 year ago
Scanner.php
6 months ago
StringInputStream.php
6 months ago
Tokenizer.php
6 months ago
TreeBuildingRules.php
6 months ago
UTF8Utils.php
6 months ago
Tokenizer.php
1216 lines
| 1 | <?php |
| 2 | |
| 3 | namespace AmeliaVendor\Masterminds\HTML5\Parser; |
| 4 | |
| 5 | use AmeliaVendor\Masterminds\HTML5\Elements; |
| 6 | |
| 7 | /** |
| 8 | * The HTML5 tokenizer. |
| 9 | * |
| 10 | * The tokenizer's role is reading data from the scanner and gathering it into |
| 11 | * semantic units. From the tokenizer, data is emitted to an event handler, |
| 12 | * which may (for example) create a DOM tree. |
| 13 | * |
| 14 | * The HTML5 specification has a detailed explanation of tokenizing HTML5. We |
| 15 | * follow that specification to the maximum extent that we can. If you find |
| 16 | * a discrepancy that is not documented, please file a bug and/or submit a |
| 17 | * patch. |
| 18 | * |
| 19 | * This tokenizer is implemented as a recursive descent parser. |
| 20 | * |
| 21 | * Within the API documentation, you may see references to the specific section |
| 22 | * of the HTML5 spec that the code attempts to reproduce. Example: 8.2.4.1. |
| 23 | * This refers to section 8.2.4.1 of the HTML5 CR specification. |
| 24 | * |
| 25 | * @see http://www.w3.org/TR/2012/CR-html5-20121217/ |
| 26 | */ |
| 27 | class Tokenizer |
| 28 | { |
| 29 | protected $scanner; |
| 30 | |
| 31 | protected $events; |
| 32 | |
| 33 | protected $tok; |
| 34 | |
| 35 | /** |
| 36 | * Buffer for text. |
| 37 | */ |
| 38 | protected $text = ''; |
| 39 | |
| 40 | // When this goes to false, the parser stops. |
| 41 | protected $carryOn = true; |
| 42 | |
| 43 | protected $textMode = 0; // TEXTMODE_NORMAL; |
| 44 | protected $untilTag = null; |
| 45 | |
| 46 | const CONFORMANT_XML = 'xml'; |
| 47 | const CONFORMANT_HTML = 'html'; |
| 48 | protected $mode = self::CONFORMANT_HTML; |
| 49 | |
| 50 | /** |
| 51 | * Create a new tokenizer. |
| 52 | * |
| 53 | * Typically, parsing a document involves creating a new tokenizer, giving |
| 54 | * it a scanner (input) and an event handler (output), and then calling |
| 55 | * the Tokenizer::parse() method.` |
| 56 | * |
| 57 | * @param Scanner $scanner A scanner initialized with an input stream. |
| 58 | * @param EventHandler $eventHandler An event handler, initialized and ready to receive events. |
| 59 | * @param string $mode |
| 60 | */ |
| 61 | public function __construct($scanner, $eventHandler, $mode = self::CONFORMANT_HTML) |
| 62 | { |
| 63 | $this->scanner = $scanner; |
| 64 | $this->events = $eventHandler; |
| 65 | $this->mode = $mode; |
| 66 | } |
| 67 | |
| 68 | /** |
| 69 | * Begin parsing. |
| 70 | * |
| 71 | * This will begin scanning the document, tokenizing as it goes. |
| 72 | * Tokens are emitted into the event handler. |
| 73 | * |
| 74 | * Tokenizing will continue until the document is completely |
| 75 | * read. Errors are emitted into the event handler, but |
| 76 | * the parser will attempt to continue parsing until the |
| 77 | * entire input stream is read. |
| 78 | */ |
| 79 | public function parse() |
| 80 | { |
| 81 | do { |
| 82 | $this->consumeData(); |
| 83 | // FIXME: Add infinite loop protection. |
| 84 | } while ($this->carryOn); |
| 85 | } |
| 86 | |
| 87 | /** |
| 88 | * Set the text mode for the character data reader. |
| 89 | * |
| 90 | * HTML5 defines three different modes for reading text: |
| 91 | * - Normal: Read until a tag is encountered. |
| 92 | * - RCDATA: Read until a tag is encountered, but skip a few otherwise- |
| 93 | * special characters. |
| 94 | * - Raw: Read until a special closing tag is encountered (viz. pre, script) |
| 95 | * |
| 96 | * This allows those modes to be set. |
| 97 | * |
| 98 | * Normally, setting is done by the event handler via a special return code on |
| 99 | * startTag(), but it can also be set manually using this function. |
| 100 | * |
| 101 | * @param int $textmode One of Elements::TEXT_*. |
| 102 | * @param string $untilTag The tag that should stop RAW or RCDATA mode. Normal mode does not |
| 103 | * use this indicator. |
| 104 | */ |
| 105 | public function setTextMode($textmode, $untilTag = null) |
| 106 | { |
| 107 | $this->textMode = $textmode & (Elements::TEXT_RAW | Elements::TEXT_RCDATA); |
| 108 | $this->untilTag = $untilTag; |
| 109 | } |
| 110 | |
| 111 | /** |
| 112 | * Consume a character and make a move. |
| 113 | * HTML5 8.2.4.1. |
| 114 | */ |
| 115 | protected function consumeData() |
| 116 | { |
| 117 | $tok = $this->scanner->current(); |
| 118 | |
| 119 | if ('&' === $tok) { |
| 120 | // Character reference |
| 121 | $ref = $this->decodeCharacterReference(); |
| 122 | $this->buffer($ref); |
| 123 | |
| 124 | $tok = $this->scanner->current(); |
| 125 | } |
| 126 | |
| 127 | // Parse tag |
| 128 | if ('<' === $tok) { |
| 129 | // Any buffered text data can go out now. |
| 130 | $this->flushBuffer(); |
| 131 | |
| 132 | $tok = $this->scanner->next(); |
| 133 | |
| 134 | if (false === $tok) { |
| 135 | // end of string |
| 136 | $this->parseError('Illegal tag opening'); |
| 137 | } elseif ('!' === $tok) { |
| 138 | $this->markupDeclaration(); |
| 139 | } elseif ('/' === $tok) { |
| 140 | $this->endTag(); |
| 141 | } elseif ('?' === $tok) { |
| 142 | $this->processingInstruction(); |
| 143 | } elseif ($this->is_alpha($tok)) { |
| 144 | $this->tagName(); |
| 145 | } else { |
| 146 | $this->parseError('Illegal tag opening'); |
| 147 | // TODO is this necessary ? |
| 148 | $this->characterData(); |
| 149 | } |
| 150 | |
| 151 | $tok = $this->scanner->current(); |
| 152 | } |
| 153 | |
| 154 | if (false === $tok) { |
| 155 | // Handle end of document |
| 156 | $this->eof(); |
| 157 | } else { |
| 158 | // Parse character |
| 159 | switch ($this->textMode) { |
| 160 | case Elements::TEXT_RAW: |
| 161 | $this->rawText($tok); |
| 162 | break; |
| 163 | |
| 164 | case Elements::TEXT_RCDATA: |
| 165 | $this->rcdata($tok); |
| 166 | break; |
| 167 | |
| 168 | default: |
| 169 | if ('<' === $tok || '&' === $tok) { |
| 170 | break; |
| 171 | } |
| 172 | |
| 173 | // NULL character |
| 174 | if ("\00" === $tok) { |
| 175 | $this->parseError('Received null character.'); |
| 176 | |
| 177 | $this->text .= $tok; |
| 178 | $this->scanner->consume(); |
| 179 | |
| 180 | break; |
| 181 | } |
| 182 | |
| 183 | $this->text .= $this->scanner->charsUntil("<&\0"); |
| 184 | } |
| 185 | } |
| 186 | |
| 187 | return $this->carryOn; |
| 188 | } |
| 189 | |
| 190 | /** |
| 191 | * Parse anything that looks like character data. |
| 192 | * |
| 193 | * Different rules apply based on the current text mode. |
| 194 | * |
| 195 | * @see Elements::TEXT_RAW Elements::TEXT_RCDATA. |
| 196 | */ |
| 197 | protected function characterData() |
| 198 | { |
| 199 | $tok = $this->scanner->current(); |
| 200 | if (false === $tok) { |
| 201 | return false; |
| 202 | } |
| 203 | switch ($this->textMode) { |
| 204 | case Elements::TEXT_RAW: |
| 205 | return $this->rawText($tok); |
| 206 | case Elements::TEXT_RCDATA: |
| 207 | return $this->rcdata($tok); |
| 208 | default: |
| 209 | if ('<' === $tok || '&' === $tok) { |
| 210 | return false; |
| 211 | } |
| 212 | |
| 213 | return $this->text($tok); |
| 214 | } |
| 215 | } |
| 216 | |
| 217 | /** |
| 218 | * This buffers the current token as character data. |
| 219 | * |
| 220 | * @param string $tok The current token. |
| 221 | * |
| 222 | * @return bool |
| 223 | */ |
| 224 | protected function text($tok) |
| 225 | { |
| 226 | // This should never happen... |
| 227 | if (false === $tok) { |
| 228 | return false; |
| 229 | } |
| 230 | |
| 231 | // NULL character |
| 232 | if ("\00" === $tok) { |
| 233 | $this->parseError('Received null character.'); |
| 234 | } |
| 235 | |
| 236 | $this->buffer($tok); |
| 237 | $this->scanner->consume(); |
| 238 | |
| 239 | return true; |
| 240 | } |
| 241 | |
| 242 | /** |
| 243 | * Read text in RAW mode. |
| 244 | * |
| 245 | * @param string $tok The current token. |
| 246 | * |
| 247 | * @return bool |
| 248 | */ |
| 249 | protected function rawText($tok) |
| 250 | { |
| 251 | if (is_null($this->untilTag)) { |
| 252 | return $this->text($tok); |
| 253 | } |
| 254 | |
| 255 | $sequence = '</' . $this->untilTag . '>'; |
| 256 | $txt = $this->readUntilSequence($sequence); |
| 257 | $this->events->text($txt); |
| 258 | $this->setTextMode(0); |
| 259 | |
| 260 | return $this->endTag(); |
| 261 | } |
| 262 | |
| 263 | /** |
| 264 | * Read text in RCDATA mode. |
| 265 | * |
| 266 | * @param string $tok The current token. |
| 267 | * |
| 268 | * @return bool |
| 269 | */ |
| 270 | protected function rcdata($tok) |
| 271 | { |
| 272 | if (is_null($this->untilTag)) { |
| 273 | return $this->text($tok); |
| 274 | } |
| 275 | |
| 276 | $sequence = '</' . $this->untilTag; |
| 277 | $txt = ''; |
| 278 | |
| 279 | $caseSensitive = !Elements::isHtml5Element($this->untilTag); |
| 280 | while (false !== $tok && !('<' == $tok && ($this->scanner->sequenceMatches($sequence, $caseSensitive)))) { |
| 281 | if ('&' == $tok) { |
| 282 | $txt .= $this->decodeCharacterReference(); |
| 283 | $tok = $this->scanner->current(); |
| 284 | } else { |
| 285 | $txt .= $tok; |
| 286 | $tok = $this->scanner->next(); |
| 287 | } |
| 288 | } |
| 289 | $len = strlen($sequence); |
| 290 | $this->scanner->consume($len); |
| 291 | $len += $this->scanner->whitespace(); |
| 292 | if ('>' !== $this->scanner->current()) { |
| 293 | $this->parseError('Unclosed RCDATA end tag'); |
| 294 | } |
| 295 | |
| 296 | $this->scanner->unconsume($len); |
| 297 | $this->events->text($txt); |
| 298 | $this->setTextMode(0); |
| 299 | |
| 300 | return $this->endTag(); |
| 301 | } |
| 302 | |
| 303 | /** |
| 304 | * If the document is read, emit an EOF event. |
| 305 | */ |
| 306 | protected function eof() |
| 307 | { |
| 308 | // fprintf(STDOUT, "EOF"); |
| 309 | $this->flushBuffer(); |
| 310 | $this->events->eof(); |
| 311 | $this->carryOn = false; |
| 312 | } |
| 313 | |
| 314 | /** |
| 315 | * Look for markup. |
| 316 | */ |
| 317 | protected function markupDeclaration() |
| 318 | { |
| 319 | $tok = $this->scanner->next(); |
| 320 | |
| 321 | // Comment: |
| 322 | if ('-' == $tok && '-' == $this->scanner->peek()) { |
| 323 | $this->scanner->consume(2); |
| 324 | |
| 325 | return $this->comment(); |
| 326 | } elseif ('D' == $tok || 'd' == $tok) { // Doctype |
| 327 | return $this->doctype(); |
| 328 | } elseif ('[' == $tok) { // CDATA section |
| 329 | return $this->cdataSection(); |
| 330 | } |
| 331 | |
| 332 | // FINISH |
| 333 | $this->parseError('Expected <!--, <![CDATA[, or <!DOCTYPE. Got <!%s', $tok); |
| 334 | $this->bogusComment('<!'); |
| 335 | |
| 336 | return true; |
| 337 | } |
| 338 | |
| 339 | /** |
| 340 | * Consume an end tag. See section 8.2.4.9. |
| 341 | */ |
| 342 | protected function endTag() |
| 343 | { |
| 344 | if ('/' != $this->scanner->current()) { |
| 345 | return false; |
| 346 | } |
| 347 | $tok = $this->scanner->next(); |
| 348 | |
| 349 | // a-zA-Z -> tagname |
| 350 | // > -> parse error |
| 351 | // EOF -> parse error |
| 352 | // -> parse error |
| 353 | if (!$this->is_alpha($tok)) { |
| 354 | $this->parseError("Expected tag name, got '%s'", $tok); |
| 355 | if ("\0" == $tok || false === $tok) { |
| 356 | return false; |
| 357 | } |
| 358 | |
| 359 | return $this->bogusComment('</'); |
| 360 | } |
| 361 | |
| 362 | $name = $this->scanner->charsUntil("\n\f \t>"); |
| 363 | $name = self::CONFORMANT_XML === $this->mode ? $name : strtolower($name); |
| 364 | // Trash whitespace. |
| 365 | $this->scanner->whitespace(); |
| 366 | |
| 367 | $tok = $this->scanner->current(); |
| 368 | if ('>' != $tok) { |
| 369 | $this->parseError("Expected >, got '%s'", $tok); |
| 370 | // We just trash stuff until we get to the next tag close. |
| 371 | $this->scanner->charsUntil('>'); |
| 372 | } |
| 373 | |
| 374 | $this->events->endTag($name); |
| 375 | $this->scanner->consume(); |
| 376 | |
| 377 | return true; |
| 378 | } |
| 379 | |
| 380 | /** |
| 381 | * Consume a tag name and body. See section 8.2.4.10. |
| 382 | */ |
| 383 | protected function tagName() |
| 384 | { |
| 385 | // We know this is at least one char. |
| 386 | $name = $this->scanner->charsWhile(':_-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz'); |
| 387 | $name = self::CONFORMANT_XML === $this->mode ? $name : strtolower($name); |
| 388 | $attributes = array(); |
| 389 | $selfClose = false; |
| 390 | |
| 391 | // Handle attribute parse exceptions here so that we can |
| 392 | // react by trying to build a sensible parse tree. |
| 393 | try { |
| 394 | do { |
| 395 | $this->scanner->whitespace(); |
| 396 | $this->attribute($attributes); |
| 397 | } while (!$this->isTagEnd($selfClose)); |
| 398 | } catch (ParseError $e) { |
| 399 | $selfClose = false; |
| 400 | } |
| 401 | |
| 402 | $mode = $this->events->startTag($name, $attributes, $selfClose); |
| 403 | |
| 404 | if (is_int($mode)) { |
| 405 | $this->setTextMode($mode, $name); |
| 406 | } |
| 407 | |
| 408 | $this->scanner->consume(); |
| 409 | |
| 410 | return true; |
| 411 | } |
| 412 | |
| 413 | /** |
| 414 | * Check if the scanner has reached the end of a tag. |
| 415 | */ |
| 416 | protected function isTagEnd(&$selfClose) |
| 417 | { |
| 418 | $tok = $this->scanner->current(); |
| 419 | if ('/' == $tok) { |
| 420 | $this->scanner->consume(); |
| 421 | $this->scanner->whitespace(); |
| 422 | $tok = $this->scanner->current(); |
| 423 | |
| 424 | if ('>' == $tok) { |
| 425 | $selfClose = true; |
| 426 | |
| 427 | return true; |
| 428 | } |
| 429 | if (false === $tok) { |
| 430 | $this->parseError('Unexpected EOF inside of tag.'); |
| 431 | |
| 432 | return true; |
| 433 | } |
| 434 | // Basically, we skip the / token and go on. |
| 435 | // See 8.2.4.43. |
| 436 | $this->parseError("Unexpected '%s' inside of a tag.", $tok); |
| 437 | |
| 438 | return false; |
| 439 | } |
| 440 | |
| 441 | if ('>' == $tok) { |
| 442 | return true; |
| 443 | } |
| 444 | if (false === $tok) { |
| 445 | $this->parseError('Unexpected EOF inside of tag.'); |
| 446 | |
| 447 | return true; |
| 448 | } |
| 449 | |
| 450 | return false; |
| 451 | } |
| 452 | |
| 453 | /** |
| 454 | * Parse attributes from inside of a tag. |
| 455 | * |
| 456 | * @param string[] $attributes |
| 457 | * |
| 458 | * @return bool |
| 459 | * |
| 460 | * @throws ParseError |
| 461 | */ |
| 462 | protected function attribute(&$attributes) |
| 463 | { |
| 464 | $tok = $this->scanner->current(); |
| 465 | if ('/' == $tok || '>' == $tok || false === $tok) { |
| 466 | return false; |
| 467 | } |
| 468 | |
| 469 | if ('<' == $tok) { |
| 470 | $this->parseError("Unexpected '<' inside of attributes list."); |
| 471 | // Push the < back onto the stack. |
| 472 | $this->scanner->unconsume(); |
| 473 | // Let the caller figure out how to handle this. |
| 474 | throw new ParseError('Start tag inside of attribute.'); |
| 475 | } |
| 476 | |
| 477 | $name = strtolower($this->scanner->charsUntil("/>=\n\f\t ")); |
| 478 | |
| 479 | if (0 == strlen($name)) { |
| 480 | $tok = $this->scanner->current(); |
| 481 | $this->parseError('Expected an attribute name, got %s.', $tok); |
| 482 | // Really, only '=' can be the char here. Everything else gets absorbed |
| 483 | // under one rule or another. |
| 484 | $name = $tok; |
| 485 | $this->scanner->consume(); |
| 486 | } |
| 487 | |
| 488 | $isValidAttribute = true; |
| 489 | // Attribute names can contain most Unicode characters for HTML5. |
| 490 | // But method "DOMElement::setAttribute" is throwing exception |
| 491 | // because of it's own internal restriction so these have to be filtered. |
| 492 | // see issue #23: https://github.com/Masterminds/html5-php/issues/23 |
| 493 | // and http://www.w3.org/TR/2011/WD-html5-20110525/syntax.html#syntax-attribute-name |
| 494 | if (preg_match("/[\x1-\x2C\\/\x3B-\x40\x5B-\x5E\x60\x7B-\x7F]/u", $name)) { |
| 495 | $this->parseError('Unexpected characters in attribute name: %s', $name); |
| 496 | $isValidAttribute = false; |
| 497 | } // There is no limitation for 1st character in HTML5. |
| 498 | // But method "DOMElement::setAttribute" is throwing exception for the |
| 499 | // characters below so they have to be filtered. |
| 500 | // see issue #23: https://github.com/Masterminds/html5-php/issues/23 |
| 501 | // and http://www.w3.org/TR/2011/WD-html5-20110525/syntax.html#syntax-attribute-name |
| 502 | elseif (preg_match('/^[0-9.-]/u', $name)) { |
| 503 | $this->parseError('Unexpected character at the begining of attribute name: %s', $name); |
| 504 | $isValidAttribute = false; |
| 505 | } |
| 506 | // 8.1.2.3 |
| 507 | $this->scanner->whitespace(); |
| 508 | |
| 509 | $val = $this->attributeValue(); |
| 510 | if ($isValidAttribute && !array_key_exists($name, $attributes)) { |
| 511 | $attributes[$name] = $val; |
| 512 | } |
| 513 | |
| 514 | return true; |
| 515 | } |
| 516 | |
| 517 | /** |
| 518 | * Consume an attribute value. See section 8.2.4.37 and after. |
| 519 | * |
| 520 | * @return string|null |
| 521 | */ |
| 522 | protected function attributeValue() |
| 523 | { |
| 524 | if ('=' != $this->scanner->current()) { |
| 525 | return null; |
| 526 | } |
| 527 | $this->scanner->consume(); |
| 528 | // 8.1.2.3 |
| 529 | $this->scanner->whitespace(); |
| 530 | |
| 531 | $tok = $this->scanner->current(); |
| 532 | switch ($tok) { |
| 533 | case "\n": |
| 534 | case "\f": |
| 535 | case ' ': |
| 536 | case "\t": |
| 537 | // Whitespace here indicates an empty value. |
| 538 | return null; |
| 539 | case '"': |
| 540 | case "'": |
| 541 | $this->scanner->consume(); |
| 542 | |
| 543 | return $this->quotedAttributeValue($tok); |
| 544 | case '>': |
| 545 | // case '/': // 8.2.4.37 seems to allow foo=/ as a valid attr. |
| 546 | $this->parseError('Expected attribute value, got tag end.'); |
| 547 | |
| 548 | return null; |
| 549 | case '=': |
| 550 | case '`': |
| 551 | $this->parseError('Expecting quotes, got %s.', $tok); |
| 552 | |
| 553 | return $this->unquotedAttributeValue(); |
| 554 | default: |
| 555 | return $this->unquotedAttributeValue(); |
| 556 | } |
| 557 | } |
| 558 | |
| 559 | /** |
| 560 | * Get an attribute value string. |
| 561 | * |
| 562 | * @param string $quote IMPORTANT: This is a series of chars! Any one of which will be considered |
| 563 | * termination of an attribute's value. E.g. "\"'" will stop at either |
| 564 | * ' or ". |
| 565 | * |
| 566 | * @return string The attribute value. |
| 567 | */ |
| 568 | protected function quotedAttributeValue($quote) |
| 569 | { |
| 570 | $stoplist = "\f" . $quote; |
| 571 | $val = ''; |
| 572 | |
| 573 | while (true) { |
| 574 | $tokens = $this->scanner->charsUntil($stoplist . '&'); |
| 575 | if (false !== $tokens) { |
| 576 | $val .= $tokens; |
| 577 | } else { |
| 578 | break; |
| 579 | } |
| 580 | |
| 581 | $tok = $this->scanner->current(); |
| 582 | if ('&' == $tok) { |
| 583 | $val .= $this->decodeCharacterReference(true); |
| 584 | continue; |
| 585 | } |
| 586 | break; |
| 587 | } |
| 588 | $this->scanner->consume(); |
| 589 | |
| 590 | return $val; |
| 591 | } |
| 592 | |
| 593 | protected function unquotedAttributeValue() |
| 594 | { |
| 595 | $val = ''; |
| 596 | $tok = $this->scanner->current(); |
| 597 | while (false !== $tok) { |
| 598 | switch ($tok) { |
| 599 | case "\n": |
| 600 | case "\f": |
| 601 | case ' ': |
| 602 | case "\t": |
| 603 | case '>': |
| 604 | break 2; |
| 605 | |
| 606 | case '&': |
| 607 | $val .= $this->decodeCharacterReference(true); |
| 608 | $tok = $this->scanner->current(); |
| 609 | |
| 610 | break; |
| 611 | |
| 612 | case "'": |
| 613 | case '"': |
| 614 | case '<': |
| 615 | case '=': |
| 616 | case '`': |
| 617 | $this->parseError('Unexpected chars in unquoted attribute value %s', $tok); |
| 618 | $val .= $tok; |
| 619 | $tok = $this->scanner->next(); |
| 620 | break; |
| 621 | |
| 622 | default: |
| 623 | $val .= $this->scanner->charsUntil("\t\n\f >&\"'<=`"); |
| 624 | |
| 625 | $tok = $this->scanner->current(); |
| 626 | } |
| 627 | } |
| 628 | |
| 629 | return $val; |
| 630 | } |
| 631 | |
| 632 | /** |
| 633 | * Consume malformed markup as if it were a comment. |
| 634 | * 8.2.4.44. |
| 635 | * |
| 636 | * The spec requires that the ENTIRE tag-like thing be enclosed inside of |
| 637 | * the comment. So this will generate comments like: |
| 638 | * |
| 639 | * <!--</+foo>--> |
| 640 | * |
| 641 | * @param string $leading Prepend any leading characters. This essentially |
| 642 | * negates the need to backtrack, but it's sort of a hack. |
| 643 | * |
| 644 | * @return bool |
| 645 | */ |
| 646 | protected function bogusComment($leading = '') |
| 647 | { |
| 648 | $comment = $leading; |
| 649 | $tokens = $this->scanner->charsUntil('>'); |
| 650 | if (false !== $tokens) { |
| 651 | $comment .= $tokens; |
| 652 | } |
| 653 | $tok = $this->scanner->current(); |
| 654 | if (false !== $tok) { |
| 655 | $comment .= $tok; |
| 656 | } |
| 657 | |
| 658 | $this->flushBuffer(); |
| 659 | $this->events->comment($comment); |
| 660 | $this->scanner->consume(); |
| 661 | |
| 662 | return true; |
| 663 | } |
| 664 | |
| 665 | /** |
| 666 | * Read a comment. |
| 667 | * Expects the first tok to be inside of the comment. |
| 668 | * |
| 669 | * @return bool |
| 670 | */ |
| 671 | protected function comment() |
| 672 | { |
| 673 | $tok = $this->scanner->current(); |
| 674 | $comment = ''; |
| 675 | |
| 676 | // <!-->. Emit an empty comment because 8.2.4.46 says to. |
| 677 | if ('>' == $tok) { |
| 678 | // Parse error. Emit the comment token. |
| 679 | $this->parseError("Expected comment data, got '>'"); |
| 680 | $this->events->comment(''); |
| 681 | $this->scanner->consume(); |
| 682 | |
| 683 | return true; |
| 684 | } |
| 685 | |
| 686 | // Replace NULL with the replacement char. |
| 687 | if ("\0" == $tok) { |
| 688 | $tok = UTF8Utils::FFFD; |
| 689 | } |
| 690 | while (!$this->isCommentEnd()) { |
| 691 | $comment .= $tok; |
| 692 | $tok = $this->scanner->next(); |
| 693 | } |
| 694 | |
| 695 | $this->events->comment($comment); |
| 696 | $this->scanner->consume(); |
| 697 | |
| 698 | return true; |
| 699 | } |
| 700 | |
| 701 | /** |
| 702 | * Check if the scanner has reached the end of a comment. |
| 703 | * |
| 704 | * @return bool |
| 705 | */ |
| 706 | protected function isCommentEnd() |
| 707 | { |
| 708 | $tok = $this->scanner->current(); |
| 709 | |
| 710 | // EOF |
| 711 | if (false === $tok) { |
| 712 | // Hit the end. |
| 713 | $this->parseError('Unexpected EOF in a comment.'); |
| 714 | |
| 715 | return true; |
| 716 | } |
| 717 | |
| 718 | // If next two tokens are not '--', not the end. |
| 719 | if ('-' != $tok || '-' != $this->scanner->peek()) { |
| 720 | return false; |
| 721 | } |
| 722 | |
| 723 | $this->scanner->consume(2); // Consume '-' and one of '!' or '>' |
| 724 | |
| 725 | // Test for '>' |
| 726 | if ('>' == $this->scanner->current()) { |
| 727 | return true; |
| 728 | } |
| 729 | // Test for '!>' |
| 730 | if ('!' == $this->scanner->current() && '>' == $this->scanner->peek()) { |
| 731 | $this->scanner->consume(); // Consume the last '>' |
| 732 | |
| 733 | return true; |
| 734 | } |
| 735 | // Unread '-' and one of '!' or '>'; |
| 736 | $this->scanner->unconsume(2); |
| 737 | |
| 738 | return false; |
| 739 | } |
| 740 | |
| 741 | /** |
| 742 | * Parse a DOCTYPE. |
| 743 | * |
| 744 | * Parse a DOCTYPE declaration. This method has strong bearing on whether or |
| 745 | * not Quirksmode is enabled on the event handler. |
| 746 | * |
| 747 | * @todo This method is a little long. Should probably refactor. |
| 748 | * |
| 749 | * @return bool |
| 750 | */ |
| 751 | protected function doctype() |
| 752 | { |
| 753 | // Check that string is DOCTYPE. |
| 754 | if ($this->scanner->sequenceMatches('DOCTYPE', false)) { |
| 755 | $this->scanner->consume(7); |
| 756 | } else { |
| 757 | $chars = $this->scanner->charsWhile('DOCTYPEdoctype'); |
| 758 | $this->parseError('Expected DOCTYPE, got %s', $chars); |
| 759 | |
| 760 | return $this->bogusComment('<!' . $chars); |
| 761 | } |
| 762 | |
| 763 | $this->scanner->whitespace(); |
| 764 | $tok = $this->scanner->current(); |
| 765 | |
| 766 | // EOF: die. |
| 767 | if (false === $tok) { |
| 768 | $this->events->doctype('html5', EventHandler::DOCTYPE_NONE, '', true); |
| 769 | $this->eof(); |
| 770 | |
| 771 | return true; |
| 772 | } |
| 773 | |
| 774 | // NULL char: convert. |
| 775 | if ("\0" === $tok) { |
| 776 | $this->parseError('Unexpected null character in DOCTYPE.'); |
| 777 | } |
| 778 | |
| 779 | $stop = " \n\f>"; |
| 780 | $doctypeName = $this->scanner->charsUntil($stop); |
| 781 | // Lowercase ASCII, replace \0 with FFFD |
| 782 | $doctypeName = strtolower(strtr($doctypeName, "\0", UTF8Utils::FFFD)); |
| 783 | |
| 784 | $tok = $this->scanner->current(); |
| 785 | |
| 786 | // If false, emit a parse error, DOCTYPE, and return. |
| 787 | if (false === $tok) { |
| 788 | $this->parseError('Unexpected EOF in DOCTYPE declaration.'); |
| 789 | $this->events->doctype($doctypeName, EventHandler::DOCTYPE_NONE, null, true); |
| 790 | |
| 791 | return true; |
| 792 | } |
| 793 | |
| 794 | // Short DOCTYPE, like <!DOCTYPE html> |
| 795 | if ('>' == $tok) { |
| 796 | // DOCTYPE without a name. |
| 797 | if (0 == strlen($doctypeName)) { |
| 798 | $this->parseError('Expected a DOCTYPE name. Got nothing.'); |
| 799 | $this->events->doctype($doctypeName, 0, null, true); |
| 800 | $this->scanner->consume(); |
| 801 | |
| 802 | return true; |
| 803 | } |
| 804 | $this->events->doctype($doctypeName); |
| 805 | $this->scanner->consume(); |
| 806 | |
| 807 | return true; |
| 808 | } |
| 809 | $this->scanner->whitespace(); |
| 810 | |
| 811 | $pub = strtoupper($this->scanner->getAsciiAlpha()); |
| 812 | $white = $this->scanner->whitespace(); |
| 813 | |
| 814 | // Get ID, and flag it as pub or system. |
| 815 | if (('PUBLIC' == $pub || 'SYSTEM' == $pub) && $white > 0) { |
| 816 | // Get the sys ID. |
| 817 | $type = 'PUBLIC' == $pub ? EventHandler::DOCTYPE_PUBLIC : EventHandler::DOCTYPE_SYSTEM; |
| 818 | $id = $this->quotedString("\0>"); |
| 819 | if (false === $id) { |
| 820 | $this->events->doctype($doctypeName, $type, $pub, false); |
| 821 | |
| 822 | return true; |
| 823 | } |
| 824 | |
| 825 | // Premature EOF. |
| 826 | if (false === $this->scanner->current()) { |
| 827 | $this->parseError('Unexpected EOF in DOCTYPE'); |
| 828 | $this->events->doctype($doctypeName, $type, $id, true); |
| 829 | |
| 830 | return true; |
| 831 | } |
| 832 | |
| 833 | // Well-formed complete DOCTYPE. |
| 834 | $this->scanner->whitespace(); |
| 835 | if ('>' == $this->scanner->current()) { |
| 836 | $this->events->doctype($doctypeName, $type, $id, false); |
| 837 | $this->scanner->consume(); |
| 838 | |
| 839 | return true; |
| 840 | } |
| 841 | |
| 842 | // If we get here, we have <!DOCTYPE foo PUBLIC "bar" SOME_JUNK |
| 843 | // Throw away the junk, parse error, quirks mode, return true. |
| 844 | $this->scanner->charsUntil('>'); |
| 845 | $this->parseError('Malformed DOCTYPE.'); |
| 846 | $this->events->doctype($doctypeName, $type, $id, true); |
| 847 | $this->scanner->consume(); |
| 848 | |
| 849 | return true; |
| 850 | } |
| 851 | |
| 852 | // Else it's a bogus DOCTYPE. |
| 853 | // Consume to > and trash. |
| 854 | $this->scanner->charsUntil('>'); |
| 855 | |
| 856 | $this->parseError('Expected PUBLIC or SYSTEM. Got %s.', $pub); |
| 857 | $this->events->doctype($doctypeName, 0, null, true); |
| 858 | $this->scanner->consume(); |
| 859 | |
| 860 | return true; |
| 861 | } |
| 862 | |
| 863 | /** |
| 864 | * Utility for reading a quoted string. |
| 865 | * |
| 866 | * @param string $stopchars Characters (in addition to a close-quote) that should stop the string. |
| 867 | * E.g. sometimes '>' is higher precedence than '"' or "'". |
| 868 | * |
| 869 | * @return mixed String if one is found (quotations omitted). |
| 870 | */ |
| 871 | protected function quotedString($stopchars) |
| 872 | { |
| 873 | $tok = $this->scanner->current(); |
| 874 | if ('"' == $tok || "'" == $tok) { |
| 875 | $this->scanner->consume(); |
| 876 | $ret = $this->scanner->charsUntil($tok . $stopchars); |
| 877 | if ($this->scanner->current() == $tok) { |
| 878 | $this->scanner->consume(); |
| 879 | } else { |
| 880 | // Parse error because no close quote. |
| 881 | $this->parseError('Expected %s, got %s', $tok, $this->scanner->current()); |
| 882 | } |
| 883 | |
| 884 | return $ret; |
| 885 | } |
| 886 | |
| 887 | return false; |
| 888 | } |
| 889 | |
| 890 | /** |
| 891 | * Handle a CDATA section. |
| 892 | * |
| 893 | * @return bool |
| 894 | */ |
| 895 | protected function cdataSection() |
| 896 | { |
| 897 | $cdata = ''; |
| 898 | $this->scanner->consume(); |
| 899 | |
| 900 | $chars = $this->scanner->charsWhile('CDAT'); |
| 901 | if ('CDATA' != $chars || '[' != $this->scanner->current()) { |
| 902 | $this->parseError('Expected [CDATA[, got %s', $chars); |
| 903 | |
| 904 | return $this->bogusComment('<![' . $chars); |
| 905 | } |
| 906 | |
| 907 | $tok = $this->scanner->next(); |
| 908 | do { |
| 909 | if (false === $tok) { |
| 910 | $this->parseError('Unexpected EOF inside CDATA.'); |
| 911 | $this->bogusComment('<![CDATA[' . $cdata); |
| 912 | |
| 913 | return true; |
| 914 | } |
| 915 | $cdata .= $tok; |
| 916 | $tok = $this->scanner->next(); |
| 917 | } while (!$this->scanner->sequenceMatches(']]>')); |
| 918 | |
| 919 | // Consume ]]> |
| 920 | $this->scanner->consume(3); |
| 921 | |
| 922 | $this->events->cdata($cdata); |
| 923 | |
| 924 | return true; |
| 925 | } |
| 926 | |
| 927 | // ================================================================ |
| 928 | // Non-HTML5 |
| 929 | // ================================================================ |
| 930 | |
| 931 | /** |
| 932 | * Handle a processing instruction. |
| 933 | * |
| 934 | * XML processing instructions are supposed to be ignored in HTML5, |
| 935 | * treated as "bogus comments". However, since we're not a user |
| 936 | * agent, we allow them. We consume until ?> and then issue a |
| 937 | * EventListener::processingInstruction() event. |
| 938 | * |
| 939 | * @return bool |
| 940 | */ |
| 941 | protected function processingInstruction() |
| 942 | { |
| 943 | if ('?' != $this->scanner->current()) { |
| 944 | return false; |
| 945 | } |
| 946 | |
| 947 | $tok = $this->scanner->next(); |
| 948 | $procName = $this->scanner->getAsciiAlpha(); |
| 949 | $white = $this->scanner->whitespace(); |
| 950 | |
| 951 | // If not a PI, send to bogusComment. |
| 952 | if (0 == strlen($procName) || 0 == $white || false == $this->scanner->current()) { |
| 953 | $this->parseError("Expected processing instruction name, got $tok"); |
| 954 | $this->bogusComment('<?' . $tok . $procName); |
| 955 | |
| 956 | return true; |
| 957 | } |
| 958 | |
| 959 | $data = ''; |
| 960 | // As long as it's not the case that the next two chars are ? and >. |
| 961 | while (!('?' == $this->scanner->current() && '>' == $this->scanner->peek())) { |
| 962 | $data .= $this->scanner->current(); |
| 963 | |
| 964 | $tok = $this->scanner->next(); |
| 965 | if (false === $tok) { |
| 966 | $this->parseError('Unexpected EOF in processing instruction.'); |
| 967 | $this->events->processingInstruction($procName, $data); |
| 968 | |
| 969 | return true; |
| 970 | } |
| 971 | } |
| 972 | |
| 973 | $this->scanner->consume(2); // Consume the closing tag |
| 974 | $this->events->processingInstruction($procName, $data); |
| 975 | |
| 976 | return true; |
| 977 | } |
| 978 | |
| 979 | // ================================================================ |
| 980 | // UTILITY FUNCTIONS |
| 981 | // ================================================================ |
| 982 | |
| 983 | /** |
| 984 | * Read from the input stream until we get to the desired sequene |
| 985 | * or hit the end of the input stream. |
| 986 | * |
| 987 | * @param string $sequence |
| 988 | * |
| 989 | * @return string |
| 990 | */ |
| 991 | protected function readUntilSequence($sequence) |
| 992 | { |
| 993 | $buffer = ''; |
| 994 | |
| 995 | // Optimization for reading larger blocks faster. |
| 996 | $first = substr($sequence, 0, 1); |
| 997 | while (false !== $this->scanner->current()) { |
| 998 | $buffer .= $this->scanner->charsUntil($first); |
| 999 | |
| 1000 | // Stop as soon as we hit the stopping condition. |
| 1001 | if ($this->scanner->sequenceMatches($sequence, false)) { |
| 1002 | return $buffer; |
| 1003 | } |
| 1004 | $buffer .= $this->scanner->current(); |
| 1005 | $this->scanner->consume(); |
| 1006 | } |
| 1007 | |
| 1008 | // If we get here, we hit the EOF. |
| 1009 | $this->parseError('Unexpected EOF during text read.'); |
| 1010 | |
| 1011 | return $buffer; |
| 1012 | } |
| 1013 | |
| 1014 | /** |
| 1015 | * Check if upcomming chars match the given sequence. |
| 1016 | * |
| 1017 | * This will read the stream for the $sequence. If it's |
| 1018 | * found, this will return true. If not, return false. |
| 1019 | * Since this unconsumes any chars it reads, the caller |
| 1020 | * will still need to read the next sequence, even if |
| 1021 | * this returns true. |
| 1022 | * |
| 1023 | * Example: $this->scanner->sequenceMatches('</script>') will |
| 1024 | * see if the input stream is at the start of a |
| 1025 | * '</script>' string. |
| 1026 | * |
| 1027 | * @param string $sequence |
| 1028 | * @param bool $caseSensitive |
| 1029 | * |
| 1030 | * @return bool |
| 1031 | */ |
| 1032 | protected function sequenceMatches($sequence, $caseSensitive = true) |
| 1033 | { |
| 1034 | @trigger_error(__METHOD__ . ' method is deprecated since version 2.4 and will be removed in 3.0. Use Scanner::sequenceMatches() instead.', E_USER_DEPRECATED); |
| 1035 | |
| 1036 | return $this->scanner->sequenceMatches($sequence, $caseSensitive); |
| 1037 | } |
| 1038 | |
| 1039 | /** |
| 1040 | * Send a TEXT event with the contents of the text buffer. |
| 1041 | * |
| 1042 | * This emits an EventHandler::text() event with the current contents of the |
| 1043 | * temporary text buffer. (The buffer is used to group as much PCDATA |
| 1044 | * as we can instead of emitting lots and lots of TEXT events.) |
| 1045 | */ |
| 1046 | protected function flushBuffer() |
| 1047 | { |
| 1048 | if ('' === $this->text) { |
| 1049 | return; |
| 1050 | } |
| 1051 | $this->events->text($this->text); |
| 1052 | $this->text = ''; |
| 1053 | } |
| 1054 | |
| 1055 | /** |
| 1056 | * Add text to the temporary buffer. |
| 1057 | * |
| 1058 | * @see flushBuffer() |
| 1059 | * |
| 1060 | * @param string $str |
| 1061 | */ |
| 1062 | protected function buffer($str) |
| 1063 | { |
| 1064 | $this->text .= $str; |
| 1065 | } |
| 1066 | |
| 1067 | /** |
| 1068 | * Emit a parse error. |
| 1069 | * |
| 1070 | * A parse error always returns false because it never consumes any |
| 1071 | * characters. |
| 1072 | * |
| 1073 | * @param string $msg |
| 1074 | * |
| 1075 | * @return string |
| 1076 | */ |
| 1077 | protected function parseError($msg) |
| 1078 | { |
| 1079 | $args = func_get_args(); |
| 1080 | |
| 1081 | if (count($args) > 1) { |
| 1082 | array_shift($args); |
| 1083 | $msg = vsprintf($msg, $args); |
| 1084 | } |
| 1085 | |
| 1086 | $line = $this->scanner->currentLine(); |
| 1087 | $col = $this->scanner->columnOffset(); |
| 1088 | $this->events->parseError($msg, $line, $col); |
| 1089 | |
| 1090 | return false; |
| 1091 | } |
| 1092 | |
| 1093 | /** |
| 1094 | * Decode a character reference and return the string. |
| 1095 | * |
| 1096 | * If $inAttribute is set to true, a bare & will be returned as-is. |
| 1097 | * |
| 1098 | * @param bool $inAttribute Set to true if the text is inside of an attribute value. |
| 1099 | * false otherwise. |
| 1100 | * |
| 1101 | * @return string |
| 1102 | */ |
| 1103 | protected function decodeCharacterReference($inAttribute = false) |
| 1104 | { |
| 1105 | // Next char after &. |
| 1106 | $tok = $this->scanner->next(); |
| 1107 | $start = $this->scanner->position(); |
| 1108 | |
| 1109 | if (false === $tok) { |
| 1110 | return '&'; |
| 1111 | } |
| 1112 | |
| 1113 | // These indicate not an entity. We return just |
| 1114 | // the &. |
| 1115 | if ("\t" === $tok || "\n" === $tok || "\f" === $tok || ' ' === $tok || '&' === $tok || '<' === $tok) { |
| 1116 | // $this->scanner->next(); |
| 1117 | return '&'; |
| 1118 | } |
| 1119 | |
| 1120 | // Numeric entity |
| 1121 | if ('#' === $tok) { |
| 1122 | $tok = $this->scanner->next(); |
| 1123 | |
| 1124 | if (false === $tok) { |
| 1125 | $this->parseError('Expected &#DEC; &#HEX;, got EOF'); |
| 1126 | $this->scanner->unconsume(1); |
| 1127 | |
| 1128 | return '&'; |
| 1129 | } |
| 1130 | |
| 1131 | // Hexadecimal encoding. |
| 1132 | // X[0-9a-fA-F]+; |
| 1133 | // x[0-9a-fA-F]+; |
| 1134 | if ('x' === $tok || 'X' === $tok) { |
| 1135 | $tok = $this->scanner->next(); // Consume x |
| 1136 | |
| 1137 | // Convert from hex code to char. |
| 1138 | $hex = $this->scanner->getHex(); |
| 1139 | if (empty($hex)) { |
| 1140 | $this->parseError('Expected &#xHEX;, got &#x%s', $tok); |
| 1141 | // We unconsume because we don't know what parser rules might |
| 1142 | // be in effect for the remaining chars. For example. '&#>' |
| 1143 | // might result in a specific parsing rule inside of tag |
| 1144 | // contexts, while not inside of pcdata context. |
| 1145 | $this->scanner->unconsume(2); |
| 1146 | |
| 1147 | return '&'; |
| 1148 | } |
| 1149 | $entity = CharacterReference::lookupHex($hex); |
| 1150 | } // Decimal encoding. |
| 1151 | // [0-9]+; |
| 1152 | else { |
| 1153 | // Convert from decimal to char. |
| 1154 | $numeric = $this->scanner->getNumeric(); |
| 1155 | if (false === $numeric) { |
| 1156 | $this->parseError('Expected &#DIGITS;, got &#%s', $tok); |
| 1157 | $this->scanner->unconsume(2); |
| 1158 | |
| 1159 | return '&'; |
| 1160 | } |
| 1161 | $entity = CharacterReference::lookupDecimal($numeric); |
| 1162 | } |
| 1163 | } elseif ('=' === $tok && $inAttribute) { |
| 1164 | return '&'; |
| 1165 | } else { // String entity. |
| 1166 | // Attempt to consume a string up to a ';'. |
| 1167 | // [a-zA-Z0-9]+; |
| 1168 | $cname = $this->scanner->getAsciiAlphaNum(); |
| 1169 | $entity = CharacterReference::lookupName($cname); |
| 1170 | |
| 1171 | // When no entity is found provide the name of the unmatched string |
| 1172 | // and continue on as the & is not part of an entity. The & will |
| 1173 | // be converted to & elsewhere. |
| 1174 | if (null === $entity) { |
| 1175 | if (!$inAttribute || '' === $cname) { |
| 1176 | $this->parseError("No match in entity table for '%s'", $cname); |
| 1177 | } |
| 1178 | $this->scanner->unconsume($this->scanner->position() - $start); |
| 1179 | |
| 1180 | return '&'; |
| 1181 | } |
| 1182 | } |
| 1183 | |
| 1184 | // The scanner has advanced the cursor for us. |
| 1185 | $tok = $this->scanner->current(); |
| 1186 | |
| 1187 | // We have an entity. We're done here. |
| 1188 | if (';' === $tok) { |
| 1189 | $this->scanner->consume(); |
| 1190 | |
| 1191 | return $entity; |
| 1192 | } |
| 1193 | |
| 1194 | // Failing to match ; means unconsume the entire string. |
| 1195 | $this->scanner->unconsume($this->scanner->position() - $start); |
| 1196 | |
| 1197 | $this->parseError('Expected &ENTITY;, got &ENTITY%s (no trailing ;) ', $tok); |
| 1198 | |
| 1199 | return '&'; |
| 1200 | } |
| 1201 | |
| 1202 | /** |
| 1203 | * Checks whether a (single-byte) character is an ASCII letter or not. |
| 1204 | * |
| 1205 | * @param string $input A single-byte string |
| 1206 | * |
| 1207 | * @return bool True if it is a letter, False otherwise |
| 1208 | */ |
| 1209 | protected function is_alpha($input) |
| 1210 | { |
| 1211 | $code = ord($input); |
| 1212 | |
| 1213 | return ($code >= 97 && $code <= 122) || ($code >= 65 && $code <= 90); |
| 1214 | } |
| 1215 | } |
| 1216 |