Complex classes like Parser often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes. You can also have a look at the cohesion graph to spot any un-connected, or weakly-connected components.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use Parser, and based on these observations, apply Extract Interface, too.
1 | <?php |
||
15 | class Parser |
||
16 | { |
||
17 | /**#@+ |
||
18 | * Boolean rules bitfield |
||
19 | */ |
||
20 | const RULE_AUTO_CLOSE = 1 << 0; |
||
21 | const RULE_AUTO_REOPEN = 1 << 1; |
||
22 | const RULE_BREAK_PARAGRAPH = 1 << 2; |
||
23 | const RULE_CREATE_PARAGRAPHS = 1 << 3; |
||
24 | const RULE_DISABLE_AUTO_BR = 1 << 4; |
||
25 | const RULE_ENABLE_AUTO_BR = 1 << 5; |
||
26 | const RULE_IGNORE_TAGS = 1 << 6; |
||
27 | const RULE_IGNORE_TEXT = 1 << 7; |
||
28 | const RULE_IGNORE_WHITESPACE = 1 << 8; |
||
29 | const RULE_IS_TRANSPARENT = 1 << 9; |
||
30 | const RULE_PREVENT_BR = 1 << 10; |
||
31 | const RULE_SUSPEND_AUTO_BR = 1 << 11; |
||
32 | const RULE_TRIM_FIRST_LINE = 1 << 12; |
||
33 | /**#@-*/ |
||
34 | |||
35 | /** |
||
36 | * Bitwise disjunction of rules related to automatic line breaks |
||
37 | */ |
||
38 | const RULES_AUTO_LINEBREAKS = self::RULE_DISABLE_AUTO_BR | self::RULE_ENABLE_AUTO_BR | self::RULE_SUSPEND_AUTO_BR; |
||
39 | |||
40 | /** |
||
41 | * Bitwise disjunction of rules that are inherited by subcontexts |
||
42 | */ |
||
43 | const RULES_INHERITANCE = self::RULE_ENABLE_AUTO_BR; |
||
44 | |||
45 | /** |
||
46 | * All the characters that are considered whitespace |
||
47 | */ |
||
48 | const WHITESPACE = " \n\t"; |
||
49 | |||
50 | /** |
||
51 | * @var array Number of open tags for each tag name |
||
52 | */ |
||
53 | protected $cntOpen; |
||
54 | |||
55 | /** |
||
56 | * @var array Number of times each tag has been used |
||
57 | */ |
||
58 | protected $cntTotal; |
||
59 | |||
60 | /** |
||
61 | * @var array Current context |
||
62 | */ |
||
63 | protected $context; |
||
64 | |||
65 | /** |
||
66 | * @var integer How hard the parser has worked on fixing bad markup so far |
||
67 | */ |
||
68 | protected $currentFixingCost; |
||
69 | |||
70 | /** |
||
71 | * @var Tag Current tag being processed |
||
72 | */ |
||
73 | protected $currentTag; |
||
74 | |||
75 | /** |
||
76 | * @var bool Whether the output contains "rich" tags, IOW any tag that is not <p> or <br/> |
||
77 | */ |
||
78 | protected $isRich; |
||
79 | |||
80 | /** |
||
81 | * @var Logger This parser's logger |
||
82 | */ |
||
83 | protected $logger; |
||
84 | |||
85 | /** |
||
86 | * @var integer How hard the parser should work on fixing bad markup |
||
87 | */ |
||
88 | public $maxFixingCost = 10000; |
||
89 | |||
90 | /** |
||
91 | * @var array Associative array of namespace prefixes in use in document (prefixes used as key) |
||
92 | */ |
||
93 | protected $namespaces; |
||
94 | |||
95 | /** |
||
96 | * @var array Stack of open tags (instances of Tag) |
||
97 | */ |
||
98 | protected $openTags; |
||
99 | |||
100 | /** |
||
101 | * @var string This parser's output |
||
102 | */ |
||
103 | protected $output; |
||
104 | |||
105 | /** |
||
106 | * @var integer Position of the cursor in the original text |
||
107 | */ |
||
108 | protected $pos; |
||
109 | |||
110 | /** |
||
111 | * @var array Array of callbacks, using plugin names as keys |
||
112 | */ |
||
113 | protected $pluginParsers = []; |
||
114 | |||
115 | /** |
||
116 | * @var array Associative array of [pluginName => pluginConfig] |
||
117 | */ |
||
118 | protected $pluginsConfig; |
||
119 | |||
120 | /** |
||
121 | * @var array Variables registered for use in filters |
||
122 | */ |
||
123 | public $registeredVars = []; |
||
124 | |||
125 | /** |
||
126 | * @var array Root context, used at the root of the document |
||
127 | */ |
||
128 | protected $rootContext; |
||
129 | |||
130 | /** |
||
131 | * @var array Tags' config |
||
132 | */ |
||
133 | protected $tagsConfig; |
||
134 | |||
135 | /** |
||
136 | * @var array Tag storage |
||
137 | */ |
||
138 | protected $tagStack; |
||
139 | |||
140 | /** |
||
141 | * @var bool Whether the tags in the stack are sorted |
||
142 | */ |
||
143 | protected $tagStackIsSorted; |
||
144 | |||
145 | /** |
||
146 | * @var string Text being parsed |
||
147 | */ |
||
148 | protected $text; |
||
149 | |||
150 | /** |
||
151 | * @var integer Length of the text being parsed |
||
152 | */ |
||
153 | protected $textLen; |
||
154 | |||
155 | /** |
||
156 | * @var integer Counter incremented everytime the parser is reset. Used to as a canary to detect |
||
157 | * whether the parser was reset during execution |
||
158 | */ |
||
159 | protected $uid = 0; |
||
160 | |||
161 | /** |
||
162 | * @var integer Position before which we output text verbatim, without paragraphs or linebreaks |
||
163 | */ |
||
164 | protected $wsPos; |
||
165 | |||
166 | /** |
||
167 | * Constructor |
||
168 | */ |
||
169 | 86 | public function __construct(array $config) |
|
178 | |||
179 | /** |
||
180 | * Serializer |
||
181 | * |
||
182 | * Returns the properties that need to persist through serialization. |
||
183 | * |
||
184 | * NOTE: using __sleep() is preferable to implementing Serializable because it leaves the choice |
||
185 | * of the serializer to the user (e.g. igbinary) |
||
186 | * |
||
187 | * @return array |
||
188 | */ |
||
189 | public function __sleep() |
||
190 | { |
||
191 | return ['pluginsConfig', 'registeredVars', 'rootContext', 'tagsConfig']; |
||
192 | } |
||
193 | |||
194 | /** |
||
195 | * Unserializer |
||
196 | * |
||
197 | * @return void |
||
198 | */ |
||
199 | 86 | public function __wakeup() |
|
203 | |||
204 | /** |
||
205 | * Reset the parser for a new parsing |
||
206 | * |
||
207 | * @param string $text Text to be parsed |
||
208 | * @return void |
||
209 | */ |
||
210 | 86 | protected function reset($text) |
|
242 | |||
243 | /** |
||
244 | * Set a tag's option |
||
245 | * |
||
246 | * This method ensures that the tag's config is a value and not a reference, to prevent |
||
247 | * potential side-effects. References contained *inside* the tag's config are left untouched |
||
248 | * |
||
249 | * @param string $tagName Tag's name |
||
250 | * @param string $optionName Option's name |
||
251 | * @param mixed $optionValue Option's value |
||
252 | * @return void |
||
253 | */ |
||
254 | protected function setTagOption($tagName, $optionName, $optionValue) |
||
255 | { |
||
256 | if (isset($this->tagsConfig[$tagName])) |
||
257 | { |
||
258 | // Copy the tag's config and remove it. That will destroy the reference |
||
259 | $tagConfig = $this->tagsConfig[$tagName]; |
||
260 | unset($this->tagsConfig[$tagName]); |
||
261 | |||
262 | // Set the new value and replace the tag's config |
||
263 | $tagConfig[$optionName] = $optionValue; |
||
264 | $this->tagsConfig[$tagName] = $tagConfig; |
||
265 | } |
||
266 | } |
||
267 | |||
268 | //========================================================================== |
||
269 | // Public API |
||
270 | //========================================================================== |
||
271 | |||
272 | /** |
||
273 | * Disable a tag |
||
274 | * |
||
275 | * @param string $tagName Name of the tag |
||
276 | * @return void |
||
277 | */ |
||
278 | public function disableTag($tagName) |
||
279 | { |
||
280 | $this->setTagOption($tagName, 'isDisabled', true); |
||
281 | } |
||
282 | |||
283 | /** |
||
284 | * Enable a tag |
||
285 | * |
||
286 | * @param string $tagName Name of the tag |
||
287 | * @return void |
||
288 | */ |
||
289 | public function enableTag($tagName) |
||
290 | { |
||
291 | if (isset($this->tagsConfig[$tagName])) |
||
292 | { |
||
293 | unset($this->tagsConfig[$tagName]['isDisabled']); |
||
294 | } |
||
295 | } |
||
296 | |||
297 | /** |
||
298 | * Get this parser's Logger instance |
||
299 | * |
||
300 | * @return Logger |
||
301 | */ |
||
302 | 6 | public function getLogger() |
|
306 | |||
307 | /** |
||
308 | * Return the last text parsed |
||
309 | * |
||
310 | * This method returns the normalized text, which may be slightly different from the original |
||
311 | * text in that EOLs are normalized to LF and other control codes are stripped. This method is |
||
312 | * meant to be used in support of processing log entries, which contain offsets based on the |
||
313 | * normalized text |
||
314 | * |
||
315 | * @see Parser::reset() |
||
316 | * |
||
317 | * @return string |
||
318 | */ |
||
319 | public function getText() |
||
320 | { |
||
321 | return $this->text; |
||
322 | } |
||
323 | |||
324 | /** |
||
325 | * Parse a text |
||
326 | * |
||
327 | * @param string $text Text to parse |
||
328 | * @return string XML representation |
||
329 | */ |
||
330 | 86 | public function parse($text) |
|
331 | { |
||
332 | // Reset the parser and save the uid |
||
333 | 86 | $this->reset($text); |
|
334 | 86 | $uid = $this->uid; |
|
335 | |||
336 | // Do the heavy lifting |
||
337 | 86 | $this->executePluginParsers(); |
|
338 | 86 | $this->processTags(); |
|
339 | |||
340 | // Finalize the document |
||
341 | 86 | $this->finalizeOutput(); |
|
342 | |||
343 | // Check the uid in case a plugin or a filter reset the parser mid-execution |
||
344 | 86 | if ($this->uid !== $uid) |
|
345 | { |
||
346 | throw new RuntimeException('The parser has been reset during execution'); |
||
347 | } |
||
348 | |||
349 | // Log a warning if the fixing cost limit was exceeded |
||
350 | 86 | if ($this->currentFixingCost > $this->maxFixingCost) |
|
351 | { |
||
352 | 1 | $this->logger->warn('Fixing cost limit exceeded'); |
|
353 | } |
||
354 | |||
355 | 86 | return $this->output; |
|
356 | } |
||
357 | |||
358 | /** |
||
359 | * Change a tag's tagLimit |
||
360 | * |
||
361 | * NOTE: the default tagLimit should generally be set during configuration instead |
||
362 | * |
||
363 | * @param string $tagName The tag's name, in UPPERCASE |
||
364 | * @param integer $tagLimit |
||
365 | * @return void |
||
366 | */ |
||
367 | public function setTagLimit($tagName, $tagLimit) |
||
368 | { |
||
369 | $this->setTagOption($tagName, 'tagLimit', $tagLimit); |
||
370 | } |
||
371 | |||
372 | /** |
||
373 | * Change a tag's nestingLimit |
||
374 | * |
||
375 | * NOTE: the default nestingLimit should generally be set during configuration instead |
||
376 | * |
||
377 | * @param string $tagName The tag's name, in UPPERCASE |
||
378 | * @param integer $nestingLimit |
||
379 | * @return void |
||
380 | */ |
||
381 | public function setNestingLimit($tagName, $nestingLimit) |
||
382 | { |
||
383 | $this->setTagOption($tagName, 'nestingLimit', $nestingLimit); |
||
384 | } |
||
385 | |||
386 | //========================================================================== |
||
387 | // Filter processing |
||
388 | //========================================================================== |
||
389 | |||
390 | /** |
||
391 | * Execute all the attribute preprocessors of given tag |
||
392 | * |
||
393 | * @private |
||
394 | * |
||
395 | * @param Tag $tag Source tag |
||
396 | * @param array $tagConfig Tag's config |
||
397 | * @return bool Unconditionally TRUE |
||
398 | */ |
||
399 | 6 | public static function executeAttributePreprocessors(Tag $tag, array $tagConfig) |
|
400 | { |
||
401 | 6 | if (!empty($tagConfig['attributePreprocessors'])) |
|
402 | { |
||
403 | 6 | foreach ($tagConfig['attributePreprocessors'] as list($attrName, $regexp, $map)) |
|
404 | { |
||
405 | 6 | if (!$tag->hasAttribute($attrName)) |
|
406 | { |
||
407 | 1 | continue; |
|
408 | } |
||
409 | |||
410 | 5 | self::executeAttributePreprocessor($tag, $attrName, $regexp, $map); |
|
411 | } |
||
412 | } |
||
413 | |||
414 | 6 | return true; |
|
415 | } |
||
416 | |||
417 | /** |
||
418 | * Execute an attribute preprocessor |
||
419 | * |
||
420 | * @param Tag $tag |
||
421 | * @param string $attrName |
||
422 | * @param string $regexp |
||
423 | * @param string[] $map |
||
424 | * @return void |
||
425 | */ |
||
426 | 5 | protected static function executeAttributePreprocessor(Tag $tag, $attrName, $regexp, $map) |
|
427 | { |
||
428 | 5 | $attrValue = $tag->getAttribute($attrName); |
|
429 | 5 | $captures = self::getNamedCaptures($attrValue, $regexp, $map); |
|
430 | 5 | foreach ($captures as $k => $v) |
|
431 | { |
||
432 | // Attribute preprocessors cannot overwrite other attributes but they can |
||
433 | // overwrite themselves |
||
434 | 4 | if ($k === $attrName || !$tag->hasAttribute($k)) |
|
435 | { |
||
436 | 4 | $tag->setAttribute($k, $v); |
|
437 | } |
||
438 | } |
||
439 | 5 | } |
|
440 | |||
441 | /** |
||
442 | * Execute a regexp and return the values of the mapped captures |
||
443 | * |
||
444 | * @param string $attrValue |
||
445 | * @param string $regexp |
||
446 | * @param string[] $map |
||
447 | * @return array |
||
448 | */ |
||
449 | 5 | protected static function getNamedCaptures($attrValue, $regexp, $map) |
|
450 | { |
||
451 | 5 | if (!preg_match($regexp, $attrValue, $m)) |
|
452 | { |
||
453 | 1 | return []; |
|
454 | } |
||
455 | |||
456 | 4 | $values = []; |
|
457 | 4 | foreach ($map as $i => $k) |
|
458 | { |
||
459 | 4 | if (isset($m[$i]) && $m[$i] !== '') |
|
460 | { |
||
461 | 4 | $values[$k] = $m[$i]; |
|
462 | } |
||
463 | } |
||
464 | |||
465 | 4 | return $values; |
|
466 | } |
||
467 | |||
468 | /** |
||
469 | * Execute a filter |
||
470 | * |
||
471 | * @see s9e\TextFormatter\Configurator\Items\ProgrammableCallback |
||
472 | * |
||
473 | * @param array $filter Programmed callback |
||
474 | * @param array $vars Variables to be used when executing the callback |
||
475 | * @return mixed Whatever the callback returns |
||
476 | */ |
||
477 | 82 | protected static function executeFilter(array $filter, array $vars) |
|
478 | { |
||
479 | 82 | $callback = $filter['callback']; |
|
480 | 82 | $params = (isset($filter['params'])) ? $filter['params'] : []; |
|
481 | |||
482 | 82 | $args = []; |
|
483 | 82 | foreach ($params as $k => $v) |
|
484 | { |
||
485 | 81 | if (is_numeric($k)) |
|
486 | { |
||
487 | // By-value param |
||
488 | 5 | $args[] = $v; |
|
489 | } |
||
490 | 80 | elseif (isset($vars[$k])) |
|
491 | { |
||
492 | // By-name param using a supplied var |
||
493 | 78 | $args[] = $vars[$k]; |
|
494 | } |
||
495 | 2 | elseif (isset($vars['registeredVars'][$k])) |
|
496 | { |
||
497 | // By-name param using a registered var |
||
498 | 1 | $args[] = $vars['registeredVars'][$k]; |
|
499 | } |
||
500 | else |
||
501 | { |
||
502 | // Unknown param |
||
503 | 81 | $args[] = null; |
|
504 | } |
||
505 | } |
||
506 | |||
507 | 82 | return call_user_func_array($callback, $args); |
|
508 | } |
||
509 | |||
510 | /** |
||
511 | * Filter the attributes of given tag |
||
512 | * |
||
513 | * @private |
||
514 | * |
||
515 | * @param Tag $tag Tag being checked |
||
516 | * @param array $tagConfig Tag's config |
||
517 | * @param array $registeredVars Array of registered vars for use in attribute filters |
||
518 | * @param Logger $logger This parser's Logger instance |
||
519 | * @return bool Whether the whole attribute set is valid |
||
520 | */ |
||
521 | 81 | public static function filterAttributes(Tag $tag, array $tagConfig, array $registeredVars, Logger $logger) |
|
522 | { |
||
523 | 81 | if (empty($tagConfig['attributes'])) |
|
524 | { |
||
525 | 71 | $tag->setAttributes([]); |
|
526 | |||
527 | 71 | return true; |
|
528 | } |
||
529 | |||
530 | // Generate values for attributes with a generator set |
||
531 | 12 | foreach ($tagConfig['attributes'] as $attrName => $attrConfig) |
|
532 | { |
||
533 | 12 | if (isset($attrConfig['generator'])) |
|
534 | { |
||
535 | 1 | $tag->setAttribute( |
|
536 | 1 | $attrName, |
|
537 | 1 | self::executeFilter( |
|
538 | 1 | $attrConfig['generator'], |
|
539 | [ |
||
540 | 1 | 'attrName' => $attrName, |
|
541 | 1 | 'logger' => $logger, |
|
542 | 12 | 'registeredVars' => $registeredVars |
|
543 | ] |
||
544 | ) |
||
545 | ); |
||
546 | } |
||
547 | } |
||
548 | |||
549 | // Filter and remove invalid attributes |
||
550 | 12 | foreach ($tag->getAttributes() as $attrName => $attrValue) |
|
551 | { |
||
552 | // Test whether this attribute exists and remove it if it doesn't |
||
553 | 11 | if (!isset($tagConfig['attributes'][$attrName])) |
|
554 | { |
||
555 | 1 | $tag->removeAttribute($attrName); |
|
556 | 1 | continue; |
|
557 | } |
||
558 | |||
559 | 11 | $attrConfig = $tagConfig['attributes'][$attrName]; |
|
560 | |||
561 | // Test whether this attribute has a filterChain |
||
562 | 11 | if (!isset($attrConfig['filterChain'])) |
|
563 | { |
||
564 | 6 | continue; |
|
565 | } |
||
566 | |||
567 | // Record the name of the attribute being filtered into the logger |
||
568 | 5 | $logger->setAttribute($attrName); |
|
569 | |||
570 | 5 | foreach ($attrConfig['filterChain'] as $filter) |
|
571 | { |
||
572 | 5 | $attrValue = self::executeFilter( |
|
573 | 5 | $filter, |
|
574 | [ |
||
575 | 5 | 'attrName' => $attrName, |
|
576 | 5 | 'attrValue' => $attrValue, |
|
577 | 5 | 'logger' => $logger, |
|
578 | 5 | 'registeredVars' => $registeredVars |
|
579 | ] |
||
580 | ); |
||
581 | |||
582 | 5 | if ($attrValue === false) |
|
583 | { |
||
584 | 3 | $tag->removeAttribute($attrName); |
|
585 | 5 | break; |
|
586 | } |
||
587 | } |
||
588 | |||
589 | // Update the attribute value if it's valid |
||
590 | 5 | if ($attrValue !== false) |
|
591 | { |
||
592 | 2 | $tag->setAttribute($attrName, $attrValue); |
|
593 | } |
||
594 | |||
595 | // Remove the attribute's name from the logger |
||
596 | 5 | $logger->unsetAttribute(); |
|
597 | } |
||
598 | |||
599 | // Iterate over the attribute definitions to handle missing attributes |
||
600 | 12 | foreach ($tagConfig['attributes'] as $attrName => $attrConfig) |
|
601 | { |
||
602 | // Test whether this attribute is missing |
||
603 | 12 | if (!$tag->hasAttribute($attrName)) |
|
604 | { |
||
605 | 5 | if (isset($attrConfig['defaultValue'])) |
|
606 | { |
||
607 | // Use the attribute's default value |
||
608 | 2 | $tag->setAttribute($attrName, $attrConfig['defaultValue']); |
|
609 | } |
||
610 | 3 | elseif (!empty($attrConfig['required'])) |
|
611 | { |
||
612 | // This attribute is missing, has no default value and is required, which means |
||
613 | // the attribute set is invalid |
||
614 | 12 | return false; |
|
615 | } |
||
616 | } |
||
617 | } |
||
618 | |||
619 | 9 | return true; |
|
620 | } |
||
621 | |||
622 | /** |
||
623 | * Execute given tag's filterChain |
||
624 | * |
||
625 | * @param Tag $tag Tag to filter |
||
626 | * @return bool Whether the tag is valid |
||
627 | */ |
||
628 | 73 | protected function filterTag(Tag $tag) |
|
629 | { |
||
630 | 73 | $tagName = $tag->getName(); |
|
631 | 73 | $tagConfig = $this->tagsConfig[$tagName]; |
|
632 | 73 | $isValid = true; |
|
633 | |||
634 | 73 | if (!empty($tagConfig['filterChain'])) |
|
635 | { |
||
636 | // Record the tag being processed into the logger it can be added to the context of |
||
637 | // messages logged during the execution |
||
638 | 72 | $this->logger->setTag($tag); |
|
639 | |||
640 | // Prepare the variables that are accessible to filters |
||
641 | $vars = [ |
||
642 | 72 | 'logger' => $this->logger, |
|
643 | 72 | 'openTags' => $this->openTags, |
|
644 | 72 | 'parser' => $this, |
|
645 | 72 | 'registeredVars' => $this->registeredVars, |
|
646 | 72 | 'tag' => $tag, |
|
647 | 72 | 'tagConfig' => $tagConfig, |
|
648 | 72 | 'text' => $this->text |
|
649 | ]; |
||
650 | |||
651 | 72 | foreach ($tagConfig['filterChain'] as $filter) |
|
652 | { |
||
653 | 72 | if (!self::executeFilter($filter, $vars)) |
|
654 | { |
||
655 | 1 | $isValid = false; |
|
656 | 72 | break; |
|
657 | } |
||
658 | } |
||
659 | |||
660 | // Remove the tag from the logger |
||
661 | 72 | $this->logger->unsetTag(); |
|
662 | } |
||
663 | |||
664 | 73 | return $isValid; |
|
665 | } |
||
666 | |||
667 | //========================================================================== |
||
668 | // Output handling |
||
669 | //========================================================================== |
||
670 | |||
671 | /** |
||
672 | * Finalize the output by appending the rest of the unprocessed text and create the root node |
||
673 | * |
||
674 | * @return void |
||
675 | */ |
||
676 | 86 | protected function finalizeOutput() |
|
677 | { |
||
678 | // Output the rest of the text and close the last paragraph |
||
679 | 86 | $this->outputText($this->textLen, 0, true); |
|
680 | |||
681 | // Remove empty tag pairs, e.g. <I><U></U></I> as well as empty paragraphs |
||
682 | do |
||
683 | { |
||
684 | 86 | $this->output = preg_replace('(<([^ />]++)[^>]*></\\1>)', '', $this->output, -1, $cnt); |
|
685 | } |
||
686 | 86 | while ($cnt > 0); |
|
687 | |||
688 | // Merge consecutive <i> tags |
||
689 | 86 | if (strpos($this->output, '</i><i>') !== false) |
|
690 | { |
||
691 | $this->output = str_replace('</i><i>', '', $this->output); |
||
692 | } |
||
693 | |||
694 | // Encode Unicode characters that are outside of the BMP |
||
695 | 86 | $this->output = Utils::encodeUnicodeSupplementaryCharacters($this->output); |
|
696 | |||
697 | // Use a <r> root if the text is rich, or <t> for plain text (including <p></p> and <br/>) |
||
698 | 86 | $tagName = ($this->isRich) ? 'r' : 't'; |
|
699 | |||
700 | // Prepare the root node with all the namespace declarations |
||
701 | 86 | $tmp = '<' . $tagName; |
|
702 | 86 | foreach (array_keys($this->namespaces) as $prefix) |
|
703 | { |
||
704 | $tmp .= ' xmlns:' . $prefix . '="urn:s9e:TextFormatter:' . $prefix . '"'; |
||
705 | } |
||
706 | |||
707 | 86 | $this->output = $tmp . '>' . $this->output . '</' . $tagName . '>'; |
|
708 | 86 | } |
|
709 | |||
710 | /** |
||
711 | * Append a tag to the output |
||
712 | * |
||
713 | * @param Tag $tag Tag to append |
||
714 | * @return void |
||
715 | */ |
||
716 | 71 | protected function outputTag(Tag $tag) |
|
717 | { |
||
718 | 71 | $this->isRich = true; |
|
719 | |||
720 | 71 | $tagName = $tag->getName(); |
|
721 | 71 | $tagPos = $tag->getPos(); |
|
722 | 71 | $tagLen = $tag->getLen(); |
|
723 | 71 | $tagFlags = $tag->getFlags(); |
|
724 | |||
725 | 71 | if ($tagFlags & self::RULE_IGNORE_WHITESPACE) |
|
726 | { |
||
727 | 5 | $skipBefore = 1; |
|
728 | 5 | $skipAfter = ($tag->isEndTag()) ? 2 : 1; |
|
729 | } |
||
730 | else |
||
731 | { |
||
732 | 71 | $skipBefore = $skipAfter = 0; |
|
733 | } |
||
734 | |||
735 | // Current paragraph must end before the tag if: |
||
736 | // - the tag is a start (or self-closing) tag and it breaks paragraphs, or |
||
737 | // - the tag is an end tag (but not self-closing) |
||
738 | 71 | $closeParagraph = false; |
|
739 | 71 | if ($tag->isStartTag()) |
|
740 | { |
||
741 | 71 | if ($tagFlags & self::RULE_BREAK_PARAGRAPH) |
|
742 | { |
||
743 | 71 | $closeParagraph = true; |
|
744 | } |
||
745 | } |
||
746 | else |
||
747 | { |
||
748 | 59 | $closeParagraph = true; |
|
749 | } |
||
750 | |||
751 | // Let the cursor catch up with this tag's position |
||
752 | 71 | $this->outputText($tagPos, $skipBefore, $closeParagraph); |
|
753 | |||
754 | // Capture the text consumed by the tag |
||
755 | 71 | $tagText = ($tagLen) |
|
756 | 56 | ? htmlspecialchars(substr($this->text, $tagPos, $tagLen), ENT_NOQUOTES, 'UTF-8') |
|
757 | 71 | : ''; |
|
758 | |||
759 | // Output current tag |
||
760 | 71 | if ($tag->isStartTag()) |
|
761 | { |
||
762 | // Handle paragraphs before opening the tag |
||
763 | 71 | if (!($tagFlags & self::RULE_BREAK_PARAGRAPH)) |
|
764 | { |
||
765 | 71 | $this->outputParagraphStart($tagPos); |
|
766 | } |
||
767 | |||
768 | // Record this tag's namespace, if applicable |
||
769 | 71 | $colonPos = strpos($tagName, ':'); |
|
770 | 71 | if ($colonPos) |
|
771 | { |
||
772 | $this->namespaces[substr($tagName, 0, $colonPos)] = 0; |
||
773 | } |
||
774 | |||
775 | // Open the start tag and add its attributes, but don't close the tag |
||
776 | 71 | $this->output .= '<' . $tagName; |
|
777 | |||
778 | // We output the attributes in lexical order. Helps canonicalizing the output and could |
||
779 | // prove useful someday |
||
780 | 71 | $attributes = $tag->getAttributes(); |
|
781 | 71 | ksort($attributes); |
|
782 | |||
783 | 71 | foreach ($attributes as $attrName => $attrValue) |
|
784 | { |
||
785 | 4 | $this->output .= ' ' . $attrName . '="' . str_replace("\n", ' ', htmlspecialchars($attrValue, ENT_COMPAT, 'UTF-8')) . '"'; |
|
786 | } |
||
787 | |||
788 | 71 | if ($tag->isSelfClosingTag()) |
|
789 | { |
||
790 | 20 | if ($tagLen) |
|
791 | { |
||
792 | 19 | $this->output .= '>' . $tagText . '</' . $tagName . '>'; |
|
793 | } |
||
794 | else |
||
795 | { |
||
796 | 20 | $this->output .= '/>'; |
|
797 | } |
||
798 | } |
||
799 | 59 | elseif ($tagLen) |
|
800 | { |
||
801 | 42 | $this->output .= '><s>' . $tagText . '</s>'; |
|
802 | } |
||
803 | else |
||
804 | { |
||
805 | 71 | $this->output .= '>'; |
|
806 | } |
||
807 | } |
||
808 | else |
||
809 | { |
||
810 | 59 | if ($tagLen) |
|
811 | { |
||
812 | 39 | $this->output .= '<e>' . $tagText . '</e>'; |
|
813 | } |
||
814 | |||
815 | 59 | $this->output .= '</' . $tagName . '>'; |
|
816 | } |
||
817 | |||
818 | // Move the cursor past the tag |
||
819 | 71 | $this->pos = $tagPos + $tagLen; |
|
820 | |||
821 | // Skip newlines (no other whitespace) after this tag |
||
822 | 71 | $this->wsPos = $this->pos; |
|
823 | 71 | while ($skipAfter && $this->wsPos < $this->textLen && $this->text[$this->wsPos] === "\n") |
|
824 | { |
||
825 | // Decrement the number of lines to skip |
||
826 | 3 | --$skipAfter; |
|
827 | |||
828 | // Move the cursor past the newline |
||
829 | 3 | ++$this->wsPos; |
|
830 | } |
||
831 | 71 | } |
|
832 | |||
833 | /** |
||
834 | * Output the text between the cursor's position (included) and given position (not included) |
||
835 | * |
||
836 | * @param integer $catchupPos Position we're catching up to |
||
837 | * @param integer $maxLines Maximum number of lines to ignore at the end of the text |
||
838 | * @param bool $closeParagraph Whether to close the paragraph at the end, if applicable |
||
839 | * @return void |
||
840 | */ |
||
841 | 86 | protected function outputText($catchupPos, $maxLines, $closeParagraph) |
|
842 | { |
||
843 | 86 | if ($closeParagraph) |
|
844 | { |
||
845 | 86 | if (!($this->context['flags'] & self::RULE_CREATE_PARAGRAPHS)) |
|
846 | { |
||
847 | 84 | $closeParagraph = false; |
|
848 | } |
||
849 | else |
||
850 | { |
||
851 | // Ignore any number of lines at the end if we're closing a paragraph |
||
852 | 3 | $maxLines = -1; |
|
853 | } |
||
854 | } |
||
855 | |||
856 | 86 | if ($this->pos >= $catchupPos) |
|
857 | { |
||
858 | // We're already there, close the paragraph if applicable and return |
||
859 | 71 | if ($closeParagraph) |
|
860 | { |
||
861 | $this->outputParagraphEnd(); |
||
862 | } |
||
863 | |||
864 | 71 | return; |
|
865 | } |
||
866 | |||
867 | // Skip over previously identified whitespace if applicable |
||
868 | 76 | if ($this->wsPos > $this->pos) |
|
869 | { |
||
870 | 3 | $skipPos = min($catchupPos, $this->wsPos); |
|
871 | 3 | $this->output .= substr($this->text, $this->pos, $skipPos - $this->pos); |
|
872 | 3 | $this->pos = $skipPos; |
|
873 | |||
874 | 3 | if ($this->pos >= $catchupPos) |
|
875 | { |
||
876 | // Skipped everything. Close the paragraph if applicable and return |
||
877 | if ($closeParagraph) |
||
878 | { |
||
879 | $this->outputParagraphEnd(); |
||
880 | } |
||
881 | |||
882 | return; |
||
883 | } |
||
884 | } |
||
885 | |||
886 | // Test whether we're even supposed to output anything |
||
887 | 76 | if ($this->context['flags'] & self::RULE_IGNORE_TEXT) |
|
888 | { |
||
889 | $catchupLen = $catchupPos - $this->pos; |
||
890 | $catchupText = substr($this->text, $this->pos, $catchupLen); |
||
891 | |||
892 | // If the catchup text is not entirely composed of whitespace, we put it inside ignore |
||
893 | // tags |
||
894 | if (strspn($catchupText, " \n\t") < $catchupLen) |
||
895 | { |
||
896 | $catchupText = '<i>' . $catchupText . '</i>'; |
||
897 | } |
||
898 | |||
899 | $this->output .= $catchupText; |
||
900 | $this->pos = $catchupPos; |
||
901 | |||
902 | if ($closeParagraph) |
||
903 | { |
||
904 | $this->outputParagraphEnd(); |
||
905 | } |
||
906 | |||
907 | return; |
||
908 | } |
||
909 | |||
910 | // Compute the amount of text to ignore at the end of the output |
||
911 | 76 | $ignorePos = $catchupPos; |
|
912 | 76 | $ignoreLen = 0; |
|
913 | |||
914 | // Ignore as many lines (including whitespace) as specified |
||
915 | 76 | while ($maxLines && --$ignorePos >= $this->pos) |
|
916 | { |
||
917 | 7 | $c = $this->text[$ignorePos]; |
|
918 | 7 | if (strpos(self::WHITESPACE, $c) === false) |
|
919 | { |
||
920 | 5 | break; |
|
921 | } |
||
922 | |||
923 | 4 | if ($c === "\n") |
|
924 | { |
||
925 | 3 | --$maxLines; |
|
926 | } |
||
927 | |||
928 | 4 | ++$ignoreLen; |
|
929 | } |
||
930 | |||
931 | // Adjust $catchupPos to ignore the text at the end |
||
932 | 76 | $catchupPos -= $ignoreLen; |
|
933 | |||
934 | // Break down the text in paragraphs if applicable |
||
935 | 76 | if ($this->context['flags'] & self::RULE_CREATE_PARAGRAPHS) |
|
936 | { |
||
937 | 3 | if (!$this->context['inParagraph']) |
|
938 | { |
||
939 | 3 | $this->outputWhitespace($catchupPos); |
|
940 | |||
941 | 3 | if ($catchupPos > $this->pos) |
|
942 | { |
||
943 | 3 | $this->outputParagraphStart($catchupPos); |
|
944 | } |
||
945 | } |
||
946 | |||
947 | // Look for a paragraph break in this text |
||
948 | 3 | $pbPos = strpos($this->text, "\n\n", $this->pos); |
|
949 | |||
950 | 3 | while ($pbPos !== false && $pbPos < $catchupPos) |
|
951 | { |
||
952 | $this->outputText($pbPos, 0, true); |
||
953 | $this->outputParagraphStart($catchupPos); |
||
954 | |||
955 | $pbPos = strpos($this->text, "\n\n", $this->pos); |
||
956 | } |
||
957 | } |
||
958 | |||
959 | // Capture, escape and output the text |
||
960 | 76 | if ($catchupPos > $this->pos) |
|
961 | { |
||
962 | 76 | $catchupText = htmlspecialchars( |
|
963 | 76 | substr($this->text, $this->pos, $catchupPos - $this->pos), |
|
964 | 76 | ENT_NOQUOTES, |
|
965 | 76 | 'UTF-8' |
|
966 | ); |
||
967 | |||
968 | // Format line breaks if applicable |
||
969 | 76 | if (($this->context['flags'] & self::RULES_AUTO_LINEBREAKS) === self::RULE_ENABLE_AUTO_BR) |
|
970 | { |
||
971 | 8 | $catchupText = str_replace("\n", "<br/>\n", $catchupText); |
|
972 | } |
||
973 | |||
974 | 76 | $this->output .= $catchupText; |
|
975 | } |
||
976 | |||
977 | // Close the paragraph if applicable |
||
978 | 76 | if ($closeParagraph) |
|
979 | { |
||
980 | 3 | $this->outputParagraphEnd(); |
|
981 | } |
||
982 | |||
983 | // Add the ignored text if applicable |
||
984 | 76 | if ($ignoreLen) |
|
985 | { |
||
986 | 4 | $this->output .= substr($this->text, $catchupPos, $ignoreLen); |
|
987 | } |
||
988 | |||
989 | // Move the cursor past the text |
||
990 | 76 | $this->pos = $catchupPos + $ignoreLen; |
|
991 | 76 | } |
|
992 | |||
993 | /** |
||
994 | * Output a linebreak tag |
||
995 | * |
||
996 | * @param Tag $tag |
||
997 | * @return void |
||
998 | */ |
||
999 | 3 | protected function outputBrTag(Tag $tag) |
|
1004 | |||
1005 | /** |
||
1006 | * Output an ignore tag |
||
1007 | * |
||
1008 | * @param Tag $tag |
||
1009 | * @return void |
||
1010 | */ |
||
1011 | 13 | protected function outputIgnoreTag(Tag $tag) |
|
1027 | |||
1028 | /** |
||
1029 | * Start a paragraph between current position and given position, if applicable |
||
1030 | * |
||
1031 | * @param integer $maxPos Rightmost position at which the paragraph can be opened |
||
1032 | * @return void |
||
1033 | */ |
||
1034 | 73 | protected function outputParagraphStart($maxPos) |
|
1035 | { |
||
1036 | // Do nothing if we're already in a paragraph, or if we don't use paragraphs |
||
1037 | 73 | if ($this->context['inParagraph'] |
|
1038 | 73 | || !($this->context['flags'] & self::RULE_CREATE_PARAGRAPHS)) |
|
1039 | { |
||
1040 | 71 | return; |
|
1041 | } |
||
1042 | |||
1043 | // Output the whitespace between $this->pos and $maxPos if applicable |
||
1044 | 3 | $this->outputWhitespace($maxPos); |
|
1045 | |||
1046 | // Open the paragraph, but only if it's not at the very end of the text |
||
1047 | 3 | if ($this->pos < $this->textLen) |
|
1048 | { |
||
1049 | 3 | $this->output .= '<p>'; |
|
1050 | 3 | $this->context['inParagraph'] = true; |
|
1051 | } |
||
1052 | 3 | } |
|
1053 | |||
1054 | /** |
||
1055 | * Close current paragraph at current position if applicable |
||
1056 | * |
||
1057 | * @return void |
||
1058 | */ |
||
1059 | 3 | protected function outputParagraphEnd() |
|
1060 | { |
||
1061 | // Do nothing if we're not in a paragraph |
||
1062 | 3 | if (!$this->context['inParagraph']) |
|
1063 | { |
||
1064 | return; |
||
1065 | } |
||
1066 | |||
1067 | 3 | $this->output .= '</p>'; |
|
1068 | 3 | $this->context['inParagraph'] = false; |
|
1069 | 3 | } |
|
1070 | |||
1071 | /** |
||
1072 | * Output the content of a verbatim tag |
||
1073 | * |
||
1074 | * @param Tag $tag |
||
1075 | * @return void |
||
1076 | */ |
||
1077 | 3 | protected function outputVerbatim(Tag $tag) |
|
1084 | |||
1085 | /** |
||
1086 | * Skip as much whitespace after current position as possible |
||
1087 | * |
||
1088 | * @param integer $maxPos Rightmost character to be skipped |
||
1089 | * @return void |
||
1090 | */ |
||
1091 | 3 | protected function outputWhitespace($maxPos) |
|
1092 | { |
||
1093 | 3 | if ($maxPos > $this->pos) |
|
1094 | { |
||
1095 | 3 | $spn = strspn($this->text, self::WHITESPACE, $this->pos, $maxPos - $this->pos); |
|
1096 | |||
1097 | 3 | if ($spn) |
|
1098 | { |
||
1099 | $this->output .= substr($this->text, $this->pos, $spn); |
||
1100 | $this->pos += $spn; |
||
1101 | } |
||
1102 | } |
||
1103 | 3 | } |
|
1104 | |||
1105 | //========================================================================== |
||
1106 | // Plugins handling |
||
1107 | //========================================================================== |
||
1108 | |||
1109 | /** |
||
1110 | * Disable a plugin |
||
1111 | * |
||
1112 | * @param string $pluginName Name of the plugin |
||
1113 | * @return void |
||
1114 | */ |
||
1115 | 5 | public function disablePlugin($pluginName) |
|
1116 | { |
||
1117 | 5 | if (isset($this->pluginsConfig[$pluginName])) |
|
1118 | { |
||
1119 | // Copy the plugin's config to remove the reference |
||
1120 | 4 | $pluginConfig = $this->pluginsConfig[$pluginName]; |
|
1121 | 4 | unset($this->pluginsConfig[$pluginName]); |
|
1122 | |||
1123 | // Update the value and replace the plugin's config |
||
1124 | 4 | $pluginConfig['isDisabled'] = true; |
|
1125 | 4 | $this->pluginsConfig[$pluginName] = $pluginConfig; |
|
1126 | } |
||
1127 | 5 | } |
|
1128 | |||
1129 | /** |
||
1130 | * Enable a plugin |
||
1131 | * |
||
1132 | * @param string $pluginName Name of the plugin |
||
1133 | * @return void |
||
1134 | */ |
||
1135 | 2 | public function enablePlugin($pluginName) |
|
1136 | { |
||
1137 | 2 | if (isset($this->pluginsConfig[$pluginName])) |
|
1138 | { |
||
1139 | 1 | $this->pluginsConfig[$pluginName]['isDisabled'] = false; |
|
1140 | } |
||
1141 | 2 | } |
|
1142 | |||
1143 | /** |
||
1144 | * Execute given plugin |
||
1145 | * |
||
1146 | * @param string $pluginName Plugin's name |
||
1147 | * @return void |
||
1148 | */ |
||
1149 | 96 | protected function executePluginParser($pluginName) |
|
1150 | { |
||
1151 | 96 | $pluginConfig = $this->pluginsConfig[$pluginName]; |
|
1152 | 96 | if (isset($pluginConfig['quickMatch']) && strpos($this->text, $pluginConfig['quickMatch']) === false) |
|
1153 | { |
||
1154 | 1 | return; |
|
1155 | } |
||
1156 | |||
1157 | 95 | $matches = []; |
|
1158 | 95 | if (isset($pluginConfig['regexp'])) |
|
1159 | { |
||
1160 | 4 | $matches = $this->getMatches($pluginConfig['regexp'], $pluginConfig['regexpLimit']); |
|
1161 | 4 | if (empty($matches)) |
|
1162 | { |
||
1163 | 1 | return; |
|
1164 | } |
||
1165 | } |
||
1166 | |||
1167 | // Execute the plugin's parser, which will add tags via $this->addStartTag() and others |
||
1168 | 94 | call_user_func($this->getPluginParser($pluginName), $this->text, $matches); |
|
1169 | 94 | } |
|
1170 | |||
1171 | /** |
||
1172 | * Execute all the plugins |
||
1173 | * |
||
1174 | * @return void |
||
1175 | */ |
||
1176 | 97 | protected function executePluginParsers() |
|
1177 | { |
||
1178 | 97 | foreach ($this->pluginsConfig as $pluginName => $pluginConfig) |
|
1179 | { |
||
1180 | 97 | if (empty($pluginConfig['isDisabled'])) |
|
1181 | { |
||
1182 | 97 | $this->executePluginParser($pluginName); |
|
1183 | } |
||
1184 | } |
||
1185 | 97 | } |
|
1186 | |||
1187 | /** |
||
1188 | * Execute given regexp and returns as many matches as given limit |
||
1189 | * |
||
1190 | * @param string $regexp |
||
1191 | * @param integer $limit |
||
1192 | * @return array |
||
1193 | */ |
||
1194 | 4 | protected function getMatches($regexp, $limit) |
|
1195 | { |
||
1196 | 4 | $cnt = preg_match_all($regexp, $this->text, $matches, PREG_SET_ORDER | PREG_OFFSET_CAPTURE); |
|
1197 | 4 | if ($cnt > $limit) |
|
1198 | { |
||
1199 | 1 | $matches = array_slice($matches, 0, $limit); |
|
1200 | } |
||
1201 | |||
1202 | 4 | return $matches; |
|
1203 | } |
||
1204 | |||
1205 | /** |
||
1206 | * Get the cached callback for given plugin's parser |
||
1207 | * |
||
1208 | * @param string $pluginName Plugin's name |
||
1209 | * @return callable |
||
1210 | */ |
||
1211 | 94 | protected function getPluginParser($pluginName) |
|
1212 | { |
||
1213 | // Cache a new instance of this plugin's parser if there isn't one already |
||
1214 | 94 | if (!isset($this->pluginParsers[$pluginName])) |
|
1215 | { |
||
1216 | 1 | $pluginConfig = $this->pluginsConfig[$pluginName]; |
|
1217 | 1 | $className = (isset($pluginConfig['className'])) |
|
1218 | 1 | ? $pluginConfig['className'] |
|
1219 | 1 | : 's9e\\TextFormatter\\Plugins\\' . $pluginName . '\\Parser'; |
|
1220 | |||
1221 | // Register the parser as a callback |
||
1222 | 1 | $this->pluginParsers[$pluginName] = [new $className($this, $pluginConfig), 'parse']; |
|
1223 | } |
||
1224 | |||
1225 | 94 | return $this->pluginParsers[$pluginName]; |
|
1226 | } |
||
1227 | |||
1228 | /** |
||
1229 | * Register a parser |
||
1230 | * |
||
1231 | * Can be used to add a new parser with no plugin config, or pre-generate a parser for an |
||
1232 | * existing plugin |
||
1233 | * |
||
1234 | * @param string $pluginName |
||
1235 | * @param callback $parser |
||
1236 | * @return void |
||
1237 | */ |
||
1238 | 89 | public function registerParser($pluginName, $parser, $regexp = null, $limit = PHP_INT_MAX) |
|
1239 | { |
||
1240 | 89 | if (!is_callable($parser)) |
|
1241 | { |
||
1242 | 1 | throw new InvalidArgumentException('Argument 1 passed to ' . __METHOD__ . ' must be a valid callback'); |
|
1243 | } |
||
1244 | // Create an empty config for this plugin to ensure it is executed |
||
1245 | 88 | if (!isset($this->pluginsConfig[$pluginName])) |
|
1246 | { |
||
1247 | 87 | $this->pluginsConfig[$pluginName] = []; |
|
1248 | } |
||
1249 | 88 | if (isset($regexp)) |
|
1250 | { |
||
1251 | $this->pluginsConfig[$pluginName]['regexp'] = $regexp; |
||
1252 | $this->pluginsConfig[$pluginName]['regexpLimit'] = $limit; |
||
1253 | } |
||
1254 | 88 | $this->pluginParsers[$pluginName] = $parser; |
|
1255 | 88 | } |
|
1256 | |||
1257 | //========================================================================== |
||
1258 | // Rules handling |
||
1259 | //========================================================================== |
||
1260 | |||
1261 | /** |
||
1262 | * Apply closeAncestor rules associated with given tag |
||
1263 | * |
||
1264 | * @param Tag $tag Tag |
||
1265 | * @return bool Whether a new tag has been added |
||
1266 | */ |
||
1267 | 70 | protected function closeAncestor(Tag $tag) |
|
1268 | { |
||
1269 | 70 | if (!empty($this->openTags)) |
|
1270 | { |
||
1271 | 40 | $tagName = $tag->getName(); |
|
1272 | 40 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1273 | |||
1274 | 40 | if (!empty($tagConfig['rules']['closeAncestor'])) |
|
1275 | { |
||
1276 | 1 | $i = count($this->openTags); |
|
1277 | |||
1278 | 1 | while (--$i >= 0) |
|
1279 | { |
||
1280 | 1 | $ancestor = $this->openTags[$i]; |
|
1281 | 1 | $ancestorName = $ancestor->getName(); |
|
1282 | |||
1283 | 1 | if (isset($tagConfig['rules']['closeAncestor'][$ancestorName])) |
|
1284 | { |
||
1285 | 1 | ++$this->currentFixingCost; |
|
1286 | |||
1287 | // We have to close this ancestor. First we reinsert this tag... |
||
1288 | 1 | $this->tagStack[] = $tag; |
|
1289 | |||
1290 | // ...then we add a new end tag for it with a better priority |
||
1291 | 1 | $this->addMagicEndTag($ancestor, $tag->getPos(), $tag->getSortPriority() - 1); |
|
1292 | |||
1293 | 1 | return true; |
|
1294 | } |
||
1295 | } |
||
1296 | } |
||
1297 | } |
||
1298 | |||
1299 | 70 | return false; |
|
1300 | } |
||
1301 | |||
1302 | /** |
||
1303 | * Apply closeParent rules associated with given tag |
||
1304 | * |
||
1305 | * @param Tag $tag Tag |
||
1306 | * @return bool Whether a new tag has been added |
||
1307 | */ |
||
1308 | 70 | protected function closeParent(Tag $tag) |
|
1309 | { |
||
1310 | 70 | if (!empty($this->openTags)) |
|
1311 | { |
||
1312 | 41 | $tagName = $tag->getName(); |
|
1313 | 41 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1314 | |||
1315 | 41 | if (!empty($tagConfig['rules']['closeParent'])) |
|
1316 | { |
||
1317 | 3 | $parent = end($this->openTags); |
|
1318 | 3 | $parentName = $parent->getName(); |
|
1319 | |||
1320 | 3 | if (isset($tagConfig['rules']['closeParent'][$parentName])) |
|
1321 | { |
||
1322 | 3 | ++$this->currentFixingCost; |
|
1323 | |||
1324 | // We have to close that parent. First we reinsert the tag... |
||
1325 | 3 | $this->tagStack[] = $tag; |
|
1326 | |||
1327 | // ...then we add a new end tag for it with a better priority |
||
1328 | 3 | $this->addMagicEndTag($parent, $tag->getPos(), $tag->getSortPriority() - 1); |
|
1329 | |||
1330 | 3 | return true; |
|
1331 | } |
||
1332 | } |
||
1333 | } |
||
1334 | |||
1335 | 70 | return false; |
|
1336 | } |
||
1337 | |||
1338 | /** |
||
1339 | * Apply the createChild rules associated with given tag |
||
1340 | * |
||
1341 | * @param Tag $tag Tag |
||
1342 | * @return void |
||
1343 | */ |
||
1344 | 71 | protected function createChild(Tag $tag) |
|
1345 | { |
||
1346 | 71 | $tagConfig = $this->tagsConfig[$tag->getName()]; |
|
1347 | 71 | if (isset($tagConfig['rules']['createChild'])) |
|
1348 | { |
||
1349 | $priority = -1000; |
||
1350 | $tagPos = $this->pos + strspn($this->text, " \n\r\t", $this->pos); |
||
1351 | foreach ($tagConfig['rules']['createChild'] as $tagName) |
||
1352 | { |
||
1353 | $this->addStartTag($tagName, $tagPos, 0, ++$priority); |
||
1354 | } |
||
1355 | } |
||
1356 | 71 | } |
|
1357 | |||
1358 | /** |
||
1359 | * Apply fosterParent rules associated with given tag |
||
1360 | * |
||
1361 | * NOTE: this rule has the potential for creating an unbounded loop, either if a tag tries to |
||
1362 | * foster itself or two or more tags try to foster each other in a loop. We mitigate the |
||
1363 | * risk by preventing a tag from creating a child of itself (the parent still gets closed) |
||
1364 | * and by checking and increasing the currentFixingCost so that a loop of multiple tags |
||
1365 | * do not run indefinitely. The default tagLimit and nestingLimit also serve to prevent the |
||
1366 | * loop from running indefinitely |
||
1367 | * |
||
1368 | * @param Tag $tag Tag |
||
1369 | * @return bool Whether a new tag has been added |
||
1370 | */ |
||
1371 | 70 | protected function fosterParent(Tag $tag) |
|
1372 | { |
||
1373 | 70 | if (!empty($this->openTags)) |
|
1374 | { |
||
1375 | 41 | $tagName = $tag->getName(); |
|
1376 | 41 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1377 | |||
1378 | 41 | if (!empty($tagConfig['rules']['fosterParent'])) |
|
1379 | { |
||
1380 | 2 | $parent = end($this->openTags); |
|
1381 | 2 | $parentName = $parent->getName(); |
|
1382 | |||
1383 | 2 | if (isset($tagConfig['rules']['fosterParent'][$parentName])) |
|
1384 | { |
||
1385 | 2 | if ($parentName !== $tagName && $this->currentFixingCost < $this->maxFixingCost) |
|
1386 | { |
||
1387 | 2 | $this->addFosterTag($tag, $parent); |
|
1388 | } |
||
1389 | |||
1390 | // Reinsert current tag |
||
1391 | 2 | $this->tagStack[] = $tag; |
|
1392 | |||
1393 | // And finally close its parent with a priority that ensures it is processed |
||
1394 | // before this tag |
||
1395 | 2 | $this->addMagicEndTag($parent, $tag->getPos(), $tag->getSortPriority() - 1); |
|
1396 | |||
1397 | // Adjust the fixing cost to account for the additional tags/processing |
||
1398 | 2 | $this->currentFixingCost += 4; |
|
1399 | |||
1400 | 2 | return true; |
|
1401 | } |
||
1402 | } |
||
1403 | } |
||
1404 | |||
1405 | 70 | return false; |
|
1406 | } |
||
1407 | |||
1408 | /** |
||
1409 | * Apply requireAncestor rules associated with given tag |
||
1410 | * |
||
1411 | * @param Tag $tag Tag |
||
1412 | * @return bool Whether this tag has an unfulfilled requireAncestor requirement |
||
1413 | */ |
||
1414 | 72 | protected function requireAncestor(Tag $tag) |
|
1415 | { |
||
1416 | 72 | $tagName = $tag->getName(); |
|
1417 | 72 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1418 | |||
1419 | 72 | if (isset($tagConfig['rules']['requireAncestor'])) |
|
1420 | { |
||
1421 | 1 | foreach ($tagConfig['rules']['requireAncestor'] as $ancestorName) |
|
1422 | { |
||
1423 | 1 | if (!empty($this->cntOpen[$ancestorName])) |
|
1424 | { |
||
1425 | 1 | return false; |
|
1426 | } |
||
1427 | } |
||
1428 | |||
1429 | 1 | $this->logger->err('Tag requires an ancestor', [ |
|
1430 | 1 | 'requireAncestor' => implode(',', $tagConfig['rules']['requireAncestor']), |
|
1431 | 1 | 'tag' => $tag |
|
1432 | ]); |
||
1433 | |||
1434 | 1 | return true; |
|
1435 | } |
||
1436 | |||
1437 | 71 | return false; |
|
1438 | } |
||
1439 | |||
1440 | //========================================================================== |
||
1441 | // Tag processing |
||
1442 | //========================================================================== |
||
1443 | |||
1444 | /** |
||
1445 | * Create and add a copy of a tag as a child of a given tag |
||
1446 | * |
||
1447 | * @param Tag $tag Current tag |
||
1448 | * @param Tag $fosterTag Tag to foster |
||
1449 | * @return void |
||
1450 | */ |
||
1451 | 2 | protected function addFosterTag(Tag $tag, Tag $fosterTag) |
|
1459 | |||
1460 | /** |
||
1461 | * Create and add an end tag for given start tag at given position |
||
1462 | * |
||
1463 | * @param Tag $startTag Start tag |
||
1464 | * @param integer $tagPos End tag's position (will be adjusted for whitespace if applicable) |
||
1465 | * @param integer $prio End tag's priority |
||
1466 | * @return Tag |
||
1467 | */ |
||
1468 | 9 | protected function addMagicEndTag(Tag $startTag, $tagPos, $prio = 0) |
|
1469 | { |
||
1470 | 9 | $tagName = $startTag->getName(); |
|
1471 | |||
1472 | // Adjust the end tag's position if whitespace is to be minimized |
||
1473 | 9 | if (($this->currentTag->getFlags() | $startTag->getFlags()) & self::RULE_IGNORE_WHITESPACE) |
|
1474 | { |
||
1475 | 3 | $tagPos = $this->getMagicEndPos($tagPos); |
|
1476 | } |
||
1477 | |||
1478 | // Add a 0-width end tag that is paired with the given start tag |
||
1479 | 9 | $endTag = $this->addEndTag($tagName, $tagPos, 0, $prio); |
|
1480 | 9 | $endTag->pairWith($startTag); |
|
1481 | |||
1482 | 9 | return $endTag; |
|
1483 | } |
||
1484 | |||
1485 | /** |
||
1486 | * Compute the position of a magic end tag, adjusted for whitespace |
||
1487 | * |
||
1488 | * @param integer $tagPos Rightmost possible position for the tag |
||
1489 | * @return integer |
||
1490 | */ |
||
1491 | 5 | protected function getMagicEndPos($tagPos) |
|
1492 | { |
||
1493 | // Back up from given position to the cursor's position until we find a character that |
||
1494 | // is not whitespace |
||
1495 | 5 | while ($tagPos > $this->pos && strpos(self::WHITESPACE, $this->text[$tagPos - 1]) !== false) |
|
1496 | { |
||
1497 | 5 | --$tagPos; |
|
1498 | } |
||
1499 | |||
1500 | 5 | return $tagPos; |
|
1501 | } |
||
1502 | |||
1503 | /** |
||
1504 | * Compute the position and priority of a magic start tag, adjusted for whitespace |
||
1505 | * |
||
1506 | * @param integer $tagPos Leftmost possible position for the tag |
||
1507 | * @return integer[] [Tag pos, priority] |
||
1508 | */ |
||
1509 | 2 | protected function getMagicStartCoords($tagPos) |
|
1510 | { |
||
1511 | 2 | if (empty($this->tagStack)) |
|
1512 | { |
||
1513 | // Set the next position outside the text boundaries |
||
1514 | $nextPos = $this->textLen + 1; |
||
1515 | $nextPrio = 0; |
||
1516 | } |
||
1517 | else |
||
1518 | { |
||
1519 | 2 | $nextTag = end($this->tagStack); |
|
1520 | 2 | $nextPos = $nextTag->getPos(); |
|
1521 | 2 | $nextPrio = $nextTag->getSortPriority(); |
|
1522 | } |
||
1523 | |||
1524 | // Find the first non-whitespace position before next tag or the end of text |
||
1525 | 2 | while ($tagPos < $nextPos && strpos(self::WHITESPACE, $this->text[$tagPos]) !== false) |
|
1526 | { |
||
1527 | ++$tagPos; |
||
1528 | } |
||
1529 | |||
1530 | // Set a priority that ensures this tag appears before the next tag |
||
1531 | 2 | $prio = ($tagPos === $nextPos) ? $nextPrio - 1 : 0; |
|
1532 | |||
1533 | 2 | return [$tagPos, $prio]; |
|
1534 | } |
||
1535 | |||
1536 | /** |
||
1537 | * Test whether given start tag is immediately followed by a closing tag |
||
1538 | * |
||
1539 | * @param Tag $tag Start tag |
||
1540 | * @return bool |
||
1541 | */ |
||
1542 | 3 | protected function isFollowedByClosingTag(Tag $tag) |
|
1546 | |||
1547 | /** |
||
1548 | * Process all tags in the stack |
||
1549 | * |
||
1550 | * @return void |
||
1551 | */ |
||
1552 | 86 | protected function processTags() |
|
1553 | { |
||
1554 | 86 | if (empty($this->tagStack)) |
|
1555 | { |
||
1556 | 6 | return; |
|
1557 | } |
||
1558 | |||
1559 | // Initialize the count tables |
||
1560 | 80 | foreach (array_keys($this->tagsConfig) as $tagName) |
|
1561 | { |
||
1562 | 73 | $this->cntOpen[$tagName] = 0; |
|
1563 | 73 | $this->cntTotal[$tagName] = 0; |
|
1564 | } |
||
1565 | |||
1566 | // Process the tag stack, close tags that were left open and repeat until done |
||
1567 | do |
||
1568 | { |
||
1569 | 80 | while (!empty($this->tagStack)) |
|
1570 | { |
||
1571 | 80 | if (!$this->tagStackIsSorted) |
|
1572 | { |
||
1573 | 80 | $this->sortTags(); |
|
1574 | } |
||
1575 | |||
1576 | 80 | $this->currentTag = array_pop($this->tagStack); |
|
1577 | 80 | $this->processCurrentTag(); |
|
1578 | } |
||
1579 | |||
1580 | // Close tags that were left open |
||
1581 | 80 | foreach ($this->openTags as $startTag) |
|
1582 | { |
||
1583 | // NOTE: we add tags in hierarchical order (ancestors to descendants) but since |
||
1584 | // the stack is processed in LIFO order, it means that tags get closed in |
||
1585 | // the correct order, from descendants to ancestors |
||
1586 | 5 | $this->addMagicEndTag($startTag, $this->textLen); |
|
1587 | } |
||
1588 | } |
||
1589 | 80 | while (!empty($this->tagStack)); |
|
1590 | 80 | } |
|
1591 | |||
1592 | /** |
||
1593 | * Process current tag |
||
1594 | * |
||
1595 | * @return void |
||
1596 | */ |
||
1597 | 80 | protected function processCurrentTag() |
|
1598 | { |
||
1599 | // Invalidate current tag if tags are disabled and current tag would not close the last open |
||
1600 | // tag and is not a system tag |
||
1601 | 80 | if (($this->context['flags'] & self::RULE_IGNORE_TAGS) |
|
1602 | 80 | && !$this->currentTag->canClose(end($this->openTags)) |
|
1603 | 80 | && !$this->currentTag->isSystemTag()) |
|
1604 | { |
||
1605 | 4 | $this->currentTag->invalidate(); |
|
1606 | } |
||
1607 | |||
1608 | 80 | $tagPos = $this->currentTag->getPos(); |
|
1609 | 80 | $tagLen = $this->currentTag->getLen(); |
|
1610 | |||
1611 | // Test whether the cursor passed this tag's position already |
||
1612 | 80 | if ($this->pos > $tagPos && !$this->currentTag->isInvalid()) |
|
1613 | { |
||
1614 | // Test whether this tag is paired with a start tag and this tag is still open |
||
1615 | 15 | $startTag = $this->currentTag->getStartTag(); |
|
1616 | |||
1617 | 15 | if ($startTag && in_array($startTag, $this->openTags, true)) |
|
1618 | { |
||
1619 | // Create an end tag that matches current tag's start tag, which consumes as much of |
||
1620 | // the same text as current tag and is paired with the same start tag |
||
1621 | 2 | $this->addEndTag( |
|
1622 | 2 | $startTag->getName(), |
|
1623 | 2 | $this->pos, |
|
1624 | 2 | max(0, $tagPos + $tagLen - $this->pos) |
|
1625 | 2 | )->pairWith($startTag); |
|
1626 | |||
1627 | // Note that current tag is not invalidated, it's merely replaced |
||
1628 | 2 | return; |
|
1629 | } |
||
1630 | |||
1631 | // If this is an ignore tag, try to ignore as much as the remaining text as possible |
||
1632 | 13 | if ($this->currentTag->isIgnoreTag()) |
|
1633 | { |
||
1634 | 2 | $ignoreLen = $tagPos + $tagLen - $this->pos; |
|
1635 | |||
1636 | 2 | if ($ignoreLen > 0) |
|
1637 | { |
||
1638 | // Create a new ignore tag and move on |
||
1639 | 1 | $this->addIgnoreTag($this->pos, $ignoreLen); |
|
1640 | |||
1641 | 1 | return; |
|
1642 | } |
||
1643 | } |
||
1644 | |||
1645 | // Skipped tags are invalidated |
||
1646 | 12 | $this->currentTag->invalidate(); |
|
1647 | } |
||
1648 | |||
1649 | 80 | if ($this->currentTag->isInvalid()) |
|
1650 | { |
||
1651 | 16 | return; |
|
1652 | } |
||
1653 | |||
1654 | 80 | if ($this->currentTag->isIgnoreTag()) |
|
1655 | { |
||
1656 | 5 | $this->outputIgnoreTag($this->currentTag); |
|
1657 | } |
||
1658 | 79 | elseif ($this->currentTag->isBrTag()) |
|
1659 | { |
||
1660 | // Output the tag if it's allowed, ignore it otherwise |
||
1661 | 4 | if (!($this->context['flags'] & self::RULE_PREVENT_BR)) |
|
1662 | { |
||
1663 | 4 | $this->outputBrTag($this->currentTag); |
|
1664 | } |
||
1665 | } |
||
1666 | 77 | elseif ($this->currentTag->isParagraphBreak()) |
|
1667 | { |
||
1668 | 3 | $this->outputText($this->currentTag->getPos(), 0, true); |
|
1669 | } |
||
1670 | 75 | elseif ($this->currentTag->isVerbatim()) |
|
1671 | { |
||
1672 | 3 | $this->outputVerbatim($this->currentTag); |
|
1673 | } |
||
1674 | 72 | elseif ($this->currentTag->isStartTag()) |
|
1675 | { |
||
1676 | 72 | $this->processStartTag($this->currentTag); |
|
1677 | } |
||
1678 | else |
||
1679 | { |
||
1680 | 59 | $this->processEndTag($this->currentTag); |
|
1681 | } |
||
1682 | 80 | } |
|
1683 | |||
1684 | /** |
||
1685 | * Process given start tag (including self-closing tags) at current position |
||
1686 | * |
||
1687 | * @param Tag $tag Start tag (including self-closing) |
||
1688 | * @return void |
||
1689 | */ |
||
1690 | 72 | protected function processStartTag(Tag $tag) |
|
1691 | { |
||
1692 | 72 | $tagName = $tag->getName(); |
|
1693 | 72 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1694 | |||
1695 | // 1. Check that this tag has not reached its global limit tagLimit |
||
1696 | // 2. Execute this tag's filterChain, which will filter/validate its attributes |
||
1697 | // 3. Apply closeParent, closeAncestor and fosterParent rules |
||
1698 | // 4. Check for nestingLimit |
||
1699 | // 5. Apply requireAncestor rules |
||
1700 | // |
||
1701 | // This order ensures that the tag is valid and within the set limits before we attempt to |
||
1702 | // close parents or ancestors. We need to close ancestors before we can check for nesting |
||
1703 | // limits, whether this tag is allowed within current context (the context may change |
||
1704 | // as ancestors are closed) or whether the required ancestors are still there (they might |
||
1705 | // have been closed by a rule.) |
||
1706 | 72 | if ($this->cntTotal[$tagName] >= $tagConfig['tagLimit']) |
|
1707 | { |
||
1708 | 1 | $this->logger->err( |
|
1709 | 1 | 'Tag limit exceeded', |
|
1710 | [ |
||
1711 | 1 | 'tag' => $tag, |
|
1712 | 1 | 'tagName' => $tagName, |
|
1713 | 1 | 'tagLimit' => $tagConfig['tagLimit'] |
|
1714 | ] |
||
1715 | ); |
||
1716 | 1 | $tag->invalidate(); |
|
1717 | |||
1718 | 1 | return; |
|
1719 | } |
||
1720 | |||
1721 | 72 | if (!$this->filterTag($tag)) |
|
1722 | { |
||
1723 | 1 | $tag->invalidate(); |
|
1724 | |||
1725 | 1 | return; |
|
1726 | } |
||
1727 | |||
1728 | 72 | if ($this->currentFixingCost < $this->maxFixingCost) |
|
1729 | { |
||
1730 | 70 | if ($this->fosterParent($tag) || $this->closeParent($tag) || $this->closeAncestor($tag)) |
|
1731 | { |
||
1732 | // This tag parent/ancestor needs to be closed, we just return (the tag is still valid) |
||
1733 | 5 | return; |
|
1734 | } |
||
1735 | } |
||
1736 | |||
1737 | 72 | if ($this->cntOpen[$tagName] >= $tagConfig['nestingLimit']) |
|
1738 | { |
||
1739 | 1 | $this->logger->err( |
|
1740 | 1 | 'Nesting limit exceeded', |
|
1741 | [ |
||
1742 | 1 | 'tag' => $tag, |
|
1743 | 1 | 'tagName' => $tagName, |
|
1744 | 1 | 'nestingLimit' => $tagConfig['nestingLimit'] |
|
1745 | ] |
||
1746 | ); |
||
1747 | 1 | $tag->invalidate(); |
|
1748 | |||
1749 | 1 | return; |
|
1750 | } |
||
1751 | |||
1752 | 72 | if (!$this->tagIsAllowed($tagName)) |
|
1753 | { |
||
1754 | 7 | $msg = 'Tag is not allowed in this context'; |
|
1755 | 7 | $context = ['tag' => $tag, 'tagName' => $tagName]; |
|
1756 | 7 | if ($tag->getLen() > 0) |
|
1757 | { |
||
1758 | 6 | $this->logger->warn($msg, $context); |
|
1759 | } |
||
1760 | else |
||
1761 | { |
||
1762 | 1 | $this->logger->debug($msg, $context); |
|
1763 | } |
||
1764 | 7 | $tag->invalidate(); |
|
1765 | |||
1766 | 7 | return; |
|
1767 | } |
||
1768 | |||
1769 | 72 | if ($this->requireAncestor($tag)) |
|
1770 | { |
||
1771 | 1 | $tag->invalidate(); |
|
1772 | |||
1773 | 1 | return; |
|
1774 | } |
||
1775 | |||
1776 | // If this tag has an autoClose rule and it's not paired with an end tag or followed by an |
||
1777 | // end tag, we replace it with a self-closing tag with the same properties |
||
1778 | 71 | if ($tag->getFlags() & self::RULE_AUTO_CLOSE |
|
1779 | 71 | && !$tag->getEndTag() |
|
1780 | 71 | && !$this->isFollowedByClosingTag($tag)) |
|
1781 | { |
||
1782 | 2 | $newTag = new Tag(Tag::SELF_CLOSING_TAG, $tagName, $tag->getPos(), $tag->getLen()); |
|
1783 | 2 | $newTag->setAttributes($tag->getAttributes()); |
|
1784 | 2 | $newTag->setFlags($tag->getFlags()); |
|
1785 | |||
1786 | 2 | $tag = $newTag; |
|
1787 | } |
||
1788 | |||
1789 | 71 | if ($tag->getFlags() & self::RULE_TRIM_FIRST_LINE |
|
1790 | 71 | && !$tag->getEndTag() |
|
1791 | 71 | && substr($this->text, $tag->getPos() + $tag->getLen(), 1) === "\n") |
|
1792 | { |
||
1793 | $this->addIgnoreTag($tag->getPos() + $tag->getLen(), 1); |
||
1794 | } |
||
1795 | |||
1796 | // This tag is valid, output it and update the context |
||
1797 | 71 | $this->outputTag($tag); |
|
1798 | 71 | $this->pushContext($tag); |
|
1799 | |||
1800 | // Apply the createChild rules if applicable |
||
1801 | 71 | $this->createChild($tag); |
|
1802 | 71 | } |
|
1803 | |||
1804 | /** |
||
1805 | * Process given end tag at current position |
||
1806 | * |
||
1807 | * @param Tag $tag end tag |
||
1808 | * @return void |
||
1809 | */ |
||
1810 | 59 | protected function processEndTag(Tag $tag) |
|
1811 | { |
||
1812 | 59 | $tagName = $tag->getName(); |
|
1813 | |||
1814 | 59 | if (empty($this->cntOpen[$tagName])) |
|
1815 | { |
||
1816 | // This is an end tag with no start tag |
||
1817 | 3 | return; |
|
1818 | } |
||
1819 | |||
1820 | /** |
||
1821 | * @var array List of tags need to be closed before given tag |
||
1822 | */ |
||
1823 | 59 | $closeTags = []; |
|
1824 | |||
1825 | // Iterate through all open tags from last to first to find a match for our tag |
||
1826 | 59 | $i = count($this->openTags); |
|
1827 | 59 | while (--$i >= 0) |
|
1828 | { |
||
1829 | 59 | $openTag = $this->openTags[$i]; |
|
1830 | |||
1831 | 59 | if ($tag->canClose($openTag)) |
|
1832 | { |
||
1833 | 59 | break; |
|
1834 | } |
||
1835 | |||
1836 | 21 | $closeTags[] = $openTag; |
|
1837 | 21 | ++$this->currentFixingCost; |
|
1838 | } |
||
1839 | |||
1840 | 59 | if ($i < 0) |
|
1841 | { |
||
1842 | // Did not find a matching tag |
||
1843 | 2 | $this->logger->debug('Skipping end tag with no start tag', ['tag' => $tag]); |
|
1844 | |||
1845 | 2 | return; |
|
1846 | } |
||
1847 | |||
1848 | // Accumulate flags to determine whether whitespace should be trimmed |
||
1849 | 59 | $flags = $tag->getFlags(); |
|
1850 | 59 | foreach ($closeTags as $openTag) |
|
1851 | { |
||
1852 | 20 | $flags |= $openTag->getFlags(); |
|
1853 | } |
||
1854 | 59 | $ignoreWhitespace = (bool) ($flags & self::RULE_IGNORE_WHITESPACE); |
|
1855 | |||
1856 | // Only reopen tags if we haven't exceeded our "fixing" budget |
||
1857 | 59 | $keepReopening = (bool) ($this->currentFixingCost < $this->maxFixingCost); |
|
1858 | |||
1859 | // Iterate over tags that are being closed, output their end tag and collect tags to be |
||
1860 | // reopened |
||
1861 | 59 | $reopenTags = []; |
|
1862 | 59 | foreach ($closeTags as $openTag) |
|
1863 | { |
||
1864 | 20 | $openTagName = $openTag->getName(); |
|
1865 | |||
1866 | // Test whether this tag should be reopened automatically |
||
1867 | 20 | if ($keepReopening) |
|
1868 | { |
||
1869 | 18 | if ($openTag->getFlags() & self::RULE_AUTO_REOPEN) |
|
1870 | { |
||
1871 | 11 | $reopenTags[] = $openTag; |
|
1872 | } |
||
1873 | else |
||
1874 | { |
||
1875 | 7 | $keepReopening = false; |
|
1876 | } |
||
1877 | } |
||
1878 | |||
1879 | // Find the earliest position we can close this open tag |
||
1880 | 20 | $tagPos = $tag->getPos(); |
|
1881 | 20 | if ($ignoreWhitespace) |
|
1882 | { |
||
1883 | 5 | $tagPos = $this->getMagicEndPos($tagPos); |
|
1884 | } |
||
1885 | |||
1886 | // Output an end tag to close this start tag, then update the context |
||
1887 | 20 | $endTag = new Tag(Tag::END_TAG, $openTagName, $tagPos, 0); |
|
1888 | 20 | $endTag->setFlags($openTag->getFlags()); |
|
1889 | 20 | $this->outputTag($endTag); |
|
1890 | 20 | $this->popContext(); |
|
1891 | } |
||
1892 | |||
1893 | // Output our tag, moving the cursor past it, then update the context |
||
1894 | 59 | $this->outputTag($tag); |
|
1895 | 59 | $this->popContext(); |
|
1896 | |||
1897 | // If our fixing budget allows it, peek at upcoming tags and remove end tags that would |
||
1898 | // close tags that are already being closed now. Also, filter our list of tags being |
||
1899 | // reopened by removing those that would immediately be closed |
||
1900 | 59 | if (!empty($closeTags) && $this->currentFixingCost < $this->maxFixingCost) |
|
1901 | { |
||
1902 | /** |
||
1903 | * @var integer Rightmost position of the portion of text to ignore |
||
1904 | */ |
||
1905 | 18 | $ignorePos = $this->pos; |
|
1906 | |||
1907 | 18 | $i = count($this->tagStack); |
|
1908 | 18 | while (--$i >= 0 && ++$this->currentFixingCost < $this->maxFixingCost) |
|
1909 | { |
||
1910 | 13 | $upcomingTag = $this->tagStack[$i]; |
|
1911 | |||
1912 | // Test whether the upcoming tag is positioned at current "ignore" position and it's |
||
1913 | // strictly an end tag (not a start tag or a self-closing tag) |
||
1914 | 13 | if ($upcomingTag->getPos() > $ignorePos |
|
1915 | 13 | || $upcomingTag->isStartTag()) |
|
1916 | { |
||
1917 | 8 | break; |
|
1918 | } |
||
1919 | |||
1920 | // Test whether this tag would close any of the tags we're about to reopen |
||
1921 | 9 | $j = count($closeTags); |
|
1922 | |||
1923 | 9 | while (--$j >= 0 && ++$this->currentFixingCost < $this->maxFixingCost) |
|
1924 | { |
||
1925 | 9 | if ($upcomingTag->canClose($closeTags[$j])) |
|
1926 | { |
||
1927 | // Remove the tag from the lists and reset the keys |
||
1928 | 8 | array_splice($closeTags, $j, 1); |
|
1929 | |||
1930 | 8 | if (isset($reopenTags[$j])) |
|
1931 | { |
||
1932 | 7 | array_splice($reopenTags, $j, 1); |
|
1933 | } |
||
1934 | |||
1935 | // Extend the ignored text to cover this tag |
||
1936 | 8 | $ignorePos = max( |
|
1937 | 8 | $ignorePos, |
|
1938 | 8 | $upcomingTag->getPos() + $upcomingTag->getLen() |
|
1939 | ); |
||
1940 | |||
1941 | 8 | break; |
|
1942 | } |
||
1943 | } |
||
1944 | } |
||
1945 | |||
1946 | 18 | if ($ignorePos > $this->pos) |
|
1947 | { |
||
1948 | /** |
||
1949 | * @todo have a method that takes (pos,len) rather than a Tag |
||
1950 | */ |
||
1951 | 8 | $this->outputIgnoreTag(new Tag(Tag::SELF_CLOSING_TAG, 'i', $this->pos, $ignorePos - $this->pos)); |
|
1952 | } |
||
1953 | } |
||
1954 | |||
1955 | // Re-add tags that need to be reopened, at current cursor position |
||
1956 | 59 | foreach ($reopenTags as $startTag) |
|
1957 | { |
||
1958 | 7 | $newTag = $this->addCopyTag($startTag, $this->pos, 0); |
|
1959 | |||
1960 | // Re-pair the new tag |
||
1961 | 7 | $endTag = $startTag->getEndTag(); |
|
1962 | 7 | if ($endTag) |
|
1963 | { |
||
1964 | 7 | $newTag->pairWith($endTag); |
|
1965 | } |
||
1966 | } |
||
1967 | 59 | } |
|
1968 | |||
1969 | /** |
||
1970 | * Update counters and replace current context with its parent context |
||
1971 | * |
||
1972 | * @return void |
||
1973 | */ |
||
1974 | 59 | protected function popContext() |
|
1980 | |||
1981 | /** |
||
1982 | * Update counters and replace current context with a new context based on given tag |
||
1983 | * |
||
1984 | * If given tag is a self-closing tag, the context won't change |
||
1985 | * |
||
1986 | * @param Tag $tag Start tag (including self-closing) |
||
1987 | * @return void |
||
1988 | */ |
||
1989 | 71 | protected function pushContext(Tag $tag) |
|
1990 | { |
||
1991 | 71 | $tagName = $tag->getName(); |
|
1992 | 71 | $tagFlags = $tag->getFlags(); |
|
1993 | 71 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1994 | |||
1995 | 71 | ++$this->cntTotal[$tagName]; |
|
1996 | |||
1997 | // If this is a self-closing tag, the context remains the same |
||
1998 | 71 | if ($tag->isSelfClosingTag()) |
|
1999 | { |
||
2000 | 20 | return; |
|
2001 | } |
||
2002 | |||
2003 | // Recompute the allowed tags |
||
2004 | 59 | $allowed = []; |
|
2005 | 59 | if ($tagFlags & self::RULE_IS_TRANSPARENT) |
|
2006 | { |
||
2007 | 2 | foreach ($this->context['allowed'] as $k => $v) |
|
2008 | { |
||
2009 | 2 | $allowed[] = $tagConfig['allowed'][$k] & $v; |
|
2010 | } |
||
2011 | } |
||
2012 | else |
||
2013 | { |
||
2014 | 58 | foreach ($this->context['allowed'] as $k => $v) |
|
2015 | { |
||
2016 | 58 | $allowed[] = $tagConfig['allowed'][$k] & (($v & 0xFF00) | ($v >> 8)); |
|
2017 | } |
||
2018 | } |
||
2019 | |||
2020 | // Use this tag's flags as a base for this context and add inherited rules |
||
2021 | 59 | $flags = $tagFlags | ($this->context['flags'] & self::RULES_INHERITANCE); |
|
2022 | |||
2023 | // RULE_DISABLE_AUTO_BR turns off RULE_ENABLE_AUTO_BR |
||
2024 | 59 | if ($flags & self::RULE_DISABLE_AUTO_BR) |
|
2025 | { |
||
2026 | 2 | $flags &= ~self::RULE_ENABLE_AUTO_BR; |
|
2027 | } |
||
2028 | |||
2029 | 59 | ++$this->cntOpen[$tagName]; |
|
2030 | 59 | $this->openTags[] = $tag; |
|
2031 | 59 | $this->context = [ |
|
2032 | 59 | 'allowed' => $allowed, |
|
2033 | 59 | 'flags' => $flags, |
|
2034 | 'inParagraph' => false, |
||
2035 | 59 | 'parentContext' => $this->context |
|
2036 | ]; |
||
2037 | 59 | } |
|
2038 | |||
2039 | /** |
||
2040 | * Return whether given tag is allowed in current context |
||
2041 | * |
||
2042 | * @param string $tagName |
||
2043 | * @return bool |
||
2044 | */ |
||
2045 | 72 | protected function tagIsAllowed($tagName) |
|
2051 | |||
2052 | //========================================================================== |
||
2053 | // Tag stack |
||
2054 | //========================================================================== |
||
2055 | |||
2056 | /** |
||
2057 | * Add a start tag |
||
2058 | * |
||
2059 | * @param string $name Name of the tag |
||
2060 | * @param integer $pos Position of the tag in the text |
||
2061 | * @param integer $len Length of text consumed by the tag |
||
2062 | * @param integer $prio Tag's priority |
||
2063 | * @return Tag |
||
2064 | */ |
||
2065 | 76 | public function addStartTag($name, $pos, $len, $prio = 0) |
|
2069 | |||
2070 | /** |
||
2071 | * Add an end tag |
||
2072 | * |
||
2073 | * @param string $name Name of the tag |
||
2074 | * @param integer $pos Position of the tag in the text |
||
2075 | * @param integer $len Length of text consumed by the tag |
||
2076 | * @param integer $prio Tag's priority |
||
2077 | * @return Tag |
||
2078 | */ |
||
2079 | 64 | public function addEndTag($name, $pos, $len, $prio = 0) |
|
2083 | |||
2084 | /** |
||
2085 | * Add a self-closing tag |
||
2086 | * |
||
2087 | * @param string $name Name of the tag |
||
2088 | * @param integer $pos Position of the tag in the text |
||
2089 | * @param integer $len Length of text consumed by the tag |
||
2090 | * @param integer $prio Tag's priority |
||
2091 | * @return Tag |
||
2092 | */ |
||
2093 | 39 | public function addSelfClosingTag($name, $pos, $len, $prio = 0) |
|
2097 | |||
2098 | /** |
||
2099 | * Add a 0-width "br" tag to force a line break at given position |
||
2100 | * |
||
2101 | * @param integer $pos Position of the tag in the text |
||
2102 | * @param integer $prio Tag's priority |
||
2103 | * @return Tag |
||
2104 | */ |
||
2105 | 6 | public function addBrTag($pos, $prio = 0) |
|
2109 | |||
2110 | /** |
||
2111 | * Add an "ignore" tag |
||
2112 | * |
||
2113 | * @param integer $pos Position of the tag in the text |
||
2114 | * @param integer $len Length of text consumed by the tag |
||
2115 | * @param integer $prio Tag's priority |
||
2116 | * @return Tag |
||
2117 | */ |
||
2118 | 7 | public function addIgnoreTag($pos, $len, $prio = 0) |
|
2122 | |||
2123 | /** |
||
2124 | * Add a paragraph break at given position |
||
2125 | * |
||
2126 | * Uses a zero-width tag that is actually never output in the result |
||
2127 | * |
||
2128 | * @param integer $pos Position of the tag in the text |
||
2129 | * @param integer $prio Tag's priority |
||
2130 | * @return Tag |
||
2131 | */ |
||
2132 | 4 | public function addParagraphBreak($pos, $prio = 0) |
|
2136 | |||
2137 | /** |
||
2138 | * Add a copy of given tag at given position and length |
||
2139 | * |
||
2140 | * @param Tag $tag Original tag |
||
2141 | * @param integer $pos Copy's position |
||
2142 | * @param integer $len Copy's length |
||
2143 | * @param integer $prio Copy's priority (same as original by default) |
||
2144 | * @return Tag Copy tag |
||
2145 | */ |
||
2146 | 13 | public function addCopyTag(Tag $tag, $pos, $len, $prio = null) |
|
2147 | { |
||
2148 | 13 | if (!isset($prio)) |
|
2149 | { |
||
2150 | 10 | $prio = $tag->getSortPriority(); |
|
2151 | } |
||
2152 | 13 | $copy = $this->addTag($tag->getType(), $tag->getName(), $pos, $len, $prio); |
|
2153 | 13 | $copy->setAttributes($tag->getAttributes()); |
|
2154 | |||
2155 | 13 | return $copy; |
|
2156 | } |
||
2157 | |||
2158 | /** |
||
2159 | * Add a tag |
||
2160 | * |
||
2161 | * @param integer $type Tag's type |
||
2162 | * @param string $name Name of the tag |
||
2163 | * @param integer $pos Position of the tag in the text |
||
2164 | * @param integer $len Length of text consumed by the tag |
||
2165 | * @param integer $prio Tag's priority |
||
2166 | * @return Tag |
||
2167 | */ |
||
2168 | 106 | protected function addTag($type, $name, $pos, $len, $prio) |
|
2169 | { |
||
2170 | // Create the tag |
||
2171 | 106 | $tag = new Tag($type, $name, $pos, $len, $prio); |
|
2172 | |||
2173 | // Set this tag's rules bitfield |
||
2174 | 106 | if (isset($this->tagsConfig[$name])) |
|
2175 | { |
||
2176 | 94 | $tag->setFlags($this->tagsConfig[$name]['rules']['flags']); |
|
2177 | } |
||
2178 | |||
2179 | // Invalidate this tag if it's an unknown tag, a disabled tag, if either of its length or |
||
2180 | // position is negative or if it's out of bounds |
||
2181 | 106 | if (!isset($this->tagsConfig[$name]) && !$tag->isSystemTag()) |
|
2182 | { |
||
2183 | 2 | $tag->invalidate(); |
|
2184 | } |
||
2185 | 104 | elseif (!empty($this->tagsConfig[$name]['isDisabled'])) |
|
2186 | { |
||
2187 | 1 | $this->logger->warn( |
|
2188 | 1 | 'Tag is disabled', |
|
2189 | [ |
||
2190 | 1 | 'tag' => $tag, |
|
2191 | 1 | 'tagName' => $name |
|
2192 | ] |
||
2193 | ); |
||
2194 | 1 | $tag->invalidate(); |
|
2195 | } |
||
2196 | 103 | elseif ($len < 0 || $pos < 0 || $pos + $len > $this->textLen) |
|
2197 | { |
||
2198 | 6 | $tag->invalidate(); |
|
2199 | } |
||
2200 | else |
||
2201 | { |
||
2202 | 99 | $this->insertTag($tag); |
|
2203 | } |
||
2204 | |||
2205 | 106 | return $tag; |
|
2206 | } |
||
2207 | |||
2208 | /** |
||
2209 | * Insert given tag in the tag stack |
||
2210 | * |
||
2211 | * @param Tag $tag |
||
2212 | * @return void |
||
2213 | */ |
||
2214 | 99 | protected function insertTag(Tag $tag) |
|
2215 | { |
||
2216 | 99 | if (!$this->tagStackIsSorted) |
|
2217 | { |
||
2218 | 99 | $this->tagStack[] = $tag; |
|
2219 | } |
||
2220 | else |
||
2221 | { |
||
2222 | // Scan the stack and copy every tag to the next slot until we find the correct index |
||
2223 | 22 | $i = count($this->tagStack); |
|
2224 | 22 | while ($i > 0 && self::compareTags($this->tagStack[$i - 1], $tag) > 0) |
|
2225 | { |
||
2226 | 3 | $this->tagStack[$i] = $this->tagStack[$i - 1]; |
|
2227 | 3 | --$i; |
|
2228 | } |
||
2229 | 22 | $this->tagStack[$i] = $tag; |
|
2230 | } |
||
2231 | 99 | } |
|
2232 | |||
2233 | /** |
||
2234 | * Add a pair of tags |
||
2235 | * |
||
2236 | * @param string $name Name of the tags |
||
2237 | * @param integer $startPos Position of the start tag |
||
2238 | * @param integer $startLen Length of the start tag |
||
2239 | * @param integer $endPos Position of the start tag |
||
2240 | * @param integer $endLen Length of the start tag |
||
2241 | * @param integer $prio Start tag's priority (the end tag will be set to minus that value) |
||
2242 | * @return Tag Start tag |
||
2243 | */ |
||
2244 | 15 | public function addTagPair($name, $startPos, $startLen, $endPos, $endLen, $prio = 0) |
|
2253 | |||
2254 | /** |
||
2255 | * Add a tag that represents a verbatim copy of the original text |
||
2256 | * |
||
2257 | * @param integer $pos Position of the tag in the text |
||
2258 | * @param integer $len Length of text consumed by the tag |
||
2259 | * @param integer $prio Tag's priority |
||
2260 | * @return Tag |
||
2261 | */ |
||
2262 | 3 | public function addVerbatim($pos, $len, $prio = 0) |
|
2266 | |||
2267 | /** |
||
2268 | * Sort tags by position and precedence |
||
2269 | * |
||
2270 | * @return void |
||
2271 | */ |
||
2272 | 86 | protected function sortTags() |
|
2277 | |||
2278 | /** |
||
2279 | * sortTags() callback |
||
2280 | * |
||
2281 | * Tags are stored as a stack, in LIFO order. We sort tags by position _descending_ so that they |
||
2282 | * are processed in the order they appear in the text. |
||
2283 | * |
||
2284 | * @param Tag $a First tag to compare |
||
2285 | * @param Tag $b Second tag to compare |
||
2286 | * @return integer |
||
2287 | */ |
||
2288 | 73 | protected static function compareTags(Tag $a, Tag $b) |
|
2289 | { |
||
2290 | 73 | $aPos = $a->getPos(); |
|
2291 | 73 | $bPos = $b->getPos(); |
|
2292 | |||
2293 | // First we order by pos descending |
||
2294 | 73 | if ($aPos !== $bPos) |
|
2295 | { |
||
2296 | 69 | return $bPos - $aPos; |
|
2297 | } |
||
2298 | |||
2299 | // If the tags start at the same position, we'll use their sortPriority if applicable. Tags |
||
2300 | // with a lower value get sorted last, which means they'll be processed first. IOW, -10 is |
||
2301 | // processed before 10 |
||
2302 | 17 | if ($a->getSortPriority() !== $b->getSortPriority()) |
|
2303 | { |
||
2341 | } |