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 | 187 | public function __construct(array $config) |
|
170 | { |
||
171 | 187 | $this->pluginsConfig = $config['plugins']; |
|
172 | 187 | $this->registeredVars = $config['registeredVars']; |
|
173 | 187 | $this->rootContext = $config['rootContext']; |
|
174 | 187 | $this->tagsConfig = $config['tags']; |
|
175 | |||
176 | 187 | $this->__wakeup(); |
|
177 | 187 | } |
|
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 | 2 | public function __sleep() |
|
190 | { |
||
191 | 2 | return ['pluginsConfig', 'registeredVars', 'rootContext', 'tagsConfig']; |
|
192 | } |
||
193 | |||
194 | /** |
||
195 | * Unserializer |
||
196 | * |
||
197 | * @return void |
||
198 | */ |
||
199 | 187 | public function __wakeup() |
|
200 | { |
||
201 | 187 | $this->logger = new Logger; |
|
202 | 187 | } |
|
203 | |||
204 | /** |
||
205 | * Reset the parser for a new parsing |
||
206 | * |
||
207 | * @param string $text Text to be parsed |
||
208 | * @return void |
||
209 | */ |
||
210 | 175 | protected function reset($text) |
|
211 | { |
||
212 | // Normalize CR/CRLF to LF, remove control characters that aren't allowed in XML |
||
213 | 175 | $text = preg_replace('/\\r\\n?/', "\n", $text); |
|
214 | 175 | $text = preg_replace('/[\\x00-\\x08\\x0B\\x0C\\x0E-\\x1F]+/S', '', $text); |
|
215 | |||
216 | // Clear the logs |
||
217 | 175 | $this->logger->clear(); |
|
218 | |||
219 | // Initialize the rest |
||
220 | 175 | $this->cntOpen = []; |
|
221 | 175 | $this->cntTotal = []; |
|
222 | 175 | $this->currentFixingCost = 0; |
|
223 | 175 | $this->currentTag = null; |
|
224 | 175 | $this->isRich = false; |
|
225 | 175 | $this->namespaces = []; |
|
226 | 175 | $this->openTags = []; |
|
227 | 175 | $this->output = ''; |
|
228 | 175 | $this->pos = 0; |
|
229 | 175 | $this->tagStack = []; |
|
230 | 175 | $this->tagStackIsSorted = false; |
|
231 | 175 | $this->text = $text; |
|
232 | 175 | $this->textLen = strlen($text); |
|
233 | 175 | $this->wsPos = 0; |
|
234 | |||
235 | // Initialize the root context |
||
236 | 175 | $this->context = $this->rootContext; |
|
237 | 175 | $this->context['inParagraph'] = false; |
|
238 | |||
239 | // Bump the UID |
||
240 | 175 | ++$this->uid; |
|
241 | 175 | } |
|
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 $optionValue Option's value |
||
252 | * @return void |
||
253 | */ |
||
254 | 7 | protected function setTagOption($tagName, $optionName, $optionValue) |
|
255 | { |
||
256 | 7 | if (isset($this->tagsConfig[$tagName])) |
|
257 | { |
||
258 | // Copy the tag's config and remove it. That will destroy the reference |
||
259 | 7 | $tagConfig = $this->tagsConfig[$tagName]; |
|
260 | 7 | unset($this->tagsConfig[$tagName]); |
|
261 | |||
262 | // Set the new value and replace the tag's config |
||
263 | 7 | $tagConfig[$optionName] = $optionValue; |
|
264 | 7 | $this->tagsConfig[$tagName] = $tagConfig; |
|
265 | } |
||
266 | 7 | } |
|
267 | |||
268 | //========================================================================== |
||
269 | // Public API |
||
270 | //========================================================================== |
||
271 | |||
272 | /** |
||
273 | * Disable a tag |
||
274 | * |
||
275 | * @param string $tagName Name of the tag |
||
276 | * @return void |
||
277 | */ |
||
278 | 3 | public function disableTag($tagName) |
|
279 | { |
||
280 | 3 | $this->setTagOption($tagName, 'isDisabled', true); |
|
281 | 3 | } |
|
282 | |||
283 | /** |
||
284 | * Enable a tag |
||
285 | * |
||
286 | * @param string $tagName Name of the tag |
||
287 | * @return void |
||
288 | */ |
||
289 | 1 | public function enableTag($tagName) |
|
290 | { |
||
291 | 1 | if (isset($this->tagsConfig[$tagName])) |
|
292 | { |
||
293 | 1 | unset($this->tagsConfig[$tagName]['isDisabled']); |
|
294 | } |
||
295 | 1 | } |
|
296 | |||
297 | /** |
||
298 | * Get this parser's Logger instance |
||
299 | * |
||
300 | * @return Logger |
||
301 | */ |
||
302 | 8 | public function getLogger() |
|
303 | { |
||
304 | 8 | return $this->logger; |
|
305 | } |
||
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 | 2 | public function getText() |
|
320 | { |
||
321 | 2 | return $this->text; |
|
322 | } |
||
323 | |||
324 | /** |
||
325 | * Parse a text |
||
326 | * |
||
327 | * @param string $text Text to parse |
||
328 | * @return string XML representation |
||
329 | */ |
||
330 | 175 | public function parse($text) |
|
331 | { |
||
332 | // Reset the parser and save the uid |
||
333 | 175 | $this->reset($text); |
|
334 | 175 | $uid = $this->uid; |
|
335 | |||
336 | // Do the heavy lifting |
||
337 | 175 | $this->executePluginParsers(); |
|
338 | 175 | $this->processTags(); |
|
339 | |||
340 | // Finalize the document |
||
341 | 175 | $this->finalizeOutput(); |
|
342 | |||
343 | // Check the uid in case a plugin or a filter reset the parser mid-execution |
||
344 | 175 | if ($this->uid !== $uid) |
|
345 | { |
||
346 | 1 | throw new RuntimeException('The parser has been reset during execution'); |
|
347 | } |
||
348 | |||
349 | // Log a warning if the fixing cost limit was exceeded |
||
350 | 175 | if ($this->currentFixingCost > $this->maxFixingCost) |
|
351 | { |
||
352 | 2 | $this->logger->warn('Fixing cost limit exceeded'); |
|
353 | } |
||
354 | |||
355 | 175 | 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 $tagLimit |
||
365 | * @return void |
||
366 | */ |
||
367 | 2 | public function setTagLimit($tagName, $tagLimit) |
|
368 | { |
||
369 | 2 | $this->setTagOption($tagName, 'tagLimit', $tagLimit); |
|
370 | 2 | } |
|
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 $nestingLimit |
||
379 | * @return void |
||
380 | */ |
||
381 | 2 | public function setNestingLimit($tagName, $nestingLimit) |
|
382 | { |
||
383 | 2 | $this->setTagOption($tagName, 'nestingLimit', $nestingLimit); |
|
384 | 2 | } |
|
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 | 7 | public static function executeAttributePreprocessors(Tag $tag, array $tagConfig) |
|
400 | { |
||
401 | 7 | if (!empty($tagConfig['attributePreprocessors'])) |
|
402 | { |
||
403 | 7 | foreach ($tagConfig['attributePreprocessors'] as list($attrName, $regexp, $map)) |
|
404 | { |
||
405 | 7 | if (!$tag->hasAttribute($attrName)) |
|
406 | { |
||
407 | 1 | continue; |
|
408 | } |
||
409 | |||
410 | 6 | self::executeAttributePreprocessor($tag, $attrName, $regexp, $map); |
|
411 | } |
||
412 | } |
||
413 | |||
414 | 7 | 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 | 6 | protected static function executeAttributePreprocessor(Tag $tag, $attrName, $regexp, $map) |
|
427 | { |
||
428 | 6 | $attrValue = $tag->getAttribute($attrName); |
|
429 | 6 | $captures = self::getNamedCaptures($attrValue, $regexp, $map); |
|
430 | 6 | foreach ($captures as $k => $v) |
|
431 | { |
||
432 | // Attribute preprocessors cannot overwrite other attributes but they can |
||
433 | // overwrite themselves |
||
434 | 5 | if ($k === $attrName || !$tag->hasAttribute($k)) |
|
435 | { |
||
436 | 5 | $tag->setAttribute($k, $v); |
|
437 | } |
||
438 | } |
||
439 | 6 | } |
|
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[] $map |
||
447 | * @return array |
||
448 | */ |
||
449 | 6 | protected static function getNamedCaptures($attrValue, $regexp, $map) |
|
450 | { |
||
451 | 6 | if (!preg_match($regexp, $attrValue, $m)) |
|
452 | { |
||
453 | 1 | return []; |
|
454 | } |
||
455 | |||
456 | 5 | $values = []; |
|
457 | 5 | foreach ($map as $i => $k) |
|
458 | { |
||
459 | 5 | if (isset($m[$i]) && $m[$i] !== '') |
|
460 | { |
||
461 | 5 | $values[$k] = $m[$i]; |
|
462 | } |
||
463 | } |
||
464 | |||
465 | 5 | 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 | 149 | protected static function executeFilter(array $filter, array $vars) |
|
478 | { |
||
479 | 149 | $callback = $filter['callback']; |
|
480 | 149 | $params = (isset($filter['params'])) ? $filter['params'] : []; |
|
481 | |||
482 | 149 | $args = []; |
|
483 | 149 | foreach ($params as $k => $v) |
|
484 | { |
||
485 | 149 | if (is_numeric($k)) |
|
486 | { |
||
487 | // By-value param |
||
488 | 5 | $args[] = $v; |
|
489 | } |
||
490 | 148 | elseif (isset($vars[$k])) |
|
491 | { |
||
492 | // By-name param using a supplied var |
||
493 | 146 | $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 | 149 | $args[] = null; |
|
504 | } |
||
505 | } |
||
506 | |||
507 | 149 | return call_user_func_array($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 void |
||
520 | */ |
||
521 | 148 | public static function filterAttributes(Tag $tag, array $tagConfig, array $registeredVars, Logger $logger) |
|
522 | { |
||
523 | 148 | if (empty($tagConfig['attributes'])) |
|
524 | { |
||
525 | 134 | $tag->setAttributes([]); |
|
526 | |||
527 | 134 | return; |
|
528 | } |
||
529 | |||
530 | // Filter and remove invalid attributes |
||
531 | 16 | foreach ($tag->getAttributes() as $attrName => $attrValue) |
|
532 | { |
||
533 | // Test whether this attribute exists and remove it if it doesn't |
||
534 | 15 | if (!isset($tagConfig['attributes'][$attrName])) |
|
535 | { |
||
536 | 2 | $tag->removeAttribute($attrName); |
|
537 | 2 | continue; |
|
538 | } |
||
539 | |||
540 | 15 | $attrConfig = $tagConfig['attributes'][$attrName]; |
|
541 | |||
542 | // Test whether this attribute has a filterChain |
||
543 | 15 | if (!isset($attrConfig['filterChain'])) |
|
544 | { |
||
545 | 10 | continue; |
|
546 | } |
||
547 | |||
548 | // Record the name of the attribute being filtered into the logger |
||
549 | 5 | $logger->setAttribute($attrName); |
|
550 | |||
551 | 5 | foreach ($attrConfig['filterChain'] as $filter) |
|
552 | { |
||
553 | 5 | $attrValue = self::executeFilter( |
|
554 | 5 | $filter, |
|
555 | [ |
||
556 | 5 | 'attrName' => $attrName, |
|
557 | 5 | 'attrValue' => $attrValue, |
|
558 | 5 | 'logger' => $logger, |
|
559 | 5 | 'registeredVars' => $registeredVars |
|
560 | ] |
||
561 | ); |
||
562 | |||
563 | 5 | if ($attrValue === false) |
|
564 | { |
||
565 | 3 | $tag->removeAttribute($attrName); |
|
566 | 5 | break; |
|
567 | } |
||
568 | } |
||
569 | |||
570 | // Update the attribute value if it's valid |
||
571 | 5 | if ($attrValue !== false) |
|
572 | { |
||
573 | 2 | $tag->setAttribute($attrName, $attrValue); |
|
574 | } |
||
575 | |||
576 | // Remove the attribute's name from the logger |
||
577 | 5 | $logger->unsetAttribute(); |
|
578 | } |
||
579 | |||
580 | // Iterate over the attribute definitions to handle missing attributes |
||
581 | 16 | foreach ($tagConfig['attributes'] as $attrName => $attrConfig) |
|
582 | { |
||
583 | // Test whether this attribute is missing |
||
584 | 16 | if (!$tag->hasAttribute($attrName)) |
|
585 | { |
||
586 | 5 | if (isset($attrConfig['defaultValue'])) |
|
587 | { |
||
588 | // Use the attribute's default value |
||
589 | 2 | $tag->setAttribute($attrName, $attrConfig['defaultValue']); |
|
590 | } |
||
591 | 3 | elseif (!empty($attrConfig['required'])) |
|
592 | { |
||
593 | // This attribute is missing, has no default value and is required, which means |
||
594 | // the attribute set is invalid |
||
595 | 16 | $tag->invalidate(); |
|
596 | } |
||
597 | } |
||
598 | } |
||
599 | 16 | } |
|
600 | |||
601 | /** |
||
602 | * Execute given tag's filterChain |
||
603 | * |
||
604 | * @param Tag $tag Tag to filter |
||
605 | * @return void |
||
606 | */ |
||
607 | 140 | protected function filterTag(Tag $tag) |
|
608 | { |
||
609 | 140 | $tagName = $tag->getName(); |
|
610 | 140 | $tagConfig = $this->tagsConfig[$tagName]; |
|
611 | 140 | if (!empty($tagConfig['filterChain'])) |
|
612 | { |
||
613 | // Record the tag being processed into the logger it can be added to the context of |
||
614 | // messages logged during the execution |
||
615 | 140 | $this->logger->setTag($tag); |
|
616 | |||
617 | // Prepare the variables that are accessible to filters |
||
618 | $vars = [ |
||
619 | 140 | 'logger' => $this->logger, |
|
620 | 140 | 'openTags' => $this->openTags, |
|
621 | 140 | 'parser' => $this, |
|
622 | 140 | 'registeredVars' => $this->registeredVars, |
|
623 | 140 | 'tag' => $tag, |
|
624 | 140 | 'tagConfig' => $tagConfig, |
|
625 | 140 | 'text' => $this->text |
|
626 | ]; |
||
627 | 140 | foreach ($tagConfig['filterChain'] as $filter) |
|
628 | { |
||
629 | 140 | if ($tag->isInvalid()) |
|
630 | { |
||
631 | 1 | break; |
|
632 | } |
||
633 | 140 | self::executeFilter($filter, $vars); |
|
634 | } |
||
635 | |||
636 | // Remove the tag from the logger |
||
637 | 140 | $this->logger->unsetTag(); |
|
638 | } |
||
639 | 140 | } |
|
640 | |||
641 | //========================================================================== |
||
642 | // Output handling |
||
643 | //========================================================================== |
||
644 | |||
645 | /** |
||
646 | * Finalize the output by appending the rest of the unprocessed text and create the root node |
||
647 | * |
||
648 | * @return void |
||
649 | */ |
||
650 | 175 | protected function finalizeOutput() |
|
651 | { |
||
652 | // Output the rest of the text and close the last paragraph |
||
653 | 175 | $this->outputText($this->textLen, 0, true); |
|
654 | |||
655 | // Remove empty tag pairs, e.g. <I><U></U></I> as well as empty paragraphs |
||
656 | do |
||
657 | { |
||
658 | 175 | $this->output = preg_replace('(<([^ />]++)[^>]*></\\1>)', '', $this->output, -1, $cnt); |
|
659 | } |
||
660 | 175 | while ($cnt > 0); |
|
661 | |||
662 | // Merge consecutive <i> tags |
||
663 | 175 | if (strpos($this->output, '</i><i>') !== false) |
|
664 | { |
||
665 | 1 | $this->output = str_replace('</i><i>', '', $this->output); |
|
666 | } |
||
667 | |||
668 | // Encode Unicode characters that are outside of the BMP |
||
669 | 175 | $this->output = Utils::encodeUnicodeSupplementaryCharacters($this->output); |
|
670 | |||
671 | // Use a <r> root if the text is rich, or <t> for plain text (including <p></p> and <br/>) |
||
672 | 175 | $tagName = ($this->isRich) ? 'r' : 't'; |
|
673 | |||
674 | // Prepare the root node with all the namespace declarations |
||
675 | 175 | $tmp = '<' . $tagName; |
|
676 | 175 | foreach (array_keys($this->namespaces) as $prefix) |
|
677 | { |
||
678 | 2 | $tmp .= ' xmlns:' . $prefix . '="urn:s9e:TextFormatter:' . $prefix . '"'; |
|
679 | } |
||
680 | |||
681 | 175 | $this->output = $tmp . '>' . $this->output . '</' . $tagName . '>'; |
|
682 | 175 | } |
|
683 | |||
684 | /** |
||
685 | * Append a tag to the output |
||
686 | * |
||
687 | * @param Tag $tag Tag to append |
||
688 | * @return void |
||
689 | */ |
||
690 | 134 | protected function outputTag(Tag $tag) |
|
691 | { |
||
692 | 134 | $this->isRich = true; |
|
693 | |||
694 | 134 | $tagName = $tag->getName(); |
|
695 | 134 | $tagPos = $tag->getPos(); |
|
696 | 134 | $tagLen = $tag->getLen(); |
|
697 | 134 | $tagFlags = $tag->getFlags(); |
|
698 | |||
699 | 134 | if ($tagFlags & self::RULE_IGNORE_WHITESPACE) |
|
700 | { |
||
701 | 11 | $skipBefore = 1; |
|
702 | 11 | $skipAfter = ($tag->isEndTag()) ? 2 : 1; |
|
703 | } |
||
704 | else |
||
705 | { |
||
706 | 129 | $skipBefore = $skipAfter = 0; |
|
707 | } |
||
708 | |||
709 | // Current paragraph must end before the tag if: |
||
710 | // - the tag is a start (or self-closing) tag and it breaks paragraphs, or |
||
711 | // - the tag is an end tag (but not self-closing) |
||
712 | 134 | $closeParagraph = false; |
|
713 | 134 | if ($tag->isStartTag()) |
|
714 | { |
||
715 | 134 | if ($tagFlags & self::RULE_BREAK_PARAGRAPH) |
|
716 | { |
||
717 | 134 | $closeParagraph = true; |
|
718 | } |
||
719 | } |
||
720 | else |
||
721 | { |
||
722 | 101 | $closeParagraph = true; |
|
723 | } |
||
724 | |||
725 | // Let the cursor catch up with this tag's position |
||
726 | 134 | $this->outputText($tagPos, $skipBefore, $closeParagraph); |
|
727 | |||
728 | // Capture the text consumed by the tag |
||
729 | 134 | $tagText = ($tagLen) |
|
730 | 95 | ? htmlspecialchars(substr($this->text, $tagPos, $tagLen), ENT_NOQUOTES, 'UTF-8') |
|
731 | 134 | : ''; |
|
732 | |||
733 | // Output current tag |
||
734 | 134 | if ($tag->isStartTag()) |
|
735 | { |
||
736 | // Handle paragraphs before opening the tag |
||
737 | 134 | if (!($tagFlags & self::RULE_BREAK_PARAGRAPH)) |
|
738 | { |
||
739 | 133 | $this->outputParagraphStart($tagPos); |
|
740 | } |
||
741 | |||
742 | // Record this tag's namespace, if applicable |
||
743 | 134 | $colonPos = strpos($tagName, ':'); |
|
744 | 134 | if ($colonPos) |
|
745 | { |
||
746 | 2 | $this->namespaces[substr($tagName, 0, $colonPos)] = 0; |
|
747 | } |
||
748 | |||
749 | // Open the start tag and add its attributes, but don't close the tag |
||
750 | 134 | $this->output .= '<' . $tagName; |
|
751 | |||
752 | // We output the attributes in lexical order. Helps canonicalizing the output and could |
||
753 | // prove useful someday |
||
754 | 134 | $attributes = $tag->getAttributes(); |
|
755 | 134 | ksort($attributes); |
|
756 | |||
757 | 134 | foreach ($attributes as $attrName => $attrValue) |
|
758 | { |
||
759 | 9 | $this->output .= ' ' . $attrName . '="' . str_replace("\n", ' ', htmlspecialchars($attrValue, ENT_COMPAT, 'UTF-8')) . '"'; |
|
760 | } |
||
761 | |||
762 | 134 | if ($tag->isSelfClosingTag()) |
|
763 | { |
||
764 | 47 | if ($tagLen) |
|
765 | { |
||
766 | 35 | $this->output .= '>' . $tagText . '</' . $tagName . '>'; |
|
767 | } |
||
768 | else |
||
769 | { |
||
770 | 47 | $this->output .= '/>'; |
|
771 | } |
||
772 | } |
||
773 | 101 | elseif ($tagLen) |
|
774 | { |
||
775 | 68 | $this->output .= '><s>' . $tagText . '</s>'; |
|
776 | } |
||
777 | else |
||
778 | { |
||
779 | 134 | $this->output .= '>'; |
|
780 | } |
||
781 | } |
||
782 | else |
||
783 | { |
||
784 | 101 | if ($tagLen) |
|
785 | { |
||
786 | 56 | $this->output .= '<e>' . $tagText . '</e>'; |
|
787 | } |
||
788 | |||
789 | 101 | $this->output .= '</' . $tagName . '>'; |
|
790 | } |
||
791 | |||
792 | // Move the cursor past the tag |
||
793 | 134 | $this->pos = $tagPos + $tagLen; |
|
794 | |||
795 | // Skip newlines (no other whitespace) after this tag |
||
796 | 134 | $this->wsPos = $this->pos; |
|
797 | 134 | while ($skipAfter && $this->wsPos < $this->textLen && $this->text[$this->wsPos] === "\n") |
|
798 | { |
||
799 | // Decrement the number of lines to skip |
||
800 | 9 | --$skipAfter; |
|
801 | |||
802 | // Move the cursor past the newline |
||
803 | 9 | ++$this->wsPos; |
|
804 | } |
||
805 | 134 | } |
|
806 | |||
807 | /** |
||
808 | * Output the text between the cursor's position (included) and given position (not included) |
||
809 | * |
||
810 | * @param integer $catchupPos Position we're catching up to |
||
811 | * @param integer $maxLines Maximum number of lines to ignore at the end of the text |
||
812 | * @param bool $closeParagraph Whether to close the paragraph at the end, if applicable |
||
813 | * @return void |
||
814 | */ |
||
815 | 175 | protected function outputText($catchupPos, $maxLines, $closeParagraph) |
|
816 | { |
||
817 | 175 | if ($closeParagraph) |
|
818 | { |
||
819 | 175 | if (!($this->context['flags'] & self::RULE_CREATE_PARAGRAPHS)) |
|
820 | { |
||
821 | 164 | $closeParagraph = false; |
|
822 | } |
||
823 | else |
||
824 | { |
||
825 | // Ignore any number of lines at the end if we're closing a paragraph |
||
826 | 18 | $maxLines = -1; |
|
827 | } |
||
828 | } |
||
829 | |||
830 | 175 | if ($this->pos >= $catchupPos) |
|
831 | { |
||
832 | // We're already there, close the paragraph if applicable and return |
||
833 | 132 | if ($closeParagraph) |
|
834 | { |
||
835 | 4 | $this->outputParagraphEnd(); |
|
836 | } |
||
837 | |||
838 | 132 | return; |
|
839 | } |
||
840 | |||
841 | // Skip over previously identified whitespace if applicable |
||
842 | 154 | if ($this->wsPos > $this->pos) |
|
843 | { |
||
844 | 9 | $skipPos = min($catchupPos, $this->wsPos); |
|
845 | 9 | $this->output .= substr($this->text, $this->pos, $skipPos - $this->pos); |
|
846 | 9 | $this->pos = $skipPos; |
|
847 | |||
848 | 9 | if ($this->pos >= $catchupPos) |
|
849 | { |
||
850 | // Skipped everything. Close the paragraph if applicable and return |
||
851 | 2 | if ($closeParagraph) |
|
852 | { |
||
853 | 1 | $this->outputParagraphEnd(); |
|
854 | } |
||
855 | |||
856 | 2 | return; |
|
857 | } |
||
858 | } |
||
859 | |||
860 | // Test whether we're even supposed to output anything |
||
861 | 154 | if ($this->context['flags'] & self::RULE_IGNORE_TEXT) |
|
862 | { |
||
863 | 3 | $catchupLen = $catchupPos - $this->pos; |
|
864 | 3 | $catchupText = substr($this->text, $this->pos, $catchupLen); |
|
865 | |||
866 | // If the catchup text is not entirely composed of whitespace, we put it inside ignore |
||
867 | // tags |
||
868 | 3 | if (strspn($catchupText, " \n\t") < $catchupLen) |
|
869 | { |
||
870 | 3 | $catchupText = '<i>' . $catchupText . '</i>'; |
|
871 | } |
||
872 | |||
873 | 3 | $this->output .= $catchupText; |
|
874 | 3 | $this->pos = $catchupPos; |
|
875 | |||
876 | 3 | if ($closeParagraph) |
|
877 | { |
||
878 | 1 | $this->outputParagraphEnd(); |
|
879 | } |
||
880 | |||
881 | 3 | return; |
|
882 | } |
||
883 | |||
884 | // Compute the amount of text to ignore at the end of the output |
||
885 | 154 | $ignorePos = $catchupPos; |
|
886 | 154 | $ignoreLen = 0; |
|
887 | |||
888 | // Ignore as many lines (including whitespace) as specified |
||
889 | 154 | while ($maxLines && --$ignorePos >= $this->pos) |
|
890 | { |
||
891 | 21 | $c = $this->text[$ignorePos]; |
|
892 | 21 | if (strpos(self::WHITESPACE, $c) === false) |
|
893 | { |
||
894 | 14 | break; |
|
895 | } |
||
896 | |||
897 | 12 | if ($c === "\n") |
|
898 | { |
||
899 | 10 | --$maxLines; |
|
900 | } |
||
901 | |||
902 | 12 | ++$ignoreLen; |
|
903 | } |
||
904 | |||
905 | // Adjust $catchupPos to ignore the text at the end |
||
906 | 154 | $catchupPos -= $ignoreLen; |
|
907 | |||
908 | // Break down the text in paragraphs if applicable |
||
909 | 154 | if ($this->context['flags'] & self::RULE_CREATE_PARAGRAPHS) |
|
910 | { |
||
911 | 15 | if (!$this->context['inParagraph']) |
|
912 | { |
||
913 | 13 | $this->outputWhitespace($catchupPos); |
|
914 | |||
915 | 13 | if ($catchupPos > $this->pos) |
|
916 | { |
||
917 | 10 | $this->outputParagraphStart($catchupPos); |
|
918 | } |
||
919 | } |
||
920 | |||
921 | // Look for a paragraph break in this text |
||
922 | 15 | $pbPos = strpos($this->text, "\n\n", $this->pos); |
|
923 | |||
924 | 15 | while ($pbPos !== false && $pbPos < $catchupPos) |
|
925 | { |
||
926 | 3 | $this->outputText($pbPos, 0, true); |
|
927 | 3 | $this->outputParagraphStart($catchupPos); |
|
928 | |||
929 | 3 | $pbPos = strpos($this->text, "\n\n", $this->pos); |
|
930 | } |
||
931 | } |
||
932 | |||
933 | // Capture, escape and output the text |
||
934 | 154 | if ($catchupPos > $this->pos) |
|
935 | { |
||
936 | 152 | $catchupText = htmlspecialchars( |
|
937 | 152 | substr($this->text, $this->pos, $catchupPos - $this->pos), |
|
938 | 152 | ENT_NOQUOTES, |
|
939 | 152 | 'UTF-8' |
|
940 | ); |
||
941 | |||
942 | // Format line breaks if applicable |
||
943 | 152 | if (($this->context['flags'] & self::RULES_AUTO_LINEBREAKS) === self::RULE_ENABLE_AUTO_BR) |
|
944 | { |
||
945 | 21 | $catchupText = str_replace("\n", "<br/>\n", $catchupText); |
|
946 | } |
||
947 | |||
948 | 152 | $this->output .= $catchupText; |
|
949 | } |
||
950 | |||
951 | // Close the paragraph if applicable |
||
952 | 154 | if ($closeParagraph) |
|
953 | { |
||
954 | 14 | $this->outputParagraphEnd(); |
|
955 | } |
||
956 | |||
957 | // Add the ignored text if applicable |
||
958 | 154 | if ($ignoreLen) |
|
959 | { |
||
960 | 12 | $this->output .= substr($this->text, $catchupPos, $ignoreLen); |
|
961 | } |
||
962 | |||
963 | // Move the cursor past the text |
||
964 | 154 | $this->pos = $catchupPos + $ignoreLen; |
|
965 | 154 | } |
|
966 | |||
967 | /** |
||
968 | * Output a linebreak tag |
||
969 | * |
||
970 | * @param Tag $tag |
||
971 | * @return void |
||
972 | */ |
||
973 | 6 | protected function outputBrTag(Tag $tag) |
|
974 | { |
||
975 | 6 | $this->outputText($tag->getPos(), 0, false); |
|
976 | 6 | $this->output .= '<br/>'; |
|
977 | 6 | } |
|
978 | |||
979 | /** |
||
980 | * Output an ignore tag |
||
981 | * |
||
982 | * @param Tag $tag |
||
983 | * @return void |
||
984 | */ |
||
985 | 18 | protected function outputIgnoreTag(Tag $tag) |
|
986 | { |
||
987 | 18 | $tagPos = $tag->getPos(); |
|
988 | 18 | $tagLen = $tag->getLen(); |
|
989 | |||
990 | // Capture the text to ignore |
||
991 | 18 | $ignoreText = substr($this->text, $tagPos, $tagLen); |
|
992 | |||
993 | // Catch up with the tag's position then output the tag |
||
994 | 18 | $this->outputText($tagPos, 0, false); |
|
995 | 18 | $this->output .= '<i>' . htmlspecialchars($ignoreText, ENT_NOQUOTES, 'UTF-8') . '</i>'; |
|
996 | 18 | $this->isRich = true; |
|
997 | |||
998 | // Move the cursor past this tag |
||
999 | 18 | $this->pos = $tagPos + $tagLen; |
|
1000 | 18 | } |
|
1001 | |||
1002 | /** |
||
1003 | * Start a paragraph between current position and given position, if applicable |
||
1004 | * |
||
1005 | * @param integer $maxPos Rightmost position at which the paragraph can be opened |
||
1006 | * @return void |
||
1007 | */ |
||
1008 | 140 | protected function outputParagraphStart($maxPos) |
|
1009 | { |
||
1010 | // Do nothing if we're already in a paragraph, or if we don't use paragraphs |
||
1011 | 140 | if ($this->context['inParagraph'] |
|
1012 | 140 | || !($this->context['flags'] & self::RULE_CREATE_PARAGRAPHS)) |
|
1013 | { |
||
1014 | 130 | return; |
|
1015 | } |
||
1016 | |||
1017 | // Output the whitespace between $this->pos and $maxPos if applicable |
||
1018 | 16 | $this->outputWhitespace($maxPos); |
|
1019 | |||
1020 | // Open the paragraph, but only if it's not at the very end of the text |
||
1021 | 16 | if ($this->pos < $this->textLen) |
|
1022 | { |
||
1023 | 16 | $this->output .= '<p>'; |
|
1024 | 16 | $this->context['inParagraph'] = true; |
|
1025 | } |
||
1026 | 16 | } |
|
1027 | |||
1028 | /** |
||
1029 | * Close current paragraph at current position if applicable |
||
1030 | * |
||
1031 | * @return void |
||
1032 | */ |
||
1033 | 18 | protected function outputParagraphEnd() |
|
1034 | { |
||
1035 | // Do nothing if we're not in a paragraph |
||
1036 | 18 | if (!$this->context['inParagraph']) |
|
1037 | { |
||
1038 | 3 | return; |
|
1039 | } |
||
1040 | |||
1041 | 16 | $this->output .= '</p>'; |
|
1042 | 16 | $this->context['inParagraph'] = false; |
|
1043 | 16 | } |
|
1044 | |||
1045 | /** |
||
1046 | * Output the content of a verbatim tag |
||
1047 | * |
||
1048 | * @param Tag $tag |
||
1049 | * @return void |
||
1050 | */ |
||
1051 | 4 | protected function outputVerbatim(Tag $tag) |
|
1052 | { |
||
1053 | 4 | $flags = $this->context['flags']; |
|
1054 | 4 | $this->context['flags'] = $tag->getFlags(); |
|
1055 | 4 | $this->outputText($this->currentTag->getPos() + $this->currentTag->getLen(), 0, false); |
|
1056 | 4 | $this->context['flags'] = $flags; |
|
1057 | 4 | } |
|
1058 | |||
1059 | /** |
||
1060 | * Skip as much whitespace after current position as possible |
||
1061 | * |
||
1062 | * @param integer $maxPos Rightmost character to be skipped |
||
1063 | * @return void |
||
1064 | */ |
||
1065 | 18 | protected function outputWhitespace($maxPos) |
|
1066 | { |
||
1067 | 18 | if ($maxPos > $this->pos) |
|
1068 | { |
||
1069 | 13 | $spn = strspn($this->text, self::WHITESPACE, $this->pos, $maxPos - $this->pos); |
|
1070 | |||
1071 | 13 | if ($spn) |
|
1072 | { |
||
1073 | 6 | $this->output .= substr($this->text, $this->pos, $spn); |
|
1074 | 6 | $this->pos += $spn; |
|
1075 | } |
||
1076 | } |
||
1077 | 18 | } |
|
1078 | |||
1079 | //========================================================================== |
||
1080 | // Plugins handling |
||
1081 | //========================================================================== |
||
1082 | |||
1083 | /** |
||
1084 | * Disable a plugin |
||
1085 | * |
||
1086 | * @param string $pluginName Name of the plugin |
||
1087 | * @return void |
||
1088 | */ |
||
1089 | 5 | public function disablePlugin($pluginName) |
|
1090 | { |
||
1091 | 5 | if (isset($this->pluginsConfig[$pluginName])) |
|
1092 | { |
||
1093 | // Copy the plugin's config to remove the reference |
||
1094 | 4 | $pluginConfig = $this->pluginsConfig[$pluginName]; |
|
1095 | 4 | unset($this->pluginsConfig[$pluginName]); |
|
1096 | |||
1097 | // Update the value and replace the plugin's config |
||
1098 | 4 | $pluginConfig['isDisabled'] = true; |
|
1099 | 4 | $this->pluginsConfig[$pluginName] = $pluginConfig; |
|
1100 | } |
||
1101 | 5 | } |
|
1102 | |||
1103 | /** |
||
1104 | * Enable a plugin |
||
1105 | * |
||
1106 | * @param string $pluginName Name of the plugin |
||
1107 | * @return void |
||
1108 | */ |
||
1109 | 2 | public function enablePlugin($pluginName) |
|
1110 | { |
||
1111 | 2 | if (isset($this->pluginsConfig[$pluginName])) |
|
1112 | { |
||
1113 | 1 | $this->pluginsConfig[$pluginName]['isDisabled'] = false; |
|
1114 | } |
||
1115 | 2 | } |
|
1116 | |||
1117 | /** |
||
1118 | * Execute given plugin |
||
1119 | * |
||
1120 | * @param string $pluginName Plugin's name |
||
1121 | * @return void |
||
1122 | */ |
||
1123 | 176 | protected function executePluginParser($pluginName) |
|
1124 | { |
||
1125 | 176 | $pluginConfig = $this->pluginsConfig[$pluginName]; |
|
1126 | 176 | if (isset($pluginConfig['quickMatch']) && strpos($this->text, $pluginConfig['quickMatch']) === false) |
|
1127 | { |
||
1128 | 1 | return; |
|
1129 | } |
||
1130 | |||
1131 | 175 | $matches = []; |
|
1132 | 175 | if (isset($pluginConfig['regexp'])) |
|
1133 | { |
||
1134 | 6 | $matches = $this->getMatches($pluginConfig['regexp'], $pluginConfig['regexpLimit']); |
|
1135 | 6 | if (empty($matches)) |
|
1136 | { |
||
1137 | 1 | return; |
|
1138 | } |
||
1139 | } |
||
1140 | |||
1141 | // Execute the plugin's parser, which will add tags via $this->addStartTag() and others |
||
1142 | 174 | call_user_func($this->getPluginParser($pluginName), $this->text, $matches); |
|
1143 | 174 | } |
|
1144 | |||
1145 | /** |
||
1146 | * Execute all the plugins |
||
1147 | * |
||
1148 | * @return void |
||
1149 | */ |
||
1150 | 186 | protected function executePluginParsers() |
|
1151 | { |
||
1152 | 186 | foreach ($this->pluginsConfig as $pluginName => $pluginConfig) |
|
1153 | { |
||
1154 | 177 | if (empty($pluginConfig['isDisabled'])) |
|
1155 | { |
||
1156 | 177 | $this->executePluginParser($pluginName); |
|
1157 | } |
||
1158 | } |
||
1159 | 186 | } |
|
1160 | |||
1161 | /** |
||
1162 | * Execute given regexp and returns as many matches as given limit |
||
1163 | * |
||
1164 | * @param string $regexp |
||
1165 | * @param integer $limit |
||
1166 | * @return array |
||
1167 | */ |
||
1168 | 6 | protected function getMatches($regexp, $limit) |
|
1169 | { |
||
1170 | 6 | $cnt = preg_match_all($regexp, $this->text, $matches, PREG_SET_ORDER | PREG_OFFSET_CAPTURE); |
|
1171 | 6 | if ($cnt > $limit) |
|
1172 | { |
||
1173 | 2 | $matches = array_slice($matches, 0, $limit); |
|
1174 | } |
||
1175 | |||
1176 | 6 | return $matches; |
|
1177 | } |
||
1178 | |||
1179 | /** |
||
1180 | * Get the cached callback for given plugin's parser |
||
1181 | * |
||
1182 | * @param string $pluginName Plugin's name |
||
1183 | * @return callable |
||
1184 | */ |
||
1185 | 174 | protected function getPluginParser($pluginName) |
|
1186 | { |
||
1187 | // Cache a new instance of this plugin's parser if there isn't one already |
||
1188 | 174 | if (!isset($this->pluginParsers[$pluginName])) |
|
1189 | { |
||
1190 | 1 | $pluginConfig = $this->pluginsConfig[$pluginName]; |
|
1191 | 1 | $className = (isset($pluginConfig['className'])) |
|
1192 | 1 | ? $pluginConfig['className'] |
|
1193 | 1 | : 's9e\\TextFormatter\\Plugins\\' . $pluginName . '\\Parser'; |
|
1194 | |||
1195 | // Register the parser as a callback |
||
1196 | 1 | $this->pluginParsers[$pluginName] = [new $className($this, $pluginConfig), 'parse']; |
|
1197 | } |
||
1198 | |||
1199 | 174 | return $this->pluginParsers[$pluginName]; |
|
1200 | } |
||
1201 | |||
1202 | /** |
||
1203 | * Register a parser |
||
1204 | * |
||
1205 | * Can be used to add a new parser with no plugin config, or pre-generate a parser for an |
||
1206 | * existing plugin |
||
1207 | * |
||
1208 | * @param string $pluginName |
||
1209 | * @param callback $parser |
||
1210 | * @return void |
||
1211 | */ |
||
1212 | 169 | public function registerParser($pluginName, $parser, $regexp = null, $limit = PHP_INT_MAX) |
|
1213 | { |
||
1214 | 169 | if (!is_callable($parser)) |
|
1215 | { |
||
1216 | 1 | throw new InvalidArgumentException('Argument 1 passed to ' . __METHOD__ . ' must be a valid callback'); |
|
1217 | } |
||
1218 | // Create an empty config for this plugin to ensure it is executed |
||
1219 | 168 | if (!isset($this->pluginsConfig[$pluginName])) |
|
1220 | { |
||
1221 | 167 | $this->pluginsConfig[$pluginName] = []; |
|
1222 | } |
||
1223 | 168 | if (isset($regexp)) |
|
1224 | { |
||
1225 | 2 | $this->pluginsConfig[$pluginName]['regexp'] = $regexp; |
|
1226 | 2 | $this->pluginsConfig[$pluginName]['regexpLimit'] = $limit; |
|
1227 | } |
||
1228 | 168 | $this->pluginParsers[$pluginName] = $parser; |
|
1229 | 168 | } |
|
1230 | |||
1231 | //========================================================================== |
||
1232 | // Rules handling |
||
1233 | //========================================================================== |
||
1234 | |||
1235 | /** |
||
1236 | * Apply closeAncestor rules associated with given tag |
||
1237 | * |
||
1238 | * @param Tag $tag Tag |
||
1239 | * @return bool Whether a new tag has been added |
||
1240 | */ |
||
1241 | 133 | protected function closeAncestor(Tag $tag) |
|
1242 | { |
||
1243 | 133 | if (!empty($this->openTags)) |
|
1244 | { |
||
1245 | 62 | $tagName = $tag->getName(); |
|
1246 | 62 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1247 | |||
1248 | 62 | if (!empty($tagConfig['rules']['closeAncestor'])) |
|
1249 | { |
||
1250 | 5 | $i = count($this->openTags); |
|
1251 | |||
1252 | 5 | while (--$i >= 0) |
|
1253 | { |
||
1254 | 5 | $ancestor = $this->openTags[$i]; |
|
1255 | 5 | $ancestorName = $ancestor->getName(); |
|
1256 | |||
1257 | 5 | if (isset($tagConfig['rules']['closeAncestor'][$ancestorName])) |
|
1258 | { |
||
1259 | 4 | ++$this->currentFixingCost; |
|
1260 | |||
1261 | // We have to close this ancestor. First we reinsert this tag... |
||
1262 | 4 | $this->tagStack[] = $tag; |
|
1263 | |||
1264 | // ...then we add a new end tag for it with a better priority |
||
1265 | 4 | $this->addMagicEndTag($ancestor, $tag->getPos(), $tag->getSortPriority() - 1); |
|
1266 | |||
1267 | 4 | return true; |
|
1268 | } |
||
1269 | } |
||
1270 | } |
||
1271 | } |
||
1272 | |||
1273 | 133 | return false; |
|
1274 | } |
||
1275 | |||
1276 | /** |
||
1277 | * Apply closeParent rules associated with given tag |
||
1278 | * |
||
1279 | * @param Tag $tag Tag |
||
1280 | * @return bool Whether a new tag has been added |
||
1281 | */ |
||
1282 | 133 | protected function closeParent(Tag $tag) |
|
1283 | { |
||
1284 | 133 | if (!empty($this->openTags)) |
|
1285 | { |
||
1286 | 66 | $tagName = $tag->getName(); |
|
1287 | 66 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1288 | |||
1289 | 66 | if (!empty($tagConfig['rules']['closeParent'])) |
|
1290 | { |
||
1291 | 7 | $parent = end($this->openTags); |
|
1292 | 7 | $parentName = $parent->getName(); |
|
1293 | |||
1294 | 7 | if (isset($tagConfig['rules']['closeParent'][$parentName])) |
|
1295 | { |
||
1296 | 6 | ++$this->currentFixingCost; |
|
1297 | |||
1298 | // We have to close that parent. First we reinsert the tag... |
||
1299 | 6 | $this->tagStack[] = $tag; |
|
1300 | |||
1301 | // ...then we add a new end tag for it with a better priority |
||
1302 | 6 | $this->addMagicEndTag($parent, $tag->getPos(), $tag->getSortPriority() - 1); |
|
1303 | |||
1304 | 6 | return true; |
|
1305 | } |
||
1306 | } |
||
1307 | } |
||
1308 | |||
1309 | 133 | return false; |
|
1310 | } |
||
1311 | |||
1312 | /** |
||
1313 | * Apply the createChild rules associated with given tag |
||
1314 | * |
||
1315 | * @param Tag $tag Tag |
||
1316 | * @return void |
||
1317 | */ |
||
1318 | 134 | protected function createChild(Tag $tag) |
|
1319 | { |
||
1320 | 134 | $tagConfig = $this->tagsConfig[$tag->getName()]; |
|
1321 | 134 | if (isset($tagConfig['rules']['createChild'])) |
|
1322 | { |
||
1323 | 3 | $priority = -1000; |
|
1324 | 3 | $tagPos = $this->pos + strspn($this->text, " \n\r\t", $this->pos); |
|
1325 | 3 | foreach ($tagConfig['rules']['createChild'] as $tagName) |
|
1326 | { |
||
1327 | 3 | $this->addStartTag($tagName, $tagPos, 0, ++$priority); |
|
1328 | } |
||
1329 | } |
||
1330 | 134 | } |
|
1331 | |||
1332 | /** |
||
1333 | * Apply fosterParent rules associated with given tag |
||
1334 | * |
||
1335 | * NOTE: this rule has the potential for creating an unbounded loop, either if a tag tries to |
||
1336 | * foster itself or two or more tags try to foster each other in a loop. We mitigate the |
||
1337 | * risk by preventing a tag from creating a child of itself (the parent still gets closed) |
||
1338 | * and by checking and increasing the currentFixingCost so that a loop of multiple tags |
||
1339 | * do not run indefinitely. The default tagLimit and nestingLimit also serve to prevent the |
||
1340 | * loop from running indefinitely |
||
1341 | * |
||
1342 | * @param Tag $tag Tag |
||
1343 | * @return bool Whether a new tag has been added |
||
1344 | */ |
||
1345 | 133 | protected function fosterParent(Tag $tag) |
|
1346 | { |
||
1347 | 133 | if (!empty($this->openTags)) |
|
1348 | { |
||
1349 | 69 | $tagName = $tag->getName(); |
|
1350 | 69 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1351 | |||
1352 | 69 | if (!empty($tagConfig['rules']['fosterParent'])) |
|
1353 | { |
||
1354 | 13 | $parent = end($this->openTags); |
|
1355 | 13 | $parentName = $parent->getName(); |
|
1356 | |||
1357 | 13 | if (isset($tagConfig['rules']['fosterParent'][$parentName])) |
|
1358 | { |
||
1359 | 12 | if ($parentName !== $tagName && $this->currentFixingCost < $this->maxFixingCost) |
|
1360 | { |
||
1361 | 11 | $this->addFosterTag($tag, $parent); |
|
1362 | } |
||
1363 | |||
1364 | // Reinsert current tag |
||
1365 | 12 | $this->tagStack[] = $tag; |
|
1366 | |||
1367 | // And finally close its parent with a priority that ensures it is processed |
||
1368 | // before this tag |
||
1369 | 12 | $this->addMagicEndTag($parent, $tag->getPos(), $tag->getSortPriority() - 1); |
|
1370 | |||
1371 | // Adjust the fixing cost to account for the additional tags/processing |
||
1372 | 12 | $this->currentFixingCost += 4; |
|
1373 | |||
1374 | 12 | return true; |
|
1375 | } |
||
1376 | } |
||
1377 | } |
||
1378 | |||
1379 | 133 | return false; |
|
1380 | } |
||
1381 | |||
1382 | /** |
||
1383 | * Apply requireAncestor rules associated with given tag |
||
1384 | * |
||
1385 | * @param Tag $tag Tag |
||
1386 | * @return bool Whether this tag has an unfulfilled requireAncestor requirement |
||
1387 | */ |
||
1388 | 136 | protected function requireAncestor(Tag $tag) |
|
1389 | { |
||
1390 | 136 | $tagName = $tag->getName(); |
|
1391 | 136 | $tagConfig = $this->tagsConfig[$tagName]; |
|
1392 | |||
1393 | 136 | if (isset($tagConfig['rules']['requireAncestor'])) |
|
1394 | { |
||
1395 | 3 | foreach ($tagConfig['rules']['requireAncestor'] as $ancestorName) |
|
1396 | { |
||
1397 | 3 | if (!empty($this->cntOpen[$ancestorName])) |
|
1398 | { |
||
1399 | 3 | return false; |
|
1400 | } |
||
1401 | } |
||
1402 | |||
1403 | 2 | $this->logger->err('Tag requires an ancestor', [ |
|
1404 | 2 | 'requireAncestor' => implode(',', $tagConfig['rules']['requireAncestor']), |
|
1405 | 2 | 'tag' => $tag |
|
1406 | ]); |
||
1407 | |||
1408 | 2 | return true; |
|
1409 | } |
||
1410 | |||
1411 | 134 | return false; |
|
1412 | } |
||
1413 | |||
1414 | //========================================================================== |
||
1415 | // Tag processing |
||
1416 | //========================================================================== |
||
1417 | |||
1418 | /** |
||
1419 | * Create and add a copy of a tag as a child of a given tag |
||
1420 | * |
||
1421 | * @param Tag $tag Current tag |
||
1422 | * @param Tag $fosterTag Tag to foster |
||
1423 | * @return void |
||
1424 | */ |
||
1425 | 11 | protected function addFosterTag(Tag $tag, Tag $fosterTag) |
|
1426 | { |
||
1427 | 11 | list($childPos, $childPrio) = $this->getMagicStartCoords($tag->getPos() + $tag->getLen()); |
|
1428 | |||
1429 | // Add a 0-width copy of the parent tag after this tag and make it depend on this tag |
||
1430 | 11 | $childTag = $this->addCopyTag($fosterTag, $childPos, 0, $childPrio); |
|
1431 | 11 | $tag->cascadeInvalidationTo($childTag); |
|
1432 | 11 | } |
|
1433 | |||
1434 | /** |
||
1435 | * Create and add an end tag for given start tag at given position |
||
1436 | * |
||
1437 | * @param Tag $startTag Start tag |
||
1438 | * @param integer $tagPos End tag's position (will be adjusted for whitespace if applicable) |
||
1439 | * @param integer $prio End tag's priority |
||
1440 | * @return Tag |
||
1441 | */ |
||
1442 | 33 | protected function addMagicEndTag(Tag $startTag, $tagPos, $prio = 0) |
|
1443 | { |
||
1444 | 33 | $tagName = $startTag->getName(); |
|
1445 | |||
1446 | // Adjust the end tag's position if whitespace is to be minimized |
||
1447 | 33 | if (($this->currentTag->getFlags() | $startTag->getFlags()) & self::RULE_IGNORE_WHITESPACE) |
|
1448 | { |
||
1449 | 3 | $tagPos = $this->getMagicEndPos($tagPos); |
|
1450 | } |
||
1451 | |||
1452 | // Add a 0-width end tag that is paired with the given start tag |
||
1453 | 33 | $endTag = $this->addEndTag($tagName, $tagPos, 0, $prio); |
|
1454 | 33 | $endTag->pairWith($startTag); |
|
1455 | |||
1456 | 33 | return $endTag; |
|
1457 | } |
||
1458 | |||
1459 | /** |
||
1460 | * Compute the position of a magic end tag, adjusted for whitespace |
||
1461 | * |
||
1462 | * @param integer $tagPos Rightmost possible position for the tag |
||
1463 | * @return integer |
||
1464 | */ |
||
1465 | 5 | protected function getMagicEndPos($tagPos) |
|
1466 | { |
||
1467 | // Back up from given position to the cursor's position until we find a character that |
||
1468 | // is not whitespace |
||
1469 | 5 | while ($tagPos > $this->pos && strpos(self::WHITESPACE, $this->text[$tagPos - 1]) !== false) |
|
1470 | { |
||
1471 | 5 | --$tagPos; |
|
1472 | } |
||
1473 | |||
1474 | 5 | return $tagPos; |
|
1475 | } |
||
1476 | |||
1477 | /** |
||
1478 | * Compute the position and priority of a magic start tag, adjusted for whitespace |
||
1479 | * |
||
1480 | * @param integer $tagPos Leftmost possible position for the tag |
||
1481 | * @return integer[] [Tag pos, priority] |
||
1482 | */ |
||
1483 | 11 | protected function getMagicStartCoords($tagPos) |
|
1484 | { |
||
1485 | 11 | if (empty($this->tagStack)) |
|
1486 | { |
||
1487 | // Set the next position outside the text boundaries |
||
1488 | 3 | $nextPos = $this->textLen + 1; |
|
1489 | 3 | $nextPrio = 0; |
|
1490 | } |
||
1491 | else |
||
1492 | { |
||
1493 | 10 | $nextTag = end($this->tagStack); |
|
1494 | 10 | $nextPos = $nextTag->getPos(); |
|
1495 | 10 | $nextPrio = $nextTag->getSortPriority(); |
|
1496 | } |
||
1497 | |||
1498 | // Find the first non-whitespace position before next tag or the end of text |
||
1499 | 11 | while ($tagPos < $nextPos && strpos(self::WHITESPACE, $this->text[$tagPos]) !== false) |
|
1500 | { |
||
1501 | 1 | ++$tagPos; |
|
1502 | } |
||
1503 | |||
1504 | // Set a priority that ensures this tag appears before the next tag |
||
1505 | 11 | $prio = ($tagPos === $nextPos) ? $nextPrio - 1 : 0; |
|
1506 | |||
1507 | 11 | return [$tagPos, $prio]; |
|
1508 | } |
||
1509 | |||
1510 | /** |
||
1511 | * Test whether given start tag is immediately followed by a closing tag |
||
1512 | * |
||
1513 | * @param Tag $tag Start tag |
||
1514 | * @return bool |
||
1515 | */ |
||
1516 | 3 | protected function isFollowedByClosingTag(Tag $tag) |
|
1520 | |||
1521 | /** |
||
1522 | * Process all tags in the stack |
||
1523 | * |
||
1524 | * @return void |
||
1525 | */ |
||
1526 | 175 | protected function processTags() |
|
1527 | { |
||
1528 | 175 | if (empty($this->tagStack)) |
|
1529 | { |
||
1530 | 22 | return; |
|
1531 | } |
||
1532 | |||
1533 | // Initialize the count tables |
||
1534 | 153 | foreach (array_keys($this->tagsConfig) as $tagName) |
|
1535 | { |
||
1536 | 139 | $this->cntOpen[$tagName] = 0; |
|
1537 | 139 | $this->cntTotal[$tagName] = 0; |
|
1538 | } |
||
1539 | |||
1540 | // Process the tag stack, close tags that were left open and repeat until done |
||
1541 | do |
||
1542 | { |
||
1543 | 153 | while (!empty($this->tagStack)) |
|
1544 | { |
||
1545 | 153 | if (!$this->tagStackIsSorted) |
|
1546 | { |
||
1547 | 153 | $this->sortTags(); |
|
1548 | } |
||
1549 | |||
1550 | 153 | $this->currentTag = array_pop($this->tagStack); |
|
1551 | 153 | $this->processCurrentTag(); |
|
1552 | } |
||
1553 | |||
1554 | // Close tags that were left open |
||
1555 | 153 | foreach ($this->openTags as $startTag) |
|
1556 | { |
||
1557 | // NOTE: we add tags in hierarchical order (ancestors to descendants) but since |
||
1558 | // the stack is processed in LIFO order, it means that tags get closed in |
||
1559 | // the correct order, from descendants to ancestors |
||
1560 | 17 | $this->addMagicEndTag($startTag, $this->textLen); |
|
1561 | } |
||
1562 | } |
||
1563 | 153 | while (!empty($this->tagStack)); |
|
1565 | |||
1566 | /** |
||
1567 | * Process current tag |
||
1568 | * |
||
1569 | * @return void |
||
1570 | */ |
||
1571 | 153 | protected function processCurrentTag() |
|
1657 | |||
1658 | /** |
||
1659 | * Process given start tag (including self-closing tags) at current position |
||
1660 | * |
||
1661 | * @param Tag $tag Start tag (including self-closing) |
||
1662 | * @return void |
||
1663 | */ |
||
1664 | 136 | protected function processStartTag(Tag $tag) |
|
1776 | |||
1777 | /** |
||
1778 | * Process given end tag at current position |
||
1779 | * |
||
1780 | * @param Tag $tag end tag |
||
1781 | * @return void |
||
1782 | */ |
||
1783 | 102 | protected function processEndTag(Tag $tag) |
|
1941 | |||
1942 | /** |
||
1943 | * Update counters and replace current context with its parent context |
||
1944 | * |
||
1945 | * @return void |
||
1946 | */ |
||
1947 | 101 | protected function popContext() |
|
1953 | |||
1954 | /** |
||
1955 | * Update counters and replace current context with a new context based on given tag |
||
1956 | * |
||
1957 | * If given tag is a self-closing tag, the context won't change |
||
1958 | * |
||
1959 | * @param Tag $tag Start tag (including self-closing) |
||
1960 | * @return void |
||
1961 | */ |
||
1962 | 134 | protected function pushContext(Tag $tag) |
|
2011 | |||
2012 | /** |
||
2013 | * Return whether given tag is allowed in current context |
||
2014 | * |
||
2015 | * @param string $tagName |
||
2016 | * @return bool |
||
2017 | */ |
||
2018 | 136 | protected function tagIsAllowed($tagName) |
|
2024 | |||
2025 | //========================================================================== |
||
2026 | // Tag stack |
||
2027 | //========================================================================== |
||
2028 | |||
2029 | /** |
||
2030 | * Add a start tag |
||
2031 | * |
||
2032 | * @param string $name Name of the tag |
||
2033 | * @param integer $pos Position of the tag in the text |
||
2034 | * @param integer $len Length of text consumed by the tag |
||
2035 | * @param integer $prio Tag's priority |
||
2036 | * @return Tag |
||
2037 | */ |
||
2038 | 119 | public function addStartTag($name, $pos, $len, $prio = 0) |
|
2042 | |||
2043 | /** |
||
2044 | * Add an end tag |
||
2045 | * |
||
2046 | * @param string $name Name of the tag |
||
2047 | * @param integer $pos Position of the tag in the text |
||
2048 | * @param integer $len Length of text consumed by the tag |
||
2049 | * @param integer $prio Tag's priority |
||
2050 | * @return Tag |
||
2051 | */ |
||
2052 | 107 | public function addEndTag($name, $pos, $len, $prio = 0) |
|
2056 | |||
2057 | /** |
||
2058 | * Add a self-closing tag |
||
2059 | * |
||
2060 | * @param string $name Name of the tag |
||
2061 | * @param integer $pos Position of the tag in the text |
||
2062 | * @param integer $len Length of text consumed by the tag |
||
2063 | * @param integer $prio Tag's priority |
||
2064 | * @return Tag |
||
2065 | */ |
||
2066 | 66 | public function addSelfClosingTag($name, $pos, $len, $prio = 0) |
|
2070 | |||
2071 | /** |
||
2072 | * Add a 0-width "br" tag to force a line break at given position |
||
2073 | * |
||
2074 | * @param integer $pos Position of the tag in the text |
||
2075 | * @param integer $prio Tag's priority |
||
2076 | * @return Tag |
||
2077 | */ |
||
2078 | 9 | public function addBrTag($pos, $prio = 0) |
|
2082 | |||
2083 | /** |
||
2084 | * Add an "ignore" tag |
||
2085 | * |
||
2086 | * @param integer $pos Position of the tag in the text |
||
2087 | * @param integer $len Length of text consumed by the tag |
||
2088 | * @param integer $prio Tag's priority |
||
2089 | * @return Tag |
||
2090 | */ |
||
2091 | 12 | public function addIgnoreTag($pos, $len, $prio = 0) |
|
2095 | |||
2096 | /** |
||
2097 | * Add a paragraph break at given position |
||
2098 | * |
||
2099 | * Uses a zero-width tag that is actually never output in the result |
||
2100 | * |
||
2101 | * @param integer $pos Position of the tag in the text |
||
2102 | * @param integer $prio Tag's priority |
||
2103 | * @return Tag |
||
2104 | */ |
||
2105 | 5 | public function addParagraphBreak($pos, $prio = 0) |
|
2109 | |||
2110 | /** |
||
2111 | * Add a copy of given tag at given position and length |
||
2112 | * |
||
2113 | * @param Tag $tag Original tag |
||
2114 | * @param integer $pos Copy's position |
||
2115 | * @param integer $len Copy's length |
||
2116 | * @param integer $prio Copy's priority (same as original by default) |
||
2117 | * @return Tag Copy tag |
||
2118 | */ |
||
2119 | 22 | public function addCopyTag(Tag $tag, $pos, $len, $prio = null) |
|
2130 | |||
2131 | /** |
||
2132 | * Add a tag |
||
2133 | * |
||
2134 | * @param integer $type Tag's type |
||
2135 | * @param string $name Name of the tag |
||
2136 | * @param integer $pos Position of the tag in the text |
||
2137 | * @param integer $len Length of text consumed by the tag |
||
2138 | * @param integer $prio Tag's priority |
||
2139 | * @return Tag |
||
2140 | */ |
||
2141 | 179 | protected function addTag($type, $name, $pos, $len, $prio) |
|
2180 | |||
2181 | /** |
||
2182 | * Insert given tag in the tag stack |
||
2183 | * |
||
2184 | * @param Tag $tag |
||
2185 | * @return void |
||
2186 | */ |
||
2187 | 172 | protected function insertTag(Tag $tag) |
|
2205 | |||
2206 | /** |
||
2207 | * Add a pair of tags |
||
2208 | * |
||
2209 | * @param string $name Name of the tags |
||
2210 | * @param integer $startPos Position of the start tag |
||
2211 | * @param integer $startLen Length of the start tag |
||
2212 | * @param integer $endPos Position of the start tag |
||
2213 | * @param integer $endLen Length of the start tag |
||
2214 | * @param integer $prio Start tag's priority (the end tag will be set to minus that value) |
||
2215 | * @return Tag Start tag |
||
2216 | */ |
||
2217 | 25 | public function addTagPair($name, $startPos, $startLen, $endPos, $endLen, $prio = 0) |
|
2226 | |||
2227 | /** |
||
2228 | * Add a tag that represents a verbatim copy of the original text |
||
2229 | * |
||
2230 | * @param integer $pos Position of the tag in the text |
||
2231 | * @param integer $len Length of text consumed by the tag |
||
2232 | * @param integer $prio Tag's priority |
||
2233 | * @return Tag |
||
2234 | */ |
||
2235 | 4 | public function addVerbatim($pos, $len, $prio = 0) |
|
2239 | |||
2240 | /** |
||
2241 | * Sort tags by position and precedence |
||
2242 | * |
||
2243 | * @return void |
||
2244 | */ |
||
2245 | 159 | protected function sortTags() |
|
2250 | |||
2251 | /** |
||
2252 | * sortTags() callback |
||
2253 | * |
||
2254 | * Tags are stored as a stack, in LIFO order. We sort tags by position _descending_ so that they |
||
2255 | * are processed in the order they appear in the text. |
||
2256 | * |
||
2257 | * @param Tag $a First tag to compare |
||
2258 | * @param Tag $b Second tag to compare |
||
2259 | * @return integer |
||
2260 | */ |
||
2261 | 122 | protected static function compareTags(Tag $a, Tag $b) |
|
2314 | } |