Deprecate Parser::fetchFile() since it's unused
[lhc/web/wiklou.git] / includes / parser / Parser.php
1 <?php
2 /**
3 * PHP parser that converts wiki markup to HTML.
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
19 *
20 * @file
21 * @ingroup Parser
22 */
23 use MediaWiki\Linker\LinkRenderer;
24 use MediaWiki\MediaWikiServices;
25 use Wikimedia\ScopedCallback;
26
27 /**
28 * @defgroup Parser Parser
29 */
30
31 /**
32 * PHP Parser - Processes wiki markup (which uses a more user-friendly
33 * syntax, such as "[[link]]" for making links), and provides a one-way
34 * transformation of that wiki markup it into (X)HTML output / markup
35 * (which in turn the browser understands, and can display).
36 *
37 * There are seven main entry points into the Parser class:
38 *
39 * - Parser::parse()
40 * produces HTML output
41 * - Parser::preSaveTransform()
42 * produces altered wiki markup
43 * - Parser::preprocess()
44 * removes HTML comments and expands templates
45 * - Parser::cleanSig() and Parser::cleanSigInSig()
46 * cleans a signature before saving it to preferences
47 * - Parser::getSection()
48 * return the content of a section from an article for section editing
49 * - Parser::replaceSection()
50 * replaces a section by number inside an article
51 * - Parser::getPreloadText()
52 * removes <noinclude> sections and <includeonly> tags
53 *
54 * Globals used:
55 * object: $wgContLang
56 *
57 * @warning $wgUser or $wgTitle or $wgRequest or $wgLang. Keep them away!
58 *
59 * @par Settings:
60 * $wgNamespacesWithSubpages
61 *
62 * @par Settings only within ParserOptions:
63 * $wgAllowExternalImages
64 * $wgAllowSpecialInclusion
65 * $wgInterwikiMagic
66 * $wgMaxArticleSize
67 *
68 * @ingroup Parser
69 */
70 class Parser {
71 /**
72 * Update this version number when the ParserOutput format
73 * changes in an incompatible way, so the parser cache
74 * can automatically discard old data.
75 */
76 const VERSION = '1.6.4';
77
78 /**
79 * Update this version number when the output of serialiseHalfParsedText()
80 * changes in an incompatible way
81 */
82 const HALF_PARSED_VERSION = 2;
83
84 # Flags for Parser::setFunctionHook
85 const SFH_NO_HASH = 1;
86 const SFH_OBJECT_ARGS = 2;
87
88 # Constants needed for external link processing
89 # Everything except bracket, space, or control characters
90 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
91 # as well as U+3000 is IDEOGRAPHIC SPACE for T21052
92 # \x{FFFD} is the Unicode replacement character, which Preprocessor_DOM
93 # uses to replace invalid HTML characters.
94 const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]';
95 # Simplified expression to match an IPv4 or IPv6 address, or
96 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
97 const EXT_LINK_ADDR = '(?:[0-9.]+|\\[(?i:[0-9a-f:.]+)\\]|[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}])';
98 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
99 // phpcs:ignore Generic.Files.LineLength
100 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]+)
101 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
102
103 # Regular expression for a non-newline space
104 const SPACE_NOT_NL = '(?:\t|&nbsp;|&\#0*160;|&\#[Xx]0*[Aa]0;|\p{Zs})';
105
106 # Flags for preprocessToDom
107 const PTD_FOR_INCLUSION = 1;
108
109 # Allowed values for $this->mOutputType
110 # Parameter to startExternalParse().
111 const OT_HTML = 1; # like parse()
112 const OT_WIKI = 2; # like preSaveTransform()
113 const OT_PREPROCESS = 3; # like preprocess()
114 const OT_MSG = 3;
115 const OT_PLAIN = 4; # like extractSections() - portions of the original are returned unchanged.
116
117 /**
118 * @var string Prefix and suffix for temporary replacement strings
119 * for the multipass parser.
120 *
121 * \x7f should never appear in input as it's disallowed in XML.
122 * Using it at the front also gives us a little extra robustness
123 * since it shouldn't match when butted up against identifier-like
124 * string constructs.
125 *
126 * Must not consist of all title characters, or else it will change
127 * the behavior of <nowiki> in a link.
128 *
129 * Must have a character that needs escaping in attributes, otherwise
130 * someone could put a strip marker in an attribute, to get around
131 * escaping quote marks, and break out of the attribute. Thus we add
132 * `'".
133 */
134 const MARKER_SUFFIX = "-QINU`\"'\x7f";
135 const MARKER_PREFIX = "\x7f'\"`UNIQ-";
136
137 # Markers used for wrapping the table of contents
138 const TOC_START = '<mw:toc>';
139 const TOC_END = '</mw:toc>';
140
141 /** @var int Assume that no output will later be saved this many seconds after parsing */
142 const MAX_TTS = 900;
143
144 # Persistent:
145 public $mTagHooks = [];
146 public $mTransparentTagHooks = [];
147 public $mFunctionHooks = [];
148 public $mFunctionSynonyms = [ 0 => [], 1 => [] ];
149 public $mFunctionTagHooks = [];
150 public $mStripList = [];
151 public $mDefaultStripList = [];
152 public $mVarCache = [];
153 public $mImageParams = [];
154 public $mImageParamsMagicArray = [];
155 public $mMarkerIndex = 0;
156 public $mFirstCall = true;
157
158 # Initialised by initialiseVariables()
159
160 /**
161 * @var MagicWordArray
162 */
163 public $mVariables;
164
165 /**
166 * @var MagicWordArray
167 */
168 public $mSubstWords;
169 # Initialised in constructor
170 public $mConf, $mExtLinkBracketedRegex, $mUrlProtocols;
171
172 # Initialized in getPreprocessor()
173 /** @var Preprocessor */
174 public $mPreprocessor;
175
176 # Cleared with clearState():
177 /**
178 * @var ParserOutput
179 */
180 public $mOutput;
181 public $mAutonumber;
182
183 /**
184 * @var StripState
185 */
186 public $mStripState;
187
188 public $mIncludeCount;
189 /**
190 * @var LinkHolderArray
191 */
192 public $mLinkHolders;
193
194 public $mLinkID;
195 public $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth;
196 public $mDefaultSort;
197 public $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
198 public $mExpensiveFunctionCount; # number of expensive parser function calls
199 public $mShowToc, $mForceTocPosition;
200
201 /**
202 * @var User
203 */
204 public $mUser; # User object; only used when doing pre-save transform
205
206 # Temporary
207 # These are variables reset at least once per parse regardless of $clearState
208
209 /**
210 * @var ParserOptions
211 */
212 public $mOptions;
213
214 /**
215 * @var Title
216 */
217 public $mTitle; # Title context, used for self-link rendering and similar things
218 public $mOutputType; # Output type, one of the OT_xxx constants
219 public $ot; # Shortcut alias, see setOutputType()
220 public $mRevisionObject; # The revision object of the specified revision ID
221 public $mRevisionId; # ID to display in {{REVISIONID}} tags
222 public $mRevisionTimestamp; # The timestamp of the specified revision ID
223 public $mRevisionUser; # User to display in {{REVISIONUSER}} tag
224 public $mRevisionSize; # Size to display in {{REVISIONSIZE}} variable
225 public $mRevIdForTs; # The revision ID which was used to fetch the timestamp
226 public $mInputSize = false; # For {{PAGESIZE}} on current page.
227
228 /**
229 * @var string Deprecated accessor for the strip marker prefix.
230 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
231 */
232 public $mUniqPrefix = self::MARKER_PREFIX;
233
234 /**
235 * @var array Array with the language name of each language link (i.e. the
236 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
237 * duplicate language links to the ParserOutput.
238 */
239 public $mLangLinkLanguages;
240
241 /**
242 * @var MapCacheLRU|null
243 * @since 1.24
244 *
245 * A cache of the current revisions of titles. Keys are $title->getPrefixedDbKey()
246 */
247 public $currentRevisionCache;
248
249 /**
250 * @var bool|string Recursive call protection.
251 * This variable should be treated as if it were private.
252 */
253 public $mInParse = false;
254
255 /** @var SectionProfiler */
256 protected $mProfiler;
257
258 /**
259 * @var LinkRenderer
260 */
261 protected $mLinkRenderer;
262
263 /**
264 * @param array $conf
265 */
266 public function __construct( $conf = [] ) {
267 $this->mConf = $conf;
268 $this->mUrlProtocols = wfUrlProtocols();
269 $this->mExtLinkBracketedRegex = '/\[(((?i)' . $this->mUrlProtocols . ')' .
270 self::EXT_LINK_ADDR .
271 self::EXT_LINK_URL_CLASS . '*)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F\\x{FFFD}]*?)\]/Su';
272 if ( isset( $conf['preprocessorClass'] ) ) {
273 $this->mPreprocessorClass = $conf['preprocessorClass'];
274 } elseif ( defined( 'HPHP_VERSION' ) ) {
275 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
276 $this->mPreprocessorClass = Preprocessor_Hash::class;
277 } elseif ( extension_loaded( 'domxml' ) ) {
278 # PECL extension that conflicts with the core DOM extension (T15770)
279 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
280 $this->mPreprocessorClass = Preprocessor_Hash::class;
281 } elseif ( extension_loaded( 'dom' ) ) {
282 $this->mPreprocessorClass = Preprocessor_DOM::class;
283 } else {
284 $this->mPreprocessorClass = Preprocessor_Hash::class;
285 }
286 wfDebug( __CLASS__ . ": using preprocessor: {$this->mPreprocessorClass}\n" );
287 }
288
289 /**
290 * Reduce memory usage to reduce the impact of circular references
291 */
292 public function __destruct() {
293 if ( isset( $this->mLinkHolders ) ) {
294 unset( $this->mLinkHolders );
295 }
296 foreach ( $this as $name => $value ) {
297 unset( $this->$name );
298 }
299 }
300
301 /**
302 * Allow extensions to clean up when the parser is cloned
303 */
304 public function __clone() {
305 $this->mInParse = false;
306
307 // T58226: When you create a reference "to" an object field, that
308 // makes the object field itself be a reference too (until the other
309 // reference goes out of scope). When cloning, any field that's a
310 // reference is copied as a reference in the new object. Both of these
311 // are defined PHP5 behaviors, as inconvenient as it is for us when old
312 // hooks from PHP4 days are passing fields by reference.
313 foreach ( [ 'mStripState', 'mVarCache' ] as $k ) {
314 // Make a non-reference copy of the field, then rebind the field to
315 // reference the new copy.
316 $tmp = $this->$k;
317 $this->$k =& $tmp;
318 unset( $tmp );
319 }
320
321 Hooks::run( 'ParserCloned', [ $this ] );
322 }
323
324 /**
325 * Do various kinds of initialisation on the first call of the parser
326 */
327 public function firstCallInit() {
328 if ( !$this->mFirstCall ) {
329 return;
330 }
331 $this->mFirstCall = false;
332
333 CoreParserFunctions::register( $this );
334 CoreTagHooks::register( $this );
335 $this->initialiseVariables();
336
337 // Avoid PHP 7.1 warning from passing $this by reference
338 $parser = $this;
339 Hooks::run( 'ParserFirstCallInit', [ &$parser ] );
340 }
341
342 /**
343 * Clear Parser state
344 *
345 * @private
346 */
347 public function clearState() {
348 if ( $this->mFirstCall ) {
349 $this->firstCallInit();
350 }
351 $this->mOutput = new ParserOutput;
352 $this->mOptions->registerWatcher( [ $this->mOutput, 'recordOption' ] );
353 $this->mAutonumber = 0;
354 $this->mIncludeCount = [];
355 $this->mLinkHolders = new LinkHolderArray( $this );
356 $this->mLinkID = 0;
357 $this->mRevisionObject = $this->mRevisionTimestamp =
358 $this->mRevisionId = $this->mRevisionUser = $this->mRevisionSize = null;
359 $this->mVarCache = [];
360 $this->mUser = null;
361 $this->mLangLinkLanguages = [];
362 $this->currentRevisionCache = null;
363
364 $this->mStripState = new StripState( $this );
365
366 # Clear these on every parse, T6549
367 $this->mTplRedirCache = $this->mTplDomCache = [];
368
369 $this->mShowToc = true;
370 $this->mForceTocPosition = false;
371 $this->mIncludeSizes = [
372 'post-expand' => 0,
373 'arg' => 0,
374 ];
375 $this->mPPNodeCount = 0;
376 $this->mGeneratedPPNodeCount = 0;
377 $this->mHighestExpansionDepth = 0;
378 $this->mDefaultSort = false;
379 $this->mHeadings = [];
380 $this->mDoubleUnderscores = [];
381 $this->mExpensiveFunctionCount = 0;
382
383 # Fix cloning
384 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) {
385 $this->mPreprocessor = null;
386 }
387
388 $this->mProfiler = new SectionProfiler();
389
390 // Avoid PHP 7.1 warning from passing $this by reference
391 $parser = $this;
392 Hooks::run( 'ParserClearState', [ &$parser ] );
393 }
394
395 /**
396 * Convert wikitext to HTML
397 * Do not call this function recursively.
398 *
399 * @param string $text Text we want to parse
400 * @param Title $title
401 * @param ParserOptions $options
402 * @param bool $linestart
403 * @param bool $clearState
404 * @param int $revid Number to pass in {{REVISIONID}}
405 * @return ParserOutput A ParserOutput
406 */
407 public function parse(
408 $text, Title $title, ParserOptions $options,
409 $linestart = true, $clearState = true, $revid = null
410 ) {
411 if ( $clearState ) {
412 // We use U+007F DELETE to construct strip markers, so we have to make
413 // sure that this character does not occur in the input text.
414 $text = strtr( $text, "\x7f", "?" );
415 $magicScopeVariable = $this->lock();
416 }
417 // Strip U+0000 NULL (T159174)
418 $text = str_replace( "\000", '', $text );
419
420 $this->startParse( $title, $options, self::OT_HTML, $clearState );
421
422 $this->currentRevisionCache = null;
423 $this->mInputSize = strlen( $text );
424 if ( $this->mOptions->getEnableLimitReport() ) {
425 $this->mOutput->resetParseStartTime();
426 }
427
428 $oldRevisionId = $this->mRevisionId;
429 $oldRevisionObject = $this->mRevisionObject;
430 $oldRevisionTimestamp = $this->mRevisionTimestamp;
431 $oldRevisionUser = $this->mRevisionUser;
432 $oldRevisionSize = $this->mRevisionSize;
433 if ( $revid !== null ) {
434 $this->mRevisionId = $revid;
435 $this->mRevisionObject = null;
436 $this->mRevisionTimestamp = null;
437 $this->mRevisionUser = null;
438 $this->mRevisionSize = null;
439 }
440
441 // Avoid PHP 7.1 warning from passing $this by reference
442 $parser = $this;
443 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
444 # No more strip!
445 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
446 $text = $this->internalParse( $text );
447 Hooks::run( 'ParserAfterParse', [ &$parser, &$text, &$this->mStripState ] );
448
449 $text = $this->internalParseHalfParsed( $text, true, $linestart );
450
451 /**
452 * A converted title will be provided in the output object if title and
453 * content conversion are enabled, the article text does not contain
454 * a conversion-suppressing double-underscore tag, and no
455 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
456 * automatic link conversion.
457 */
458 if ( !( $options->getDisableTitleConversion()
459 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
460 || isset( $this->mDoubleUnderscores['notitleconvert'] )
461 || $this->mOutput->getDisplayTitle() !== false )
462 ) {
463 $convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
464 if ( $convruletitle ) {
465 $this->mOutput->setTitleText( $convruletitle );
466 } else {
467 $titleText = $this->getConverterLanguage()->convertTitle( $title );
468 $this->mOutput->setTitleText( $titleText );
469 }
470 }
471
472 # Compute runtime adaptive expiry if set
473 $this->mOutput->finalizeAdaptiveCacheExpiry();
474
475 # Warn if too many heavyweight parser functions were used
476 if ( $this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit() ) {
477 $this->limitationWarn( 'expensive-parserfunction',
478 $this->mExpensiveFunctionCount,
479 $this->mOptions->getExpensiveParserFunctionLimit()
480 );
481 }
482
483 # Information on limits, for the benefit of users who try to skirt them
484 if ( $this->mOptions->getEnableLimitReport() ) {
485 $text .= $this->makeLimitReport();
486 }
487
488 # Wrap non-interface parser output in a <div> so it can be targeted
489 # with CSS (T37247)
490 $class = $this->mOptions->getWrapOutputClass();
491 if ( $class !== false && !$this->mOptions->getInterfaceMessage() ) {
492 $text = Html::rawElement( 'div', [ 'class' => $class ], $text );
493 }
494
495 $this->mOutput->setText( $text );
496
497 $this->mRevisionId = $oldRevisionId;
498 $this->mRevisionObject = $oldRevisionObject;
499 $this->mRevisionTimestamp = $oldRevisionTimestamp;
500 $this->mRevisionUser = $oldRevisionUser;
501 $this->mRevisionSize = $oldRevisionSize;
502 $this->mInputSize = false;
503 $this->currentRevisionCache = null;
504
505 return $this->mOutput;
506 }
507
508 /**
509 * Set the limit report data in the current ParserOutput, and return the
510 * limit report HTML comment.
511 *
512 * @return string
513 */
514 protected function makeLimitReport() {
515 global $wgShowHostnames;
516
517 $maxIncludeSize = $this->mOptions->getMaxIncludeSize();
518
519 $cpuTime = $this->mOutput->getTimeSinceStart( 'cpu' );
520 if ( $cpuTime !== null ) {
521 $this->mOutput->setLimitReportData( 'limitreport-cputime',
522 sprintf( "%.3f", $cpuTime )
523 );
524 }
525
526 $wallTime = $this->mOutput->getTimeSinceStart( 'wall' );
527 $this->mOutput->setLimitReportData( 'limitreport-walltime',
528 sprintf( "%.3f", $wallTime )
529 );
530
531 $this->mOutput->setLimitReportData( 'limitreport-ppvisitednodes',
532 [ $this->mPPNodeCount, $this->mOptions->getMaxPPNodeCount() ]
533 );
534 $this->mOutput->setLimitReportData( 'limitreport-ppgeneratednodes',
535 [ $this->mGeneratedPPNodeCount, $this->mOptions->getMaxGeneratedPPNodeCount() ]
536 );
537 $this->mOutput->setLimitReportData( 'limitreport-postexpandincludesize',
538 [ $this->mIncludeSizes['post-expand'], $maxIncludeSize ]
539 );
540 $this->mOutput->setLimitReportData( 'limitreport-templateargumentsize',
541 [ $this->mIncludeSizes['arg'], $maxIncludeSize ]
542 );
543 $this->mOutput->setLimitReportData( 'limitreport-expansiondepth',
544 [ $this->mHighestExpansionDepth, $this->mOptions->getMaxPPExpandDepth() ]
545 );
546 $this->mOutput->setLimitReportData( 'limitreport-expensivefunctioncount',
547 [ $this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit() ]
548 );
549
550 foreach ( $this->mStripState->getLimitReport() as list( $key, $value ) ) {
551 $this->mOutput->setLimitReportData( $key, $value );
552 }
553
554 Hooks::run( 'ParserLimitReportPrepare', [ $this, $this->mOutput ] );
555
556 $limitReport = "NewPP limit report\n";
557 if ( $wgShowHostnames ) {
558 $limitReport .= 'Parsed by ' . wfHostname() . "\n";
559 }
560 $limitReport .= 'Cached time: ' . $this->mOutput->getCacheTime() . "\n";
561 $limitReport .= 'Cache expiry: ' . $this->mOutput->getCacheExpiry() . "\n";
562 $limitReport .= 'Dynamic content: ' .
563 ( $this->mOutput->hasDynamicContent() ? 'true' : 'false' ) .
564 "\n";
565
566 foreach ( $this->mOutput->getLimitReportData() as $key => $value ) {
567 if ( Hooks::run( 'ParserLimitReportFormat',
568 [ $key, &$value, &$limitReport, false, false ]
569 ) ) {
570 $keyMsg = wfMessage( $key )->inLanguage( 'en' )->useDatabase( false );
571 $valueMsg = wfMessage( [ "$key-value-text", "$key-value" ] )
572 ->inLanguage( 'en' )->useDatabase( false );
573 if ( !$valueMsg->exists() ) {
574 $valueMsg = new RawMessage( '$1' );
575 }
576 if ( !$keyMsg->isDisabled() && !$valueMsg->isDisabled() ) {
577 $valueMsg->params( $value );
578 $limitReport .= "{$keyMsg->text()}: {$valueMsg->text()}\n";
579 }
580 }
581 }
582 // Since we're not really outputting HTML, decode the entities and
583 // then re-encode the things that need hiding inside HTML comments.
584 $limitReport = htmlspecialchars_decode( $limitReport );
585 // Run deprecated hook
586 Hooks::run( 'ParserLimitReport', [ $this, &$limitReport ], '1.22' );
587
588 // Sanitize for comment. Note '‐' in the replacement is U+2010,
589 // which looks much like the problematic '-'.
590 $limitReport = str_replace( [ '-', '&' ], [ '‐', '&amp;' ], $limitReport );
591 $text = "\n<!-- \n$limitReport-->\n";
592
593 // Add on template profiling data in human/machine readable way
594 $dataByFunc = $this->mProfiler->getFunctionStats();
595 uasort( $dataByFunc, function ( $a, $b ) {
596 return $a['real'] < $b['real']; // descending order
597 } );
598 $profileReport = [];
599 foreach ( array_slice( $dataByFunc, 0, 10 ) as $item ) {
600 $profileReport[] = sprintf( "%6.2f%% %8.3f %6d %s",
601 $item['%real'], $item['real'], $item['calls'],
602 htmlspecialchars( $item['name'] ) );
603 }
604 $text .= "<!--\nTransclusion expansion time report (%,ms,calls,template)\n";
605 $text .= implode( "\n", $profileReport ) . "\n-->\n";
606
607 $this->mOutput->setLimitReportData( 'limitreport-timingprofile', $profileReport );
608
609 // Add other cache related metadata
610 if ( $wgShowHostnames ) {
611 $this->mOutput->setLimitReportData( 'cachereport-origin', wfHostname() );
612 }
613 $this->mOutput->setLimitReportData( 'cachereport-timestamp',
614 $this->mOutput->getCacheTime() );
615 $this->mOutput->setLimitReportData( 'cachereport-ttl',
616 $this->mOutput->getCacheExpiry() );
617 $this->mOutput->setLimitReportData( 'cachereport-transientcontent',
618 $this->mOutput->hasDynamicContent() );
619
620 if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) {
621 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' .
622 $this->mTitle->getPrefixedDBkey() );
623 }
624 return $text;
625 }
626
627 /**
628 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
629 * can be called from an extension tag hook.
630 *
631 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
632 * instead, which means that lists and links have not been fully parsed yet,
633 * and strip markers are still present.
634 *
635 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
636 *
637 * Use this function if you're a parser tag hook and you want to parse
638 * wikitext before or after applying additional transformations, and you
639 * intend to *return the result as hook output*, which will cause it to go
640 * through the rest of parsing process automatically.
641 *
642 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
643 * $text are not expanded
644 *
645 * @param string $text Text extension wants to have parsed
646 * @param bool|PPFrame $frame The frame to use for expanding any template variables
647 * @return string UNSAFE half-parsed HTML
648 */
649 public function recursiveTagParse( $text, $frame = false ) {
650 // Avoid PHP 7.1 warning from passing $this by reference
651 $parser = $this;
652 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
653 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
654 $text = $this->internalParse( $text, false, $frame );
655 return $text;
656 }
657
658 /**
659 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
660 * point can be called from an extension tag hook.
661 *
662 * The output of this function is fully-parsed HTML that is safe for output.
663 * If you're a parser tag hook, you might want to use recursiveTagParse()
664 * instead.
665 *
666 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
667 * $text are not expanded
668 *
669 * @since 1.25
670 *
671 * @param string $text Text extension wants to have parsed
672 * @param bool|PPFrame $frame The frame to use for expanding any template variables
673 * @return string Fully parsed HTML
674 */
675 public function recursiveTagParseFully( $text, $frame = false ) {
676 $text = $this->recursiveTagParse( $text, $frame );
677 $text = $this->internalParseHalfParsed( $text, false );
678 return $text;
679 }
680
681 /**
682 * Expand templates and variables in the text, producing valid, static wikitext.
683 * Also removes comments.
684 * Do not call this function recursively.
685 * @param string $text
686 * @param Title $title
687 * @param ParserOptions $options
688 * @param int|null $revid
689 * @param bool|PPFrame $frame
690 * @return mixed|string
691 */
692 public function preprocess( $text, Title $title = null,
693 ParserOptions $options, $revid = null, $frame = false
694 ) {
695 $magicScopeVariable = $this->lock();
696 $this->startParse( $title, $options, self::OT_PREPROCESS, true );
697 if ( $revid !== null ) {
698 $this->mRevisionId = $revid;
699 }
700 // Avoid PHP 7.1 warning from passing $this by reference
701 $parser = $this;
702 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
703 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
704 $text = $this->replaceVariables( $text, $frame );
705 $text = $this->mStripState->unstripBoth( $text );
706 return $text;
707 }
708
709 /**
710 * Recursive parser entry point that can be called from an extension tag
711 * hook.
712 *
713 * @param string $text Text to be expanded
714 * @param bool|PPFrame $frame The frame to use for expanding any template variables
715 * @return string
716 * @since 1.19
717 */
718 public function recursivePreprocess( $text, $frame = false ) {
719 $text = $this->replaceVariables( $text, $frame );
720 $text = $this->mStripState->unstripBoth( $text );
721 return $text;
722 }
723
724 /**
725 * Process the wikitext for the "?preload=" feature. (T7210)
726 *
727 * "<noinclude>", "<includeonly>" etc. are parsed as for template
728 * transclusion, comments, templates, arguments, tags hooks and parser
729 * functions are untouched.
730 *
731 * @param string $text
732 * @param Title $title
733 * @param ParserOptions $options
734 * @param array $params
735 * @return string
736 */
737 public function getPreloadText( $text, Title $title, ParserOptions $options, $params = [] ) {
738 $msg = new RawMessage( $text );
739 $text = $msg->params( $params )->plain();
740
741 # Parser (re)initialisation
742 $magicScopeVariable = $this->lock();
743 $this->startParse( $title, $options, self::OT_PLAIN, true );
744
745 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES;
746 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
747 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
748 $text = $this->mStripState->unstripBoth( $text );
749 return $text;
750 }
751
752 /**
753 * Set the current user.
754 * Should only be used when doing pre-save transform.
755 *
756 * @param User|null $user User object or null (to reset)
757 */
758 public function setUser( $user ) {
759 $this->mUser = $user;
760 }
761
762 /**
763 * Set the context title
764 *
765 * @param Title $t
766 */
767 public function setTitle( $t ) {
768 if ( !$t ) {
769 $t = Title::newFromText( 'NO TITLE' );
770 }
771
772 if ( $t->hasFragment() ) {
773 # Strip the fragment to avoid various odd effects
774 $this->mTitle = $t->createFragmentTarget( '' );
775 } else {
776 $this->mTitle = $t;
777 }
778 }
779
780 /**
781 * Accessor for the Title object
782 *
783 * @return Title
784 */
785 public function getTitle() {
786 return $this->mTitle;
787 }
788
789 /**
790 * Accessor/mutator for the Title object
791 *
792 * @param Title $x Title object or null to just get the current one
793 * @return Title
794 */
795 public function Title( $x = null ) {
796 return wfSetVar( $this->mTitle, $x );
797 }
798
799 /**
800 * Set the output type
801 *
802 * @param int $ot New value
803 */
804 public function setOutputType( $ot ) {
805 $this->mOutputType = $ot;
806 # Shortcut alias
807 $this->ot = [
808 'html' => $ot == self::OT_HTML,
809 'wiki' => $ot == self::OT_WIKI,
810 'pre' => $ot == self::OT_PREPROCESS,
811 'plain' => $ot == self::OT_PLAIN,
812 ];
813 }
814
815 /**
816 * Accessor/mutator for the output type
817 *
818 * @param int|null $x New value or null to just get the current one
819 * @return int
820 */
821 public function OutputType( $x = null ) {
822 return wfSetVar( $this->mOutputType, $x );
823 }
824
825 /**
826 * Get the ParserOutput object
827 *
828 * @return ParserOutput
829 */
830 public function getOutput() {
831 return $this->mOutput;
832 }
833
834 /**
835 * Get the ParserOptions object
836 *
837 * @return ParserOptions
838 */
839 public function getOptions() {
840 return $this->mOptions;
841 }
842
843 /**
844 * Accessor/mutator for the ParserOptions object
845 *
846 * @param ParserOptions $x New value or null to just get the current one
847 * @return ParserOptions Current ParserOptions object
848 */
849 public function Options( $x = null ) {
850 return wfSetVar( $this->mOptions, $x );
851 }
852
853 /**
854 * @return int
855 */
856 public function nextLinkID() {
857 return $this->mLinkID++;
858 }
859
860 /**
861 * @param int $id
862 */
863 public function setLinkID( $id ) {
864 $this->mLinkID = $id;
865 }
866
867 /**
868 * Get a language object for use in parser functions such as {{FORMATNUM:}}
869 * @return Language
870 */
871 public function getFunctionLang() {
872 return $this->getTargetLanguage();
873 }
874
875 /**
876 * Get the target language for the content being parsed. This is usually the
877 * language that the content is in.
878 *
879 * @since 1.19
880 *
881 * @throws MWException
882 * @return Language
883 */
884 public function getTargetLanguage() {
885 $target = $this->mOptions->getTargetLanguage();
886
887 if ( $target !== null ) {
888 return $target;
889 } elseif ( $this->mOptions->getInterfaceMessage() ) {
890 return $this->mOptions->getUserLangObj();
891 } elseif ( is_null( $this->mTitle ) ) {
892 throw new MWException( __METHOD__ . ': $this->mTitle is null' );
893 }
894
895 return $this->mTitle->getPageLanguage();
896 }
897
898 /**
899 * Get the language object for language conversion
900 * @return Language|null
901 */
902 public function getConverterLanguage() {
903 return $this->getTargetLanguage();
904 }
905
906 /**
907 * Get a User object either from $this->mUser, if set, or from the
908 * ParserOptions object otherwise
909 *
910 * @return User
911 */
912 public function getUser() {
913 if ( !is_null( $this->mUser ) ) {
914 return $this->mUser;
915 }
916 return $this->mOptions->getUser();
917 }
918
919 /**
920 * Get a preprocessor object
921 *
922 * @return Preprocessor
923 */
924 public function getPreprocessor() {
925 if ( !isset( $this->mPreprocessor ) ) {
926 $class = $this->mPreprocessorClass;
927 $this->mPreprocessor = new $class( $this );
928 }
929 return $this->mPreprocessor;
930 }
931
932 /**
933 * Get a LinkRenderer instance to make links with
934 *
935 * @since 1.28
936 * @return LinkRenderer
937 */
938 public function getLinkRenderer() {
939 if ( !$this->mLinkRenderer ) {
940 $this->mLinkRenderer = MediaWikiServices::getInstance()
941 ->getLinkRendererFactory()->create();
942 $this->mLinkRenderer->setStubThreshold(
943 $this->getOptions()->getStubThreshold()
944 );
945 }
946
947 return $this->mLinkRenderer;
948 }
949
950 /**
951 * Replaces all occurrences of HTML-style comments and the given tags
952 * in the text with a random marker and returns the next text. The output
953 * parameter $matches will be an associative array filled with data in
954 * the form:
955 *
956 * @code
957 * 'UNIQ-xxxxx' => [
958 * 'element',
959 * 'tag content',
960 * [ 'param' => 'x' ],
961 * '<element param="x">tag content</element>' ]
962 * @endcode
963 *
964 * @param array $elements List of element names. Comments are always extracted.
965 * @param string $text Source text string.
966 * @param array &$matches Out parameter, Array: extracted tags
967 * @return string Stripped text
968 */
969 public static function extractTagsAndParams( $elements, $text, &$matches ) {
970 static $n = 1;
971 $stripped = '';
972 $matches = [];
973
974 $taglist = implode( '|', $elements );
975 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
976
977 while ( $text != '' ) {
978 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
979 $stripped .= $p[0];
980 if ( count( $p ) < 5 ) {
981 break;
982 }
983 if ( count( $p ) > 5 ) {
984 # comment
985 $element = $p[4];
986 $attributes = '';
987 $close = '';
988 $inside = $p[5];
989 } else {
990 # tag
991 $element = $p[1];
992 $attributes = $p[2];
993 $close = $p[3];
994 $inside = $p[4];
995 }
996
997 $marker = self::MARKER_PREFIX . "-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
998 $stripped .= $marker;
999
1000 if ( $close === '/>' ) {
1001 # Empty element tag, <tag />
1002 $content = null;
1003 $text = $inside;
1004 $tail = null;
1005 } else {
1006 if ( $element === '!--' ) {
1007 $end = '/(-->)/';
1008 } else {
1009 $end = "/(<\\/$element\\s*>)/i";
1010 }
1011 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
1012 $content = $q[0];
1013 if ( count( $q ) < 3 ) {
1014 # No end tag -- let it run out to the end of the text.
1015 $tail = '';
1016 $text = '';
1017 } else {
1018 $tail = $q[1];
1019 $text = $q[2];
1020 }
1021 }
1022
1023 $matches[$marker] = [ $element,
1024 $content,
1025 Sanitizer::decodeTagAttributes( $attributes ),
1026 "<$element$attributes$close$content$tail" ];
1027 }
1028 return $stripped;
1029 }
1030
1031 /**
1032 * Get a list of strippable XML-like elements
1033 *
1034 * @return array
1035 */
1036 public function getStripList() {
1037 return $this->mStripList;
1038 }
1039
1040 /**
1041 * Add an item to the strip state
1042 * Returns the unique tag which must be inserted into the stripped text
1043 * The tag will be replaced with the original text in unstrip()
1044 *
1045 * @param string $text
1046 *
1047 * @return string
1048 */
1049 public function insertStripItem( $text ) {
1050 $marker = self::MARKER_PREFIX . "-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
1051 $this->mMarkerIndex++;
1052 $this->mStripState->addGeneral( $marker, $text );
1053 return $marker;
1054 }
1055
1056 /**
1057 * parse the wiki syntax used to render tables
1058 *
1059 * @private
1060 * @param string $text
1061 * @return string
1062 */
1063 public function doTableStuff( $text ) {
1064 $lines = StringUtils::explode( "\n", $text );
1065 $out = '';
1066 $td_history = []; # Is currently a td tag open?
1067 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1068 $tr_history = []; # Is currently a tr tag open?
1069 $tr_attributes = []; # history of tr attributes
1070 $has_opened_tr = []; # Did this table open a <tr> element?
1071 $indent_level = 0; # indent level of the table
1072
1073 foreach ( $lines as $outLine ) {
1074 $line = trim( $outLine );
1075
1076 if ( $line === '' ) { # empty line, go to next line
1077 $out .= $outLine . "\n";
1078 continue;
1079 }
1080
1081 $first_character = $line[0];
1082 $first_two = substr( $line, 0, 2 );
1083 $matches = [];
1084
1085 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1086 # First check if we are starting a new table
1087 $indent_level = strlen( $matches[1] );
1088
1089 $attributes = $this->mStripState->unstripBoth( $matches[2] );
1090 $attributes = Sanitizer::fixTagAttributes( $attributes, 'table' );
1091
1092 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1093 array_push( $td_history, false );
1094 array_push( $last_tag_history, '' );
1095 array_push( $tr_history, false );
1096 array_push( $tr_attributes, '' );
1097 array_push( $has_opened_tr, false );
1098 } elseif ( count( $td_history ) == 0 ) {
1099 # Don't do any of the following
1100 $out .= $outLine . "\n";
1101 continue;
1102 } elseif ( $first_two === '|}' ) {
1103 # We are ending a table
1104 $line = '</table>' . substr( $line, 2 );
1105 $last_tag = array_pop( $last_tag_history );
1106
1107 if ( !array_pop( $has_opened_tr ) ) {
1108 $line = "<tr><td></td></tr>{$line}";
1109 }
1110
1111 if ( array_pop( $tr_history ) ) {
1112 $line = "</tr>{$line}";
1113 }
1114
1115 if ( array_pop( $td_history ) ) {
1116 $line = "</{$last_tag}>{$line}";
1117 }
1118 array_pop( $tr_attributes );
1119 if ( $indent_level > 0 ) {
1120 $outLine = rtrim( $line ) . str_repeat( '</dd></dl>', $indent_level );
1121 } else {
1122 $outLine = $line;
1123 }
1124 } elseif ( $first_two === '|-' ) {
1125 # Now we have a table row
1126 $line = preg_replace( '#^\|-+#', '', $line );
1127
1128 # Whats after the tag is now only attributes
1129 $attributes = $this->mStripState->unstripBoth( $line );
1130 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
1131 array_pop( $tr_attributes );
1132 array_push( $tr_attributes, $attributes );
1133
1134 $line = '';
1135 $last_tag = array_pop( $last_tag_history );
1136 array_pop( $has_opened_tr );
1137 array_push( $has_opened_tr, true );
1138
1139 if ( array_pop( $tr_history ) ) {
1140 $line = '</tr>';
1141 }
1142
1143 if ( array_pop( $td_history ) ) {
1144 $line = "</{$last_tag}>{$line}";
1145 }
1146
1147 $outLine = $line;
1148 array_push( $tr_history, false );
1149 array_push( $td_history, false );
1150 array_push( $last_tag_history, '' );
1151 } elseif ( $first_character === '|'
1152 || $first_character === '!'
1153 || $first_two === '|+'
1154 ) {
1155 # This might be cell elements, td, th or captions
1156 if ( $first_two === '|+' ) {
1157 $first_character = '+';
1158 $line = substr( $line, 2 );
1159 } else {
1160 $line = substr( $line, 1 );
1161 }
1162
1163 // Implies both are valid for table headings.
1164 if ( $first_character === '!' ) {
1165 $line = StringUtils::replaceMarkup( '!!', '||', $line );
1166 }
1167
1168 # Split up multiple cells on the same line.
1169 # FIXME : This can result in improper nesting of tags processed
1170 # by earlier parser steps.
1171 $cells = explode( '||', $line );
1172
1173 $outLine = '';
1174
1175 # Loop through each table cell
1176 foreach ( $cells as $cell ) {
1177 $previous = '';
1178 if ( $first_character !== '+' ) {
1179 $tr_after = array_pop( $tr_attributes );
1180 if ( !array_pop( $tr_history ) ) {
1181 $previous = "<tr{$tr_after}>\n";
1182 }
1183 array_push( $tr_history, true );
1184 array_push( $tr_attributes, '' );
1185 array_pop( $has_opened_tr );
1186 array_push( $has_opened_tr, true );
1187 }
1188
1189 $last_tag = array_pop( $last_tag_history );
1190
1191 if ( array_pop( $td_history ) ) {
1192 $previous = "</{$last_tag}>\n{$previous}";
1193 }
1194
1195 if ( $first_character === '|' ) {
1196 $last_tag = 'td';
1197 } elseif ( $first_character === '!' ) {
1198 $last_tag = 'th';
1199 } elseif ( $first_character === '+' ) {
1200 $last_tag = 'caption';
1201 } else {
1202 $last_tag = '';
1203 }
1204
1205 array_push( $last_tag_history, $last_tag );
1206
1207 # A cell could contain both parameters and data
1208 $cell_data = explode( '|', $cell, 2 );
1209
1210 # T2553: Note that a '|' inside an invalid link should not
1211 # be mistaken as delimiting cell parameters
1212 # Bug T153140: Neither should language converter markup.
1213 if ( preg_match( '/\[\[|-\{/', $cell_data[0] ) === 1 ) {
1214 $cell = "{$previous}<{$last_tag}>" . trim( $cell );
1215 } elseif ( count( $cell_data ) == 1 ) {
1216 // Whitespace in cells is trimmed
1217 $cell = "{$previous}<{$last_tag}>" . trim( $cell_data[0] );
1218 } else {
1219 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1220 $attributes = Sanitizer::fixTagAttributes( $attributes, $last_tag );
1221 // Whitespace in cells is trimmed
1222 $cell = "{$previous}<{$last_tag}{$attributes}>" . trim( $cell_data[1] );
1223 }
1224
1225 $outLine .= $cell;
1226 array_push( $td_history, true );
1227 }
1228 }
1229 $out .= $outLine . "\n";
1230 }
1231
1232 # Closing open td, tr && table
1233 while ( count( $td_history ) > 0 ) {
1234 if ( array_pop( $td_history ) ) {
1235 $out .= "</td>\n";
1236 }
1237 if ( array_pop( $tr_history ) ) {
1238 $out .= "</tr>\n";
1239 }
1240 if ( !array_pop( $has_opened_tr ) ) {
1241 $out .= "<tr><td></td></tr>\n";
1242 }
1243
1244 $out .= "</table>\n";
1245 }
1246
1247 # Remove trailing line-ending (b/c)
1248 if ( substr( $out, -1 ) === "\n" ) {
1249 $out = substr( $out, 0, -1 );
1250 }
1251
1252 # special case: don't return empty table
1253 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1254 $out = '';
1255 }
1256
1257 return $out;
1258 }
1259
1260 /**
1261 * Helper function for parse() that transforms wiki markup into half-parsed
1262 * HTML. Only called for $mOutputType == self::OT_HTML.
1263 *
1264 * @private
1265 *
1266 * @param string $text The text to parse
1267 * @param bool $isMain Whether this is being called from the main parse() function
1268 * @param PPFrame|bool $frame A pre-processor frame
1269 *
1270 * @return string
1271 */
1272 public function internalParse( $text, $isMain = true, $frame = false ) {
1273 $origText = $text;
1274
1275 // Avoid PHP 7.1 warning from passing $this by reference
1276 $parser = $this;
1277
1278 # Hook to suspend the parser in this state
1279 if ( !Hooks::run( 'ParserBeforeInternalParse', [ &$parser, &$text, &$this->mStripState ] ) ) {
1280 return $text;
1281 }
1282
1283 # if $frame is provided, then use $frame for replacing any variables
1284 if ( $frame ) {
1285 # use frame depth to infer how include/noinclude tags should be handled
1286 # depth=0 means this is the top-level document; otherwise it's an included document
1287 if ( !$frame->depth ) {
1288 $flag = 0;
1289 } else {
1290 $flag = self::PTD_FOR_INCLUSION;
1291 }
1292 $dom = $this->preprocessToDom( $text, $flag );
1293 $text = $frame->expand( $dom );
1294 } else {
1295 # if $frame is not provided, then use old-style replaceVariables
1296 $text = $this->replaceVariables( $text );
1297 }
1298
1299 Hooks::run( 'InternalParseBeforeSanitize', [ &$parser, &$text, &$this->mStripState ] );
1300 $text = Sanitizer::removeHTMLtags(
1301 $text,
1302 [ $this, 'attributeStripCallback' ],
1303 false,
1304 array_keys( $this->mTransparentTagHooks ),
1305 [],
1306 [ $this, 'addTrackingCategory' ]
1307 );
1308 Hooks::run( 'InternalParseBeforeLinks', [ &$parser, &$text, &$this->mStripState ] );
1309
1310 # Tables need to come after variable replacement for things to work
1311 # properly; putting them before other transformations should keep
1312 # exciting things like link expansions from showing up in surprising
1313 # places.
1314 $text = $this->doTableStuff( $text );
1315
1316 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1317
1318 $text = $this->doDoubleUnderscore( $text );
1319
1320 $text = $this->doHeadings( $text );
1321 $text = $this->replaceInternalLinks( $text );
1322 $text = $this->doAllQuotes( $text );
1323 $text = $this->replaceExternalLinks( $text );
1324
1325 # replaceInternalLinks may sometimes leave behind
1326 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1327 $text = str_replace( self::MARKER_PREFIX . 'NOPARSE', '', $text );
1328
1329 $text = $this->doMagicLinks( $text );
1330 $text = $this->formatHeadings( $text, $origText, $isMain );
1331
1332 return $text;
1333 }
1334
1335 /**
1336 * Helper function for parse() that transforms half-parsed HTML into fully
1337 * parsed HTML.
1338 *
1339 * @param string $text
1340 * @param bool $isMain
1341 * @param bool $linestart
1342 * @return string
1343 */
1344 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1345 $text = $this->mStripState->unstripGeneral( $text );
1346
1347 // Avoid PHP 7.1 warning from passing $this by reference
1348 $parser = $this;
1349
1350 if ( $isMain ) {
1351 Hooks::run( 'ParserAfterUnstrip', [ &$parser, &$text ] );
1352 }
1353
1354 # Clean up special characters, only run once, next-to-last before doBlockLevels
1355 $fixtags = [
1356 # French spaces, last one Guillemet-left
1357 # only if there is something before the space
1358 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;',
1359 # french spaces, Guillemet-right
1360 '/(\\302\\253) /' => '\\1&#160;',
1361 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, T13874.
1362 ];
1363 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
1364
1365 $text = $this->doBlockLevels( $text, $linestart );
1366
1367 $this->replaceLinkHolders( $text );
1368
1369 /**
1370 * The input doesn't get language converted if
1371 * a) It's disabled
1372 * b) Content isn't converted
1373 * c) It's a conversion table
1374 * d) it is an interface message (which is in the user language)
1375 */
1376 if ( !( $this->mOptions->getDisableContentConversion()
1377 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) )
1378 ) {
1379 if ( !$this->mOptions->getInterfaceMessage() ) {
1380 # The position of the convert() call should not be changed. it
1381 # assumes that the links are all replaced and the only thing left
1382 # is the <nowiki> mark.
1383 $text = $this->getConverterLanguage()->convert( $text );
1384 }
1385 }
1386
1387 $text = $this->mStripState->unstripNoWiki( $text );
1388
1389 if ( $isMain ) {
1390 Hooks::run( 'ParserBeforeTidy', [ &$parser, &$text ] );
1391 }
1392
1393 $text = $this->replaceTransparentTags( $text );
1394 $text = $this->mStripState->unstripGeneral( $text );
1395
1396 $text = Sanitizer::normalizeCharReferences( $text );
1397
1398 if ( MWTidy::isEnabled() ) {
1399 if ( $this->mOptions->getTidy() ) {
1400 $text = MWTidy::tidy( $text );
1401 }
1402 } else {
1403 # attempt to sanitize at least some nesting problems
1404 # (T4702 and quite a few others)
1405 $tidyregs = [
1406 # ''Something [http://www.cool.com cool''] -->
1407 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1408 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1409 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1410 # fix up an anchor inside another anchor, only
1411 # at least for a single single nested link (T5695)
1412 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1413 '\\1\\2</a>\\3</a>\\1\\4</a>',
1414 # fix div inside inline elements- doBlockLevels won't wrap a line which
1415 # contains a div, so fix it up here; replace
1416 # div with escaped text
1417 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1418 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
1419 # remove empty italic or bold tag pairs, some
1420 # introduced by rules above
1421 '/<([bi])><\/\\1>/' => '',
1422 ];
1423
1424 $text = preg_replace(
1425 array_keys( $tidyregs ),
1426 array_values( $tidyregs ),
1427 $text );
1428 }
1429
1430 if ( $isMain ) {
1431 Hooks::run( 'ParserAfterTidy', [ &$parser, &$text ] );
1432 }
1433
1434 return $text;
1435 }
1436
1437 /**
1438 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1439 * magic external links.
1440 *
1441 * DML
1442 * @private
1443 *
1444 * @param string $text
1445 *
1446 * @return string
1447 */
1448 public function doMagicLinks( $text ) {
1449 $prots = wfUrlProtocolsWithoutProtRel();
1450 $urlChar = self::EXT_LINK_URL_CLASS;
1451 $addr = self::EXT_LINK_ADDR;
1452 $space = self::SPACE_NOT_NL; # non-newline space
1453 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1454 $spaces = "$space++"; # possessive match of 1 or more spaces
1455 $text = preg_replace_callback(
1456 '!(?: # Start cases
1457 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1458 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1459 (\b # m[3]: Free external links
1460 (?i:$prots)
1461 ($addr$urlChar*) # m[4]: Post-protocol path
1462 ) |
1463 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1464 ([0-9]+)\b |
1465 \bISBN $spaces ( # m[6]: ISBN, capture number
1466 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1467 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1468 [0-9Xx] # check digit
1469 )\b
1470 )!xu", [ $this, 'magicLinkCallback' ], $text );
1471 return $text;
1472 }
1473
1474 /**
1475 * @throws MWException
1476 * @param array $m
1477 * @return string HTML
1478 */
1479 public function magicLinkCallback( $m ) {
1480 if ( isset( $m[1] ) && $m[1] !== '' ) {
1481 # Skip anchor
1482 return $m[0];
1483 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1484 # Skip HTML element
1485 return $m[0];
1486 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1487 # Free external link
1488 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1489 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1490 # RFC or PMID
1491 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1492 if ( !$this->mOptions->getMagicRFCLinks() ) {
1493 return $m[0];
1494 }
1495 $keyword = 'RFC';
1496 $urlmsg = 'rfcurl';
1497 $cssClass = 'mw-magiclink-rfc';
1498 $trackingCat = 'magiclink-tracking-rfc';
1499 $id = $m[5];
1500 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1501 if ( !$this->mOptions->getMagicPMIDLinks() ) {
1502 return $m[0];
1503 }
1504 $keyword = 'PMID';
1505 $urlmsg = 'pubmedurl';
1506 $cssClass = 'mw-magiclink-pmid';
1507 $trackingCat = 'magiclink-tracking-pmid';
1508 $id = $m[5];
1509 } else {
1510 throw new MWException( __METHOD__ . ': unrecognised match type "' .
1511 substr( $m[0], 0, 20 ) . '"' );
1512 }
1513 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1514 $this->addTrackingCategory( $trackingCat );
1515 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1516 } elseif ( isset( $m[6] ) && $m[6] !== ''
1517 && $this->mOptions->getMagicISBNLinks()
1518 ) {
1519 # ISBN
1520 $isbn = $m[6];
1521 $space = self::SPACE_NOT_NL; # non-newline space
1522 $isbn = preg_replace( "/$space/", ' ', $isbn );
1523 $num = strtr( $isbn, [
1524 '-' => '',
1525 ' ' => '',
1526 'x' => 'X',
1527 ] );
1528 $this->addTrackingCategory( 'magiclink-tracking-isbn' );
1529 return $this->getLinkRenderer()->makeKnownLink(
1530 SpecialPage::getTitleFor( 'Booksources', $num ),
1531 "ISBN $isbn",
1532 [
1533 'class' => 'internal mw-magiclink-isbn',
1534 'title' => false // suppress title attribute
1535 ]
1536 );
1537 } else {
1538 return $m[0];
1539 }
1540 }
1541
1542 /**
1543 * Make a free external link, given a user-supplied URL
1544 *
1545 * @param string $url
1546 * @param int $numPostProto
1547 * The number of characters after the protocol.
1548 * @return string HTML
1549 * @private
1550 */
1551 public function makeFreeExternalLink( $url, $numPostProto ) {
1552 $trail = '';
1553
1554 # The characters '<' and '>' (which were escaped by
1555 # removeHTMLtags()) should not be included in
1556 # URLs, per RFC 2396.
1557 # Make &nbsp; terminate a URL as well (bug T84937)
1558 $m2 = [];
1559 if ( preg_match(
1560 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1561 $url,
1562 $m2,
1563 PREG_OFFSET_CAPTURE
1564 ) ) {
1565 $trail = substr( $url, $m2[0][1] ) . $trail;
1566 $url = substr( $url, 0, $m2[0][1] );
1567 }
1568
1569 # Move trailing punctuation to $trail
1570 $sep = ',;\.:!?';
1571 # If there is no left bracket, then consider right brackets fair game too
1572 if ( strpos( $url, '(' ) === false ) {
1573 $sep .= ')';
1574 }
1575
1576 $urlRev = strrev( $url );
1577 $numSepChars = strspn( $urlRev, $sep );
1578 # Don't break a trailing HTML entity by moving the ; into $trail
1579 # This is in hot code, so use substr_compare to avoid having to
1580 # create a new string object for the comparison
1581 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1582 # more optimization: instead of running preg_match with a $
1583 # anchor, which can be slow, do the match on the reversed
1584 # string starting at the desired offset.
1585 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1586 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1587 $numSepChars--;
1588 }
1589 }
1590 if ( $numSepChars ) {
1591 $trail = substr( $url, -$numSepChars ) . $trail;
1592 $url = substr( $url, 0, -$numSepChars );
1593 }
1594
1595 # Verify that we still have a real URL after trail removal, and
1596 # not just lone protocol
1597 if ( strlen( $trail ) >= $numPostProto ) {
1598 return $url . $trail;
1599 }
1600
1601 $url = Sanitizer::cleanUrl( $url );
1602
1603 # Is this an external image?
1604 $text = $this->maybeMakeExternalImage( $url );
1605 if ( $text === false ) {
1606 # Not an image, make a link
1607 $text = Linker::makeExternalLink( $url,
1608 $this->getConverterLanguage()->markNoConversion( $url, true ),
1609 true, 'free',
1610 $this->getExternalLinkAttribs( $url ), $this->mTitle );
1611 # Register it in the output object...
1612 $this->mOutput->addExternalLink( $url );
1613 }
1614 return $text . $trail;
1615 }
1616
1617 /**
1618 * Parse headers and return html
1619 *
1620 * @private
1621 *
1622 * @param string $text
1623 *
1624 * @return string
1625 */
1626 public function doHeadings( $text ) {
1627 for ( $i = 6; $i >= 1; --$i ) {
1628 $h = str_repeat( '=', $i );
1629 // Trim non-newline whitespace from headings
1630 // Using \s* will break for: "==\n===\n" and parse as <h2>=</h2>
1631 $text = preg_replace( "/^(?:$h)[ \\t]*(.+?)[ \\t]*(?:$h)\\s*$/m", "<h$i>\\1</h$i>", $text );
1632 }
1633 return $text;
1634 }
1635
1636 /**
1637 * Replace single quotes with HTML markup
1638 * @private
1639 *
1640 * @param string $text
1641 *
1642 * @return string The altered text
1643 */
1644 public function doAllQuotes( $text ) {
1645 $outtext = '';
1646 $lines = StringUtils::explode( "\n", $text );
1647 foreach ( $lines as $line ) {
1648 $outtext .= $this->doQuotes( $line ) . "\n";
1649 }
1650 $outtext = substr( $outtext, 0, -1 );
1651 return $outtext;
1652 }
1653
1654 /**
1655 * Helper function for doAllQuotes()
1656 *
1657 * @param string $text
1658 *
1659 * @return string
1660 */
1661 public function doQuotes( $text ) {
1662 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1663 $countarr = count( $arr );
1664 if ( $countarr == 1 ) {
1665 return $text;
1666 }
1667
1668 // First, do some preliminary work. This may shift some apostrophes from
1669 // being mark-up to being text. It also counts the number of occurrences
1670 // of bold and italics mark-ups.
1671 $numbold = 0;
1672 $numitalics = 0;
1673 for ( $i = 1; $i < $countarr; $i += 2 ) {
1674 $thislen = strlen( $arr[$i] );
1675 // If there are ever four apostrophes, assume the first is supposed to
1676 // be text, and the remaining three constitute mark-up for bold text.
1677 // (T15227: ''''foo'''' turns into ' ''' foo ' ''')
1678 if ( $thislen == 4 ) {
1679 $arr[$i - 1] .= "'";
1680 $arr[$i] = "'''";
1681 $thislen = 3;
1682 } elseif ( $thislen > 5 ) {
1683 // If there are more than 5 apostrophes in a row, assume they're all
1684 // text except for the last 5.
1685 // (T15227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1686 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1687 $arr[$i] = "'''''";
1688 $thislen = 5;
1689 }
1690 // Count the number of occurrences of bold and italics mark-ups.
1691 if ( $thislen == 2 ) {
1692 $numitalics++;
1693 } elseif ( $thislen == 3 ) {
1694 $numbold++;
1695 } elseif ( $thislen == 5 ) {
1696 $numitalics++;
1697 $numbold++;
1698 }
1699 }
1700
1701 // If there is an odd number of both bold and italics, it is likely
1702 // that one of the bold ones was meant to be an apostrophe followed
1703 // by italics. Which one we cannot know for certain, but it is more
1704 // likely to be one that has a single-letter word before it.
1705 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1706 $firstsingleletterword = -1;
1707 $firstmultiletterword = -1;
1708 $firstspace = -1;
1709 for ( $i = 1; $i < $countarr; $i += 2 ) {
1710 if ( strlen( $arr[$i] ) == 3 ) {
1711 $x1 = substr( $arr[$i - 1], -1 );
1712 $x2 = substr( $arr[$i - 1], -2, 1 );
1713 if ( $x1 === ' ' ) {
1714 if ( $firstspace == -1 ) {
1715 $firstspace = $i;
1716 }
1717 } elseif ( $x2 === ' ' ) {
1718 $firstsingleletterword = $i;
1719 // if $firstsingleletterword is set, we don't
1720 // look at the other options, so we can bail early.
1721 break;
1722 } else {
1723 if ( $firstmultiletterword == -1 ) {
1724 $firstmultiletterword = $i;
1725 }
1726 }
1727 }
1728 }
1729
1730 // If there is a single-letter word, use it!
1731 if ( $firstsingleletterword > -1 ) {
1732 $arr[$firstsingleletterword] = "''";
1733 $arr[$firstsingleletterword - 1] .= "'";
1734 } elseif ( $firstmultiletterword > -1 ) {
1735 // If not, but there's a multi-letter word, use that one.
1736 $arr[$firstmultiletterword] = "''";
1737 $arr[$firstmultiletterword - 1] .= "'";
1738 } elseif ( $firstspace > -1 ) {
1739 // ... otherwise use the first one that has neither.
1740 // (notice that it is possible for all three to be -1 if, for example,
1741 // there is only one pentuple-apostrophe in the line)
1742 $arr[$firstspace] = "''";
1743 $arr[$firstspace - 1] .= "'";
1744 }
1745 }
1746
1747 // Now let's actually convert our apostrophic mush to HTML!
1748 $output = '';
1749 $buffer = '';
1750 $state = '';
1751 $i = 0;
1752 foreach ( $arr as $r ) {
1753 if ( ( $i % 2 ) == 0 ) {
1754 if ( $state === 'both' ) {
1755 $buffer .= $r;
1756 } else {
1757 $output .= $r;
1758 }
1759 } else {
1760 $thislen = strlen( $r );
1761 if ( $thislen == 2 ) {
1762 if ( $state === 'i' ) {
1763 $output .= '</i>';
1764 $state = '';
1765 } elseif ( $state === 'bi' ) {
1766 $output .= '</i>';
1767 $state = 'b';
1768 } elseif ( $state === 'ib' ) {
1769 $output .= '</b></i><b>';
1770 $state = 'b';
1771 } elseif ( $state === 'both' ) {
1772 $output .= '<b><i>' . $buffer . '</i>';
1773 $state = 'b';
1774 } else { // $state can be 'b' or ''
1775 $output .= '<i>';
1776 $state .= 'i';
1777 }
1778 } elseif ( $thislen == 3 ) {
1779 if ( $state === 'b' ) {
1780 $output .= '</b>';
1781 $state = '';
1782 } elseif ( $state === 'bi' ) {
1783 $output .= '</i></b><i>';
1784 $state = 'i';
1785 } elseif ( $state === 'ib' ) {
1786 $output .= '</b>';
1787 $state = 'i';
1788 } elseif ( $state === 'both' ) {
1789 $output .= '<i><b>' . $buffer . '</b>';
1790 $state = 'i';
1791 } else { // $state can be 'i' or ''
1792 $output .= '<b>';
1793 $state .= 'b';
1794 }
1795 } elseif ( $thislen == 5 ) {
1796 if ( $state === 'b' ) {
1797 $output .= '</b><i>';
1798 $state = 'i';
1799 } elseif ( $state === 'i' ) {
1800 $output .= '</i><b>';
1801 $state = 'b';
1802 } elseif ( $state === 'bi' ) {
1803 $output .= '</i></b>';
1804 $state = '';
1805 } elseif ( $state === 'ib' ) {
1806 $output .= '</b></i>';
1807 $state = '';
1808 } elseif ( $state === 'both' ) {
1809 $output .= '<i><b>' . $buffer . '</b></i>';
1810 $state = '';
1811 } else { // ($state == '')
1812 $buffer = '';
1813 $state = 'both';
1814 }
1815 }
1816 }
1817 $i++;
1818 }
1819 // Now close all remaining tags. Notice that the order is important.
1820 if ( $state === 'b' || $state === 'ib' ) {
1821 $output .= '</b>';
1822 }
1823 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1824 $output .= '</i>';
1825 }
1826 if ( $state === 'bi' ) {
1827 $output .= '</b>';
1828 }
1829 // There might be lonely ''''', so make sure we have a buffer
1830 if ( $state === 'both' && $buffer ) {
1831 $output .= '<b><i>' . $buffer . '</i></b>';
1832 }
1833 return $output;
1834 }
1835
1836 /**
1837 * Replace external links (REL)
1838 *
1839 * Note: this is all very hackish and the order of execution matters a lot.
1840 * Make sure to run tests/parser/parserTests.php if you change this code.
1841 *
1842 * @private
1843 *
1844 * @param string $text
1845 *
1846 * @throws MWException
1847 * @return string
1848 */
1849 public function replaceExternalLinks( $text ) {
1850 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1851 if ( $bits === false ) {
1852 throw new MWException( "PCRE needs to be compiled with "
1853 . "--enable-unicode-properties in order for MediaWiki to function" );
1854 }
1855 $s = array_shift( $bits );
1856
1857 $i = 0;
1858 while ( $i < count( $bits ) ) {
1859 $url = $bits[$i++];
1860 $i++; // protocol
1861 $text = $bits[$i++];
1862 $trail = $bits[$i++];
1863
1864 # The characters '<' and '>' (which were escaped by
1865 # removeHTMLtags()) should not be included in
1866 # URLs, per RFC 2396.
1867 $m2 = [];
1868 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1869 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1870 $url = substr( $url, 0, $m2[0][1] );
1871 }
1872
1873 # If the link text is an image URL, replace it with an <img> tag
1874 # This happened by accident in the original parser, but some people used it extensively
1875 $img = $this->maybeMakeExternalImage( $text );
1876 if ( $img !== false ) {
1877 $text = $img;
1878 }
1879
1880 $dtrail = '';
1881
1882 # Set linktype for CSS
1883 $linktype = 'text';
1884
1885 # No link text, e.g. [http://domain.tld/some.link]
1886 if ( $text == '' ) {
1887 # Autonumber
1888 $langObj = $this->getTargetLanguage();
1889 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1890 $linktype = 'autonumber';
1891 } else {
1892 # Have link text, e.g. [http://domain.tld/some.link text]s
1893 # Check for trail
1894 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1895 }
1896
1897 $text = $this->getConverterLanguage()->markNoConversion( $text );
1898
1899 $url = Sanitizer::cleanUrl( $url );
1900
1901 # Use the encoded URL
1902 # This means that users can paste URLs directly into the text
1903 # Funny characters like ö aren't valid in URLs anyway
1904 # This was changed in August 2004
1905 $s .= Linker::makeExternalLink( $url, $text, false, $linktype,
1906 $this->getExternalLinkAttribs( $url ), $this->mTitle ) . $dtrail . $trail;
1907
1908 # Register link in the output object.
1909 $this->mOutput->addExternalLink( $url );
1910 }
1911
1912 return $s;
1913 }
1914
1915 /**
1916 * Get the rel attribute for a particular external link.
1917 *
1918 * @since 1.21
1919 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1920 * nofollow if appropriate
1921 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1922 * @return string|null Rel attribute for $url
1923 */
1924 public static function getExternalLinkRel( $url = false, $title = null ) {
1925 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1926 $ns = $title ? $title->getNamespace() : false;
1927 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1928 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1929 ) {
1930 return 'nofollow';
1931 }
1932 return null;
1933 }
1934
1935 /**
1936 * Get an associative array of additional HTML attributes appropriate for a
1937 * particular external link. This currently may include rel => nofollow
1938 * (depending on configuration, namespace, and the URL's domain) and/or a
1939 * target attribute (depending on configuration).
1940 *
1941 * @param string $url URL to extract the domain from for rel =>
1942 * nofollow if appropriate
1943 * @return array Associative array of HTML attributes
1944 */
1945 public function getExternalLinkAttribs( $url ) {
1946 $attribs = [];
1947 $rel = self::getExternalLinkRel( $url, $this->mTitle );
1948
1949 $target = $this->mOptions->getExternalLinkTarget();
1950 if ( $target ) {
1951 $attribs['target'] = $target;
1952 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1953 // T133507. New windows can navigate parent cross-origin.
1954 // Including noreferrer due to lacking browser
1955 // support of noopener. Eventually noreferrer should be removed.
1956 if ( $rel !== '' ) {
1957 $rel .= ' ';
1958 }
1959 $rel .= 'noreferrer noopener';
1960 }
1961 }
1962 $attribs['rel'] = $rel;
1963 return $attribs;
1964 }
1965
1966 /**
1967 * Replace unusual escape codes in a URL with their equivalent characters
1968 *
1969 * This generally follows the syntax defined in RFC 3986, with special
1970 * consideration for HTTP query strings.
1971 *
1972 * @param string $url
1973 * @return string
1974 */
1975 public static function normalizeLinkUrl( $url ) {
1976 # First, make sure unsafe characters are encoded
1977 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
1978 function ( $m ) {
1979 return rawurlencode( $m[0] );
1980 },
1981 $url
1982 );
1983
1984 $ret = '';
1985 $end = strlen( $url );
1986
1987 # Fragment part - 'fragment'
1988 $start = strpos( $url, '#' );
1989 if ( $start !== false && $start < $end ) {
1990 $ret = self::normalizeUrlComponent(
1991 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
1992 $end = $start;
1993 }
1994
1995 # Query part - 'query' minus &=+;
1996 $start = strpos( $url, '?' );
1997 if ( $start !== false && $start < $end ) {
1998 $ret = self::normalizeUrlComponent(
1999 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
2000 $end = $start;
2001 }
2002
2003 # Scheme and path part - 'pchar'
2004 # (we assume no userinfo or encoded colons in the host)
2005 $ret = self::normalizeUrlComponent(
2006 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
2007
2008 return $ret;
2009 }
2010
2011 private static function normalizeUrlComponent( $component, $unsafe ) {
2012 $callback = function ( $matches ) use ( $unsafe ) {
2013 $char = urldecode( $matches[0] );
2014 $ord = ord( $char );
2015 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
2016 # Unescape it
2017 return $char;
2018 } else {
2019 # Leave it escaped, but use uppercase for a-f
2020 return strtoupper( $matches[0] );
2021 }
2022 };
2023 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
2024 }
2025
2026 /**
2027 * make an image if it's allowed, either through the global
2028 * option, through the exception, or through the on-wiki whitelist
2029 *
2030 * @param string $url
2031 *
2032 * @return string
2033 */
2034 private function maybeMakeExternalImage( $url ) {
2035 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
2036 $imagesexception = !empty( $imagesfrom );
2037 $text = false;
2038 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2039 if ( $imagesexception && is_array( $imagesfrom ) ) {
2040 $imagematch = false;
2041 foreach ( $imagesfrom as $match ) {
2042 if ( strpos( $url, $match ) === 0 ) {
2043 $imagematch = true;
2044 break;
2045 }
2046 }
2047 } elseif ( $imagesexception ) {
2048 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2049 } else {
2050 $imagematch = false;
2051 }
2052
2053 if ( $this->mOptions->getAllowExternalImages()
2054 || ( $imagesexception && $imagematch )
2055 ) {
2056 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
2057 # Image found
2058 $text = Linker::makeExternalImage( $url );
2059 }
2060 }
2061 if ( !$text && $this->mOptions->getEnableImageWhitelist()
2062 && preg_match( self::EXT_IMAGE_REGEX, $url )
2063 ) {
2064 $whitelist = explode(
2065 "\n",
2066 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2067 );
2068
2069 foreach ( $whitelist as $entry ) {
2070 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2071 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
2072 continue;
2073 }
2074 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2075 # Image matches a whitelist entry
2076 $text = Linker::makeExternalImage( $url );
2077 break;
2078 }
2079 }
2080 }
2081 return $text;
2082 }
2083
2084 /**
2085 * Process [[ ]] wikilinks
2086 *
2087 * @param string $s
2088 *
2089 * @return string Processed text
2090 *
2091 * @private
2092 */
2093 public function replaceInternalLinks( $s ) {
2094 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
2095 return $s;
2096 }
2097
2098 /**
2099 * Process [[ ]] wikilinks (RIL)
2100 * @param string &$s
2101 * @throws MWException
2102 * @return LinkHolderArray
2103 *
2104 * @private
2105 */
2106 public function replaceInternalLinks2( &$s ) {
2107 global $wgExtraInterlanguageLinkPrefixes;
2108
2109 static $tc = false, $e1, $e1_img;
2110 # the % is needed to support urlencoded titles as well
2111 if ( !$tc ) {
2112 $tc = Title::legalChars() . '#%';
2113 # Match a link having the form [[namespace:link|alternate]]trail
2114 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2115 # Match cases where there is no "]]", which might still be images
2116 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2117 }
2118
2119 $holders = new LinkHolderArray( $this );
2120
2121 # split the entire text string on occurrences of [[
2122 $a = StringUtils::explode( '[[', ' ' . $s );
2123 # get the first element (all text up to first [[), and remove the space we added
2124 $s = $a->current();
2125 $a->next();
2126 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2127 $s = substr( $s, 1 );
2128
2129 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2130 $e2 = null;
2131 if ( $useLinkPrefixExtension ) {
2132 # Match the end of a line for a word that's not followed by whitespace,
2133 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2134 global $wgContLang;
2135 $charset = $wgContLang->linkPrefixCharset();
2136 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2137 }
2138
2139 if ( is_null( $this->mTitle ) ) {
2140 throw new MWException( __METHOD__ . ": \$this->mTitle is null\n" );
2141 }
2142 $nottalk = !$this->mTitle->isTalkPage();
2143
2144 if ( $useLinkPrefixExtension ) {
2145 $m = [];
2146 if ( preg_match( $e2, $s, $m ) ) {
2147 $first_prefix = $m[2];
2148 } else {
2149 $first_prefix = false;
2150 }
2151 } else {
2152 $prefix = '';
2153 }
2154
2155 $useSubpages = $this->areSubpagesAllowed();
2156
2157 # Loop for each link
2158 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2159 # Check for excessive memory usage
2160 if ( $holders->isBig() ) {
2161 # Too big
2162 # Do the existence check, replace the link holders and clear the array
2163 $holders->replace( $s );
2164 $holders->clear();
2165 }
2166
2167 if ( $useLinkPrefixExtension ) {
2168 if ( preg_match( $e2, $s, $m ) ) {
2169 $prefix = $m[2];
2170 $s = $m[1];
2171 } else {
2172 $prefix = '';
2173 }
2174 # first link
2175 if ( $first_prefix ) {
2176 $prefix = $first_prefix;
2177 $first_prefix = false;
2178 }
2179 }
2180
2181 $might_be_img = false;
2182
2183 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2184 $text = $m[2];
2185 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2186 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2187 # the real problem is with the $e1 regex
2188 # See T1500.
2189 # Still some problems for cases where the ] is meant to be outside punctuation,
2190 # and no image is in sight. See T4095.
2191 if ( $text !== ''
2192 && substr( $m[3], 0, 1 ) === ']'
2193 && strpos( $text, '[' ) !== false
2194 ) {
2195 $text .= ']'; # so that replaceExternalLinks($text) works later
2196 $m[3] = substr( $m[3], 1 );
2197 }
2198 # fix up urlencoded title texts
2199 if ( strpos( $m[1], '%' ) !== false ) {
2200 # Should anchors '#' also be rejected?
2201 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2202 }
2203 $trail = $m[3];
2204 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2205 # Invalid, but might be an image with a link in its caption
2206 $might_be_img = true;
2207 $text = $m[2];
2208 if ( strpos( $m[1], '%' ) !== false ) {
2209 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2210 }
2211 $trail = "";
2212 } else { # Invalid form; output directly
2213 $s .= $prefix . '[[' . $line;
2214 continue;
2215 }
2216
2217 $origLink = ltrim( $m[1], ' ' );
2218
2219 # Don't allow internal links to pages containing
2220 # PROTO: where PROTO is a valid URL protocol; these
2221 # should be external links.
2222 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $origLink ) ) {
2223 $s .= $prefix . '[[' . $line;
2224 continue;
2225 }
2226
2227 # Make subpage if necessary
2228 if ( $useSubpages ) {
2229 $link = $this->maybeDoSubpageLink( $origLink, $text );
2230 } else {
2231 $link = $origLink;
2232 }
2233
2234 // \x7f isn't a default legal title char, so most likely strip
2235 // markers will force us into the "invalid form" path above. But,
2236 // just in case, let's assert that xmlish tags aren't valid in
2237 // the title position.
2238 $unstrip = $this->mStripState->killMarkers( $link );
2239 $noMarkers = ( $unstrip === $link );
2240
2241 $nt = $noMarkers ? Title::newFromText( $link ) : null;
2242 if ( $nt === null ) {
2243 $s .= $prefix . '[[' . $line;
2244 continue;
2245 }
2246
2247 $ns = $nt->getNamespace();
2248 $iw = $nt->getInterwiki();
2249
2250 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2251
2252 if ( $might_be_img ) { # if this is actually an invalid link
2253 if ( $ns == NS_FILE && $noforce ) { # but might be an image
2254 $found = false;
2255 while ( true ) {
2256 # look at the next 'line' to see if we can close it there
2257 $a->next();
2258 $next_line = $a->current();
2259 if ( $next_line === false || $next_line === null ) {
2260 break;
2261 }
2262 $m = explode( ']]', $next_line, 3 );
2263 if ( count( $m ) == 3 ) {
2264 # the first ]] closes the inner link, the second the image
2265 $found = true;
2266 $text .= "[[{$m[0]}]]{$m[1]}";
2267 $trail = $m[2];
2268 break;
2269 } elseif ( count( $m ) == 2 ) {
2270 # if there's exactly one ]] that's fine, we'll keep looking
2271 $text .= "[[{$m[0]}]]{$m[1]}";
2272 } else {
2273 # if $next_line is invalid too, we need look no further
2274 $text .= '[[' . $next_line;
2275 break;
2276 }
2277 }
2278 if ( !$found ) {
2279 # we couldn't find the end of this imageLink, so output it raw
2280 # but don't ignore what might be perfectly normal links in the text we've examined
2281 $holders->merge( $this->replaceInternalLinks2( $text ) );
2282 $s .= "{$prefix}[[$link|$text";
2283 # note: no $trail, because without an end, there *is* no trail
2284 continue;
2285 }
2286 } else { # it's not an image, so output it raw
2287 $s .= "{$prefix}[[$link|$text";
2288 # note: no $trail, because without an end, there *is* no trail
2289 continue;
2290 }
2291 }
2292
2293 $wasblank = ( $text == '' );
2294 if ( $wasblank ) {
2295 $text = $link;
2296 if ( !$noforce ) {
2297 # Strip off leading ':'
2298 $text = substr( $text, 1 );
2299 }
2300 } else {
2301 # T6598 madness. Handle the quotes only if they come from the alternate part
2302 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2303 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2304 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2305 $text = $this->doQuotes( $text );
2306 }
2307
2308 # Link not escaped by : , create the various objects
2309 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2310 # Interwikis
2311 if (
2312 $iw && $this->mOptions->getInterwikiMagic() && $nottalk && (
2313 Language::fetchLanguageName( $iw, null, 'mw' ) ||
2314 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2315 )
2316 ) {
2317 # T26502: filter duplicates
2318 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
2319 $this->mLangLinkLanguages[$iw] = true;
2320 $this->mOutput->addLanguageLink( $nt->getFullText() );
2321 }
2322
2323 /**
2324 * Strip the whitespace interwiki links produce, see T10897
2325 */
2326 $s = rtrim( $s . $prefix ) . $trail; # T175416
2327 continue;
2328 }
2329
2330 if ( $ns == NS_FILE ) {
2331 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
2332 if ( $wasblank ) {
2333 # if no parameters were passed, $text
2334 # becomes something like "File:Foo.png",
2335 # which we don't want to pass on to the
2336 # image generator
2337 $text = '';
2338 } else {
2339 # recursively parse links inside the image caption
2340 # actually, this will parse them in any other parameters, too,
2341 # but it might be hard to fix that, and it doesn't matter ATM
2342 $text = $this->replaceExternalLinks( $text );
2343 $holders->merge( $this->replaceInternalLinks2( $text ) );
2344 }
2345 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2346 $s .= $prefix . $this->armorLinks(
2347 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2348 continue;
2349 }
2350 } elseif ( $ns == NS_CATEGORY ) {
2351 /**
2352 * Strip the whitespace Category links produce, see T2087
2353 */
2354 $s = rtrim( $s . $prefix ) . $trail; # T2087, T87753
2355
2356 if ( $wasblank ) {
2357 $sortkey = $this->getDefaultSort();
2358 } else {
2359 $sortkey = $text;
2360 }
2361 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
2362 $sortkey = str_replace( "\n", '', $sortkey );
2363 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
2364 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
2365
2366 continue;
2367 }
2368 }
2369
2370 # Self-link checking. For some languages, variants of the title are checked in
2371 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2372 # for linking to a different variant.
2373 if ( $ns != NS_SPECIAL && $nt->equals( $this->mTitle ) && !$nt->hasFragment() ) {
2374 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail );
2375 continue;
2376 }
2377
2378 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2379 # @todo FIXME: Should do batch file existence checks, see comment below
2380 if ( $ns == NS_MEDIA ) {
2381 # Give extensions a chance to select the file revision for us
2382 $options = [];
2383 $descQuery = false;
2384 Hooks::run( 'BeforeParserFetchFileAndTitle',
2385 [ $this, $nt, &$options, &$descQuery ] );
2386 # Fetch and register the file (file title may be different via hooks)
2387 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2388 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2389 $s .= $prefix . $this->armorLinks(
2390 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2391 continue;
2392 }
2393
2394 # Some titles, such as valid special pages or files in foreign repos, should
2395 # be shown as bluelinks even though they're not included in the page table
2396 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2397 # batch file existence checks for NS_FILE and NS_MEDIA
2398 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2399 $this->mOutput->addLink( $nt );
2400 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2401 } else {
2402 # Links will be added to the output link list after checking
2403 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2404 }
2405 }
2406 return $holders;
2407 }
2408
2409 /**
2410 * Render a forced-blue link inline; protect against double expansion of
2411 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2412 * Since this little disaster has to split off the trail text to avoid
2413 * breaking URLs in the following text without breaking trails on the
2414 * wiki links, it's been made into a horrible function.
2415 *
2416 * @param Title $nt
2417 * @param string $text
2418 * @param string $trail
2419 * @param string $prefix
2420 * @return string HTML-wikitext mix oh yuck
2421 */
2422 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2423 list( $inside, $trail ) = Linker::splitTrail( $trail );
2424
2425 if ( $text == '' ) {
2426 $text = htmlspecialchars( $nt->getPrefixedText() );
2427 }
2428
2429 $link = $this->getLinkRenderer()->makeKnownLink(
2430 $nt, new HtmlArmor( "$prefix$text$inside" )
2431 );
2432
2433 return $this->armorLinks( $link ) . $trail;
2434 }
2435
2436 /**
2437 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2438 * going to go through further parsing steps before inline URL expansion.
2439 *
2440 * Not needed quite as much as it used to be since free links are a bit
2441 * more sensible these days. But bracketed links are still an issue.
2442 *
2443 * @param string $text More-or-less HTML
2444 * @return string Less-or-more HTML with NOPARSE bits
2445 */
2446 public function armorLinks( $text ) {
2447 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/',
2448 self::MARKER_PREFIX . "NOPARSE$1", $text );
2449 }
2450
2451 /**
2452 * Return true if subpage links should be expanded on this page.
2453 * @return bool
2454 */
2455 public function areSubpagesAllowed() {
2456 # Some namespaces don't allow subpages
2457 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2458 }
2459
2460 /**
2461 * Handle link to subpage if necessary
2462 *
2463 * @param string $target The source of the link
2464 * @param string &$text The link text, modified as necessary
2465 * @return string The full name of the link
2466 * @private
2467 */
2468 public function maybeDoSubpageLink( $target, &$text ) {
2469 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2470 }
2471
2472 /**
2473 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2474 *
2475 * @param string $text
2476 * @param bool $linestart Whether or not this is at the start of a line.
2477 * @private
2478 * @return string The lists rendered as HTML
2479 */
2480 public function doBlockLevels( $text, $linestart ) {
2481 return BlockLevelPass::doBlockLevels( $text, $linestart );
2482 }
2483
2484 /**
2485 * Return value of a magic variable (like PAGENAME)
2486 *
2487 * @private
2488 *
2489 * @param string $index Magic variable identifier as mapped in MagicWord::$mVariableIDs
2490 * @param bool|PPFrame $frame
2491 *
2492 * @throws MWException
2493 * @return string
2494 */
2495 public function getVariableValue( $index, $frame = false ) {
2496 global $wgContLang, $wgSitename, $wgServer, $wgServerName;
2497 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2498
2499 if ( is_null( $this->mTitle ) ) {
2500 // If no title set, bad things are going to happen
2501 // later. Title should always be set since this
2502 // should only be called in the middle of a parse
2503 // operation (but the unit-tests do funky stuff)
2504 throw new MWException( __METHOD__ . ' Should only be '
2505 . ' called while parsing (no title set)' );
2506 }
2507
2508 // Avoid PHP 7.1 warning from passing $this by reference
2509 $parser = $this;
2510
2511 /**
2512 * Some of these require message or data lookups and can be
2513 * expensive to check many times.
2514 */
2515 if ( Hooks::run( 'ParserGetVariableValueVarCache', [ &$parser, &$this->mVarCache ] ) ) {
2516 if ( isset( $this->mVarCache[$index] ) ) {
2517 return $this->mVarCache[$index];
2518 }
2519 }
2520
2521 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2522 Hooks::run( 'ParserGetVariableValueTs', [ &$parser, &$ts ] );
2523
2524 $pageLang = $this->getFunctionLang();
2525
2526 switch ( $index ) {
2527 case '!':
2528 $value = '|';
2529 break;
2530 case 'currentmonth':
2531 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'm' ), true );
2532 break;
2533 case 'currentmonth1':
2534 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'n' ), true );
2535 break;
2536 case 'currentmonthname':
2537 $value = $pageLang->getMonthName( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2538 break;
2539 case 'currentmonthnamegen':
2540 $value = $pageLang->getMonthNameGen( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2541 break;
2542 case 'currentmonthabbrev':
2543 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2544 break;
2545 case 'currentday':
2546 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'j' ), true );
2547 break;
2548 case 'currentday2':
2549 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'd' ), true );
2550 break;
2551 case 'localmonth':
2552 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'm' ), true );
2553 break;
2554 case 'localmonth1':
2555 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'n' ), true );
2556 break;
2557 case 'localmonthname':
2558 $value = $pageLang->getMonthName( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2559 break;
2560 case 'localmonthnamegen':
2561 $value = $pageLang->getMonthNameGen( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2562 break;
2563 case 'localmonthabbrev':
2564 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2565 break;
2566 case 'localday':
2567 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'j' ), true );
2568 break;
2569 case 'localday2':
2570 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'd' ), true );
2571 break;
2572 case 'pagename':
2573 $value = wfEscapeWikiText( $this->mTitle->getText() );
2574 break;
2575 case 'pagenamee':
2576 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() );
2577 break;
2578 case 'fullpagename':
2579 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2580 break;
2581 case 'fullpagenamee':
2582 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() );
2583 break;
2584 case 'subpagename':
2585 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2586 break;
2587 case 'subpagenamee':
2588 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() );
2589 break;
2590 case 'rootpagename':
2591 $value = wfEscapeWikiText( $this->mTitle->getRootText() );
2592 break;
2593 case 'rootpagenamee':
2594 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2595 ' ',
2596 '_',
2597 $this->mTitle->getRootText()
2598 ) ) );
2599 break;
2600 case 'basepagename':
2601 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2602 break;
2603 case 'basepagenamee':
2604 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2605 ' ',
2606 '_',
2607 $this->mTitle->getBaseText()
2608 ) ) );
2609 break;
2610 case 'talkpagename':
2611 if ( $this->mTitle->canHaveTalkPage() ) {
2612 $talkPage = $this->mTitle->getTalkPage();
2613 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2614 } else {
2615 $value = '';
2616 }
2617 break;
2618 case 'talkpagenamee':
2619 if ( $this->mTitle->canHaveTalkPage() ) {
2620 $talkPage = $this->mTitle->getTalkPage();
2621 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2622 } else {
2623 $value = '';
2624 }
2625 break;
2626 case 'subjectpagename':
2627 $subjPage = $this->mTitle->getSubjectPage();
2628 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2629 break;
2630 case 'subjectpagenamee':
2631 $subjPage = $this->mTitle->getSubjectPage();
2632 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2633 break;
2634 case 'pageid': // requested in T25427
2635 $pageid = $this->getTitle()->getArticleID();
2636 if ( $pageid == 0 ) {
2637 # 0 means the page doesn't exist in the database,
2638 # which means the user is previewing a new page.
2639 # The vary-revision flag must be set, because the magic word
2640 # will have a different value once the page is saved.
2641 $this->mOutput->setFlag( 'vary-revision' );
2642 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2643 }
2644 $value = $pageid ? $pageid : null;
2645 break;
2646 case 'revisionid':
2647 # Let the edit saving system know we should parse the page
2648 # *after* a revision ID has been assigned.
2649 $this->mOutput->setFlag( 'vary-revision-id' );
2650 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision-id...\n" );
2651 $value = $this->mRevisionId;
2652 if ( !$value && $this->mOptions->getSpeculativeRevIdCallback() ) {
2653 $value = call_user_func( $this->mOptions->getSpeculativeRevIdCallback() );
2654 $this->mOutput->setSpeculativeRevIdUsed( $value );
2655 }
2656 break;
2657 case 'revisionday':
2658 $value = (int)$this->getRevisionTimestampSubstring( 6, 2, self::MAX_TTS, $index );
2659 break;
2660 case 'revisionday2':
2661 $value = $this->getRevisionTimestampSubstring( 6, 2, self::MAX_TTS, $index );
2662 break;
2663 case 'revisionmonth':
2664 $value = $this->getRevisionTimestampSubstring( 4, 2, self::MAX_TTS, $index );
2665 break;
2666 case 'revisionmonth1':
2667 $value = (int)$this->getRevisionTimestampSubstring( 4, 2, self::MAX_TTS, $index );
2668 break;
2669 case 'revisionyear':
2670 $value = $this->getRevisionTimestampSubstring( 0, 4, self::MAX_TTS, $index );
2671 break;
2672 case 'revisiontimestamp':
2673 # Let the edit saving system know we should parse the page
2674 # *after* a revision ID has been assigned. This is for null edits.
2675 $this->mOutput->setFlag( 'vary-revision' );
2676 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2677 $value = $this->getRevisionTimestamp();
2678 break;
2679 case 'revisionuser':
2680 # Let the edit saving system know we should parse the page
2681 # *after* a revision ID has been assigned for null edits.
2682 $this->mOutput->setFlag( 'vary-user' );
2683 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-user...\n" );
2684 $value = $this->getRevisionUser();
2685 break;
2686 case 'revisionsize':
2687 $value = $this->getRevisionSize();
2688 break;
2689 case 'namespace':
2690 $value = str_replace( '_', ' ', $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2691 break;
2692 case 'namespacee':
2693 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2694 break;
2695 case 'namespacenumber':
2696 $value = $this->mTitle->getNamespace();
2697 break;
2698 case 'talkspace':
2699 $value = $this->mTitle->canHaveTalkPage()
2700 ? str_replace( '_', ' ', $this->mTitle->getTalkNsText() )
2701 : '';
2702 break;
2703 case 'talkspacee':
2704 $value = $this->mTitle->canHaveTalkPage() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2705 break;
2706 case 'subjectspace':
2707 $value = str_replace( '_', ' ', $this->mTitle->getSubjectNsText() );
2708 break;
2709 case 'subjectspacee':
2710 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2711 break;
2712 case 'currentdayname':
2713 $value = $pageLang->getWeekdayName( (int)MWTimestamp::getInstance( $ts )->format( 'w' ) + 1 );
2714 break;
2715 case 'currentyear':
2716 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'Y' ), true );
2717 break;
2718 case 'currenttime':
2719 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2720 break;
2721 case 'currenthour':
2722 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'H' ), true );
2723 break;
2724 case 'currentweek':
2725 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2726 # int to remove the padding
2727 $value = $pageLang->formatNum( (int)MWTimestamp::getInstance( $ts )->format( 'W' ) );
2728 break;
2729 case 'currentdow':
2730 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'w' ) );
2731 break;
2732 case 'localdayname':
2733 $value = $pageLang->getWeekdayName(
2734 (int)MWTimestamp::getLocalInstance( $ts )->format( 'w' ) + 1
2735 );
2736 break;
2737 case 'localyear':
2738 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'Y' ), true );
2739 break;
2740 case 'localtime':
2741 $value = $pageLang->time(
2742 MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' ),
2743 false,
2744 false
2745 );
2746 break;
2747 case 'localhour':
2748 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'H' ), true );
2749 break;
2750 case 'localweek':
2751 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2752 # int to remove the padding
2753 $value = $pageLang->formatNum( (int)MWTimestamp::getLocalInstance( $ts )->format( 'W' ) );
2754 break;
2755 case 'localdow':
2756 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'w' ) );
2757 break;
2758 case 'numberofarticles':
2759 $value = $pageLang->formatNum( SiteStats::articles() );
2760 break;
2761 case 'numberoffiles':
2762 $value = $pageLang->formatNum( SiteStats::images() );
2763 break;
2764 case 'numberofusers':
2765 $value = $pageLang->formatNum( SiteStats::users() );
2766 break;
2767 case 'numberofactiveusers':
2768 $value = $pageLang->formatNum( SiteStats::activeUsers() );
2769 break;
2770 case 'numberofpages':
2771 $value = $pageLang->formatNum( SiteStats::pages() );
2772 break;
2773 case 'numberofadmins':
2774 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2775 break;
2776 case 'numberofedits':
2777 $value = $pageLang->formatNum( SiteStats::edits() );
2778 break;
2779 case 'currenttimestamp':
2780 $value = wfTimestamp( TS_MW, $ts );
2781 break;
2782 case 'localtimestamp':
2783 $value = MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' );
2784 break;
2785 case 'currentversion':
2786 $value = SpecialVersion::getVersion();
2787 break;
2788 case 'articlepath':
2789 return $wgArticlePath;
2790 case 'sitename':
2791 return $wgSitename;
2792 case 'server':
2793 return $wgServer;
2794 case 'servername':
2795 return $wgServerName;
2796 case 'scriptpath':
2797 return $wgScriptPath;
2798 case 'stylepath':
2799 return $wgStylePath;
2800 case 'directionmark':
2801 return $pageLang->getDirMark();
2802 case 'contentlanguage':
2803 global $wgLanguageCode;
2804 return $wgLanguageCode;
2805 case 'pagelanguage':
2806 $value = $pageLang->getCode();
2807 break;
2808 case 'cascadingsources':
2809 $value = CoreParserFunctions::cascadingsources( $this );
2810 break;
2811 default:
2812 $ret = null;
2813 Hooks::run(
2814 'ParserGetVariableValueSwitch',
2815 [ &$parser, &$this->mVarCache, &$index, &$ret, &$frame ]
2816 );
2817