Merge "Show a warning in edit preview when a template loop is detected"
[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 * $wgMaxArticleSize
67 *
68 * @ingroup Parser
69 */
70 class Parser {
71 /**
72 * Update this version number when the ParserOutput format
73 * changes in an incompatible way, so the parser cache
74 * can automatically discard old data.
75 */
76 const VERSION = '1.6.4';
77
78 /**
79 * Update this version number when the output of serialiseHalfParsedText()
80 * changes in an incompatible way
81 */
82 const HALF_PARSED_VERSION = 2;
83
84 # Flags for Parser::setFunctionHook
85 const SFH_NO_HASH = 1;
86 const SFH_OBJECT_ARGS = 2;
87
88 # Constants needed for external link processing
89 # Everything except bracket, space, or control characters
90 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
91 # as well as U+3000 is IDEOGRAPHIC SPACE for T21052
92 # \x{FFFD} is the Unicode replacement character, which Preprocessor_DOM
93 # uses to replace invalid HTML characters.
94 const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]';
95 # Simplified expression to match an IPv4 or IPv6 address, or
96 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
97 const EXT_LINK_ADDR = '(?:[0-9.]+|\\[(?i:[0-9a-f:.]+)\\]|[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}])';
98 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
99 // @codingStandardsIgnoreStart Generic.Files.LineLength
100 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]+)
101 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
102 // @codingStandardsIgnoreEnd
103
104 # Regular expression for a non-newline space
105 const SPACE_NOT_NL = '(?:\t|&nbsp;|&\#0*160;|&\#[Xx]0*[Aa]0;|\p{Zs})';
106
107 # Flags for preprocessToDom
108 const PTD_FOR_INCLUSION = 1;
109
110 # Allowed values for $this->mOutputType
111 # Parameter to startExternalParse().
112 const OT_HTML = 1; # like parse()
113 const OT_WIKI = 2; # like preSaveTransform()
114 const OT_PREPROCESS = 3; # like preprocess()
115 const OT_MSG = 3;
116 const OT_PLAIN = 4; # like extractSections() - portions of the original are returned unchanged.
117
118 /**
119 * @var string Prefix and suffix for temporary replacement strings
120 * for the multipass parser.
121 *
122 * \x7f should never appear in input as it's disallowed in XML.
123 * Using it at the front also gives us a little extra robustness
124 * since it shouldn't match when butted up against identifier-like
125 * string constructs.
126 *
127 * Must not consist of all title characters, or else it will change
128 * the behavior of <nowiki> in a link.
129 *
130 * Must have a character that needs escaping in attributes, otherwise
131 * someone could put a strip marker in an attribute, to get around
132 * escaping quote marks, and break out of the attribute. Thus we add
133 * `'".
134 */
135 const MARKER_SUFFIX = "-QINU`\"'\x7f";
136 const MARKER_PREFIX = "\x7f'\"`UNIQ-";
137
138 # Markers used for wrapping the table of contents
139 const TOC_START = '<mw:toc>';
140 const TOC_END = '</mw:toc>';
141
142 # Persistent:
143 public $mTagHooks = [];
144 public $mTransparentTagHooks = [];
145 public $mFunctionHooks = [];
146 public $mFunctionSynonyms = [ 0 => [], 1 => [] ];
147 public $mFunctionTagHooks = [];
148 public $mStripList = [];
149 public $mDefaultStripList = [];
150 public $mVarCache = [];
151 public $mImageParams = [];
152 public $mImageParamsMagicArray = [];
153 public $mMarkerIndex = 0;
154 public $mFirstCall = true;
155
156 # Initialised by initialiseVariables()
157
158 /**
159 * @var MagicWordArray
160 */
161 public $mVariables;
162
163 /**
164 * @var MagicWordArray
165 */
166 public $mSubstWords;
167 # Initialised in constructor
168 public $mConf, $mExtLinkBracketedRegex, $mUrlProtocols;
169
170 # Initialized in getPreprocessor()
171 /** @var Preprocessor */
172 public $mPreprocessor;
173
174 # Cleared with clearState():
175 /**
176 * @var ParserOutput
177 */
178 public $mOutput;
179 public $mAutonumber;
180
181 /**
182 * @var StripState
183 */
184 public $mStripState;
185
186 public $mIncludeCount;
187 /**
188 * @var LinkHolderArray
189 */
190 public $mLinkHolders;
191
192 public $mLinkID;
193 public $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth;
194 public $mDefaultSort;
195 public $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
196 public $mExpensiveFunctionCount; # number of expensive parser function calls
197 public $mShowToc, $mForceTocPosition;
198
199 /**
200 * @var User
201 */
202 public $mUser; # User object; only used when doing pre-save transform
203
204 # Temporary
205 # These are variables reset at least once per parse regardless of $clearState
206
207 /**
208 * @var ParserOptions
209 */
210 public $mOptions;
211
212 /**
213 * @var Title
214 */
215 public $mTitle; # Title context, used for self-link rendering and similar things
216 public $mOutputType; # Output type, one of the OT_xxx constants
217 public $ot; # Shortcut alias, see setOutputType()
218 public $mRevisionObject; # The revision object of the specified revision ID
219 public $mRevisionId; # ID to display in {{REVISIONID}} tags
220 public $mRevisionTimestamp; # The timestamp of the specified revision ID
221 public $mRevisionUser; # User to display in {{REVISIONUSER}} tag
222 public $mRevisionSize; # Size to display in {{REVISIONSIZE}} variable
223 public $mRevIdForTs; # The revision ID which was used to fetch the timestamp
224 public $mInputSize = false; # For {{PAGESIZE}} on current page.
225
226 /**
227 * @var string Deprecated accessor for the strip marker prefix.
228 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
229 */
230 public $mUniqPrefix = self::MARKER_PREFIX;
231
232 /**
233 * @var array Array with the language name of each language link (i.e. the
234 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
235 * duplicate language links to the ParserOutput.
236 */
237 public $mLangLinkLanguages;
238
239 /**
240 * @var MapCacheLRU|null
241 * @since 1.24
242 *
243 * A cache of the current revisions of titles. Keys are $title->getPrefixedDbKey()
244 */
245 public $currentRevisionCache;
246
247 /**
248 * @var bool|string Recursive call protection.
249 * This variable should be treated as if it were private.
250 */
251 public $mInParse = false;
252
253 /** @var SectionProfiler */
254 protected $mProfiler;
255
256 /**
257 * @var LinkRenderer
258 */
259 protected $mLinkRenderer;
260
261 /**
262 * @param array $conf
263 */
264 public function __construct( $conf = [] ) {
265 $this->mConf = $conf;
266 $this->mUrlProtocols = wfUrlProtocols();
267 $this->mExtLinkBracketedRegex = '/\[(((?i)' . $this->mUrlProtocols . ')' .
268 self::EXT_LINK_ADDR .
269 self::EXT_LINK_URL_CLASS . '*)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F\\x{FFFD}]*?)\]/Su';
270 if ( isset( $conf['preprocessorClass'] ) ) {
271 $this->mPreprocessorClass = $conf['preprocessorClass'];
272 } elseif ( defined( 'HPHP_VERSION' ) ) {
273 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
274 $this->mPreprocessorClass = 'Preprocessor_Hash';
275 } elseif ( extension_loaded( 'domxml' ) ) {
276 # PECL extension that conflicts with the core DOM extension (T15770)
277 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
278 $this->mPreprocessorClass = 'Preprocessor_Hash';
279 } elseif ( extension_loaded( 'dom' ) ) {
280 $this->mPreprocessorClass = 'Preprocessor_DOM';
281 } else {
282 $this->mPreprocessorClass = 'Preprocessor_Hash';
283 }
284 wfDebug( __CLASS__ . ": using preprocessor: {$this->mPreprocessorClass}\n" );
285 }
286
287 /**
288 * Reduce memory usage to reduce the impact of circular references
289 */
290 public function __destruct() {
291 if ( isset( $this->mLinkHolders ) ) {
292 unset( $this->mLinkHolders );
293 }
294 foreach ( $this as $name => $value ) {
295 unset( $this->$name );
296 }
297 }
298
299 /**
300 * Allow extensions to clean up when the parser is cloned
301 */
302 public function __clone() {
303 $this->mInParse = false;
304
305 // T58226: When you create a reference "to" an object field, that
306 // makes the object field itself be a reference too (until the other
307 // reference goes out of scope). When cloning, any field that's a
308 // reference is copied as a reference in the new object. Both of these
309 // are defined PHP5 behaviors, as inconvenient as it is for us when old
310 // hooks from PHP4 days are passing fields by reference.
311 foreach ( [ 'mStripState', 'mVarCache' ] as $k ) {
312 // Make a non-reference copy of the field, then rebind the field to
313 // reference the new copy.
314 $tmp = $this->$k;
315 $this->$k =& $tmp;
316 unset( $tmp );
317 }
318
319 Hooks::run( 'ParserCloned', [ $this ] );
320 }
321
322 /**
323 * Do various kinds of initialisation on the first call of the parser
324 */
325 public function firstCallInit() {
326 if ( !$this->mFirstCall ) {
327 return;
328 }
329 $this->mFirstCall = false;
330
331 CoreParserFunctions::register( $this );
332 CoreTagHooks::register( $this );
333 $this->initialiseVariables();
334
335 // Avoid PHP 7.1 warning from passing $this by reference
336 $parser = $this;
337 Hooks::run( 'ParserFirstCallInit', [ &$parser ] );
338 }
339
340 /**
341 * Clear Parser state
342 *
343 * @private
344 */
345 public function clearState() {
346 if ( $this->mFirstCall ) {
347 $this->firstCallInit();
348 }
349 $this->mOutput = new ParserOutput;
350 $this->mOptions->registerWatcher( [ $this->mOutput, 'recordOption' ] );
351 $this->mAutonumber = 0;
352 $this->mIncludeCount = [];
353 $this->mLinkHolders = new LinkHolderArray( $this );
354 $this->mLinkID = 0;
355 $this->mRevisionObject = $this->mRevisionTimestamp =
356 $this->mRevisionId = $this->mRevisionUser = $this->mRevisionSize = null;
357 $this->mVarCache = [];
358 $this->mUser = null;
359 $this->mLangLinkLanguages = [];
360 $this->currentRevisionCache = null;
361
362 $this->mStripState = new StripState;
363
364 # Clear these on every parse, T6549
365 $this->mTplRedirCache = $this->mTplDomCache = [];
366
367 $this->mShowToc = true;
368 $this->mForceTocPosition = false;
369 $this->mIncludeSizes = [
370 'post-expand' => 0,
371 'arg' => 0,
372 ];
373 $this->mPPNodeCount = 0;
374 $this->mGeneratedPPNodeCount = 0;
375 $this->mHighestExpansionDepth = 0;
376 $this->mDefaultSort = false;
377 $this->mHeadings = [];
378 $this->mDoubleUnderscores = [];
379 $this->mExpensiveFunctionCount = 0;
380
381 # Fix cloning
382 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) {
383 $this->mPreprocessor = null;
384 }
385
386 $this->mProfiler = new SectionProfiler();
387
388 // Avoid PHP 7.1 warning from passing $this by reference
389 $parser = $this;
390 Hooks::run( 'ParserClearState', [ &$parser ] );
391 }
392
393 /**
394 * Convert wikitext to HTML
395 * Do not call this function recursively.
396 *
397 * @param string $text Text we want to parse
398 * @param Title $title
399 * @param ParserOptions $options
400 * @param bool $linestart
401 * @param bool $clearState
402 * @param int $revid Number to pass in {{REVISIONID}}
403 * @return ParserOutput A ParserOutput
404 */
405 public function parse(
406 $text, Title $title, ParserOptions $options,
407 $linestart = true, $clearState = true, $revid = null
408 ) {
409 /**
410 * First pass--just handle <nowiki> sections, pass the rest off
411 * to internalParse() which does all the real work.
412 */
413
414 global $wgShowHostnames;
415
416 if ( $clearState ) {
417 // We use U+007F DELETE to construct strip markers, so we have to make
418 // sure that this character does not occur in the input text.
419 $text = strtr( $text, "\x7f", "?" );
420 $magicScopeVariable = $this->lock();
421 }
422 // Strip U+0000 NULL (T159174)
423 $text = str_replace( "\000", '', $text );
424
425 $this->startParse( $title, $options, self::OT_HTML, $clearState );
426
427 $this->currentRevisionCache = null;
428 $this->mInputSize = strlen( $text );
429 if ( $this->mOptions->getEnableLimitReport() ) {
430 $this->mOutput->resetParseStartTime();
431 }
432
433 $oldRevisionId = $this->mRevisionId;
434 $oldRevisionObject = $this->mRevisionObject;
435 $oldRevisionTimestamp = $this->mRevisionTimestamp;
436 $oldRevisionUser = $this->mRevisionUser;
437 $oldRevisionSize = $this->mRevisionSize;
438 if ( $revid !== null ) {
439 $this->mRevisionId = $revid;
440 $this->mRevisionObject = null;
441 $this->mRevisionTimestamp = null;
442 $this->mRevisionUser = null;
443 $this->mRevisionSize = null;
444 }
445
446 // Avoid PHP 7.1 warning from passing $this by reference
447 $parser = $this;
448 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
449 # No more strip!
450 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
451 $text = $this->internalParse( $text );
452 Hooks::run( 'ParserAfterParse', [ &$parser, &$text, &$this->mStripState ] );
453
454 $text = $this->internalParseHalfParsed( $text, true, $linestart );
455
456 /**
457 * A converted title will be provided in the output object if title and
458 * content conversion are enabled, the article text does not contain
459 * a conversion-suppressing double-underscore tag, and no
460 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
461 * automatic link conversion.
462 */
463 if ( !( $options->getDisableTitleConversion()
464 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
465 || isset( $this->mDoubleUnderscores['notitleconvert'] )
466 || $this->mOutput->getDisplayTitle() !== false )
467 ) {
468 $convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
469 if ( $convruletitle ) {
470 $this->mOutput->setTitleText( $convruletitle );
471 } else {
472 $titleText = $this->getConverterLanguage()->convertTitle( $title );
473 $this->mOutput->setTitleText( $titleText );
474 }
475 }
476
477 # Done parsing! Compute runtime adaptive expiry if set
478 $this->mOutput->finalizeAdaptiveCacheExpiry();
479
480 # Warn if too many heavyweight parser functions were used
481 if ( $this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit() ) {
482 $this->limitationWarn( 'expensive-parserfunction',
483 $this->mExpensiveFunctionCount,
484 $this->mOptions->getExpensiveParserFunctionLimit()
485 );
486 }
487
488 # Information on include size limits, for the benefit of users who try to skirt them
489 if ( $this->mOptions->getEnableLimitReport() ) {
490 $max = $this->mOptions->getMaxIncludeSize();
491
492 $cpuTime = $this->mOutput->getTimeSinceStart( 'cpu' );
493 if ( $cpuTime !== null ) {
494 $this->mOutput->setLimitReportData( 'limitreport-cputime',
495 sprintf( "%.3f", $cpuTime )
496 );
497 }
498
499 $wallTime = $this->mOutput->getTimeSinceStart( 'wall' );
500 $this->mOutput->setLimitReportData( 'limitreport-walltime',
501 sprintf( "%.3f", $wallTime )
502 );
503
504 $this->mOutput->setLimitReportData( 'limitreport-ppvisitednodes',
505 [ $this->mPPNodeCount, $this->mOptions->getMaxPPNodeCount() ]
506 );
507 $this->mOutput->setLimitReportData( 'limitreport-ppgeneratednodes',
508 [ $this->mGeneratedPPNodeCount, $this->mOptions->getMaxGeneratedPPNodeCount() ]
509 );
510 $this->mOutput->setLimitReportData( 'limitreport-postexpandincludesize',
511 [ $this->mIncludeSizes['post-expand'], $max ]
512 );
513 $this->mOutput->setLimitReportData( 'limitreport-templateargumentsize',
514 [ $this->mIncludeSizes['arg'], $max ]
515 );
516 $this->mOutput->setLimitReportData( 'limitreport-expansiondepth',
517 [ $this->mHighestExpansionDepth, $this->mOptions->getMaxPPExpandDepth() ]
518 );
519 $this->mOutput->setLimitReportData( 'limitreport-expensivefunctioncount',
520 [ $this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit() ]
521 );
522 Hooks::run( 'ParserLimitReportPrepare', [ $this, $this->mOutput ] );
523
524 $limitReport = "NewPP limit report\n";
525 if ( $wgShowHostnames ) {
526 $limitReport .= 'Parsed by ' . wfHostname() . "\n";
527 }
528 $limitReport .= 'Cached time: ' . $this->mOutput->getCacheTime() . "\n";
529 $limitReport .= 'Cache expiry: ' . $this->mOutput->getCacheExpiry() . "\n";
530 $limitReport .= 'Dynamic content: ' .
531 ( $this->mOutput->hasDynamicContent() ? 'true' : 'false' ) .
532 "\n";
533
534 foreach ( $this->mOutput->getLimitReportData() as $key => $value ) {
535 if ( Hooks::run( 'ParserLimitReportFormat',
536 [ $key, &$value, &$limitReport, false, false ]
537 ) ) {
538 $keyMsg = wfMessage( $key )->inLanguage( 'en' )->useDatabase( false );
539 $valueMsg = wfMessage( [ "$key-value-text", "$key-value" ] )
540 ->inLanguage( 'en' )->useDatabase( false );
541 if ( !$valueMsg->exists() ) {
542 $valueMsg = new RawMessage( '$1' );
543 }
544 if ( !$keyMsg->isDisabled() && !$valueMsg->isDisabled() ) {
545 $valueMsg->params( $value );
546 $limitReport .= "{$keyMsg->text()}: {$valueMsg->text()}\n";
547 }
548 }
549 }
550 // Since we're not really outputting HTML, decode the entities and
551 // then re-encode the things that need hiding inside HTML comments.
552 $limitReport = htmlspecialchars_decode( $limitReport );
553 // Run deprecated hook
554 Hooks::run( 'ParserLimitReport', [ $this, &$limitReport ], '1.22' );
555
556 // Sanitize for comment. Note '‐' in the replacement is U+2010,
557 // which looks much like the problematic '-'.
558 $limitReport = str_replace( [ '-', '&' ], [ '‐', '&amp;' ], $limitReport );
559 $text .= "\n<!-- \n$limitReport-->\n";
560
561 // Add on template profiling data in human/machine readable way
562 $dataByFunc = $this->mProfiler->getFunctionStats();
563 uasort( $dataByFunc, function ( $a, $b ) {
564 return $a['real'] < $b['real']; // descending order
565 } );
566 $profileReport = [];
567 foreach ( array_slice( $dataByFunc, 0, 10 ) as $item ) {
568 $profileReport[] = sprintf( "%6.2f%% %8.3f %6d %s",
569 $item['%real'], $item['real'], $item['calls'],
570 htmlspecialchars( $item['name'] ) );
571 }
572 $text .= "<!--\nTransclusion expansion time report (%,ms,calls,template)\n";
573 $text .= implode( "\n", $profileReport ) . "\n-->\n";
574
575 $this->mOutput->setLimitReportData( 'limitreport-timingprofile', $profileReport );
576
577 // Add other cache related metadata
578 if ( $wgShowHostnames ) {
579 $this->mOutput->setLimitReportData( 'cachereport-origin', wfHostname() );
580 }
581 $this->mOutput->setLimitReportData( 'cachereport-timestamp',
582 $this->mOutput->getCacheTime() );
583 $this->mOutput->setLimitReportData( 'cachereport-ttl',
584 $this->mOutput->getCacheExpiry() );
585 $this->mOutput->setLimitReportData( 'cachereport-transientcontent',
586 $this->mOutput->hasDynamicContent() );
587
588 if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) {
589 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' .
590 $this->mTitle->getPrefixedDBkey() );
591 }
592 }
593
594 # Wrap non-interface parser output in a <div> so it can be targeted
595 # with CSS (T37247)
596 $class = $this->mOptions->getWrapOutputClass();
597 if ( $class !== false && !$this->mOptions->getInterfaceMessage() ) {
598 $text = Html::rawElement( 'div', [ 'class' => $class ], $text );
599 }
600
601 $this->mOutput->setText( $text );
602
603 $this->mRevisionId = $oldRevisionId;
604 $this->mRevisionObject = $oldRevisionObject;
605 $this->mRevisionTimestamp = $oldRevisionTimestamp;
606 $this->mRevisionUser = $oldRevisionUser;
607 $this->mRevisionSize = $oldRevisionSize;
608 $this->mInputSize = false;
609 $this->currentRevisionCache = null;
610
611 return $this->mOutput;
612 }
613
614 /**
615 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
616 * can be called from an extension tag hook.
617 *
618 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
619 * instead, which means that lists and links have not been fully parsed yet,
620 * and strip markers are still present.
621 *
622 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
623 *
624 * Use this function if you're a parser tag hook and you want to parse
625 * wikitext before or after applying additional transformations, and you
626 * intend to *return the result as hook output*, which will cause it to go
627 * through the rest of parsing process automatically.
628 *
629 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
630 * $text are not expanded
631 *
632 * @param string $text Text extension wants to have parsed
633 * @param bool|PPFrame $frame The frame to use for expanding any template variables
634 * @return string UNSAFE half-parsed HTML
635 */
636 public function recursiveTagParse( $text, $frame = false ) {
637 // Avoid PHP 7.1 warning from passing $this by reference
638 $parser = $this;
639 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
640 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
641 $text = $this->internalParse( $text, false, $frame );
642 return $text;
643 }
644
645 /**
646 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
647 * point can be called from an extension tag hook.
648 *
649 * The output of this function is fully-parsed HTML that is safe for output.
650 * If you're a parser tag hook, you might want to use recursiveTagParse()
651 * instead.
652 *
653 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
654 * $text are not expanded
655 *
656 * @since 1.25
657 *
658 * @param string $text Text extension wants to have parsed
659 * @param bool|PPFrame $frame The frame to use for expanding any template variables
660 * @return string Fully parsed HTML
661 */
662 public function recursiveTagParseFully( $text, $frame = false ) {
663 $text = $this->recursiveTagParse( $text, $frame );
664 $text = $this->internalParseHalfParsed( $text, false );
665 return $text;
666 }
667
668 /**
669 * Expand templates and variables in the text, producing valid, static wikitext.
670 * Also removes comments.
671 * Do not call this function recursively.
672 * @param string $text
673 * @param Title $title
674 * @param ParserOptions $options
675 * @param int|null $revid
676 * @param bool|PPFrame $frame
677 * @return mixed|string
678 */
679 public function preprocess( $text, Title $title = null,
680 ParserOptions $options, $revid = null, $frame = false
681 ) {
682 $magicScopeVariable = $this->lock();
683 $this->startParse( $title, $options, self::OT_PREPROCESS, true );
684 if ( $revid !== null ) {
685 $this->mRevisionId = $revid;
686 }
687 // Avoid PHP 7.1 warning from passing $this by reference
688 $parser = $this;
689 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
690 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
691 $text = $this->replaceVariables( $text, $frame );
692 $text = $this->mStripState->unstripBoth( $text );
693 return $text;
694 }
695
696 /**
697 * Recursive parser entry point that can be called from an extension tag
698 * hook.
699 *
700 * @param string $text Text to be expanded
701 * @param bool|PPFrame $frame The frame to use for expanding any template variables
702 * @return string
703 * @since 1.19
704 */
705 public function recursivePreprocess( $text, $frame = false ) {
706 $text = $this->replaceVariables( $text, $frame );
707 $text = $this->mStripState->unstripBoth( $text );
708 return $text;
709 }
710
711 /**
712 * Process the wikitext for the "?preload=" feature. (T7210)
713 *
714 * "<noinclude>", "<includeonly>" etc. are parsed as for template
715 * transclusion, comments, templates, arguments, tags hooks and parser
716 * functions are untouched.
717 *
718 * @param string $text
719 * @param Title $title
720 * @param ParserOptions $options
721 * @param array $params
722 * @return string
723 */
724 public function getPreloadText( $text, Title $title, ParserOptions $options, $params = [] ) {
725 $msg = new RawMessage( $text );
726 $text = $msg->params( $params )->plain();
727
728 # Parser (re)initialisation
729 $magicScopeVariable = $this->lock();
730 $this->startParse( $title, $options, self::OT_PLAIN, true );
731
732 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES;
733 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
734 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
735 $text = $this->mStripState->unstripBoth( $text );
736 return $text;
737 }
738
739 /**
740 * Set the current user.
741 * Should only be used when doing pre-save transform.
742 *
743 * @param User|null $user User object or null (to reset)
744 */
745 public function setUser( $user ) {
746 $this->mUser = $user;
747 }
748
749 /**
750 * Set the context title
751 *
752 * @param Title $t
753 */
754 public function setTitle( $t ) {
755 if ( !$t ) {
756 $t = Title::newFromText( 'NO TITLE' );
757 }
758
759 if ( $t->hasFragment() ) {
760 # Strip the fragment to avoid various odd effects
761 $this->mTitle = $t->createFragmentTarget( '' );
762 } else {
763 $this->mTitle = $t;
764 }
765 }
766
767 /**
768 * Accessor for the Title object
769 *
770 * @return Title
771 */
772 public function getTitle() {
773 return $this->mTitle;
774 }
775
776 /**
777 * Accessor/mutator for the Title object
778 *
779 * @param Title $x Title object or null to just get the current one
780 * @return Title
781 */
782 public function Title( $x = null ) {
783 return wfSetVar( $this->mTitle, $x );
784 }
785
786 /**
787 * Set the output type
788 *
789 * @param int $ot New value
790 */
791 public function setOutputType( $ot ) {
792 $this->mOutputType = $ot;
793 # Shortcut alias
794 $this->ot = [
795 'html' => $ot == self::OT_HTML,
796 'wiki' => $ot == self::OT_WIKI,
797 'pre' => $ot == self::OT_PREPROCESS,
798 'plain' => $ot == self::OT_PLAIN,
799 ];
800 }
801
802 /**
803 * Accessor/mutator for the output type
804 *
805 * @param int|null $x New value or null to just get the current one
806 * @return int
807 */
808 public function OutputType( $x = null ) {
809 return wfSetVar( $this->mOutputType, $x );
810 }
811
812 /**
813 * Get the ParserOutput object
814 *
815 * @return ParserOutput
816 */
817 public function getOutput() {
818 return $this->mOutput;
819 }
820
821 /**
822 * Get the ParserOptions object
823 *
824 * @return ParserOptions
825 */
826 public function getOptions() {
827 return $this->mOptions;
828 }
829
830 /**
831 * Accessor/mutator for the ParserOptions object
832 *
833 * @param ParserOptions $x New value or null to just get the current one
834 * @return ParserOptions Current ParserOptions object
835 */
836 public function Options( $x = null ) {
837 return wfSetVar( $this->mOptions, $x );
838 }
839
840 /**
841 * @return int
842 */
843 public function nextLinkID() {
844 return $this->mLinkID++;
845 }
846
847 /**
848 * @param int $id
849 */
850 public function setLinkID( $id ) {
851 $this->mLinkID = $id;
852 }
853
854 /**
855 * Get a language object for use in parser functions such as {{FORMATNUM:}}
856 * @return Language
857 */
858 public function getFunctionLang() {
859 return $this->getTargetLanguage();
860 }
861
862 /**
863 * Get the target language for the content being parsed. This is usually the
864 * language that the content is in.
865 *
866 * @since 1.19
867 *
868 * @throws MWException
869 * @return Language
870 */
871 public function getTargetLanguage() {
872 $target = $this->mOptions->getTargetLanguage();
873
874 if ( $target !== null ) {
875 return $target;
876 } elseif ( $this->mOptions->getInterfaceMessage() ) {
877 return $this->mOptions->getUserLangObj();
878 } elseif ( is_null( $this->mTitle ) ) {
879 throw new MWException( __METHOD__ . ': $this->mTitle is null' );
880 }
881
882 return $this->mTitle->getPageLanguage();
883 }
884
885 /**
886 * Get the language object for language conversion
887 * @return Language|null
888 */
889 public function getConverterLanguage() {
890 return $this->getTargetLanguage();
891 }
892
893 /**
894 * Get a User object either from $this->mUser, if set, or from the
895 * ParserOptions object otherwise
896 *
897 * @return User
898 */
899 public function getUser() {
900 if ( !is_null( $this->mUser ) ) {
901 return $this->mUser;
902 }
903 return $this->mOptions->getUser();
904 }
905
906 /**
907 * Get a preprocessor object
908 *
909 * @return Preprocessor
910 */
911 public function getPreprocessor() {
912 if ( !isset( $this->mPreprocessor ) ) {
913 $class = $this->mPreprocessorClass;
914 $this->mPreprocessor = new $class( $this );
915 }
916 return $this->mPreprocessor;
917 }
918
919 /**
920 * Get a LinkRenderer instance to make links with
921 *
922 * @since 1.28
923 * @return LinkRenderer
924 */
925 public function getLinkRenderer() {
926 if ( !$this->mLinkRenderer ) {
927 $this->mLinkRenderer = MediaWikiServices::getInstance()
928 ->getLinkRendererFactory()->create();
929 $this->mLinkRenderer->setStubThreshold(
930 $this->getOptions()->getStubThreshold()
931 );
932 }
933
934 return $this->mLinkRenderer;
935 }
936
937 /**
938 * Replaces all occurrences of HTML-style comments and the given tags
939 * in the text with a random marker and returns the next text. The output
940 * parameter $matches will be an associative array filled with data in
941 * the form:
942 *
943 * @code
944 * 'UNIQ-xxxxx' => [
945 * 'element',
946 * 'tag content',
947 * [ 'param' => 'x' ],
948 * '<element param="x">tag content</element>' ]
949 * @endcode
950 *
951 * @param array $elements List of element names. Comments are always extracted.
952 * @param string $text Source text string.
953 * @param array &$matches Out parameter, Array: extracted tags
954 * @return string Stripped text
955 */
956 public static function extractTagsAndParams( $elements, $text, &$matches ) {
957 static $n = 1;
958 $stripped = '';
959 $matches = [];
960
961 $taglist = implode( '|', $elements );
962 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
963
964 while ( $text != '' ) {
965 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
966 $stripped .= $p[0];
967 if ( count( $p ) < 5 ) {
968 break;
969 }
970 if ( count( $p ) > 5 ) {
971 # comment
972 $element = $p[4];
973 $attributes = '';
974 $close = '';
975 $inside = $p[5];
976 } else {
977 # tag
978 $element = $p[1];
979 $attributes = $p[2];
980 $close = $p[3];
981 $inside = $p[4];
982 }
983
984 $marker = self::MARKER_PREFIX . "-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
985 $stripped .= $marker;
986
987 if ( $close === '/>' ) {
988 # Empty element tag, <tag />
989 $content = null;
990 $text = $inside;
991 $tail = null;
992 } else {
993 if ( $element === '!--' ) {
994 $end = '/(-->)/';
995 } else {
996 $end = "/(<\\/$element\\s*>)/i";
997 }
998 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
999 $content = $q[0];
1000 if ( count( $q ) < 3 ) {
1001 # No end tag -- let it run out to the end of the text.
1002 $tail = '';
1003 $text = '';
1004 } else {
1005 $tail = $q[1];
1006 $text = $q[2];
1007 }
1008 }
1009
1010 $matches[$marker] = [ $element,
1011 $content,
1012 Sanitizer::decodeTagAttributes( $attributes ),
1013 "<$element$attributes$close$content$tail" ];
1014 }
1015 return $stripped;
1016 }
1017
1018 /**
1019 * Get a list of strippable XML-like elements
1020 *
1021 * @return array
1022 */
1023 public function getStripList() {
1024 return $this->mStripList;
1025 }
1026
1027 /**
1028 * Add an item to the strip state
1029 * Returns the unique tag which must be inserted into the stripped text
1030 * The tag will be replaced with the original text in unstrip()
1031 *
1032 * @param string $text
1033 *
1034 * @return string
1035 */
1036 public function insertStripItem( $text ) {
1037 $marker = self::MARKER_PREFIX . "-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
1038 $this->mMarkerIndex++;
1039 $this->mStripState->addGeneral( $marker, $text );
1040 return $marker;
1041 }
1042
1043 /**
1044 * parse the wiki syntax used to render tables
1045 *
1046 * @private
1047 * @param string $text
1048 * @return string
1049 */
1050 public function doTableStuff( $text ) {
1051 $lines = StringUtils::explode( "\n", $text );
1052 $out = '';
1053 $td_history = []; # Is currently a td tag open?
1054 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1055 $tr_history = []; # Is currently a tr tag open?
1056 $tr_attributes = []; # history of tr attributes
1057 $has_opened_tr = []; # Did this table open a <tr> element?
1058 $indent_level = 0; # indent level of the table
1059
1060 foreach ( $lines as $outLine ) {
1061 $line = trim( $outLine );
1062
1063 if ( $line === '' ) { # empty line, go to next line
1064 $out .= $outLine . "\n";
1065 continue;
1066 }
1067
1068 $first_character = $line[0];
1069 $first_two = substr( $line, 0, 2 );
1070 $matches = [];
1071
1072 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1073 # First check if we are starting a new table
1074 $indent_level = strlen( $matches[1] );
1075
1076 $attributes = $this->mStripState->unstripBoth( $matches[2] );
1077 $attributes = Sanitizer::fixTagAttributes( $attributes, 'table' );
1078
1079 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1080 array_push( $td_history, false );
1081 array_push( $last_tag_history, '' );
1082 array_push( $tr_history, false );
1083 array_push( $tr_attributes, '' );
1084 array_push( $has_opened_tr, false );
1085 } elseif ( count( $td_history ) == 0 ) {
1086 # Don't do any of the following
1087 $out .= $outLine . "\n";
1088 continue;
1089 } elseif ( $first_two === '|}' ) {
1090 # We are ending a table
1091 $line = '</table>' . substr( $line, 2 );
1092 $last_tag = array_pop( $last_tag_history );
1093
1094 if ( !array_pop( $has_opened_tr ) ) {
1095 $line = "<tr><td></td></tr>{$line}";
1096 }
1097
1098 if ( array_pop( $tr_history ) ) {
1099 $line = "</tr>{$line}";
1100 }
1101
1102 if ( array_pop( $td_history ) ) {
1103 $line = "</{$last_tag}>{$line}";
1104 }
1105 array_pop( $tr_attributes );
1106 $outLine = $line . str_repeat( '</dd></dl>', $indent_level );
1107 } elseif ( $first_two === '|-' ) {
1108 # Now we have a table row
1109 $line = preg_replace( '#^\|-+#', '', $line );
1110
1111 # Whats after the tag is now only attributes
1112 $attributes = $this->mStripState->unstripBoth( $line );
1113 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
1114 array_pop( $tr_attributes );
1115 array_push( $tr_attributes, $attributes );
1116
1117 $line = '';
1118 $last_tag = array_pop( $last_tag_history );
1119 array_pop( $has_opened_tr );
1120 array_push( $has_opened_tr, true );
1121
1122 if ( array_pop( $tr_history ) ) {
1123 $line = '</tr>';
1124 }
1125
1126 if ( array_pop( $td_history ) ) {
1127 $line = "</{$last_tag}>{$line}";
1128 }
1129
1130 $outLine = $line;
1131 array_push( $tr_history, false );
1132 array_push( $td_history, false );
1133 array_push( $last_tag_history, '' );
1134 } elseif ( $first_character === '|'
1135 || $first_character === '!'
1136 || $first_two === '|+'
1137 ) {
1138 # This might be cell elements, td, th or captions
1139 if ( $first_two === '|+' ) {
1140 $first_character = '+';
1141 $line = substr( $line, 2 );
1142 } else {
1143 $line = substr( $line, 1 );
1144 }
1145
1146 // Implies both are valid for table headings.
1147 if ( $first_character === '!' ) {
1148 $line = StringUtils::replaceMarkup( '!!', '||', $line );
1149 }
1150
1151 # Split up multiple cells on the same line.
1152 # FIXME : This can result in improper nesting of tags processed
1153 # by earlier parser steps.
1154 $cells = explode( '||', $line );
1155
1156 $outLine = '';
1157
1158 # Loop through each table cell
1159 foreach ( $cells as $cell ) {
1160 $previous = '';
1161 if ( $first_character !== '+' ) {
1162 $tr_after = array_pop( $tr_attributes );
1163 if ( !array_pop( $tr_history ) ) {
1164 $previous = "<tr{$tr_after}>\n";
1165 }
1166 array_push( $tr_history, true );
1167 array_push( $tr_attributes, '' );
1168 array_pop( $has_opened_tr );
1169 array_push( $has_opened_tr, true );
1170 }
1171
1172 $last_tag = array_pop( $last_tag_history );
1173
1174 if ( array_pop( $td_history ) ) {
1175 $previous = "</{$last_tag}>\n{$previous}";
1176 }
1177
1178 if ( $first_character === '|' ) {
1179 $last_tag = 'td';
1180 } elseif ( $first_character === '!' ) {
1181 $last_tag = 'th';
1182 } elseif ( $first_character === '+' ) {
1183 $last_tag = 'caption';
1184 } else {
1185 $last_tag = '';
1186 }
1187
1188 array_push( $last_tag_history, $last_tag );
1189
1190 # A cell could contain both parameters and data
1191 $cell_data = explode( '|', $cell, 2 );
1192
1193 # T2553: Note that a '|' inside an invalid link should not
1194 # be mistaken as delimiting cell parameters
1195 # Bug T153140: Neither should language converter markup.
1196 if ( preg_match( '/\[\[|-\{/', $cell_data[0] ) === 1 ) {
1197 $cell = "{$previous}<{$last_tag}>{$cell}";
1198 } elseif ( count( $cell_data ) == 1 ) {
1199 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
1200 } else {
1201 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1202 $attributes = Sanitizer::fixTagAttributes( $attributes, $last_tag );
1203 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
1204 }
1205
1206 $outLine .= $cell;
1207 array_push( $td_history, true );
1208 }
1209 }
1210 $out .= $outLine . "\n";
1211 }
1212
1213 # Closing open td, tr && table
1214 while ( count( $td_history ) > 0 ) {
1215 if ( array_pop( $td_history ) ) {
1216 $out .= "</td>\n";
1217 }
1218 if ( array_pop( $tr_history ) ) {
1219 $out .= "</tr>\n";
1220 }
1221 if ( !array_pop( $has_opened_tr ) ) {
1222 $out .= "<tr><td></td></tr>\n";
1223 }
1224
1225 $out .= "</table>\n";
1226 }
1227
1228 # Remove trailing line-ending (b/c)
1229 if ( substr( $out, -1 ) === "\n" ) {
1230 $out = substr( $out, 0, -1 );
1231 }
1232
1233 # special case: don't return empty table
1234 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1235 $out = '';
1236 }
1237
1238 return $out;
1239 }
1240
1241 /**
1242 * Helper function for parse() that transforms wiki markup into half-parsed
1243 * HTML. Only called for $mOutputType == self::OT_HTML.
1244 *
1245 * @private
1246 *
1247 * @param string $text The text to parse
1248 * @param bool $isMain Whether this is being called from the main parse() function
1249 * @param PPFrame|bool $frame A pre-processor frame
1250 *
1251 * @return string
1252 */
1253 public function internalParse( $text, $isMain = true, $frame = false ) {
1254 $origText = $text;
1255
1256 // Avoid PHP 7.1 warning from passing $this by reference
1257 $parser = $this;
1258
1259 # Hook to suspend the parser in this state
1260 if ( !Hooks::run( 'ParserBeforeInternalParse', [ &$parser, &$text, &$this->mStripState ] ) ) {
1261 return $text;
1262 }
1263
1264 # if $frame is provided, then use $frame for replacing any variables
1265 if ( $frame ) {
1266 # use frame depth to infer how include/noinclude tags should be handled
1267 # depth=0 means this is the top-level document; otherwise it's an included document
1268 if ( !$frame->depth ) {
1269 $flag = 0;
1270 } else {
1271 $flag = self::PTD_FOR_INCLUSION;
1272 }
1273 $dom = $this->preprocessToDom( $text, $flag );
1274 $text = $frame->expand( $dom );
1275 } else {
1276 # if $frame is not provided, then use old-style replaceVariables
1277 $text = $this->replaceVariables( $text );
1278 }
1279
1280 Hooks::run( 'InternalParseBeforeSanitize', [ &$parser, &$text, &$this->mStripState ] );
1281 $text = Sanitizer::removeHTMLtags(
1282 $text,
1283 [ $this, 'attributeStripCallback' ],
1284 false,
1285 array_keys( $this->mTransparentTagHooks ),
1286 [],
1287 [ $this, 'addTrackingCategory' ]
1288 );
1289 Hooks::run( 'InternalParseBeforeLinks', [ &$parser, &$text, &$this->mStripState ] );
1290
1291 # Tables need to come after variable replacement for things to work
1292 # properly; putting them before other transformations should keep
1293 # exciting things like link expansions from showing up in surprising
1294 # places.
1295 $text = $this->doTableStuff( $text );
1296
1297 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1298
1299 $text = $this->doDoubleUnderscore( $text );
1300
1301 $text = $this->doHeadings( $text );
1302 $text = $this->replaceInternalLinks( $text );
1303 $text = $this->doAllQuotes( $text );
1304 $text = $this->replaceExternalLinks( $text );
1305
1306 # replaceInternalLinks may sometimes leave behind
1307 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1308 $text = str_replace( self::MARKER_PREFIX . 'NOPARSE', '', $text );
1309
1310 $text = $this->doMagicLinks( $text );
1311 $text = $this->formatHeadings( $text, $origText, $isMain );
1312
1313 return $text;
1314 }
1315
1316 /**
1317 * Helper function for parse() that transforms half-parsed HTML into fully
1318 * parsed HTML.
1319 *
1320 * @param string $text
1321 * @param bool $isMain
1322 * @param bool $linestart
1323 * @return string
1324 */
1325 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1326 $text = $this->mStripState->unstripGeneral( $text );
1327
1328 // Avoid PHP 7.1 warning from passing $this by reference
1329 $parser = $this;
1330
1331 if ( $isMain ) {
1332 Hooks::run( 'ParserAfterUnstrip', [ &$parser, &$text ] );
1333 }
1334
1335 # Clean up special characters, only run once, next-to-last before doBlockLevels
1336 $fixtags = [
1337 # French spaces, last one Guillemet-left
1338 # only if there is something before the space
1339 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;',
1340 # french spaces, Guillemet-right
1341 '/(\\302\\253) /' => '\\1&#160;',
1342 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, T13874.
1343 ];
1344 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
1345
1346 $text = $this->doBlockLevels( $text, $linestart );
1347
1348 $this->replaceLinkHolders( $text );
1349
1350 /**
1351 * The input doesn't get language converted if
1352 * a) It's disabled
1353 * b) Content isn't converted
1354 * c) It's a conversion table
1355 * d) it is an interface message (which is in the user language)
1356 */
1357 if ( !( $this->mOptions->getDisableContentConversion()
1358 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) )
1359 ) {
1360 if ( !$this->mOptions->getInterfaceMessage() ) {
1361 # The position of the convert() call should not be changed. it
1362 # assumes that the links are all replaced and the only thing left
1363 # is the <nowiki> mark.
1364 $text = $this->getConverterLanguage()->convert( $text );
1365 }
1366 }
1367
1368 $text = $this->mStripState->unstripNoWiki( $text );
1369
1370 if ( $isMain ) {
1371 Hooks::run( 'ParserBeforeTidy', [ &$parser, &$text ] );
1372 }
1373
1374 $text = $this->replaceTransparentTags( $text );
1375 $text = $this->mStripState->unstripGeneral( $text );
1376
1377 $text = Sanitizer::normalizeCharReferences( $text );
1378
1379 if ( MWTidy::isEnabled() ) {
1380 if ( $this->mOptions->getTidy() ) {
1381 $text = MWTidy::tidy( $text );
1382 }
1383 } else {
1384 # attempt to sanitize at least some nesting problems
1385 # (T4702 and quite a few others)
1386 $tidyregs = [
1387 # ''Something [http://www.cool.com cool''] -->
1388 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1389 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1390 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1391 # fix up an anchor inside another anchor, only
1392 # at least for a single single nested link (T5695)
1393 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1394 '\\1\\2</a>\\3</a>\\1\\4</a>',
1395 # fix div inside inline elements- doBlockLevels won't wrap a line which
1396 # contains a div, so fix it up here; replace
1397 # div with escaped text
1398 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1399 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
1400 # remove empty italic or bold tag pairs, some
1401 # introduced by rules above
1402 '/<([bi])><\/\\1>/' => '',
1403 ];
1404
1405 $text = preg_replace(
1406 array_keys( $tidyregs ),
1407 array_values( $tidyregs ),
1408 $text );
1409 }
1410
1411 if ( $isMain ) {
1412 Hooks::run( 'ParserAfterTidy', [ &$parser, &$text ] );
1413 }
1414
1415 return $text;
1416 }
1417
1418 /**
1419 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1420 * magic external links.
1421 *
1422 * DML
1423 * @private
1424 *
1425 * @param string $text
1426 *
1427 * @return string
1428 */
1429 public function doMagicLinks( $text ) {
1430 $prots = wfUrlProtocolsWithoutProtRel();
1431 $urlChar = self::EXT_LINK_URL_CLASS;
1432 $addr = self::EXT_LINK_ADDR;
1433 $space = self::SPACE_NOT_NL; # non-newline space
1434 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1435 $spaces = "$space++"; # possessive match of 1 or more spaces
1436 $text = preg_replace_callback(
1437 '!(?: # Start cases
1438 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1439 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1440 (\b # m[3]: Free external links
1441 (?i:$prots)
1442 ($addr$urlChar*) # m[4]: Post-protocol path
1443 ) |
1444 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1445 ([0-9]+)\b |
1446 \bISBN $spaces ( # m[6]: ISBN, capture number
1447 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1448 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1449 [0-9Xx] # check digit
1450 )\b
1451 )!xu", [ $this, 'magicLinkCallback' ], $text );
1452 return $text;
1453 }
1454
1455 /**
1456 * @throws MWException
1457 * @param array $m
1458 * @return HTML|string
1459 */
1460 public function magicLinkCallback( $m ) {
1461 if ( isset( $m[1] ) && $m[1] !== '' ) {
1462 # Skip anchor
1463 return $m[0];
1464 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1465 # Skip HTML element
1466 return $m[0];
1467 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1468 # Free external link
1469 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1470 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1471 # RFC or PMID
1472 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1473 if ( !$this->mOptions->getMagicRFCLinks() ) {
1474 return $m[0];
1475 }
1476 $keyword = 'RFC';
1477 $urlmsg = 'rfcurl';
1478 $cssClass = 'mw-magiclink-rfc';
1479 $trackingCat = 'magiclink-tracking-rfc';
1480 $id = $m[5];
1481 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1482 if ( !$this->mOptions->getMagicPMIDLinks() ) {
1483 return $m[0];
1484 }
1485 $keyword = 'PMID';
1486 $urlmsg = 'pubmedurl';
1487 $cssClass = 'mw-magiclink-pmid';
1488 $trackingCat = 'magiclink-tracking-pmid';
1489 $id = $m[5];
1490 } else {
1491 throw new MWException( __METHOD__ . ': unrecognised match type "' .
1492 substr( $m[0], 0, 20 ) . '"' );
1493 }
1494 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1495 $this->addTrackingCategory( $trackingCat );
1496 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1497 } elseif ( isset( $m[6] ) && $m[6] !== ''
1498 && $this->mOptions->getMagicISBNLinks()
1499 ) {
1500 # ISBN
1501 $isbn = $m[6];
1502 $space = self::SPACE_NOT_NL; # non-newline space
1503 $isbn = preg_replace( "/$space/", ' ', $isbn );
1504 $num = strtr( $isbn, [
1505 '-' => '',
1506 ' ' => '',
1507 'x' => 'X',
1508 ] );
1509 $this->addTrackingCategory( 'magiclink-tracking-isbn' );
1510 return $this->getLinkRenderer()->makeKnownLink(
1511 SpecialPage::getTitleFor( 'Booksources', $num ),
1512 "ISBN $isbn",
1513 [
1514 'class' => 'internal mw-magiclink-isbn',
1515 'title' => false // suppress title attribute
1516 ]
1517 );
1518 } else {
1519 return $m[0];
1520 }
1521 }
1522
1523 /**
1524 * Make a free external link, given a user-supplied URL
1525 *
1526 * @param string $url
1527 * @param int $numPostProto
1528 * The number of characters after the protocol.
1529 * @return string HTML
1530 * @private
1531 */
1532 public function makeFreeExternalLink( $url, $numPostProto ) {
1533 $trail = '';
1534
1535 # The characters '<' and '>' (which were escaped by
1536 # removeHTMLtags()) should not be included in
1537 # URLs, per RFC 2396.
1538 # Make &nbsp; terminate a URL as well (bug T84937)
1539 $m2 = [];
1540 if ( preg_match(
1541 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1542 $url,
1543 $m2,
1544 PREG_OFFSET_CAPTURE
1545 ) ) {
1546 $trail = substr( $url, $m2[0][1] ) . $trail;
1547 $url = substr( $url, 0, $m2[0][1] );
1548 }
1549
1550 # Move trailing punctuation to $trail
1551 $sep = ',;\.:!?';
1552 # If there is no left bracket, then consider right brackets fair game too
1553 if ( strpos( $url, '(' ) === false ) {
1554 $sep .= ')';
1555 }
1556
1557 $urlRev = strrev( $url );
1558 $numSepChars = strspn( $urlRev, $sep );
1559 # Don't break a trailing HTML entity by moving the ; into $trail
1560 # This is in hot code, so use substr_compare to avoid having to
1561 # create a new string object for the comparison
1562 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1563 # more optimization: instead of running preg_match with a $
1564 # anchor, which can be slow, do the match on the reversed
1565 # string starting at the desired offset.
1566 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1567 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1568 $numSepChars--;
1569 }
1570 }
1571 if ( $numSepChars ) {
1572 $trail = substr( $url, -$numSepChars ) . $trail;
1573 $url = substr( $url, 0, -$numSepChars );
1574 }
1575
1576 # Verify that we still have a real URL after trail removal, and
1577 # not just lone protocol
1578 if ( strlen( $trail ) >= $numPostProto ) {
1579 return $url . $trail;
1580 }
1581
1582 $url = Sanitizer::cleanUrl( $url );
1583
1584 # Is this an external image?
1585 $text = $this->maybeMakeExternalImage( $url );
1586 if ( $text === false ) {
1587 # Not an image, make a link
1588 $text = Linker::makeExternalLink( $url,
1589 $this->getConverterLanguage()->markNoConversion( $url, true ),
1590 true, 'free',
1591 $this->getExternalLinkAttribs( $url ), $this->mTitle );
1592 # Register it in the output object...
1593 $this->mOutput->addExternalLink( $url );
1594 }
1595 return $text . $trail;
1596 }
1597
1598 /**
1599 * Parse headers and return html
1600 *
1601 * @private
1602 *
1603 * @param string $text
1604 *
1605 * @return string
1606 */
1607 public function doHeadings( $text ) {
1608 for ( $i = 6; $i >= 1; --$i ) {
1609 $h = str_repeat( '=', $i );
1610 $text = preg_replace( "/^$h(.+)$h\\s*$/m", "<h$i>\\1</h$i>", $text );
1611 }
1612 return $text;
1613 }
1614
1615 /**
1616 * Replace single quotes with HTML markup
1617 * @private
1618 *
1619 * @param string $text
1620 *
1621 * @return string The altered text
1622 */
1623 public function doAllQuotes( $text ) {
1624 $outtext = '';
1625 $lines = StringUtils::explode( "\n", $text );
1626 foreach ( $lines as $line ) {
1627 $outtext .= $this->doQuotes( $line ) . "\n";
1628 }
1629 $outtext = substr( $outtext, 0, -1 );
1630 return $outtext;
1631 }
1632
1633 /**
1634 * Helper function for doAllQuotes()
1635 *
1636 * @param string $text
1637 *
1638 * @return string
1639 */
1640 public function doQuotes( $text ) {
1641 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1642 $countarr = count( $arr );
1643 if ( $countarr == 1 ) {
1644 return $text;
1645 }
1646
1647 // First, do some preliminary work. This may shift some apostrophes from
1648 // being mark-up to being text. It also counts the number of occurrences
1649 // of bold and italics mark-ups.
1650 $numbold = 0;
1651 $numitalics = 0;
1652 for ( $i = 1; $i < $countarr; $i += 2 ) {
1653 $thislen = strlen( $arr[$i] );
1654 // If there are ever four apostrophes, assume the first is supposed to
1655 // be text, and the remaining three constitute mark-up for bold text.
1656 // (T15227: ''''foo'''' turns into ' ''' foo ' ''')
1657 if ( $thislen == 4 ) {
1658 $arr[$i - 1] .= "'";
1659 $arr[$i] = "'''";
1660 $thislen = 3;
1661 } elseif ( $thislen > 5 ) {
1662 // If there are more than 5 apostrophes in a row, assume they're all
1663 // text except for the last 5.
1664 // (T15227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1665 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1666 $arr[$i] = "'''''";
1667 $thislen = 5;
1668 }
1669 // Count the number of occurrences of bold and italics mark-ups.
1670 if ( $thislen == 2 ) {
1671 $numitalics++;
1672 } elseif ( $thislen == 3 ) {
1673 $numbold++;
1674 } elseif ( $thislen == 5 ) {
1675 $numitalics++;
1676 $numbold++;
1677 }
1678 }
1679
1680 // If there is an odd number of both bold and italics, it is likely
1681 // that one of the bold ones was meant to be an apostrophe followed
1682 // by italics. Which one we cannot know for certain, but it is more
1683 // likely to be one that has a single-letter word before it.
1684 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1685 $firstsingleletterword = -1;
1686 $firstmultiletterword = -1;
1687 $firstspace = -1;
1688 for ( $i = 1; $i < $countarr; $i += 2 ) {
1689 if ( strlen( $arr[$i] ) == 3 ) {
1690 $x1 = substr( $arr[$i - 1], -1 );
1691 $x2 = substr( $arr[$i - 1], -2, 1 );
1692 if ( $x1 === ' ' ) {
1693 if ( $firstspace == -1 ) {
1694 $firstspace = $i;
1695 }
1696 } elseif ( $x2 === ' ' ) {
1697 $firstsingleletterword = $i;
1698 // if $firstsingleletterword is set, we don't
1699 // look at the other options, so we can bail early.
1700 break;
1701 } else {
1702 if ( $firstmultiletterword == -1 ) {
1703 $firstmultiletterword = $i;
1704 }
1705 }
1706 }
1707 }
1708
1709 // If there is a single-letter word, use it!
1710 if ( $firstsingleletterword > -1 ) {
1711 $arr[$firstsingleletterword] = "''";
1712 $arr[$firstsingleletterword - 1] .= "'";
1713 } elseif ( $firstmultiletterword > -1 ) {
1714 // If not, but there's a multi-letter word, use that one.
1715 $arr[$firstmultiletterword] = "''";
1716 $arr[$firstmultiletterword - 1] .= "'";
1717 } elseif ( $firstspace > -1 ) {
1718 // ... otherwise use the first one that has neither.
1719 // (notice that it is possible for all three to be -1 if, for example,
1720 // there is only one pentuple-apostrophe in the line)
1721 $arr[$firstspace] = "''";
1722 $arr[$firstspace - 1] .= "'";
1723 }
1724 }
1725
1726 // Now let's actually convert our apostrophic mush to HTML!
1727 $output = '';
1728 $buffer = '';
1729 $state = '';
1730 $i = 0;
1731 foreach ( $arr as $r ) {
1732 if ( ( $i % 2 ) == 0 ) {
1733 if ( $state === 'both' ) {
1734 $buffer .= $r;
1735 } else {
1736 $output .= $r;
1737 }
1738 } else {
1739 $thislen = strlen( $r );
1740 if ( $thislen == 2 ) {
1741 if ( $state === 'i' ) {
1742 $output .= '</i>';
1743 $state = '';
1744 } elseif ( $state === 'bi' ) {
1745 $output .= '</i>';
1746 $state = 'b';
1747 } elseif ( $state === 'ib' ) {
1748 $output .= '</b></i><b>';
1749 $state = 'b';
1750 } elseif ( $state === 'both' ) {
1751 $output .= '<b><i>' . $buffer . '</i>';
1752 $state = 'b';
1753 } else { // $state can be 'b' or ''
1754 $output .= '<i>';
1755 $state .= 'i';
1756 }
1757 } elseif ( $thislen == 3 ) {
1758 if ( $state === 'b' ) {
1759 $output .= '</b>';
1760 $state = '';
1761 } elseif ( $state === 'bi' ) {
1762 $output .= '</i></b><i>';
1763 $state = 'i';
1764 } elseif ( $state === 'ib' ) {
1765 $output .= '</b>';
1766 $state = 'i';
1767 } elseif ( $state === 'both' ) {
1768 $output .= '<i><b>' . $buffer . '</b>';
1769 $state = 'i';
1770 } else { // $state can be 'i' or ''
1771 $output .= '<b>';
1772 $state .= 'b';
1773 }
1774 } elseif ( $thislen == 5 ) {
1775 if ( $state === 'b' ) {
1776 $output .= '</b><i>';
1777 $state = 'i';
1778 } elseif ( $state === 'i' ) {
1779 $output .= '</i><b>';
1780 $state = 'b';
1781 } elseif ( $state === 'bi' ) {
1782 $output .= '</i></b>';
1783 $state = '';
1784 } elseif ( $state === 'ib' ) {
1785 $output .= '</b></i>';
1786 $state = '';
1787 } elseif ( $state === 'both' ) {
1788 $output .= '<i><b>' . $buffer . '</b></i>';
1789 $state = '';
1790 } else { // ($state == '')
1791 $buffer = '';
1792 $state = 'both';
1793 }
1794 }
1795 }
1796 $i++;
1797 }
1798 // Now close all remaining tags. Notice that the order is important.
1799 if ( $state === 'b' || $state === 'ib' ) {
1800 $output .= '</b>';
1801 }
1802 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1803 $output .= '</i>';
1804 }
1805 if ( $state === 'bi' ) {
1806 $output .= '</b>';
1807 }
1808 // There might be lonely ''''', so make sure we have a buffer
1809 if ( $state === 'both' && $buffer ) {
1810 $output .= '<b><i>' . $buffer . '</i></b>';
1811 }
1812 return $output;
1813 }
1814
1815 /**
1816 * Replace external links (REL)
1817 *
1818 * Note: this is all very hackish and the order of execution matters a lot.
1819 * Make sure to run tests/parser/parserTests.php if you change this code.
1820 *
1821 * @private
1822 *
1823 * @param string $text
1824 *
1825 * @throws MWException
1826 * @return string
1827 */
1828 public function replaceExternalLinks( $text ) {
1829 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1830 if ( $bits === false ) {
1831 throw new MWException( "PCRE needs to be compiled with "
1832 . "--enable-unicode-properties in order for MediaWiki to function" );
1833 }
1834 $s = array_shift( $bits );
1835
1836 $i = 0;
1837 while ( $i < count( $bits ) ) {
1838 $url = $bits[$i++];
1839 $i++; // protocol
1840 $text = $bits[$i++];
1841 $trail = $bits[$i++];
1842
1843 # The characters '<' and '>' (which were escaped by
1844 # removeHTMLtags()) should not be included in
1845 # URLs, per RFC 2396.
1846 $m2 = [];
1847 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1848 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1849 $url = substr( $url, 0, $m2[0][1] );
1850 }
1851
1852 # If the link text is an image URL, replace it with an <img> tag
1853 # This happened by accident in the original parser, but some people used it extensively
1854 $img = $this->maybeMakeExternalImage( $text );
1855 if ( $img !== false ) {
1856 $text = $img;
1857 }
1858
1859 $dtrail = '';
1860
1861 # Set linktype for CSS - if URL==text, link is essentially free
1862 $linktype = ( $text === $url ) ? 'free' : 'text';
1863
1864 # No link text, e.g. [http://domain.tld/some.link]
1865 if ( $text == '' ) {
1866 # Autonumber
1867 $langObj = $this->getTargetLanguage();
1868 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1869 $linktype = 'autonumber';
1870 } else {
1871 # Have link text, e.g. [http://domain.tld/some.link text]s
1872 # Check for trail
1873 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1874 }
1875
1876 $text = $this->getConverterLanguage()->markNoConversion( $text );
1877
1878 $url = Sanitizer::cleanUrl( $url );
1879
1880 # Use the encoded URL
1881 # This means that users can paste URLs directly into the text
1882 # Funny characters like ö aren't valid in URLs anyway
1883 # This was changed in August 2004
1884 $s .= Linker::makeExternalLink( $url, $text, false, $linktype,
1885 $this->getExternalLinkAttribs( $url ), $this->mTitle ) . $dtrail . $trail;
1886
1887 # Register link in the output object.
1888 $this->mOutput->addExternalLink( $url );
1889 }
1890
1891 return $s;
1892 }
1893
1894 /**
1895 * Get the rel attribute for a particular external link.
1896 *
1897 * @since 1.21
1898 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1899 * nofollow if appropriate
1900 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1901 * @return string|null Rel attribute for $url
1902 */
1903 public static function getExternalLinkRel( $url = false, $title = null ) {
1904 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1905 $ns = $title ? $title->getNamespace() : false;
1906 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1907 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1908 ) {
1909 return 'nofollow';
1910 }
1911 return null;
1912 }
1913
1914 /**
1915 * Get an associative array of additional HTML attributes appropriate for a
1916 * particular external link. This currently may include rel => nofollow
1917 * (depending on configuration, namespace, and the URL's domain) and/or a
1918 * target attribute (depending on configuration).
1919 *
1920 * @param string $url URL to extract the domain from for rel =>
1921 * nofollow if appropriate
1922 * @return array Associative array of HTML attributes
1923 */
1924 public function getExternalLinkAttribs( $url ) {
1925 $attribs = [];
1926 $rel = self::getExternalLinkRel( $url, $this->mTitle );
1927
1928 $target = $this->mOptions->getExternalLinkTarget();
1929 if ( $target ) {
1930 $attribs['target'] = $target;
1931 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1932 // T133507. New windows can navigate parent cross-origin.
1933 // Including noreferrer due to lacking browser
1934 // support of noopener. Eventually noreferrer should be removed.
1935 if ( $rel !== '' ) {
1936 $rel .= ' ';
1937 }
1938 $rel .= 'noreferrer noopener';
1939 }
1940 }
1941 $attribs['rel'] = $rel;
1942 return $attribs;
1943 }
1944
1945 /**
1946 * Replace unusual escape codes in a URL with their equivalent characters
1947 *
1948 * This generally follows the syntax defined in RFC 3986, with special
1949 * consideration for HTTP query strings.
1950 *
1951 * @param string $url
1952 * @return string
1953 */
1954 public static function normalizeLinkUrl( $url ) {
1955 # First, make sure unsafe characters are encoded
1956 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
1957 function ( $m ) {
1958 return rawurlencode( $m[0] );
1959 },
1960 $url
1961 );
1962
1963 $ret = '';
1964 $end = strlen( $url );
1965
1966 # Fragment part - 'fragment'
1967 $start = strpos( $url, '#' );
1968 if ( $start !== false && $start < $end ) {
1969 $ret = self::normalizeUrlComponent(
1970 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
1971 $end = $start;
1972 }
1973
1974 # Query part - 'query' minus &=+;
1975 $start = strpos( $url, '?' );
1976 if ( $start !== false && $start < $end ) {
1977 $ret = self::normalizeUrlComponent(
1978 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
1979 $end = $start;
1980 }
1981
1982 # Scheme and path part - 'pchar'
1983 # (we assume no userinfo or encoded colons in the host)
1984 $ret = self::normalizeUrlComponent(
1985 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
1986
1987 return $ret;
1988 }
1989
1990 private static function normalizeUrlComponent( $component, $unsafe ) {
1991 $callback = function ( $matches ) use ( $unsafe ) {
1992 $char = urldecode( $matches[0] );
1993 $ord = ord( $char );
1994 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
1995 # Unescape it
1996 return $char;
1997 } else {
1998 # Leave it escaped, but use uppercase for a-f
1999 return strtoupper( $matches[0] );
2000 }
2001 };
2002 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
2003 }
2004
2005 /**
2006 * make an image if it's allowed, either through the global
2007 * option, through the exception, or through the on-wiki whitelist
2008 *
2009 * @param string $url
2010 *
2011 * @return string
2012 */
2013 private function maybeMakeExternalImage( $url ) {
2014 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
2015 $imagesexception = !empty( $imagesfrom );
2016 $text = false;
2017 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2018 if ( $imagesexception && is_array( $imagesfrom ) ) {
2019 $imagematch = false;
2020 foreach ( $imagesfrom as $match ) {
2021 if ( strpos( $url, $match ) === 0 ) {
2022 $imagematch = true;
2023 break;
2024 }
2025 }
2026 } elseif ( $imagesexception ) {
2027 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2028 } else {
2029 $imagematch = false;
2030 }
2031
2032 if ( $this->mOptions->getAllowExternalImages()
2033 || ( $imagesexception && $imagematch )
2034 ) {
2035 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
2036 # Image found
2037 $text = Linker::makeExternalImage( $url );
2038 }
2039 }
2040 if ( !$text && $this->mOptions->getEnableImageWhitelist()
2041 && preg_match( self::EXT_IMAGE_REGEX, $url )
2042 ) {
2043 $whitelist = explode(
2044 "\n",
2045 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2046 );
2047
2048 foreach ( $whitelist as $entry ) {
2049 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2050 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
2051 continue;
2052 }
2053 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2054 # Image matches a whitelist entry
2055 $text = Linker::makeExternalImage( $url );
2056 break;
2057 }
2058 }
2059 }
2060 return $text;
2061 }
2062
2063 /**
2064 * Process [[ ]] wikilinks
2065 *
2066 * @param string $s
2067 *
2068 * @return string Processed text
2069 *
2070 * @private
2071 */
2072 public function replaceInternalLinks( $s ) {
2073 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
2074 return $s;
2075 }
2076
2077 /**
2078 * Process [[ ]] wikilinks (RIL)
2079 * @param string &$s
2080 * @throws MWException
2081 * @return LinkHolderArray
2082 *
2083 * @private
2084 */
2085 public function replaceInternalLinks2( &$s ) {
2086 global $wgExtraInterlanguageLinkPrefixes;
2087
2088 static $tc = false, $e1, $e1_img;
2089 # the % is needed to support urlencoded titles as well
2090 if ( !$tc ) {
2091 $tc = Title::legalChars() . '#%';
2092 # Match a link having the form [[namespace:link|alternate]]trail
2093 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2094 # Match cases where there is no "]]", which might still be images
2095 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2096 }
2097
2098 $holders = new LinkHolderArray( $this );
2099
2100 # split the entire text string on occurrences of [[
2101 $a = StringUtils::explode( '[[', ' ' . $s );
2102 # get the first element (all text up to first [[), and remove the space we added
2103 $s = $a->current();
2104 $a->next();
2105 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2106 $s = substr( $s, 1 );
2107
2108 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2109 $e2 = null;
2110 if ( $useLinkPrefixExtension ) {
2111 # Match the end of a line for a word that's not followed by whitespace,
2112 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2113 global $wgContLang;
2114 $charset = $wgContLang->linkPrefixCharset();
2115 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2116 }
2117
2118 if ( is_null( $this->mTitle ) ) {
2119 throw new MWException( __METHOD__ . ": \$this->mTitle is null\n" );
2120 }
2121 $nottalk = !$this->mTitle->isTalkPage();
2122
2123 if ( $useLinkPrefixExtension ) {
2124 $m = [];
2125 if ( preg_match( $e2, $s, $m ) ) {
2126 $first_prefix = $m[2];
2127 } else {
2128 $first_prefix = false;
2129 }
2130 } else {
2131 $prefix = '';
2132 }
2133
2134 $useSubpages = $this->areSubpagesAllowed();
2135
2136 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2137 # Loop for each link
2138 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2139 // @codingStandardsIgnoreEnd
2140
2141 # Check for excessive memory usage
2142 if ( $holders->isBig() ) {
2143 # Too big
2144 # Do the existence check, replace the link holders and clear the array
2145 $holders->replace( $s );
2146 $holders->clear();
2147 }
2148
2149 if ( $useLinkPrefixExtension ) {
2150 if ( preg_match( $e2, $s, $m ) ) {
2151 $prefix = $m[2];
2152 $s = $m[1];
2153 } else {
2154 $prefix = '';
2155 }
2156 # first link
2157 if ( $first_prefix ) {
2158 $prefix = $first_prefix;
2159 $first_prefix = false;
2160 }
2161 }
2162
2163 $might_be_img = false;
2164
2165 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2166 $text = $m[2];
2167 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2168 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2169 # the real problem is with the $e1 regex
2170 # See T1500.
2171 # Still some problems for cases where the ] is meant to be outside punctuation,
2172 # and no image is in sight. See T4095.
2173 if ( $text !== ''
2174 && substr( $m[3], 0, 1 ) === ']'
2175 && strpos( $text, '[' ) !== false
2176 ) {
2177 $text .= ']'; # so that replaceExternalLinks($text) works later
2178 $m[3] = substr( $m[3], 1 );
2179 }
2180 # fix up urlencoded title texts
2181 if ( strpos( $m[1], '%' ) !== false ) {
2182 # Should anchors '#' also be rejected?
2183 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2184 }
2185 $trail = $m[3];
2186 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2187 # Invalid, but might be an image with a link in its caption
2188 $might_be_img = true;
2189 $text = $m[2];
2190 if ( strpos( $m[1], '%' ) !== false ) {
2191 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2192 }
2193 $trail = "";
2194 } else { # Invalid form; output directly
2195 $s .= $prefix . '[[' . $line;
2196 continue;
2197 }
2198
2199 $origLink = ltrim( $m[1], ' ' );
2200
2201 # Don't allow internal links to pages containing
2202 # PROTO: where PROTO is a valid URL protocol; these
2203 # should be external links.
2204 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $origLink ) ) {
2205 $s .= $prefix . '[[' . $line;
2206 continue;
2207 }
2208
2209 # Make subpage if necessary
2210 if ( $useSubpages ) {
2211 $link = $this->maybeDoSubpageLink( $origLink, $text );
2212 } else {
2213 $link = $origLink;
2214 }
2215
2216 $unstrip = $this->mStripState->unstripNoWiki( $link );
2217 $nt = is_string( $unstrip ) ? Title::newFromText( $unstrip ) : null;
2218 if ( $nt === null ) {
2219 $s .= $prefix . '[[' . $line;
2220 continue;
2221 }
2222
2223 $ns = $nt->getNamespace();
2224 $iw = $nt->getInterwiki();
2225
2226 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2227
2228 if ( $might_be_img ) { # if this is actually an invalid link
2229 if ( $ns == NS_FILE && $noforce ) { # but might be an image
2230 $found = false;
2231 while ( true ) {
2232 # look at the next 'line' to see if we can close it there
2233 $a->next();
2234 $next_line = $a->current();
2235 if ( $next_line === false || $next_line === null ) {
2236 break;
2237 }
2238 $m = explode( ']]', $next_line, 3 );
2239 if ( count( $m ) == 3 ) {
2240 # the first ]] closes the inner link, the second the image
2241 $found = true;
2242 $text .= "[[{$m[0]}]]{$m[1]}";
2243 $trail = $m[2];
2244 break;
2245 } elseif ( count( $m ) == 2 ) {
2246 # if there's exactly one ]] that's fine, we'll keep looking
2247 $text .= "[[{$m[0]}]]{$m[1]}";
2248 } else {
2249 # if $next_line is invalid too, we need look no further
2250 $text .= '[[' . $next_line;
2251 break;
2252 }
2253 }
2254 if ( !$found ) {
2255 # we couldn't find the end of this imageLink, so output it raw
2256 # but don't ignore what might be perfectly normal links in the text we've examined
2257 $holders->merge( $this->replaceInternalLinks2( $text ) );
2258 $s .= "{$prefix}[[$link|$text";
2259 # note: no $trail, because without an end, there *is* no trail
2260 continue;
2261 }
2262 } else { # it's not an image, so output it raw
2263 $s .= "{$prefix}[[$link|$text";
2264 # note: no $trail, because without an end, there *is* no trail
2265 continue;
2266 }
2267 }
2268
2269 $wasblank = ( $text == '' );
2270 if ( $wasblank ) {
2271 $text = $link;
2272 if ( !$noforce ) {
2273 # Strip off leading ':'
2274 $text = substr( $text, 1 );
2275 }
2276 } else {
2277 # T6598 madness. Handle the quotes only if they come from the alternate part
2278 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2279 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2280 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2281 $text = $this->doQuotes( $text );
2282 }
2283
2284 # Link not escaped by : , create the various objects
2285 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2286 # Interwikis
2287 if (
2288 $iw && $this->mOptions->getInterwikiMagic() && $nottalk && (
2289 Language::fetchLanguageName( $iw, null, 'mw' ) ||
2290 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2291 )
2292 ) {
2293 # T26502: filter duplicates
2294 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
2295 $this->mLangLinkLanguages[$iw] = true;
2296 $this->mOutput->addLanguageLink( $nt->getFullText() );
2297 }
2298
2299 $s = rtrim( $s . $prefix );
2300 $s .= trim( $trail, "\n" ) == '' ? '' : $prefix . $trail;
2301 continue;
2302 }
2303
2304 if ( $ns == NS_FILE ) {
2305 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
2306 if ( $wasblank ) {
2307 # if no parameters were passed, $text
2308 # becomes something like "File:Foo.png",
2309 # which we don't want to pass on to the
2310 # image generator
2311 $text = '';
2312 } else {
2313 # recursively parse links inside the image caption
2314 # actually, this will parse them in any other parameters, too,
2315 # but it might be hard to fix that, and it doesn't matter ATM
2316 $text = $this->replaceExternalLinks( $text );
2317 $holders->merge( $this->replaceInternalLinks2( $text ) );
2318 }
2319 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2320 $s .= $prefix . $this->armorLinks(
2321 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2322 continue;
2323 }
2324 } elseif ( $ns == NS_CATEGORY ) {
2325 $s = rtrim( $s . "\n" ); # T2087
2326
2327 if ( $wasblank ) {
2328 $sortkey = $this->getDefaultSort();
2329 } else {
2330 $sortkey = $text;
2331 }
2332 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
2333 $sortkey = str_replace( "\n", '', $sortkey );
2334 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
2335 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
2336
2337 /**
2338 * Strip the whitespace Category links produce, see T2087
2339 */
2340 $s .= trim( $prefix . $trail, "\n" ) == '' ? '' : $prefix . $trail;
2341
2342 continue;
2343 }
2344 }
2345
2346 # Self-link checking. For some languages, variants of the title are checked in
2347 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2348 # for linking to a different variant.
2349 if ( $ns != NS_SPECIAL && $nt->equals( $this->mTitle ) && !$nt->hasFragment() ) {
2350 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail );
2351 continue;
2352 }
2353
2354 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2355 # @todo FIXME: Should do batch file existence checks, see comment below
2356 if ( $ns == NS_MEDIA ) {
2357 # Give extensions a chance to select the file revision for us
2358 $options = [];
2359 $descQuery = false;
2360 Hooks::run( 'BeforeParserFetchFileAndTitle',
2361 [ $this, $nt, &$options, &$descQuery ] );
2362 # Fetch and register the file (file title may be different via hooks)
2363 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2364 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2365 $s .= $prefix . $this->armorLinks(
2366 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2367 continue;
2368 }
2369
2370 # Some titles, such as valid special pages or files in foreign repos, should
2371 # be shown as bluelinks even though they're not included in the page table
2372 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2373 # batch file existence checks for NS_FILE and NS_MEDIA
2374 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2375 $this->mOutput->addLink( $nt );
2376 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2377 } else {
2378 # Links will be added to the output link list after checking
2379 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2380 }
2381 }
2382 return $holders;
2383 }
2384
2385 /**
2386 * Render a forced-blue link inline; protect against double expansion of
2387 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2388 * Since this little disaster has to split off the trail text to avoid
2389 * breaking URLs in the following text without breaking trails on the
2390 * wiki links, it's been made into a horrible function.
2391 *
2392 * @param Title $nt
2393 * @param string $text
2394 * @param string $trail
2395 * @param string $prefix
2396 * @return string HTML-wikitext mix oh yuck
2397 */
2398 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2399 list( $inside, $trail ) = Linker::splitTrail( $trail );
2400
2401 if ( $text == '' ) {
2402 $text = htmlspecialchars( $nt->getPrefixedText() );
2403 }
2404
2405 $link = $this->getLinkRenderer()->makeKnownLink(
2406 $nt, new HtmlArmor( "$prefix$text$inside" )
2407 );
2408
2409 return $this->armorLinks( $link ) . $trail;
2410 }
2411
2412 /**
2413 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2414 * going to go through further parsing steps before inline URL expansion.
2415 *
2416 * Not needed quite as much as it used to be since free links are a bit
2417 * more sensible these days. But bracketed links are still an issue.
2418 *
2419 * @param string $text More-or-less HTML
2420 * @return string Less-or-more HTML with NOPARSE bits
2421 */
2422 public function armorLinks( $text ) {
2423 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/',
2424 self::MARKER_PREFIX . "NOPARSE$1", $text );
2425 }
2426
2427 /**
2428 * Return true if subpage links should be expanded on this page.
2429 * @return bool
2430 */
2431 public function areSubpagesAllowed() {
2432 # Some namespaces don't allow subpages
2433 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2434 }
2435
2436 /**
2437 * Handle link to subpage if necessary
2438 *
2439 * @param string $target The source of the link
2440 * @param string &$text The link text, modified as necessary
2441 * @return string The full name of the link
2442 * @private
2443 */
2444 public function maybeDoSubpageLink( $target, &$text ) {
2445 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2446 }
2447
2448 /**
2449 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2450 *
2451 * @param string $text
2452 * @param bool $linestart Whether or not this is at the start of a line.
2453 * @private
2454 * @return string The lists rendered as HTML
2455 */
2456 public function doBlockLevels( $text, $linestart ) {
2457 return BlockLevelPass::doBlockLevels( $text, $linestart );
2458 }
2459
2460 /**
2461 * Return value of a magic variable (like PAGENAME)
2462 *
2463 * @private
2464 *
2465 * @param string $index Magic variable identifier as mapped in MagicWord::$mVariableIDs
2466 * @param bool|PPFrame $frame
2467 *
2468 * @throws MWException
2469 * @return string
2470 */
2471 public function getVariableValue( $index, $frame = false ) {
2472 global $wgContLang, $wgSitename, $wgServer, $wgServerName;
2473 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2474
2475 if ( is_null( $this->mTitle ) ) {
2476 // If no title set, bad things are going to happen
2477 // later. Title should always be set since this
2478 // should only be called in the middle of a parse
2479 // operation (but the unit-tests do funky stuff)
2480 throw new MWException( __METHOD__ . ' Should only be '
2481 . ' called while parsing (no title set)' );
2482 }
2483
2484 // Avoid PHP 7.1 warning from passing $this by reference
2485 $parser = $this;
2486
2487 /**
2488 * Some of these require message or data lookups and can be
2489 * expensive to check many times.
2490 */
2491 if ( Hooks::run( 'ParserGetVariableValueVarCache', [ &$parser, &$this->mVarCache ] ) ) {
2492 if ( isset( $this->mVarCache[$index] ) ) {
2493 return $this->mVarCache[$index];
2494 }
2495 }
2496
2497 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2498 Hooks::run( 'ParserGetVariableValueTs', [ &$parser, &$ts ] );
2499
2500 $pageLang = $this->getFunctionLang();
2501
2502 switch ( $index ) {
2503 case '!':
2504 $value = '|';
2505 break;
2506 case 'currentmonth':
2507 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'm' ) );
2508 break;
2509 case 'currentmonth1':
2510 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2511 break;
2512 case 'currentmonthname':
2513 $value = $pageLang->getMonthName( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2514 break;
2515 case 'currentmonthnamegen':
2516 $value = $pageLang->getMonthNameGen( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2517 break;
2518 case 'currentmonthabbrev':
2519 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2520 break;
2521 case 'currentday':
2522 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'j' ) );
2523 break;
2524 case 'currentday2':
2525 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'd' ) );
2526 break;
2527 case 'localmonth':
2528 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'm' ) );
2529 break;
2530 case 'localmonth1':
2531 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2532 break;
2533 case 'localmonthname':
2534 $value = $pageLang->getMonthName( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2535 break;
2536 case 'localmonthnamegen':
2537 $value = $pageLang->getMonthNameGen( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2538 break;
2539 case 'localmonthabbrev':
2540 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2541 break;
2542 case 'localday':
2543 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'j' ) );
2544 break;
2545 case 'localday2':
2546 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'd' ) );
2547 break;
2548 case 'pagename':
2549 $value = wfEscapeWikiText( $this->mTitle->getText() );
2550 break;
2551 case 'pagenamee':
2552 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() );
2553 break;
2554 case 'fullpagename':
2555 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2556 break;
2557 case 'fullpagenamee':
2558 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() );
2559 break;
2560 case 'subpagename':
2561 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2562 break;
2563 case 'subpagenamee':
2564 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() );
2565 break;
2566 case 'rootpagename':
2567 $value = wfEscapeWikiText( $this->mTitle->getRootText() );
2568 break;
2569 case 'rootpagenamee':
2570 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2571 ' ',
2572 '_',
2573 $this->mTitle->getRootText()
2574 ) ) );
2575 break;
2576 case 'basepagename':
2577 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2578 break;
2579 case 'basepagenamee':
2580 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2581 ' ',
2582 '_',
2583 $this->mTitle->getBaseText()
2584 ) ) );
2585 break;
2586 case 'talkpagename':
2587 if ( $this->mTitle->canHaveTalkPage() ) {
2588 $talkPage = $this->mTitle->getTalkPage();
2589 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2590 } else {
2591 $value = '';
2592 }
2593 break;
2594 case 'talkpagenamee':
2595 if ( $this->mTitle->canHaveTalkPage() ) {
2596 $talkPage = $this->mTitle->getTalkPage();
2597 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2598 } else {
2599 $value = '';
2600 }
2601 break;
2602 case 'subjectpagename':
2603 $subjPage = $this->mTitle->getSubjectPage();
2604 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2605 break;
2606 case 'subjectpagenamee':
2607 $subjPage = $this->mTitle->getSubjectPage();
2608 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2609 break;
2610 case 'pageid': // requested in T25427
2611 $pageid = $this->getTitle()->getArticleID();
2612 if ( $pageid == 0 ) {
2613 # 0 means the page doesn't exist in the database,
2614 # which means the user is previewing a new page.
2615 # The vary-revision flag must be set, because the magic word
2616 # will have a different value once the page is saved.
2617 $this->mOutput->setFlag( 'vary-revision' );
2618 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2619 }
2620 $value = $pageid ? $pageid : null;
2621 break;
2622 case 'revisionid':
2623 # Let the edit saving system know we should parse the page
2624 # *after* a revision ID has been assigned.
2625 $this->mOutput->setFlag( 'vary-revision-id' );
2626 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision-id...\n" );
2627 $value = $this->mRevisionId;
2628 if ( !$value && $this->mOptions->getSpeculativeRevIdCallback() ) {
2629 $value = call_user_func( $this->mOptions->getSpeculativeRevIdCallback() );
2630 $this->mOutput->setSpeculativeRevIdUsed( $value );
2631 }
2632 break;
2633 case 'revisionday':
2634 # Let the edit saving system know we should parse the page
2635 # *after* a revision ID has been assigned. This is for null edits.
2636 $this->mOutput->setFlag( 'vary-revision' );
2637 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2638 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2639 break;
2640 case 'revisionday2':
2641 # Let the edit saving system know we should parse the page
2642 # *after* a revision ID has been assigned. This is for null edits.
2643 $this->mOutput->setFlag( 'vary-revision' );
2644 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2645 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2646 break;
2647 case 'revisionmonth':
2648 # Let the edit saving system know we should parse the page
2649 # *after* a revision ID has been assigned. This is for null edits.
2650 $this->mOutput->setFlag( 'vary-revision' );
2651 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2652 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2653 break;
2654 case 'revisionmonth1':
2655 # Let the edit saving system know we should parse the page
2656 # *after* a revision ID has been assigned. This is for null edits.
2657 $this->mOutput->setFlag( 'vary-revision' );
2658 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2659 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2660 break;
2661 case 'revisionyear':
2662 # Let the edit saving system know we should parse the page
2663 # *after* a revision ID has been assigned. This is for null edits.
2664 $this->mOutput->setFlag( 'vary-revision' );
2665 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2666 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2667 break;
2668 case 'revisiontimestamp':
2669 # Let the edit saving system know we should parse the page
2670 # *after* a revision ID has been assigned. This is for null edits.
2671 $this->mOutput->setFlag( 'vary-revision' );
2672 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2673 $value = $this->getRevisionTimestamp();
2674 break;
2675 case 'revisionuser':
2676 # Let the edit saving system know we should parse the page
2677 # *after* a revision ID has been assigned for null edits.
2678 $this->mOutput->setFlag( 'vary-user' );
2679 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-user...\n" );
2680 $value = $this->getRevisionUser();
2681 break;
2682 case 'revisionsize':
2683 $value = $this->getRevisionSize();
2684 break;
2685 case 'namespace':
2686 $value = str_replace( '_', ' ', $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2687 break;
2688 case 'namespacee':
2689 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2690 break;
2691 case 'namespacenumber':
2692 $value = $this->mTitle->getNamespace();
2693 break;
2694 case 'talkspace':
2695 $value = $this->mTitle->canHaveTalkPage()
2696 ? str_replace( '_', ' ', $this->mTitle->getTalkNsText() )
2697 : '';
2698 break;
2699 case 'talkspacee':
2700 $value = $this->mTitle->canHaveTalkPage() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2701 break;
2702 case 'subjectspace':
2703 $value = str_replace( '_', ' ', $this->mTitle->getSubjectNsText() );
2704 break;
2705 case 'subjectspacee':
2706 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2707 break;
2708 case 'currentdayname':
2709 $value = $pageLang->getWeekdayName( (int)MWTimestamp::getInstance( $ts )->format( 'w' ) + 1 );
2710 break;
2711 case 'currentyear':
2712 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'Y' ), true );
2713 break;
2714 case 'currenttime':
2715 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2716 break;
2717 case 'currenthour':
2718 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'H' ), true );
2719 break;
2720 case 'currentweek':
2721 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2722 # int to remove the padding
2723 $value = $pageLang->formatNum( (int)MWTimestamp::getInstance( $ts )->format( 'W' ) );
2724 break;
2725 case 'currentdow':
2726 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'w' ) );
2727 break;
2728 case 'localdayname':
2729 $value = $pageLang->getWeekdayName(
2730 (int)MWTimestamp::getLocalInstance( $ts )->format( 'w' ) + 1
2731 );
2732 break;
2733 case 'localyear':
2734 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'Y' ), true );
2735 break;
2736 case 'localtime':
2737 $value = $pageLang->time(
2738 MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' ),
2739 false,
2740 false
2741 );
2742 break;
2743 case 'localhour':
2744 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'H' ), true );
2745 break;
2746 case 'localweek':
2747 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2748 # int to remove the padding
2749 $value = $pageLang->formatNum( (int)MWTimestamp::getLocalInstance( $ts )->format( 'W' ) );
2750 break;
2751 case 'localdow':
2752 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'w' ) );
2753 break;
2754 case 'numberofarticles':
2755 $value = $pageLang->formatNum( SiteStats::articles() );
2756 break;
2757 case 'numberoffiles':
2758 $value = $pageLang->formatNum( SiteStats::images() );
2759 break;
2760 case 'numberofusers':
2761 $value = $pageLang->formatNum( SiteStats::users() );
2762 break;
2763 case 'numberofactiveusers':
2764 $value = $pageLang->formatNum( SiteStats::activeUsers() );
2765 break;
2766 case 'numberofpages':
2767 $value = $pageLang->formatNum( SiteStats::pages() );
2768 break;
2769 case 'numberofadmins':
2770 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2771 break;
2772 case 'numberofedits':
2773 $value = $pageLang->formatNum( SiteStats::edits() );
2774 break;
2775 case 'currenttimestamp':
2776 $value = wfTimestamp( TS_MW, $ts );
2777 break;
2778 case 'localtimestamp':
2779 $value = MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' );
2780 break;
2781 case 'currentversion':
2782 $value = SpecialVersion::getVersion();
2783 break;
2784 case 'articlepath':
2785 return $wgArticlePath;
2786 case 'sitename':
2787 return $wgSitename;
2788 case 'server':
2789 return $wgServer;
2790 case 'servername':
2791 return $wgServerName;
2792 case 'scriptpath':
2793 return $wgScriptPath;
2794 case 'stylepath':
2795 return $wgStylePath;
2796 case 'directionmark':
2797 return $pageLang->getDirMark();
2798 case 'contentlanguage':
2799 global $wgLanguageCode;
2800 return $wgLanguageCode;
2801 case 'pagelanguage':
2802 $value = $pageLang->getCode();
2803 break;
2804 case 'cascadingsources':
2805 $value = CoreParserFunctions::cascadingsources( $this );
2806 break;
2807 default:
2808 $ret = null;
2809 Hooks::run(
2810 'ParserGetVariableValueSwitch',
2811 [ &$parser, &$this->mVarCache, &$index, &$ret, &$frame ]