dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
Parser: Emit deprecation warnings for ParserLimitReport hook
[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 = Parser::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 * Get a random string
741 *
742 * @return string
743 * @deprecated since 1.26; use wfRandomString() instead.
744 */
745 public static function getRandomString() {
746 wfDeprecated( __METHOD__, '1.26' );
747 return wfRandomString( 16 );
748 }
749
750 /**
751 * Set the current user.
752 * Should only be used when doing pre-save transform.
753 *
754 * @param User|null $user User object or null (to reset)
755 */
756 public function setUser( $user ) {
757 $this->mUser = $user;
758 }
759
760 /**
761 * Accessor for mUniqPrefix.
762 *
763 * @return string
764 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
765 */
766 public function uniqPrefix() {
767 wfDeprecated( __METHOD__, '1.26' );
768 return self::MARKER_PREFIX;
769 }
770
771 /**
772 * Set the context title
773 *
774 * @param Title $t
775 */
776 public function setTitle( $t ) {
777 if ( !$t ) {
778 $t = Title::newFromText( 'NO TITLE' );
779 }
780
781 if ( $t->hasFragment() ) {
782 # Strip the fragment to avoid various odd effects
783 $this->mTitle = $t->createFragmentTarget( '' );
784 } else {
785 $this->mTitle = $t;
786 }
787 }
788
789 /**
790 * Accessor for the Title object
791 *
792 * @return Title
793 */
794 public function getTitle() {
795 return $this->mTitle;
796 }
797
798 /**
799 * Accessor/mutator for the Title object
800 *
801 * @param Title $x Title object or null to just get the current one
802 * @return Title
803 */
804 public function Title( $x = null ) {
805 return wfSetVar( $this->mTitle, $x );
806 }
807
808 /**
809 * Set the output type
810 *
811 * @param int $ot New value
812 */
813 public function setOutputType( $ot ) {
814 $this->mOutputType = $ot;
815 # Shortcut alias
816 $this->ot = [
817 'html' => $ot == self::OT_HTML,
818 'wiki' => $ot == self::OT_WIKI,
819 'pre' => $ot == self::OT_PREPROCESS,
820 'plain' => $ot == self::OT_PLAIN,
821 ];
822 }
823
824 /**
825 * Accessor/mutator for the output type
826 *
827 * @param int|null $x New value or null to just get the current one
828 * @return int
829 */
830 public function OutputType( $x = null ) {
831 return wfSetVar( $this->mOutputType, $x );
832 }
833
834 /**
835 * Get the ParserOutput object
836 *
837 * @return ParserOutput
838 */
839 public function getOutput() {
840 return $this->mOutput;
841 }
842
843 /**
844 * Get the ParserOptions object
845 *
846 * @return ParserOptions
847 */
848 public function getOptions() {
849 return $this->mOptions;
850 }
851
852 /**
853 * Accessor/mutator for the ParserOptions object
854 *
855 * @param ParserOptions $x New value or null to just get the current one
856 * @return ParserOptions Current ParserOptions object
857 */
858 public function Options( $x = null ) {
859 return wfSetVar( $this->mOptions, $x );
860 }
861
862 /**
863 * @return int
864 */
865 public function nextLinkID() {
866 return $this->mLinkID++;
867 }
868
869 /**
870 * @param int $id
871 */
872 public function setLinkID( $id ) {
873 $this->mLinkID = $id;
874 }
875
876 /**
877 * Get a language object for use in parser functions such as {{FORMATNUM:}}
878 * @return Language
879 */
880 public function getFunctionLang() {
881 return $this->getTargetLanguage();
882 }
883
884 /**
885 * Get the target language for the content being parsed. This is usually the
886 * language that the content is in.
887 *
888 * @since 1.19
889 *
890 * @throws MWException
891 * @return Language
892 */
893 public function getTargetLanguage() {
894 $target = $this->mOptions->getTargetLanguage();
895
896 if ( $target !== null ) {
897 return $target;
898 } elseif ( $this->mOptions->getInterfaceMessage() ) {
899 return $this->mOptions->getUserLangObj();
900 } elseif ( is_null( $this->mTitle ) ) {
901 throw new MWException( __METHOD__ . ': $this->mTitle is null' );
902 }
903
904 return $this->mTitle->getPageLanguage();
905 }
906
907 /**
908 * Get the language object for language conversion
909 * @return Language|null
910 */
911 public function getConverterLanguage() {
912 return $this->getTargetLanguage();
913 }
914
915 /**
916 * Get a User object either from $this->mUser, if set, or from the
917 * ParserOptions object otherwise
918 *
919 * @return User
920 */
921 public function getUser() {
922 if ( !is_null( $this->mUser ) ) {
923 return $this->mUser;
924 }
925 return $this->mOptions->getUser();
926 }
927
928 /**
929 * Get a preprocessor object
930 *
931 * @return Preprocessor
932 */
933 public function getPreprocessor() {
934 if ( !isset( $this->mPreprocessor ) ) {
935 $class = $this->mPreprocessorClass;
936 $this->mPreprocessor = new $class( $this );
937 }
938 return $this->mPreprocessor;
939 }
940
941 /**
942 * Get a LinkRenderer instance to make links with
943 *
944 * @since 1.28
945 * @return LinkRenderer
946 */
947 public function getLinkRenderer() {
948 if ( !$this->mLinkRenderer ) {
949 $this->mLinkRenderer = MediaWikiServices::getInstance()
950 ->getLinkRendererFactory()->create();
951 $this->mLinkRenderer->setStubThreshold(
952 $this->getOptions()->getStubThreshold()
953 );
954 }
955
956 return $this->mLinkRenderer;
957 }
958
959 /**
960 * Replaces all occurrences of HTML-style comments and the given tags
961 * in the text with a random marker and returns the next text. The output
962 * parameter $matches will be an associative array filled with data in
963 * the form:
964 *
965 * @code
966 * 'UNIQ-xxxxx' => [
967 * 'element',
968 * 'tag content',
969 * [ 'param' => 'x' ],
970 * '<element param="x">tag content</element>' ]
971 * @endcode
972 *
973 * @param array $elements List of element names. Comments are always extracted.
974 * @param string $text Source text string.
975 * @param array $matches Out parameter, Array: extracted tags
976 * @param string|null $uniq_prefix
977 * @return string Stripped text
978 * @since 1.26 The uniq_prefix argument is deprecated.
979 */
980 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = null ) {
981 if ( $uniq_prefix !== null ) {
982 wfDeprecated( __METHOD__ . ' called with $prefix argument', '1.26' );
983 }
984 static $n = 1;
985 $stripped = '';
986 $matches = [];
987
988 $taglist = implode( '|', $elements );
989 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
990
991 while ( $text != '' ) {
992 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
993 $stripped .= $p[0];
994 if ( count( $p ) < 5 ) {
995 break;
996 }
997 if ( count( $p ) > 5 ) {
998 # comment
999 $element = $p[4];
1000 $attributes = '';
1001 $close = '';
1002 $inside = $p[5];
1003 } else {
1004 # tag
1005 $element = $p[1];
1006 $attributes = $p[2];
1007 $close = $p[3];
1008 $inside = $p[4];
1009 }
1010
1011 $marker = self::MARKER_PREFIX . "-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
1012 $stripped .= $marker;
1013
1014 if ( $close === '/>' ) {
1015 # Empty element tag, <tag />
1016 $content = null;
1017 $text = $inside;
1018 $tail = null;
1019 } else {
1020 if ( $element === '!--' ) {
1021 $end = '/(-->)/';
1022 } else {
1023 $end = "/(<\\/$element\\s*>)/i";
1024 }
1025 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
1026 $content = $q[0];
1027 if ( count( $q ) < 3 ) {
1028 # No end tag -- let it run out to the end of the text.
1029 $tail = '';
1030 $text = '';
1031 } else {
1032 $tail = $q[1];
1033 $text = $q[2];
1034 }
1035 }
1036
1037 $matches[$marker] = [ $element,
1038 $content,
1039 Sanitizer::decodeTagAttributes( $attributes ),
1040 "<$element$attributes$close$content$tail" ];
1041 }
1042 return $stripped;
1043 }
1044
1045 /**
1046 * Get a list of strippable XML-like elements
1047 *
1048 * @return array
1049 */
1050 public function getStripList() {
1051 return $this->mStripList;
1052 }
1053
1054 /**
1055 * Add an item to the strip state
1056 * Returns the unique tag which must be inserted into the stripped text
1057 * The tag will be replaced with the original text in unstrip()
1058 *
1059 * @param string $text
1060 *
1061 * @return string
1062 */
1063 public function insertStripItem( $text ) {
1064 $marker = self::MARKER_PREFIX . "-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
1065 $this->mMarkerIndex++;
1066 $this->mStripState->addGeneral( $marker, $text );
1067 return $marker;
1068 }
1069
1070 /**
1071 * parse the wiki syntax used to render tables
1072 *
1073 * @private
1074 * @param string $text
1075 * @return string
1076 */
1077 public function doTableStuff( $text ) {
1078
1079 $lines = StringUtils::explode( "\n", $text );
1080 $out = '';
1081 $td_history = []; # Is currently a td tag open?
1082 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1083 $tr_history = []; # Is currently a tr tag open?
1084 $tr_attributes = []; # history of tr attributes
1085 $has_opened_tr = []; # Did this table open a <tr> element?
1086 $indent_level = 0; # indent level of the table
1087
1088 foreach ( $lines as $outLine ) {
1089 $line = trim( $outLine );
1090
1091 if ( $line === '' ) { # empty line, go to next line
1092 $out .= $outLine . "\n";
1093 continue;
1094 }
1095
1096 $first_character = $line[0];
1097 $first_two = substr( $line, 0, 2 );
1098 $matches = [];
1099
1100 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1101 # First check if we are starting a new table
1102 $indent_level = strlen( $matches[1] );
1103
1104 $attributes = $this->mStripState->unstripBoth( $matches[2] );
1105 $attributes = Sanitizer::fixTagAttributes( $attributes, 'table' );
1106
1107 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1108 array_push( $td_history, false );
1109 array_push( $last_tag_history, '' );
1110 array_push( $tr_history, false );
1111 array_push( $tr_attributes, '' );
1112 array_push( $has_opened_tr, false );
1113 } elseif ( count( $td_history ) == 0 ) {
1114 # Don't do any of the following
1115 $out .= $outLine . "\n";
1116 continue;
1117 } elseif ( $first_two === '|}' ) {
1118 # We are ending a table
1119 $line = '</table>' . substr( $line, 2 );
1120 $last_tag = array_pop( $last_tag_history );
1121
1122 if ( !array_pop( $has_opened_tr ) ) {
1123 $line = "<tr><td></td></tr>{$line}";
1124 }
1125
1126 if ( array_pop( $tr_history ) ) {
1127 $line = "</tr>{$line}";
1128 }
1129
1130 if ( array_pop( $td_history ) ) {
1131 $line = "</{$last_tag}>{$line}";
1132 }
1133 array_pop( $tr_attributes );
1134 $outLine = $line . str_repeat( '</dd></dl>', $indent_level );
1135 } elseif ( $first_two === '|-' ) {
1136 # Now we have a table row
1137 $line = preg_replace( '#^\|-+#', '', $line );
1138
1139 # Whats after the tag is now only attributes
1140 $attributes = $this->mStripState->unstripBoth( $line );
1141 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
1142 array_pop( $tr_attributes );
1143 array_push( $tr_attributes, $attributes );
1144
1145 $line = '';
1146 $last_tag = array_pop( $last_tag_history );
1147 array_pop( $has_opened_tr );
1148 array_push( $has_opened_tr, true );
1149
1150 if ( array_pop( $tr_history ) ) {
1151 $line = '</tr>';
1152 }
1153
1154 if ( array_pop( $td_history ) ) {
1155 $line = "</{$last_tag}>{$line}";
1156 }
1157
1158 $outLine = $line;
1159 array_push( $tr_history, false );
1160 array_push( $td_history, false );
1161 array_push( $last_tag_history, '' );
1162 } elseif ( $first_character === '|'
1163 || $first_character === '!'
1164 || $first_two === '|+'
1165 ) {
1166 # This might be cell elements, td, th or captions
1167 if ( $first_two === '|+' ) {
1168 $first_character = '+';
1169 $line = substr( $line, 2 );
1170 } else {
1171 $line = substr( $line, 1 );
1172 }
1173
1174 // Implies both are valid for table headings.
1175 if ( $first_character === '!' ) {
1176 $line = StringUtils::replaceMarkup( '!!', '||', $line );
1177 }
1178
1179 # Split up multiple cells on the same line.
1180 # FIXME : This can result in improper nesting of tags processed
1181 # by earlier parser steps.
1182 $cells = explode( '||', $line );
1183
1184 $outLine = '';
1185
1186 # Loop through each table cell
1187 foreach ( $cells as $cell ) {
1188 $previous = '';
1189 if ( $first_character !== '+' ) {
1190 $tr_after = array_pop( $tr_attributes );
1191 if ( !array_pop( $tr_history ) ) {
1192 $previous = "<tr{$tr_after}>\n";
1193 }
1194 array_push( $tr_history, true );
1195 array_push( $tr_attributes, '' );
1196 array_pop( $has_opened_tr );
1197 array_push( $has_opened_tr, true );
1198 }
1199
1200 $last_tag = array_pop( $last_tag_history );
1201
1202 if ( array_pop( $td_history ) ) {
1203 $previous = "</{$last_tag}>\n{$previous}";
1204 }
1205
1206 if ( $first_character === '|' ) {
1207 $last_tag = 'td';
1208 } elseif ( $first_character === '!' ) {
1209 $last_tag = 'th';
1210 } elseif ( $first_character === '+' ) {
1211 $last_tag = 'caption';
1212 } else {
1213 $last_tag = '';
1214 }
1215
1216 array_push( $last_tag_history, $last_tag );
1217
1218 # A cell could contain both parameters and data
1219 $cell_data = explode( '|', $cell, 2 );
1220
1221 # T2553: Note that a '|' inside an invalid link should not
1222 # be mistaken as delimiting cell parameters
1223 # Bug T153140: Neither should language converter markup.
1224 if ( preg_match( '/\[\[|-\{/', $cell_data[0] ) === 1 ) {
1225 $cell = "{$previous}<{$last_tag}>{$cell}";
1226 } elseif ( count( $cell_data ) == 1 ) {
1227 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
1228 } else {
1229 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1230 $attributes = Sanitizer::fixTagAttributes( $attributes, $last_tag );
1231 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
1232 }
1233
1234 $outLine .= $cell;
1235 array_push( $td_history, true );
1236 }
1237 }
1238 $out .= $outLine . "\n";
1239 }
1240
1241 # Closing open td, tr && table
1242 while ( count( $td_history ) > 0 ) {
1243 if ( array_pop( $td_history ) ) {
1244 $out .= "</td>\n";
1245 }
1246 if ( array_pop( $tr_history ) ) {
1247 $out .= "</tr>\n";
1248 }
1249 if ( !array_pop( $has_opened_tr ) ) {
1250 $out .= "<tr><td></td></tr>\n";
1251 }
1252
1253 $out .= "</table>\n";
1254 }
1255
1256 # Remove trailing line-ending (b/c)
1257 if ( substr( $out, -1 ) === "\n" ) {
1258 $out = substr( $out, 0, -1 );
1259 }
1260
1261 # special case: don't return empty table
1262 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1263 $out = '';
1264 }
1265
1266 return $out;
1267 }
1268
1269 /**
1270 * Helper function for parse() that transforms wiki markup into half-parsed
1271 * HTML. Only called for $mOutputType == self::OT_HTML.
1272 *
1273 * @private
1274 *
1275 * @param string $text The text to parse
1276 * @param bool $isMain Whether this is being called from the main parse() function
1277 * @param PPFrame|bool $frame A pre-processor frame
1278 *
1279 * @return string
1280 */
1281 public function internalParse( $text, $isMain = true, $frame = false ) {
1282
1283 $origText = $text;
1284
1285 // Avoid PHP 7.1 warning from passing $this by reference
1286 $parser = $this;
1287
1288 # Hook to suspend the parser in this state
1289 if ( !Hooks::run( 'ParserBeforeInternalParse', [ &$parser, &$text, &$this->mStripState ] ) ) {
1290 return $text;
1291 }
1292
1293 # if $frame is provided, then use $frame for replacing any variables
1294 if ( $frame ) {
1295 # use frame depth to infer how include/noinclude tags should be handled
1296 # depth=0 means this is the top-level document; otherwise it's an included document
1297 if ( !$frame->depth ) {
1298 $flag = 0;
1299 } else {
1300 $flag = Parser::PTD_FOR_INCLUSION;
1301 }
1302 $dom = $this->preprocessToDom( $text, $flag );
1303 $text = $frame->expand( $dom );
1304 } else {
1305 # if $frame is not provided, then use old-style replaceVariables
1306 $text = $this->replaceVariables( $text );
1307 }
1308
1309 Hooks::run( 'InternalParseBeforeSanitize', [ &$parser, &$text, &$this->mStripState ] );
1310 $text = Sanitizer::removeHTMLtags(
1311 $text,
1312 [ $this, 'attributeStripCallback' ],
1313 false,
1314 array_keys( $this->mTransparentTagHooks ),
1315 [],
1316 [ $this, 'addTrackingCategory' ]
1317 );
1318 Hooks::run( 'InternalParseBeforeLinks', [ &$parser, &$text, &$this->mStripState ] );
1319
1320 # Tables need to come after variable replacement for things to work
1321 # properly; putting them before other transformations should keep
1322 # exciting things like link expansions from showing up in surprising
1323 # places.
1324 $text = $this->doTableStuff( $text );
1325
1326 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1327
1328 $text = $this->doDoubleUnderscore( $text );
1329
1330 $text = $this->doHeadings( $text );
1331 $text = $this->replaceInternalLinks( $text );
1332 $text = $this->doAllQuotes( $text );
1333 $text = $this->replaceExternalLinks( $text );
1334
1335 # replaceInternalLinks may sometimes leave behind
1336 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1337 $text = str_replace( self::MARKER_PREFIX . 'NOPARSE', '', $text );
1338
1339 $text = $this->doMagicLinks( $text );
1340 $text = $this->formatHeadings( $text, $origText, $isMain );
1341
1342 return $text;
1343 }
1344
1345 /**
1346 * Helper function for parse() that transforms half-parsed HTML into fully
1347 * parsed HTML.
1348 *
1349 * @param string $text
1350 * @param bool $isMain
1351 * @param bool $linestart
1352 * @return string
1353 */
1354 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1355 $text = $this->mStripState->unstripGeneral( $text );
1356
1357 // Avoid PHP 7.1 warning from passing $this by reference
1358 $parser = $this;
1359
1360 if ( $isMain ) {
1361 Hooks::run( 'ParserAfterUnstrip', [ &$parser, &$text ] );
1362 }
1363
1364 # Clean up special characters, only run once, next-to-last before doBlockLevels
1365 $fixtags = [
1366 # French spaces, last one Guillemet-left
1367 # only if there is something before the space
1368 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;',
1369 # french spaces, Guillemet-right
1370 '/(\\302\\253) /' => '\\1&#160;',
1371 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, T13874.
1372 ];
1373 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
1374
1375 $text = $this->doBlockLevels( $text, $linestart );
1376
1377 $this->replaceLinkHolders( $text );
1378
1379 /**
1380 * The input doesn't get language converted if
1381 * a) It's disabled
1382 * b) Content isn't converted
1383 * c) It's a conversion table
1384 * d) it is an interface message (which is in the user language)
1385 */
1386 if ( !( $this->mOptions->getDisableContentConversion()
1387 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) )
1388 ) {
1389 if ( !$this->mOptions->getInterfaceMessage() ) {
1390 # The position of the convert() call should not be changed. it
1391 # assumes that the links are all replaced and the only thing left
1392 # is the <nowiki> mark.
1393 $text = $this->getConverterLanguage()->convert( $text );
1394 }
1395 }
1396
1397 $text = $this->mStripState->unstripNoWiki( $text );
1398
1399 if ( $isMain ) {
1400 Hooks::run( 'ParserBeforeTidy', [ &$parser, &$text ] );
1401 }
1402
1403 $text = $this->replaceTransparentTags( $text );
1404 $text = $this->mStripState->unstripGeneral( $text );
1405
1406 $text = Sanitizer::normalizeCharReferences( $text );
1407
1408 if ( MWTidy::isEnabled() ) {
1409 if ( $this->mOptions->getTidy() ) {
1410 $text = MWTidy::tidy( $text );
1411 }
1412 } else {
1413 # attempt to sanitize at least some nesting problems
1414 # (T4702 and quite a few others)
1415 $tidyregs = [
1416 # ''Something [http://www.cool.com cool''] -->
1417 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1418 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1419 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1420 # fix up an anchor inside another anchor, only
1421 # at least for a single single nested link (T5695)
1422 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1423 '\\1\\2</a>\\3</a>\\1\\4</a>',
1424 # fix div inside inline elements- doBlockLevels won't wrap a line which
1425 # contains a div, so fix it up here; replace
1426 # div with escaped text
1427 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1428 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
1429 # remove empty italic or bold tag pairs, some
1430 # introduced by rules above
1431 '/<([bi])><\/\\1>/' => '',
1432 ];
1433
1434 $text = preg_replace(
1435 array_keys( $tidyregs ),
1436 array_values( $tidyregs ),
1437 $text );
1438 }
1439
1440 if ( $isMain ) {
1441 Hooks::run( 'ParserAfterTidy', [ &$parser, &$text ] );
1442 }
1443
1444 return $text;
1445 }
1446
1447 /**
1448 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1449 * magic external links.
1450 *
1451 * DML
1452 * @private
1453 *
1454 * @param string $text
1455 *
1456 * @return string
1457 */
1458 public function doMagicLinks( $text ) {
1459 $prots = wfUrlProtocolsWithoutProtRel();
1460 $urlChar = self::EXT_LINK_URL_CLASS;
1461 $addr = self::EXT_LINK_ADDR;
1462 $space = self::SPACE_NOT_NL; # non-newline space
1463 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1464 $spaces = "$space++"; # possessive match of 1 or more spaces
1465 $text = preg_replace_callback(
1466 '!(?: # Start cases
1467 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1468 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1469 (\b # m[3]: Free external links
1470 (?i:$prots)
1471 ($addr$urlChar*) # m[4]: Post-protocol path
1472 ) |
1473 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1474 ([0-9]+)\b |
1475 \bISBN $spaces ( # m[6]: ISBN, capture number
1476 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1477 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1478 [0-9Xx] # check digit
1479 )\b
1480 )!xu", [ $this, 'magicLinkCallback' ], $text );
1481 return $text;
1482 }
1483
1484 /**
1485 * @throws MWException
1486 * @param array $m
1487 * @return HTML|string
1488 */
1489 public function magicLinkCallback( $m ) {
1490 if ( isset( $m[1] ) && $m[1] !== '' ) {
1491 # Skip anchor
1492 return $m[0];
1493 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1494 # Skip HTML element
1495 return $m[0];
1496 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1497 # Free external link
1498 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1499 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1500 # RFC or PMID
1501 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1502 if ( !$this->mOptions->getMagicRFCLinks() ) {
1503 return $m[0];
1504 }
1505 $keyword = 'RFC';
1506 $urlmsg = 'rfcurl';
1507 $cssClass = 'mw-magiclink-rfc';
1508 $trackingCat = 'magiclink-tracking-rfc';
1509 $id = $m[5];
1510 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1511 if ( !$this->mOptions->getMagicPMIDLinks() ) {
1512 return $m[0];
1513 }
1514 $keyword = 'PMID';
1515 $urlmsg = 'pubmedurl';
1516 $cssClass = 'mw-magiclink-pmid';
1517 $trackingCat = 'magiclink-tracking-pmid';
1518 $id = $m[5];
1519 } else {
1520 throw new MWException( __METHOD__ . ': unrecognised match type "' .
1521 substr( $m[0], 0, 20 ) . '"' );
1522 }
1523 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1524 $this->addTrackingCategory( $trackingCat );
1525 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1526 } elseif ( isset( $m[6] ) && $m[6] !== ''
1527 && $this->mOptions->getMagicISBNLinks()
1528 ) {
1529 # ISBN
1530 $isbn = $m[6];
1531 $space = self::SPACE_NOT_NL; # non-newline space
1532 $isbn = preg_replace( "/$space/", ' ', $isbn );
1533 $num = strtr( $isbn, [
1534 '-' => '',
1535 ' ' => '',
1536 'x' => 'X',
1537 ] );
1538 $this->addTrackingCategory( 'magiclink-tracking-isbn' );
1539 return $this->getLinkRenderer()->makeKnownLink(
1540 SpecialPage::getTitleFor( 'Booksources', $num ),
1541 "ISBN $isbn",
1542 [
1543 'class' => 'internal mw-magiclink-isbn',
1544 'title' => false // suppress title attribute
1545 ]
1546 );
1547 } else {
1548 return $m[0];
1549 }
1550 }
1551
1552 /**
1553 * Make a free external link, given a user-supplied URL
1554 *
1555 * @param string $url
1556 * @param int $numPostProto
1557 * The number of characters after the protocol.
1558 * @return string HTML
1559 * @private
1560 */
1561 public function makeFreeExternalLink( $url, $numPostProto ) {
1562 $trail = '';
1563
1564 # The characters '<' and '>' (which were escaped by
1565 # removeHTMLtags()) should not be included in
1566 # URLs, per RFC 2396.
1567 # Make &nbsp; terminate a URL as well (bug T84937)
1568 $m2 = [];
1569 if ( preg_match(
1570 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1571 $url,
1572 $m2,
1573 PREG_OFFSET_CAPTURE
1574 ) ) {
1575 $trail = substr( $url, $m2[0][1] ) . $trail;
1576 $url = substr( $url, 0, $m2[0][1] );
1577 }
1578
1579 # Move trailing punctuation to $trail
1580 $sep = ',;\.:!?';
1581 # If there is no left bracket, then consider right brackets fair game too
1582 if ( strpos( $url, '(' ) === false ) {
1583 $sep .= ')';
1584 }
1585
1586 $urlRev = strrev( $url );
1587 $numSepChars = strspn( $urlRev, $sep );
1588 # Don't break a trailing HTML entity by moving the ; into $trail
1589 # This is in hot code, so use substr_compare to avoid having to
1590 # create a new string object for the comparison
1591 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1592 # more optimization: instead of running preg_match with a $
1593 # anchor, which can be slow, do the match on the reversed
1594 # string starting at the desired offset.
1595 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1596 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1597 $numSepChars--;
1598 }
1599 }
1600 if ( $numSepChars ) {
1601 $trail = substr( $url, -$numSepChars ) . $trail;
1602 $url = substr( $url, 0, -$numSepChars );
1603 }
1604
1605 # Verify that we still have a real URL after trail removal, and
1606 # not just lone protocol
1607 if ( strlen( $trail ) >= $numPostProto ) {
1608 return $url . $trail;
1609 }
1610
1611 $url = Sanitizer::cleanUrl( $url );
1612
1613 # Is this an external image?
1614 $text = $this->maybeMakeExternalImage( $url );
1615 if ( $text === false ) {
1616 # Not an image, make a link
1617 $text = Linker::makeExternalLink( $url,
1618 $this->getConverterLanguage()->markNoConversion( $url, true ),
1619 true, 'free',
1620 $this->getExternalLinkAttribs( $url ), $this->mTitle );
1621 # Register it in the output object...
1622 $this->mOutput->addExternalLink( $url );
1623 }
1624 return $text . $trail;
1625 }
1626
1627 /**
1628 * Parse headers and return html
1629 *
1630 * @private
1631 *
1632 * @param string $text
1633 *
1634 * @return string
1635 */
1636 public function doHeadings( $text ) {
1637 for ( $i = 6; $i >= 1; --$i ) {
1638 $h = str_repeat( '=', $i );
1639 $text = preg_replace( "/^$h(.+)$h\\s*$/m", "<h$i>\\1</h$i>", $text );
1640 }
1641 return $text;
1642 }
1643
1644 /**
1645 * Replace single quotes with HTML markup
1646 * @private
1647 *
1648 * @param string $text
1649 *
1650 * @return string The altered text
1651 */
1652 public function doAllQuotes( $text ) {
1653 $outtext = '';
1654 $lines = StringUtils::explode( "\n", $text );
1655 foreach ( $lines as $line ) {
1656 $outtext .= $this->doQuotes( $line ) . "\n";
1657 }
1658 $outtext = substr( $outtext, 0, -1 );
1659 return $outtext;
1660 }
1661
1662 /**
1663 * Helper function for doAllQuotes()
1664 *
1665 * @param string $text
1666 *
1667 * @return string
1668 */
1669 public function doQuotes( $text ) {
1670 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1671 $countarr = count( $arr );
1672 if ( $countarr == 1 ) {
1673 return $text;
1674 }
1675
1676 // First, do some preliminary work. This may shift some apostrophes from
1677 // being mark-up to being text. It also counts the number of occurrences
1678 // of bold and italics mark-ups.
1679 $numbold = 0;
1680 $numitalics = 0;
1681 for ( $i = 1; $i < $countarr; $i += 2 ) {
1682 $thislen = strlen( $arr[$i] );
1683 // If there are ever four apostrophes, assume the first is supposed to
1684 // be text, and the remaining three constitute mark-up for bold text.
1685 // (T15227: ''''foo'''' turns into ' ''' foo ' ''')
1686 if ( $thislen == 4 ) {
1687 $arr[$i - 1] .= "'";
1688 $arr[$i] = "'''";
1689 $thislen = 3;
1690 } elseif ( $thislen > 5 ) {
1691 // If there are more than 5 apostrophes in a row, assume they're all
1692 // text except for the last 5.
1693 // (T15227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1694 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1695 $arr[$i] = "'''''";
1696 $thislen = 5;
1697 }
1698 // Count the number of occurrences of bold and italics mark-ups.
1699 if ( $thislen == 2 ) {
1700 $numitalics++;
1701 } elseif ( $thislen == 3 ) {
1702 $numbold++;
1703 } elseif ( $thislen == 5 ) {
1704 $numitalics++;
1705 $numbold++;
1706 }
1707 }
1708
1709 // If there is an odd number of both bold and italics, it is likely
1710 // that one of the bold ones was meant to be an apostrophe followed
1711 // by italics. Which one we cannot know for certain, but it is more
1712 // likely to be one that has a single-letter word before it.
1713 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1714 $firstsingleletterword = -1;
1715 $firstmultiletterword = -1;
1716 $firstspace = -1;
1717 for ( $i = 1; $i < $countarr; $i += 2 ) {
1718 if ( strlen( $arr[$i] ) == 3 ) {
1719 $x1 = substr( $arr[$i - 1], -1 );
1720 $x2 = substr( $arr[$i - 1], -2, 1 );
1721 if ( $x1 === ' ' ) {
1722 if ( $firstspace == -1 ) {
1723 $firstspace = $i;
1724 }
1725 } elseif ( $x2 === ' ' ) {
1726 $firstsingleletterword = $i;
1727 // if $firstsingleletterword is set, we don't
1728 // look at the other options, so we can bail early.
1729 break;
1730 } else {
1731 if ( $firstmultiletterword == -1 ) {
1732 $firstmultiletterword = $i;
1733 }
1734 }
1735 }
1736 }
1737
1738 // If there is a single-letter word, use it!
1739 if ( $firstsingleletterword > -1 ) {
1740 $arr[$firstsingleletterword] = "''";
1741 $arr[$firstsingleletterword - 1] .= "'";
1742 } elseif ( $firstmultiletterword > -1 ) {
1743 // If not, but there's a multi-letter word, use that one.
1744 $arr[$firstmultiletterword] = "''";
1745 $arr[$firstmultiletterword - 1] .= "'";
1746 } elseif ( $firstspace > -1 ) {
1747 // ... otherwise use the first one that has neither.
1748 // (notice that it is possible for all three to be -1 if, for example,
1749 // there is only one pentuple-apostrophe in the line)
1750 $arr[$firstspace] = "''";
1751 $arr[$firstspace - 1] .= "'";
1752 }
1753 }
1754
1755 // Now let's actually convert our apostrophic mush to HTML!
1756 $output = '';
1757 $buffer = '';
1758 $state = '';
1759 $i = 0;
1760 foreach ( $arr as $r ) {
1761 if ( ( $i % 2 ) == 0 ) {
1762 if ( $state === 'both' ) {
1763 $buffer .= $r;
1764 } else {
1765 $output .= $r;
1766 }
1767 } else {
1768 $thislen = strlen( $r );
1769 if ( $thislen == 2 ) {
1770 if ( $state === 'i' ) {
1771 $output .= '</i>';
1772 $state = '';
1773 } elseif ( $state === 'bi' ) {
1774 $output .= '</i>';
1775 $state = 'b';
1776 } elseif ( $state === 'ib' ) {
1777 $output .= '</b></i><b>';
1778 $state = 'b';
1779 } elseif ( $state === 'both' ) {
1780 $output .= '<b><i>' . $buffer . '</i>';
1781 $state = 'b';
1782 } else { // $state can be 'b' or ''
1783 $output .= '<i>';
1784 $state .= 'i';
1785 }
1786 } elseif ( $thislen == 3 ) {
1787 if ( $state === 'b' ) {
1788 $output .= '</b>';
1789 $state = '';
1790 } elseif ( $state === 'bi' ) {
1791 $output .= '</i></b><i>';
1792 $state = 'i';
1793 } elseif ( $state === 'ib' ) {
1794 $output .= '</b>';
1795 $state = 'i';
1796 } elseif ( $state === 'both' ) {
1797 $output .= '<i><b>' . $buffer . '</b>';
1798 $state = 'i';
1799 } else { // $state can be 'i' or ''
1800 $output .= '<b>';
1801 $state .= 'b';
1802 }
1803 } elseif ( $thislen == 5 ) {
1804 if ( $state === 'b' ) {
1805 $output .= '</b><i>';
1806 $state = 'i';
1807 } elseif ( $state === 'i' ) {
1808 $output .= '</i><b>';
1809 $state = 'b';
1810 } elseif ( $state === 'bi' ) {
1811 $output .= '</i></b>';
1812 $state = '';
1813 } elseif ( $state === 'ib' ) {
1814 $output .= '</b></i>';
1815 $state = '';
1816 } elseif ( $state === 'both' ) {
1817 $output .= '<i><b>' . $buffer . '</b></i>';
1818 $state = '';
1819 } else { // ($state == '')
1820 $buffer = '';
1821 $state = 'both';
1822 }
1823 }
1824 }
1825 $i++;
1826 }
1827 // Now close all remaining tags. Notice that the order is important.
1828 if ( $state === 'b' || $state === 'ib' ) {
1829 $output .= '</b>';
1830 }
1831 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1832 $output .= '</i>';
1833 }
1834 if ( $state === 'bi' ) {
1835 $output .= '</b>';
1836 }
1837 // There might be lonely ''''', so make sure we have a buffer
1838 if ( $state === 'both' && $buffer ) {
1839 $output .= '<b><i>' . $buffer . '</i></b>';
1840 }
1841 return $output;
1842 }
1843
1844 /**
1845 * Replace external links (REL)
1846 *
1847 * Note: this is all very hackish and the order of execution matters a lot.
1848 * Make sure to run tests/parser/parserTests.php if you change this code.
1849 *
1850 * @private
1851 *
1852 * @param string $text
1853 *
1854 * @throws MWException
1855 * @return string
1856 */
1857 public function replaceExternalLinks( $text ) {
1858
1859 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1860 if ( $bits === false ) {
1861 throw new MWException( "PCRE needs to be compiled with "
1862 . "--enable-unicode-properties in order for MediaWiki to function" );
1863 }
1864 $s = array_shift( $bits );
1865
1866 $i = 0;
1867 while ( $i < count( $bits ) ) {
1868 $url = $bits[$i++];
1869 $i++; // protocol
1870 $text = $bits[$i++];
1871 $trail = $bits[$i++];
1872
1873 # The characters '<' and '>' (which were escaped by
1874 # removeHTMLtags()) should not be included in
1875 # URLs, per RFC 2396.
1876 $m2 = [];
1877 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1878 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1879 $url = substr( $url, 0, $m2[0][1] );
1880 }
1881
1882 # If the link text is an image URL, replace it with an <img> tag
1883 # This happened by accident in the original parser, but some people used it extensively
1884 $img = $this->maybeMakeExternalImage( $text );
1885 if ( $img !== false ) {
1886 $text = $img;
1887 }
1888
1889 $dtrail = '';
1890
1891 # Set linktype for CSS - if URL==text, link is essentially free
1892 $linktype = ( $text === $url ) ? 'free' : 'text';
1893
1894 # No link text, e.g. [http://domain.tld/some.link]
1895 if ( $text == '' ) {
1896 # Autonumber
1897 $langObj = $this->getTargetLanguage();
1898 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1899 $linktype = 'autonumber';
1900 } else {
1901 # Have link text, e.g. [http://domain.tld/some.link text]s
1902 # Check for trail
1903 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1904 }
1905
1906 $text = $this->getConverterLanguage()->markNoConversion( $text );
1907
1908 $url = Sanitizer::cleanUrl( $url );
1909
1910 # Use the encoded URL
1911 # This means that users can paste URLs directly into the text
1912 # Funny characters like ö aren't valid in URLs anyway
1913 # This was changed in August 2004
1914 $s .= Linker::makeExternalLink( $url, $text, false, $linktype,
1915 $this->getExternalLinkAttribs( $url ), $this->mTitle ) . $dtrail . $trail;
1916
1917 # Register link in the output object.
1918 $this->mOutput->addExternalLink( $url );
1919 }
1920
1921 return $s;
1922 }
1923
1924 /**
1925 * Get the rel attribute for a particular external link.
1926 *
1927 * @since 1.21
1928 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1929 * nofollow if appropriate
1930 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1931 * @return string|null Rel attribute for $url
1932 */
1933 public static function getExternalLinkRel( $url = false, $title = null ) {
1934 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1935 $ns = $title ? $title->getNamespace() : false;
1936 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1937 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1938 ) {
1939 return 'nofollow';
1940 }
1941 return null;
1942 }
1943
1944 /**
1945 * Get an associative array of additional HTML attributes appropriate for a
1946 * particular external link. This currently may include rel => nofollow
1947 * (depending on configuration, namespace, and the URL's domain) and/or a
1948 * target attribute (depending on configuration).
1949 *
1950 * @param string $url URL to extract the domain from for rel =>
1951 * nofollow if appropriate
1952 * @return array Associative array of HTML attributes
1953 */
1954 public function getExternalLinkAttribs( $url ) {
1955 $attribs = [];
1956 $rel = self::getExternalLinkRel( $url, $this->mTitle );
1957
1958 $target = $this->mOptions->getExternalLinkTarget();
1959 if ( $target ) {
1960 $attribs['target'] = $target;
1961 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1962 // T133507. New windows can navigate parent cross-origin.
1963 // Including noreferrer due to lacking browser
1964 // support of noopener. Eventually noreferrer should be removed.
1965 if ( $rel !== '' ) {
1966 $rel .= ' ';
1967 }
1968 $rel .= 'noreferrer noopener';
1969 }
1970 }
1971 $attribs['rel'] = $rel;
1972 return $attribs;
1973 }
1974
1975 /**
1976 * Replace unusual escape codes in a URL with their equivalent characters
1977 *
1978 * This generally follows the syntax defined in RFC 3986, with special
1979 * consideration for HTTP query strings.
1980 *
1981 * @param string $url
1982 * @return string
1983 */
1984 public static function normalizeLinkUrl( $url ) {
1985 # First, make sure unsafe characters are encoded
1986 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
1987 function ( $m ) {
1988 return rawurlencode( $m[0] );
1989 },
1990 $url
1991 );
1992
1993 $ret = '';
1994 $end = strlen( $url );
1995
1996 # Fragment part - 'fragment'
1997 $start = strpos( $url, '#' );
1998 if ( $start !== false && $start < $end ) {
1999 $ret = self::normalizeUrlComponent(
2000 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
2001 $end = $start;
2002 }
2003
2004 # Query part - 'query' minus &=+;
2005 $start = strpos( $url, '?' );
2006 if ( $start !== false && $start < $end ) {
2007 $ret = self::normalizeUrlComponent(
2008 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
2009 $end = $start;
2010 }
2011
2012 # Scheme and path part - 'pchar'
2013 # (we assume no userinfo or encoded colons in the host)
2014 $ret = self::normalizeUrlComponent(
2015 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
2016
2017 return $ret;
2018 }
2019
2020 private static function normalizeUrlComponent( $component, $unsafe ) {
2021 $callback = function ( $matches ) use ( $unsafe ) {
2022 $char = urldecode( $matches[0] );
2023 $ord = ord( $char );
2024 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
2025 # Unescape it
2026 return $char;
2027 } else {
2028 # Leave it escaped, but use uppercase for a-f
2029 return strtoupper( $matches[0] );
2030 }
2031 };
2032 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
2033 }
2034
2035 /**
2036 * make an image if it's allowed, either through the global
2037 * option, through the exception, or through the on-wiki whitelist
2038 *
2039 * @param string $url
2040 *
2041 * @return string
2042 */
2043 private function maybeMakeExternalImage( $url ) {
2044 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
2045 $imagesexception = !empty( $imagesfrom );
2046 $text = false;
2047 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2048 if ( $imagesexception && is_array( $imagesfrom ) ) {
2049 $imagematch = false;
2050 foreach ( $imagesfrom as $match ) {
2051 if ( strpos( $url, $match ) === 0 ) {
2052 $imagematch = true;
2053 break;
2054 }
2055 }
2056 } elseif ( $imagesexception ) {
2057 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2058 } else {
2059 $imagematch = false;
2060 }
2061
2062 if ( $this->mOptions->getAllowExternalImages()
2063 || ( $imagesexception && $imagematch )
2064 ) {
2065 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
2066 # Image found
2067 $text = Linker::makeExternalImage( $url );
2068 }
2069 }
2070 if ( !$text && $this->mOptions->getEnableImageWhitelist()
2071 && preg_match( self::EXT_IMAGE_REGEX, $url )
2072 ) {
2073 $whitelist = explode(
2074 "\n",
2075 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2076 );
2077
2078 foreach ( $whitelist as $entry ) {
2079 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2080 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
2081 continue;
2082 }
2083 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2084 # Image matches a whitelist entry
2085 $text = Linker::makeExternalImage( $url );
2086 break;
2087 }
2088 }
2089 }
2090 return $text;
2091 }
2092
2093 /**
2094 * Process [[ ]] wikilinks
2095 *
2096 * @param string $s
2097 *
2098 * @return string Processed text
2099 *
2100 * @private
2101 */
2102 public function replaceInternalLinks( $s ) {
2103 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
2104 return $s;
2105 }
2106
2107 /**
2108 * Process [[ ]] wikilinks (RIL)
2109 * @param string $s
2110 * @throws MWException
2111 * @return LinkHolderArray
2112 *
2113 * @private
2114 */
2115 public function replaceInternalLinks2( &$s ) {
2116 global $wgExtraInterlanguageLinkPrefixes;
2117
2118 static $tc = false, $e1, $e1_img;
2119 # the % is needed to support urlencoded titles as well
2120 if ( !$tc ) {
2121 $tc = Title::legalChars() . '#%';
2122 # Match a link having the form [[namespace:link|alternate]]trail
2123 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2124 # Match cases where there is no "]]", which might still be images
2125 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2126 }
2127
2128 $holders = new LinkHolderArray( $this );
2129
2130 # split the entire text string on occurrences of [[
2131 $a = StringUtils::explode( '[[', ' ' . $s );
2132 # get the first element (all text up to first [[), and remove the space we added
2133 $s = $a->current();
2134 $a->next();
2135 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2136 $s = substr( $s, 1 );
2137
2138 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2139 $e2 = null;
2140 if ( $useLinkPrefixExtension ) {
2141 # Match the end of a line for a word that's not followed by whitespace,
2142 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2143 global $wgContLang;
2144 $charset = $wgContLang->linkPrefixCharset();
2145 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2146 }
2147
2148 if ( is_null( $this->mTitle ) ) {
2149 throw new MWException( __METHOD__ . ": \$this->mTitle is null\n" );
2150 }
2151 $nottalk = !$this->mTitle->isTalkPage();
2152
2153 if ( $useLinkPrefixExtension ) {
2154 $m = [];
2155 if ( preg_match( $e2, $s, $m ) ) {
2156 $first_prefix = $m[2];
2157 } else {
2158 $first_prefix = false;
2159 }
2160 } else {
2161 $prefix = '';
2162 }
2163
2164 $useSubpages = $this->areSubpagesAllowed();
2165
2166 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2167 # Loop for each link
2168 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2169 // @codingStandardsIgnoreEnd
2170
2171 # Check for excessive memory usage
2172 if ( $holders->isBig() ) {
2173 # Too big
2174 # Do the existence check, replace the link holders and clear the array
2175 $holders->replace( $s );
2176 $holders->clear();
2177 }
2178
2179 if ( $useLinkPrefixExtension ) {
2180 if ( preg_match( $e2, $s, $m ) ) {
2181 $prefix = $m[2];
2182 $s = $m[1];
2183 } else {
2184 $prefix = '';
2185 }
2186 # first link
2187 if ( $first_prefix ) {
2188 $prefix = $first_prefix;
2189 $first_prefix = false;
2190 }
2191 }
2192
2193 $might_be_img = false;
2194
2195 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2196 $text = $m[2];
2197 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2198 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2199 # the real problem is with the $e1 regex
2200 # See T1500.
2201 # Still some problems for cases where the ] is meant to be outside punctuation,
2202 # and no image is in sight. See T4095.
2203 if ( $text !== ''
2204 && substr( $m[3], 0, 1 ) === ']'
2205 && strpos( $text, '[' ) !== false
2206 ) {
2207 $text .= ']'; # so that replaceExternalLinks($text) works later
2208 $m[3] = substr( $m[3], 1 );
2209 }
2210 # fix up urlencoded title texts
2211 if ( strpos( $m[1], '%' ) !== false ) {
2212 # Should anchors '#' also be rejected?
2213 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2214 }
2215 $trail = $m[3];
2216 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2217 # Invalid, but might be an image with a link in its caption
2218 $might_be_img = true;
2219 $text = $m[2];
2220 if ( strpos( $m[1], '%' ) !== false ) {
2221 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2222 }
2223 $trail = "";
2224 } else { # Invalid form; output directly
2225 $s .= $prefix . '[[' . $line;
2226 continue;
2227 }
2228
2229 $origLink = ltrim( $m[1], ' ' );
2230
2231 # Don't allow internal links to pages containing
2232 # PROTO: where PROTO is a valid URL protocol; these
2233 # should be external links.
2234 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $origLink ) ) {
2235 $s .= $prefix . '[[' . $line;
2236 continue;
2237 }
2238
2239 # Make subpage if necessary
2240 if ( $useSubpages ) {
2241 $link = $this->maybeDoSubpageLink( $origLink, $text );
2242 } else {
2243 $link = $origLink;
2244 }
2245
2246 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2247 if ( !$noforce ) {
2248 # Strip off leading ':'
2249 $link = substr( $link, 1 );
2250 }
2251
2252 $unstrip = $this->mStripState->unstripNoWiki( $link );
2253 $nt = is_string( $unstrip ) ? Title::newFromText( $unstrip ) : null;
2254 if ( $nt === null ) {
2255 $s .= $prefix . '[[' . $line;
2256 continue;
2257 }
2258
2259 $ns = $nt->getNamespace();
2260 $iw = $nt->getInterwiki();
2261
2262 if ( $might_be_img ) { # if this is actually an invalid link
2263 if ( $ns == NS_FILE && $noforce ) { # but might be an image
2264 $found = false;
2265 while ( true ) {
2266 # look at the next 'line' to see if we can close it there
2267 $a->next();
2268 $next_line = $a->current();
2269 if ( $next_line === false || $next_line === null ) {
2270 break;
2271 }
2272 $m = explode( ']]', $next_line, 3 );
2273 if ( count( $m ) == 3 ) {
2274 # the first ]] closes the inner link, the second the image
2275 $found = true;
2276 $text .= "[[{$m[0]}]]{$m[1]}";
2277 $trail = $m[2];
2278 break;
2279 } elseif ( count( $m ) == 2 ) {
2280 # if there's exactly one ]] that's fine, we'll keep looking
2281 $text .= "[[{$m[0]}]]{$m[1]}";
2282 } else {
2283 # if $next_line is invalid too, we need look no further
2284 $text .= '[[' . $next_line;
2285 break;
2286 }
2287 }
2288 if ( !$found ) {
2289 # we couldn't find the end of this imageLink, so output it raw
2290 # but don't ignore what might be perfectly normal links in the text we've examined
2291 $holders->merge( $this->replaceInternalLinks2( $text ) );
2292 $s .= "{$prefix}[[$link|$text";
2293 # note: no $trail, because without an end, there *is* no trail
2294 continue;
2295 }
2296 } else { # it's not an image, so output it raw
2297 $s .= "{$prefix}[[$link|$text";
2298 # note: no $trail, because without an end, there *is* no trail
2299 continue;
2300 }
2301 }
2302
2303 $wasblank = ( $text == '' );
2304 if ( $wasblank ) {
2305 $text = $link;
2306 } else {
2307 # T6598 madness. Handle the quotes only if they come from the alternate part
2308 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2309 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2310 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2311 $text = $this->doQuotes( $text );
2312 }
2313
2314 # Link not escaped by : , create the various objects
2315 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2316 # Interwikis
2317 if (
2318 $iw && $this->mOptions->getInterwikiMagic() && $nottalk && (
2319 Language::fetchLanguageName( $iw, null, 'mw' ) ||
2320 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2321 )
2322 ) {
2323 # T26502: filter duplicates
2324 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
2325 $this->mLangLinkLanguages[$iw] = true;
2326 $this->mOutput->addLanguageLink( $nt->getFullText() );
2327 }
2328
2329 $s = rtrim( $s . $prefix );
2330 $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail;
2331 continue;
2332 }
2333
2334 if ( $ns == NS_FILE ) {
2335 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
2336 if ( $wasblank ) {
2337 # if no parameters were passed, $text
2338 # becomes something like "File:Foo.png",
2339 # which we don't want to pass on to the
2340 # image generator
2341 $text = '';
2342 } else {
2343 # recursively parse links inside the image caption
2344 # actually, this will parse them in any other parameters, too,
2345 # but it might be hard to fix that, and it doesn't matter ATM
2346 $text = $this->replaceExternalLinks( $text );
2347 $holders->merge( $this->replaceInternalLinks2( $text ) );
2348 }
2349 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2350 $s .= $prefix . $this->armorLinks(
2351 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2352 continue;
2353 }
2354 } elseif ( $ns == NS_CATEGORY ) {
2355 $s = rtrim( $s . "\n" ); # T2087
2356
2357 if ( $wasblank ) {
2358 $sortkey = $this->getDefaultSort();
2359 } else {
2360 $sortkey = $text;
2361 }
2362 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
2363 $sortkey = str_replace( "\n", '', $sortkey );
2364 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
2365 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
2366
2367 /**
2368 * Strip the whitespace Category links produce, see T2087
2369 */
2370 $s .= trim( $prefix . $trail, "\n" ) == '' ? '' : $prefix . $trail;
2371
2372 continue;
2373 }
2374 }
2375
2376 # Self-link checking. For some languages, variants of the title are checked in
2377 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2378 # for linking to a different variant.
2379 if ( $ns != NS_SPECIAL && $nt->equals( $this->mTitle ) && !$nt->hasFragment() ) {
2380 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail );
2381 continue;
2382 }
2383
2384 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2385 # @todo FIXME: Should do batch file existence checks, see comment below
2386 if ( $ns == NS_MEDIA ) {
2387 # Give extensions a chance to select the file revision for us
2388 $options = [];
2389 $descQuery = false;
2390 Hooks::run( 'BeforeParserFetchFileAndTitle',
2391 [ $this, $nt, &$options, &$descQuery ] );
2392 # Fetch and register the file (file title may be different via hooks)
2393 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2394 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2395 $s .= $prefix . $this->armorLinks(
2396 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2397 continue;
2398 }
2399
2400 # Some titles, such as valid special pages or files in foreign repos, should
2401 # be shown as bluelinks even though they're not included in the page table
2402 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2403 # batch file existence checks for NS_FILE and NS_MEDIA
2404 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2405 $this->mOutput->addLink( $nt );
2406 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2407 } else {
2408 # Links will be added to the output link list after checking
2409 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2410 }
2411 }
2412 return $holders;
2413 }
2414
2415 /**
2416 * Render a forced-blue link inline; protect against double expansion of
2417 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2418 * Since this little disaster has to split off the trail text to avoid
2419 * breaking URLs in the following text without breaking trails on the
2420 * wiki links, it's been made into a horrible function.
2421 *
2422 * @param Title $nt
2423 * @param string $text
2424 * @param string $trail
2425 * @param string $prefix
2426 * @return string HTML-wikitext mix oh yuck
2427 */
2428 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2429 list( $inside, $trail ) = Linker::splitTrail( $trail );
2430
2431 if ( $text == '' ) {
2432 $text = htmlspecialchars( $nt->getPrefixedText() );
2433 }
2434
2435 $link = $this->getLinkRenderer()->makeKnownLink(
2436 $nt, new HtmlArmor( "$prefix$text$inside" )
2437 );
2438
2439 return $this->armorLinks( $link ) . $trail;
2440 }
2441
2442 /**
2443 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2444 * going to go through further parsing steps before inline URL expansion.
2445 *
2446 * Not needed quite as much as it used to be since free links are a bit
2447 * more sensible these days. But bracketed links are still an issue.
2448 *
2449 * @param string $text More-or-less HTML
2450 * @return string Less-or-more HTML with NOPARSE bits
2451 */
2452 public function armorLinks( $text ) {
2453 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/',
2454 self::MARKER_PREFIX . "NOPARSE$1", $text );
2455 }
2456
2457 /**
2458 * Return true if subpage links should be expanded on this page.
2459 * @return bool
2460 */
2461 public function areSubpagesAllowed() {
2462 # Some namespaces don't allow subpages
2463 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2464 }
2465
2466 /**
2467 * Handle link to subpage if necessary
2468 *
2469 * @param string $target The source of the link
2470 * @param string &$text The link text, modified as necessary
2471 * @return string The full name of the link
2472 * @private
2473 */
2474 public function maybeDoSubpageLink( $target, &$text ) {
2475 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2476 }
2477
2478 /**
2479 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2480 *
2481 * @param string $text
2482 * @param bool $linestart Whether or not this is at the start of a line.
2483 * @private
2484 * @return string The lists rendered as HTML
2485 */
2486 public function doBlockLevels( $text, $linestart ) {
2487 return BlockLevelPass::doBlockLevels( $text, $linestart );
2488 }
2489
2490 /**
2491 * Return value of a magic variable (like PAGENAME)
2492 *
2493 * @private
2494 *
2495 * @param string $index Magic variable identifier as mapped in MagicWord::$mVariableIDs
2496 * @param bool|PPFrame $frame
2497 *
2498 * @throws MWException
2499 * @return string
2500 */
2501 public function getVariableValue( $index, $frame = false ) {
2502 global $wgContLang, $wgSitename, $wgServer, $wgServerName;
2503 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2504
2505 if ( is_null( $this->mTitle ) ) {
2506 // If no title set, bad things are going to happen
2507 // later. Title should always be set since this
2508 // should only be called in the middle of a parse
2509 // operation (but the unit-tests do funky stuff)
2510 throw new MWException( __METHOD__ . ' Should only be '
2511 . ' called while parsing (no title set)' );
2512 }
2513
2514 // Avoid PHP 7.1 warning from passing $this by reference
2515 $parser = $this;
2516
2517 /**
2518 * Some of these require message or data lookups and can be
2519 * expensive to check many times.
2520 */
2521 if ( Hooks::run( 'ParserGetVariableValueVarCache', [ &$parser, &$this->mVarCache ] ) ) {
2522 if ( isset( $this->mVarCache[$index] ) ) {
2523 return $this->mVarCache[$index];
2524 }
2525 }
2526
2527 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2528 Hooks::run( 'ParserGetVariableValueTs', [ &$parser, &$ts ] );
2529
2530 $pageLang = $this->getFunctionLang();
2531
2532 switch ( $index ) {
2533 case '!':
2534 $value = '|';
2535 break;
2536 case 'currentmonth':
2537 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'm' ) );
2538 break;
2539 case 'currentmonth1':
2540 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2541 break;
2542 case 'currentmonthname':
2543 $value = $pageLang->getMonthName( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2544 break;
2545 case 'currentmonthnamegen':
2546 $value = $pageLang->getMonthNameGen( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2547 break;
2548 case 'currentmonthabbrev':
2549 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2550 break;
2551 case 'currentday':
2552 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'j' ) );
2553 break;
2554 case 'currentday2':
2555 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'd' ) );
2556 break;
2557 case 'localmonth':
2558 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'm' ) );
2559 break;
2560 case 'localmonth1':
2561 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2562 break;
2563 case 'localmonthname':
2564 $value = $pageLang->getMonthName( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2565 break;
2566 case 'localmonthnamegen':
2567 $value = $pageLang->getMonthNameGen( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2568 break;
2569 case 'localmonthabbrev':
2570 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2571 break;
2572 case 'localday':
2573 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'j' ) );
2574 break;
2575 case 'localday2':
2576 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'd' ) );
2577 break;
2578 case 'pagename':
2579 $value = wfEscapeWikiText( $this->mTitle->getText() );
2580 break;
2581 case 'pagenamee':
2582 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() );
2583 break;
2584 case 'fullpagename':
2585 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2586 break;
2587 case 'fullpagenamee':
2588 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() );
2589 break;
2590 case 'subpagename':
2591 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2592 break;
2593 case 'subpagenamee':
2594 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() );
2595 break;
2596 case 'rootpagename':
2597 $value = wfEscapeWikiText( $this->mTitle->getRootText() );
2598 break;
2599 case 'rootpagenamee':
2600 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2601 ' ',
2602 '_',
2603 $this->mTitle->getRootText()
2604 ) ) );
2605 break;
2606 case 'basepagename':
2607 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2608 break;
2609 case 'basepagenamee':
2610 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2611 ' ',
2612 '_',
2613 $this->mTitle->getBaseText()
2614 ) ) );
2615 break;
2616 case 'talkpagename':
2617 if ( $this->mTitle->canTalk() ) {
2618 $talkPage = $this->mTitle->getTalkPage();
2619 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2620 } else {
2621 $value = '';
2622 }
2623 break;
2624 case 'talkpagenamee':
2625 if ( $this->mTitle->canTalk() ) {
2626 $talkPage = $this->mTitle->getTalkPage();
2627 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2628 } else {
2629 $value = '';
2630 }
2631 break;
2632 case 'subjectpagename':
2633 $subjPage = $this->mTitle->getSubjectPage();
2634 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2635 break;
2636 case 'subjectpagenamee':
2637 $subjPage = $this->mTitle->getSubjectPage();
2638 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2639 break;
2640 case 'pageid': // requested in T25427
2641 $pageid = $this->getTitle()->getArticleID();
2642 if ( $pageid == 0 ) {
2643 # 0 means the page doesn't exist in the database,
2644 # which means the user is previewing a new page.
2645 # The vary-revision flag must be set, because the magic word
2646 # will have a different value once the page is saved.
2647 $this->mOutput->setFlag( 'vary-revision' );
2648 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2649 }
2650 $value = $pageid ? $pageid : null;
2651 break;
2652 case 'revisionid':
2653 # Let the edit saving system know we should parse the page
2654 # *after* a revision ID has been assigned.
2655 $this->mOutput->setFlag( 'vary-revision-id' );
2656 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision-id...\n" );
2657 $value = $this->mRevisionId;
2658 if ( !$value && $this->mOptions->getSpeculativeRevIdCallback() ) {
2659 $value = call_user_func( $this->mOptions->getSpeculativeRevIdCallback() );
2660 $this->mOutput->setSpeculativeRevIdUsed( $value );
2661 }
2662 break;
2663 case 'revisionday':
2664 # Let the edit saving system know we should parse the page
2665 # *after* a revision ID has been assigned. This is for null edits.
2666 $this->mOutput->setFlag( 'vary-revision' );
2667 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2668 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2669 break;
2670 case 'revisionday2':
2671 # Let the edit saving system know we should parse the page
2672 # *after* a revision ID has been assigned. This is for null edits.
2673 $this->mOutput->setFlag( 'vary-revision' );
2674 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2675 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2676 break;
2677 case 'revisionmonth':
2678 # Let the edit saving system know we should parse the page
2679 # *after* a revision ID has been assigned. This is for null edits.
2680 $this->mOutput->setFlag( 'vary-revision' );
2681 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2682 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2683 break;
2684 case 'revisionmonth1':
2685 # Let the edit saving system know we should parse the page
2686 # *after* a revision ID has been assigned. This is for null edits.
2687 $this->mOutput->setFlag( 'vary-revision' );
2688 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2689 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2690 break;
2691 case 'revisionyear':
2692 # Let the edit saving system know we should parse the page
2693 # *after* a revision ID has been assigned. This is for null edits.
2694 $this->mOutput->setFlag( 'vary-revision' );
2695 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2696 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2697 break;
2698 case 'revisiontimestamp':
2699 # Let the edit saving system know we should parse the page
2700 # *after* a revision ID has been assigned. This is for null edits.
2701 $this->mOutput->setFlag( 'vary-revision' );
2702 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2703 $value = $this->getRevisionTimestamp();
2704 break;
2705 case 'revisionuser':
2706 # Let the edit saving system know we should parse the page
2707 # *after* a revision ID has been assigned for null edits.
2708 $this->mOutput->setFlag( 'vary-user' );
2709 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-user...\n" );
2710 $value = $this->getRevisionUser();
2711 break;
2712 case 'revisionsize':
2713 $value = $this->getRevisionSize();
2714 break;
2715 case 'namespace':
2716 $value = str_replace( '_', ' ', $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2717 break;
2718 case 'namespacee':
2719 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2720 break;
2721 case 'namespacenumber':
2722 $value = $this->mTitle->getNamespace();
2723 break;
2724 case 'talkspace':
2725 $value = $this->mTitle->canTalk()
2726 ? str_replace( '_', ' ', $this->mTitle->getTalkNsText() )
2727 : '';
2728 break;
2729 case 'talkspacee':
2730 $value = $this->mTitle->canTalk() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2731 break;
2732 case 'subjectspace':
2733 $value = str_replace( '_', ' ', $this->mTitle->getSubjectNsText() );
2734 break;
2735 case 'subjectspacee':
2736 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2737 break;
2738 case 'currentdayname':
2739 $value = $pageLang->getWeekdayName( (int)MWTimestamp::getInstance( $ts )->format( 'w' ) + 1 );
2740 break;
2741 case 'currentyear':
2742 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'Y' ), true );
2743 break;
2744 case 'currenttime':
2745 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2746 break;
2747 case 'currenthour':
2748 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'H' ), true );
2749 break;
2750 case 'currentweek':
2751 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2752 # int to remove the padding
2753 $value = $pageLang->formatNum( (int)MWTimestamp::getInstance( $ts )->format( 'W' ) );
2754 break;
2755 case 'currentdow':
2756 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'w' ) );
2757 break;
2758 case 'localdayname':
2759 $value = $pageLang->getWeekdayName(
2760 (int)MWTimestamp::getLocalInstance( $ts )->format( 'w' ) + 1
2761 );
2762 break;
2763 case 'localyear':
2764 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'Y' ), true );
2765 break;
2766 case 'localtime':
2767 $value = $pageLang->time(
2768 MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' ),
2769 false,
2770 false
2771 );
2772 break;
2773 case 'localhour':
2774 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'H' ), true );
2775 break;
2776 case 'localweek':
2777 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2778 # int to remove the padding
2779 $value = $pageLang->formatNum( (int)MWTimestamp::getLocalInstance( $ts )->format( 'W' ) );
2780 break;
2781 case 'localdow':
2782 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'w' ) );
2783 break;
2784 case 'numberofarticles':
2785 $value = $pageLang->formatNum( SiteStats::articles() );
2786 break;
2787 case 'numberoffiles':
2788 $value = $pageLang->formatNum( SiteStats::images() );
2789 break;
2790 case 'numberofusers':
2791 $value = $pageLang->formatNum( SiteStats::users() );
2792 break;
2793 case 'numberofactiveusers':
2794 $value = $pageLang->formatNum( SiteStats::activeUsers() );
2795 break;
2796 case 'numberofpages':
2797 $value = $pageLang->formatNum( SiteStats::pages() );
2798 break;
2799 case 'numberofadmins':
2800 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2801 break;
2802 case 'numberofedits':
2803 $value = $pageLang->formatNum( SiteStats::edits() );
2804 break;
2805 case 'currenttimestamp':
2806 $value = wfTimestamp( TS_MW, $ts );
2807 break;
2808 case 'localtimestamp':
2809 $value = MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' );
2810 break;
2811 case 'currentversion':
2812 $value = SpecialVersion::getVersion();
2813 break;
2814 case 'articlepath':
2815 return $wgArticlePath;
2816 case 'sitename':
2817 return