phpunit: Avoid use of deprecated getMock for PHPUnit 5 compat
[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 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 Hooks::run( 'ParserLimitReport', [ $this, &$limitReport ] );
554
555 // Sanitize for comment. Note '‐' in the replacement is U+2010,
556 // which looks much like the problematic '-'.
557 $limitReport = str_replace( [ '-', '&' ], [ '‐', '&amp;' ], $limitReport );
558 $text .= "\n<!-- \n$limitReport-->\n";
559
560 // Add on template profiling data in human/machine readable way
561 $dataByFunc = $this->mProfiler->getFunctionStats();
562 uasort( $dataByFunc, function ( $a, $b ) {
563 return $a['real'] < $b['real']; // descending order
564 } );
565 $profileReport = [];
566 foreach ( array_slice( $dataByFunc, 0, 10 ) as $item ) {
567 $profileReport[] = sprintf( "%6.2f%% %8.3f %6d %s",
568 $item['%real'], $item['real'], $item['calls'],
569 htmlspecialchars( $item['name'] ) );
570 }
571 $text .= "<!--\nTransclusion expansion time report (%,ms,calls,template)\n";
572 $text .= implode( "\n", $profileReport ) . "\n-->\n";
573
574 $this->mOutput->setLimitReportData( 'limitreport-timingprofile', $profileReport );
575
576 // Add other cache related metadata
577 if ( $wgShowHostnames ) {
578 $this->mOutput->setLimitReportData( 'cachereport-origin', wfHostname() );
579 }
580 $this->mOutput->setLimitReportData( 'cachereport-timestamp',
581 $this->mOutput->getCacheTime() );
582 $this->mOutput->setLimitReportData( 'cachereport-ttl',
583 $this->mOutput->getCacheExpiry() );
584 $this->mOutput->setLimitReportData( 'cachereport-transientcontent',
585 $this->mOutput->hasDynamicContent() );
586
587 if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) {
588 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' .
589 $this->mTitle->getPrefixedDBkey() );
590 }
591 }
592 $this->mOutput->setText( $text );
593
594 $this->mRevisionId = $oldRevisionId;
595 $this->mRevisionObject = $oldRevisionObject;
596 $this->mRevisionTimestamp = $oldRevisionTimestamp;
597 $this->mRevisionUser = $oldRevisionUser;
598 $this->mRevisionSize = $oldRevisionSize;
599 $this->mInputSize = false;
600 $this->currentRevisionCache = null;
601
602 return $this->mOutput;
603 }
604
605 /**
606 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
607 * can be called from an extension tag hook.
608 *
609 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
610 * instead, which means that lists and links have not been fully parsed yet,
611 * and strip markers are still present.
612 *
613 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
614 *
615 * Use this function if you're a parser tag hook and you want to parse
616 * wikitext before or after applying additional transformations, and you
617 * intend to *return the result as hook output*, which will cause it to go
618 * through the rest of parsing process automatically.
619 *
620 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
621 * $text are not expanded
622 *
623 * @param string $text Text extension wants to have parsed
624 * @param bool|PPFrame $frame The frame to use for expanding any template variables
625 * @return string UNSAFE half-parsed HTML
626 */
627 public function recursiveTagParse( $text, $frame = false ) {
628 // Avoid PHP 7.1 warning from passing $this by reference
629 $parser = $this;
630 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
631 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
632 $text = $this->internalParse( $text, false, $frame );
633 return $text;
634 }
635
636 /**
637 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
638 * point can be called from an extension tag hook.
639 *
640 * The output of this function is fully-parsed HTML that is safe for output.
641 * If you're a parser tag hook, you might want to use recursiveTagParse()
642 * instead.
643 *
644 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
645 * $text are not expanded
646 *
647 * @since 1.25
648 *
649 * @param string $text Text extension wants to have parsed
650 * @param bool|PPFrame $frame The frame to use for expanding any template variables
651 * @return string Fully parsed HTML
652 */
653 public function recursiveTagParseFully( $text, $frame = false ) {
654 $text = $this->recursiveTagParse( $text, $frame );
655 $text = $this->internalParseHalfParsed( $text, false );
656 return $text;
657 }
658
659 /**
660 * Expand templates and variables in the text, producing valid, static wikitext.
661 * Also removes comments.
662 * Do not call this function recursively.
663 * @param string $text
664 * @param Title $title
665 * @param ParserOptions $options
666 * @param int|null $revid
667 * @param bool|PPFrame $frame
668 * @return mixed|string
669 */
670 public function preprocess( $text, Title $title = null,
671 ParserOptions $options, $revid = null, $frame = false
672 ) {
673 $magicScopeVariable = $this->lock();
674 $this->startParse( $title, $options, self::OT_PREPROCESS, true );
675 if ( $revid !== null ) {
676 $this->mRevisionId = $revid;
677 }
678 // Avoid PHP 7.1 warning from passing $this by reference
679 $parser = $this;
680 Hooks::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState ] );
681 Hooks::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState ] );
682 $text = $this->replaceVariables( $text, $frame );
683 $text = $this->mStripState->unstripBoth( $text );
684 return $text;
685 }
686
687 /**
688 * Recursive parser entry point that can be called from an extension tag
689 * hook.
690 *
691 * @param string $text Text to be expanded
692 * @param bool|PPFrame $frame The frame to use for expanding any template variables
693 * @return string
694 * @since 1.19
695 */
696 public function recursivePreprocess( $text, $frame = false ) {
697 $text = $this->replaceVariables( $text, $frame );
698 $text = $this->mStripState->unstripBoth( $text );
699 return $text;
700 }
701
702 /**
703 * Process the wikitext for the "?preload=" feature. (T7210)
704 *
705 * "<noinclude>", "<includeonly>" etc. are parsed as for template
706 * transclusion, comments, templates, arguments, tags hooks and parser
707 * functions are untouched.
708 *
709 * @param string $text
710 * @param Title $title
711 * @param ParserOptions $options
712 * @param array $params
713 * @return string
714 */
715 public function getPreloadText( $text, Title $title, ParserOptions $options, $params = [] ) {
716 $msg = new RawMessage( $text );
717 $text = $msg->params( $params )->plain();
718
719 # Parser (re)initialisation
720 $magicScopeVariable = $this->lock();
721 $this->startParse( $title, $options, self::OT_PLAIN, true );
722
723 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES;
724 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
725 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
726 $text = $this->mStripState->unstripBoth( $text );
727 return $text;
728 }
729
730 /**
731 * Get a random string
732 *
733 * @return string
734 * @deprecated since 1.26; use wfRandomString() instead.
735 */
736 public static function getRandomString() {
737 wfDeprecated( __METHOD__, '1.26' );
738 return wfRandomString( 16 );
739 }
740
741 /**
742 * Set the current user.
743 * Should only be used when doing pre-save transform.
744 *
745 * @param User|null $user User object or null (to reset)
746 */
747 public function setUser( $user ) {
748 $this->mUser = $user;
749 }
750
751 /**
752 * Accessor for mUniqPrefix.
753 *
754 * @return string
755 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
756 */
757 public function uniqPrefix() {
758 wfDeprecated( __METHOD__, '1.26' );
759 return self::MARKER_PREFIX;
760 }
761
762 /**
763 * Set the context title
764 *
765 * @param Title $t
766 */
767 public function setTitle( $t ) {
768 if ( !$t ) {
769 $t = Title::newFromText( 'NO TITLE' );
770 }
771
772 if ( $t->hasFragment() ) {
773 # Strip the fragment to avoid various odd effects
774 $this->mTitle = $t->createFragmentTarget( '' );
775 } else {
776 $this->mTitle = $t;
777 }
778 }
779
780 /**
781 * Accessor for the Title object
782 *
783 * @return Title
784 */
785 public function getTitle() {
786 return $this->mTitle;
787 }
788
789 /**
790 * Accessor/mutator for the Title object
791 *
792 * @param Title $x Title object or null to just get the current one
793 * @return Title
794 */
795 public function Title( $x = null ) {
796 return wfSetVar( $this->mTitle, $x );
797 }
798
799 /**
800 * Set the output type
801 *
802 * @param int $ot New value
803 */
804 public function setOutputType( $ot ) {
805 $this->mOutputType = $ot;
806 # Shortcut alias
807 $this->ot = [
808 'html' => $ot == self::OT_HTML,
809 'wiki' => $ot == self::OT_WIKI,
810 'pre' => $ot == self::OT_PREPROCESS,
811 'plain' => $ot == self::OT_PLAIN,
812 ];
813 }
814
815 /**
816 * Accessor/mutator for the output type
817 *
818 * @param int|null $x New value or null to just get the current one
819 * @return int
820 */
821 public function OutputType( $x = null ) {
822 return wfSetVar( $this->mOutputType, $x );
823 }
824
825 /**
826 * Get the ParserOutput object
827 *
828 * @return ParserOutput
829 */
830 public function getOutput() {
831 return $this->mOutput;
832 }
833
834 /**
835 * Get the ParserOptions object
836 *
837 * @return ParserOptions
838 */
839 public function getOptions() {
840 return $this->mOptions;
841 }
842
843 /**
844 * Accessor/mutator for the ParserOptions object
845 *
846 * @param ParserOptions $x New value or null to just get the current one
847 * @return ParserOptions Current ParserOptions object
848 */
849 public function Options( $x = null ) {
850 return wfSetVar( $this->mOptions, $x );
851 }
852
853 /**
854 * @return int
855 */
856 public function nextLinkID() {
857 return $this->mLinkID++;
858 }
859
860 /**
861 * @param int $id
862 */
863 public function setLinkID( $id ) {
864 $this->mLinkID = $id;
865 }
866
867 /**
868 * Get a language object for use in parser functions such as {{FORMATNUM:}}
869 * @return Language
870 */
871 public function getFunctionLang() {
872 return $this->getTargetLanguage();
873 }
874
875 /**
876 * Get the target language for the content being parsed. This is usually the
877 * language that the content is in.
878 *
879 * @since 1.19
880 *
881 * @throws MWException
882 * @return Language
883 */
884 public function getTargetLanguage() {
885 $target = $this->mOptions->getTargetLanguage();
886
887 if ( $target !== null ) {
888 return $target;
889 } elseif ( $this->mOptions->getInterfaceMessage() ) {
890 return $this->mOptions->getUserLangObj();
891 } elseif ( is_null( $this->mTitle ) ) {
892 throw new MWException( __METHOD__ . ': $this->mTitle is null' );
893 }
894
895 return $this->mTitle->getPageLanguage();
896 }
897
898 /**
899 * Get the language object for language conversion
900 * @return Language|null
901 */
902 public function getConverterLanguage() {
903 return $this->getTargetLanguage();
904 }
905
906 /**
907 * Get a User object either from $this->mUser, if set, or from the
908 * ParserOptions object otherwise
909 *
910 * @return User
911 */
912 public function getUser() {
913 if ( !is_null( $this->mUser ) ) {
914 return $this->mUser;
915 }
916 return $this->mOptions->getUser();
917 }
918
919 /**
920 * Get a preprocessor object
921 *
922 * @return Preprocessor
923 */
924 public function getPreprocessor() {
925 if ( !isset( $this->mPreprocessor ) ) {
926 $class = $this->mPreprocessorClass;
927 $this->mPreprocessor = new $class( $this );
928 }
929 return $this->mPreprocessor;
930 }
931
932 /**
933 * Get a LinkRenderer instance to make links with
934 *
935 * @since 1.28
936 * @return LinkRenderer
937 */
938 public function getLinkRenderer() {
939 if ( !$this->mLinkRenderer ) {
940 $this->mLinkRenderer = MediaWikiServices::getInstance()
941 ->getLinkRendererFactory()->create();
942 $this->mLinkRenderer->setStubThreshold(
943 $this->getOptions()->getStubThreshold()
944 );
945 }
946
947 return $this->mLinkRenderer;
948 }
949
950 /**
951 * Replaces all occurrences of HTML-style comments and the given tags
952 * in the text with a random marker and returns the next text. The output
953 * parameter $matches will be an associative array filled with data in
954 * the form:
955 *
956 * @code
957 * 'UNIQ-xxxxx' => [
958 * 'element',
959 * 'tag content',
960 * [ 'param' => 'x' ],
961 * '<element param="x">tag content</element>' ]
962 * @endcode
963 *
964 * @param array $elements List of element names. Comments are always extracted.
965 * @param string $text Source text string.
966 * @param array $matches Out parameter, Array: extracted tags
967 * @param string|null $uniq_prefix
968 * @return string Stripped text
969 * @since 1.26 The uniq_prefix argument is deprecated.
970 */
971 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = null ) {
972 if ( $uniq_prefix !== null ) {
973 wfDeprecated( __METHOD__ . ' called with $prefix argument', '1.26' );
974 }
975 static $n = 1;
976 $stripped = '';
977 $matches = [];
978
979 $taglist = implode( '|', $elements );
980 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
981
982 while ( $text != '' ) {
983 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
984 $stripped .= $p[0];
985 if ( count( $p ) < 5 ) {
986 break;
987 }
988 if ( count( $p ) > 5 ) {
989 # comment
990 $element = $p[4];
991 $attributes = '';
992 $close = '';
993 $inside = $p[5];
994 } else {
995 # tag
996 $element = $p[1];
997 $attributes = $p[2];
998 $close = $p[3];
999 $inside = $p[4];
1000 }
1001
1002 $marker = self::MARKER_PREFIX . "-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
1003 $stripped .= $marker;
1004
1005 if ( $close === '/>' ) {
1006 # Empty element tag, <tag />
1007 $content = null;
1008 $text = $inside;
1009 $tail = null;
1010 } else {
1011 if ( $element === '!--' ) {
1012 $end = '/(-->)/';
1013 } else {
1014 $end = "/(<\\/$element\\s*>)/i";
1015 }
1016 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
1017 $content = $q[0];
1018 if ( count( $q ) < 3 ) {
1019 # No end tag -- let it run out to the end of the text.
1020 $tail = '';
1021 $text = '';
1022 } else {
1023 $tail = $q[1];
1024 $text = $q[2];
1025 }
1026 }
1027
1028 $matches[$marker] = [ $element,
1029 $content,
1030 Sanitizer::decodeTagAttributes( $attributes ),
1031 "<$element$attributes$close$content$tail" ];
1032 }
1033 return $stripped;
1034 }
1035
1036 /**
1037 * Get a list of strippable XML-like elements
1038 *
1039 * @return array
1040 */
1041 public function getStripList() {
1042 return $this->mStripList;
1043 }
1044
1045 /**
1046 * Add an item to the strip state
1047 * Returns the unique tag which must be inserted into the stripped text
1048 * The tag will be replaced with the original text in unstrip()
1049 *
1050 * @param string $text
1051 *
1052 * @return string
1053 */
1054 public function insertStripItem( $text ) {
1055 $marker = self::MARKER_PREFIX . "-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
1056 $this->mMarkerIndex++;
1057 $this->mStripState->addGeneral( $marker, $text );
1058 return $marker;
1059 }
1060
1061 /**
1062 * parse the wiki syntax used to render tables
1063 *
1064 * @private
1065 * @param string $text
1066 * @return string
1067 */
1068 public function doTableStuff( $text ) {
1069
1070 $lines = StringUtils::explode( "\n", $text );
1071 $out = '';
1072 $td_history = []; # Is currently a td tag open?
1073 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1074 $tr_history = []; # Is currently a tr tag open?
1075 $tr_attributes = []; # history of tr attributes
1076 $has_opened_tr = []; # Did this table open a <tr> element?
1077 $indent_level = 0; # indent level of the table
1078
1079 foreach ( $lines as $outLine ) {
1080 $line = trim( $outLine );
1081
1082 if ( $line === '' ) { # empty line, go to next line
1083 $out .= $outLine . "\n";
1084 continue;
1085 }
1086
1087 $first_character = $line[0];
1088 $first_two = substr( $line, 0, 2 );
1089 $matches = [];
1090
1091 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1092 # First check if we are starting a new table
1093 $indent_level = strlen( $matches[1] );
1094
1095 $attributes = $this->mStripState->unstripBoth( $matches[2] );
1096 $attributes = Sanitizer::fixTagAttributes( $attributes, 'table' );
1097
1098 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1099 array_push( $td_history, false );
1100 array_push( $last_tag_history, '' );
1101 array_push( $tr_history, false );
1102 array_push( $tr_attributes, '' );
1103 array_push( $has_opened_tr, false );
1104 } elseif ( count( $td_history ) == 0 ) {
1105 # Don't do any of the following
1106 $out .= $outLine . "\n";
1107 continue;
1108 } elseif ( $first_two === '|}' ) {
1109 # We are ending a table
1110 $line = '</table>' . substr( $line, 2 );
1111 $last_tag = array_pop( $last_tag_history );
1112
1113 if ( !array_pop( $has_opened_tr ) ) {
1114 $line = "<tr><td></td></tr>{$line}";
1115 }
1116
1117 if ( array_pop( $tr_history ) ) {
1118 $line = "</tr>{$line}";
1119 }
1120
1121 if ( array_pop( $td_history ) ) {
1122 $line = "</{$last_tag}>{$line}";
1123 }
1124 array_pop( $tr_attributes );
1125 $outLine = $line . str_repeat( '</dd></dl>', $indent_level );
1126 } elseif ( $first_two === '|-' ) {
1127 # Now we have a table row
1128 $line = preg_replace( '#^\|-+#', '', $line );
1129
1130 # Whats after the tag is now only attributes
1131 $attributes = $this->mStripState->unstripBoth( $line );
1132 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
1133 array_pop( $tr_attributes );
1134 array_push( $tr_attributes, $attributes );
1135
1136 $line = '';
1137 $last_tag = array_pop( $last_tag_history );
1138 array_pop( $has_opened_tr );
1139 array_push( $has_opened_tr, true );
1140
1141 if ( array_pop( $tr_history ) ) {
1142 $line = '</tr>';
1143 }
1144
1145 if ( array_pop( $td_history ) ) {
1146 $line = "</{$last_tag}>{$line}";
1147 }
1148
1149 $outLine = $line;
1150 array_push( $tr_history, false );
1151 array_push( $td_history, false );
1152 array_push( $last_tag_history, '' );
1153 } elseif ( $first_character === '|'
1154 || $first_character === '!'
1155 || $first_two === '|+'
1156 ) {
1157 # This might be cell elements, td, th or captions
1158 if ( $first_two === '|+' ) {
1159 $first_character = '+';
1160 $line = substr( $line, 2 );
1161 } else {
1162 $line = substr( $line, 1 );
1163 }
1164
1165 // Implies both are valid for table headings.
1166 if ( $first_character === '!' ) {
1167 $line = StringUtils::replaceMarkup( '!!', '||', $line );
1168 }
1169
1170 # Split up multiple cells on the same line.
1171 # FIXME : This can result in improper nesting of tags processed
1172 # by earlier parser steps.
1173 $cells = explode( '||', $line );
1174
1175 $outLine = '';
1176
1177 # Loop through each table cell
1178 foreach ( $cells as $cell ) {
1179 $previous = '';
1180 if ( $first_character !== '+' ) {
1181 $tr_after = array_pop( $tr_attributes );
1182 if ( !array_pop( $tr_history ) ) {
1183 $previous = "<tr{$tr_after}>\n";
1184 }
1185 array_push( $tr_history, true );
1186 array_push( $tr_attributes, '' );
1187 array_pop( $has_opened_tr );
1188 array_push( $has_opened_tr, true );
1189 }
1190
1191 $last_tag = array_pop( $last_tag_history );
1192
1193 if ( array_pop( $td_history ) ) {
1194 $previous = "</{$last_tag}>\n{$previous}";
1195 }
1196
1197 if ( $first_character === '|' ) {
1198 $last_tag = 'td';
1199 } elseif ( $first_character === '!' ) {
1200 $last_tag = 'th';
1201 } elseif ( $first_character === '+' ) {
1202 $last_tag = 'caption';
1203 } else {
1204 $last_tag = '';
1205 }
1206
1207 array_push( $last_tag_history, $last_tag );
1208
1209 # A cell could contain both parameters and data
1210 $cell_data = explode( '|', $cell, 2 );
1211
1212 # T2553: Note that a '|' inside an invalid link should not
1213 # be mistaken as delimiting cell parameters
1214 # Bug T153140: Neither should language converter markup.
1215 if ( preg_match( '/\[\[|-\{/', $cell_data[0] ) === 1 ) {
1216 $cell = "{$previous}<{$last_tag}>{$cell}";
1217 } elseif ( count( $cell_data ) == 1 ) {
1218 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
1219 } else {
1220 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1221 $attributes = Sanitizer::fixTagAttributes( $attributes, $last_tag );
1222 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
1223 }
1224
1225 $outLine .= $cell;
1226 array_push( $td_history, true );
1227 }
1228 }
1229 $out .= $outLine . "\n";
1230 }
1231
1232 # Closing open td, tr && table
1233 while ( count( $td_history ) > 0 ) {
1234 if ( array_pop( $td_history ) ) {
1235 $out .= "</td>\n";
1236 }
1237 if ( array_pop( $tr_history ) ) {
1238 $out .= "</tr>\n";
1239 }
1240 if ( !array_pop( $has_opened_tr ) ) {
1241 $out .= "<tr><td></td></tr>\n";
1242 }
1243
1244 $out .= "</table>\n";
1245 }
1246
1247 # Remove trailing line-ending (b/c)
1248 if ( substr( $out, -1 ) === "\n" ) {
1249 $out = substr( $out, 0, -1 );
1250 }
1251
1252 # special case: don't return empty table
1253 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1254 $out = '';
1255 }
1256
1257 return $out;
1258 }
1259
1260 /**
1261 * Helper function for parse() that transforms wiki markup into half-parsed
1262 * HTML. Only called for $mOutputType == self::OT_HTML.
1263 *
1264 * @private
1265 *
1266 * @param string $text The text to parse
1267 * @param bool $isMain Whether this is being called from the main parse() function
1268 * @param PPFrame|bool $frame A pre-processor frame
1269 *
1270 * @return string
1271 */
1272 public function internalParse( $text, $isMain = true, $frame = false ) {
1273
1274 $origText = $text;
1275
1276 // Avoid PHP 7.1 warning from passing $this by reference
1277 $parser = $this;
1278
1279 # Hook to suspend the parser in this state
1280 if ( !Hooks::run( 'ParserBeforeInternalParse', [ &$parser, &$text, &$this->mStripState ] ) ) {
1281 return $text;
1282 }
1283
1284 # if $frame is provided, then use $frame for replacing any variables
1285 if ( $frame ) {
1286 # use frame depth to infer how include/noinclude tags should be handled
1287 # depth=0 means this is the top-level document; otherwise it's an included document
1288 if ( !$frame->depth ) {
1289 $flag = 0;
1290 } else {
1291 $flag = Parser::PTD_FOR_INCLUSION;
1292 }
1293 $dom = $this->preprocessToDom( $text, $flag );
1294 $text = $frame->expand( $dom );
1295 } else {
1296 # if $frame is not provided, then use old-style replaceVariables
1297 $text = $this->replaceVariables( $text );
1298 }
1299
1300 Hooks::run( 'InternalParseBeforeSanitize', [ &$parser, &$text, &$this->mStripState ] );
1301 $text = Sanitizer::removeHTMLtags(
1302 $text,
1303 [ $this, 'attributeStripCallback' ],
1304 false,
1305 array_keys( $this->mTransparentTagHooks ),
1306 [],
1307 [ $this, 'addTrackingCategory' ]
1308 );
1309 Hooks::run( 'InternalParseBeforeLinks', [ &$parser, &$text, &$this->mStripState ] );
1310
1311 # Tables need to come after variable replacement for things to work
1312 # properly; putting them before other transformations should keep
1313 # exciting things like link expansions from showing up in surprising
1314 # places.
1315 $text = $this->doTableStuff( $text );
1316
1317 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1318
1319 $text = $this->doDoubleUnderscore( $text );
1320
1321 $text = $this->doHeadings( $text );
1322 $text = $this->replaceInternalLinks( $text );
1323 $text = $this->doAllQuotes( $text );
1324 $text = $this->replaceExternalLinks( $text );
1325
1326 # replaceInternalLinks may sometimes leave behind
1327 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1328 $text = str_replace( self::MARKER_PREFIX . 'NOPARSE', '', $text );
1329
1330 $text = $this->doMagicLinks( $text );
1331 $text = $this->formatHeadings( $text, $origText, $isMain );
1332
1333 return $text;
1334 }
1335
1336 /**
1337 * Helper function for parse() that transforms half-parsed HTML into fully
1338 * parsed HTML.
1339 *
1340 * @param string $text
1341 * @param bool $isMain
1342 * @param bool $linestart
1343 * @return string
1344 */
1345 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1346 $text = $this->mStripState->unstripGeneral( $text );
1347
1348 // Avoid PHP 7.1 warning from passing $this by reference
1349 $parser = $this;
1350
1351 if ( $isMain ) {
1352 Hooks::run( 'ParserAfterUnstrip', [ &$parser, &$text ] );
1353 }
1354
1355 # Clean up special characters, only run once, next-to-last before doBlockLevels
1356 $fixtags = [
1357 # French spaces, last one Guillemet-left
1358 # only if there is something before the space
1359 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;',
1360 # french spaces, Guillemet-right
1361 '/(\\302\\253) /' => '\\1&#160;',
1362 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, T13874.
1363 ];
1364 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
1365
1366 $text = $this->doBlockLevels( $text, $linestart );
1367
1368 $this->replaceLinkHolders( $text );
1369
1370 /**
1371 * The input doesn't get language converted if
1372 * a) It's disabled
1373 * b) Content isn't converted
1374 * c) It's a conversion table
1375 * d) it is an interface message (which is in the user language)
1376 */
1377 if ( !( $this->mOptions->getDisableContentConversion()
1378 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) )
1379 ) {
1380 if ( !$this->mOptions->getInterfaceMessage() ) {
1381 # The position of the convert() call should not be changed. it
1382 # assumes that the links are all replaced and the only thing left
1383 # is the <nowiki> mark.
1384 $text = $this->getConverterLanguage()->convert( $text );
1385 }
1386 }
1387
1388 $text = $this->mStripState->unstripNoWiki( $text );
1389
1390 if ( $isMain ) {
1391 Hooks::run( 'ParserBeforeTidy', [ &$parser, &$text ] );
1392 }
1393
1394 $text = $this->replaceTransparentTags( $text );
1395 $text = $this->mStripState->unstripGeneral( $text );
1396
1397 $text = Sanitizer::normalizeCharReferences( $text );
1398
1399 if ( MWTidy::isEnabled() ) {
1400 if ( $this->mOptions->getTidy() ) {
1401 $text = MWTidy::tidy( $text );
1402 }
1403 } else {
1404 # attempt to sanitize at least some nesting problems
1405 # (T4702 and quite a few others)
1406 $tidyregs = [
1407 # ''Something [http://www.cool.com cool''] -->
1408 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1409 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1410 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1411 # fix up an anchor inside another anchor, only
1412 # at least for a single single nested link (T5695)
1413 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1414 '\\1\\2</a>\\3</a>\\1\\4</a>',
1415 # fix div inside inline elements- doBlockLevels won't wrap a line which
1416 # contains a div, so fix it up here; replace
1417 # div with escaped text
1418 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1419 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
1420 # remove empty italic or bold tag pairs, some
1421 # introduced by rules above
1422 '/<([bi])><\/\\1>/' => '',
1423 ];
1424
1425 $text = preg_replace(
1426 array_keys( $tidyregs ),
1427 array_values( $tidyregs ),
1428 $text );
1429 }
1430
1431 if ( $isMain ) {
1432 Hooks::run( 'ParserAfterTidy', [ &$parser, &$text ] );
1433 }
1434
1435 return $text;
1436 }
1437
1438 /**
1439 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1440 * magic external links.
1441 *
1442 * DML
1443 * @private
1444 *
1445 * @param string $text
1446 *
1447 * @return string
1448 */
1449 public function doMagicLinks( $text ) {
1450 $prots = wfUrlProtocolsWithoutProtRel();
1451 $urlChar = self::EXT_LINK_URL_CLASS;
1452 $addr = self::EXT_LINK_ADDR;
1453 $space = self::SPACE_NOT_NL; # non-newline space
1454 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1455 $spaces = "$space++"; # possessive match of 1 or more spaces
1456 $text = preg_replace_callback(
1457 '!(?: # Start cases
1458 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1459 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1460 (\b # m[3]: Free external links
1461 (?i:$prots)
1462 ($addr$urlChar*) # m[4]: Post-protocol path
1463 ) |
1464 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1465 ([0-9]+)\b |
1466 \bISBN $spaces ( # m[6]: ISBN, capture number
1467 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1468 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1469 [0-9Xx] # check digit
1470 )\b
1471 )!xu", [ $this, 'magicLinkCallback' ], $text );
1472 return $text;
1473 }
1474
1475 /**
1476 * @throws MWException
1477 * @param array $m
1478 * @return HTML|string
1479 */
1480 public function magicLinkCallback( $m ) {
1481 if ( isset( $m[1] ) && $m[1] !== '' ) {
1482 # Skip anchor
1483 return $m[0];
1484 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1485 # Skip HTML element
1486 return $m[0];
1487 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1488 # Free external link
1489 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1490 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1491 # RFC or PMID
1492 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1493 if ( !$this->mOptions->getMagicRFCLinks() ) {
1494 return $m[0];
1495 }
1496 $keyword = 'RFC';
1497 $urlmsg = 'rfcurl';
1498 $cssClass = 'mw-magiclink-rfc';
1499 $trackingCat = 'magiclink-tracking-rfc';
1500 $id = $m[5];
1501 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1502 if ( !$this->mOptions->getMagicPMIDLinks() ) {
1503 return $m[0];
1504 }
1505 $keyword = 'PMID';
1506 $urlmsg = 'pubmedurl';
1507 $cssClass = 'mw-magiclink-pmid';
1508 $trackingCat = 'magiclink-tracking-pmid';
1509 $id = $m[5];
1510 } else {
1511 throw new MWException( __METHOD__ . ': unrecognised match type "' .
1512 substr( $m[0], 0, 20 ) . '"' );
1513 }
1514 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1515 $this->addTrackingCategory( $trackingCat );
1516 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1517 } elseif ( isset( $m[6] ) && $m[6] !== ''
1518 && $this->mOptions->getMagicISBNLinks()
1519 ) {
1520 # ISBN
1521 $isbn = $m[6];
1522 $space = self::SPACE_NOT_NL; # non-newline space
1523 $isbn = preg_replace( "/$space/", ' ', $isbn );
1524 $num = strtr( $isbn, [
1525 '-' => '',
1526 ' ' => '',
1527 'x' => 'X',
1528 ] );
1529 $this->addTrackingCategory( 'magiclink-tracking-isbn' );
1530 return $this->getLinkRenderer()->makeKnownLink(
1531 SpecialPage::getTitleFor( 'Booksources', $num ),
1532 "ISBN $isbn",
1533 [
1534 'class' => 'internal mw-magiclink-isbn',
1535 'title' => false // suppress title attribute
1536 ]
1537 );
1538 } else {
1539 return $m[0];
1540 }
1541 }
1542
1543 /**
1544 * Make a free external link, given a user-supplied URL
1545 *
1546 * @param string $url
1547 * @param int $numPostProto
1548 * The number of characters after the protocol.
1549 * @return string HTML
1550 * @private
1551 */
1552 public function makeFreeExternalLink( $url, $numPostProto ) {
1553 $trail = '';
1554
1555 # The characters '<' and '>' (which were escaped by
1556 # removeHTMLtags()) should not be included in
1557 # URLs, per RFC 2396.
1558 # Make &nbsp; terminate a URL as well (bug T84937)
1559 $m2 = [];
1560 if ( preg_match(
1561 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1562 $url,
1563 $m2,
1564 PREG_OFFSET_CAPTURE
1565 ) ) {
1566 $trail = substr( $url, $m2[0][1] ) . $trail;
1567 $url = substr( $url, 0, $m2[0][1] );
1568 }
1569
1570 # Move trailing punctuation to $trail
1571 $sep = ',;\.:!?';
1572 # If there is no left bracket, then consider right brackets fair game too
1573 if ( strpos( $url, '(' ) === false ) {
1574 $sep .= ')';
1575 }
1576
1577 $urlRev = strrev( $url );
1578 $numSepChars = strspn( $urlRev, $sep );
1579 # Don't break a trailing HTML entity by moving the ; into $trail
1580 # This is in hot code, so use substr_compare to avoid having to
1581 # create a new string object for the comparison
1582 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1583 # more optimization: instead of running preg_match with a $
1584 # anchor, which can be slow, do the match on the reversed
1585 # string starting at the desired offset.
1586 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1587 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1588 $numSepChars--;
1589 }
1590 }
1591 if ( $numSepChars ) {
1592 $trail = substr( $url, -$numSepChars ) . $trail;
1593 $url = substr( $url, 0, -$numSepChars );
1594 }
1595
1596 # Verify that we still have a real URL after trail removal, and
1597 # not just lone protocol
1598 if ( strlen( $trail ) >= $numPostProto ) {
1599 return $url . $trail;
1600 }
1601
1602 $url = Sanitizer::cleanUrl( $url );
1603
1604 # Is this an external image?
1605 $text = $this->maybeMakeExternalImage( $url );
1606 if ( $text === false ) {
1607 # Not an image, make a link
1608 $text = Linker::makeExternalLink( $url,
1609 $this->getConverterLanguage()->markNoConversion( $url, true ),
1610 true, 'free',
1611 $this->getExternalLinkAttribs( $url ), $this->mTitle );
1612 # Register it in the output object...
1613 # Replace unnecessary URL escape codes with their equivalent characters
1614 $pasteurized = self::normalizeLinkUrl( $url );
1615 $this->mOutput->addExternalLink( $pasteurized );
1616 }
1617 return $text . $trail;
1618 }
1619
1620 /**
1621 * Parse headers and return html
1622 *
1623 * @private
1624 *
1625 * @param string $text
1626 *
1627 * @return string
1628 */
1629 public function doHeadings( $text ) {
1630 for ( $i = 6; $i >= 1; --$i ) {
1631 $h = str_repeat( '=', $i );
1632 $text = preg_replace( "/^$h(.+)$h\\s*$/m", "<h$i>\\1</h$i>", $text );
1633 }
1634 return $text;
1635 }
1636
1637 /**
1638 * Replace single quotes with HTML markup
1639 * @private
1640 *
1641 * @param string $text
1642 *
1643 * @return string The altered text
1644 */
1645 public function doAllQuotes( $text ) {
1646 $outtext = '';
1647 $lines = StringUtils::explode( "\n", $text );
1648 foreach ( $lines as $line ) {
1649 $outtext .= $this->doQuotes( $line ) . "\n";
1650 }
1651 $outtext = substr( $outtext, 0, -1 );
1652 return $outtext;
1653 }
1654
1655 /**
1656 * Helper function for doAllQuotes()
1657 *
1658 * @param string $text
1659 *
1660 * @return string
1661 */
1662 public function doQuotes( $text ) {
1663 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1664 $countarr = count( $arr );
1665 if ( $countarr == 1 ) {
1666 return $text;
1667 }
1668
1669 // First, do some preliminary work. This may shift some apostrophes from
1670 // being mark-up to being text. It also counts the number of occurrences
1671 // of bold and italics mark-ups.
1672 $numbold = 0;
1673 $numitalics = 0;
1674 for ( $i = 1; $i < $countarr; $i += 2 ) {
1675 $thislen = strlen( $arr[$i] );
1676 // If there are ever four apostrophes, assume the first is supposed to
1677 // be text, and the remaining three constitute mark-up for bold text.
1678 // (T15227: ''''foo'''' turns into ' ''' foo ' ''')
1679 if ( $thislen == 4 ) {
1680 $arr[$i - 1] .= "'";
1681 $arr[$i] = "'''";
1682 $thislen = 3;
1683 } elseif ( $thislen > 5 ) {
1684 // If there are more than 5 apostrophes in a row, assume they're all
1685 // text except for the last 5.
1686 // (T15227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1687 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1688 $arr[$i] = "'''''";
1689 $thislen = 5;
1690 }
1691 // Count the number of occurrences of bold and italics mark-ups.
1692 if ( $thislen == 2 ) {
1693 $numitalics++;
1694 } elseif ( $thislen == 3 ) {
1695 $numbold++;
1696 } elseif ( $thislen == 5 ) {
1697 $numitalics++;
1698 $numbold++;
1699 }
1700 }
1701
1702 // If there is an odd number of both bold and italics, it is likely
1703 // that one of the bold ones was meant to be an apostrophe followed
1704 // by italics. Which one we cannot know for certain, but it is more
1705 // likely to be one that has a single-letter word before it.
1706 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1707 $firstsingleletterword = -1;
1708 $firstmultiletterword = -1;
1709 $firstspace = -1;
1710 for ( $i = 1; $i < $countarr; $i += 2 ) {
1711 if ( strlen( $arr[$i] ) == 3 ) {
1712 $x1 = substr( $arr[$i - 1], -1 );
1713 $x2 = substr( $arr[$i - 1], -2, 1 );
1714 if ( $x1 === ' ' ) {
1715 if ( $firstspace == -1 ) {
1716 $firstspace = $i;
1717 }
1718 } elseif ( $x2 === ' ' ) {
1719 $firstsingleletterword = $i;
1720 // if $firstsingleletterword is set, we don't
1721 // look at the other options, so we can bail early.
1722 break;
1723 } else {
1724 if ( $firstmultiletterword == -1 ) {
1725 $firstmultiletterword = $i;
1726 }
1727 }
1728 }
1729 }
1730
1731 // If there is a single-letter word, use it!
1732 if ( $firstsingleletterword > -1 ) {
1733 $arr[$firstsingleletterword] = "''";
1734 $arr[$firstsingleletterword - 1] .= "'";
1735 } elseif ( $firstmultiletterword > -1 ) {
1736 // If not, but there's a multi-letter word, use that one.
1737 $arr[$firstmultiletterword] = "''";
1738 $arr[$firstmultiletterword - 1] .= "'";
1739 } elseif ( $firstspace > -1 ) {
1740 // ... otherwise use the first one that has neither.
1741 // (notice that it is possible for all three to be -1 if, for example,
1742 // there is only one pentuple-apostrophe in the line)
1743 $arr[$firstspace] = "''";
1744 $arr[$firstspace - 1] .= "'";
1745 }
1746 }
1747
1748 // Now let's actually convert our apostrophic mush to HTML!
1749 $output = '';
1750 $buffer = '';
1751 $state = '';
1752 $i = 0;
1753 foreach ( $arr as $r ) {
1754 if ( ( $i % 2 ) == 0 ) {
1755 if ( $state === 'both' ) {
1756 $buffer .= $r;
1757 } else {
1758 $output .= $r;
1759 }
1760 } else {
1761 $thislen = strlen( $r );
1762 if ( $thislen == 2 ) {
1763 if ( $state === 'i' ) {
1764 $output .= '</i>';
1765 $state = '';
1766 } elseif ( $state === 'bi' ) {
1767 $output .= '</i>';
1768 $state = 'b';
1769 } elseif ( $state === 'ib' ) {
1770 $output .= '</b></i><b>';
1771 $state = 'b';
1772 } elseif ( $state === 'both' ) {
1773 $output .= '<b><i>' . $buffer . '</i>';
1774 $state = 'b';
1775 } else { // $state can be 'b' or ''
1776 $output .= '<i>';
1777 $state .= 'i';
1778 }
1779 } elseif ( $thislen == 3 ) {
1780 if ( $state === 'b' ) {
1781 $output .= '</b>';
1782 $state = '';
1783 } elseif ( $state === 'bi' ) {
1784 $output .= '</i></b><i>';
1785 $state = 'i';
1786 } elseif ( $state === 'ib' ) {
1787 $output .= '</b>';
1788 $state = 'i';
1789 } elseif ( $state === 'both' ) {
1790 $output .= '<i><b>' . $buffer . '</b>';
1791 $state = 'i';
1792 } else { // $state can be 'i' or ''
1793 $output .= '<b>';
1794 $state .= 'b';
1795 }
1796 } elseif ( $thislen == 5 ) {
1797 if ( $state === 'b' ) {
1798 $output .= '</b><i>';
1799 $state = 'i';
1800 } elseif ( $state === 'i' ) {
1801 $output .= '</i><b>';
1802 $state = 'b';
1803 } elseif ( $state === 'bi' ) {
1804 $output .= '</i></b>';
1805 $state = '';
1806 } elseif ( $state === 'ib' ) {
1807 $output .= '</b></i>';
1808 $state = '';
1809 } elseif ( $state === 'both' ) {
1810 $output .= '<i><b>' . $buffer . '</b></i>';
1811 $state = '';
1812 } else { // ($state == '')
1813 $buffer = '';
1814 $state = 'both';
1815 }
1816 }
1817 }
1818 $i++;
1819 }
1820 // Now close all remaining tags. Notice that the order is important.
1821 if ( $state === 'b' || $state === 'ib' ) {
1822 $output .= '</b>';
1823 }
1824 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1825 $output .= '</i>';
1826 }
1827 if ( $state === 'bi' ) {
1828 $output .= '</b>';
1829 }
1830 // There might be lonely ''''', so make sure we have a buffer
1831 if ( $state === 'both' && $buffer ) {
1832 $output .= '<b><i>' . $buffer . '</i></b>';
1833 }
1834 return $output;
1835 }
1836
1837 /**
1838 * Replace external links (REL)
1839 *
1840 * Note: this is all very hackish and the order of execution matters a lot.
1841 * Make sure to run tests/parser/parserTests.php if you change this code.
1842 *
1843 * @private
1844 *
1845 * @param string $text
1846 *
1847 * @throws MWException
1848 * @return string
1849 */
1850 public function replaceExternalLinks( $text ) {
1851
1852 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1853 if ( $bits === false ) {
1854 throw new MWException( "PCRE needs to be compiled with "
1855 . "--enable-unicode-properties in order for MediaWiki to function" );
1856 }
1857 $s = array_shift( $bits );
1858
1859 $i = 0;
1860 while ( $i < count( $bits ) ) {
1861 $url = $bits[$i++];
1862 $i++; // protocol
1863 $text = $bits[$i++];
1864 $trail = $bits[$i++];
1865
1866 # The characters '<' and '>' (which were escaped by
1867 # removeHTMLtags()) should not be included in
1868 # URLs, per RFC 2396.
1869 $m2 = [];
1870 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1871 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1872 $url = substr( $url, 0, $m2[0][1] );
1873 }
1874
1875 # If the link text is an image URL, replace it with an <img> tag
1876 # This happened by accident in the original parser, but some people used it extensively
1877 $img = $this->maybeMakeExternalImage( $text );
1878 if ( $img !== false ) {
1879 $text = $img;
1880 }
1881
1882 $dtrail = '';
1883
1884 # Set linktype for CSS - if URL==text, link is essentially free
1885 $linktype = ( $text === $url ) ? 'free' : 'text';
1886
1887 # No link text, e.g. [http://domain.tld/some.link]
1888 if ( $text == '' ) {
1889 # Autonumber
1890 $langObj = $this->getTargetLanguage();
1891 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1892 $linktype = 'autonumber';
1893 } else {
1894 # Have link text, e.g. [http://domain.tld/some.link text]s
1895 # Check for trail
1896 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1897 }
1898
1899 $text = $this->getConverterLanguage()->markNoConversion( $text );
1900
1901 $url = Sanitizer::cleanUrl( $url );
1902
1903 # Use the encoded URL
1904 # This means that users can paste URLs directly into the text
1905 # Funny characters like ö aren't valid in URLs anyway
1906 # This was changed in August 2004
1907 $s .= Linker::makeExternalLink( $url, $text, false, $linktype,
1908 $this->getExternalLinkAttribs( $url ), $this->mTitle ) . $dtrail . $trail;
1909
1910 # Register link in the output object.
1911 # Replace unnecessary URL escape codes with the referenced character
1912 # This prevents spammers from hiding links from the filters
1913 $pasteurized = self::normalizeLinkUrl( $url );
1914 $this->mOutput->addExternalLink( $pasteurized );
1915 }
1916
1917 return $s;
1918 }
1919
1920 /**
1921 * Get the rel attribute for a particular external link.
1922 *
1923 * @since 1.21
1924 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1925 * nofollow if appropriate
1926 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1927 * @return string|null Rel attribute for $url
1928 */
1929 public static function getExternalLinkRel( $url = false, $title = null ) {
1930 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1931 $ns = $title ? $title->getNamespace() : false;
1932 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1933 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1934 ) {
1935 return 'nofollow';
1936 }
1937 return null;
1938 }
1939
1940 /**
1941 * Get an associative array of additional HTML attributes appropriate for a
1942 * particular external link. This currently may include rel => nofollow
1943 * (depending on configuration, namespace, and the URL's domain) and/or a
1944 * target attribute (depending on configuration).
1945 *
1946 * @param string $url URL to extract the domain from for rel =>
1947 * nofollow if appropriate
1948 * @return array Associative array of HTML attributes
1949 */
1950 public function getExternalLinkAttribs( $url ) {
1951 $attribs = [];
1952 $rel = self::getExternalLinkRel( $url, $this->mTitle );
1953
1954 $target = $this->mOptions->getExternalLinkTarget();
1955 if ( $target ) {
1956 $attribs['target'] = $target;
1957 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1958 // T133507. New windows can navigate parent cross-origin.
1959 // Including noreferrer due to lacking browser
1960 // support of noopener. Eventually noreferrer should be removed.
1961 if ( $rel !== '' ) {
1962 $rel .= ' ';
1963 }
1964 $rel .= 'noreferrer noopener';
1965 }
1966 }
1967 $attribs['rel'] = $rel;
1968 return $attribs;
1969 }
1970
1971 /**
1972 * Replace unusual escape codes in a URL with their equivalent characters
1973 *
1974 * This generally follows the syntax defined in RFC 3986, with special
1975 * consideration for HTTP query strings.
1976 *
1977 * @param string $url
1978 * @return string
1979 */
1980 public static function normalizeLinkUrl( $url ) {
1981 # First, make sure unsafe characters are encoded
1982 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
1983 function ( $m ) {
1984 return rawurlencode( $m[0] );
1985 },
1986 $url
1987 );
1988
1989 $ret = '';
1990 $end = strlen( $url );
1991
1992 # Fragment part - 'fragment'
1993 $start = strpos( $url, '#' );
1994 if ( $start !== false && $start < $end ) {
1995 $ret = self::normalizeUrlComponent(
1996 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
1997 $end = $start;
1998 }
1999
2000 # Query part - 'query' minus &=+;
2001 $start = strpos( $url, '?' );
2002 if ( $start !== false && $start < $end ) {
2003 $ret = self::normalizeUrlComponent(
2004 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
2005 $end = $start;
2006 }
2007
2008 # Scheme and path part - 'pchar'
2009 # (we assume no userinfo or encoded colons in the host)
2010 $ret = self::normalizeUrlComponent(
2011 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
2012
2013 return $ret;
2014 }
2015
2016 private static function normalizeUrlComponent( $component, $unsafe ) {
2017 $callback = function ( $matches ) use ( $unsafe ) {
2018 $char = urldecode( $matches[0] );
2019 $ord = ord( $char );
2020 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
2021 # Unescape it
2022 return $char;
2023 } else {
2024 # Leave it escaped, but use uppercase for a-f
2025 return strtoupper( $matches[0] );
2026 }
2027 };
2028 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
2029 }
2030
2031 /**
2032 * make an image if it's allowed, either through the global
2033 * option, through the exception, or through the on-wiki whitelist
2034 *
2035 * @param string $url
2036 *
2037 * @return string
2038 */
2039 private function maybeMakeExternalImage( $url ) {
2040 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
2041 $imagesexception = !empty( $imagesfrom );
2042 $text = false;
2043 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2044 if ( $imagesexception && is_array( $imagesfrom ) ) {
2045 $imagematch = false;
2046 foreach ( $imagesfrom as $match ) {
2047 if ( strpos( $url, $match ) === 0 ) {
2048 $imagematch = true;
2049 break;
2050 }
2051 }
2052 } elseif ( $imagesexception ) {
2053 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2054 } else {
2055 $imagematch = false;
2056 }
2057
2058 if ( $this->mOptions->getAllowExternalImages()
2059 || ( $imagesexception && $imagematch )
2060 ) {
2061 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
2062 # Image found
2063 $text = Linker::makeExternalImage( $url );
2064 }
2065 }
2066 if ( !$text && $this->mOptions->getEnableImageWhitelist()
2067 && preg_match( self::EXT_IMAGE_REGEX, $url )
2068 ) {
2069 $whitelist = explode(
2070 "\n",
2071 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2072 );
2073
2074 foreach ( $whitelist as $entry ) {
2075 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2076 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
2077 continue;
2078 }
2079 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2080 # Image matches a whitelist entry
2081 $text = Linker::makeExternalImage( $url );
2082 break;
2083 }
2084 }
2085 }
2086 return $text;
2087 }
2088
2089 /**
2090 * Process [[ ]] wikilinks
2091 *
2092 * @param string $s
2093 *
2094 * @return string Processed text
2095 *
2096 * @private
2097 */
2098 public function replaceInternalLinks( $s ) {
2099 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
2100 return $s;
2101 }
2102
2103 /**
2104 * Process [[ ]] wikilinks (RIL)
2105 * @param string $s
2106 * @throws MWException
2107 * @return LinkHolderArray
2108 *
2109 * @private
2110 */
2111 public function replaceInternalLinks2( &$s ) {
2112 global $wgExtraInterlanguageLinkPrefixes;
2113
2114 static $tc = false, $e1, $e1_img;
2115 # the % is needed to support urlencoded titles as well
2116 if ( !$tc ) {
2117 $tc = Title::legalChars() . '#%';
2118 # Match a link having the form [[namespace:link|alternate]]trail
2119 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2120 # Match cases where there is no "]]", which might still be images
2121 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2122 }
2123
2124 $holders = new LinkHolderArray( $this );
2125
2126 # split the entire text string on occurrences of [[
2127 $a = StringUtils::explode( '[[', ' ' . $s );
2128 # get the first element (all text up to first [[), and remove the space we added
2129 $s = $a->current();
2130 $a->next();
2131 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2132 $s = substr( $s, 1 );
2133
2134 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2135 $e2 = null;
2136 if ( $useLinkPrefixExtension ) {
2137 # Match the end of a line for a word that's not followed by whitespace,
2138 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2139 global $wgContLang;
2140 $charset = $wgContLang->linkPrefixCharset();
2141 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2142 }
2143
2144 if ( is_null( $this->mTitle ) ) {
2145 throw new MWException( __METHOD__ . ": \$this->mTitle is null\n" );
2146 }
2147 $nottalk = !$this->mTitle->isTalkPage();
2148
2149 if ( $useLinkPrefixExtension ) {
2150 $m = [];
2151 if ( preg_match( $e2, $s, $m ) ) {
2152 $first_prefix = $m[2];
2153 } else {
2154 $first_prefix = false;
2155 }
2156 } else {
2157 $prefix = '';
2158 }
2159
2160 $useSubpages = $this->areSubpagesAllowed();
2161
2162 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2163 # Loop for each link
2164 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2165 // @codingStandardsIgnoreEnd
2166
2167 # Check for excessive memory usage
2168 if ( $holders->isBig() ) {
2169 # Too big
2170 # Do the existence check, replace the link holders and clear the array
2171 $holders->replace( $s );
2172 $holders->clear();
2173 }
2174
2175 if ( $useLinkPrefixExtension ) {
2176 if ( preg_match( $e2, $s, $m ) ) {
2177 $prefix = $m[2];
2178 $s = $m[1];
2179 } else {
2180 $prefix = '';
2181 }
2182 # first link
2183 if ( $first_prefix ) {
2184 $prefix = $first_prefix;
2185 $first_prefix = false;
2186 }
2187 }
2188
2189 $might_be_img = false;
2190
2191 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2192 $text = $m[2];
2193 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2194 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2195 # the real problem is with the $e1 regex
2196 # See T1500.
2197 # Still some problems for cases where the ] is meant to be outside punctuation,
2198 # and no image is in sight. See T4095.
2199 if ( $text !== ''
2200 && substr( $m[3], 0, 1 ) === ']'
2201 && strpos( $text, '[' ) !== false
2202 ) {
2203 $text .= ']'; # so that replaceExternalLinks($text) works later
2204 $m[3] = substr( $m[3], 1 );
2205 }
2206 # fix up urlencoded title texts
2207 if ( strpos( $m[1], '%' ) !== false ) {
2208 # Should anchors '#' also be rejected?
2209 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2210 }
2211 $trail = $m[3];
2212 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2213 # Invalid, but might be an image with a link in its caption
2214 $might_be_img = true;
2215 $text = $m[2];
2216 if ( strpos( $m[1], '%' ) !== false ) {
2217 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2218 }
2219 $trail = "";
2220 } else { # Invalid form; output directly
2221 $s .= $prefix . '[[' . $line;
2222 continue;
2223 }
2224
2225 $origLink = ltrim( $m[1], ' ' );
2226
2227 # Don't allow internal links to pages containing
2228 # PROTO: where PROTO is a valid URL protocol; these
2229 # should be external links.
2230 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $origLink ) ) {
2231 $s .= $prefix . '[[' . $line;
2232 continue;
2233 }
2234
2235 # Make subpage if necessary
2236 if ( $useSubpages ) {
2237 $link = $this->maybeDoSubpageLink( $origLink, $text );
2238 } else {
2239 $link = $origLink;
2240 }
2241
2242 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2243 if ( !$noforce ) {
2244 # Strip off leading ':'
2245 $link = substr( $link, 1 );
2246 }
2247
2248 $unstrip = $this->mStripState->unstripNoWiki( $link );
2249 $nt = is_string( $unstrip ) ? Title::newFromText( $unstrip ) : null;
2250 if ( $nt === null ) {
2251 $s .= $prefix . '[[' . $line;
2252 continue;
2253 }
2254
2255 $ns = $nt->getNamespace();
2256 $iw = $nt->getInterwiki();
2257
2258 if ( $might_be_img ) { # if this is actually an invalid link
2259 if ( $ns == NS_FILE && $noforce ) { # but might be an image
2260 $found = false;
2261 while ( true ) {
2262 # look at the next 'line' to see if we can close it there
2263 $a->next();
2264 $next_line = $a->current();
2265 if ( $next_line === false || $next_line === null ) {
2266 break;
2267 }
2268 $m = explode( ']]', $next_line, 3 );
2269 if ( count( $m ) == 3 ) {
2270 # the first ]] closes the inner link, the second the image
2271 $found = true;
2272 $text .= "[[{$m[0]}]]{$m[1]}";
2273 $trail = $m[2];
2274 break;
2275 } elseif ( count( $m ) == 2 ) {
2276 # if there's exactly one ]] that's fine, we'll keep looking
2277 $text .= "[[{$m[0]}]]{$m[1]}";
2278 } else {
2279 # if $next_line is invalid too, we need look no further
2280 $text .= '[[' . $next_line;
2281 break;
2282 }
2283 }
2284 if ( !$found ) {
2285 # we couldn't find the end of this imageLink, so output it raw
2286 # but don't ignore what might be perfectly normal links in the text we've examined
2287 $holders->merge( $this->replaceInternalLinks2( $text ) );
2288 $s .= "{$prefix}[[$link|$text";
2289 # note: no $trail, because without an end, there *is* no trail
2290 continue;
2291 }
2292 } else { # it's not an image, so output it raw
2293 $s .= "{$prefix}[[$link|$text";
2294 # note: no $trail, because without an end, there *is* no trail
2295 continue;
2296 }
2297 }
2298
2299 $wasblank = ( $text == '' );
2300 if ( $wasblank ) {
2301 $text = $link;
2302 } else {
2303 # T6598 madness. Handle the quotes only if they come from the alternate part
2304 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2305 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2306 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2307 $text = $this->doQuotes( $text );
2308 }
2309
2310 # Link not escaped by : , create the various objects
2311 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2312 # Interwikis
2313 if (
2314 $iw && $this->mOptions->getInterwikiMagic() && $nottalk && (
2315 Language::fetchLanguageName( $iw, null, 'mw' ) ||
2316 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2317 )
2318 ) {
2319 # T26502: filter duplicates
2320 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
2321 $this->mLangLinkLanguages[$iw] = true;
2322 $this->mOutput->addLanguageLink( $nt->getFullText() );
2323 }
2324
2325 $s = rtrim( $s . $prefix );
2326 $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail;
2327 continue;
2328 }
2329
2330 if ( $ns == NS_FILE ) {
2331 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
2332 if ( $wasblank ) {
2333 # if no parameters were passed, $text
2334 # becomes something like "File:Foo.png",
2335 # which we don't want to pass on to the
2336 # image generator
2337 $text = '';
2338 } else {
2339 # recursively parse links inside the image caption
2340 # actually, this will parse them in any other parameters, too,
2341 # but it might be hard to fix that, and it doesn't matter ATM
2342 $text = $this->replaceExternalLinks( $text );
2343 $holders->merge( $this->replaceInternalLinks2( $text ) );
2344 }
2345 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2346 $s .= $prefix . $this->armorLinks(
2347 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2348 continue;
2349 }
2350 } elseif ( $ns == NS_CATEGORY ) {
2351 $s = rtrim( $s . "\n" ); # T2087
2352
2353 if ( $wasblank ) {
2354 $sortkey = $this->getDefaultSort();
2355 } else {
2356 $sortkey = $text;
2357 }
2358 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
2359 $sortkey = str_replace( "\n", '', $sortkey );
2360 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
2361 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
2362
2363 /**
2364 * Strip the whitespace Category links produce, see T2087
2365 */
2366 $s .= trim( $prefix . $trail, "\n" ) == '' ? '' : $prefix . $trail;
2367
2368 continue;
2369 }
2370 }
2371
2372 # Self-link checking. For some languages, variants of the title are checked in
2373 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2374 # for linking to a different variant.
2375 if ( $ns != NS_SPECIAL && $nt->equals( $this->mTitle ) && !$nt->hasFragment() ) {
2376 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail );
2377 continue;
2378 }
2379
2380 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2381 # @todo FIXME: Should do batch file existence checks, see comment below
2382 if ( $ns == NS_MEDIA ) {
2383 # Give extensions a chance to select the file revision for us
2384 $options = [];
2385 $descQuery = false;
2386 Hooks::run( 'BeforeParserFetchFileAndTitle',
2387 [ $this, $nt, &$options, &$descQuery ] );
2388 # Fetch and register the file (file title may be different via hooks)
2389 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2390 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2391 $s .= $prefix . $this->armorLinks(
2392 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2393 continue;
2394 }
2395
2396 # Some titles, such as valid special pages or files in foreign repos, should
2397 # be shown as bluelinks even though they're not included in the page table
2398 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2399 # batch file existence checks for NS_FILE and NS_MEDIA
2400 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2401 $this->mOutput->addLink( $nt );
2402 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2403 } else {
2404 # Links will be added to the output link list after checking
2405 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2406 }
2407 }
2408 return $holders;
2409 }
2410
2411 /**
2412 * Render a forced-blue link inline; protect against double expansion of
2413 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2414 * Since this little disaster has to split off the trail text to avoid
2415 * breaking URLs in the following text without breaking trails on the
2416 * wiki links, it's been made into a horrible function.
2417 *
2418 * @param Title $nt
2419 * @param string $text
2420 * @param string $trail
2421 * @param string $prefix
2422 * @return string HTML-wikitext mix oh yuck
2423 */
2424 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2425 list( $inside, $trail ) = Linker::splitTrail( $trail );
2426
2427 if ( $text == '' ) {
2428 $text = htmlspecialchars( $nt->getPrefixedText() );
2429 }
2430
2431 $link = $this->getLinkRenderer()->makeKnownLink(
2432 $nt, new HtmlArmor( "$prefix$text$inside" )
2433 );
2434
2435 return $this->armorLinks( $link ) . $trail;
2436 }
2437
2438 /**
2439 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2440 * going to go through further parsing steps before inline URL expansion.
2441 *
2442 * Not needed quite as much as it used to be since free links are a bit
2443 * more sensible these days. But bracketed links are still an issue.
2444 *
2445 * @param string $text More-or-less HTML
2446 * @return string Less-or-more HTML with NOPARSE bits
2447 */
2448 public function armorLinks( $text ) {
2449 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/',
2450 self::MARKER_PREFIX . "NOPARSE$1", $text );
2451 }
2452
2453 /**
2454 * Return true if subpage links should be expanded on this page.
2455 * @return bool
2456 */
2457 public function areSubpagesAllowed() {
2458 # Some namespaces don't allow subpages
2459 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2460 }
2461
2462 /**
2463 * Handle link to subpage if necessary
2464 *
2465 * @param string $target The source of the link
2466 * @param string &$text The link text, modified as necessary
2467 * @return string The full name of the link
2468 * @private
2469 */
2470 public function maybeDoSubpageLink( $target, &$text ) {
2471 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2472 }
2473
2474 /**
2475 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2476 *
2477 * @param string $text
2478 * @param bool $linestart Whether or not this is at the start of a line.
2479 * @private
2480 * @return string The lists rendered as HTML
2481 */
2482 public function doBlockLevels( $text, $linestart ) {
2483 return BlockLevelPass::doBlockLevels( $text, $linestart );
2484 }
2485
2486 /**
2487 * Return value of a magic variable (like PAGENAME)
2488 *
2489 * @private
2490 *
2491 * @param string $index Magic variable identifier as mapped in MagicWord::$mVariableIDs
2492 * @param bool|PPFrame $frame
2493 *
2494 * @throws MWException
2495 * @return string
2496 */
2497 public function getVariableValue( $index, $frame = false ) {
2498 global $wgContLang, $wgSitename, $wgServer, $wgServerName;
2499 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2500
2501 if ( is_null( $this->mTitle ) ) {
2502 // If no title set, bad things are going to happen
2503 // later. Title should always be set since this
2504 // should only be called in the middle of a parse
2505 // operation (but the unit-tests do funky stuff)
2506 throw new MWException( __METHOD__ . ' Should only be '
2507 . ' called while parsing (no title set)' );
2508 }
2509
2510 // Avoid PHP 7.1 warning from passing $this by reference
2511 $parser = $this;
2512
2513 /**
2514 * Some of these require message or data lookups and can be
2515 * expensive to check many times.
2516 */
2517 if ( Hooks::run( 'ParserGetVariableValueVarCache', [ &$parser, &$this->mVarCache ] ) ) {
2518 if ( isset( $this->mVarCache[$index] ) ) {
2519 return $this->mVarCache[$index];
2520 }
2521 }
2522
2523 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2524 Hooks::run( 'ParserGetVariableValueTs', [ &$parser, &$ts ] );
2525
2526 $pageLang = $this->getFunctionLang();
2527
2528 switch ( $index ) {
2529 case '!':
2530 $value = '|';
2531 break;
2532 case 'currentmonth':
2533 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'm' ) );
2534 break;
2535 case 'currentmonth1':
2536 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2537 break;
2538 case 'currentmonthname':
2539 $value = $pageLang->getMonthName( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2540 break;
2541 case 'currentmonthnamegen':
2542 $value = $pageLang->getMonthNameGen( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2543 break;
2544 case 'currentmonthabbrev':
2545 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2546 break;
2547 case 'currentday':
2548 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'j' ) );
2549 break;
2550 case 'currentday2':
2551 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'd' ) );
2552 break;
2553 case 'localmonth':
2554 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'm' ) );
2555 break;
2556 case 'localmonth1':
2557 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2558 break;
2559 case 'localmonthname':
2560 $value = $pageLang->getMonthName( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2561 break;
2562 case 'localmonthnamegen':
2563 $value = $pageLang->getMonthNameGen( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2564 break;
2565 case 'localmonthabbrev':
2566 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2567 break;
2568 case 'localday':
2569 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'j' ) );
2570 break;
2571 case 'localday2':
2572 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'd' ) );
2573 break;
2574 case 'pagename':
2575 $value = wfEscapeWikiText( $this->mTitle->getText() );
2576 break;
2577 case 'pagenamee':
2578 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() );
2579 break;
2580 case 'fullpagename':
2581 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2582 break;
2583 case 'fullpagenamee':
2584 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() );
2585 break;
2586 case 'subpagename':
2587 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2588 break;
2589 case 'subpagenamee':
2590 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() );
2591 break;
2592 case 'rootpagename':
2593 $value = wfEscapeWikiText( $this->mTitle->getRootText() );
2594 break;
2595 case 'rootpagenamee':
2596 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2597 ' ',
2598 '_',
2599 $this->mTitle->getRootText()
2600 ) ) );
2601 break;
2602 case 'basepagename':
2603 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2604 break;
2605 case 'basepagenamee':
2606 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2607 ' ',
2608 '_',
2609 $this->mTitle->getBaseText()
2610 ) ) );
2611 break;
2612 case 'talkpagename':
2613 if ( $this->mTitle->canTalk() ) {
2614 $talkPage = $this->mTitle->getTalkPage();
2615 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2616 } else {
2617 $value = '';
2618 }
2619 break;
2620 case 'talkpagenamee':
2621 if ( $this->mTitle->canTalk() ) {
2622 $talkPage = $this->mTitle->getTalkPage();
2623 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2624 } else {
2625 $value = '';
2626 }
2627 break;
2628 case 'subjectpagename':
2629 $subjPage = $this->mTitle->getSubjectPage();
2630 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2631 break;
2632 case 'subjectpagenamee':
2633 $subjPage = $this->mTitle->getSubjectPage();
2634 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2635 break;
2636 case 'pageid': // requested in T25427
2637 $pageid = $this->getTitle()->getArticleID();
2638 if ( $pageid == 0 ) {
2639 # 0 means the page doesn't exist in the database,
2640 # which means the user is previewing a new page.
2641 # The vary-revision flag must be set, because the magic word
2642 # will have a different value once the page is saved.
2643 $this->mOutput->setFlag( 'vary-revision' );
2644 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2645 }
2646 $value = $pageid ? $pageid : null;
2647 break;
2648 case 'revisionid':
2649 # Let the edit saving system know we should parse the page
2650 # *after* a revision ID has been assigned.
2651 $this->mOutput->setFlag( 'vary-revision-id' );
2652 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision-id...\n" );
2653 $value = $this->mRevisionId;
2654 if ( !$value && $this->mOptions->getSpeculativeRevIdCallback() ) {
2655 $value = call_user_func( $this->mOptions->getSpeculativeRevIdCallback() );
2656 $this->mOutput->setSpeculativeRevIdUsed( $value );
2657 }
2658 break;
2659 case 'revisionday':
2660 # Let the edit saving system know we should parse the page
2661 # *after* a revision ID has been assigned. This is for null edits.
2662 $this->mOutput->setFlag( 'vary-revision' );
2663 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2664 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2665 break;
2666 case 'revisionday2':
2667 # Let the edit saving system know we should parse the page
2668 # *after* a revision ID has been assigned. This is for null edits.
2669 $this->mOutput->setFlag( 'vary-revision' );
2670 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2671 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2672 break;
2673 case 'revisionmonth':
2674 # Let the edit saving system know we should parse the page
2675 # *after* a revision ID has been assigned. This is for null edits.
2676 $this->mOutput->setFlag( 'vary-revision' );
2677 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2678 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2679 break;
2680 case 'revisionmonth1':
2681 # Let the edit saving system know we should parse the page
2682 # *after* a revision ID has been assigned. This is for null edits.
2683 $this->mOutput->setFlag( 'vary-revision' );
2684 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2685 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2686 break;
2687 case 'revisionyear':
2688 # Let the edit saving system know we should parse the page
2689 # *after* a revision ID has been assigned. This is for null edits.
2690 $this->mOutput->setFlag( 'vary-revision' );
2691 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2692 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2693 break;
2694 case 'revisiontimestamp':
2695 # Let the edit saving system know we should parse the page
2696 # *after* a revision ID has been assigned. This is for null edits.
2697 $this->mOutput->setFlag( 'vary-revision' );
2698 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2699 $value = $this->getRevisionTimestamp();
2700 break;
2701 case 'revisionuser':
2702 # Let the edit saving system know we should parse the page
2703 # *after* a revision ID has been assigned for null edits.
2704 $this->mOutput->setFlag( 'vary-user' );
2705 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-user...\n" );
2706 $value = $this->getRevisionUser();
2707 break;
2708 case 'revisionsize':
2709 $value = $this->getRevisionSize();
2710 break;
2711 case 'namespace':
2712 $value = str_replace( '_', ' ', $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2713 break;
2714 case 'namespacee':
2715 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2716 break;
2717 case 'namespacenumber':
2718 $value = $this->mTitle->getNamespace();
2719 break;
2720 case 'talkspace':
2721 $value = $this->mTitle->canTalk()
2722 ? str_replace( '_', ' ', $this->mTitle->getTalkNsText() )
2723 : '';
2724 break;
2725 case 'talkspacee':
2726 $value = $this->mTitle->canTalk() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2727 break;
2728 case 'subjectspace':
2729 $value = str_replace( '_', ' ', $this->mTitle->getSubjectNsText() );
2730 break;
2731 case 'subjectspacee':
2732 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2733 break;
2734 case 'currentdayname':
2735 $value = $pageLang->getWeekdayName( (int)MWTimestamp::getInstance( $ts )->format( 'w' ) + 1 );
2736 break;
2737 case 'currentyear':
2738 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'Y' ), true );
2739 break;
2740 case 'currenttime':
2741 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2742 break;
2743 case 'currenthour':
2744 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'H' ), true );
2745 break;
2746 case 'currentweek':
2747 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2748 # int to remove the padding
2749 $value = $pageLang->formatNum( (int)MWTimestamp::getInstance( $ts )->format( 'W' ) );
2750 break;
2751 case 'currentdow':
2752 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'w' ) );
2753 break;
2754 case 'localdayname':
2755 $value = $pageLang->getWeekdayName(
2756 (int)MWTimestamp::getLocalInstance( $ts )->format( 'w' ) + 1
2757 );
2758 break;
2759 case 'localyear':
2760 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'Y' ), true );
2761 break;
2762 case 'localtime':
2763 $value = $pageLang->time(
2764 MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' ),
2765 false,
2766 false
2767 );
2768 break;
2769 case 'localhour':
2770 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'H' ), true );
2771 break;
2772 case 'localweek':
2773 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2774 # int to remove the padding
2775 $value = $pageLang->formatNum( (int)MWTimestamp::getLocalInstance( $ts )->format( 'W' ) );
2776 break;
2777 case 'localdow':
2778 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'w' ) );
2779 break;
2780 case 'numberofarticles':
2781 $value = $pageLang->formatNum( SiteStats::articles() );
2782 break;
2783 case 'numberoffiles':
2784 $value = $pageLang->formatNum( SiteStats::images() );
2785 break;
2786 case 'numberofusers':
2787 $value = $pageLang->formatNum( SiteStats::users() );
2788 break;
2789 case 'numberofactiveusers':
2790 $value = $pageLang->formatNum( SiteStats::activeUsers() );
2791 break;
2792 case 'numberofpages':
2793 $value = $pageLang->formatNum( SiteStats::pages() );
2794 break;
2795 case 'numberofadmins':
2796 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2797 break;
2798 case 'numberofedits':
2799 $value = $pageLang->formatNum( SiteStats::edits() );
2800 break;
2801 case 'currenttimestamp':
2802 $value = wfTimestamp( TS_MW, $ts );
2803 break;
2804 case 'localtimestamp':
2805 $value = MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' );
2806 break;
2807 case 'currentversion':
2808 $value = SpecialVersion::getVersion();
2809 break;
2810 case 'articlepath':
2811 return $wgArticlePath;
2812 case 'sitename':
2813 return $wgSitename;
2814 case 'server':
2815 return $wgServer;
2816 case 'servername':
2817 return $wgServerName;