Revert "Move NewPP limit report HTML comments to JS variables" and followups
[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 * $wgMaxArticleSize
67 *
68 * @ingroup Parser
69 */
70 class Parser {
71 /**
72 * Update this version number when the ParserOutput format
73 * changes in an incompatible way, so the parser cache
74 * can automatically discard old data.
75 */
76 const VERSION = '1.6.4';
77
78 /**
79 * Update this version number when the output of serialiseHalfParsedText()
80 * changes in an incompatible way
81 */
82 const HALF_PARSED_VERSION = 2;
83
84 # Flags for Parser::setFunctionHook
85 const SFH_NO_HASH = 1;
86 const SFH_OBJECT_ARGS = 2;
87
88 # Constants needed for external link processing
89 # Everything except bracket, space, or control characters
90 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
91 # as well as U+3000 is IDEOGRAPHIC SPACE for bug 19052
92 const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F\p{Zs}]';
93 # Simplified expression to match an IPv4 or IPv6 address, or
94 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
95 const EXT_LINK_ADDR = '(?:[0-9.]+|\\[(?i:[0-9a-f:.]+)\\]|[^][<>"\\x00-\\x20\\x7F\p{Zs}])';
96 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
97 // @codingStandardsIgnoreStart Generic.Files.LineLength
98 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}]+)
99 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
100 // @codingStandardsIgnoreEnd
101
102 # Regular expression for a non-newline space
103 const SPACE_NOT_NL = '(?:\t|&nbsp;|&\#0*160;|&\#[Xx]0*[Aa]0;|\p{Zs})';
104
105 # Flags for preprocessToDom
106 const PTD_FOR_INCLUSION = 1;
107
108 # Allowed values for $this->mOutputType
109 # Parameter to startExternalParse().
110 const OT_HTML = 1; # like parse()
111 const OT_WIKI = 2; # like preSaveTransform()
112 const OT_PREPROCESS = 3; # like preprocess()
113 const OT_MSG = 3;
114 const OT_PLAIN = 4; # like extractSections() - portions of the original are returned unchanged.
115
116 /**
117 * @var string Prefix and suffix for temporary replacement strings
118 * for the multipass parser.
119 *
120 * \x7f should never appear in input as it's disallowed in XML.
121 * Using it at the front also gives us a little extra robustness
122 * since it shouldn't match when butted up against identifier-like
123 * string constructs.
124 *
125 * Must not consist of all title characters, or else it will change
126 * the behavior of <nowiki> in a link.
127 *
128 * Must have a character that needs escaping in attributes, otherwise
129 * someone could put a strip marker in an attribute, to get around
130 * escaping quote marks, and break out of the attribute. Thus we add
131 * `'".
132 */
133 const MARKER_SUFFIX = "-QINU`\"'\x7f";
134 const MARKER_PREFIX = "\x7f'\"`UNIQ-";
135
136 # Markers used for wrapping the table of contents
137 const TOC_START = '<mw:toc>';
138 const TOC_END = '</mw:toc>';
139
140 # Persistent:
141 public $mTagHooks = [];
142 public $mTransparentTagHooks = [];
143 public $mFunctionHooks = [];
144 public $mFunctionSynonyms = [ 0 => [], 1 => [] ];
145 public $mFunctionTagHooks = [];
146 public $mStripList = [];
147 public $mDefaultStripList = [];
148 public $mVarCache = [];
149 public $mImageParams = [];
150 public $mImageParamsMagicArray = [];
151 public $mMarkerIndex = 0;
152 public $mFirstCall = true;
153
154 # Initialised by initialiseVariables()
155
156 /**
157 * @var MagicWordArray
158 */
159 public $mVariables;
160
161 /**
162 * @var MagicWordArray
163 */
164 public $mSubstWords;
165 # Initialised in constructor
166 public $mConf, $mExtLinkBracketedRegex, $mUrlProtocols;
167
168 # Initialized in getPreprocessor()
169 /** @var Preprocessor */
170 public $mPreprocessor;
171
172 # Cleared with clearState():
173 /**
174 * @var ParserOutput
175 */
176 public $mOutput;
177 public $mAutonumber;
178
179 /**
180 * @var StripState
181 */
182 public $mStripState;
183
184 public $mIncludeCount;
185 /**
186 * @var LinkHolderArray
187 */
188 public $mLinkHolders;
189
190 public $mLinkID;
191 public $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth;
192 public $mDefaultSort;
193 public $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
194 public $mExpensiveFunctionCount; # number of expensive parser function calls
195 public $mShowToc, $mForceTocPosition;
196
197 /**
198 * @var User
199 */
200 public $mUser; # User object; only used when doing pre-save transform
201
202 # Temporary
203 # These are variables reset at least once per parse regardless of $clearState
204
205 /**
206 * @var ParserOptions
207 */
208 public $mOptions;
209
210 /**
211 * @var Title
212 */
213 public $mTitle; # Title context, used for self-link rendering and similar things
214 public $mOutputType; # Output type, one of the OT_xxx constants
215 public $ot; # Shortcut alias, see setOutputType()
216 public $mRevisionObject; # The revision object of the specified revision ID
217 public $mRevisionId; # ID to display in {{REVISIONID}} tags
218 public $mRevisionTimestamp; # The timestamp of the specified revision ID
219 public $mRevisionUser; # User to display in {{REVISIONUSER}} tag
220 public $mRevisionSize; # Size to display in {{REVISIONSIZE}} variable
221 public $mRevIdForTs; # The revision ID which was used to fetch the timestamp
222 public $mInputSize = false; # For {{PAGESIZE}} on current page.
223
224 /**
225 * @var string Deprecated accessor for the strip marker prefix.
226 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
227 **/
228 public $mUniqPrefix = Parser::MARKER_PREFIX;
229
230 /**
231 * @var array Array with the language name of each language link (i.e. the
232 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
233 * duplicate language links to the ParserOutput.
234 */
235 public $mLangLinkLanguages;
236
237 /**
238 * @var MapCacheLRU|null
239 * @since 1.24
240 *
241 * A cache of the current revisions of titles. Keys are $title->getPrefixedDbKey()
242 */
243 public $currentRevisionCache;
244
245 /**
246 * @var bool Recursive call protection.
247 * This variable should be treated as if it were private.
248 */
249 public $mInParse = false;
250
251 /** @var SectionProfiler */
252 protected $mProfiler;
253
254 /**
255 * @var LinkRenderer
256 */
257 protected $mLinkRenderer;
258
259 /**
260 * @param array $conf
261 */
262 public function __construct( $conf = [] ) {
263 $this->mConf = $conf;
264 $this->mUrlProtocols = wfUrlProtocols();
265 $this->mExtLinkBracketedRegex = '/\[(((?i)' . $this->mUrlProtocols . ')' .
266 self::EXT_LINK_ADDR .
267 self::EXT_LINK_URL_CLASS . '*)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F]*?)\]/Su';
268 if ( isset( $conf['preprocessorClass'] ) ) {
269 $this->mPreprocessorClass = $conf['preprocessorClass'];
270 } elseif ( defined( 'HPHP_VERSION' ) ) {
271 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
272 $this->mPreprocessorClass = 'Preprocessor_Hash';
273 } elseif ( extension_loaded( 'domxml' ) ) {
274 # PECL extension that conflicts with the core DOM extension (bug 13770)
275 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
276 $this->mPreprocessorClass = 'Preprocessor_Hash';
277 } elseif ( extension_loaded( 'dom' ) ) {
278 $this->mPreprocessorClass = 'Preprocessor_DOM';
279 } else {
280 $this->mPreprocessorClass = 'Preprocessor_Hash';
281 }
282 wfDebug( __CLASS__ . ": using preprocessor: {$this->mPreprocessorClass}\n" );
283 }
284
285 /**
286 * Reduce memory usage to reduce the impact of circular references
287 */
288 public function __destruct() {
289 if ( isset( $this->mLinkHolders ) ) {
290 unset( $this->mLinkHolders );
291 }
292 foreach ( $this as $name => $value ) {
293 unset( $this->$name );
294 }
295 }
296
297 /**
298 * Allow extensions to clean up when the parser is cloned
299 */
300 public function __clone() {
301 $this->mInParse = false;
302
303 // Bug 56226: When you create a reference "to" an object field, that
304 // makes the object field itself be a reference too (until the other
305 // reference goes out of scope). When cloning, any field that's a
306 // reference is copied as a reference in the new object. Both of these
307 // are defined PHP5 behaviors, as inconvenient as it is for us when old
308 // hooks from PHP4 days are passing fields by reference.
309 foreach ( [ 'mStripState', 'mVarCache' ] as $k ) {
310 // Make a non-reference copy of the field, then rebind the field to
311 // reference the new copy.
312 $tmp = $this->$k;
313 $this->$k =& $tmp;
314 unset( $tmp );
315 }
316
317 Hooks::run( 'ParserCloned', [ $this ] );
318 }
319
320 /**
321 * Do various kinds of initialisation on the first call of the parser
322 */
323 public function firstCallInit() {
324 if ( !$this->mFirstCall ) {
325 return;
326 }
327 $this->mFirstCall = false;
328
329 CoreParserFunctions::register( $this );
330 CoreTagHooks::register( $this );
331 $this->initialiseVariables();
332
333 Hooks::run( 'ParserFirstCallInit', [ &$this ] );
334 }
335
336 /**
337 * Clear Parser state
338 *
339 * @private
340 */
341 public function clearState() {
342 if ( $this->mFirstCall ) {
343 $this->firstCallInit();
344 }
345 $this->mOutput = new ParserOutput;
346 $this->mOptions->registerWatcher( [ $this->mOutput, 'recordOption' ] );
347 $this->mAutonumber = 0;
348 $this->mIncludeCount = [];
349 $this->mLinkHolders = new LinkHolderArray( $this );
350 $this->mLinkID = 0;
351 $this->mRevisionObject = $this->mRevisionTimestamp =
352 $this->mRevisionId = $this->mRevisionUser = $this->mRevisionSize = null;
353 $this->mVarCache = [];
354 $this->mUser = null;
355 $this->mLangLinkLanguages = [];
356 $this->currentRevisionCache = null;
357
358 $this->mStripState = new StripState;
359
360 # Clear these on every parse, bug 4549
361 $this->mTplRedirCache = $this->mTplDomCache = [];
362
363 $this->mShowToc = true;
364 $this->mForceTocPosition = false;
365 $this->mIncludeSizes = [
366 'post-expand' => 0,
367 'arg' => 0,
368 ];
369 $this->mPPNodeCount = 0;
370 $this->mGeneratedPPNodeCount = 0;
371 $this->mHighestExpansionDepth = 0;
372 $this->mDefaultSort = false;
373 $this->mHeadings = [];
374 $this->mDoubleUnderscores = [];
375 $this->mExpensiveFunctionCount = 0;
376
377 # Fix cloning
378 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) {
379 $this->mPreprocessor = null;
380 }
381
382 $this->mProfiler = new SectionProfiler();
383
384 Hooks::run( 'ParserClearState', [ &$this ] );
385 }
386
387 /**
388 * Convert wikitext to HTML
389 * Do not call this function recursively.
390 *
391 * @param string $text Text we want to parse
392 * @param Title $title
393 * @param ParserOptions $options
394 * @param bool $linestart
395 * @param bool $clearState
396 * @param int $revid Number to pass in {{REVISIONID}}
397 * @return ParserOutput A ParserOutput
398 */
399 public function parse(
400 $text, Title $title, ParserOptions $options,
401 $linestart = true, $clearState = true, $revid = null
402 ) {
403 /**
404 * First pass--just handle <nowiki> sections, pass the rest off
405 * to internalParse() which does all the real work.
406 */
407
408 global $wgShowHostnames;
409
410 if ( $clearState ) {
411 // We use U+007F DELETE to construct strip markers, so we have to make
412 // sure that this character does not occur in the input text.
413 $text = strtr( $text, "\x7f", "?" );
414 $magicScopeVariable = $this->lock();
415 }
416
417 $this->startParse( $title, $options, self::OT_HTML, $clearState );
418
419 $this->currentRevisionCache = null;
420 $this->mInputSize = strlen( $text );
421 if ( $this->mOptions->getEnableLimitReport() ) {
422 $this->mOutput->resetParseStartTime();
423 }
424
425 $oldRevisionId = $this->mRevisionId;
426 $oldRevisionObject = $this->mRevisionObject;
427 $oldRevisionTimestamp = $this->mRevisionTimestamp;
428 $oldRevisionUser = $this->mRevisionUser;
429 $oldRevisionSize = $this->mRevisionSize;
430 if ( $revid !== null ) {
431 $this->mRevisionId = $revid;
432 $this->mRevisionObject = null;
433 $this->mRevisionTimestamp = null;
434 $this->mRevisionUser = null;
435 $this->mRevisionSize = null;
436 }
437
438 Hooks::run( 'ParserBeforeStrip', [ &$this, &$text, &$this->mStripState ] );
439 # No more strip!
440 Hooks::run( 'ParserAfterStrip', [ &$this, &$text, &$this->mStripState ] );
441 $text = $this->internalParse( $text );
442 Hooks::run( 'ParserAfterParse', [ &$this, &$text, &$this->mStripState ] );
443
444 $text = $this->internalParseHalfParsed( $text, true, $linestart );
445
446 /**
447 * A converted title will be provided in the output object if title and
448 * content conversion are enabled, the article text does not contain
449 * a conversion-suppressing double-underscore tag, and no
450 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
451 * automatic link conversion.
452 */
453 if ( !( $options->getDisableTitleConversion()
454 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
455 || isset( $this->mDoubleUnderscores['notitleconvert'] )
456 || $this->mOutput->getDisplayTitle() !== false )
457 ) {
458 $convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
459 if ( $convruletitle ) {
460 $this->mOutput->setTitleText( $convruletitle );
461 } else {
462 $titleText = $this->getConverterLanguage()->convertTitle( $title );
463 $this->mOutput->setTitleText( $titleText );
464 }
465 }
466
467 # Done parsing! Compute runtime adaptive expiry if set
468 $this->mOutput->finalizeAdaptiveCacheExpiry();
469
470 # Warn if too many heavyweight parser functions were used
471 if ( $this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit() ) {
472 $this->limitationWarn( 'expensive-parserfunction',
473 $this->mExpensiveFunctionCount,
474 $this->mOptions->getExpensiveParserFunctionLimit()
475 );
476 }
477
478 # Information on include size limits, for the benefit of users who try to skirt them
479 if ( $this->mOptions->getEnableLimitReport() ) {
480 $max = $this->mOptions->getMaxIncludeSize();
481
482 $cpuTime = $this->mOutput->getTimeSinceStart( 'cpu' );
483 if ( $cpuTime !== null ) {
484 $this->mOutput->setLimitReportData( 'limitreport-cputime',
485 sprintf( "%.3f", $cpuTime )
486 );
487 }
488
489 $wallTime = $this->mOutput->getTimeSinceStart( 'wall' );
490 $this->mOutput->setLimitReportData( 'limitreport-walltime',
491 sprintf( "%.3f", $wallTime )
492 );
493
494 $this->mOutput->setLimitReportData( 'limitreport-ppvisitednodes',
495 [ $this->mPPNodeCount, $this->mOptions->getMaxPPNodeCount() ]
496 );
497 $this->mOutput->setLimitReportData( 'limitreport-ppgeneratednodes',
498 [ $this->mGeneratedPPNodeCount, $this->mOptions->getMaxGeneratedPPNodeCount() ]
499 );
500 $this->mOutput->setLimitReportData( 'limitreport-postexpandincludesize',
501 [ $this->mIncludeSizes['post-expand'], $max ]
502 );
503 $this->mOutput->setLimitReportData( 'limitreport-templateargumentsize',
504 [ $this->mIncludeSizes['arg'], $max ]
505 );
506 $this->mOutput->setLimitReportData( 'limitreport-expansiondepth',
507 [ $this->mHighestExpansionDepth, $this->mOptions->getMaxPPExpandDepth() ]
508 );
509 $this->mOutput->setLimitReportData( 'limitreport-expensivefunctioncount',
510 [ $this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit() ]
511 );
512 Hooks::run( 'ParserLimitReportPrepare', [ $this, $this->mOutput ] );
513
514 $limitReport = "NewPP limit report\n";
515 if ( $wgShowHostnames ) {
516 $limitReport .= 'Parsed by ' . wfHostname() . "\n";
517 }
518 $limitReport .= 'Cached time: ' . $this->mOutput->getCacheTime() . "\n";
519 $limitReport .= 'Cache expiry: ' . $this->mOutput->getCacheExpiry() . "\n";
520 $limitReport .= 'Dynamic content: ' .
521 ( $this->mOutput->hasDynamicContent() ? 'true' : 'false' ) .
522 "\n";
523
524 foreach ( $this->mOutput->getLimitReportData() as $key => $value ) {
525 if ( Hooks::run( 'ParserLimitReportFormat',
526 [ $key, &$value, &$limitReport, false, false ]
527 ) ) {
528 $keyMsg = wfMessage( $key )->inLanguage( 'en' )->useDatabase( false );
529 $valueMsg = wfMessage( [ "$key-value-text", "$key-value" ] )
530 ->inLanguage( 'en' )->useDatabase( false );
531 if ( !$valueMsg->exists() ) {
532 $valueMsg = new RawMessage( '$1' );
533 }
534 if ( !$keyMsg->isDisabled() && !$valueMsg->isDisabled() ) {
535 $valueMsg->params( $value );
536 $limitReport .= "{$keyMsg->text()}: {$valueMsg->text()}\n";
537 }
538 }
539 }
540 // Since we're not really outputting HTML, decode the entities and
541 // then re-encode the things that need hiding inside HTML comments.
542 $limitReport = htmlspecialchars_decode( $limitReport );
543 Hooks::run( 'ParserLimitReport', [ $this, &$limitReport ] );
544
545 // Sanitize for comment. Note '‐' in the replacement is U+2010,
546 // which looks much like the problematic '-'.
547 $limitReport = str_replace( [ '-', '&' ], [ '‐', '&amp;' ], $limitReport );
548 $text .= "\n<!-- \n$limitReport-->\n";
549
550 // Add on template profiling data
551 $dataByFunc = $this->mProfiler->getFunctionStats();
552 uasort( $dataByFunc, function ( $a, $b ) {
553 return $a['real'] < $b['real']; // descending order
554 } );
555 $profileReport = "Transclusion expansion time report (%,ms,calls,template)\n";
556 foreach ( array_slice( $dataByFunc, 0, 10 ) as $item ) {
557 $profileReport .= sprintf( "%6.2f%% %8.3f %6d - %s\n",
558 $item['%real'], $item['real'], $item['calls'],
559 htmlspecialchars( $item['name'] ) );
560 }
561 $text .= "\n<!-- \n$profileReport-->\n";
562
563 if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) {
564 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' .
565 $this->mTitle->getPrefixedDBkey() );
566 }
567 }
568 $this->mOutput->setText( $text );
569
570 $this->mRevisionId = $oldRevisionId;
571 $this->mRevisionObject = $oldRevisionObject;
572 $this->mRevisionTimestamp = $oldRevisionTimestamp;
573 $this->mRevisionUser = $oldRevisionUser;
574 $this->mRevisionSize = $oldRevisionSize;
575 $this->mInputSize = false;
576 $this->currentRevisionCache = null;
577
578 return $this->mOutput;
579 }
580
581 /**
582 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
583 * can be called from an extension tag hook.
584 *
585 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
586 * instead, which means that lists and links have not been fully parsed yet,
587 * and strip markers are still present.
588 *
589 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
590 *
591 * Use this function if you're a parser tag hook and you want to parse
592 * wikitext before or after applying additional transformations, and you
593 * intend to *return the result as hook output*, which will cause it to go
594 * through the rest of parsing process automatically.
595 *
596 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
597 * $text are not expanded
598 *
599 * @param string $text Text extension wants to have parsed
600 * @param bool|PPFrame $frame The frame to use for expanding any template variables
601 * @return string UNSAFE half-parsed HTML
602 */
603 public function recursiveTagParse( $text, $frame = false ) {
604 Hooks::run( 'ParserBeforeStrip', [ &$this, &$text, &$this->mStripState ] );
605 Hooks::run( 'ParserAfterStrip', [ &$this, &$text, &$this->mStripState ] );
606 $text = $this->internalParse( $text, false, $frame );
607 return $text;
608 }
609
610 /**
611 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
612 * point can be called from an extension tag hook.
613 *
614 * The output of this function is fully-parsed HTML that is safe for output.
615 * If you're a parser tag hook, you might want to use recursiveTagParse()
616 * instead.
617 *
618 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
619 * $text are not expanded
620 *
621 * @since 1.25
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 Fully parsed HTML
626 */
627 public function recursiveTagParseFully( $text, $frame = false ) {
628 $text = $this->recursiveTagParse( $text, $frame );
629 $text = $this->internalParseHalfParsed( $text, false );
630 return $text;
631 }
632
633 /**
634 * Expand templates and variables in the text, producing valid, static wikitext.
635 * Also removes comments.
636 * Do not call this function recursively.
637 * @param string $text
638 * @param Title $title
639 * @param ParserOptions $options
640 * @param int|null $revid
641 * @param bool|PPFrame $frame
642 * @return mixed|string
643 */
644 public function preprocess( $text, Title $title = null,
645 ParserOptions $options, $revid = null, $frame = false
646 ) {
647 $magicScopeVariable = $this->lock();
648 $this->startParse( $title, $options, self::OT_PREPROCESS, true );
649 if ( $revid !== null ) {
650 $this->mRevisionId = $revid;
651 }
652 Hooks::run( 'ParserBeforeStrip', [ &$this, &$text, &$this->mStripState ] );
653 Hooks::run( 'ParserAfterStrip', [ &$this, &$text, &$this->mStripState ] );
654 $text = $this->replaceVariables( $text, $frame );
655 $text = $this->mStripState->unstripBoth( $text );
656 return $text;
657 }
658
659 /**
660 * Recursive parser entry point that can be called from an extension tag
661 * hook.
662 *
663 * @param string $text Text to be expanded
664 * @param bool|PPFrame $frame The frame to use for expanding any template variables
665 * @return string
666 * @since 1.19
667 */
668 public function recursivePreprocess( $text, $frame = false ) {
669 $text = $this->replaceVariables( $text, $frame );
670 $text = $this->mStripState->unstripBoth( $text );
671 return $text;
672 }
673
674 /**
675 * Process the wikitext for the "?preload=" feature. (bug 5210)
676 *
677 * "<noinclude>", "<includeonly>" etc. are parsed as for template
678 * transclusion, comments, templates, arguments, tags hooks and parser
679 * functions are untouched.
680 *
681 * @param string $text
682 * @param Title $title
683 * @param ParserOptions $options
684 * @param array $params
685 * @return string
686 */
687 public function getPreloadText( $text, Title $title, ParserOptions $options, $params = [] ) {
688 $msg = new RawMessage( $text );
689 $text = $msg->params( $params )->plain();
690
691 # Parser (re)initialisation
692 $magicScopeVariable = $this->lock();
693 $this->startParse( $title, $options, self::OT_PLAIN, true );
694
695 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES;
696 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
697 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
698 $text = $this->mStripState->unstripBoth( $text );
699 return $text;
700 }
701
702 /**
703 * Get a random string
704 *
705 * @return string
706 * @deprecated since 1.26; use wfRandomString() instead.
707 */
708 public static function getRandomString() {
709 wfDeprecated( __METHOD__, '1.26' );
710 return wfRandomString( 16 );
711 }
712
713 /**
714 * Set the current user.
715 * Should only be used when doing pre-save transform.
716 *
717 * @param User|null $user User object or null (to reset)
718 */
719 public function setUser( $user ) {
720 $this->mUser = $user;
721 }
722
723 /**
724 * Accessor for mUniqPrefix.
725 *
726 * @return string
727 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
728 */
729 public function uniqPrefix() {
730 wfDeprecated( __METHOD__, '1.26' );
731 return self::MARKER_PREFIX;
732 }
733
734 /**
735 * Set the context title
736 *
737 * @param Title $t
738 */
739 public function setTitle( $t ) {
740 if ( !$t ) {
741 $t = Title::newFromText( 'NO TITLE' );
742 }
743
744 if ( $t->hasFragment() ) {
745 # Strip the fragment to avoid various odd effects
746 $this->mTitle = $t->createFragmentTarget( '' );
747 } else {
748 $this->mTitle = $t;
749 }
750 }
751
752 /**
753 * Accessor for the Title object
754 *
755 * @return Title
756 */
757 public function getTitle() {
758 return $this->mTitle;
759 }
760
761 /**
762 * Accessor/mutator for the Title object
763 *
764 * @param Title $x Title object or null to just get the current one
765 * @return Title
766 */
767 public function Title( $x = null ) {
768 return wfSetVar( $this->mTitle, $x );
769 }
770
771 /**
772 * Set the output type
773 *
774 * @param int $ot New value
775 */
776 public function setOutputType( $ot ) {
777 $this->mOutputType = $ot;
778 # Shortcut alias
779 $this->ot = [
780 'html' => $ot == self::OT_HTML,
781 'wiki' => $ot == self::OT_WIKI,
782 'pre' => $ot == self::OT_PREPROCESS,
783 'plain' => $ot == self::OT_PLAIN,
784 ];
785 }
786
787 /**
788 * Accessor/mutator for the output type
789 *
790 * @param int|null $x New value or null to just get the current one
791 * @return int
792 */
793 public function OutputType( $x = null ) {
794 return wfSetVar( $this->mOutputType, $x );
795 }
796
797 /**
798 * Get the ParserOutput object
799 *
800 * @return ParserOutput
801 */
802 public function getOutput() {
803 return $this->mOutput;
804 }
805
806 /**
807 * Get the ParserOptions object
808 *
809 * @return ParserOptions
810 */
811 public function getOptions() {
812 return $this->mOptions;
813 }
814
815 /**
816 * Accessor/mutator for the ParserOptions object
817 *
818 * @param ParserOptions $x New value or null to just get the current one
819 * @return ParserOptions Current ParserOptions object
820 */
821 public function Options( $x = null ) {
822 return wfSetVar( $this->mOptions, $x );
823 }
824
825 /**
826 * @return int
827 */
828 public function nextLinkID() {
829 return $this->mLinkID++;
830 }
831
832 /**
833 * @param int $id
834 */
835 public function setLinkID( $id ) {
836 $this->mLinkID = $id;
837 }
838
839 /**
840 * Get a language object for use in parser functions such as {{FORMATNUM:}}
841 * @return Language
842 */
843 public function getFunctionLang() {
844 return $this->getTargetLanguage();
845 }
846
847 /**
848 * Get the target language for the content being parsed. This is usually the
849 * language that the content is in.
850 *
851 * @since 1.19
852 *
853 * @throws MWException
854 * @return Language
855 */
856 public function getTargetLanguage() {
857 $target = $this->mOptions->getTargetLanguage();
858
859 if ( $target !== null ) {
860 return $target;
861 } elseif ( $this->mOptions->getInterfaceMessage() ) {
862 return $this->mOptions->getUserLangObj();
863 } elseif ( is_null( $this->mTitle ) ) {
864 throw new MWException( __METHOD__ . ': $this->mTitle is null' );
865 }
866
867 return $this->mTitle->getPageLanguage();
868 }
869
870 /**
871 * Get the language object for language conversion
872 * @return Language|null
873 */
874 public function getConverterLanguage() {
875 return $this->getTargetLanguage();
876 }
877
878 /**
879 * Get a User object either from $this->mUser, if set, or from the
880 * ParserOptions object otherwise
881 *
882 * @return User
883 */
884 public function getUser() {
885 if ( !is_null( $this->mUser ) ) {
886 return $this->mUser;
887 }
888 return $this->mOptions->getUser();
889 }
890
891 /**
892 * Get a preprocessor object
893 *
894 * @return Preprocessor
895 */
896 public function getPreprocessor() {
897 if ( !isset( $this->mPreprocessor ) ) {
898 $class = $this->mPreprocessorClass;
899 $this->mPreprocessor = new $class( $this );
900 }
901 return $this->mPreprocessor;
902 }
903
904 /**
905 * Get a LinkRenderer instance to make links with
906 *
907 * @since 1.28
908 * @return LinkRenderer
909 */
910 public function getLinkRenderer() {
911 if ( !$this->mLinkRenderer ) {
912 $this->mLinkRenderer = MediaWikiServices::getInstance()
913 ->getLinkRendererFactory()->create();
914 $this->mLinkRenderer->setStubThreshold(
915 $this->getOptions()->getStubThreshold()
916 );
917 }
918
919 return $this->mLinkRenderer;
920 }
921
922 /**
923 * Replaces all occurrences of HTML-style comments and the given tags
924 * in the text with a random marker and returns the next text. The output
925 * parameter $matches will be an associative array filled with data in
926 * the form:
927 *
928 * @code
929 * 'UNIQ-xxxxx' => [
930 * 'element',
931 * 'tag content',
932 * [ 'param' => 'x' ],
933 * '<element param="x">tag content</element>' ]
934 * @endcode
935 *
936 * @param array $elements List of element names. Comments are always extracted.
937 * @param string $text Source text string.
938 * @param array $matches Out parameter, Array: extracted tags
939 * @param string|null $uniq_prefix
940 * @return string Stripped text
941 * @since 1.26 The uniq_prefix argument is deprecated.
942 */
943 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = null ) {
944 if ( $uniq_prefix !== null ) {
945 wfDeprecated( __METHOD__ . ' called with $prefix argument', '1.26' );
946 }
947 static $n = 1;
948 $stripped = '';
949 $matches = [];
950
951 $taglist = implode( '|', $elements );
952 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
953
954 while ( $text != '' ) {
955 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
956 $stripped .= $p[0];
957 if ( count( $p ) < 5 ) {
958 break;
959 }
960 if ( count( $p ) > 5 ) {
961 # comment
962 $element = $p[4];
963 $attributes = '';
964 $close = '';
965 $inside = $p[5];
966 } else {
967 # tag
968 $element = $p[1];
969 $attributes = $p[2];
970 $close = $p[3];
971 $inside = $p[4];
972 }
973
974 $marker = self::MARKER_PREFIX . "-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
975 $stripped .= $marker;
976
977 if ( $close === '/>' ) {
978 # Empty element tag, <tag />
979 $content = null;
980 $text = $inside;
981 $tail = null;
982 } else {
983 if ( $element === '!--' ) {
984 $end = '/(-->)/';
985 } else {
986 $end = "/(<\\/$element\\s*>)/i";
987 }
988 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
989 $content = $q[0];
990 if ( count( $q ) < 3 ) {
991 # No end tag -- let it run out to the end of the text.
992 $tail = '';
993 $text = '';
994 } else {
995 $tail = $q[1];
996 $text = $q[2];
997 }
998 }
999
1000 $matches[$marker] = [ $element,
1001 $content,
1002 Sanitizer::decodeTagAttributes( $attributes ),
1003 "<$element$attributes$close$content$tail" ];
1004 }
1005 return $stripped;
1006 }
1007
1008 /**
1009 * Get a list of strippable XML-like elements
1010 *
1011 * @return array
1012 */
1013 public function getStripList() {
1014 return $this->mStripList;
1015 }
1016
1017 /**
1018 * Add an item to the strip state
1019 * Returns the unique tag which must be inserted into the stripped text
1020 * The tag will be replaced with the original text in unstrip()
1021 *
1022 * @param string $text
1023 *
1024 * @return string
1025 */
1026 public function insertStripItem( $text ) {
1027 $marker = self::MARKER_PREFIX . "-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
1028 $this->mMarkerIndex++;
1029 $this->mStripState->addGeneral( $marker, $text );
1030 return $marker;
1031 }
1032
1033 /**
1034 * parse the wiki syntax used to render tables
1035 *
1036 * @private
1037 * @param string $text
1038 * @return string
1039 */
1040 public function doTableStuff( $text ) {
1041
1042 $lines = StringUtils::explode( "\n", $text );
1043 $out = '';
1044 $td_history = []; # Is currently a td tag open?
1045 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1046 $tr_history = []; # Is currently a tr tag open?
1047 $tr_attributes = []; # history of tr attributes
1048 $has_opened_tr = []; # Did this table open a <tr> element?
1049 $indent_level = 0; # indent level of the table
1050
1051 foreach ( $lines as $outLine ) {
1052 $line = trim( $outLine );
1053
1054 if ( $line === '' ) { # empty line, go to next line
1055 $out .= $outLine . "\n";
1056 continue;
1057 }
1058
1059 $first_character = $line[0];
1060 $first_two = substr( $line, 0, 2 );
1061 $matches = [];
1062
1063 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1064 # First check if we are starting a new table
1065 $indent_level = strlen( $matches[1] );
1066
1067 $attributes = $this->mStripState->unstripBoth( $matches[2] );
1068 $attributes = Sanitizer::fixTagAttributes( $attributes, 'table' );
1069
1070 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1071 array_push( $td_history, false );
1072 array_push( $last_tag_history, '' );
1073 array_push( $tr_history, false );
1074 array_push( $tr_attributes, '' );
1075 array_push( $has_opened_tr, false );
1076 } elseif ( count( $td_history ) == 0 ) {
1077 # Don't do any of the following
1078 $out .= $outLine . "\n";
1079 continue;
1080 } elseif ( $first_two === '|}' ) {
1081 # We are ending a table
1082 $line = '</table>' . substr( $line, 2 );
1083 $last_tag = array_pop( $last_tag_history );
1084
1085 if ( !array_pop( $has_opened_tr ) ) {
1086 $line = "<tr><td></td></tr>{$line}";
1087 }
1088
1089 if ( array_pop( $tr_history ) ) {
1090 $line = "</tr>{$line}";
1091 }
1092
1093 if ( array_pop( $td_history ) ) {
1094 $line = "</{$last_tag}>{$line}";
1095 }
1096 array_pop( $tr_attributes );
1097 $outLine = $line . str_repeat( '</dd></dl>', $indent_level );
1098 } elseif ( $first_two === '|-' ) {
1099 # Now we have a table row
1100 $line = preg_replace( '#^\|-+#', '', $line );
1101
1102 # Whats after the tag is now only attributes
1103 $attributes = $this->mStripState->unstripBoth( $line );
1104 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
1105 array_pop( $tr_attributes );
1106 array_push( $tr_attributes, $attributes );
1107
1108 $line = '';
1109 $last_tag = array_pop( $last_tag_history );
1110 array_pop( $has_opened_tr );
1111 array_push( $has_opened_tr, true );
1112
1113 if ( array_pop( $tr_history ) ) {
1114 $line = '</tr>';
1115 }
1116
1117 if ( array_pop( $td_history ) ) {
1118 $line = "</{$last_tag}>{$line}";
1119 }
1120
1121 $outLine = $line;
1122 array_push( $tr_history, false );
1123 array_push( $td_history, false );
1124 array_push( $last_tag_history, '' );
1125 } elseif ( $first_character === '|'
1126 || $first_character === '!'
1127 || $first_two === '|+'
1128 ) {
1129 # This might be cell elements, td, th or captions
1130 if ( $first_two === '|+' ) {
1131 $first_character = '+';
1132 $line = substr( $line, 2 );
1133 } else {
1134 $line = substr( $line, 1 );
1135 }
1136
1137 // Implies both are valid for table headings.
1138 if ( $first_character === '!' ) {
1139 $line = StringUtils::replaceMarkup( '!!', '||', $line );
1140 }
1141
1142 # Split up multiple cells on the same line.
1143 # FIXME : This can result in improper nesting of tags processed
1144 # by earlier parser steps.
1145 $cells = explode( '||', $line );
1146
1147 $outLine = '';
1148
1149 # Loop through each table cell
1150 foreach ( $cells as $cell ) {
1151 $previous = '';
1152 if ( $first_character !== '+' ) {
1153 $tr_after = array_pop( $tr_attributes );
1154 if ( !array_pop( $tr_history ) ) {
1155 $previous = "<tr{$tr_after}>\n";
1156 }
1157 array_push( $tr_history, true );
1158 array_push( $tr_attributes, '' );
1159 array_pop( $has_opened_tr );
1160 array_push( $has_opened_tr, true );
1161 }
1162
1163 $last_tag = array_pop( $last_tag_history );
1164
1165 if ( array_pop( $td_history ) ) {
1166 $previous = "</{$last_tag}>\n{$previous}";
1167 }
1168
1169 if ( $first_character === '|' ) {
1170 $last_tag = 'td';
1171 } elseif ( $first_character === '!' ) {
1172 $last_tag = 'th';
1173 } elseif ( $first_character === '+' ) {
1174 $last_tag = 'caption';
1175 } else {
1176 $last_tag = '';
1177 }
1178
1179 array_push( $last_tag_history, $last_tag );
1180
1181 # A cell could contain both parameters and data
1182 $cell_data = explode( '|', $cell, 2 );
1183
1184 # Bug 553: Note that a '|' inside an invalid link should not
1185 # be mistaken as delimiting cell parameters
1186 if ( strpos( $cell_data[0], '[[' ) !== false ) {
1187 $cell = "{$previous}<{$last_tag}>{$cell}";
1188 } elseif ( count( $cell_data ) == 1 ) {
1189 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
1190 } else {
1191 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1192 $attributes = Sanitizer::fixTagAttributes( $attributes, $last_tag );
1193 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
1194 }
1195
1196 $outLine .= $cell;
1197 array_push( $td_history, true );
1198 }
1199 }
1200 $out .= $outLine . "\n";
1201 }
1202
1203 # Closing open td, tr && table
1204 while ( count( $td_history ) > 0 ) {
1205 if ( array_pop( $td_history ) ) {
1206 $out .= "</td>\n";
1207 }
1208 if ( array_pop( $tr_history ) ) {
1209 $out .= "</tr>\n";
1210 }
1211 if ( !array_pop( $has_opened_tr ) ) {
1212 $out .= "<tr><td></td></tr>\n";
1213 }
1214
1215 $out .= "</table>\n";
1216 }
1217
1218 # Remove trailing line-ending (b/c)
1219 if ( substr( $out, -1 ) === "\n" ) {
1220 $out = substr( $out, 0, -1 );
1221 }
1222
1223 # special case: don't return empty table
1224 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1225 $out = '';
1226 }
1227
1228 return $out;
1229 }
1230
1231 /**
1232 * Helper function for parse() that transforms wiki markup into half-parsed
1233 * HTML. Only called for $mOutputType == self::OT_HTML.
1234 *
1235 * @private
1236 *
1237 * @param string $text The text to parse
1238 * @param bool $isMain Whether this is being called from the main parse() function
1239 * @param PPFrame|bool $frame A pre-processor frame
1240 *
1241 * @return string
1242 */
1243 public function internalParse( $text, $isMain = true, $frame = false ) {
1244
1245 $origText = $text;
1246
1247 # Hook to suspend the parser in this state
1248 if ( !Hooks::run( 'ParserBeforeInternalParse', [ &$this, &$text, &$this->mStripState ] ) ) {
1249 return $text;
1250 }
1251
1252 # if $frame is provided, then use $frame for replacing any variables
1253 if ( $frame ) {
1254 # use frame depth to infer how include/noinclude tags should be handled
1255 # depth=0 means this is the top-level document; otherwise it's an included document
1256 if ( !$frame->depth ) {
1257 $flag = 0;
1258 } else {
1259 $flag = Parser::PTD_FOR_INCLUSION;
1260 }
1261 $dom = $this->preprocessToDom( $text, $flag );
1262 $text = $frame->expand( $dom );
1263 } else {
1264 # if $frame is not provided, then use old-style replaceVariables
1265 $text = $this->replaceVariables( $text );
1266 }
1267
1268 Hooks::run( 'InternalParseBeforeSanitize', [ &$this, &$text, &$this->mStripState ] );
1269 $text = Sanitizer::removeHTMLtags(
1270 $text,
1271 [ &$this, 'attributeStripCallback' ],
1272 false,
1273 array_keys( $this->mTransparentTagHooks ),
1274 [],
1275 [ &$this, 'addTrackingCategory' ]
1276 );
1277 Hooks::run( 'InternalParseBeforeLinks', [ &$this, &$text, &$this->mStripState ] );
1278
1279 # Tables need to come after variable replacement for things to work
1280 # properly; putting them before other transformations should keep
1281 # exciting things like link expansions from showing up in surprising
1282 # places.
1283 $text = $this->doTableStuff( $text );
1284
1285 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1286
1287 $text = $this->doDoubleUnderscore( $text );
1288
1289 $text = $this->doHeadings( $text );
1290 $text = $this->replaceInternalLinks( $text );
1291 $text = $this->doAllQuotes( $text );
1292 $text = $this->replaceExternalLinks( $text );
1293
1294 # replaceInternalLinks may sometimes leave behind
1295 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1296 $text = str_replace( self::MARKER_PREFIX . 'NOPARSE', '', $text );
1297
1298 $text = $this->doMagicLinks( $text );
1299 $text = $this->formatHeadings( $text, $origText, $isMain );
1300
1301 return $text;
1302 }
1303
1304 /**
1305 * Helper function for parse() that transforms half-parsed HTML into fully
1306 * parsed HTML.
1307 *
1308 * @param string $text
1309 * @param bool $isMain
1310 * @param bool $linestart
1311 * @return string
1312 */
1313 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1314 $text = $this->mStripState->unstripGeneral( $text );
1315
1316 if ( $isMain ) {
1317 Hooks::run( 'ParserAfterUnstrip', [ &$this, &$text ] );
1318 }
1319
1320 # Clean up special characters, only run once, next-to-last before doBlockLevels
1321 $fixtags = [
1322 # french spaces, last one Guillemet-left
1323 # only if there is something before the space
1324 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;',
1325 # french spaces, Guillemet-right
1326 '/(\\302\\253) /' => '\\1&#160;',
1327 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, bug #11874.
1328 ];
1329 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
1330
1331 $text = $this->doBlockLevels( $text, $linestart );
1332
1333 $this->replaceLinkHolders( $text );
1334
1335 /**
1336 * The input doesn't get language converted if
1337 * a) It's disabled
1338 * b) Content isn't converted
1339 * c) It's a conversion table
1340 * d) it is an interface message (which is in the user language)
1341 */
1342 if ( !( $this->mOptions->getDisableContentConversion()
1343 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) )
1344 ) {
1345 if ( !$this->mOptions->getInterfaceMessage() ) {
1346 # The position of the convert() call should not be changed. it
1347 # assumes that the links are all replaced and the only thing left
1348 # is the <nowiki> mark.
1349 $text = $this->getConverterLanguage()->convert( $text );
1350 }
1351 }
1352
1353 $text = $this->mStripState->unstripNoWiki( $text );
1354
1355 if ( $isMain ) {
1356 Hooks::run( 'ParserBeforeTidy', [ &$this, &$text ] );
1357 }
1358
1359 $text = $this->replaceTransparentTags( $text );
1360 $text = $this->mStripState->unstripGeneral( $text );
1361
1362 $text = Sanitizer::normalizeCharReferences( $text );
1363
1364 if ( MWTidy::isEnabled() ) {
1365 if ( $this->mOptions->getTidy() ) {
1366 $text = MWTidy::tidy( $text );
1367 }
1368 } else {
1369 # attempt to sanitize at least some nesting problems
1370 # (bug #2702 and quite a few others)
1371 $tidyregs = [
1372 # ''Something [http://www.cool.com cool''] -->
1373 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1374 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1375 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1376 # fix up an anchor inside another anchor, only
1377 # at least for a single single nested link (bug 3695)
1378 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1379 '\\1\\2</a>\\3</a>\\1\\4</a>',
1380 # fix div inside inline elements- doBlockLevels won't wrap a line which
1381 # contains a div, so fix it up here; replace
1382 # div with escaped text
1383 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1384 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
1385 # remove empty italic or bold tag pairs, some
1386 # introduced by rules above
1387 '/<([bi])><\/\\1>/' => '',
1388 ];
1389
1390 $text = preg_replace(
1391 array_keys( $tidyregs ),
1392 array_values( $tidyregs ),
1393 $text );
1394 }
1395
1396 if ( $isMain ) {
1397 Hooks::run( 'ParserAfterTidy', [ &$this, &$text ] );
1398 }
1399
1400 return $text;
1401 }
1402
1403 /**
1404 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1405 * magic external links.
1406 *
1407 * DML
1408 * @private
1409 *
1410 * @param string $text
1411 *
1412 * @return string
1413 */
1414 public function doMagicLinks( $text ) {
1415 $prots = wfUrlProtocolsWithoutProtRel();
1416 $urlChar = self::EXT_LINK_URL_CLASS;
1417 $addr = self::EXT_LINK_ADDR;
1418 $space = self::SPACE_NOT_NL; # non-newline space
1419 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1420 $spaces = "$space++"; # possessive match of 1 or more spaces
1421 $text = preg_replace_callback(
1422 '!(?: # Start cases
1423 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1424 (<.*?>) | # m[2]: Skip stuff inside
1425 # HTML elements' . "
1426 (\b(?i:$prots)($addr$urlChar*)) | # m[3]: Free external links
1427 # m[4]: Post-protocol path
1428 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1429 ([0-9]+)\b |
1430 \bISBN $spaces ( # m[6]: ISBN, capture number
1431 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1432 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1433 [0-9Xx] # check digit
1434 )\b
1435 )!xu", [ &$this, 'magicLinkCallback' ], $text );
1436 return $text;
1437 }
1438
1439 /**
1440 * @throws MWException
1441 * @param array $m
1442 * @return HTML|string
1443 */
1444 public function magicLinkCallback( $m ) {
1445 if ( isset( $m[1] ) && $m[1] !== '' ) {
1446 # Skip anchor
1447 return $m[0];
1448 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1449 # Skip HTML element
1450 return $m[0];
1451 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1452 # Free external link
1453 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1454 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1455 # RFC or PMID
1456 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1457 if ( !$this->mOptions->getMagicRFCLinks() ) {
1458 return $m[0];
1459 }
1460 $keyword = 'RFC';
1461 $urlmsg = 'rfcurl';
1462 $cssClass = 'mw-magiclink-rfc';
1463 $trackingCat = 'magiclink-tracking-rfc';
1464 $id = $m[5];
1465 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1466 if ( !$this->mOptions->getMagicPMIDLinks() ) {
1467 return $m[0];
1468 }
1469 $keyword = 'PMID';
1470 $urlmsg = 'pubmedurl';
1471 $cssClass = 'mw-magiclink-pmid';
1472 $trackingCat = 'magiclink-tracking-pmid';
1473 $id = $m[5];
1474 } else {
1475 throw new MWException( __METHOD__ . ': unrecognised match type "' .
1476 substr( $m[0], 0, 20 ) . '"' );
1477 }
1478 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1479 $this->addTrackingCategory( $trackingCat );
1480 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1481 } elseif ( isset( $m[6] ) && $m[6] !== ''
1482 && $this->mOptions->getMagicISBNLinks()
1483 ) {
1484 # ISBN
1485 $isbn = $m[6];
1486 $space = self::SPACE_NOT_NL; # non-newline space
1487 $isbn = preg_replace( "/$space/", ' ', $isbn );
1488 $num = strtr( $isbn, [
1489 '-' => '',
1490 ' ' => '',
1491 'x' => 'X',
1492 ] );
1493 $this->addTrackingCategory( 'magiclink-tracking-isbn' );
1494 return $this->getLinkRenderer()->makeKnownLink(
1495 SpecialPage::getTitleFor( 'Booksources', $num ),
1496 "ISBN $isbn",
1497 [
1498 'class' => 'internal mw-magiclink-isbn',
1499 'title' => false // suppress title attribute
1500 ]
1501 );
1502 } else {
1503 return $m[0];
1504 }
1505 }
1506
1507 /**
1508 * Make a free external link, given a user-supplied URL
1509 *
1510 * @param string $url
1511 * @param int $numPostProto
1512 * The number of characters after the protocol.
1513 * @return string HTML
1514 * @private
1515 */
1516 public function makeFreeExternalLink( $url, $numPostProto ) {
1517 $trail = '';
1518
1519 # The characters '<' and '>' (which were escaped by
1520 # removeHTMLtags()) should not be included in
1521 # URLs, per RFC 2396.
1522 # Make &nbsp; terminate a URL as well (bug T84937)
1523 $m2 = [];
1524 if ( preg_match(
1525 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1526 $url,
1527 $m2,
1528 PREG_OFFSET_CAPTURE
1529 ) ) {
1530 $trail = substr( $url, $m2[0][1] ) . $trail;
1531 $url = substr( $url, 0, $m2[0][1] );
1532 }
1533
1534 # Move trailing punctuation to $trail
1535 $sep = ',;\.:!?';
1536 # If there is no left bracket, then consider right brackets fair game too
1537 if ( strpos( $url, '(' ) === false ) {
1538 $sep .= ')';
1539 }
1540
1541 $urlRev = strrev( $url );
1542 $numSepChars = strspn( $urlRev, $sep );
1543 # Don't break a trailing HTML entity by moving the ; into $trail
1544 # This is in hot code, so use substr_compare to avoid having to
1545 # create a new string object for the comparison
1546 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1547 # more optimization: instead of running preg_match with a $
1548 # anchor, which can be slow, do the match on the reversed
1549 # string starting at the desired offset.
1550 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1551 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1552 $numSepChars--;
1553 }
1554 }
1555 if ( $numSepChars ) {
1556 $trail = substr( $url, -$numSepChars ) . $trail;
1557 $url = substr( $url, 0, -$numSepChars );
1558 }
1559
1560 # Verify that we still have a real URL after trail removal, and
1561 # not just lone protocol
1562 if ( strlen( $trail ) >= $numPostProto ) {
1563 return $url . $trail;
1564 }
1565
1566 $url = Sanitizer::cleanUrl( $url );
1567
1568 # Is this an external image?
1569 $text = $this->maybeMakeExternalImage( $url );
1570 if ( $text === false ) {
1571 # Not an image, make a link
1572 $text = Linker::makeExternalLink( $url,
1573 $this->getConverterLanguage()->markNoConversion( $url, true ),
1574 true, 'free',
1575 $this->getExternalLinkAttribs( $url ), $this->mTitle );
1576 # Register it in the output object...
1577 # Replace unnecessary URL escape codes with their equivalent characters
1578 $pasteurized = self::normalizeLinkUrl( $url );
1579 $this->mOutput->addExternalLink( $pasteurized );
1580 }
1581 return $text . $trail;
1582 }
1583
1584 /**
1585 * Parse headers and return html
1586 *
1587 * @private
1588 *
1589 * @param string $text
1590 *
1591 * @return string
1592 */
1593 public function doHeadings( $text ) {
1594 for ( $i = 6; $i >= 1; --$i ) {
1595 $h = str_repeat( '=', $i );
1596 $text = preg_replace( "/^$h(.+)$h\\s*$/m", "<h$i>\\1</h$i>", $text );
1597 }
1598 return $text;
1599 }
1600
1601 /**
1602 * Replace single quotes with HTML markup
1603 * @private
1604 *
1605 * @param string $text
1606 *
1607 * @return string The altered text
1608 */
1609 public function doAllQuotes( $text ) {
1610 $outtext = '';
1611 $lines = StringUtils::explode( "\n", $text );
1612 foreach ( $lines as $line ) {
1613 $outtext .= $this->doQuotes( $line ) . "\n";
1614 }
1615 $outtext = substr( $outtext, 0, -1 );
1616 return $outtext;
1617 }
1618
1619 /**
1620 * Helper function for doAllQuotes()
1621 *
1622 * @param string $text
1623 *
1624 * @return string
1625 */
1626 public function doQuotes( $text ) {
1627 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1628 $countarr = count( $arr );
1629 if ( $countarr == 1 ) {
1630 return $text;
1631 }
1632
1633 // First, do some preliminary work. This may shift some apostrophes from
1634 // being mark-up to being text. It also counts the number of occurrences
1635 // of bold and italics mark-ups.
1636 $numbold = 0;
1637 $numitalics = 0;
1638 for ( $i = 1; $i < $countarr; $i += 2 ) {
1639 $thislen = strlen( $arr[$i] );
1640 // If there are ever four apostrophes, assume the first is supposed to
1641 // be text, and the remaining three constitute mark-up for bold text.
1642 // (bug 13227: ''''foo'''' turns into ' ''' foo ' ''')
1643 if ( $thislen == 4 ) {
1644 $arr[$i - 1] .= "'";
1645 $arr[$i] = "'''";
1646 $thislen = 3;
1647 } elseif ( $thislen > 5 ) {
1648 // If there are more than 5 apostrophes in a row, assume they're all
1649 // text except for the last 5.
1650 // (bug 13227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1651 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1652 $arr[$i] = "'''''";
1653 $thislen = 5;
1654 }
1655 // Count the number of occurrences of bold and italics mark-ups.
1656 if ( $thislen == 2 ) {
1657 $numitalics++;
1658 } elseif ( $thislen == 3 ) {
1659 $numbold++;
1660 } elseif ( $thislen == 5 ) {
1661 $numitalics++;
1662 $numbold++;
1663 }
1664 }
1665
1666 // If there is an odd number of both bold and italics, it is likely
1667 // that one of the bold ones was meant to be an apostrophe followed
1668 // by italics. Which one we cannot know for certain, but it is more
1669 // likely to be one that has a single-letter word before it.
1670 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1671 $firstsingleletterword = -1;
1672 $firstmultiletterword = -1;
1673 $firstspace = -1;
1674 for ( $i = 1; $i < $countarr; $i += 2 ) {
1675 if ( strlen( $arr[$i] ) == 3 ) {
1676 $x1 = substr( $arr[$i - 1], -1 );
1677 $x2 = substr( $arr[$i - 1], -2, 1 );
1678 if ( $x1 === ' ' ) {
1679 if ( $firstspace == -1 ) {
1680 $firstspace = $i;
1681 }
1682 } elseif ( $x2 === ' ' ) {
1683 $firstsingleletterword = $i;
1684 // if $firstsingleletterword is set, we don't
1685 // look at the other options, so we can bail early.
1686 break;
1687 } else {
1688 if ( $firstmultiletterword == -1 ) {
1689 $firstmultiletterword = $i;
1690 }
1691 }
1692 }
1693 }
1694
1695 // If there is a single-letter word, use it!
1696 if ( $firstsingleletterword > -1 ) {
1697 $arr[$firstsingleletterword] = "''";
1698 $arr[$firstsingleletterword - 1] .= "'";
1699 } elseif ( $firstmultiletterword > -1 ) {
1700 // If not, but there's a multi-letter word, use that one.
1701 $arr[$firstmultiletterword] = "''";
1702 $arr[$firstmultiletterword - 1] .= "'";
1703 } elseif ( $firstspace > -1 ) {
1704 // ... otherwise use the first one that has neither.
1705 // (notice that it is possible for all three to be -1 if, for example,
1706 // there is only one pentuple-apostrophe in the line)
1707 $arr[$firstspace] = "''";
1708 $arr[$firstspace - 1] .= "'";
1709 }
1710 }
1711
1712 // Now let's actually convert our apostrophic mush to HTML!
1713 $output = '';
1714 $buffer = '';
1715 $state = '';
1716 $i = 0;
1717 foreach ( $arr as $r ) {
1718 if ( ( $i % 2 ) == 0 ) {
1719 if ( $state === 'both' ) {
1720 $buffer .= $r;
1721 } else {
1722 $output .= $r;
1723 }
1724 } else {
1725 $thislen = strlen( $r );
1726 if ( $thislen == 2 ) {
1727 if ( $state === 'i' ) {
1728 $output .= '</i>';
1729 $state = '';
1730 } elseif ( $state === 'bi' ) {
1731 $output .= '</i>';
1732 $state = 'b';
1733 } elseif ( $state === 'ib' ) {
1734 $output .= '</b></i><b>';
1735 $state = 'b';
1736 } elseif ( $state === 'both' ) {
1737 $output .= '<b><i>' . $buffer . '</i>';
1738 $state = 'b';
1739 } else { // $state can be 'b' or ''
1740 $output .= '<i>';
1741 $state .= 'i';
1742 }
1743 } elseif ( $thislen == 3 ) {
1744 if ( $state === 'b' ) {
1745 $output .= '</b>';
1746 $state = '';
1747 } elseif ( $state === 'bi' ) {
1748 $output .= '</i></b><i>';
1749 $state = 'i';
1750 } elseif ( $state === 'ib' ) {
1751 $output .= '</b>';
1752 $state = 'i';
1753 } elseif ( $state === 'both' ) {
1754 $output .= '<i><b>' . $buffer . '</b>';
1755 $state = 'i';
1756 } else { // $state can be 'i' or ''
1757 $output .= '<b>';
1758 $state .= 'b';
1759 }
1760 } elseif ( $thislen == 5 ) {
1761 if ( $state === 'b' ) {
1762 $output .= '</b><i>';
1763 $state = 'i';
1764 } elseif ( $state === 'i' ) {
1765 $output .= '</i><b>';
1766 $state = 'b';
1767 } elseif ( $state === 'bi' ) {
1768 $output .= '</i></b>';
1769 $state = '';
1770 } elseif ( $state === 'ib' ) {
1771 $output .= '</b></i>';
1772 $state = '';
1773 } elseif ( $state === 'both' ) {
1774 $output .= '<i><b>' . $buffer . '</b></i>';
1775 $state = '';
1776 } else { // ($state == '')
1777 $buffer = '';
1778 $state = 'both';
1779 }
1780 }
1781 }
1782 $i++;
1783 }
1784 // Now close all remaining tags. Notice that the order is important.
1785 if ( $state === 'b' || $state === 'ib' ) {
1786 $output .= '</b>';
1787 }
1788 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1789 $output .= '</i>';
1790 }
1791 if ( $state === 'bi' ) {
1792 $output .= '</b>';
1793 }
1794 // There might be lonely ''''', so make sure we have a buffer
1795 if ( $state === 'both' && $buffer ) {
1796 $output .= '<b><i>' . $buffer . '</i></b>';
1797 }
1798 return $output;
1799 }
1800
1801 /**
1802 * Replace external links (REL)
1803 *
1804 * Note: this is all very hackish and the order of execution matters a lot.
1805 * Make sure to run tests/parser/parserTests.php if you change this code.
1806 *
1807 * @private
1808 *
1809 * @param string $text
1810 *
1811 * @throws MWException
1812 * @return string
1813 */
1814 public function replaceExternalLinks( $text ) {
1815
1816 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1817 if ( $bits === false ) {
1818 throw new MWException( "PCRE needs to be compiled with "
1819 . "--enable-unicode-properties in order for MediaWiki to function" );
1820 }
1821 $s = array_shift( $bits );
1822
1823 $i = 0;
1824 while ( $i < count( $bits ) ) {
1825 $url = $bits[$i++];
1826 $i++; // protocol
1827 $text = $bits[$i++];
1828 $trail = $bits[$i++];
1829
1830 # The characters '<' and '>' (which were escaped by
1831 # removeHTMLtags()) should not be included in
1832 # URLs, per RFC 2396.
1833 $m2 = [];
1834 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1835 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1836 $url = substr( $url, 0, $m2[0][1] );
1837 }
1838
1839 # If the link text is an image URL, replace it with an <img> tag
1840 # This happened by accident in the original parser, but some people used it extensively
1841 $img = $this->maybeMakeExternalImage( $text );
1842 if ( $img !== false ) {
1843 $text = $img;
1844 }
1845
1846 $dtrail = '';
1847
1848 # Set linktype for CSS - if URL==text, link is essentially free
1849 $linktype = ( $text === $url ) ? 'free' : 'text';
1850
1851 # No link text, e.g. [http://domain.tld/some.link]
1852 if ( $text == '' ) {
1853 # Autonumber
1854 $langObj = $this->getTargetLanguage();
1855 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1856 $linktype = 'autonumber';
1857 } else {
1858 # Have link text, e.g. [http://domain.tld/some.link text]s
1859 # Check for trail
1860 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1861 }
1862
1863 $text = $this->getConverterLanguage()->markNoConversion( $text );
1864
1865 $url = Sanitizer::cleanUrl( $url );
1866
1867 # Use the encoded URL
1868 # This means that users can paste URLs directly into the text
1869 # Funny characters like ö aren't valid in URLs anyway
1870 # This was changed in August 2004
1871 $s .= Linker::makeExternalLink( $url, $text, false, $linktype,
1872 $this->getExternalLinkAttribs( $url ), $this->mTitle ) . $dtrail . $trail;
1873
1874 # Register link in the output object.
1875 # Replace unnecessary URL escape codes with the referenced character
1876 # This prevents spammers from hiding links from the filters
1877 $pasteurized = self::normalizeLinkUrl( $url );
1878 $this->mOutput->addExternalLink( $pasteurized );
1879 }
1880
1881 return $s;
1882 }
1883
1884 /**
1885 * Get the rel attribute for a particular external link.
1886 *
1887 * @since 1.21
1888 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1889 * nofollow if appropriate
1890 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1891 * @return string|null Rel attribute for $url
1892 */
1893 public static function getExternalLinkRel( $url = false, $title = null ) {
1894 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1895 $ns = $title ? $title->getNamespace() : false;
1896 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1897 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1898 ) {
1899 return 'nofollow';
1900 }
1901 return null;
1902 }
1903
1904 /**
1905 * Get an associative array of additional HTML attributes appropriate for a
1906 * particular external link. This currently may include rel => nofollow
1907 * (depending on configuration, namespace, and the URL's domain) and/or a
1908 * target attribute (depending on configuration).
1909 *
1910 * @param string $url URL to extract the domain from for rel =>
1911 * nofollow if appropriate
1912 * @return array Associative array of HTML attributes
1913 */
1914 public function getExternalLinkAttribs( $url ) {
1915 $attribs = [];
1916 $rel = self::getExternalLinkRel( $url, $this->mTitle );
1917
1918 $target = $this->mOptions->getExternalLinkTarget();
1919 if ( $target ) {
1920 $attribs['target'] = $target;
1921 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1922 // T133507. New windows can navigate parent cross-origin.
1923 // Including noreferrer due to lacking browser
1924 // support of noopener. Eventually noreferrer should be removed.
1925 if ( $rel !== '' ) {
1926 $rel .= ' ';
1927 }
1928 $rel .= 'noreferrer noopener';
1929 }
1930 }
1931 $attribs['rel'] = $rel;
1932 return $attribs;
1933 }
1934
1935 /**
1936 * Replace unusual escape codes in a URL with their equivalent characters
1937 *
1938 * @deprecated since 1.24, use normalizeLinkUrl
1939 * @param string $url
1940 * @return string
1941 */
1942 public static function replaceUnusualEscapes( $url ) {
1943 wfDeprecated( __METHOD__, '1.24' );
1944 return self::normalizeLinkUrl( $url );
1945 }
1946
1947 /**
1948 * Replace unusual escape codes in a URL with their equivalent characters
1949 *
1950 * This generally follows the syntax defined in RFC 3986, with special
1951 * consideration for HTTP query strings.
1952 *
1953 * @param string $url
1954 * @return string
1955 */
1956 public static function normalizeLinkUrl( $url ) {
1957 # First, make sure unsafe characters are encoded
1958 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
1959 function ( $m ) {
1960 return rawurlencode( $m[0] );
1961 },
1962 $url
1963 );
1964
1965 $ret = '';
1966 $end = strlen( $url );
1967
1968 # Fragment part - 'fragment'
1969 $start = strpos( $url, '#' );
1970 if ( $start !== false && $start < $end ) {
1971 $ret = self::normalizeUrlComponent(
1972 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
1973 $end = $start;
1974 }
1975
1976 # Query part - 'query' minus &=+;
1977 $start = strpos( $url, '?' );
1978 if ( $start !== false && $start < $end ) {
1979 $ret = self::normalizeUrlComponent(
1980 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
1981 $end = $start;
1982 }
1983
1984 # Scheme and path part - 'pchar'
1985 # (we assume no userinfo or encoded colons in the host)
1986 $ret = self::normalizeUrlComponent(
1987 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
1988
1989 return $ret;
1990 }
1991
1992 private static function normalizeUrlComponent( $component, $unsafe ) {
1993 $callback = function ( $matches ) use ( $unsafe ) {
1994 $char = urldecode( $matches[0] );
1995 $ord = ord( $char );
1996 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
1997 # Unescape it
1998 return $char;
1999 } else {
2000 # Leave it escaped, but use uppercase for a-f
2001 return strtoupper( $matches[0] );
2002 }
2003 };
2004 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
2005 }
2006
2007 /**
2008 * make an image if it's allowed, either through the global
2009 * option, through the exception, or through the on-wiki whitelist
2010 *
2011 * @param string $url
2012 *
2013 * @return string
2014 */
2015 private function maybeMakeExternalImage( $url ) {
2016 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
2017 $imagesexception = !empty( $imagesfrom );
2018 $text = false;
2019 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2020 if ( $imagesexception && is_array( $imagesfrom ) ) {
2021 $imagematch = false;
2022 foreach ( $imagesfrom as $match ) {
2023 if ( strpos( $url, $match ) === 0 ) {
2024 $imagematch = true;
2025 break;
2026 }
2027 }
2028 } elseif ( $imagesexception ) {
2029 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2030 } else {
2031 $imagematch = false;
2032 }
2033
2034 if ( $this->mOptions->getAllowExternalImages()
2035 || ( $imagesexception && $imagematch )
2036 ) {
2037 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
2038 # Image found
2039 $text = Linker::makeExternalImage( $url );
2040 }
2041 }
2042 if ( !$text && $this->mOptions->getEnableImageWhitelist()
2043 && preg_match( self::EXT_IMAGE_REGEX, $url )
2044 ) {
2045 $whitelist = explode(
2046 "\n",
2047 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2048 );
2049
2050 foreach ( $whitelist as $entry ) {
2051 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2052 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
2053 continue;
2054 }
2055 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2056 # Image matches a whitelist entry
2057 $text = Linker::makeExternalImage( $url );
2058 break;
2059 }
2060 }
2061 }
2062 return $text;
2063 }
2064
2065 /**
2066 * Process [[ ]] wikilinks
2067 *
2068 * @param string $s
2069 *
2070 * @return string Processed text
2071 *
2072 * @private
2073 */
2074 public function replaceInternalLinks( $s ) {
2075 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
2076 return $s;
2077 }
2078
2079 /**
2080 * Process [[ ]] wikilinks (RIL)
2081 * @param string $s
2082 * @throws MWException
2083 * @return LinkHolderArray
2084 *
2085 * @private
2086 */
2087 public function replaceInternalLinks2( &$s ) {
2088 global $wgExtraInterlanguageLinkPrefixes;
2089
2090 static $tc = false, $e1, $e1_img;
2091 # the % is needed to support urlencoded titles as well
2092 if ( !$tc ) {
2093 $tc = Title::legalChars() . '#%';
2094 # Match a link having the form [[namespace:link|alternate]]trail
2095 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2096 # Match cases where there is no "]]", which might still be images
2097 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2098 }
2099
2100 $holders = new LinkHolderArray( $this );
2101
2102 # split the entire text string on occurrences of [[
2103 $a = StringUtils::explode( '[[', ' ' . $s );
2104 # get the first element (all text up to first [[), and remove the space we added
2105 $s = $a->current();
2106 $a->next();
2107 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2108 $s = substr( $s, 1 );
2109
2110 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2111 $e2 = null;
2112 if ( $useLinkPrefixExtension ) {
2113 # Match the end of a line for a word that's not followed by whitespace,
2114 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2115 global $wgContLang;
2116 $charset = $wgContLang->linkPrefixCharset();
2117 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2118 }
2119
2120 if ( is_null( $this->mTitle ) ) {
2121 throw new MWException( __METHOD__ . ": \$this->mTitle is null\n" );
2122 }
2123 $nottalk = !$this->mTitle->isTalkPage();
2124
2125 if ( $useLinkPrefixExtension ) {
2126 $m = [];
2127 if ( preg_match( $e2, $s, $m ) ) {
2128 $first_prefix = $m[2];
2129 } else {
2130 $first_prefix = false;
2131 }
2132 } else {
2133 $prefix = '';
2134 }
2135
2136 $useSubpages = $this->areSubpagesAllowed();
2137
2138 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2139 # Loop for each link
2140 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2141 // @codingStandardsIgnoreEnd
2142
2143 # Check for excessive memory usage
2144 if ( $holders->isBig() ) {
2145 # Too big
2146 # Do the existence check, replace the link holders and clear the array
2147 $holders->replace( $s );
2148 $holders->clear();
2149 }
2150
2151 if ( $useLinkPrefixExtension ) {
2152 if ( preg_match( $e2, $s, $m ) ) {
2153 $prefix = $m[2];
2154 $s = $m[1];
2155 } else {
2156 $prefix = '';
2157 }
2158 # first link
2159 if ( $first_prefix ) {
2160 $prefix = $first_prefix;
2161 $first_prefix = false;
2162 }
2163 }
2164
2165 $might_be_img = false;
2166
2167 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2168 $text = $m[2];
2169 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2170 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2171 # the real problem is with the $e1 regex
2172 # See bug 1300.
2173 # Still some problems for cases where the ] is meant to be outside punctuation,
2174 # and no image is in sight. See bug 2095.
2175 if ( $text !== ''
2176 && substr( $m[3], 0, 1 ) === ']'
2177 && strpos( $text, '[' ) !== false
2178 ) {
2179 $text .= ']'; # so that replaceExternalLinks($text) works later
2180 $m[3] = substr( $m[3], 1 );
2181 }
2182 # fix up urlencoded title texts
2183 if ( strpos( $m[1], '%' ) !== false ) {
2184 # Should anchors '#' also be rejected?
2185 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2186 }
2187 $trail = $m[3];
2188 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2189 # Invalid, but might be an image with a link in its caption
2190 $might_be_img = true;
2191 $text = $m[2];
2192 if ( strpos( $m[1], '%' ) !== false ) {
2193 $m[1] = str_replace( [ '<', '>' ], [ '&lt;', '&gt;' ], rawurldecode( $m[1] ) );
2194 }
2195 $trail = "";
2196 } else { # Invalid form; output directly
2197 $s .= $prefix . '[[' . $line;
2198 continue;
2199 }
2200
2201 $origLink = $m[1];
2202
2203 # Don't allow internal links to pages containing
2204 # PROTO: where PROTO is a valid URL protocol; these
2205 # should be external links.
2206 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $origLink ) ) {
2207 $s .= $prefix . '[[' . $line;
2208 continue;
2209 }
2210
2211 # Make subpage if necessary
2212 if ( $useSubpages ) {
2213 $link = $this->maybeDoSubpageLink( $origLink, $text );
2214 } else {
2215 $link = $origLink;
2216 }
2217
2218 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2219 if ( !$noforce ) {
2220 # Strip off leading ':'
2221 $link = substr( $link, 1 );
2222 }
2223
2224 $unstrip = $this->mStripState->unstripNoWiki( $link );
2225 $nt = is_string( $unstrip ) ? Title::newFromText( $unstrip ) : null;
2226 if ( $nt === null ) {
2227 $s .= $prefix . '[[' . $line;
2228 continue;
2229 }
2230
2231 $ns = $nt->getNamespace();
2232 $iw = $nt->getInterwiki();
2233
2234 if ( $might_be_img ) { # if this is actually an invalid link
2235 if ( $ns == NS_FILE && $noforce ) { # but might be an image
2236 $found = false;
2237 while ( true ) {
2238 # look at the next 'line' to see if we can close it there
2239 $a->next();
2240 $next_line = $a->current();
2241 if ( $next_line === false || $next_line === null ) {
2242 break;
2243 }
2244 $m = explode( ']]', $next_line, 3 );
2245 if ( count( $m ) == 3 ) {
2246 # the first ]] closes the inner link, the second the image
2247 $found = true;
2248 $text .= "[[{$m[0]}]]{$m[1]}";
2249 $trail = $m[2];
2250 break;
2251 } elseif ( count( $m ) == 2 ) {
2252 # if there's exactly one ]] that's fine, we'll keep looking
2253 $text .= "[[{$m[0]}]]{$m[1]}";
2254 } else {
2255 # if $next_line is invalid too, we need look no further
2256 $text .= '[[' . $next_line;
2257 break;
2258 }
2259 }
2260 if ( !$found ) {
2261 # we couldn't find the end of this imageLink, so output it raw
2262 # but don't ignore what might be perfectly normal links in the text we've examined
2263 $holders->merge( $this->replaceInternalLinks2( $text ) );
2264 $s .= "{$prefix}[[$link|$text";
2265 # note: no $trail, because without an end, there *is* no trail
2266 continue;
2267 }
2268 } else { # it's not an image, so output it raw
2269 $s .= "{$prefix}[[$link|$text";
2270 # note: no $trail, because without an end, there *is* no trail
2271 continue;
2272 }
2273 }
2274
2275 $wasblank = ( $text == '' );
2276 if ( $wasblank ) {
2277 $text = $link;
2278 } else {
2279 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
2280 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2281 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2282 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2283 $text = $this->doQuotes( $text );
2284 }
2285
2286 # Link not escaped by : , create the various objects
2287 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2288 # Interwikis
2289 if (
2290 $iw && $this->mOptions->getInterwikiMagic() && $nottalk && (
2291 Language::fetchLanguageName( $iw, null, 'mw' ) ||
2292 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2293 )
2294 ) {
2295 # Bug 24502: filter duplicates
2296 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
2297 $this->mLangLinkLanguages[$iw] = true;
2298 $this->mOutput->addLanguageLink( $nt->getFullText() );
2299 }
2300
2301 $s = rtrim( $s . $prefix );
2302 $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail;
2303 continue;
2304 }
2305
2306 if ( $ns == NS_FILE ) {
2307 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
2308 if ( $wasblank ) {
2309 # if no parameters were passed, $text
2310 # becomes something like "File:Foo.png",
2311 # which we don't want to pass on to the
2312 # image generator
2313 $text = '';
2314 } else {
2315 # recursively parse links inside the image caption
2316 # actually, this will parse them in any other parameters, too,
2317 # but it might be hard to fix that, and it doesn't matter ATM
2318 $text = $this->replaceExternalLinks( $text );
2319 $holders->merge( $this->replaceInternalLinks2( $text ) );
2320 }
2321 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2322 $s .= $prefix . $this->armorLinks(
2323 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2324 continue;
2325 }
2326 } elseif ( $ns == NS_CATEGORY ) {
2327 $s = rtrim( $s . "\n" ); # bug 87
2328
2329 if ( $wasblank ) {
2330 $sortkey = $this->getDefaultSort();
2331 } else {
2332 $sortkey = $text;
2333 }
2334 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
2335 $sortkey = str_replace( "\n", '', $sortkey );
2336 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
2337 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
2338
2339 /**
2340 * Strip the whitespace Category links produce, see bug 87
2341 */
2342 $s .= trim( $prefix . $trail, "\n" ) == '' ? '' : $prefix . $trail;
2343
2344 continue;
2345 }
2346 }
2347
2348 # Self-link checking. For some languages, variants of the title are checked in
2349 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2350 # for linking to a different variant.
2351 if ( $ns != NS_SPECIAL && $nt->equals( $this->mTitle ) && !$nt->hasFragment() ) {
2352 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail );
2353 continue;
2354 }
2355
2356 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2357 # @todo FIXME: Should do batch file existence checks, see comment below
2358 if ( $ns == NS_MEDIA ) {
2359 # Give extensions a chance to select the file revision for us
2360 $options = [];
2361 $descQuery = false;
2362 Hooks::run( 'BeforeParserFetchFileAndTitle',
2363 [ $this, $nt, &$options, &$descQuery ] );
2364 # Fetch and register the file (file title may be different via hooks)
2365 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2366 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2367 $s .= $prefix . $this->armorLinks(
2368 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2369 continue;
2370 }
2371
2372 # Some titles, such as valid special pages or files in foreign repos, should
2373 # be shown as bluelinks even though they're not included in the page table
2374 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2375 # batch file existence checks for NS_FILE and NS_MEDIA
2376 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2377 $this->mOutput->addLink( $nt );
2378 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2379 } else {
2380 # Links will be added to the output link list after checking
2381 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2382 }
2383 }
2384 return $holders;
2385 }
2386
2387 /**
2388 * Render a forced-blue link inline; protect against double expansion of
2389 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2390 * Since this little disaster has to split off the trail text to avoid
2391 * breaking URLs in the following text without breaking trails on the
2392 * wiki links, it's been made into a horrible function.
2393 *
2394 * @param Title $nt
2395 * @param string $text
2396 * @param string $trail
2397 * @param string $prefix
2398 * @return string HTML-wikitext mix oh yuck
2399 */
2400 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2401 list( $inside, $trail ) = Linker::splitTrail( $trail );
2402
2403 if ( $text == '' ) {
2404 $text = htmlspecialchars( $nt->getPrefixedText() );
2405 }
2406
2407 $link = $this->getLinkRenderer()->makeKnownLink(
2408 $nt, new HtmlArmor( "$prefix$text$inside" )
2409 );
2410
2411 return $this->armorLinks( $link ) . $trail;
2412 }
2413
2414 /**
2415 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2416 * going to go through further parsing steps before inline URL expansion.
2417 *
2418 * Not needed quite as much as it used to be since free links are a bit
2419 * more sensible these days. But bracketed links are still an issue.
2420 *
2421 * @param string $text More-or-less HTML
2422 * @return string Less-or-more HTML with NOPARSE bits
2423 */
2424 public function armorLinks( $text ) {
2425 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/',
2426 self::MARKER_PREFIX . "NOPARSE$1", $text );
2427 }
2428
2429 /**
2430 * Return true if subpage links should be expanded on this page.
2431 * @return bool
2432 */
2433 public function areSubpagesAllowed() {
2434 # Some namespaces don't allow subpages
2435 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2436 }
2437
2438 /**
2439 * Handle link to subpage if necessary
2440 *
2441 * @param string $target The source of the link
2442 * @param string &$text The link text, modified as necessary
2443 * @return string The full name of the link
2444 * @private
2445 */
2446 public function maybeDoSubpageLink( $target, &$text ) {
2447 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2448 }
2449
2450 /**
2451 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2452 *
2453 * @param string $text
2454 * @param bool $linestart Whether or not this is at the start of a line.
2455 * @private
2456 * @return string The lists rendered as HTML
2457 */
2458 public function doBlockLevels( $text, $linestart ) {
2459 return BlockLevelPass::doBlockLevels( $text, $linestart );
2460 }
2461
2462 /**
2463 * Return value of a magic variable (like PAGENAME)
2464 *
2465 * @private
2466 *
2467 * @param int $index
2468 * @param bool|PPFrame $frame
2469 *
2470 * @throws MWException
2471 * @return string
2472 */
2473 public function getVariableValue( $index, $frame = false ) {
2474 global $wgContLang, $wgSitename, $wgServer, $wgServerName;
2475 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2476
2477 if ( is_null( $this->mTitle ) ) {
2478 // If no title set, bad things are going to happen
2479 // later. Title should always be set since this
2480 // should only be called in the middle of a parse
2481 // operation (but the unit-tests do funky stuff)
2482 throw new MWException( __METHOD__ . ' Should only be '
2483 . ' called while parsing (no title set)' );
2484 }
2485
2486 /**
2487 * Some of these require message or data lookups and can be
2488 * expensive to check many times.
2489 */
2490 if ( Hooks::run( 'ParserGetVariableValueVarCache', [ &$this, &$this->mVarCache ] ) ) {
2491 if ( isset( $this->mVarCache[$index] ) ) {
2492 return $this->mVarCache[$index];
2493 }
2494 }
2495
2496 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2497 Hooks::run( 'ParserGetVariableValueTs', [ &$this, &$ts ] );
2498
2499 $pageLang = $this->getFunctionLang();
2500
2501 switch ( $index ) {
2502 case '!':
2503 $value = '|';
2504 break;
2505 case 'currentmonth':
2506 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'm' ) );
2507 break;
2508 case 'currentmonth1':
2509 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2510 break;
2511 case 'currentmonthname':
2512 $value = $pageLang->getMonthName( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2513 break;
2514 case 'currentmonthnamegen':
2515 $value = $pageLang->getMonthNameGen( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2516 break;
2517 case 'currentmonthabbrev':
2518 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getInstance( $ts )->format( 'n' ) );
2519 break;
2520 case 'currentday':
2521 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'j' ) );
2522 break;
2523 case 'currentday2':
2524 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'd' ) );
2525 break;
2526 case 'localmonth':
2527 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'm' ) );
2528 break;
2529 case 'localmonth1':
2530 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2531 break;
2532 case 'localmonthname':
2533 $value = $pageLang->getMonthName( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2534 break;
2535 case 'localmonthnamegen':
2536 $value = $pageLang->getMonthNameGen( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2537 break;
2538 case 'localmonthabbrev':
2539 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2540 break;
2541 case 'localday':
2542 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'j' ) );
2543 break;
2544 case 'localday2':
2545 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'd' ) );
2546 break;
2547 case 'pagename':
2548 $value = wfEscapeWikiText( $this->mTitle->getText() );
2549 break;
2550 case 'pagenamee':
2551 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() );
2552 break;
2553 case 'fullpagename':
2554 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2555 break;
2556 case 'fullpagenamee':
2557 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() );
2558 break;
2559 case 'subpagename':
2560 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2561 break;
2562 case 'subpagenamee':
2563 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() );
2564 break;
2565 case 'rootpagename':
2566 $value = wfEscapeWikiText( $this->mTitle->getRootText() );
2567 break;
2568 case 'rootpagenamee':
2569 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2570 ' ',
2571 '_',
2572 $this->mTitle->getRootText()
2573 ) ) );
2574 break;
2575 case 'basepagename':
2576 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2577 break;
2578 case 'basepagenamee':
2579 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2580 ' ',
2581 '_',
2582 $this->mTitle->getBaseText()
2583 ) ) );
2584 break;
2585 case 'talkpagename':
2586 if ( $this->mTitle->canTalk() ) {
2587 $talkPage = $this->mTitle->getTalkPage();
2588 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2589 } else {
2590 $value = '';
2591 }
2592 break;
2593 case 'talkpagenamee':
2594 if ( $this->mTitle->canTalk() ) {
2595 $talkPage = $this->mTitle->getTalkPage();
2596 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2597 } else {
2598 $value = '';
2599 }
2600 break;
2601 case 'subjectpagename':
2602 $subjPage = $this->mTitle->getSubjectPage();
2603 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2604 break;
2605 case 'subjectpagenamee':
2606 $subjPage = $this->mTitle->getSubjectPage();
2607 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2608 break;
2609 case 'pageid': // requested in bug 23427
2610 $pageid = $this->getTitle()->getArticleID();
2611 if ( $pageid == 0 ) {
2612 # 0 means the page doesn't exist in the database,
2613 # which means the user is previewing a new page.
2614 # The vary-revision flag must be set, because the magic word
2615 # will have a different value once the page is saved.
2616 $this->mOutput->setFlag( 'vary-revision' );
2617 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2618 }
2619 $value = $pageid ? $pageid : null;
2620 break;
2621 case 'revisionid':
2622 # Let the edit saving system know we should parse the page
2623 # *after* a revision ID has been assigned.
2624 $this->mOutput->setFlag( 'vary-revision-id' );
2625 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision-id...\n" );
2626 $value = $this->mRevisionId;
2627 if ( !$value && $this->mOptions->getSpeculativeRevIdCallback() ) {
2628 $value = call_user_func( $this->mOptions->getSpeculativeRevIdCallback() );
2629 $this->mOutput->setSpeculativeRevIdUsed( $value );
2630 }
2631 break;
2632 case 'revisionday':
2633 # Let the edit saving system know we should parse the page
2634 # *after* a revision ID has been assigned. This is for null edits.
2635 $this->mOutput->setFlag( 'vary-revision' );
2636 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2637 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2638 break;
2639 case 'revisionday2':
2640 # Let the edit saving system know we should parse the page
2641 # *after* a revision ID has been assigned. This is for null edits.
2642 $this->mOutput->setFlag( 'vary-revision' );
2643 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2644 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2645 break;
2646 case 'revisionmonth':
2647 # Let the edit saving system know we should parse the page
2648 # *after* a revision ID has been assigned. This is for null edits.
2649 $this->mOutput->setFlag( 'vary-revision' );
2650 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2651 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2652 break;
2653 case 'revisionmonth1':
2654 # Let the edit saving system know we should parse the page
2655 # *after* a revision ID has been assigned. This is for null edits.
2656 $this->mOutput->setFlag( 'vary-revision' );
2657 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2658 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2659 break;
2660 case 'revisionyear':
2661 # Let the edit saving system know we should parse the page
2662 # *after* a revision ID has been assigned. This is for null edits.
2663 $this->mOutput->setFlag( 'vary-revision' );
2664 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2665 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2666 break;
2667 case 'revisiontimestamp':
2668 # Let the edit saving system know we should parse the page
2669 # *after* a revision ID has been assigned. This is for null edits.
2670 $this->mOutput->setFlag( 'vary-revision' );
2671 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2672 $value = $this->getRevisionTimestamp();
2673 break;
2674 case 'revisionuser':
2675 # Let the edit saving system know we should parse the page
2676 # *after* a revision ID has been assigned for null edits.
2677 $this->mOutput->setFlag( 'vary-user' );
2678 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-user...\n" );
2679 $value = $this->getRevisionUser();
2680 break;
2681 case 'revisionsize':
2682 $value = $this->getRevisionSize();
2683 break;
2684 case 'namespace':
2685 $value = str_replace( '_', ' ', $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2686 break;
2687 case 'namespacee':
2688 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2689 break;
2690 case 'namespacenumber':
2691 $value = $this->mTitle->getNamespace();
2692 break;
2693 case 'talkspace':
2694 $value = $this->mTitle->canTalk()
2695 ? str_replace( '_', ' ', $this->mTitle->getTalkNsText() )
2696 : '';
2697 break;
2698 case 'talkspacee':
2699 $value = $this->mTitle->canTalk() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2700 break;
2701 case 'subjectspace':
2702 $value = str_replace( '_', ' ', $this->mTitle->getSubjectNsText() );
2703 break;
2704 case 'subjectspacee':
2705 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2706 break;
2707 case 'currentdayname':
2708 $value = $pageLang->getWeekdayName( (int)MWTimestamp::getInstance( $ts )->format( 'w' ) + 1 );
2709 break;
2710 case 'currentyear':
2711 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'Y' ), true );
2712 break;
2713 case 'currenttime':
2714 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2715 break;
2716 case 'currenthour':
2717 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'H' ), true );
2718 break;
2719 case 'currentweek':
2720 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2721 # int to remove the padding
2722 $value = $pageLang->formatNum( (int)MWTimestamp::getInstance( $ts )->format( 'W' ) );
2723 break;
2724 case 'currentdow':
2725 $value = $pageLang->formatNum( MWTimestamp::getInstance( $ts )->format( 'w' ) );
2726 break;
2727 case 'localdayname':
2728 $value = $pageLang->getWeekdayName(
2729 (int)MWTimestamp::getLocalInstance( $ts )->format( 'w' ) + 1
2730 );
2731 break;
2732 case 'localyear':
2733 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'Y' ), true );
2734 break;
2735 case 'localtime':
2736 $value = $pageLang->time(
2737 MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' ),
2738 false,
2739 false
2740 );
2741 break;
2742 case 'localhour':
2743 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'H' ), true );
2744 break;
2745 case 'localweek':
2746 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2747 # int to remove the padding
2748 $value = $pageLang->formatNum( (int)MWTimestamp::getLocalInstance( $ts )->format( 'W' ) );
2749 break;
2750 case 'localdow':
2751 $value = $pageLang->formatNum( MWTimestamp::getLocalInstance( $ts )->format( 'w' ) );
2752 break;
2753 case 'numberofarticles':
2754 $value = $pageLang->formatNum( SiteStats::articles() );
2755 break;
2756 case 'numberoffiles':
2757 $value = $pageLang->formatNum( SiteStats::images() );
2758 break;
2759 case 'numberofusers':
2760 $value = $pageLang->formatNum( SiteStats::users() );
2761 break;
2762 case 'numberofactiveusers':
2763 $value = $pageLang->formatNum( SiteStats::activeUsers() );
2764 break;
2765 case 'numberofpages':
2766 $value = $pageLang->formatNum( SiteStats::pages() );
2767 break;
2768 case 'numberofadmins':
2769 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2770 break;
2771 case 'numberofedits':
2772 $value = $pageLang->formatNum( SiteStats::edits() );
2773 break;
2774 case 'currenttimestamp':
2775 $value = wfTimestamp( TS_MW, $ts );
2776 break;
2777 case 'localtimestamp':
2778 $value = MWTimestamp::getLocalInstance( $ts )->format( 'YmdHis' );
2779 break;
2780 case 'currentversion':
2781 $value = SpecialVersion::getVersion();
2782 break;
2783 case 'articlepath':
2784 return $wgArticlePath;
2785 case 'sitename':
2786 return $wgSitename;
2787 case 'server':
2788 return $wgServer;
2789 case 'servername':
2790 return $wgServerName;
2791 case 'scriptpath':
2792 return $wgScriptPath;
2793 case 'stylepath':
2794 return $wgStylePath;
2795 case 'directionmark':
2796 return $pageLang->getDirMark();
2797 case 'contentlanguage':
2798 global $wgLanguageCode;
2799 return $wgLanguageCode;
2800 case 'cascadingsources':
2801 $value = CoreParserFunctions::cascadingsources( $this );
2802 break;
2803 default:
2804 $ret = null;
2805 Hooks::run(
2806 'ParserGetVariableValueSwitch',
2807 [ &$this, &$this->mVarCache, &$index, &$ret, &$frame ]
2808 );
2809
2810 return $ret;
2811 }
2812
2813 if ( $index ) {
2814 $this->mVarCache[$index] = $value;
2815 }
2816
2817 return $value;
2818 }
2819
2820 /**
2821 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers