3 * PHP parser that converts wiki markup to HTML.
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.
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.
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
23 use MediaWiki\Linker\LinkRenderer
;
24 use MediaWiki\MediaWikiServices
;
25 use MediaWiki\Special\SpecialPageFactory
;
26 use Wikimedia\ScopedCallback
;
29 * @defgroup Parser Parser
33 * PHP Parser - Processes wiki markup (which uses a more user-friendly
34 * syntax, such as "[[link]]" for making links), and provides a one-way
35 * transformation of that wiki markup it into (X)HTML output / markup
36 * (which in turn the browser understands, and can display).
38 * There are seven main entry points into the Parser class:
41 * produces HTML output
42 * - Parser::preSaveTransform()
43 * produces altered wiki markup
44 * - Parser::preprocess()
45 * removes HTML comments and expands templates
46 * - Parser::cleanSig() and Parser::cleanSigInSig()
47 * cleans a signature before saving it to preferences
48 * - Parser::getSection()
49 * return the content of a section from an article for section editing
50 * - Parser::replaceSection()
51 * replaces a section by number inside an article
52 * - Parser::getPreloadText()
53 * removes <noinclude> sections and <includeonly> tags
55 * @warning $wgUser or $wgTitle or $wgRequest or $wgLang. Keep them away!
58 * $wgNamespacesWithSubpages
60 * @par Settings only within ParserOptions:
61 * $wgAllowExternalImages
62 * $wgAllowSpecialInclusion
70 * Update this version number when the ParserOutput format
71 * changes in an incompatible way, so the parser cache
72 * can automatically discard old data.
74 const VERSION
= '1.6.4';
77 * Update this version number when the output of serialiseHalfParsedText()
78 * changes in an incompatible way
80 const HALF_PARSED_VERSION
= 2;
82 # Flags for Parser::setFunctionHook
83 const SFH_NO_HASH
= 1;
84 const SFH_OBJECT_ARGS
= 2;
86 # Constants needed for external link processing
87 # Everything except bracket, space, or control characters
88 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
89 # as well as U+3000 is IDEOGRAPHIC SPACE for T21052
90 # \x{FFFD} is the Unicode replacement character, which Preprocessor_DOM
91 # uses to replace invalid HTML characters.
92 const EXT_LINK_URL_CLASS
= '[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]';
93 # Simplified expression to match an IPv4 or IPv6 address, or
94 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
95 const EXT_LINK_ADDR
= '(?:[0-9.]+|\\[(?i:[0-9a-f:.]+)\\]|[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}])';
96 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
97 // phpcs:ignore Generic.Files.LineLength
98 const EXT_IMAGE_REGEX
= '/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]+)
99 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
101 # Regular expression for a non-newline space
102 const SPACE_NOT_NL
= '(?:\t| |&\#0*160;|&\#[Xx]0*[Aa]0;|\p{Zs})';
104 # Flags for preprocessToDom
105 const PTD_FOR_INCLUSION
= 1;
107 # Allowed values for $this->mOutputType
108 # Parameter to startExternalParse().
109 const OT_HTML
= 1; # like parse()
110 const OT_WIKI
= 2; # like preSaveTransform()
111 const OT_PREPROCESS
= 3; # like preprocess()
113 const OT_PLAIN
= 4; # like extractSections() - portions of the original are returned unchanged.
116 * @var string Prefix and suffix for temporary replacement strings
117 * for the multipass parser.
119 * \x7f should never appear in input as it's disallowed in XML.
120 * Using it at the front also gives us a little extra robustness
121 * since it shouldn't match when butted up against identifier-like
124 * Must not consist of all title characters, or else it will change
125 * the behavior of <nowiki> in a link.
127 * Must have a character that needs escaping in attributes, otherwise
128 * someone could put a strip marker in an attribute, to get around
129 * escaping quote marks, and break out of the attribute. Thus we add
132 const MARKER_SUFFIX
= "-QINU`\"'\x7f";
133 const MARKER_PREFIX
= "\x7f'\"`UNIQ-";
135 # Markers used for wrapping the table of contents
136 const TOC_START
= '<mw:toc>';
137 const TOC_END
= '</mw:toc>';
139 /** @var int Assume that no output will later be saved this many seconds after parsing */
143 public $mTagHooks = [];
144 public $mTransparentTagHooks = [];
145 public $mFunctionHooks = [];
146 public $mFunctionSynonyms = [ 0 => [], 1 => [] ];
147 public $mFunctionTagHooks = [];
148 public $mStripList = [];
149 public $mDefaultStripList = [];
150 public $mVarCache = [];
151 public $mImageParams = [];
152 public $mImageParamsMagicArray = [];
153 public $mMarkerIndex = 0;
155 * @var bool Whether firstCallInit still needs to be called
157 public $mFirstCall = true;
159 # Initialised by initialiseVariables()
162 * @var MagicWordArray
167 * @var MagicWordArray
170 # Initialised in constructor
171 public $mConf, $mExtLinkBracketedRegex, $mUrlProtocols;
173 # Initialized in getPreprocessor()
174 /** @var Preprocessor */
175 public $mPreprocessor;
177 # Cleared with clearState():
189 public $mIncludeCount;
191 * @var LinkHolderArray
193 public $mLinkHolders;
196 public $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth;
197 public $mDefaultSort;
198 public $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
199 public $mExpensiveFunctionCount; # number of expensive parser function calls
200 public $mShowToc, $mForceTocPosition;
205 public $mUser; # User object; only used when doing pre-save transform
208 # These are variables reset at least once per parse regardless of $clearState
218 public $mTitle; # Title context, used for self-link rendering and similar things
219 public $mOutputType; # Output type, one of the OT_xxx constants
220 public $ot; # Shortcut alias, see setOutputType()
221 public $mRevisionObject; # The revision object of the specified revision ID
222 public $mRevisionId; # ID to display in {{REVISIONID}} tags
223 public $mRevisionTimestamp; # The timestamp of the specified revision ID
224 public $mRevisionUser; # User to display in {{REVISIONUSER}} tag
225 public $mRevisionSize; # Size to display in {{REVISIONSIZE}} variable
226 public $mRevIdForTs; # The revision ID which was used to fetch the timestamp
227 public $mInputSize = false; # For {{PAGESIZE}} on current page.
230 * @var string Deprecated accessor for the strip marker prefix.
231 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
233 public $mUniqPrefix = self
::MARKER_PREFIX
;
236 * @var array Array with the language name of each language link (i.e. the
237 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
238 * duplicate language links to the ParserOutput.
240 public $mLangLinkLanguages;
243 * @var MapCacheLRU|null
246 * A cache of the current revisions of titles. Keys are $title->getPrefixedDbKey()
248 public $currentRevisionCache;
251 * @var bool|string Recursive call protection.
252 * This variable should be treated as if it were private.
254 public $mInParse = false;
256 /** @var SectionProfiler */
257 protected $mProfiler;
262 protected $mLinkRenderer;
264 /** @var MagicWordFactory */
265 private $magicWordFactory;
270 /** @var ParserFactory */
273 /** @var SpecialPageFactory */
274 private $specialPageFactory;
277 * @param array $conf See $wgParserConf documentation
278 * @param MagicWordFactory|null $magicWordFactory
279 * @param Language|null $contLang Content language
280 * @param ParserFactory|null $factory
281 * @param string|null $urlProtocols As returned from wfUrlProtocols()
282 * @param SpecialPageFactory|null $spFactory
284 public function __construct(
285 array $conf = [], MagicWordFactory
$magicWordFactory = null, Language
$contLang = null,
286 ParserFactory
$factory = null, $urlProtocols = null, SpecialPageFactory
$spFactory = null
288 $this->mConf
= $conf;
289 $this->mUrlProtocols
= $urlProtocols ??
wfUrlProtocols();
290 $this->mExtLinkBracketedRegex
= '/\[(((?i)' . $this->mUrlProtocols
. ')' .
291 self
::EXT_LINK_ADDR
.
292 self
::EXT_LINK_URL_CLASS
. '*)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F\\x{FFFD}]*?)\]/Su';
293 if ( isset( $conf['preprocessorClass'] ) ) {
294 $this->mPreprocessorClass
= $conf['preprocessorClass'];
295 } elseif ( defined( 'HPHP_VERSION' ) ) {
296 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
297 $this->mPreprocessorClass
= Preprocessor_Hash
::class;
298 } elseif ( extension_loaded( 'domxml' ) ) {
299 # PECL extension that conflicts with the core DOM extension (T15770)
300 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
301 $this->mPreprocessorClass
= Preprocessor_Hash
::class;
302 } elseif ( extension_loaded( 'dom' ) ) {
303 $this->mPreprocessorClass
= Preprocessor_DOM
::class;
305 $this->mPreprocessorClass
= Preprocessor_Hash
::class;
307 wfDebug( __CLASS__
. ": using preprocessor: {$this->mPreprocessorClass}\n" );
309 $services = MediaWikiServices
::getInstance();
310 $this->magicWordFactory
= $magicWordFactory ??
311 $services->getMagicWordFactory();
313 $this->contLang
= $contLang ??
$services->getContentLanguage();
315 $this->factory
= $factory ??
$services->getParserFactory();
316 $this->specialPageFactory
= $spFactory ??
$services->getSpecialPageFactory();
320 * Reduce memory usage to reduce the impact of circular references
322 public function __destruct() {
323 if ( isset( $this->mLinkHolders
) ) {
324 unset( $this->mLinkHolders
);
326 foreach ( $this as $name => $value ) {
327 unset( $this->$name );
332 * Allow extensions to clean up when the parser is cloned
334 public function __clone() {
335 $this->mInParse
= false;
337 // T58226: When you create a reference "to" an object field, that
338 // makes the object field itself be a reference too (until the other
339 // reference goes out of scope). When cloning, any field that's a
340 // reference is copied as a reference in the new object. Both of these
341 // are defined PHP5 behaviors, as inconvenient as it is for us when old
342 // hooks from PHP4 days are passing fields by reference.
343 foreach ( [ 'mStripState', 'mVarCache' ] as $k ) {
344 // Make a non-reference copy of the field, then rebind the field to
345 // reference the new copy.
351 Hooks
::run( 'ParserCloned', [ $this ] );
355 * Do various kinds of initialisation on the first call of the parser
357 public function firstCallInit() {
358 if ( !$this->mFirstCall
) {
361 $this->mFirstCall
= false;
363 CoreParserFunctions
::register( $this );
364 CoreTagHooks
::register( $this );
365 $this->initialiseVariables();
367 // Avoid PHP 7.1 warning from passing $this by reference
369 Hooks
::run( 'ParserFirstCallInit', [ &$parser ] );
377 public function clearState() {
378 $this->firstCallInit();
379 $this->mOutput
= new ParserOutput
;
380 $this->mOptions
->registerWatcher( [ $this->mOutput
, 'recordOption' ] );
381 $this->mAutonumber
= 0;
382 $this->mIncludeCount
= [];
383 $this->mLinkHolders
= new LinkHolderArray( $this );
385 $this->mRevisionObject
= $this->mRevisionTimestamp
=
386 $this->mRevisionId
= $this->mRevisionUser
= $this->mRevisionSize
= null;
387 $this->mVarCache
= [];
389 $this->mLangLinkLanguages
= [];
390 $this->currentRevisionCache
= null;
392 $this->mStripState
= new StripState( $this );
394 # Clear these on every parse, T6549
395 $this->mTplRedirCache
= $this->mTplDomCache
= [];
397 $this->mShowToc
= true;
398 $this->mForceTocPosition
= false;
399 $this->mIncludeSizes
= [
403 $this->mPPNodeCount
= 0;
404 $this->mGeneratedPPNodeCount
= 0;
405 $this->mHighestExpansionDepth
= 0;
406 $this->mDefaultSort
= false;
407 $this->mHeadings
= [];
408 $this->mDoubleUnderscores
= [];
409 $this->mExpensiveFunctionCount
= 0;
412 if ( isset( $this->mPreprocessor
) && $this->mPreprocessor
->parser
!== $this ) {
413 $this->mPreprocessor
= null;
416 $this->mProfiler
= new SectionProfiler();
418 // Avoid PHP 7.1 warning from passing $this by reference
420 Hooks
::run( 'ParserClearState', [ &$parser ] );
424 * Convert wikitext to HTML
425 * Do not call this function recursively.
427 * @param string $text Text we want to parse
428 * @param Title $title
429 * @param ParserOptions $options
430 * @param bool $linestart
431 * @param bool $clearState
432 * @param int|null $revid Number to pass in {{REVISIONID}}
433 * @return ParserOutput A ParserOutput
435 public function parse(
436 $text, Title
$title, ParserOptions
$options,
437 $linestart = true, $clearState = true, $revid = null
440 // We use U+007F DELETE to construct strip markers, so we have to make
441 // sure that this character does not occur in the input text.
442 $text = strtr( $text, "\x7f", "?" );
443 $magicScopeVariable = $this->lock();
445 // Strip U+0000 NULL (T159174)
446 $text = str_replace( "\000", '', $text );
448 $this->startParse( $title, $options, self
::OT_HTML
, $clearState );
450 $this->currentRevisionCache
= null;
451 $this->mInputSize
= strlen( $text );
452 if ( $this->mOptions
->getEnableLimitReport() ) {
453 $this->mOutput
->resetParseStartTime();
456 $oldRevisionId = $this->mRevisionId
;
457 $oldRevisionObject = $this->mRevisionObject
;
458 $oldRevisionTimestamp = $this->mRevisionTimestamp
;
459 $oldRevisionUser = $this->mRevisionUser
;
460 $oldRevisionSize = $this->mRevisionSize
;
461 if ( $revid !== null ) {
462 $this->mRevisionId
= $revid;
463 $this->mRevisionObject
= null;
464 $this->mRevisionTimestamp
= null;
465 $this->mRevisionUser
= null;
466 $this->mRevisionSize
= null;
469 // Avoid PHP 7.1 warning from passing $this by reference
471 Hooks
::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState
] );
473 Hooks
::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState
] );
474 $text = $this->internalParse( $text );
475 Hooks
::run( 'ParserAfterParse', [ &$parser, &$text, &$this->mStripState
] );
477 $text = $this->internalParseHalfParsed( $text, true, $linestart );
480 * A converted title will be provided in the output object if title and
481 * content conversion are enabled, the article text does not contain
482 * a conversion-suppressing double-underscore tag, and no
483 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
484 * automatic link conversion.
486 if ( !( $options->getDisableTitleConversion()
487 ||
isset( $this->mDoubleUnderscores
['nocontentconvert'] )
488 ||
isset( $this->mDoubleUnderscores
['notitleconvert'] )
489 ||
$this->mOutput
->getDisplayTitle() !== false )
491 $convruletitle = $this->getTargetLanguage()->getConvRuleTitle();
492 if ( $convruletitle ) {
493 $this->mOutput
->setTitleText( $convruletitle );
495 $titleText = $this->getTargetLanguage()->convertTitle( $title );
496 $this->mOutput
->setTitleText( $titleText );
500 # Compute runtime adaptive expiry if set
501 $this->mOutput
->finalizeAdaptiveCacheExpiry();
503 # Warn if too many heavyweight parser functions were used
504 if ( $this->mExpensiveFunctionCount
> $this->mOptions
->getExpensiveParserFunctionLimit() ) {
505 $this->limitationWarn( 'expensive-parserfunction',
506 $this->mExpensiveFunctionCount
,
507 $this->mOptions
->getExpensiveParserFunctionLimit()
511 # Information on limits, for the benefit of users who try to skirt them
512 if ( $this->mOptions
->getEnableLimitReport() ) {
513 $text .= $this->makeLimitReport();
516 # Wrap non-interface parser output in a <div> so it can be targeted
518 $class = $this->mOptions
->getWrapOutputClass();
519 if ( $class !== false && !$this->mOptions
->getInterfaceMessage() ) {
520 $this->mOutput
->addWrapperDivClass( $class );
523 $this->mOutput
->setText( $text );
525 $this->mRevisionId
= $oldRevisionId;
526 $this->mRevisionObject
= $oldRevisionObject;
527 $this->mRevisionTimestamp
= $oldRevisionTimestamp;
528 $this->mRevisionUser
= $oldRevisionUser;
529 $this->mRevisionSize
= $oldRevisionSize;
530 $this->mInputSize
= false;
531 $this->currentRevisionCache
= null;
533 return $this->mOutput
;
537 * Set the limit report data in the current ParserOutput, and return the
538 * limit report HTML comment.
542 protected function makeLimitReport() {
543 global $wgShowHostnames;
545 $maxIncludeSize = $this->mOptions
->getMaxIncludeSize();
547 $cpuTime = $this->mOutput
->getTimeSinceStart( 'cpu' );
548 if ( $cpuTime !== null ) {
549 $this->mOutput
->setLimitReportData( 'limitreport-cputime',
550 sprintf( "%.3f", $cpuTime )
554 $wallTime = $this->mOutput
->getTimeSinceStart( 'wall' );
555 $this->mOutput
->setLimitReportData( 'limitreport-walltime',
556 sprintf( "%.3f", $wallTime )
559 $this->mOutput
->setLimitReportData( 'limitreport-ppvisitednodes',
560 [ $this->mPPNodeCount
, $this->mOptions
->getMaxPPNodeCount() ]
562 $this->mOutput
->setLimitReportData( 'limitreport-ppgeneratednodes',
563 [ $this->mGeneratedPPNodeCount
, $this->mOptions
->getMaxGeneratedPPNodeCount() ]
565 $this->mOutput
->setLimitReportData( 'limitreport-postexpandincludesize',
566 [ $this->mIncludeSizes
['post-expand'], $maxIncludeSize ]
568 $this->mOutput
->setLimitReportData( 'limitreport-templateargumentsize',
569 [ $this->mIncludeSizes
['arg'], $maxIncludeSize ]
571 $this->mOutput
->setLimitReportData( 'limitreport-expansiondepth',
572 [ $this->mHighestExpansionDepth
, $this->mOptions
->getMaxPPExpandDepth() ]
574 $this->mOutput
->setLimitReportData( 'limitreport-expensivefunctioncount',
575 [ $this->mExpensiveFunctionCount
, $this->mOptions
->getExpensiveParserFunctionLimit() ]
578 foreach ( $this->mStripState
->getLimitReport() as list( $key, $value ) ) {
579 $this->mOutput
->setLimitReportData( $key, $value );
582 Hooks
::run( 'ParserLimitReportPrepare', [ $this, $this->mOutput
] );
584 $limitReport = "NewPP limit report\n";
585 if ( $wgShowHostnames ) {
586 $limitReport .= 'Parsed by ' . wfHostname() . "\n";
588 $limitReport .= 'Cached time: ' . $this->mOutput
->getCacheTime() . "\n";
589 $limitReport .= 'Cache expiry: ' . $this->mOutput
->getCacheExpiry() . "\n";
590 $limitReport .= 'Dynamic content: ' .
591 ( $this->mOutput
->hasDynamicContent() ?
'true' : 'false' ) .
594 foreach ( $this->mOutput
->getLimitReportData() as $key => $value ) {
595 if ( Hooks
::run( 'ParserLimitReportFormat',
596 [ $key, &$value, &$limitReport, false, false ]
598 $keyMsg = wfMessage( $key )->inLanguage( 'en' )->useDatabase( false );
599 $valueMsg = wfMessage( [ "$key-value-text", "$key-value" ] )
600 ->inLanguage( 'en' )->useDatabase( false );
601 if ( !$valueMsg->exists() ) {
602 $valueMsg = new RawMessage( '$1' );
604 if ( !$keyMsg->isDisabled() && !$valueMsg->isDisabled() ) {
605 $valueMsg->params( $value );
606 $limitReport .= "{$keyMsg->text()}: {$valueMsg->text()}\n";
610 // Since we're not really outputting HTML, decode the entities and
611 // then re-encode the things that need hiding inside HTML comments.
612 $limitReport = htmlspecialchars_decode( $limitReport );
613 // Run deprecated hook
614 Hooks
::run( 'ParserLimitReport', [ $this, &$limitReport ], '1.22' );
616 // Sanitize for comment. Note '‐' in the replacement is U+2010,
617 // which looks much like the problematic '-'.
618 $limitReport = str_replace( [ '-', '&' ], [ '‐', '&' ], $limitReport );
619 $text = "\n<!-- \n$limitReport-->\n";
621 // Add on template profiling data in human/machine readable way
622 $dataByFunc = $this->mProfiler
->getFunctionStats();
623 uasort( $dataByFunc, function ( $a, $b ) {
624 return $b['real'] <=> $a['real']; // descending order
627 foreach ( array_slice( $dataByFunc, 0, 10 ) as $item ) {
628 $profileReport[] = sprintf( "%6.2f%% %8.3f %6d %s",
629 $item['%real'], $item['real'], $item['calls'],
630 htmlspecialchars( $item['name'] ) );
632 $text .= "<!--\nTransclusion expansion time report (%,ms,calls,template)\n";
633 $text .= implode( "\n", $profileReport ) . "\n-->\n";
635 $this->mOutput
->setLimitReportData( 'limitreport-timingprofile', $profileReport );
637 // Add other cache related metadata
638 if ( $wgShowHostnames ) {
639 $this->mOutput
->setLimitReportData( 'cachereport-origin', wfHostname() );
641 $this->mOutput
->setLimitReportData( 'cachereport-timestamp',
642 $this->mOutput
->getCacheTime() );
643 $this->mOutput
->setLimitReportData( 'cachereport-ttl',
644 $this->mOutput
->getCacheExpiry() );
645 $this->mOutput
->setLimitReportData( 'cachereport-transientcontent',
646 $this->mOutput
->hasDynamicContent() );
648 if ( $this->mGeneratedPPNodeCount
> $this->mOptions
->getMaxGeneratedPPNodeCount() / 10 ) {
649 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount
. ' ' .
650 $this->mTitle
->getPrefixedDBkey() );
656 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
657 * can be called from an extension tag hook.
659 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
660 * instead, which means that lists and links have not been fully parsed yet,
661 * and strip markers are still present.
663 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
665 * Use this function if you're a parser tag hook and you want to parse
666 * wikitext before or after applying additional transformations, and you
667 * intend to *return the result as hook output*, which will cause it to go
668 * through the rest of parsing process automatically.
670 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
671 * $text are not expanded
673 * @param string $text Text extension wants to have parsed
674 * @param bool|PPFrame $frame The frame to use for expanding any template variables
675 * @return string UNSAFE half-parsed HTML
677 public function recursiveTagParse( $text, $frame = false ) {
678 // Avoid PHP 7.1 warning from passing $this by reference
680 Hooks
::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState
] );
681 Hooks
::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState
] );
682 $text = $this->internalParse( $text, false, $frame );
687 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
688 * point can be called from an extension tag hook.
690 * The output of this function is fully-parsed HTML that is safe for output.
691 * If you're a parser tag hook, you might want to use recursiveTagParse()
694 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
695 * $text are not expanded
699 * @param string $text Text extension wants to have parsed
700 * @param bool|PPFrame $frame The frame to use for expanding any template variables
701 * @return string Fully parsed HTML
703 public function recursiveTagParseFully( $text, $frame = false ) {
704 $text = $this->recursiveTagParse( $text, $frame );
705 $text = $this->internalParseHalfParsed( $text, false );
710 * Expand templates and variables in the text, producing valid, static wikitext.
711 * Also removes comments.
712 * Do not call this function recursively.
713 * @param string $text
714 * @param Title|null $title
715 * @param ParserOptions $options
716 * @param int|null $revid
717 * @param bool|PPFrame $frame
718 * @return mixed|string
720 public function preprocess( $text, Title
$title = null,
721 ParserOptions
$options, $revid = null, $frame = false
723 $magicScopeVariable = $this->lock();
724 $this->startParse( $title, $options, self
::OT_PREPROCESS
, true );
725 if ( $revid !== null ) {
726 $this->mRevisionId
= $revid;
728 // Avoid PHP 7.1 warning from passing $this by reference
730 Hooks
::run( 'ParserBeforeStrip', [ &$parser, &$text, &$this->mStripState
] );
731 Hooks
::run( 'ParserAfterStrip', [ &$parser, &$text, &$this->mStripState
] );
732 $text = $this->replaceVariables( $text, $frame );
733 $text = $this->mStripState
->unstripBoth( $text );
738 * Recursive parser entry point that can be called from an extension tag
741 * @param string $text Text to be expanded
742 * @param bool|PPFrame $frame The frame to use for expanding any template variables
746 public function recursivePreprocess( $text, $frame = false ) {
747 $text = $this->replaceVariables( $text, $frame );
748 $text = $this->mStripState
->unstripBoth( $text );
753 * Process the wikitext for the "?preload=" feature. (T7210)
755 * "<noinclude>", "<includeonly>" etc. are parsed as for template
756 * transclusion, comments, templates, arguments, tags hooks and parser
757 * functions are untouched.
759 * @param string $text
760 * @param Title $title
761 * @param ParserOptions $options
762 * @param array $params
765 public function getPreloadText( $text, Title
$title, ParserOptions
$options, $params = [] ) {
766 $msg = new RawMessage( $text );
767 $text = $msg->params( $params )->plain();
769 # Parser (re)initialisation
770 $magicScopeVariable = $this->lock();
771 $this->startParse( $title, $options, self
::OT_PLAIN
, true );
773 $flags = PPFrame
::NO_ARGS | PPFrame
::NO_TEMPLATES
;
774 $dom = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
775 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
776 $text = $this->mStripState
->unstripBoth( $text );
781 * Set the current user.
782 * Should only be used when doing pre-save transform.
784 * @param User|null $user User object or null (to reset)
786 public function setUser( $user ) {
787 $this->mUser
= $user;
791 * Set the context title
795 public function setTitle( $t ) {
797 $t = Title
::newFromText( 'NO TITLE' );
800 if ( $t->hasFragment() ) {
801 # Strip the fragment to avoid various odd effects
802 $this->mTitle
= $t->createFragmentTarget( '' );
809 * Accessor for the Title object
813 public function getTitle() {
814 return $this->mTitle
;
818 * Accessor/mutator for the Title object
820 * @param Title|null $x Title object or null to just get the current one
823 public function Title( $x = null ) {
824 return wfSetVar( $this->mTitle
, $x );
828 * Set the output type
830 * @param int $ot New value
832 public function setOutputType( $ot ) {
833 $this->mOutputType
= $ot;
836 'html' => $ot == self
::OT_HTML
,
837 'wiki' => $ot == self
::OT_WIKI
,
838 'pre' => $ot == self
::OT_PREPROCESS
,
839 'plain' => $ot == self
::OT_PLAIN
,
844 * Accessor/mutator for the output type
846 * @param int|null $x New value or null to just get the current one
849 public function OutputType( $x = null ) {
850 return wfSetVar( $this->mOutputType
, $x );
854 * Get the ParserOutput object
856 * @return ParserOutput
858 public function getOutput() {
859 return $this->mOutput
;
863 * Get the ParserOptions object
865 * @return ParserOptions
867 public function getOptions() {
868 return $this->mOptions
;
872 * Accessor/mutator for the ParserOptions object
874 * @param ParserOptions|null $x New value or null to just get the current one
875 * @return ParserOptions Current ParserOptions object
877 public function Options( $x = null ) {
878 return wfSetVar( $this->mOptions
, $x );
884 public function nextLinkID() {
885 return $this->mLinkID++
;
891 public function setLinkID( $id ) {
892 $this->mLinkID
= $id;
896 * Get a language object for use in parser functions such as {{FORMATNUM:}}
899 public function getFunctionLang() {
900 return $this->getTargetLanguage();
904 * Get the target language for the content being parsed. This is usually the
905 * language that the content is in.
909 * @throws MWException
912 public function getTargetLanguage() {
913 $target = $this->mOptions
->getTargetLanguage();
915 if ( $target !== null ) {
917 } elseif ( $this->mOptions
->getInterfaceMessage() ) {
918 return $this->mOptions
->getUserLangObj();
919 } elseif ( is_null( $this->mTitle
) ) {
920 throw new MWException( __METHOD__
. ': $this->mTitle is null' );
923 return $this->mTitle
->getPageLanguage();
927 * Get the language object for language conversion
928 * @deprecated since 1.32, just use getTargetLanguage()
929 * @return Language|null
931 public function getConverterLanguage() {
932 return $this->getTargetLanguage();
936 * Get a User object either from $this->mUser, if set, or from the
937 * ParserOptions object otherwise
941 public function getUser() {
942 if ( !is_null( $this->mUser
) ) {
945 return $this->mOptions
->getUser();
949 * Get a preprocessor object
951 * @return Preprocessor
953 public function getPreprocessor() {
954 if ( !isset( $this->mPreprocessor
) ) {
955 $class = $this->mPreprocessorClass
;
956 $this->mPreprocessor
= new $class( $this );
958 return $this->mPreprocessor
;
962 * Get a LinkRenderer instance to make links with
965 * @return LinkRenderer
967 public function getLinkRenderer() {
968 if ( !$this->mLinkRenderer
) {
969 $this->mLinkRenderer
= MediaWikiServices
::getInstance()
970 ->getLinkRendererFactory()->create();
971 $this->mLinkRenderer
->setStubThreshold(
972 $this->getOptions()->getStubThreshold()
976 return $this->mLinkRenderer
;
980 * Get the MagicWordFactory that this Parser is using
983 * @return MagicWordFactory
985 public function getMagicWordFactory() {
986 return $this->magicWordFactory
;
990 * Get the content language that this Parser is using
995 public function getContentLanguage() {
996 return $this->contLang
;
1000 * Replaces all occurrences of HTML-style comments and the given tags
1001 * in the text with a random marker and returns the next text. The output
1002 * parameter $matches will be an associative array filled with data in
1009 * [ 'param' => 'x' ],
1010 * '<element param="x">tag content</element>' ]
1013 * @param array $elements List of element names. Comments are always extracted.
1014 * @param string $text Source text string.
1015 * @param array &$matches Out parameter, Array: extracted tags
1016 * @return string Stripped text
1018 public static function extractTagsAndParams( $elements, $text, &$matches ) {
1023 $taglist = implode( '|', $elements );
1024 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
1026 while ( $text != '' ) {
1027 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE
);
1029 if ( count( $p ) < 5 ) {
1032 if ( count( $p ) > 5 ) {
1041 $attributes = $p[2];
1046 $marker = self
::MARKER_PREFIX
. "-$element-" . sprintf( '%08X', $n++
) . self
::MARKER_SUFFIX
;
1047 $stripped .= $marker;
1049 if ( $close === '/>' ) {
1050 # Empty element tag, <tag />
1055 if ( $element === '!--' ) {
1058 $end = "/(<\\/$element\\s*>)/i";
1060 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE
);
1062 if ( count( $q ) < 3 ) {
1063 # No end tag -- let it run out to the end of the text.
1072 $matches[$marker] = [ $element,
1074 Sanitizer
::decodeTagAttributes( $attributes ),
1075 "<$element$attributes$close$content$tail" ];
1081 * Get a list of strippable XML-like elements
1085 public function getStripList() {
1086 return $this->mStripList
;
1090 * Add an item to the strip state
1091 * Returns the unique tag which must be inserted into the stripped text
1092 * The tag will be replaced with the original text in unstrip()
1094 * @param string $text
1098 public function insertStripItem( $text ) {
1099 $marker = self
::MARKER_PREFIX
. "-item-{$this->mMarkerIndex}-" . self
::MARKER_SUFFIX
;
1100 $this->mMarkerIndex++
;
1101 $this->mStripState
->addGeneral( $marker, $text );
1106 * parse the wiki syntax used to render tables
1109 * @param string $text
1112 public function doTableStuff( $text ) {
1113 $lines = StringUtils
::explode( "\n", $text );
1115 $td_history = []; # Is currently a td tag open?
1116 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1117 $tr_history = []; # Is currently a tr tag open?
1118 $tr_attributes = []; # history of tr attributes
1119 $has_opened_tr = []; # Did this table open a <tr> element?
1120 $indent_level = 0; # indent level of the table
1122 foreach ( $lines as $outLine ) {
1123 $line = trim( $outLine );
1125 if ( $line === '' ) { # empty line, go to next line
1126 $out .= $outLine . "\n";
1130 $first_character = $line[0];
1131 $first_two = substr( $line, 0, 2 );
1134 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1135 # First check if we are starting a new table
1136 $indent_level = strlen( $matches[1] );
1138 $attributes = $this->mStripState
->unstripBoth( $matches[2] );
1139 $attributes = Sanitizer
::fixTagAttributes( $attributes, 'table' );
1141 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1142 array_push( $td_history, false );
1143 array_push( $last_tag_history, '' );
1144 array_push( $tr_history, false );
1145 array_push( $tr_attributes, '' );
1146 array_push( $has_opened_tr, false );
1147 } elseif ( count( $td_history ) == 0 ) {
1148 # Don't do any of the following
1149 $out .= $outLine . "\n";
1151 } elseif ( $first_two === '|}' ) {
1152 # We are ending a table
1153 $line = '</table>' . substr( $line, 2 );
1154 $last_tag = array_pop( $last_tag_history );
1156 if ( !array_pop( $has_opened_tr ) ) {
1157 $line = "<tr><td></td></tr>{$line}";
1160 if ( array_pop( $tr_history ) ) {
1161 $line = "</tr>{$line}";
1164 if ( array_pop( $td_history ) ) {
1165 $line = "</{$last_tag}>{$line}";
1167 array_pop( $tr_attributes );
1168 if ( $indent_level > 0 ) {
1169 $outLine = rtrim( $line ) . str_repeat( '</dd></dl>', $indent_level );
1173 } elseif ( $first_two === '|-' ) {
1174 # Now we have a table row
1175 $line = preg_replace( '#^\|-+#', '', $line );
1177 # Whats after the tag is now only attributes
1178 $attributes = $this->mStripState
->unstripBoth( $line );
1179 $attributes = Sanitizer
::fixTagAttributes( $attributes, 'tr' );
1180 array_pop( $tr_attributes );
1181 array_push( $tr_attributes, $attributes );
1184 $last_tag = array_pop( $last_tag_history );
1185 array_pop( $has_opened_tr );
1186 array_push( $has_opened_tr, true );
1188 if ( array_pop( $tr_history ) ) {
1192 if ( array_pop( $td_history ) ) {
1193 $line = "</{$last_tag}>{$line}";
1197 array_push( $tr_history, false );
1198 array_push( $td_history, false );
1199 array_push( $last_tag_history, '' );
1200 } elseif ( $first_character === '|'
1201 ||
$first_character === '!'
1202 ||
$first_two === '|+'
1204 # This might be cell elements, td, th or captions
1205 if ( $first_two === '|+' ) {
1206 $first_character = '+';
1207 $line = substr( $line, 2 );
1209 $line = substr( $line, 1 );
1212 // Implies both are valid for table headings.
1213 if ( $first_character === '!' ) {
1214 $line = StringUtils
::replaceMarkup( '!!', '||', $line );
1217 # Split up multiple cells on the same line.
1218 # FIXME : This can result in improper nesting of tags processed
1219 # by earlier parser steps.
1220 $cells = explode( '||', $line );
1224 # Loop through each table cell
1225 foreach ( $cells as $cell ) {
1227 if ( $first_character !== '+' ) {
1228 $tr_after = array_pop( $tr_attributes );
1229 if ( !array_pop( $tr_history ) ) {
1230 $previous = "<tr{$tr_after}>\n";
1232 array_push( $tr_history, true );
1233 array_push( $tr_attributes, '' );
1234 array_pop( $has_opened_tr );
1235 array_push( $has_opened_tr, true );
1238 $last_tag = array_pop( $last_tag_history );
1240 if ( array_pop( $td_history ) ) {
1241 $previous = "</{$last_tag}>\n{$previous}";
1244 if ( $first_character === '|' ) {
1246 } elseif ( $first_character === '!' ) {
1248 } elseif ( $first_character === '+' ) {
1249 $last_tag = 'caption';
1254 array_push( $last_tag_history, $last_tag );
1256 # A cell could contain both parameters and data
1257 $cell_data = explode( '|', $cell, 2 );
1259 # T2553: Note that a '|' inside an invalid link should not
1260 # be mistaken as delimiting cell parameters
1261 # Bug T153140: Neither should language converter markup.
1262 if ( preg_match( '/\[\[|-\{/', $cell_data[0] ) === 1 ) {
1263 $cell = "{$previous}<{$last_tag}>" . trim( $cell );
1264 } elseif ( count( $cell_data ) == 1 ) {
1265 // Whitespace in cells is trimmed
1266 $cell = "{$previous}<{$last_tag}>" . trim( $cell_data[0] );
1268 $attributes = $this->mStripState
->unstripBoth( $cell_data[0] );
1269 $attributes = Sanitizer
::fixTagAttributes( $attributes, $last_tag );
1270 // Whitespace in cells is trimmed
1271 $cell = "{$previous}<{$last_tag}{$attributes}>" . trim( $cell_data[1] );
1275 array_push( $td_history, true );
1278 $out .= $outLine . "\n";
1281 # Closing open td, tr && table
1282 while ( count( $td_history ) > 0 ) {
1283 if ( array_pop( $td_history ) ) {
1286 if ( array_pop( $tr_history ) ) {
1289 if ( !array_pop( $has_opened_tr ) ) {
1290 $out .= "<tr><td></td></tr>\n";
1293 $out .= "</table>\n";
1296 # Remove trailing line-ending (b/c)
1297 if ( substr( $out, -1 ) === "\n" ) {
1298 $out = substr( $out, 0, -1 );
1301 # special case: don't return empty table
1302 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1310 * Helper function for parse() that transforms wiki markup into half-parsed
1311 * HTML. Only called for $mOutputType == self::OT_HTML.
1315 * @param string $text The text to parse
1316 * @param bool $isMain Whether this is being called from the main parse() function
1317 * @param PPFrame|bool $frame A pre-processor frame
1321 public function internalParse( $text, $isMain = true, $frame = false ) {
1324 // Avoid PHP 7.1 warning from passing $this by reference
1327 # Hook to suspend the parser in this state
1328 if ( !Hooks
::run( 'ParserBeforeInternalParse', [ &$parser, &$text, &$this->mStripState
] ) ) {
1332 # if $frame is provided, then use $frame for replacing any variables
1334 # use frame depth to infer how include/noinclude tags should be handled
1335 # depth=0 means this is the top-level document; otherwise it's an included document
1336 if ( !$frame->depth
) {
1339 $flag = self
::PTD_FOR_INCLUSION
;
1341 $dom = $this->preprocessToDom( $text, $flag );
1342 $text = $frame->expand( $dom );
1344 # if $frame is not provided, then use old-style replaceVariables
1345 $text = $this->replaceVariables( $text );
1348 Hooks
::run( 'InternalParseBeforeSanitize', [ &$parser, &$text, &$this->mStripState
] );
1349 $text = Sanitizer
::removeHTMLtags(
1351 [ $this, 'attributeStripCallback' ],
1353 array_keys( $this->mTransparentTagHooks
),
1355 [ $this, 'addTrackingCategory' ]
1357 Hooks
::run( 'InternalParseBeforeLinks', [ &$parser, &$text, &$this->mStripState
] );
1359 # Tables need to come after variable replacement for things to work
1360 # properly; putting them before other transformations should keep
1361 # exciting things like link expansions from showing up in surprising
1363 $text = $this->doTableStuff( $text );
1365 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1367 $text = $this->doDoubleUnderscore( $text );
1369 $text = $this->doHeadings( $text );
1370 $text = $this->replaceInternalLinks( $text );
1371 $text = $this->doAllQuotes( $text );
1372 $text = $this->replaceExternalLinks( $text );
1374 # replaceInternalLinks may sometimes leave behind
1375 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1376 $text = str_replace( self
::MARKER_PREFIX
. 'NOPARSE', '', $text );
1378 $text = $this->doMagicLinks( $text );
1379 $text = $this->formatHeadings( $text, $origText, $isMain );
1385 * Helper function for parse() that transforms half-parsed HTML into fully
1388 * @param string $text
1389 * @param bool $isMain
1390 * @param bool $linestart
1393 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1394 $text = $this->mStripState
->unstripGeneral( $text );
1396 // Avoid PHP 7.1 warning from passing $this by reference
1400 Hooks
::run( 'ParserAfterUnstrip', [ &$parser, &$text ] );
1403 # Clean up special characters, only run once, next-to-last before doBlockLevels
1404 $text = Sanitizer
::armorFrenchSpaces( $text );
1406 $text = $this->doBlockLevels( $text, $linestart );
1408 $this->replaceLinkHolders( $text );
1411 * The input doesn't get language converted if
1413 * b) Content isn't converted
1414 * c) It's a conversion table
1415 * d) it is an interface message (which is in the user language)
1417 if ( !( $this->mOptions
->getDisableContentConversion()
1418 ||
isset( $this->mDoubleUnderscores
['nocontentconvert'] ) )
1420 if ( !$this->mOptions
->getInterfaceMessage() ) {
1421 # The position of the convert() call should not be changed. it
1422 # assumes that the links are all replaced and the only thing left
1423 # is the <nowiki> mark.
1424 $text = $this->getTargetLanguage()->convert( $text );
1428 $text = $this->mStripState
->unstripNoWiki( $text );
1431 Hooks
::run( 'ParserBeforeTidy', [ &$parser, &$text ] );
1434 $text = $this->replaceTransparentTags( $text );
1435 $text = $this->mStripState
->unstripGeneral( $text );
1437 $text = Sanitizer
::normalizeCharReferences( $text );
1439 if ( MWTidy
::isEnabled() ) {
1440 if ( $this->mOptions
->getTidy() ) {
1441 $text = MWTidy
::tidy( $text );
1444 # attempt to sanitize at least some nesting problems
1445 # (T4702 and quite a few others)
1447 # ''Something [http://www.cool.com cool''] -->
1448 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1449 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1450 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1451 # fix up an anchor inside another anchor, only
1452 # at least for a single single nested link (T5695)
1453 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1454 '\\1\\2</a>\\3</a>\\1\\4</a>',
1455 # fix div inside inline elements- doBlockLevels won't wrap a line which
1456 # contains a div, so fix it up here; replace
1457 # div with escaped text
1458 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1459 '\\1\\3<div\\5>\\6</div>\\8\\9',
1460 # remove empty italic or bold tag pairs, some
1461 # introduced by rules above
1462 '/<([bi])><\/\\1>/' => '',
1465 $text = preg_replace(
1466 array_keys( $tidyregs ),
1467 array_values( $tidyregs ),
1472 Hooks
::run( 'ParserAfterTidy', [ &$parser, &$text ] );
1479 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1480 * magic external links.
1485 * @param string $text
1489 public function doMagicLinks( $text ) {
1490 $prots = wfUrlProtocolsWithoutProtRel();
1491 $urlChar = self
::EXT_LINK_URL_CLASS
;
1492 $addr = self
::EXT_LINK_ADDR
;
1493 $space = self
::SPACE_NOT_NL
; # non-newline space
1494 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1495 $spaces = "$space++"; # possessive match of 1 or more spaces
1496 $text = preg_replace_callback(
1498 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1499 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1500 (\b # m[3]: Free external links
1502 ($addr$urlChar*) # m[4]: Post-protocol path
1504 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1506 \bISBN $spaces ( # m[6]: ISBN, capture number
1507 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1508 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1509 [0-9Xx] # check digit
1511 )!xu", [ $this, 'magicLinkCallback' ], $text );
1516 * @throws MWException
1518 * @return string HTML
1520 public function magicLinkCallback( $m ) {
1521 if ( isset( $m[1] ) && $m[1] !== '' ) {
1524 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1527 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1528 # Free external link
1529 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1530 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1532 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1533 if ( !$this->mOptions
->getMagicRFCLinks() ) {
1538 $cssClass = 'mw-magiclink-rfc';
1539 $trackingCat = 'magiclink-tracking-rfc';
1541 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1542 if ( !$this->mOptions
->getMagicPMIDLinks() ) {
1546 $urlmsg = 'pubmedurl';
1547 $cssClass = 'mw-magiclink-pmid';
1548 $trackingCat = 'magiclink-tracking-pmid';
1551 throw new MWException( __METHOD__
. ': unrecognised match type "' .
1552 substr( $m[0], 0, 20 ) . '"' );
1554 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1555 $this->addTrackingCategory( $trackingCat );
1556 return Linker
::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle
);
1557 } elseif ( isset( $m[6] ) && $m[6] !== ''
1558 && $this->mOptions
->getMagicISBNLinks()
1562 $space = self
::SPACE_NOT_NL
; # non-newline space
1563 $isbn = preg_replace( "/$space/", ' ', $isbn );
1564 $num = strtr( $isbn, [
1569 $this->addTrackingCategory( 'magiclink-tracking-isbn' );
1570 return $this->getLinkRenderer()->makeKnownLink(
1571 SpecialPage
::getTitleFor( 'Booksources', $num ),
1574 'class' => 'internal mw-magiclink-isbn',
1575 'title' => false // suppress title attribute
1584 * Make a free external link, given a user-supplied URL
1586 * @param string $url
1587 * @param int $numPostProto
1588 * The number of characters after the protocol.
1589 * @return string HTML
1592 public function makeFreeExternalLink( $url, $numPostProto ) {
1595 # The characters '<' and '>' (which were escaped by
1596 # removeHTMLtags()) should not be included in
1597 # URLs, per RFC 2396.
1598 # Make terminate a URL as well (bug T84937)
1601 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1606 $trail = substr( $url, $m2[0][1] ) . $trail;
1607 $url = substr( $url, 0, $m2[0][1] );
1610 # Move trailing punctuation to $trail
1612 # If there is no left bracket, then consider right brackets fair game too
1613 if ( strpos( $url, '(' ) === false ) {
1617 $urlRev = strrev( $url );
1618 $numSepChars = strspn( $urlRev, $sep );
1619 # Don't break a trailing HTML entity by moving the ; into $trail
1620 # This is in hot code, so use substr_compare to avoid having to
1621 # create a new string object for the comparison
1622 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1623 # more optimization: instead of running preg_match with a $
1624 # anchor, which can be slow, do the match on the reversed
1625 # string starting at the desired offset.
1626 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1627 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1631 if ( $numSepChars ) {
1632 $trail = substr( $url, -$numSepChars ) . $trail;
1633 $url = substr( $url, 0, -$numSepChars );
1636 # Verify that we still have a real URL after trail removal, and
1637 # not just lone protocol
1638 if ( strlen( $trail ) >= $numPostProto ) {
1639 return $url . $trail;
1642 $url = Sanitizer
::cleanUrl( $url );
1644 # Is this an external image?
1645 $text = $this->maybeMakeExternalImage( $url );
1646 if ( $text === false ) {
1647 # Not an image, make a link
1648 $text = Linker
::makeExternalLink( $url,
1649 $this->getTargetLanguage()->getConverter()->markNoConversion( $url ),
1651 $this->getExternalLinkAttribs( $url ), $this->mTitle
);
1652 # Register it in the output object...
1653 $this->mOutput
->addExternalLink( $url );
1655 return $text . $trail;
1659 * Parse headers and return html
1663 * @param string $text
1667 public function doHeadings( $text ) {
1668 for ( $i = 6; $i >= 1; --$i ) {
1669 $h = str_repeat( '=', $i );
1670 // Trim non-newline whitespace from headings
1671 // Using \s* will break for: "==\n===\n" and parse as <h2>=</h2>
1672 $text = preg_replace( "/^(?:$h)[ \\t]*(.+?)[ \\t]*(?:$h)\\s*$/m", "<h$i>\\1</h$i>", $text );
1678 * Replace single quotes with HTML markup
1681 * @param string $text
1683 * @return string The altered text
1685 public function doAllQuotes( $text ) {
1687 $lines = StringUtils
::explode( "\n", $text );
1688 foreach ( $lines as $line ) {
1689 $outtext .= $this->doQuotes( $line ) . "\n";
1691 $outtext = substr( $outtext, 0, -1 );
1696 * Helper function for doAllQuotes()
1698 * @param string $text
1702 public function doQuotes( $text ) {
1703 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE
);
1704 $countarr = count( $arr );
1705 if ( $countarr == 1 ) {
1709 // First, do some preliminary work. This may shift some apostrophes from
1710 // being mark-up to being text. It also counts the number of occurrences
1711 // of bold and italics mark-ups.
1714 for ( $i = 1; $i < $countarr; $i +
= 2 ) {
1715 $thislen = strlen( $arr[$i] );
1716 // If there are ever four apostrophes, assume the first is supposed to
1717 // be text, and the remaining three constitute mark-up for bold text.
1718 // (T15227: ''''foo'''' turns into ' ''' foo ' ''')
1719 if ( $thislen == 4 ) {
1720 $arr[$i - 1] .= "'";
1723 } elseif ( $thislen > 5 ) {
1724 // If there are more than 5 apostrophes in a row, assume they're all
1725 // text except for the last 5.
1726 // (T15227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1727 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1731 // Count the number of occurrences of bold and italics mark-ups.
1732 if ( $thislen == 2 ) {
1734 } elseif ( $thislen == 3 ) {
1736 } elseif ( $thislen == 5 ) {
1742 // If there is an odd number of both bold and italics, it is likely
1743 // that one of the bold ones was meant to be an apostrophe followed
1744 // by italics. Which one we cannot know for certain, but it is more
1745 // likely to be one that has a single-letter word before it.
1746 if ( ( $numbold %
2 == 1 ) && ( $numitalics %
2 == 1 ) ) {
1747 $firstsingleletterword = -1;
1748 $firstmultiletterword = -1;
1750 for ( $i = 1; $i < $countarr; $i +
= 2 ) {
1751 if ( strlen( $arr[$i] ) == 3 ) {
1752 $x1 = substr( $arr[$i - 1], -1 );
1753 $x2 = substr( $arr[$i - 1], -2, 1 );
1754 if ( $x1 === ' ' ) {
1755 if ( $firstspace == -1 ) {
1758 } elseif ( $x2 === ' ' ) {
1759 $firstsingleletterword = $i;
1760 // if $firstsingleletterword is set, we don't
1761 // look at the other options, so we can bail early.
1764 if ( $firstmultiletterword == -1 ) {
1765 $firstmultiletterword = $i;
1771 // If there is a single-letter word, use it!
1772 if ( $firstsingleletterword > -1 ) {
1773 $arr[$firstsingleletterword] = "''";
1774 $arr[$firstsingleletterword - 1] .= "'";
1775 } elseif ( $firstmultiletterword > -1 ) {
1776 // If not, but there's a multi-letter word, use that one.
1777 $arr[$firstmultiletterword] = "''";
1778 $arr[$firstmultiletterword - 1] .= "'";
1779 } elseif ( $firstspace > -1 ) {
1780 // ... otherwise use the first one that has neither.
1781 // (notice that it is possible for all three to be -1 if, for example,
1782 // there is only one pentuple-apostrophe in the line)
1783 $arr[$firstspace] = "''";
1784 $arr[$firstspace - 1] .= "'";
1788 // Now let's actually convert our apostrophic mush to HTML!
1793 foreach ( $arr as $r ) {
1794 if ( ( $i %
2 ) == 0 ) {
1795 if ( $state === 'both' ) {
1801 $thislen = strlen( $r );
1802 if ( $thislen == 2 ) {
1803 if ( $state === 'i' ) {
1806 } elseif ( $state === 'bi' ) {
1809 } elseif ( $state === 'ib' ) {
1810 $output .= '</b></i><b>';
1812 } elseif ( $state === 'both' ) {
1813 $output .= '<b><i>' . $buffer . '</i>';
1815 } else { // $state can be 'b' or ''
1819 } elseif ( $thislen == 3 ) {
1820 if ( $state === 'b' ) {
1823 } elseif ( $state === 'bi' ) {
1824 $output .= '</i></b><i>';
1826 } elseif ( $state === 'ib' ) {
1829 } elseif ( $state === 'both' ) {
1830 $output .= '<i><b>' . $buffer . '</b>';
1832 } else { // $state can be 'i' or ''
1836 } elseif ( $thislen == 5 ) {
1837 if ( $state === 'b' ) {
1838 $output .= '</b><i>';
1840 } elseif ( $state === 'i' ) {
1841 $output .= '</i><b>';
1843 } elseif ( $state === 'bi' ) {
1844 $output .= '</i></b>';
1846 } elseif ( $state === 'ib' ) {
1847 $output .= '</b></i>';
1849 } elseif ( $state === 'both' ) {
1850 $output .= '<i><b>' . $buffer . '</b></i>';
1852 } else { // ($state == '')
1860 // Now close all remaining tags. Notice that the order is important.
1861 if ( $state === 'b' ||
$state === 'ib' ) {
1864 if ( $state === 'i' ||
$state === 'bi' ||
$state === 'ib' ) {
1867 if ( $state === 'bi' ) {
1870 // There might be lonely ''''', so make sure we have a buffer
1871 if ( $state === 'both' && $buffer ) {
1872 $output .= '<b><i>' . $buffer . '</i></b>';
1878 * Replace external links (REL)
1880 * Note: this is all very hackish and the order of execution matters a lot.
1881 * Make sure to run tests/parser/parserTests.php if you change this code.
1885 * @param string $text
1887 * @throws MWException
1890 public function replaceExternalLinks( $text ) {
1891 $bits = preg_split( $this->mExtLinkBracketedRegex
, $text, -1, PREG_SPLIT_DELIM_CAPTURE
);
1892 if ( $bits === false ) {
1893 throw new MWException( "PCRE needs to be compiled with "
1894 . "--enable-unicode-properties in order for MediaWiki to function" );
1896 $s = array_shift( $bits );
1899 while ( $i < count( $bits ) ) {
1902 $text = $bits[$i++
];
1903 $trail = $bits[$i++
];
1905 # The characters '<' and '>' (which were escaped by
1906 # removeHTMLtags()) should not be included in
1907 # URLs, per RFC 2396.
1909 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE
) ) {
1910 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1911 $url = substr( $url, 0, $m2[0][1] );
1914 # If the link text is an image URL, replace it with an <img> tag
1915 # This happened by accident in the original parser, but some people used it extensively
1916 $img = $this->maybeMakeExternalImage( $text );
1917 if ( $img !== false ) {
1923 # Set linktype for CSS
1926 # No link text, e.g. [http://domain.tld/some.link]
1927 if ( $text == '' ) {
1929 $langObj = $this->getTargetLanguage();
1930 $text = '[' . $langObj->formatNum( ++
$this->mAutonumber
) . ']';
1931 $linktype = 'autonumber';
1933 # Have link text, e.g. [http://domain.tld/some.link text]s
1935 list( $dtrail, $trail ) = Linker
::splitTrail( $trail );
1938 // Excluding protocol-relative URLs may avoid many false positives.
1939 if ( preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
1940 $text = $this->getTargetLanguage()->getConverter()->markNoConversion( $text );
1943 $url = Sanitizer
::cleanUrl( $url );
1945 # Use the encoded URL
1946 # This means that users can paste URLs directly into the text
1947 # Funny characters like ö aren't valid in URLs anyway
1948 # This was changed in August 2004
1949 $s .= Linker
::makeExternalLink( $url, $text, false, $linktype,
1950 $this->getExternalLinkAttribs( $url ), $this->mTitle
) . $dtrail . $trail;
1952 # Register link in the output object.
1953 $this->mOutput
->addExternalLink( $url );
1960 * Get the rel attribute for a particular external link.
1963 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1964 * nofollow if appropriate
1965 * @param Title|null $title Optional Title, for wgNoFollowNsExceptions lookups
1966 * @return string|null Rel attribute for $url
1968 public static function getExternalLinkRel( $url = false, $title = null ) {
1969 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1970 $ns = $title ?
$title->getNamespace() : false;
1971 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1972 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1980 * Get an associative array of additional HTML attributes appropriate for a
1981 * particular external link. This currently may include rel => nofollow
1982 * (depending on configuration, namespace, and the URL's domain) and/or a
1983 * target attribute (depending on configuration).
1985 * @param string $url URL to extract the domain from for rel =>
1986 * nofollow if appropriate
1987 * @return array Associative array of HTML attributes
1989 public function getExternalLinkAttribs( $url ) {
1991 $rel = self
::getExternalLinkRel( $url, $this->mTitle
);
1993 $target = $this->mOptions
->getExternalLinkTarget();
1995 $attribs['target'] = $target;
1996 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1997 // T133507. New windows can navigate parent cross-origin.
1998 // Including noreferrer due to lacking browser
1999 // support of noopener. Eventually noreferrer should be removed.
2000 if ( $rel !== '' ) {
2003 $rel .= 'noreferrer noopener';
2006 $attribs['rel'] = $rel;
2011 * Replace unusual escape codes in a URL with their equivalent characters
2013 * This generally follows the syntax defined in RFC 3986, with special
2014 * consideration for HTTP query strings.
2016 * @param string $url
2019 public static function normalizeLinkUrl( $url ) {
2020 # First, make sure unsafe characters are encoded
2021 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
2023 return rawurlencode( $m[0] );
2029 $end = strlen( $url );
2031 # Fragment part - 'fragment'
2032 $start = strpos( $url, '#' );
2033 if ( $start !== false && $start < $end ) {
2034 $ret = self
::normalizeUrlComponent(
2035 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
2039 # Query part - 'query' minus &=+;
2040 $start = strpos( $url, '?' );
2041 if ( $start !== false && $start < $end ) {
2042 $ret = self
::normalizeUrlComponent(
2043 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
2047 # Scheme and path part - 'pchar'
2048 # (we assume no userinfo or encoded colons in the host)
2049 $ret = self
::normalizeUrlComponent(
2050 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
2055 private static function normalizeUrlComponent( $component, $unsafe ) {
2056 $callback = function ( $matches ) use ( $unsafe ) {
2057 $char = urldecode( $matches[0] );
2058 $ord = ord( $char );
2059 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
2063 # Leave it escaped, but use uppercase for a-f
2064 return strtoupper( $matches[0] );
2067 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
2071 * make an image if it's allowed, either through the global
2072 * option, through the exception, or through the on-wiki whitelist
2074 * @param string $url
2078 private function maybeMakeExternalImage( $url ) {
2079 $imagesfrom = $this->mOptions
->getAllowExternalImagesFrom();
2080 $imagesexception = !empty( $imagesfrom );
2082 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2083 if ( $imagesexception && is_array( $imagesfrom ) ) {
2084 $imagematch = false;
2085 foreach ( $imagesfrom as $match ) {
2086 if ( strpos( $url, $match ) === 0 ) {
2091 } elseif ( $imagesexception ) {
2092 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2094 $imagematch = false;
2097 if ( $this->mOptions
->getAllowExternalImages()
2098 ||
( $imagesexception && $imagematch )
2100 if ( preg_match( self
::EXT_IMAGE_REGEX
, $url ) ) {
2102 $text = Linker
::makeExternalImage( $url );
2105 if ( !$text && $this->mOptions
->getEnableImageWhitelist()
2106 && preg_match( self
::EXT_IMAGE_REGEX
, $url )
2108 $whitelist = explode(
2110 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2113 foreach ( $whitelist as $entry ) {
2114 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2115 if ( strpos( $entry, '#' ) === 0 ||
$entry === '' ) {
2118 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2119 # Image matches a whitelist entry
2120 $text = Linker
::makeExternalImage( $url );
2129 * Process [[ ]] wikilinks
2133 * @return string Processed text
2137 public function replaceInternalLinks( $s ) {
2138 $this->mLinkHolders
->merge( $this->replaceInternalLinks2( $s ) );
2143 * Process [[ ]] wikilinks (RIL)
2145 * @throws MWException
2146 * @return LinkHolderArray
2150 public function replaceInternalLinks2( &$s ) {
2151 global $wgExtraInterlanguageLinkPrefixes;
2153 static $tc = false, $e1, $e1_img;
2154 # the % is needed to support urlencoded titles as well
2156 $tc = Title
::legalChars() . '#%';
2157 # Match a link having the form [[namespace:link|alternate]]trail
2158 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2159 # Match cases where there is no "]]", which might still be images
2160 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2163 $holders = new LinkHolderArray( $this );
2165 # split the entire text string on occurrences of [[
2166 $a = StringUtils
::explode( '[[', ' ' . $s );
2167 # get the first element (all text up to first [[), and remove the space we added
2170 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2171 $s = substr( $s, 1 );
2173 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2175 if ( $useLinkPrefixExtension ) {
2176 # Match the end of a line for a word that's not followed by whitespace,
2177 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2178 $charset = $this->contLang
->linkPrefixCharset();
2179 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2182 if ( is_null( $this->mTitle
) ) {
2183 throw new MWException( __METHOD__
. ": \$this->mTitle is null\n" );
2185 $nottalk = !$this->mTitle
->isTalkPage();
2187 if ( $useLinkPrefixExtension ) {
2189 if ( preg_match( $e2, $s, $m ) ) {
2190 $first_prefix = $m[2];
2192 $first_prefix = false;
2198 $useSubpages = $this->areSubpagesAllowed();
2200 # Loop for each link
2201 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2202 # Check for excessive memory usage
2203 if ( $holders->isBig() ) {
2205 # Do the existence check, replace the link holders and clear the array
2206 $holders->replace( $s );
2210 if ( $useLinkPrefixExtension ) {
2211 if ( preg_match( $e2, $s, $m ) ) {
2218 if ( $first_prefix ) {
2219 $prefix = $first_prefix;
2220 $first_prefix = false;
2224 $might_be_img = false;
2226 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2228 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2229 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2230 # the real problem is with the $e1 regex
2232 # Still some problems for cases where the ] is meant to be outside punctuation,
2233 # and no image is in sight. See T4095.
2235 && substr( $m[3], 0, 1 ) === ']'
2236 && strpos( $text, '[' ) !== false
2238 $text .= ']'; # so that replaceExternalLinks($text) works later
2239 $m[3] = substr( $m[3], 1 );
2241 # fix up urlencoded title texts
2242 if ( strpos( $m[1], '%' ) !== false ) {
2243 # Should anchors '#' also be rejected?
2244 $m[1] = str_replace( [ '<', '>' ], [ '<', '>' ], rawurldecode( $m[1] ) );
2247 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2248 # Invalid, but might be an image with a link in its caption
2249 $might_be_img = true;
2251 if ( strpos( $m[1], '%' ) !== false ) {
2252 $m[1] = str_replace( [ '<', '>' ], [ '<', '>' ], rawurldecode( $m[1] ) );
2255 } else { # Invalid form; output directly
2256 $s .= $prefix . '[[' . $line;
2260 $origLink = ltrim( $m[1], ' ' );
2262 # Don't allow internal links to pages containing
2263 # PROTO: where PROTO is a valid URL protocol; these
2264 # should be external links.
2265 if ( preg_match( '/^(?i:' . $this->mUrlProtocols
. ')/', $origLink ) ) {
2266 $s .= $prefix . '[[' . $line;
2270 # Make subpage if necessary
2271 if ( $useSubpages ) {
2272 $link = $this->maybeDoSubpageLink( $origLink, $text );
2277 // \x7f isn't a default legal title char, so most likely strip
2278 // markers will force us into the "invalid form" path above. But,
2279 // just in case, let's assert that xmlish tags aren't valid in
2280 // the title position.
2281 $unstrip = $this->mStripState
->killMarkers( $link );
2282 $noMarkers = ( $unstrip === $link );
2284 $nt = $noMarkers ? Title
::newFromText( $link ) : null;
2285 if ( $nt === null ) {
2286 $s .= $prefix . '[[' . $line;
2290 $ns = $nt->getNamespace();
2291 $iw = $nt->getInterwiki();
2293 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2295 if ( $might_be_img ) { # if this is actually an invalid link
2296 if ( $ns == NS_FILE
&& $noforce ) { # but might be an image
2299 # look at the next 'line' to see if we can close it there
2301 $next_line = $a->current();
2302 if ( $next_line === false ||
$next_line === null ) {
2305 $m = explode( ']]', $next_line, 3 );
2306 if ( count( $m ) == 3 ) {
2307 # the first ]] closes the inner link, the second the image
2309 $text .= "[[{$m[0]}]]{$m[1]}";
2312 } elseif ( count( $m ) == 2 ) {
2313 # if there's exactly one ]] that's fine, we'll keep looking
2314 $text .= "[[{$m[0]}]]{$m[1]}";
2316 # if $next_line is invalid too, we need look no further
2317 $text .= '[[' . $next_line;
2322 # we couldn't find the end of this imageLink, so output it raw
2323 # but don't ignore what might be perfectly normal links in the text we've examined
2324 $holders->merge( $this->replaceInternalLinks2( $text ) );
2325 $s .= "{$prefix}[[$link|$text";
2326 # note: no $trail, because without an end, there *is* no trail
2329 } else { # it's not an image, so output it raw
2330 $s .= "{$prefix}[[$link|$text";
2331 # note: no $trail, because without an end, there *is* no trail
2336 $wasblank = ( $text == '' );
2340 # Strip off leading ':'
2341 $text = substr( $text, 1 );
2344 # T6598 madness. Handle the quotes only if they come from the alternate part
2345 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2346 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2347 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2348 $text = $this->doQuotes( $text );
2351 # Link not escaped by : , create the various objects
2352 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2355 $iw && $this->mOptions
->getInterwikiMagic() && $nottalk && (
2356 Language
::fetchLanguageName( $iw, null, 'mw' ) ||
2357 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2360 # T26502: filter duplicates
2361 if ( !isset( $this->mLangLinkLanguages
[$iw] ) ) {
2362 $this->mLangLinkLanguages
[$iw] = true;
2363 $this->mOutput
->addLanguageLink( $nt->getFullText() );
2367 * Strip the whitespace interwiki links produce, see T10897
2369 $s = rtrim( $s . $prefix ) . $trail; # T175416
2373 if ( $ns == NS_FILE
) {
2374 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle
) ) {
2376 # if no parameters were passed, $text
2377 # becomes something like "File:Foo.png",
2378 # which we don't want to pass on to the
2382 # recursively parse links inside the image caption
2383 # actually, this will parse them in any other parameters, too,
2384 # but it might be hard to fix that, and it doesn't matter ATM
2385 $text = $this->replaceExternalLinks( $text );
2386 $holders->merge( $this->replaceInternalLinks2( $text ) );
2388 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2389 $s .= $prefix . $this->armorLinks(
2390 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2393 } elseif ( $ns == NS_CATEGORY
) {
2395 * Strip the whitespace Category links produce, see T2087
2397 $s = rtrim( $s . $prefix ) . $trail; # T2087, T87753
2400 $sortkey = $this->getDefaultSort();
2404 $sortkey = Sanitizer
::decodeCharReferences( $sortkey );
2405 $sortkey = str_replace( "\n", '', $sortkey );
2406 $sortkey = $this->getTargetLanguage()->convertCategoryKey( $sortkey );
2407 $this->mOutput
->addCategory( $nt->getDBkey(), $sortkey );
2413 # Self-link checking. For some languages, variants of the title are checked in
2414 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2415 # for linking to a different variant.
2416 if ( $ns != NS_SPECIAL
&& $nt->equals( $this->mTitle
) && !$nt->hasFragment() ) {
2417 $s .= $prefix . Linker
::makeSelfLinkObj( $nt, $text, '', $trail );
2421 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2422 # @todo FIXME: Should do batch file existence checks, see comment below
2423 if ( $ns == NS_MEDIA
) {
2424 # Give extensions a chance to select the file revision for us
2427 Hooks
::run( 'BeforeParserFetchFileAndTitle',
2428 [ $this, $nt, &$options, &$descQuery ] );
2429 # Fetch and register the file (file title may be different via hooks)
2430 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2431 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2432 $s .= $prefix . $this->armorLinks(
2433 Linker
::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2437 # Some titles, such as valid special pages or files in foreign repos, should
2438 # be shown as bluelinks even though they're not included in the page table
2439 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2440 # batch file existence checks for NS_FILE and NS_MEDIA
2441 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2442 $this->mOutput
->addLink( $nt );
2443 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2445 # Links will be added to the output link list after checking
2446 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2453 * Render a forced-blue link inline; protect against double expansion of
2454 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2455 * Since this little disaster has to split off the trail text to avoid
2456 * breaking URLs in the following text without breaking trails on the
2457 * wiki links, it's been made into a horrible function.
2460 * @param string $text
2461 * @param string $trail
2462 * @param string $prefix
2463 * @return string HTML-wikitext mix oh yuck
2465 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2466 list( $inside, $trail ) = Linker
::splitTrail( $trail );
2468 if ( $text == '' ) {
2469 $text = htmlspecialchars( $nt->getPrefixedText() );
2472 $link = $this->getLinkRenderer()->makeKnownLink(
2473 $nt, new HtmlArmor( "$prefix$text$inside" )
2476 return $this->armorLinks( $link ) . $trail;
2480 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2481 * going to go through further parsing steps before inline URL expansion.
2483 * Not needed quite as much as it used to be since free links are a bit
2484 * more sensible these days. But bracketed links are still an issue.
2486 * @param string $text More-or-less HTML
2487 * @return string Less-or-more HTML with NOPARSE bits
2489 public function armorLinks( $text ) {
2490 return preg_replace( '/\b((?i)' . $this->mUrlProtocols
. ')/',
2491 self
::MARKER_PREFIX
. "NOPARSE$1", $text );
2495 * Return true if subpage links should be expanded on this page.
2498 public function areSubpagesAllowed() {
2499 # Some namespaces don't allow subpages
2500 return MWNamespace
::hasSubpages( $this->mTitle
->getNamespace() );
2504 * Handle link to subpage if necessary
2506 * @param string $target The source of the link
2507 * @param string &$text The link text, modified as necessary
2508 * @return string The full name of the link
2511 public function maybeDoSubpageLink( $target, &$text ) {
2512 return Linker
::normalizeSubpageLink( $this->mTitle
, $target, $text );
2516 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2518 * @param string $text
2519 * @param bool $linestart Whether or not this is at the start of a line.
2521 * @return string The lists rendered as HTML
2523 public function doBlockLevels( $text, $linestart ) {
2524 return BlockLevelPass
::doBlockLevels( $text, $linestart );
2528 * Return value of a magic variable (like PAGENAME)
2532 * @param string $index Magic variable identifier as mapped in MagicWordFactory::$mVariableIDs
2533 * @param bool|PPFrame $frame
2535 * @throws MWException
2538 public function getVariableValue( $index, $frame = false ) {
2539 global $wgSitename, $wgServer, $wgServerName;
2540 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2542 if ( is_null( $this->mTitle
) ) {
2543 // If no title set, bad things are going to happen
2544 // later. Title should always be set since this
2545 // should only be called in the middle of a parse
2546 // operation (but the unit-tests do funky stuff)
2547 throw new MWException( __METHOD__
. ' Should only be '
2548 . ' called while parsing (no title set)' );
2551 // Avoid PHP 7.1 warning from passing $this by reference
2555 * Some of these require message or data lookups and can be
2556 * expensive to check many times.
2558 if ( Hooks
::run( 'ParserGetVariableValueVarCache', [ &$parser, &$this->mVarCache
] ) ) {
2559 if ( isset( $this->mVarCache
[$index] ) ) {
2560 return $this->mVarCache
[$index];
2564 $ts = wfTimestamp( TS_UNIX
, $this->mOptions
->getTimestamp() );
2565 Hooks
::run( 'ParserGetVariableValueTs', [ &$parser, &$ts ] );
2567 $pageLang = $this->getFunctionLang();
2573 case 'currentmonth':
2574 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'm' ), true );
2576 case 'currentmonth1':
2577 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'n' ), true );
2579 case 'currentmonthname':
2580 $value = $pageLang->getMonthName( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2582 case 'currentmonthnamegen':
2583 $value = $pageLang->getMonthNameGen( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2585 case 'currentmonthabbrev':
2586 $value = $pageLang->getMonthAbbreviation( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2589 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'j' ), true );
2592 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'd' ), true );
2595 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'm' ), true );
2598 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ), true );
2600 case 'localmonthname':
2601 $value = $pageLang->getMonthName( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2603 case 'localmonthnamegen':
2604 $value = $pageLang->getMonthNameGen( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2606 case 'localmonthabbrev':
2607 $value = $pageLang->getMonthAbbreviation( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2610 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'j' ), true );
2613 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'd' ), true );
2616 $value = wfEscapeWikiText( $this->mTitle
->getText() );
2619 $value = wfEscapeWikiText( $this->mTitle
->getPartialURL() );
2621 case 'fullpagename':
2622 $value = wfEscapeWikiText( $this->mTitle
->getPrefixedText() );
2624 case 'fullpagenamee':
2625 $value = wfEscapeWikiText( $this->mTitle
->getPrefixedURL() );
2628 $value = wfEscapeWikiText( $this->mTitle
->getSubpageText() );
2630 case 'subpagenamee':
2631 $value = wfEscapeWikiText( $this->mTitle
->getSubpageUrlForm() );
2633 case 'rootpagename':
2634 $value = wfEscapeWikiText( $this->mTitle
->getRootText() );
2636 case 'rootpagenamee':
2637 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2640 $this->mTitle
->getRootText()
2643 case 'basepagename':
2644 $value = wfEscapeWikiText( $this->mTitle
->getBaseText() );
2646 case 'basepagenamee':
2647 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2650 $this->mTitle
->getBaseText()
2653 case 'talkpagename':
2654 if ( $this->mTitle
->canHaveTalkPage() ) {
2655 $talkPage = $this->mTitle
->getTalkPage();
2656 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2661 case 'talkpagenamee':
2662 if ( $this->mTitle
->canHaveTalkPage() ) {
2663 $talkPage = $this->mTitle
->getTalkPage();
2664 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2669 case 'subjectpagename':
2670 $subjPage = $this->mTitle
->getSubjectPage();
2671 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2673 case 'subjectpagenamee':
2674 $subjPage = $this->mTitle
->getSubjectPage();
2675 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2677 case 'pageid': // requested in T25427
2678 $pageid = $this->getTitle()->getArticleID();
2679 if ( $pageid == 0 ) {
2680 # 0 means the page doesn't exist in the database,
2681 # which means the user is previewing a new page.
2682 # The vary-revision flag must be set, because the magic word
2683 # will have a different value once the page is saved.
2684 $this->mOutput
->setFlag( 'vary-revision' );
2685 wfDebug( __METHOD__
. ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2687 $value = $pageid ?
: null;
2690 # Let the edit saving system know we should parse the page
2691 # *after* a revision ID has been assigned.
2692 $this->mOutput
->setFlag( 'vary-revision-id' );
2693 wfDebug( __METHOD__
. ": {{REVISIONID}} used, setting vary-revision-id...\n" );
2694 $value = $this->mRevisionId
;
2695 if ( !$value && $this->mOptions
->getSpeculativeRevIdCallback() ) {
2696 $value = call_user_func( $this->mOptions
->getSpeculativeRevIdCallback() );
2697 $this->mOutput
->setSpeculativeRevIdUsed( $value );
2701 $value = (int)$this->getRevisionTimestampSubstring( 6, 2, self
::MAX_TTS
, $index );
2703 case 'revisionday2':
2704 $value = $this->getRevisionTimestampSubstring( 6, 2, self
::MAX_TTS
, $index );
2706 case 'revisionmonth':
2707 $value = $this->getRevisionTimestampSubstring( 4, 2, self
::MAX_TTS
, $index );
2709 case 'revisionmonth1':
2710 $value = (int)$this->getRevisionTimestampSubstring( 4, 2, self
::MAX_TTS
, $index );
2712 case 'revisionyear':
2713 $value = $this->getRevisionTimestampSubstring( 0, 4, self
::MAX_TTS
, $index );
2715 case 'revisiontimestamp':
2716 # Let the edit saving system know we should parse the page
2717 # *after* a revision ID has been assigned. This is for null edits.
2718 $this->mOutput
->setFlag( 'vary-revision' );
2719 wfDebug( __METHOD__
. ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2720 $value = $this->getRevisionTimestamp();
2722 case 'revisionuser':
2723 # Let the edit saving system know we should parse the page
2724 # *after* a revision ID has been assigned for null edits.
2725 $this->mOutput
->setFlag( 'vary-user' );
2726 wfDebug( __METHOD__
. ": {{REVISIONUSER}} used, setting vary-user...\n" );
2727 $value = $this->getRevisionUser();
2729 case 'revisionsize':
2730 $value = $this->getRevisionSize();
2733 $value = str_replace( '_', ' ',
2734 $this->contLang
->getNsText( $this->mTitle
->getNamespace() ) );
2737 $value = wfUrlencode( $this->contLang
->getNsText( $this->mTitle
->getNamespace() ) );
2739 case 'namespacenumber':
2740 $value = $this->mTitle
->getNamespace();
2743 $value = $this->mTitle
->canHaveTalkPage()
2744 ?
str_replace( '_', ' ', $this->mTitle
->getTalkNsText() )
2748 $value = $this->mTitle
->canHaveTalkPage() ?
wfUrlencode( $this->mTitle
->getTalkNsText() ) : '';
2750 case 'subjectspace':
2751 $value = str_replace( '_', ' ', $this->mTitle
->getSubjectNsText() );
2753 case 'subjectspacee':
2754 $value = ( wfUrlencode( $this->mTitle
->getSubjectNsText() ) );
2756 case 'currentdayname':
2757 $value = $pageLang->getWeekdayName( (int)MWTimestamp
::getInstance( $ts )->format( 'w' ) +
1 );
2760 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'Y' ), true );
2763 $value = $pageLang->time( wfTimestamp( TS_MW
, $ts ), false, false );
2766 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'H' ), true );
2769 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2770 # int to remove the padding
2771 $value = $pageLang->formatNum( (int)MWTimestamp
::getInstance( $ts )->format( 'W' ) );
2774 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'w' ) );
2776 case 'localdayname':
2777 $value = $pageLang->getWeekdayName(
2778 (int)MWTimestamp
::getLocalInstance( $ts )->format( 'w' ) +
1
2782 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'Y' ), true );
2785 $value = $pageLang->time(
2786 MWTimestamp
::getLocalInstance( $ts )->format( 'YmdHis' ),
2792 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'H' ), true );
2795 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2796 # int to remove the padding
2797 $value = $pageLang->formatNum( (int)MWTimestamp
::getLocalInstance( $ts )->format( 'W' ) );
2800 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'w' ) );
2802 case 'numberofarticles':
2803 $value = $pageLang->formatNum( SiteStats
::articles() );
2805 case 'numberoffiles':
2806 $value = $pageLang->formatNum( SiteStats
::images() );
2808 case 'numberofusers':
2809 $value = $pageLang->formatNum( SiteStats
::users() );
2811 case 'numberofactiveusers':
2812 $value = $pageLang->formatNum( SiteStats
::activeUsers() );
2814 case 'numberofpages':
2815 $value = $pageLang->formatNum( SiteStats
::pages() );
2817 case 'numberofadmins':
2818 $value = $pageLang->formatNum( SiteStats
::numberingroup( 'sysop' ) );
2820 case 'numberofedits':
2821 $value = $pageLang->formatNum( SiteStats
::edits() );
2823 case 'currenttimestamp':
2824 $value = wfTimestamp( TS_MW
, $ts );
2826 case 'localtimestamp':
2827 $value = MWTimestamp
::getLocalInstance( $ts )->format( 'YmdHis' );
2829 case 'currentversion':
2830 $value = SpecialVersion
::getVersion();
2833 return $wgArticlePath;
2839 return $wgServerName;
2841 return $wgScriptPath;
2843 return $wgStylePath;
2844 case 'directionmark':
2845 return $pageLang->getDirMark();
2846 case 'contentlanguage':
2847 global $wgLanguageCode;
2848 return $wgLanguageCode;
2849 case 'pagelanguage':
2850 $value = $pageLang->getCode();
2852 case 'cascadingsources':
2853 $value = CoreParserFunctions
::cascadingsources( $this );
2858 'ParserGetVariableValueSwitch',
2859 [ &$parser, &$this->mVarCache
, &$index, &$ret, &$frame ]
2866 $this->mVarCache
[$index] = $value;
2875 * @param int $mtts Max time-till-save; sets vary-revision if result might change by then
2876 * @param string $variable Parser variable name
2879 private function getRevisionTimestampSubstring( $start, $len, $mtts, $variable ) {
2880 # Get the timezone-adjusted timestamp to be used for this revision
2881 $resNow = substr( $this->getRevisionTimestamp(), $start, $len );
2882 # Possibly set vary-revision if there is not yet an associated revision
2883 if ( !$this->getRevisionObject() ) {
2884 # Get the timezone-adjusted timestamp $mtts seconds in the future
2886 $this->contLang
->userAdjust( wfTimestamp( TS_MW
, time() +
$mtts ), '' ),
2891 if ( $resNow !== $resThen ) {
2892 # Let the edit saving system know we should parse the page
2893 # *after* a revision ID has been assigned. This is for null edits.
2894 $this->mOutput
->setFlag( 'vary-revision' );
2895 wfDebug( __METHOD__
. ": $variable used, setting vary-revision...\n" );
2903 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2907 public function initialiseVariables() {
2908 $variableIDs = $this->magicWordFactory
->getVariableIDs();
2909 $substIDs = $this->magicWordFactory
->getSubstIDs();
2911 $this->mVariables
= $this->magicWordFactory
->newArray( $variableIDs );
2912 $this->mSubstWords
= $this->magicWordFactory
->newArray( $substIDs );
2916 * Preprocess some wikitext and return the document tree.
2917 * This is the ghost of replace_variables().
2919 * @param string $text The text to parse
2920 * @param int $flags Bitwise combination of:
2921 * - self::PTD_FOR_INCLUSION: Handle "<noinclude>" and "<includeonly>" as if the text is being
2922 * included. Default is to assume a direct page view.
2924 * The generated DOM tree must depend only on the input text and the flags.
2925 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of T6899.
2927 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2928 * change in the DOM tree for a given text, must be passed through the section identifier
2929 * in the section edit link and thus back to extractSections().
2931 * The output of this function is currently only cached in process memory, but a persistent
2932 * cache may be implemented at a later date which takes further advantage of these strict
2933 * dependency requirements.
2937 public function preprocessToDom( $text, $flags = 0 ) {
2938 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
2943 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2949 public static function splitWhitespace( $s ) {
2950 $ltrimmed = ltrim( $s );
2951 $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) );
2952 $trimmed = rtrim( $ltrimmed );
2953 $diff = strlen( $ltrimmed ) - strlen( $trimmed );
2955 $w2 = substr( $ltrimmed, -$diff );
2959 return [ $w1, $trimmed, $w2 ];
2963 * Replace magic variables, templates, and template arguments
2964 * with the appropriate text. Templates are substituted recursively,
2965 * taking care to avoid infinite loops.
2967 * Note that the substitution depends on value of $mOutputType:
2968 * self::OT_WIKI: only {{subst:}} templates
2969 * self::OT_PREPROCESS: templates but not extension tags
2970 * self::OT_HTML: all templates and extension tags
2972 * @param string $text The text to transform
2973 * @param bool|PPFrame $frame Object describing the arguments passed to the
2974 * template. Arguments may also be provided as an associative array, as
2975 * was the usual case before MW1.12. Providing arguments this way may be
2976 * useful for extensions wishing to perform variable replacement
2978 * @param bool $argsOnly Only do argument (triple-brace) expansion, not
2979 * double-brace expansion.
2982 public function replaceVariables( $text, $frame = false, $argsOnly = false ) {
2983 # Is there any text? Also, Prevent too big inclusions!
2984 $textSize = strlen( $text );
2985 if ( $textSize < 1 ||
$textSize > $this->mOptions
->getMaxIncludeSize() ) {
2989 if ( $frame === false ) {
2990 $frame = $this->getPreprocessor()->newFrame();
2991 } elseif ( !( $frame instanceof PPFrame
) ) {
2992 wfDebug( __METHOD__
. " called using plain parameters instead of "
2993 . "a PPFrame instance. Creating custom frame.\n" );
2994 $frame = $this->getPreprocessor()->newCustomFrame( $frame );
2997 $dom = $this->preprocessToDom( $text );
2998 $flags = $argsOnly ? PPFrame
::NO_TEMPLATES
: 0;
2999 $text = $frame->expand( $dom, $flags );
3005 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
3007 * @param array $args
3011 public static function createAssocArgs( $args ) {
3014 foreach ( $args as $arg ) {
3015 $eqpos = strpos( $arg, '=' );
3016 if ( $eqpos === false ) {
3017 $assocArgs[$index++
] = $arg;
3019 $name = trim( substr( $arg, 0, $eqpos ) );
3020 $value = trim( substr( $arg, $eqpos +
1 ) );
3021 if ( $value === false ) {
3024 if ( $name !== false ) {
3025 $assocArgs[$name] = $value;
3034 * Warn the user when a parser limitation is reached
3035 * Will warn at most once the user per limitation type
3037 * The results are shown during preview and run through the Parser (See EditPage.php)
3039 * @param string $limitationType Should be one of:
3040 * 'expensive-parserfunction' (corresponding messages:
3041 * 'expensive-parserfunction-warning',
3042 * 'expensive-parserfunction-category')
3043 * 'post-expand-template-argument' (corresponding messages:
3044 * 'post-expand-template-argument-warning',
3045 * 'post-expand-template-argument-category')
3046 * 'post-expand-template-inclusion' (corresponding messages:
3047 * 'post-expand-template-inclusion-warning',
3048 * 'post-expand-template-inclusion-category')
3049 * 'node-count-exceeded' (corresponding messages:
3050 * 'node-count-exceeded-warning',
3051 * 'node-count-exceeded-category')
3052 * 'expansion-depth-exceeded' (corresponding messages:
3053 * 'expansion-depth-exceeded-warning',
3054 * 'expansion-depth-exceeded-category')
3055 * @param string|int|null $current Current value
3056 * @param string|int|null $max Maximum allowed, when an explicit limit has been
3057 * exceeded, provide the values (optional)
3059 public function limitationWarn( $limitationType, $current = '', $max = '' ) {
3060 # does no harm if $current and $max are present but are unnecessary for the message
3061 # Not doing ->inLanguage( $this->mOptions->getUserLangObj() ), since this is shown
3062 # only during preview, and that would split the parser cache unnecessarily.
3063 $warning = wfMessage( "$limitationType-warning" )->numParams( $current, $max )
3065 $this->mOutput
->addWarning( $warning );
3066 $this->addTrackingCategory( "$limitationType-category" );
3070 * Return the text of a template, after recursively
3071 * replacing any variables or templates within the template.
3073 * @param array $piece The parts of the template
3074 * $piece['title']: the title, i.e. the part before the |
3075 * $piece['parts']: the parameter array
3076 * $piece['lineStart']: whether the brace was at the start of a line
3077 * @param PPFrame $frame The current frame, contains template arguments
3079 * @return string The text of the template
3081 public function braceSubstitution( $piece, $frame ) {
3084 // $text has been filled
3086 // wiki markup in $text should be escaped
3088 // $text is HTML, armour it against wikitext transformation
3090 // Force interwiki transclusion to be done in raw mode not rendered
3091 $forceRawInterwiki = false;
3092 // $text is a DOM node needing expansion in a child frame
3093 $isChildObj = false;
3094 // $text is a DOM node needing expansion in the current frame
3095 $isLocalObj = false;
3097 # Title object, where $text came from
3100 # $part1 is the bit before the first |, and must contain only title characters.
3101 # Various prefixes will be stripped from it later.
3102 $titleWithSpaces = $frame->expand( $piece['title'] );
3103 $part1 = trim( $titleWithSpaces );
3106 # Original title text preserved for various purposes
3107 $originalTitle = $part1;
3109 # $args is a list of argument nodes, starting from index 0, not including $part1
3110 # @todo FIXME: If piece['parts'] is null then the call to getLength()
3111 # below won't work b/c this $args isn't an object
3112 $args = ( null == $piece['parts'] ) ?
[] : $piece['parts'];
3114 $profileSection = null; // profile templates
3118 $substMatch = $this->mSubstWords
->matchStartAndRemove( $part1 );
3120 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3121 # Decide whether to expand template or keep wikitext as-is.
3122 if ( $this->ot
['wiki'] ) {
3123 if ( $substMatch === false ) {
3124 $literal = true; # literal when in PST with no prefix
3126 $literal = false; # expand when in PST with subst: or safesubst:
3129 if ( $substMatch == 'subst' ) {
3130 $literal = true; # literal when not in PST with plain subst:
3132 $literal = false; # expand when not in PST with safesubst: or no prefix
3136 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3143 if ( !$found && $args->getLength() == 0 ) {
3144 $id = $this->mVariables
->matchStartToEnd( $part1 );
3145 if ( $id !== false ) {
3146 $text = $this->getVariableValue( $id, $frame );
3147 if ( $this->magicWordFactory
->getCacheTTL( $id ) > -1 ) {
3148 $this->mOutput
->updateCacheExpiry(
3149 $this->magicWordFactory
->getCacheTTL( $id ) );
3155 # MSG, MSGNW and RAW
3158 $mwMsgnw = $this->magicWordFactory
->get( 'msgnw' );
3159 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
3162 # Remove obsolete MSG:
3163 $mwMsg = $this->magicWordFactory
->get( 'msg' );
3164 $mwMsg->matchStartAndRemove( $part1 );
3168 $mwRaw = $this->magicWordFactory
->get( 'raw' );
3169 if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
3170 $forceRawInterwiki = true;
3176 $colonPos = strpos( $part1, ':' );
3177 if ( $colonPos !== false ) {
3178 $func = substr( $part1, 0, $colonPos );
3179 $funcArgs = [ trim( substr( $part1, $colonPos +
1 ) ) ];
3180 $argsLength = $args->getLength();
3181 for ( $i = 0; $i < $argsLength; $i++
) {
3182 $funcArgs[] = $args->item( $i );
3185 $result = $this->callParserFunction( $frame, $func, $funcArgs );
3187 // Extract any forwarded flags
3188 if ( isset( $result['title'] ) ) {
3189 $title = $result['title'];
3191 if ( isset( $result['found'] ) ) {
3192 $found = $result['found'];
3194 if ( array_key_exists( 'text', $result ) ) {
3196 $text = $result['text'];
3198 if ( isset( $result['nowiki'] ) ) {
3199 $nowiki = $result['nowiki'];
3201 if ( isset( $result['isHTML'] ) ) {
3202 $isHTML = $result['isHTML'];
3204 if ( isset( $result['forceRawInterwiki'] ) ) {
3205 $forceRawInterwiki = $result['forceRawInterwiki'];
3207 if ( isset( $result['isChildObj'] ) ) {
3208 $isChildObj = $result['isChildObj'];
3210 if ( isset( $result['isLocalObj'] ) ) {
3211 $isLocalObj = $result['isLocalObj'];
3216 # Finish mangling title and then check for loops.
3217 # Set $title to a Title object and $titleText to the PDBK
3220 # Split the title into page and subpage
3222 $relative = $this->maybeDoSubpageLink( $part1, $subpage );
3223 if ( $part1 !== $relative ) {
3225 $ns = $this->mTitle
->getNamespace();
3227 $title = Title
::newFromText( $part1, $ns );
3229 $titleText = $title->getPrefixedText();
3230 # Check for language variants if the template is not found
3231 if ( $this->getTargetLanguage()->hasVariants() && $title->getArticleID() == 0 ) {
3232 $this->getTargetLanguage()->findVariantLink( $part1, $title, true );
3234 # Do recursion depth check
3235 $limit = $this->mOptions
->getMaxTemplateDepth();
3236 if ( $frame->depth
>= $limit ) {
3238 $text = '<span class="error">'
3239 . wfMessage( 'parser-template-recursion-depth-warning' )
3240 ->numParams( $limit )->inContentLanguage()->text()
3246 # Load from database
3247 if ( !$found && $title ) {
3248 $profileSection = $this->mProfiler
->scopedProfileIn( $title->getPrefixedDBkey() );
3249 if ( !$title->isExternal() ) {
3250 if ( $title->isSpecialPage()
3251 && $this->mOptions
->getAllowSpecialInclusion()
3252 && $this->ot
['html']
3254 $specialPage = $this->specialPageFactory
->getPage( $title->getDBkey() );
3255 // Pass the template arguments as URL parameters.
3256 // "uselang" will have no effect since the Language object
3257 // is forced to the one defined in ParserOptions.
3259 $argsLength = $args->getLength();
3260 for ( $i = 0; $i < $argsLength; $i++
) {
3261 $bits = $args->item( $i )->splitArg();
3262 if ( strval( $bits['index'] ) === '' ) {
3263 $name = trim( $frame->expand( $bits['name'], PPFrame
::STRIP_COMMENTS
) );
3264 $value = trim( $frame->expand( $bits['value'] ) );
3265 $pageArgs[$name] = $value;
3269 // Create a new context to execute the special page
3270 $context = new RequestContext
;
3271 $context->setTitle( $title );
3272 $context->setRequest( new FauxRequest( $pageArgs ) );
3273 if ( $specialPage && $specialPage->maxIncludeCacheTime() === 0 ) {
3274 $context->setUser( $this->getUser() );
3276 // If this page is cached, then we better not be per user.
3277 $context->setUser( User
::newFromName( '127.0.0.1', false ) );
3279 $context->setLanguage( $this->mOptions
->getUserLangObj() );
3280 $ret = $this->specialPageFactory
->capturePath( $title, $context, $this->getLinkRenderer() );
3282 $text = $context->getOutput()->getHTML();
3283 $this->mOutput
->addOutputPageMetadata( $context->getOutput() );
3286 if ( $specialPage && $specialPage->maxIncludeCacheTime() !== false ) {
3287 $this->mOutput
->updateRuntimeAdaptiveExpiry(
3288 $specialPage->maxIncludeCacheTime()
3292 } elseif ( MWNamespace
::isNonincludable( $title->getNamespace() ) ) {
3293 $found = false; # access denied
3294 wfDebug( __METHOD__
. ": template inclusion denied for " .
3295 $title->getPrefixedDBkey() . "\n" );
3297 list( $text, $title ) = $this->getTemplateDom( $title );
3298 if ( $text !== false ) {
3304 # If the title is valid but undisplayable, make a link to it
3305 if ( !$found && ( $this->ot
['html'] ||
$this->ot
['pre'] ) ) {
3306 $text = "[[:$titleText]]";
3309 } elseif ( $title->isTrans() ) {
3310 # Interwiki transclusion
3311 if ( $this->ot
['html'] && !$forceRawInterwiki ) {
3312 $text = $this->interwikiTransclude( $title, 'render' );
3315 $text = $this->interwikiTransclude( $title, 'raw' );
3316 # Preprocess it like a template
3317 $text = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
3323 # Do infinite loop check
3324 # This has to be done after redirect resolution to avoid infinite loops via redirects
3325 if ( !$frame->loopCheck( $title ) ) {
3327 $text = '<span class="error">'
3328 . wfMessage( 'parser-template-loop-warning', $titleText )->inContentLanguage()->text()
3330 $this->addTrackingCategory( 'template-loop-category' );
3331 $this->mOutput
->addWarning( wfMessage( 'template-loop-warning',
3332 wfEscapeWikiText( $titleText ) )->text() );
3333 wfDebug( __METHOD__
. ": template loop broken at '$titleText'\n" );
3337 # If we haven't found text to substitute by now, we're done
3338 # Recover the source wikitext and return it
3340 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3341 if ( $profileSection ) {
3342 $this->mProfiler
->scopedProfileOut( $profileSection );
3344 return [ 'object' => $text ];
3347 # Expand DOM-style return values in a child frame
3348 if ( $isChildObj ) {
3349 # Clean up argument array
3350 $newFrame = $frame->newChild( $args, $title );
3353 $text = $newFrame->expand( $text, PPFrame
::RECOVER_ORIG
);
3354 } elseif ( $titleText !== false && $newFrame->isEmpty() ) {
3355 # Expansion is eligible for the empty-frame cache
3356 $text = $newFrame->cachedExpand( $titleText, $text );
3358 # Uncached expansion
3359 $text = $newFrame->expand( $text );
3362 if ( $isLocalObj && $nowiki ) {
3363 $text = $frame->expand( $text, PPFrame
::RECOVER_ORIG
);
3364 $isLocalObj = false;
3367 if ( $profileSection ) {
3368 $this->mProfiler
->scopedProfileOut( $profileSection );
3371 # Replace raw HTML by a placeholder
3373 $text = $this->insertStripItem( $text );
3374 } elseif ( $nowiki && ( $this->ot
['html'] ||
$this->ot
['pre'] ) ) {
3375 # Escape nowiki-style return values
3376 $text = wfEscapeWikiText( $text );
3377 } elseif ( is_string( $text )
3378 && !$piece['lineStart']
3379 && preg_match( '/^(?:{\\||:|;|#|\*)/', $text )
3381 # T2529: if the template begins with a table or block-level
3382 # element, it should be treated as beginning a new line.
3383 # This behavior is somewhat controversial.
3384 $text = "\n" . $text;
3387 if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
3388 # Error, oversize inclusion
3389 if ( $titleText !== false ) {
3390 # Make a working, properly escaped link if possible (T25588)
3391 $text = "[[:$titleText]]";
3393 # This will probably not be a working link, but at least it may
3394 # provide some hint of where the problem is
3395 preg_replace( '/^:/', '', $originalTitle );
3396 $text = "[[:$originalTitle]]";
3398 $text .= $this->insertStripItem( '<!-- WARNING: template omitted, '
3399 . 'post-expand include size too large -->' );
3400 $this->limitationWarn( 'post-expand-template-inclusion' );
3403 if ( $isLocalObj ) {
3404 $ret = [ 'object' => $text ];
3406 $ret = [ 'text' => $text ];
3413 * Call a parser function and return an array with text and flags.
3415 * The returned array will always contain a boolean 'found', indicating
3416 * whether the parser function was found or not. It may also contain the
3418 * text: string|object, resulting wikitext or PP DOM object
3419 * isHTML: bool, $text is HTML, armour it against wikitext transformation
3420 * isChildObj: bool, $text is a DOM node needing expansion in a child frame
3421 * isLocalObj: bool, $text is a DOM node needing expansion in the current frame
3422 * nowiki: bool, wiki markup in $text should be escaped
3425 * @param PPFrame $frame The current frame, contains template arguments
3426 * @param string $function Function name
3427 * @param array $args Arguments to the function
3428 * @throws MWException
3431 public function callParserFunction( $frame, $function, array $args = [] ) {
3432 # Case sensitive functions
3433 if ( isset( $this->mFunctionSynonyms
[1][$function] ) ) {
3434 $function = $this->mFunctionSynonyms
[1][$function];
3436 # Case insensitive functions
3437 $function = $this->contLang
->lc( $function );
3438 if ( isset( $this->mFunctionSynonyms
[0][$function] ) ) {
3439 $function = $this->mFunctionSynonyms
[0][$function];
3441 return [ 'found' => false ];
3445 list( $callback, $flags ) = $this->mFunctionHooks
[$function];
3447 // Avoid PHP 7.1 warning from passing $this by reference
3450 $allArgs = [ &$parser ];
3451 if ( $flags & self
::SFH_OBJECT_ARGS
) {
3452 # Convert arguments to PPNodes and collect for appending to $allArgs
3454 foreach ( $args as $k => $v ) {
3455 if ( $v instanceof PPNode ||
$k === 0 ) {
3458 $funcArgs[] = $this->mPreprocessor
->newPartNodeArray( [ $k => $v ] )->item( 0 );
3462 # Add a frame parameter, and pass the arguments as an array
3463 $allArgs[] = $frame;
3464 $allArgs[] = $funcArgs;
3466 # Convert arguments to plain text and append to $allArgs
3467 foreach ( $args as $k => $v ) {
3468 if ( $v instanceof PPNode
) {
3469 $allArgs[] = trim( $frame->expand( $v ) );
3470 } elseif ( is_int( $k ) && $k >= 0 ) {
3471 $allArgs[] = trim( $v );
3473 $allArgs[] = trim( "$k=$v" );
3478 $result = $callback( ...$allArgs );
3480 # The interface for function hooks allows them to return a wikitext
3481 # string or an array containing the string and any flags. This mungs
3482 # things around to match what this method should return.
3483 if ( !is_array( $result ) ) {
3489 if ( isset( $result[0] ) && !isset( $result['text'] ) ) {
3490 $result['text'] = $result[0];
3492 unset( $result[0] );
3499 $preprocessFlags = 0;
3500 if ( isset( $result['noparse'] ) ) {
3501 $noparse = $result['noparse'];
3503 if ( isset( $result['preprocessFlags'] ) ) {
3504 $preprocessFlags = $result['preprocessFlags'];
3508 $result['text'] = $this->preprocessToDom( $result['text'], $preprocessFlags );
3509 $result['isChildObj'] = true;
3516 * Get the semi-parsed DOM representation of a template with a given title,
3517 * and its redirect destination title. Cached.
3519 * @param Title $title
3523 public function getTemplateDom( $title ) {
3524 $cacheTitle = $title;
3525 $titleText = $title->getPrefixedDBkey();
3527 if ( isset( $this->mTplRedirCache
[$titleText] ) ) {
3528 list( $ns, $dbk ) = $this->mTplRedirCache
[$titleText];
3529 $title = Title
::makeTitle( $ns, $dbk );
3530 $titleText = $title->getPrefixedDBkey();
3532 if ( isset( $this->mTplDomCache
[$titleText] ) ) {
3533 return [ $this->mTplDomCache
[$titleText], $title ];
3536 # Cache miss, go to the database
3537 list( $text, $title ) = $this->fetchTemplateAndTitle( $title );
3539 if ( $text === false ) {
3540 $this->mTplDomCache
[$titleText] = false;
3541 return [ false, $title ];
3544 $dom = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
3545 $this->mTplDomCache
[$titleText] = $dom;
3547 if ( !$title->equals( $cacheTitle ) ) {
3548 $this->mTplRedirCache
[$cacheTitle->getPrefixedDBkey()] =
3549 [ $title->getNamespace(), $title->getDBkey() ];
3552 return [ $dom, $title ];
3556 * Fetch the current revision of a given title. Note that the revision
3557 * (and even the title) may not exist in the database, so everything
3558 * contributing to the output of the parser should use this method
3559 * where possible, rather than getting the revisions themselves. This
3560 * method also caches its results, so using it benefits performance.
3563 * @param Title $title
3566 public function fetchCurrentRevisionOfTitle( $title ) {
3567 $cacheKey = $title->getPrefixedDBkey();
3568 if ( !$this->currentRevisionCache
) {
3569 $this->currentRevisionCache
= new MapCacheLRU( 100 );
3571 if ( !$this->currentRevisionCache
->has( $cacheKey ) ) {
3572 $this->currentRevisionCache
->set( $cacheKey,
3573 // Defaults to Parser::statelessFetchRevision()
3574 call_user_func( $this->mOptions
->getCurrentRevisionCallback(), $title, $this )
3577 return $this->currentRevisionCache
->get( $cacheKey );
3581 * Wrapper around Revision::newFromTitle to allow passing additional parameters
3582 * without passing them on to it.
3585 * @param Title $title
3586 * @param Parser|bool $parser
3587 * @return Revision|bool False if missing
3589 public static function statelessFetchRevision( Title
$title, $parser = false ) {
3590 $rev = Revision
::newKnownCurrent( wfGetDB( DB_REPLICA
), $title );
3596 * Fetch the unparsed text of a template and register a reference to it.
3597 * @param Title $title
3598 * @return array ( string or false, Title )
3600 public function fetchTemplateAndTitle( $title ) {
3601 // Defaults to Parser::statelessFetchTemplate()
3602 $templateCb = $this->mOptions
->getTemplateCallback();
3603 $stuff = call_user_func( $templateCb, $title, $this );
3604 // We use U+007F DELETE to distinguish strip markers from regular text.
3605 $text = $stuff['text'];
3606 if ( is_string( $stuff['text'] ) ) {
3607 $text = strtr( $text, "\x7f", "?" );
3609 $finalTitle = $stuff['finalTitle'] ??
$title;
3610 if ( isset( $stuff['deps'] ) ) {
3611 foreach ( $stuff['deps'] as $dep ) {
3612 $this->mOutput
->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
3613 if ( $dep['title']->equals( $this->getTitle() ) ) {
3614 // If we transclude ourselves, the final result
3615 // will change based on the new version of the page
3616 $this->mOutput
->setFlag( 'vary-revision' );
3620 return [ $text, $finalTitle ];
3624 * Fetch the unparsed text of a template and register a reference to it.
3625 * @param Title $title
3626 * @return string|bool
3628 public function fetchTemplate( $title ) {
3629 return $this->fetchTemplateAndTitle( $title )[0];
3633 * Static function to get a template
3634 * Can be overridden via ParserOptions::setTemplateCallback().
3636 * @param Title $title
3637 * @param bool|Parser $parser
3641 public static function statelessFetchTemplate( $title, $parser = false ) {
3642 $text = $skip = false;
3643 $finalTitle = $title;
3646 # Loop to fetch the article, with up to 1 redirect
3647 // phpcs:ignore Generic.CodeAnalysis.ForLoopWithTestFunctionCall
3648 for ( $i = 0; $i < 2 && is_object( $title ); $i++
) {
3649 # Give extensions a chance to select the revision instead
3650 $id = false; # Assume current
3651 Hooks
::run( 'BeforeParserFetchTemplateAndtitle',
3652 [ $parser, $title, &$skip, &$id ] );
3658 'page_id' => $title->getArticleID(),
3665 $rev = Revision
::newFromId( $id );
3666 } elseif ( $parser ) {
3667 $rev = $parser->fetchCurrentRevisionOfTitle( $title );
3669 $rev = Revision
::newFromTitle( $title );
3671 $rev_id = $rev ?
$rev->getId() : 0;
3672 # If there is no current revision, there is no page
3673 if ( $id === false && !$rev ) {
3674 $linkCache = MediaWikiServices
::getInstance()->getLinkCache();
3675 $linkCache->addBadLinkObj( $title );
3680 'page_id' => $title->getArticleID(),
3681 'rev_id' => $rev_id ];
3682 if ( $rev && !$title->equals( $rev->getTitle() ) ) {
3683 # We fetched a rev from a different title; register it too...
3685 'title' => $rev->getTitle(),
3686 'page_id' => $rev->getPage(),
3687 'rev_id' => $rev_id ];
3691 $content = $rev->getContent();
3692 $text = $content ?
$content->getWikitextForTransclusion() : null;
3694 Hooks
::run( 'ParserFetchTemplate',
3695 [ $parser, $title, $rev, &$text, &$deps ] );
3697 if ( $text === false ||
$text === null ) {
3701 } elseif ( $title->getNamespace() == NS_MEDIAWIKI
) {
3702 $message = wfMessage( MediaWikiServices
::getInstance()->getContentLanguage()->
3703 lcfirst( $title->getText() ) )->inContentLanguage();
3704 if ( !$message->exists() ) {
3708 $content = $message->content();
3709 $text = $message->plain();
3717 $finalTitle = $title;
3718 $title = $content->getRedirectTarget();
3722 'finalTitle' => $finalTitle,
3727 * Fetch a file and its title and register a reference to it.
3728 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3729 * @param Title $title
3730 * @param array $options Array of options to RepoGroup::findFile
3732 * @deprecated since 1.32, use fetchFileAndTitle instead
3734 public function fetchFile( $title, $options = [] ) {
3735 wfDeprecated( __METHOD__
, '1.32' );
3736 return $this->fetchFileAndTitle( $title, $options )[0];
3740 * Fetch a file and its title and register a reference to it.
3741 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3742 * @param Title $title
3743 * @param array $options Array of options to RepoGroup::findFile
3744 * @return array ( File or false, Title of file )
3746 public function fetchFileAndTitle( $title, $options = [] ) {
3747 $file = $this->fetchFileNoRegister( $title, $options );
3749 $time = $file ?
$file->getTimestamp() : false;
3750 $sha1 = $file ?
$file->getSha1() : false;
3751 # Register the file as a dependency...
3752 $this->mOutput
->addImage( $title->getDBkey(), $time, $sha1 );
3753 if ( $file && !$title->equals( $file->getTitle() ) ) {
3754 # Update fetched file title
3755 $title = $file->getTitle();
3756 $this->mOutput
->addImage( $title->getDBkey(), $time, $sha1 );
3758 return [ $file, $title ];
3762 * Helper function for fetchFileAndTitle.
3764 * Also useful if you need to fetch a file but not use it yet,
3765 * for example to get the file's handler.
3767 * @param Title $title
3768 * @param array $options Array of options to RepoGroup::findFile
3771 protected function fetchFileNoRegister( $title, $options = [] ) {
3772 if ( isset( $options['broken'] ) ) {
3773 $file = false; // broken thumbnail forced by hook
3774 } elseif ( isset( $options['sha1'] ) ) { // get by (sha1,timestamp)
3775 $file = RepoGroup
::singleton()->findFileFromKey( $options['sha1'], $options );
3776 } else { // get by (name,timestamp)
3777 $file = wfFindFile( $title, $options );
3783 * Transclude an interwiki link.
3785 * @param Title $title
3786 * @param string $action Usually one of (raw, render)
3790 public function interwikiTransclude( $title, $action ) {
3791 global $wgEnableScaryTranscluding, $wgTranscludeCacheExpiry;
3793 if ( !$wgEnableScaryTranscluding ) {
3794 return wfMessage( 'scarytranscludedisabled' )->inContentLanguage()->text();
3797 $url = $title->getFullURL( [ 'action' => $action ] );
3798 if ( strlen( $url ) > 1024 ) {
3799 return wfMessage( 'scarytranscludetoolong' )->inContentLanguage()->text();
3802 $wikiId = $title->getTransWikiID(); // remote wiki ID or false
3804 $fname = __METHOD__
;
3805 $cache = MediaWikiServices
::getInstance()->getMainWANObjectCache();
3807 $data = $cache->getWithSetCallback(
3808 $cache->makeGlobalKey(
3809 'interwiki-transclude',
3810 ( $wikiId !== false ) ?
$wikiId : 'external',
3813 $wgTranscludeCacheExpiry,
3814 function ( $oldValue, &$ttl ) use ( $url, $fname, $cache ) {
3815 $req = MWHttpRequest
::factory( $url, [], $fname );
3817 $status = $req->execute(); // Status object
3818 if ( !$status->isOK() ) {
3819 $ttl = $cache::TTL_UNCACHEABLE
;
3820 } elseif ( $req->getResponseHeader( 'X-Database-Lagged' ) !== null ) {
3821 $ttl = min( $cache::TTL_LAGGED
, $ttl );
3825 'text' => $status->isOK() ?
$req->getContent() : null,
3826 'code' => $req->getStatus()
3830 'checkKeys' => ( $wikiId !== false )
3831 ?
[ $cache->makeGlobalKey( 'interwiki-page', $wikiId, $title->getDBkey() ) ]
3833 'pcGroup' => 'interwiki-transclude:5',
3834 'pcTTL' => $cache::TTL_PROC_LONG
3838 if ( is_string( $data['text'] ) ) {
3839 $text = $data['text'];
3840 } elseif ( $data['code'] != 200 ) {
3841 // Though we failed to fetch the content, this status is useless.
3842 $text = wfMessage( 'scarytranscludefailed-httpstatus' )
3843 ->params( $url, $data['code'] )->inContentLanguage()->text();
3845 $text = wfMessage( 'scarytranscludefailed', $url )->inContentLanguage()->text();
3852 * Triple brace replacement -- used for template arguments
3855 * @param array $piece
3856 * @param PPFrame $frame
3860 public function argSubstitution( $piece, $frame ) {
3862 $parts = $piece['parts'];
3863 $nameWithSpaces = $frame->expand( $piece['title'] );
3864 $argName = trim( $nameWithSpaces );
3866 $text = $frame->getArgument( $argName );
3867 if ( $text === false && $parts->getLength() > 0
3868 && ( $this->ot
['html']
3870 ||
( $this->ot
['wiki'] && $frame->isTemplate() )
3873 # No match in frame, use the supplied default
3874 $object = $parts->item( 0 )->getChildren();
3876 if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
3877 $error = '<!-- WARNING: argument omitted, expansion size too large -->';
3878 $this->limitationWarn( 'post-expand-template-argument' );
3881 if ( $text === false && $object === false ) {
3883 $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts );
3885 if ( $error !== false ) {
3888 if ( $object !== false ) {
3889 $ret = [ 'object' => $object ];
3891 $ret = [ 'text' => $text ];
3898 * Return the text to be used for a given extension tag.
3899 * This is the ghost of strip().
3901 * @param array $params Associative array of parameters:
3902 * name PPNode for the tag name
3903 * attr PPNode for unparsed text where tag attributes are thought to be
3904 * attributes Optional associative array of parsed attributes
3905 * inner Contents of extension element
3906 * noClose Original text did not have a close tag
3907 * @param PPFrame $frame
3909 * @throws MWException
3912 public function extensionSubstitution( $params, $frame ) {
3913 static $errorStr = '<span class="error">';
3914 static $errorLen = 20;
3916 $name = $frame->expand( $params['name'] );
3917 if ( substr( $name, 0, $errorLen ) === $errorStr ) {
3918 // Probably expansion depth or node count exceeded. Just punt the
3923 $attrText = !isset( $params['attr'] ) ?
null : $frame->expand( $params['attr'] );
3924 if ( substr( $attrText, 0, $errorLen ) === $errorStr ) {
3929 // We can't safely check if the expansion for $content resulted in an
3930 // error, because the content could happen to be the error string
3932 $content = !isset( $params['inner'] ) ?
null : $frame->expand( $params['inner'] );
3934 $marker = self
::MARKER_PREFIX
. "-$name-"
3935 . sprintf( '%08X', $this->mMarkerIndex++
) . self
::MARKER_SUFFIX
;
3937 $isFunctionTag = isset( $this->mFunctionTagHooks
[strtolower( $name )] ) &&
3938 ( $this->ot
['html'] ||
$this->ot
['pre'] );
3939 if ( $isFunctionTag ) {
3940 $markerType = 'none';
3942 $markerType = 'general';
3944 if ( $this->ot
['html'] ||
$isFunctionTag ) {
3945 $name = strtolower( $name );
3946 $attributes = Sanitizer
::decodeTagAttributes( $attrText );
3947 if ( isset( $params['attributes'] ) ) {
3948 $attributes = $attributes +
$params['attributes'];
3951 if ( isset( $this->mTagHooks
[$name] ) ) {
3952 $output = call_user_func_array( $this->mTagHooks
[$name],
3953 [ $content, $attributes, $this, $frame ] );
3954 } elseif ( isset( $this->mFunctionTagHooks
[$name] ) ) {
3955 list( $callback, ) = $this->mFunctionTagHooks
[$name];
3957 // Avoid PHP 7.1 warning from passing $this by reference
3959 $output = call_user_func_array( $callback, [ &$parser, $frame, $content, $attributes ] );
3961 $output = '<span class="error">Invalid tag extension name: ' .
3962 htmlspecialchars( $name ) . '</span>';
3965 if ( is_array( $output ) ) {
3968 $output = $flags[0];
3969 if ( isset( $flags['markerType'] ) ) {
3970 $markerType = $flags['markerType'];
3974 if ( is_null( $attrText ) ) {
3977 if ( isset( $params['attributes'] ) ) {
3978 foreach ( $params['attributes'] as $attrName => $attrValue ) {
3979 $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' .
3980 htmlspecialchars( $attrValue ) . '"';
3983 if ( $content === null ) {
3984 $output = "<$name$attrText/>";
3986 $close = is_null( $params['close'] ) ?
'' : $frame->expand( $params['close'] );
3987 if ( substr( $close, 0, $errorLen ) === $errorStr ) {
3991 $output = "<$name$attrText>$content$close";
3995 if ( $markerType === 'none' ) {
3997 } elseif ( $markerType === 'nowiki' ) {
3998 $this->mStripState
->addNoWiki( $marker, $output );
3999 } elseif ( $markerType === 'general' ) {
4000 $this->mStripState
->addGeneral( $marker, $output );
4002 throw new MWException( __METHOD__
. ': invalid marker type' );
4008 * Increment an include size counter
4010 * @param string $type The type of expansion
4011 * @param int $size The size of the text
4012 * @return bool False if this inclusion would take it over the maximum, true otherwise
4014 public function incrementIncludeSize( $type, $size ) {
4015 if ( $this->mIncludeSizes
[$type] +
$size > $this->mOptions
->getMaxIncludeSize() ) {
4018 $this->mIncludeSizes
[$type] +
= $size;
4024 * Increment the expensive function count
4026 * @return bool False if the limit has been exceeded
4028 public function incrementExpensiveFunctionCount() {
4029 $this->mExpensiveFunctionCount++
;
4030 return $this->mExpensiveFunctionCount
<= $this->mOptions
->getExpensiveParserFunctionLimit();
4034 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
4035 * Fills $this->mDoubleUnderscores, returns the modified text
4037 * @param string $text
4041 public function doDoubleUnderscore( $text ) {
4042 # The position of __TOC__ needs to be recorded
4043 $mw = $this->magicWordFactory
->get( 'toc' );
4044 if ( $mw->match( $text ) ) {
4045 $this->mShowToc
= true;
4046 $this->mForceTocPosition
= true;
4048 # Set a placeholder. At the end we'll fill it in with the TOC.
4049 $text = $mw->replace( '<!--MWTOC\'"-->', $text, 1 );
4051 # Only keep the first one.
4052 $text = $mw->replace( '', $text );
4055 # Now match and remove the rest of them
4056 $mwa = $this->magicWordFactory
->getDoubleUnderscoreArray();
4057 $this->mDoubleUnderscores
= $mwa->matchAndRemove( $text );
4059 if ( isset( $this->mDoubleUnderscores
['nogallery'] ) ) {
4060 $this->mOutput
->mNoGallery
= true;
4062 if ( isset( $this->mDoubleUnderscores
['notoc'] ) && !$this->mForceTocPosition
) {
4063 $this->mShowToc
= false;
4065 if ( isset( $this->mDoubleUnderscores
['hiddencat'] )
4066 && $this->mTitle
->getNamespace() == NS_CATEGORY
4068 $this->addTrackingCategory( 'hidden-category-category' );
4070 # (T10068) Allow control over whether robots index a page.
4071 # __INDEX__ always overrides __NOINDEX__, see T16899
4072 if ( isset( $this->mDoubleUnderscores
['noindex'] ) && $this->mTitle
->canUseNoindex() ) {
4073 $this->mOutput
->setIndexPolicy( 'noindex' );
4074 $this->addTrackingCategory( 'noindex-category' );
4076 if ( isset( $this->mDoubleUnderscores
['index'] ) && $this->mTitle
->canUseNoindex() ) {
4077 $this->mOutput
->setIndexPolicy( 'index' );
4078 $this->addTrackingCategory( 'index-category' );
4081 # Cache all double underscores in the database
4082 foreach ( $this->mDoubleUnderscores
as $key => $val ) {
4083 $this->mOutput
->setProperty( $key, '' );
4090 * @see ParserOutput::addTrackingCategory()
4091 * @param string $msg Message key
4092 * @return bool Whether the addition was successful
4094 public function addTrackingCategory( $msg ) {
4095 return $this->mOutput
->addTrackingCategory( $msg, $this->mTitle
);
4099 * This function accomplishes several tasks:
4100 * 1) Auto-number headings if that option is enabled
4101 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
4102 * 3) Add a Table of contents on the top for users who have enabled the option
4103 * 4) Auto-anchor headings
4105 * It loops through all headlines, collects the necessary data, then splits up the
4106 * string and re-inserts the newly formatted headlines.
4108 * @param string $text
4109 * @param string $origText Original, untouched wikitext
4110 * @param bool $isMain
4111 * @return mixed|string
4114 public function formatHeadings( $text, $origText, $isMain = true ) {
4115 global $wgMaxTocLevel;
4117 # Inhibit editsection links if requested in the page
4118 if ( isset( $this->mDoubleUnderscores
['noeditsection'] ) ) {
4119 $maybeShowEditLink = false;
4121 $maybeShowEditLink = true; /* Actual presence will depend on post-cache transforms */
4124 # Get all headlines for numbering them and adding funky stuff like [edit]
4125 # links - this is for later, but we need the number of headlines right now
4126 # NOTE: white space in headings have been trimmed in doHeadings. They shouldn't
4127 # be trimmed here since whitespace in HTML headings is significant.
4129 $numMatches = preg_match_all(
4130 '/<H(?P<level>[1-6])(?P<attrib>.*?>)(?P<header>[\s\S]*?)<\/H[1-6] *>/i',
4135 # if there are fewer than 4 headlines in the article, do not show TOC
4136 # unless it's been explicitly enabled.
4137 $enoughToc = $this->mShowToc
&&
4138 ( ( $numMatches >= 4 ) ||
$this->mForceTocPosition
);
4140 # Allow user to stipulate that a page should have a "new section"
4141 # link added via __NEWSECTIONLINK__
4142 if ( isset( $this->mDoubleUnderscores
['newsectionlink'] ) ) {
4143 $this->mOutput
->setNewSection( true );
4146 # Allow user to remove the "new section"
4147 # link via __NONEWSECTIONLINK__
4148 if ( isset( $this->mDoubleUnderscores
['nonewsectionlink'] ) ) {
4149 $this->mOutput
->hideNewSection( true );
4152 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
4153 # override above conditions and always show TOC above first header
4154 if ( isset( $this->mDoubleUnderscores
['forcetoc'] ) ) {
4155 $this->mShowToc
= true;
4163 # Ugh .. the TOC should have neat indentation levels which can be
4164 # passed to the skin functions. These are determined here
4168 $sublevelCount = [];
4174 $markerRegex = self
::MARKER_PREFIX
. "-h-(\d+)-" . self
::MARKER_SUFFIX
;
4175 $baseTitleText = $this->mTitle
->getPrefixedDBkey();
4176 $oldType = $this->mOutputType
;
4177 $this->setOutputType( self
::OT_WIKI
);
4178 $frame = $this->getPreprocessor()->newFrame();
4179 $root = $this->preprocessToDom( $origText );
4180 $node = $root->getFirstChild();
4185 $headlines = $numMatches !== false ?
$matches[3] : [];
4187 foreach ( $headlines as $headline ) {
4188 $isTemplate = false;
4190 $sectionIndex = false;
4192 $markerMatches = [];
4193 if ( preg_match( "/^$markerRegex/", $headline, $markerMatches ) ) {
4194 $serial = $markerMatches[1];
4195 list( $titleText, $sectionIndex ) = $this->mHeadings
[$serial];
4196 $isTemplate = ( $titleText != $baseTitleText );
4197 $headline = preg_replace( "/^$markerRegex\\s*/", "", $headline );
4201 $prevlevel = $level;
4203 $level = $matches[1][$headlineCount];
4205 if ( $level > $prevlevel ) {
4206 # Increase TOC level
4208 $sublevelCount[$toclevel] = 0;
4209 if ( $toclevel < $wgMaxTocLevel ) {
4210 $prevtoclevel = $toclevel;
4211 $toc .= Linker
::tocIndent();
4214 } elseif ( $level < $prevlevel && $toclevel > 1 ) {
4215 # Decrease TOC level, find level to jump to
4217 for ( $i = $toclevel; $i > 0; $i-- ) {
4218 if ( $levelCount[$i] == $level ) {
4219 # Found last matching level
4222 } elseif ( $levelCount[$i] < $level ) {
4223 # Found first matching level below current level
4231 if ( $toclevel < $wgMaxTocLevel ) {
4232 if ( $prevtoclevel < $wgMaxTocLevel ) {
4233 # Unindent only if the previous toc level was shown :p
4234 $toc .= Linker
::tocUnindent( $prevtoclevel - $toclevel );
4235 $prevtoclevel = $toclevel;
4237 $toc .= Linker
::tocLineEnd();
4241 # No change in level, end TOC line
4242 if ( $toclevel < $wgMaxTocLevel ) {
4243 $toc .= Linker
::tocLineEnd();
4247 $levelCount[$toclevel] = $level;
4249 # count number of headlines for each level
4250 $sublevelCount[$toclevel]++
;
4252 for ( $i = 1; $i <= $toclevel; $i++
) {
4253 if ( !empty( $sublevelCount[$i] ) ) {
4257 $numbering .= $this->getTargetLanguage()->formatNum( $sublevelCount[$i] );
4262 # The safe header is a version of the header text safe to use for links
4264 # Remove link placeholders by the link text.
4265 # <!--LINK number-->
4267 # link text with suffix
4268 # Do this before unstrip since link text can contain strip markers
4269 $safeHeadline = $this->replaceLinkHoldersText( $headline );
4271 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4272 $safeHeadline = $this->mStripState
->unstripBoth( $safeHeadline );
4274 # Remove any <style> or <script> tags (T198618)
4275 $safeHeadline = preg_replace(
4276 '#<(style|script)(?: [^>]*[^>/])?>.*?</\1>#is',
4281 # Strip out HTML (first regex removes any tag not allowed)
4283 # * <sup> and <sub> (T10393)
4287 # * <span dir="rtl"> and <span dir="ltr"> (T37167)
4288 # * <s> and <strike> (T35715)
4289 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
4290 # to allow setting directionality in toc items.
4291 $tocline = preg_replace(
4293 '#<(?!/?(span|sup|sub|bdi|i|b|s|strike)(?: [^>]*)?>).*?>#',
4294 '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|bdi|i|b|s|strike))(?: .*?)?>#'
4300 # Strip '<span></span>', which is the result from the above if
4301 # <span id="foo"></span> is used to produce an additional anchor
4303 $tocline = str_replace( '<span></span>', '', $tocline );
4305 $tocline = trim( $tocline );
4307 # For the anchor, strip out HTML-y stuff period
4308 $safeHeadline = preg_replace( '/<.*?>/', '', $safeHeadline );
4309 $safeHeadline = Sanitizer
::normalizeSectionNameWhitespace( $safeHeadline );
4311 # Save headline for section edit hint before it's escaped
4312 $headlineHint = $safeHeadline;
4314 # Decode HTML entities
4315 $safeHeadline = Sanitizer
::decodeCharReferences( $safeHeadline );
4317 $safeHeadline = self
::normalizeSectionName( $safeHeadline );
4319 $fallbackHeadline = Sanitizer
::escapeIdForAttribute( $safeHeadline, Sanitizer
::ID_FALLBACK
);
4320 $linkAnchor = Sanitizer
::escapeIdForLink( $safeHeadline );
4321 $safeHeadline = Sanitizer
::escapeIdForAttribute( $safeHeadline, Sanitizer
::ID_PRIMARY
);
4322 if ( $fallbackHeadline === $safeHeadline ) {
4323 # No reason to have both (in fact, we can't)
4324 $fallbackHeadline = false;
4327 # HTML IDs must be case-insensitively unique for IE compatibility (T12721).
4328 # @todo FIXME: We may be changing them depending on the current locale.
4329 $arrayKey = strtolower( $safeHeadline );
4330 if ( $fallbackHeadline === false ) {
4331 $fallbackArrayKey = false;
4333 $fallbackArrayKey = strtolower( $fallbackHeadline );
4336 # Create the anchor for linking from the TOC to the section
4337 $anchor = $safeHeadline;
4338 $fallbackAnchor = $fallbackHeadline;
4339 if ( isset( $refers[$arrayKey] ) ) {
4340 // phpcs:ignore Generic.CodeAnalysis.ForLoopWithTestFunctionCall,Generic.Formatting.DisallowMultipleStatements
4341 for ( $i = 2; isset( $refers["${arrayKey}_$i"] ); ++
$i );
4343 $linkAnchor .= "_$i";
4344 $refers["${arrayKey}_$i"] = true;
4346 $refers[$arrayKey] = true;
4348 if ( $fallbackHeadline !== false && isset( $refers[$fallbackArrayKey] ) ) {
4349 // phpcs:ignore Generic.CodeAnalysis.ForLoopWithTestFunctionCall,Generic.Formatting.DisallowMultipleStatements
4350 for ( $i = 2; isset( $refers["${fallbackArrayKey}_$i"] ); ++
$i );
4351 $fallbackAnchor .= "_$i";
4352 $refers["${fallbackArrayKey}_$i"] = true;
4354 $refers[$fallbackArrayKey] = true;
4357 # Don't number the heading if it is the only one (looks silly)
4358 if ( count( $matches[3] ) > 1 && $this->mOptions
->getNumberHeadings() ) {
4359 # the two are different if the line contains a link
4360 $headline = Html
::element(
4362 [ 'class' => 'mw-headline-number' ],
4364 ) . ' ' . $headline;
4367 if ( $enoughToc && ( !isset( $wgMaxTocLevel ) ||
$toclevel < $wgMaxTocLevel ) ) {
4368 $toc .= Linker
::tocLine( $linkAnchor, $tocline,
4369 $numbering, $toclevel, ( $isTemplate ?
false : $sectionIndex ) );
4372 # Add the section to the section tree
4373 # Find the DOM node for this header
4374 $noOffset = ( $isTemplate ||
$sectionIndex === false );
4375 while ( $node && !$noOffset ) {
4376 if ( $node->getName() === 'h' ) {
4377 $bits = $node->splitHeading();
4378 if ( $bits['i'] == $sectionIndex ) {
4382 $byteOffset +
= mb_strlen( $this->mStripState
->unstripBoth(
4383 $frame->expand( $node, PPFrame
::RECOVER_ORIG
) ) );
4384 $node = $node->getNextSibling();
4387 'toclevel' => $toclevel,
4390 'number' => $numbering,
4391 'index' => ( $isTemplate ?
'T-' : '' ) . $sectionIndex,
4392 'fromtitle' => $titleText,
4393 'byteoffset' => ( $noOffset ?
null : $byteOffset ),
4394 'anchor' => $anchor,
4397 # give headline the correct <h#> tag
4398 if ( $maybeShowEditLink && $sectionIndex !== false ) {
4399 // Output edit section links as markers with styles that can be customized by skins
4400 if ( $isTemplate ) {
4401 # Put a T flag in the section identifier, to indicate to extractSections()
4402 # that sections inside <includeonly> should be counted.
4403 $editsectionPage = $titleText;
4404 $editsectionSection = "T-$sectionIndex";
4405 $editsectionContent = null;
4407 $editsectionPage = $this->mTitle
->getPrefixedText();
4408 $editsectionSection = $sectionIndex;
4409 $editsectionContent = $headlineHint;
4411 // We use a bit of pesudo-xml for editsection markers. The
4412 // language converter is run later on. Using a UNIQ style marker
4413 // leads to the converter screwing up the tokens when it
4414 // converts stuff. And trying to insert strip tags fails too. At
4415 // this point all real inputted tags have already been escaped,
4416 // so we don't have to worry about a user trying to input one of
4417 // these markers directly. We use a page and section attribute
4418 // to stop the language converter from converting these
4419 // important bits of data, but put the headline hint inside a
4420 // content block because the language converter is supposed to
4421 // be able to convert that piece of data.
4422 // Gets replaced with html in ParserOutput::getText
4423 $editlink = '<mw:editsection page="' . htmlspecialchars( $editsectionPage );
4424 $editlink .= '" section="' . htmlspecialchars( $editsectionSection ) . '"';
4425 if ( $editsectionContent !== null ) {
4426 $editlink .= '>' . $editsectionContent . '</mw:editsection>';
4433 $head[$headlineCount] = Linker
::makeHeadline( $level,
4434 $matches['attrib'][$headlineCount], $anchor, $headline,
4435 $editlink, $fallbackAnchor );
4440 $this->setOutputType( $oldType );
4442 # Never ever show TOC if no headers
4443 if ( $numVisible < 1 ) {
4448 if ( $prevtoclevel > 0 && $prevtoclevel < $wgMaxTocLevel ) {
4449 $toc .= Linker
::tocUnindent( $prevtoclevel - 1 );
4451 $toc = Linker
::tocList( $toc, $this->mOptions
->getUserLangObj() );
4452 $this->mOutput
->setTOCHTML( $toc );
4453 $toc = self
::TOC_START
. $toc . self
::TOC_END
;
4457 $this->mOutput
->setSections( $tocraw );
4460 # split up and insert constructed headlines
4461 $blocks = preg_split( '/<H[1-6].*?>[\s\S]*?<\/H[1-6]>/i', $text );
4464 // build an array of document sections
4466 foreach ( $blocks as $block ) {
4467 // $head is zero-based, sections aren't.
4468 if ( empty( $head[$i - 1] ) ) {
4469 $sections[$i] = $block;
4471 $sections[$i] = $head[$i - 1] . $block;
4475 * Send a hook, one per section.
4476 * The idea here is to be able to make section-level DIVs, but to do so in a
4477 * lower-impact, more correct way than r50769
4480 * $section : the section number
4481 * &$sectionContent : ref to the content of the section
4482 * $maybeShowEditLinks : boolean describing whether this section has an edit link
4484 Hooks
::run( 'ParserSectionCreate', [ $this, $i, &$sections[$i], $maybeShowEditLink ] );
4489 if ( $enoughToc && $isMain && !$this->mForceTocPosition
) {
4490 // append the TOC at the beginning
4491 // Top anchor now in skin
4492 $sections[0] = $sections[0] . $toc . "\n";
4495 $full .= implode( '', $sections );
4497 if ( $this->mForceTocPosition
) {
4498 return str_replace( '<!--MWTOC\'"-->', $toc, $full );
4505 * Transform wiki markup when saving a page by doing "\r\n" -> "\n"
4506 * conversion, substituting signatures, {{subst:}} templates, etc.
4508 * @param string $text The text to transform
4509 * @param Title $title The Title object for the current article
4510 * @param User $user The User object describing the current user
4511 * @param ParserOptions $options Parsing options
4512 * @param bool $clearState Whether to clear the parser state first
4513 * @return string The altered wiki markup
4515 public function preSaveTransform( $text, Title
$title, User
$user,
4516 ParserOptions
$options, $clearState = true
4518 if ( $clearState ) {
4519 $magicScopeVariable = $this->lock();
4521 $this->startParse( $title, $options, self
::OT_WIKI
, $clearState );
4522 $this->setUser( $user );
4524 // Strip U+0000 NULL (T159174)
4525 $text = str_replace( "\000", '', $text );
4527 // We still normalize line endings for backwards-compatibility
4528 // with other code that just calls PST, but this should already
4529 // be handled in TextContent subclasses
4530 $text = TextContent
::normalizeLineEndings( $text );
4532 if ( $options->getPreSaveTransform() ) {
4533 $text = $this->pstPass2( $text, $user );
4535 $text = $this->mStripState
->unstripBoth( $text );
4537 $this->setUser( null ); # Reset
4543 * Pre-save transform helper function
4545 * @param string $text
4550 private function pstPass2( $text, $user ) {
4551 # Note: This is the timestamp saved as hardcoded wikitext to the database, we use
4552 # $this->contLang here in order to give everyone the same signature and use the default one
4553 # rather than the one selected in each user's preferences. (see also T14815)
4554 $ts = $this->mOptions
->getTimestamp();
4555 $timestamp = MWTimestamp
::getLocalInstance( $ts );
4556 $ts = $timestamp->format( 'YmdHis' );
4557 $tzMsg = $timestamp->getTimezoneMessage()->inContentLanguage()->text();
4559 $d = $this->contLang
->timeanddate( $ts, false, false ) . " ($tzMsg)";
4561 # Variable replacement
4562 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4563 $text = $this->replaceVariables( $text );
4565 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4566 # which may corrupt this parser instance via its wfMessage()->text() call-
4569 if ( strpos( $text, '~~~' ) !== false ) {
4570 $sigText = $this->getUserSig( $user );
4571 $text = strtr( $text, [
4573 '~~~~' => "$sigText $d",
4576 # The main two signature forms used above are time-sensitive
4577 $this->mOutput
->setFlag( 'user-signature' );
4580 # Context links ("pipe tricks"): [[|name]] and [[name (context)|]]
4581 $tc = '[' . Title
::legalChars() . ']';
4582 $nc = '[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can use non-ascii!
4584 // [[ns:page (context)|]]
4585 $p1 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\))\\|]]/";
4586 // [[ns:page(context)|]] (double-width brackets, added in r40257)
4587 $p4 = "/\[\[(:?$nc+:|:|)($tc+?)( ?($tc+))\\|]]/";
4588 // [[ns:page (context), context|]] (using either single or double-width comma)
4589 $p3 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\)|)((?:, |,)$tc+|)\\|]]/";
4590 // [[|page]] (reverse pipe trick: add context from page title)
4591 $p2 = "/\[\[\\|($tc+)]]/";
4593 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4594 $text = preg_replace( $p1, '[[\\1\\2\\3|\\2]]', $text );
4595 $text = preg_replace( $p4, '[[\\1\\2\\3|\\2]]', $text );
4596 $text = preg_replace( $p3, '[[\\1\\2\\3\\4|\\2]]', $text );
4598 $t = $this->mTitle
->getText();
4600 if ( preg_match( "/^($nc+:|)$tc+?( \\($tc+\\))$/", $t, $m ) ) {
4601 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4602 } elseif ( preg_match( "/^($nc+:|)$tc+?(, $tc+|)$/", $t, $m ) && "$m[1]$m[2]" != '' ) {
4603 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4605 # if there's no context, don't bother duplicating the title
4606 $text = preg_replace( $p2, '[[\\1]]', $text );
4613 * Fetch the user's signature text, if any, and normalize to
4614 * validated, ready-to-insert wikitext.
4615 * If you have pre-fetched the nickname or the fancySig option, you can
4616 * specify them here to save a database query.
4617 * Do not reuse this parser instance after calling getUserSig(),
4618 * as it may have changed if it's the $wgParser.
4620 * @param User &$user
4621 * @param string|bool $nickname Nickname to use or false to use user's default nickname
4622 * @param bool|null $fancySig whether the nicknname is the complete signature
4623 * or null to use default value
4626 public function getUserSig( &$user, $nickname = false, $fancySig = null ) {
4627 global $wgMaxSigChars;
4629 $username = $user->getName();
4631 # If not given, retrieve from the user object.
4632 if ( $nickname === false ) {
4633 $nickname = $user->getOption( 'nickname' );
4636 if ( is_null( $fancySig ) ) {
4637 $fancySig = $user->getBoolOption( 'fancysig' );
4640 $nickname = $nickname == null ?
$username : $nickname;
4642 if ( mb_strlen( $nickname ) > $wgMaxSigChars ) {
4643 $nickname = $username;
4644 wfDebug( __METHOD__
. ": $username has overlong signature.\n" );
4645 } elseif ( $fancySig !== false ) {
4646 # Sig. might contain markup; validate this
4647 if ( $this->validateSig( $nickname ) !== false ) {
4648 # Validated; clean up (if needed) and return it
4649 return $this->cleanSig( $nickname, true );
4651 # Failed to validate; fall back to the default
4652 $nickname = $username;
4653 wfDebug( __METHOD__
. ": $username has bad XML tags in signature.\n" );
4657 # Make sure nickname doesnt get a sig in a sig
4658 $nickname = self
::cleanSigInSig( $nickname );
4660 # If we're still here, make it a link to the user page
4661 $userText = wfEscapeWikiText( $username );
4662 $nickText = wfEscapeWikiText( $nickname );
4663 $msgName = $user->isAnon() ?
'signature-anon' : 'signature';
4665 return wfMessage( $msgName, $userText, $nickText )->inContentLanguage()
4666 ->title( $this->getTitle() )->text();
4670 * Check that the user's signature contains no bad XML
4672 * @param string $text
4673 * @return string|bool An expanded string, or false if invalid.
4675 public function validateSig( $text ) {
4676 return Xml
::isWellFormedXmlFragment( $text ) ?
$text : false;
4680 * Clean up signature text
4682 * 1) Strip 3, 4 or 5 tildes out of signatures @see cleanSigInSig
4683 * 2) Substitute all transclusions
4685 * @param string $text
4686 * @param bool $parsing Whether we're cleaning (preferences save) or parsing
4687 * @return string Signature text
4689 public function cleanSig( $text, $parsing = false ) {
4692 $magicScopeVariable = $this->lock();
4693 $this->startParse( $wgTitle, new ParserOptions
, self
::OT_PREPROCESS
, true );
4696 # Option to disable this feature
4697 if ( !$this->mOptions
->getCleanSignatures() ) {
4701 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4702 # => Move this logic to braceSubstitution()
4703 $substWord = $this->magicWordFactory
->get( 'subst' );
4704 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4705 $substText = '{{' . $substWord->getSynonym( 0 );
4707 $text = preg_replace( $substRegex, $substText, $text );
4708 $text = self
::cleanSigInSig( $text );
4709 $dom = $this->preprocessToDom( $text );
4710 $frame = $this->getPreprocessor()->newFrame();
4711 $text = $frame->expand( $dom );
4714 $text = $this->mStripState
->unstripBoth( $text );
4721 * Strip 3, 4 or 5 tildes out of signatures.
4723 * @param string $text
4724 * @return string Signature text with /~{3,5}/ removed
4726 public static function cleanSigInSig( $text ) {
4727 $text = preg_replace( '/~{3,5}/', '', $text );
4732 * Set up some variables which are usually set up in parse()
4733 * so that an external function can call some class members with confidence
4735 * @param Title|null $title
4736 * @param ParserOptions $options
4737 * @param int $outputType
4738 * @param bool $clearState
4740 public function startExternalParse( Title
$title = null, ParserOptions
$options,
4741 $outputType, $clearState = true
4743 $this->startParse( $title, $options, $outputType, $clearState );
4747 * @param Title|null $title
4748 * @param ParserOptions $options
4749 * @param int $outputType
4750 * @param bool $clearState
4752 private function startParse( Title
$title = null, ParserOptions
$options,
4753 $outputType, $clearState = true
4755 $this->setTitle( $title );
4756 $this->mOptions
= $options;
4757 $this->setOutputType( $outputType );
4758 if ( $clearState ) {
4759 $this->clearState();
4764 * Wrapper for preprocess()
4766 * @param string $text The text to preprocess
4767 * @param ParserOptions $options
4768 * @param Title|null $title Title object or null to use $wgTitle
4771 public function transformMsg( $text, $options, $title = null ) {
4772 static $executing = false;
4774 # Guard against infinite recursion
4785 $text = $this->preprocess( $text, $title, $options );
4792 * Create an HTML-style tag, e.g. "<yourtag>special text</yourtag>"
4793 * The callback should have the following form:
4794 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4796 * Transform and return $text. Use $parser for any required context, e.g. use
4797 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4799 * Hooks may return extended information by returning an array, of which the
4800 * first numbered element (index 0) must be the return string, and all other
4801 * entries are extracted into local variables within an internal function
4802 * in the Parser class.
4804 * This interface (introduced r61913) appears to be undocumented, but
4805 * 'markerType' is used by some core tag hooks to override which strip
4806 * array their results are placed in. **Use great caution if attempting
4807 * this interface, as it is not documented and injudicious use could smash
4808 * private variables.**
4810 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4811 * @param callable $callback The callback function (and object) to use for the tag
4812 * @throws MWException
4813 * @return callable|null The old value of the mTagHooks array associated with the hook
4815 public function setHook( $tag, callable
$callback ) {
4816 $tag = strtolower( $tag );
4817 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4818 throw new MWException( "Invalid character {$m[0]} in setHook('$tag', ...) call" );
4820 $oldVal = $this->mTagHooks
[$tag] ??
null;
4821 $this->mTagHooks
[$tag] = $callback;
4822 if ( !in_array( $tag, $this->mStripList
) ) {
4823 $this->mStripList
[] = $tag;
4830 * As setHook(), but letting the contents be parsed.
4832 * Transparent tag hooks are like regular XML-style tag hooks, except they
4833 * operate late in the transformation sequence, on HTML instead of wikitext.
4835 * This is probably obsoleted by things dealing with parser frames?
4836 * The only extension currently using it is geoserver.
4839 * @todo better document or deprecate this
4841 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4842 * @param callable $callback The callback function (and object) to use for the tag
4843 * @throws MWException
4844 * @return callable|null The old value of the mTagHooks array associated with the hook
4846 public function setTransparentTagHook( $tag, callable
$callback ) {
4847 $tag = strtolower( $tag );
4848 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4849 throw new MWException( "Invalid character {$m[0]} in setTransparentHook('$tag', ...) call" );
4851 $oldVal = $this->mTransparentTagHooks
[$tag] ??
null;
4852 $this->mTransparentTagHooks
[$tag] = $callback;
4858 * Remove all tag hooks
4860 public function clearTagHooks() {
4861 $this->mTagHooks
= [];
4862 $this->mFunctionTagHooks
= [];
4863 $this->mStripList
= $this->mDefaultStripList
;
4867 * Create a function, e.g. {{sum:1|2|3}}
4868 * The callback function should have the form:
4869 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4871 * Or with Parser::SFH_OBJECT_ARGS:
4872 * function myParserFunction( $parser, $frame, $args ) { ... }
4874 * The callback may either return the text result of the function, or an array with the text
4875 * in element 0, and a number of flags in the other elements. The names of the flags are
4876 * specified in the keys. Valid flags are:
4877 * found The text returned is valid, stop processing the template. This
4879 * nowiki Wiki markup in the return value should be escaped
4880 * isHTML The returned text is HTML, armour it against wikitext transformation
4882 * @param string $id The magic word ID
4883 * @param callable $callback The callback function (and object) to use
4884 * @param int $flags A combination of the following flags:
4885 * Parser::SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4887 * Parser::SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text.
4888 * This allows for conditional expansion of the parse tree, allowing you to eliminate dead
4889 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4890 * the arguments, and to control the way they are expanded.
4892 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4893 * arguments, for instance:
4894 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4896 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4897 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4898 * working if/when this is changed.
4900 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4903 * Please read the documentation in includes/parser/Preprocessor.php for more information
4904 * about the methods available in PPFrame and PPNode.
4906 * @throws MWException
4907 * @return string|callable The old callback function for this name, if any
4909 public function setFunctionHook( $id, callable
$callback, $flags = 0 ) {
4910 $oldVal = isset( $this->mFunctionHooks
[$id] ) ?
$this->mFunctionHooks
[$id][0] : null;
4911 $this->mFunctionHooks
[$id] = [ $callback, $flags ];
4913 # Add to function cache
4914 $mw = $this->magicWordFactory
->get( $id );
4916 throw new MWException( __METHOD__
. '() expecting a magic word identifier.' );
4919 $synonyms = $mw->getSynonyms();
4920 $sensitive = intval( $mw->isCaseSensitive() );
4922 foreach ( $synonyms as $syn ) {
4924 if ( !$sensitive ) {
4925 $syn = $this->contLang
->lc( $syn );
4928 if ( !( $flags & self
::SFH_NO_HASH
) ) {
4931 # Remove trailing colon
4932 if ( substr( $syn, -1, 1 ) === ':' ) {
4933 $syn = substr( $syn, 0, -1 );
4935 $this->mFunctionSynonyms
[$sensitive][$syn] = $id;
4941 * Get all registered function hook identifiers
4945 public function getFunctionHooks() {
4946 $this->firstCallInit();
4947 return array_keys( $this->mFunctionHooks
);
4951 * Create a tag function, e.g. "<test>some stuff</test>".
4952 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4953 * Unlike parser functions, their content is not preprocessed.
4954 * @param string $tag
4955 * @param callable $callback
4957 * @throws MWException
4960 public function setFunctionTagHook( $tag, callable
$callback, $flags ) {
4961 $tag = strtolower( $tag );
4962 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4963 throw new MWException( "Invalid character {$m[0]} in setFunctionTagHook('$tag', ...) call" );
4965 $old = $this->mFunctionTagHooks
[$tag] ??
null;
4966 $this->mFunctionTagHooks
[$tag] = [ $callback, $flags ];
4968 if ( !in_array( $tag, $this->mStripList
) ) {
4969 $this->mStripList
[] = $tag;
4976 * Replace "<!--LINK-->" link placeholders with actual links, in the buffer
4977 * Placeholders created in Linker::link()
4979 * @param string &$text
4980 * @param int $options
4982 public function replaceLinkHolders( &$text, $options = 0 ) {
4983 $this->mLinkHolders
->replace( $text );
4987 * Replace "<!--LINK-->" link placeholders with plain text of links
4988 * (not HTML-formatted).
4990 * @param string $text
4993 public function replaceLinkHoldersText( $text ) {
4994 return $this->mLinkHolders
->replaceText( $text );
4998 * Renders an image gallery from a text with one line per image.
4999 * text labels may be given by using |-style alternative text. E.g.
5000 * Image:one.jpg|The number "1"
5001 * Image:tree.jpg|A tree
5002 * given as text will return the HTML of a gallery with two images,
5003 * labeled 'The number "1"' and
5006 * @param string $text
5007 * @param array $params
5008 * @return string HTML
5010 public function renderImageGallery( $text, $params ) {
5012 if ( isset( $params['mode'] ) ) {
5013 $mode = $params['mode'];
5017 $ig = ImageGalleryBase
::factory( $mode );
5018 } catch ( Exception
$e ) {
5019 // If invalid type set, fallback to default.
5020 $ig = ImageGalleryBase
::factory( false );
5023 $ig->setContextTitle( $this->mTitle
);
5024 $ig->setShowBytes( false );
5025 $ig->setShowDimensions( false );
5026 $ig->setShowFilename( false );
5027 $ig->setParser( $this );
5028 $ig->setHideBadImages();
5029 $ig->setAttributes( Sanitizer
::validateTagAttributes( $params, 'ul' ) );
5031 if ( isset( $params['showfilename'] ) ) {
5032 $ig->setShowFilename( true );
5034 $ig->setShowFilename( false );
5036 if ( isset( $params['caption'] ) ) {
5037 $caption = $params['caption'];
5038 $caption = htmlspecialchars( $caption );
5039 $caption = $this->replaceInternalLinks( $caption );
5040 $ig->setCaptionHtml( $caption );
5042 if ( isset( $params['perrow'] ) ) {
5043 $ig->setPerRow( $params['perrow'] );
5045 if ( isset( $params['widths'] ) ) {
5046 $ig->setWidths( $params['widths'] );
5048 if ( isset( $params['heights'] ) ) {
5049 $ig->setHeights( $params['heights'] );
5051 $ig->setAdditionalOptions( $params );
5053 // Avoid PHP 7.1 warning from passing $this by reference
5055 Hooks
::run( 'BeforeParserrenderImageGallery', [ &$parser, &$ig ] );
5057 $lines = StringUtils
::explode( "\n", $text );
5058 foreach ( $lines as $line ) {
5059 # match lines like these:
5060 # Image:someimage.jpg|This is some image
5062 preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
5064 if ( count( $matches ) == 0 ) {
5068 if ( strpos( $matches[0], '%' ) !== false ) {
5069 $matches[1] = rawurldecode( $matches[1] );
5071 $title = Title
::newFromText( $matches[1], NS_FILE
);
5072 if ( is_null( $title ) ) {
5073 # Bogus title. Ignore these so we don't bomb out later.
5077 # We need to get what handler the file uses, to figure out parameters.
5078 # Note, a hook can overide the file name, and chose an entirely different
5079 # file (which potentially could be of a different type and have different handler).
5082 Hooks
::run( 'BeforeParserFetchFileAndTitle',
5083 [ $this, $title, &$options, &$descQuery ] );
5084 # Don't register it now, as TraditionalImageGallery does that later.
5085 $file = $this->fetchFileNoRegister( $title, $options );
5086 $handler = $file ?
$file->getHandler() : false;
5089 'img_alt' => 'gallery-internal-alt',
5090 'img_link' => 'gallery-internal-link',
5093 $paramMap = $paramMap +
$handler->getParamMap();
5094 // We don't want people to specify per-image widths.
5095 // Additionally the width parameter would need special casing anyhow.
5096 unset( $paramMap['img_width'] );
5099 $mwArray = $this->magicWordFactory
->newArray( array_keys( $paramMap ) );
5104 $handlerOptions = [];
5105 if ( isset( $matches[3] ) ) {
5106 // look for an |alt= definition while trying not to break existing
5107 // captions with multiple pipes (|) in it, until a more sensible grammar
5108 // is defined for images in galleries
5110 // FIXME: Doing recursiveTagParse at this stage, and the trim before
5111 // splitting on '|' is a bit odd, and different from makeImage.
5112 $matches[3] = $this->recursiveTagParse( trim( $matches[3] ) );
5113 // Protect LanguageConverter markup
5114 $parameterMatches = StringUtils
::delimiterExplode(
5115 '-{', '}-', '|', $matches[3], true /* nested */
5118 foreach ( $parameterMatches as $parameterMatch ) {
5119 list( $magicName, $match ) = $mwArray->matchVariableStartToEnd( $parameterMatch );
5121 $paramName = $paramMap[$magicName];
5123 switch ( $paramName ) {
5124 case 'gallery-internal-alt':
5125 $alt = $this->stripAltText( $match, false );
5127 case 'gallery-internal-link':
5128 $linkValue = strip_tags( $this->replaceLinkHoldersText( $match ) );
5129 $chars = self
::EXT_LINK_URL_CLASS
;
5130 $addr = self
::EXT_LINK_ADDR
;
5131 $prots = $this->mUrlProtocols
;
5132 // check to see if link matches an absolute url, if not then it must be a wiki link.
5133 if ( preg_match( '/^-{R|(.*)}-$/', $linkValue ) ) {
5134 // Result of LanguageConverter::markNoConversion
5135 // invoked on an external link.
5136 $linkValue = substr( $linkValue, 4, -2 );
5138 if ( preg_match( "/^($prots)$addr$chars*$/u", $linkValue ) ) {
5140 $this->mOutput
->addExternalLink( $link );
5142 $localLinkTitle = Title
::newFromText( $linkValue );
5143 if ( $localLinkTitle !== null ) {
5144 $this->mOutput
->addLink( $localLinkTitle );
5145 $link = $localLinkTitle->getLinkURL();
5150 // Must be a handler specific parameter.
5151 if ( $handler->validateParam( $paramName, $match ) ) {
5152 $handlerOptions[$paramName] = $match;
5154 // Guess not, consider it as caption.
5155 wfDebug( "$parameterMatch failed parameter validation\n" );
5156 $label = $parameterMatch;
5162 $label = $parameterMatch;
5167 $ig->add( $title, $label, $alt, $link, $handlerOptions );
5169 $html = $ig->toHTML();
5170 Hooks
::run( 'AfterParserFetchFileAndTitle', [ $this, $ig, &$html ] );
5175 * @param MediaHandler $handler
5178 public function getImageParams( $handler ) {
5180 $handlerClass = get_class( $handler );
5184 if ( !isset( $this->mImageParams
[$handlerClass] ) ) {
5185 # Initialise static lists
5186 static $internalParamNames = [
5187 'horizAlign' => [ 'left', 'right', 'center', 'none' ],
5188 'vertAlign' => [ 'baseline', 'sub', 'super', 'top', 'text-top', 'middle',
5189 'bottom', 'text-bottom' ],
5190 'frame' => [ 'thumbnail', 'manualthumb', 'framed', 'frameless',
5191 'upright', 'border', 'link', 'alt', 'class' ],
5193 static $internalParamMap;
5194 if ( !$internalParamMap ) {
5195 $internalParamMap = [];
5196 foreach ( $internalParamNames as $type => $names ) {
5197 foreach ( $names as $name ) {
5198 // For grep: img_left, img_right, img_center, img_none,
5199 // img_baseline, img_sub, img_super, img_top, img_text_top, img_middle,
5200 // img_bottom, img_text_bottom,
5201 // img_thumbnail, img_manualthumb, img_framed, img_frameless, img_upright,
5202 // img_border, img_link, img_alt, img_class
5203 $magicName = str_replace( '-', '_', "img_$name" );
5204 $internalParamMap[$magicName] = [ $type, $name ];
5209 # Add handler params
5210 $paramMap = $internalParamMap;
5212 $handlerParamMap = $handler->getParamMap();
5213 foreach ( $handlerParamMap as $magic => $paramName ) {
5214 $paramMap[$magic] = [ 'handler', $paramName ];
5217 $this->mImageParams
[$handlerClass] = $paramMap;
5218 $this->mImageParamsMagicArray
[$handlerClass] =
5219 $this->magicWordFactory
->newArray( array_keys( $paramMap ) );
5221 return [ $this->mImageParams
[$handlerClass], $this->mImageParamsMagicArray
[$handlerClass] ];
5225 * Parse image options text and use it to make an image
5227 * @param Title $title
5228 * @param string $options
5229 * @param LinkHolderArray|bool $holders
5230 * @return string HTML
5232 public function makeImage( $title, $options, $holders = false ) {
5233 # Check if the options text is of the form "options|alt text"
5235 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5236 # * left no resizing, just left align. label is used for alt= only
5237 # * right same, but right aligned
5238 # * none same, but not aligned
5239 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5240 # * center center the image
5241 # * frame Keep original image size, no magnify-button.
5242 # * framed Same as "frame"
5243 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5244 # * upright reduce width for upright images, rounded to full __0 px
5245 # * border draw a 1px border around the image
5246 # * alt Text for HTML alt attribute (defaults to empty)
5247 # * class Set a class for img node
5248 # * link Set the target of the image link. Can be external, interwiki, or local
5249 # vertical-align values (no % or length right now):
5259 # Protect LanguageConverter markup when splitting into parts
5260 $parts = StringUtils
::delimiterExplode(
5261 '-{', '}-', '|', $options, true /* allow nesting */
5264 # Give extensions a chance to select the file revision for us
5267 Hooks
::run( 'BeforeParserFetchFileAndTitle',
5268 [ $this, $title, &$options, &$descQuery ] );
5269 # Fetch and register the file (file title may be different via hooks)
5270 list( $file, $title ) = $this->fetchFileAndTitle( $title, $options );
5273 $handler = $file ?
$file->getHandler() : false;
5275 list( $paramMap, $mwArray ) = $this->getImageParams( $handler );
5278 $this->addTrackingCategory( 'broken-file-category' );
5281 # Process the input parameters
5283 $params = [ 'frame' => [], 'handler' => [],
5284 'horizAlign' => [], 'vertAlign' => [] ];
5285 $seenformat = false;
5286 foreach ( $parts as $part ) {
5287 $part = trim( $part );
5288 list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part );
5290 if ( isset( $paramMap[$magicName] ) ) {
5291 list( $type, $paramName ) = $paramMap[$magicName];
5293 # Special case; width and height come in one variable together
5294 if ( $type === 'handler' && $paramName === 'width' ) {
5295 $parsedWidthParam = self
::parseWidthParam( $value );
5296 if ( isset( $parsedWidthParam['width'] ) ) {
5297 $width = $parsedWidthParam['width'];
5298 if ( $handler->validateParam( 'width', $width ) ) {
5299 $params[$type]['width'] = $width;
5303 if ( isset( $parsedWidthParam['height'] ) ) {
5304 $height = $parsedWidthParam['height'];
5305 if ( $handler->validateParam( 'height', $height ) ) {
5306 $params[$type]['height'] = $height;
5310 # else no validation -- T15436
5312 if ( $type === 'handler' ) {
5313 # Validate handler parameter
5314 $validated = $handler->validateParam( $paramName, $value );
5316 # Validate internal parameters
5317 switch ( $paramName ) {
5321 # @todo FIXME: Possibly check validity here for
5322 # manualthumb? downstream behavior seems odd with
5323 # missing manual thumbs.
5325 $value = $this->stripAltText( $value, $holders );
5328 $chars = self
::EXT_LINK_URL_CLASS
;
5329 $addr = self
::EXT_LINK_ADDR
;
5330 $prots = $this->mUrlProtocols
;
5331 if ( $value === '' ) {
5332 $paramName = 'no-link';
5335 } elseif ( preg_match( "/^((?i)$prots)/", $value ) ) {
5336 if ( preg_match( "/^((?i)$prots)$addr$chars*$/u", $value, $m ) ) {
5337 $paramName = 'link-url';
5338 $this->mOutput
->addExternalLink( $value );
5339 if ( $this->mOptions
->getExternalLinkTarget() ) {
5340 $params[$type]['link-target'] = $this->mOptions
->getExternalLinkTarget();
5345 $linkTitle = Title
::newFromText( $value );
5347 $paramName = 'link-title';
5348 $value = $linkTitle;
5349 $this->mOutput
->addLink( $linkTitle );
5357 // use first appearing option, discard others.
5358 $validated = !$seenformat;
5362 # Most other things appear to be empty or numeric...
5363 $validated = ( $value === false ||
is_numeric( trim( $value ) ) );
5368 $params[$type][$paramName] = $value;
5372 if ( !$validated ) {
5377 # Process alignment parameters
5378 if ( $params['horizAlign'] ) {
5379 $params['frame']['align'] = key( $params['horizAlign'] );
5381 if ( $params['vertAlign'] ) {
5382 $params['frame']['valign'] = key( $params['vertAlign'] );
5385 $params['frame']['caption'] = $caption;
5387 # Will the image be presented in a frame, with the caption below?
5388 $imageIsFramed = isset( $params['frame']['frame'] )
5389 ||
isset( $params['frame']['framed'] )
5390 ||
isset( $params['frame']['thumbnail'] )
5391 ||
isset( $params['frame']['manualthumb'] );
5393 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5394 # came to also set the caption, ordinary text after the image -- which
5395 # makes no sense, because that just repeats the text multiple times in
5396 # screen readers. It *also* came to set the title attribute.
5397 # Now that we have an alt attribute, we should not set the alt text to
5398 # equal the caption: that's worse than useless, it just repeats the
5399 # text. This is the framed/thumbnail case. If there's no caption, we
5400 # use the unnamed parameter for alt text as well, just for the time be-
5401 # ing, if the unnamed param is set and the alt param is not.
5402 # For the future, we need to figure out if we want to tweak this more,
5403 # e.g., introducing a title= parameter for the title; ignoring the un-
5404 # named parameter entirely for images without a caption; adding an ex-
5405 # plicit caption= parameter and preserving the old magic unnamed para-
5407 if ( $imageIsFramed ) { # Framed image
5408 if ( $caption === '' && !isset( $params['frame']['alt'] ) ) {
5409 # No caption or alt text, add the filename as the alt text so
5410 # that screen readers at least get some description of the image
5411 $params['frame']['alt'] = $title->getText();
5413 # Do not set $params['frame']['title'] because tooltips don't make sense
5415 } else { # Inline image
5416 if ( !isset( $params['frame']['alt'] ) ) {
5417 # No alt text, use the "caption" for the alt text
5418 if ( $caption !== '' ) {
5419 $params['frame']['alt'] = $this->stripAltText( $caption, $holders );
5421 # No caption, fall back to using the filename for the
5423 $params['frame']['alt'] = $title->getText();
5426 # Use the "caption" for the tooltip text
5427 $params['frame']['title'] = $this->stripAltText( $caption, $holders );
5430 Hooks
::run( 'ParserMakeImageParams', [ $title, $file, &$params, $this ] );
5432 # Linker does the rest
5433 $time = $options['time'] ??
false;
5434 $ret = Linker
::makeImageLink( $this, $title, $file, $params['frame'], $params['handler'],
5435 $time, $descQuery, $this->mOptions
->getThumbSize() );
5437 # Give the handler a chance to modify the parser object
5439 $handler->parserTransformHook( $this, $file );
5446 * @param string $caption
5447 * @param LinkHolderArray|bool $holders
5448 * @return mixed|string
5450 protected function stripAltText( $caption, $holders ) {
5451 # Strip bad stuff out of the title (tooltip). We can't just use
5452 # replaceLinkHoldersText() here, because if this function is called
5453 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5455 $tooltip = $holders->replaceText( $caption );
5457 $tooltip = $this->replaceLinkHoldersText( $caption );
5460 # make sure there are no placeholders in thumbnail attributes
5461 # that are later expanded to html- so expand them now and
5463 $tooltip = $this->mStripState
->unstripBoth( $tooltip );
5464 $tooltip = Sanitizer
::stripAllTags( $tooltip );
5470 * Set a flag in the output object indicating that the content is dynamic and
5471 * shouldn't be cached.
5472 * @deprecated since 1.28; use getOutput()->updateCacheExpiry()
5474 public function disableCache() {
5475 wfDebug( "Parser output marked as uncacheable.\n" );
5476 if ( !$this->mOutput
) {
5477 throw new MWException( __METHOD__
.
5478 " can only be called when actually parsing something" );
5480 $this->mOutput
->updateCacheExpiry( 0 ); // new style, for consistency
5484 * Callback from the Sanitizer for expanding items found in HTML attribute
5485 * values, so they can be safely tested and escaped.
5487 * @param string &$text
5488 * @param bool|PPFrame $frame
5491 public function attributeStripCallback( &$text, $frame = false ) {
5492 $text = $this->replaceVariables( $text, $frame );
5493 $text = $this->mStripState
->unstripBoth( $text );
5502 public function getTags() {
5503 $this->firstCallInit();
5505 array_keys( $this->mTransparentTagHooks
),
5506 array_keys( $this->mTagHooks
),
5507 array_keys( $this->mFunctionTagHooks
)
5515 public function getFunctionSynonyms() {
5516 $this->firstCallInit();
5517 return $this->mFunctionSynonyms
;
5524 public function getUrlProtocols() {
5525 return $this->mUrlProtocols
;
5529 * Replace transparent tags in $text with the values given by the callbacks.
5531 * Transparent tag hooks are like regular XML-style tag hooks, except they
5532 * operate late in the transformation sequence, on HTML instead of wikitext.
5534 * @param string $text
5538 public function replaceTransparentTags( $text ) {
5540 $elements = array_keys( $this->mTransparentTagHooks
);
5541 $text = self
::extractTagsAndParams( $elements, $text, $matches );
5544 foreach ( $matches as $marker => $data ) {
5545 list( $element, $content, $params, $tag ) = $data;
5546 $tagName = strtolower( $element );
5547 if ( isset( $this->mTransparentTagHooks
[$tagName] ) ) {
5548 $output = call_user_func_array(
5549 $this->mTransparentTagHooks
[$tagName],
5550 [ $content, $params, $this ]
5555 $replacements[$marker] = $output;
5557 return strtr( $text, $replacements );
5561 * Break wikitext input into sections, and either pull or replace
5562 * some particular section's text.
5564 * External callers should use the getSection and replaceSection methods.
5566 * @param string $text Page wikitext
5567 * @param string|int $sectionId A section identifier string of the form:
5568 * "<flag1> - <flag2> - ... - <section number>"
5570 * Currently the only recognised flag is "T", which means the target section number
5571 * was derived during a template inclusion parse, in other words this is a template
5572 * section edit link. If no flags are given, it was an ordinary section edit link.
5573 * This flag is required to avoid a section numbering mismatch when a section is
5574 * enclosed by "<includeonly>" (T8563).
5576 * The section number 0 pulls the text before the first heading; other numbers will
5577 * pull the given section along with its lower-level subsections. If the section is
5578 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5580 * Section 0 is always considered to exist, even if it only contains the empty
5581 * string. If $text is the empty string and section 0 is replaced, $newText is
5584 * @param string $mode One of "get" or "replace"
5585 * @param string $newText Replacement text for section data.
5586 * @return string For "get", the extracted section text.
5587 * for "replace", the whole page with the section replaced.
5589 private function extractSections( $text, $sectionId, $mode, $newText = '' ) {
5590 global $wgTitle; # not generally used but removes an ugly failure mode
5592 $magicScopeVariable = $this->lock();
5593 $this->startParse( $wgTitle, new ParserOptions
, self
::OT_PLAIN
, true );
5595 $frame = $this->getPreprocessor()->newFrame();
5597 # Process section extraction flags
5599 $sectionParts = explode( '-', $sectionId );
5600 $sectionIndex = array_pop( $sectionParts );
5601 foreach ( $sectionParts as $part ) {
5602 if ( $part === 'T' ) {
5603 $flags |
= self
::PTD_FOR_INCLUSION
;
5607 # Check for empty input
5608 if ( strval( $text ) === '' ) {
5609 # Only sections 0 and T-0 exist in an empty document
5610 if ( $sectionIndex == 0 ) {
5611 if ( $mode === 'get' ) {
5617 if ( $mode === 'get' ) {
5625 # Preprocess the text
5626 $root = $this->preprocessToDom( $text, $flags );
5628 # <h> nodes indicate section breaks
5629 # They can only occur at the top level, so we can find them by iterating the root's children
5630 $node = $root->getFirstChild();
5632 # Find the target section
5633 if ( $sectionIndex == 0 ) {
5634 # Section zero doesn't nest, level=big
5635 $targetLevel = 1000;
5638 if ( $node->getName() === 'h' ) {
5639 $bits = $node->splitHeading();
5640 if ( $bits['i'] == $sectionIndex ) {
5641 $targetLevel = $bits['level'];
5645 if ( $mode === 'replace' ) {
5646 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5648 $node = $node->getNextSibling();
5654 if ( $mode === 'get' ) {
5661 # Find the end of the section, including nested sections
5663 if ( $node->getName() === 'h' ) {
5664 $bits = $node->splitHeading();
5665 $curLevel = $bits['level'];
5666 if ( $bits['i'] != $sectionIndex && $curLevel <= $targetLevel ) {
5670 if ( $mode === 'get' ) {
5671 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5673 $node = $node->getNextSibling();
5676 # Write out the remainder (in replace mode only)
5677 if ( $mode === 'replace' ) {
5678 # Output the replacement text
5679 # Add two newlines on -- trailing whitespace in $newText is conventionally
5680 # stripped by the editor, so we need both newlines to restore the paragraph gap
5681 # Only add trailing whitespace if there is newText
5682 if ( $newText != "" ) {
5683 $outText .= $newText . "\n\n";
5687 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5688 $node = $node->getNextSibling();
5692 if ( is_string( $outText ) ) {
5693 # Re-insert stripped tags
5694 $outText = rtrim( $this->mStripState
->unstripBoth( $outText ) );
5701 * This function returns the text of a section, specified by a number ($section).
5702 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5703 * the first section before any such heading (section 0).
5705 * If a section contains subsections, these are also returned.
5707 * @param string $text Text to look in
5708 * @param string|int $sectionId Section identifier as a number or string
5709 * (e.g. 0, 1 or 'T-1').
5710 * @param string $defaultText Default to return if section is not found
5712 * @return string Text of the requested section
5714 public function getSection( $text, $sectionId, $defaultText = '' ) {
5715 return $this->extractSections( $text, $sectionId, 'get', $defaultText );
5719 * This function returns $oldtext after the content of the section
5720 * specified by $section has been replaced with $text. If the target
5721 * section does not exist, $oldtext is returned unchanged.
5723 * @param string $oldText Former text of the article
5724 * @param string|int $sectionId Section identifier as a number or string
5725 * (e.g. 0, 1 or 'T-1').
5726 * @param string $newText Replacing text
5728 * @return string Modified text
5730 public function replaceSection( $oldText, $sectionId, $newText ) {
5731 return $this->extractSections( $oldText, $sectionId, 'replace', $newText );
5735 * Get the ID of the revision we are parsing
5739 public function getRevisionId() {
5740 return $this->mRevisionId
;
5744 * Get the revision object for $this->mRevisionId
5746 * @return Revision|null Either a Revision object or null
5747 * @since 1.23 (public since 1.23)
5749 public function getRevisionObject() {
5750 if ( !is_null( $this->mRevisionObject
) ) {
5751 return $this->mRevisionObject
;
5753 if ( is_null( $this->mRevisionId
) ) {
5757 $rev = call_user_func(
5758 $this->mOptions
->getCurrentRevisionCallback(), $this->getTitle(), $this
5761 # If the parse is for a new revision, then the callback should have
5762 # already been set to force the object and should match mRevisionId.
5763 # If not, try to fetch by mRevisionId for sanity.
5764 if ( $rev && $rev->getId() != $this->mRevisionId
) {
5765 $rev = Revision
::newFromId( $this->mRevisionId
);
5768 $this->mRevisionObject
= $rev;
5770 return $this->mRevisionObject
;
5774 * Get the timestamp associated with the current revision, adjusted for
5775 * the default server-local timestamp
5778 public function getRevisionTimestamp() {
5779 if ( is_null( $this->mRevisionTimestamp
) ) {
5780 $revObject = $this->getRevisionObject();
5781 $timestamp = $revObject ?
$revObject->getTimestamp() : wfTimestampNow();
5783 # The cryptic '' timezone parameter tells to use the site-default
5784 # timezone offset instead of the user settings.
5785 # Since this value will be saved into the parser cache, served
5786 # to other users, and potentially even used inside links and such,
5787 # it needs to be consistent for all visitors.
5788 $this->mRevisionTimestamp
= $this->contLang
->userAdjust( $timestamp, '' );
5790 return $this->mRevisionTimestamp
;
5794 * Get the name of the user that edited the last revision
5796 * @return string User name
5798 public function getRevisionUser() {
5799 if ( is_null( $this->mRevisionUser
) ) {
5800 $revObject = $this->getRevisionObject();
5802 # if this template is subst: the revision id will be blank,
5803 # so just use the current user's name
5805 $this->mRevisionUser
= $revObject->getUserText();
5806 } elseif ( $this->ot
['wiki'] ||
$this->mOptions
->getIsPreview() ) {
5807 $this->mRevisionUser
= $this->getUser()->getName();
5810 return $this->mRevisionUser
;
5814 * Get the size of the revision
5816 * @return int|null Revision size
5818 public function getRevisionSize() {
5819 if ( is_null( $this->mRevisionSize
) ) {
5820 $revObject = $this->getRevisionObject();
5822 # if this variable is subst: the revision id will be blank,
5823 # so just use the parser input size, because the own substituation
5824 # will change the size.
5826 $this->mRevisionSize
= $revObject->getSize();
5828 $this->mRevisionSize
= $this->mInputSize
;
5831 return $this->mRevisionSize
;
5835 * Mutator for $mDefaultSort
5837 * @param string $sort New value
5839 public function setDefaultSort( $sort ) {
5840 $this->mDefaultSort
= $sort;
5841 $this->mOutput
->setProperty( 'defaultsort', $sort );
5845 * Accessor for $mDefaultSort
5846 * Will use the empty string if none is set.
5848 * This value is treated as a prefix, so the
5849 * empty string is equivalent to sorting by
5854 public function getDefaultSort() {
5855 if ( $this->mDefaultSort
!== false ) {
5856 return $this->mDefaultSort
;
5863 * Accessor for $mDefaultSort
5864 * Unlike getDefaultSort(), will return false if none is set
5866 * @return string|bool
5868 public function getCustomDefaultSort() {
5869 return $this->mDefaultSort
;
5872 private static function getSectionNameFromStrippedText( $text ) {
5873 $text = Sanitizer
::normalizeSectionNameWhitespace( $text );
5874 $text = Sanitizer
::decodeCharReferences( $text );
5875 $text = self
::normalizeSectionName( $text );
5879 private static function makeAnchor( $sectionName ) {
5880 return '#' . Sanitizer
::escapeIdForLink( $sectionName );
5883 private static function makeLegacyAnchor( $sectionName ) {
5884 global $wgFragmentMode;
5885 if ( isset( $wgFragmentMode[1] ) && $wgFragmentMode[1] === 'legacy' ) {
5886 // ForAttribute() and ForLink() are the same for legacy encoding
5887 $id = Sanitizer
::escapeIdForAttribute( $sectionName, Sanitizer
::ID_FALLBACK
);
5889 $id = Sanitizer
::escapeIdForLink( $sectionName );
5896 * Try to guess the section anchor name based on a wikitext fragment
5897 * presumably extracted from a heading, for example "Header" from
5900 * @param string $text
5901 * @return string Anchor (starting with '#')
5903 public function guessSectionNameFromWikiText( $text ) {
5904 # Strip out wikitext links(they break the anchor)
5905 $text = $this->stripSectionName( $text );
5906 $sectionName = self
::getSectionNameFromStrippedText( $text );
5907 return self
::makeAnchor( $sectionName );
5911 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5912 * instead, if possible. For use in redirects, since various versions
5913 * of Microsoft browsers interpret Location: headers as something other
5914 * than UTF-8, resulting in breakage.
5916 * @param string $text The section name
5917 * @return string Anchor (starting with '#')
5919 public function guessLegacySectionNameFromWikiText( $text ) {
5920 # Strip out wikitext links(they break the anchor)
5921 $text = $this->stripSectionName( $text );
5922 $sectionName = self
::getSectionNameFromStrippedText( $text );
5923 return self
::makeLegacyAnchor( $sectionName );
5927 * Like guessSectionNameFromWikiText(), but takes already-stripped text as input.
5928 * @param string $text Section name (plain text)
5929 * @return string Anchor (starting with '#')
5931 public static function guessSectionNameFromStrippedText( $text ) {
5932 $sectionName = self
::getSectionNameFromStrippedText( $text );
5933 return self
::makeAnchor( $sectionName );
5937 * Apply the same normalization as code making links to this section would
5939 * @param string $text
5942 private static function normalizeSectionName( $text ) {
5943 # T90902: ensure the same normalization is applied for IDs as to links
5944 $titleParser = MediaWikiServices
::getInstance()->getTitleParser();
5947 $parts = $titleParser->splitTitleString( "#$text" );
5948 } catch ( MalformedTitleException
$ex ) {
5951 return $parts['fragment'];
5955 * Strips a text string of wikitext for use in a section anchor
5957 * Accepts a text string and then removes all wikitext from the
5958 * string and leaves only the resultant text (i.e. the result of
5959 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5960 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5961 * to create valid section anchors by mimicing the output of the
5962 * parser when headings are parsed.
5964 * @param string $text Text string to be stripped of wikitext
5965 * for use in a Section anchor
5966 * @return string Filtered text string
5968 public function stripSectionName( $text ) {
5969 # Strip internal link markup
5970 $text = preg_replace( '/\[\[:?([^[|]+)\|([^[]+)\]\]/', '$2', $text );
5971 $text = preg_replace( '/\[\[:?([^[]+)\|?\]\]/', '$1', $text );
5973 # Strip external link markup
5974 # @todo FIXME: Not tolerant to blank link text
5975 # I.E. [https://www.mediawiki.org] will render as [1] or something depending
5976 # on how many empty links there are on the page - need to figure that out.
5977 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols
. ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5979 # Parse wikitext quotes (italics & bold)
5980 $text = $this->doQuotes( $text );
5983 $text = StringUtils
::delimiterReplace( '<', '>', '', $text );
5988 * strip/replaceVariables/unstrip for preprocessor regression testing
5990 * @param string $text
5991 * @param Title $title
5992 * @param ParserOptions $options
5993 * @param int $outputType
5997 public function testSrvus( $text, Title
$title, ParserOptions
$options,
5998 $outputType = self
::OT_HTML
6000 $magicScopeVariable = $this->lock();
6001 $this->startParse( $title, $options, $outputType, true );
6003 $text = $this->replaceVariables( $text );
6004 $text = $this->mStripState
->unstripBoth( $text );
6005 $text = Sanitizer
::removeHTMLtags( $text );
6010 * @param string $text
6011 * @param Title $title
6012 * @param ParserOptions $options
6015 public function testPst( $text, Title
$title, ParserOptions
$options ) {
6016 return $this->preSaveTransform( $text, $title, $options->getUser(), $options );
6020 * @param string $text
6021 * @param Title $title
6022 * @param ParserOptions $options
6025 public function testPreprocess( $text, Title
$title, ParserOptions
$options ) {
6026 return $this->testSrvus( $text, $title, $options, self
::OT_PREPROCESS
);
6030 * Call a callback function on all regions of the given text that are not
6031 * inside strip markers, and replace those regions with the return value
6032 * of the callback. For example, with input:
6036 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
6037 * two strings will be replaced with the value returned by the callback in
6041 * @param callable $callback
6045 public function markerSkipCallback( $s, $callback ) {
6048 while ( $i < strlen( $s ) ) {
6049 $markerStart = strpos( $s, self
::MARKER_PREFIX
, $i );
6050 if ( $markerStart === false ) {
6051 $out .= call_user_func( $callback, substr( $s, $i ) );
6054 $out .= call_user_func( $callback, substr( $s, $i, $markerStart - $i ) );
6055 $markerEnd = strpos( $s, self
::MARKER_SUFFIX
, $markerStart );
6056 if ( $markerEnd === false ) {
6057 $out .= substr( $s, $markerStart );
6060 $markerEnd +
= strlen( self
::MARKER_SUFFIX
);
6061 $out .= substr( $s, $markerStart, $markerEnd - $markerStart );
6070 * Remove any strip markers found in the given text.
6072 * @param string $text
6075 public function killMarkers( $text ) {
6076 return $this->mStripState
->killMarkers( $text );
6080 * Save the parser state required to convert the given half-parsed text to
6081 * HTML. "Half-parsed" in this context means the output of
6082 * recursiveTagParse() or internalParse(). This output has strip markers
6083 * from replaceVariables (extensionSubstitution() etc.), and link
6084 * placeholders from replaceLinkHolders().
6086 * Returns an array which can be serialized and stored persistently. This
6087 * array can later be loaded into another parser instance with
6088 * unserializeHalfParsedText(). The text can then be safely incorporated into
6089 * the return value of a parser hook.
6091 * @deprecated since 1.31
6092 * @param string $text
6096 public function serializeHalfParsedText( $text ) {
6097 wfDeprecated( __METHOD__
, '1.31' );
6100 'version' => self
::HALF_PARSED_VERSION
,
6101 'stripState' => $this->mStripState
->getSubState( $text ),
6102 'linkHolders' => $this->mLinkHolders
->getSubArray( $text )
6108 * Load the parser state given in the $data array, which is assumed to
6109 * have been generated by serializeHalfParsedText(). The text contents is
6110 * extracted from the array, and its markers are transformed into markers
6111 * appropriate for the current Parser instance. This transformed text is
6112 * returned, and can be safely included in the return value of a parser
6115 * If the $data array has been stored persistently, the caller should first
6116 * check whether it is still valid, by calling isValidHalfParsedText().
6118 * @deprecated since 1.31
6119 * @param array $data Serialized data
6120 * @throws MWException
6123 public function unserializeHalfParsedText( $data ) {
6124 wfDeprecated( __METHOD__
, '1.31' );
6125 if ( !isset( $data['version'] ) ||
$data['version'] != self
::HALF_PARSED_VERSION
) {
6126 throw new MWException( __METHOD__
. ': invalid version' );
6129 # First, extract the strip state.
6130 $texts = [ $data['text'] ];
6131 $texts = $this->mStripState
->merge( $data['stripState'], $texts );
6133 # Now renumber links
6134 $texts = $this->mLinkHolders
->mergeForeign( $data['linkHolders'], $texts );
6136 # Should be good to go.
6141 * Returns true if the given array, presumed to be generated by
6142 * serializeHalfParsedText(), is compatible with the current version of the
6145 * @deprecated since 1.31
6146 * @param array $data
6150 public function isValidHalfParsedText( $data ) {
6151 wfDeprecated( __METHOD__
, '1.31' );
6152 return isset( $data['version'] ) && $data['version'] == self
::HALF_PARSED_VERSION
;
6156 * Parsed a width param of imagelink like 300px or 200x300px
6158 * @param string $value
6159 * @param bool $parseHeight
6164 public static function parseWidthParam( $value, $parseHeight = true ) {
6165 $parsedWidthParam = [];
6166 if ( $value === '' ) {
6167 return $parsedWidthParam;
6170 # (T15500) In both cases (width/height and width only),
6171 # permit trailing "px" for backward compatibility.
6172 if ( $parseHeight && preg_match( '/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/', $value, $m ) ) {
6173 $width = intval( $m[1] );
6174 $height = intval( $m[2] );
6175 $parsedWidthParam['width'] = $width;
6176 $parsedWidthParam['height'] = $height;
6177 } elseif ( preg_match( '/^[0-9]*\s*(?:px)?\s*$/', $value ) ) {
6178 $width = intval( $value );
6179 $parsedWidthParam['width'] = $width;
6181 return $parsedWidthParam;
6185 * Lock the current instance of the parser.
6187 * This is meant to stop someone from calling the parser
6188 * recursively and messing up all the strip state.
6190 * @throws MWException If parser is in a parse
6191 * @return ScopedCallback The lock will be released once the return value goes out of scope.
6193 protected function lock() {
6194 if ( $this->mInParse
) {
6195 throw new MWException( "Parser state cleared while parsing. "
6196 . "Did you call Parser::parse recursively? Lock is held by: " . $this->mInParse
);
6199 // Save the backtrace when locking, so that if some code tries locking again,
6200 // we can print the lock owner's backtrace for easier debugging
6202 $this->mInParse
= $e->getTraceAsString();
6204 $recursiveCheck = new ScopedCallback( function () {
6205 $this->mInParse
= false;
6208 return $recursiveCheck;
6212 * Strip outer <p></p> tag from the HTML source of a single paragraph.
6214 * Returns original HTML if the <p/> tag has any attributes, if there's no wrapping <p/> tag,
6215 * or if there is more than one <p/> tag in the input HTML.
6217 * @param string $html
6221 public static function stripOuterParagraph( $html ) {
6223 if ( preg_match( '/^<p>(.*)\n?<\/p>\n?$/sU', $html, $m ) ) {
6224 if ( strpos( $m[1], '</p>' ) === false ) {
6233 * Return this parser if it is not doing anything, otherwise
6234 * get a fresh parser. You can use this method by doing
6235 * $myParser = $wgParser->getFreshParser(), or more simply
6236 * $wgParser->getFreshParser()->parse( ... );
6237 * if you're unsure if $wgParser is safe to use.
6240 * @return Parser A parser object that is not parsing anything
6242 public function getFreshParser() {
6243 if ( $this->mInParse
) {
6244 return $this->factory
->create();
6251 * Set's up the PHP implementation of OOUI for use in this request
6252 * and instructs OutputPage to enable OOUI for itself.
6256 public function enableOOUI() {
6257 OutputPage
::setupOOUI();
6258 $this->mOutput
->setEnableOOUI( true );