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
;
27 * @defgroup Parser Parser
31 * PHP Parser - Processes wiki markup (which uses a more user-friendly
32 * syntax, such as "[[link]]" for making links), and provides a one-way
33 * transformation of that wiki markup it into (X)HTML output / markup
34 * (which in turn the browser understands, and can display).
36 * There are seven main entry points into the Parser class:
39 * produces HTML output
40 * - Parser::preSaveTransform()
41 * produces altered wiki markup
42 * - Parser::preprocess()
43 * removes HTML comments and expands templates
44 * - Parser::cleanSig() and Parser::cleanSigInSig()
45 * cleans a signature before saving it to preferences
46 * - Parser::getSection()
47 * return the content of a section from an article for section editing
48 * - Parser::replaceSection()
49 * replaces a section by number inside an article
50 * - Parser::getPreloadText()
51 * removes <noinclude> sections and <includeonly> tags
56 * @warning $wgUser or $wgTitle or $wgRequest or $wgLang. Keep them away!
59 * $wgNamespacesWithSubpages
61 * @par Settings only within ParserOptions:
62 * $wgAllowExternalImages
63 * $wgAllowSpecialInclusion
71 * Update this version number when the ParserOutput format
72 * changes in an incompatible way, so the parser cache
73 * can automatically discard old data.
75 const VERSION
= '1.6.4';
78 * Update this version number when the output of serialiseHalfParsedText()
79 * changes in an incompatible way
81 const HALF_PARSED_VERSION
= 2;
83 # Flags for Parser::setFunctionHook
84 const SFH_NO_HASH
= 1;
85 const SFH_OBJECT_ARGS
= 2;
87 # Constants needed for external link processing
88 # Everything except bracket, space, or control characters
89 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
90 # as well as U+3000 is IDEOGRAPHIC SPACE for bug 19052
91 const EXT_LINK_URL_CLASS
= '[^][<>"\\x00-\\x20\\x7F\p{Zs}]';
92 # Simplified expression to match an IPv4 or IPv6 address, or
93 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
94 const EXT_LINK_ADDR
= '(?:[0-9.]+|\\[(?i:[0-9a-f:.]+)\\]|[^][<>"\\x00-\\x20\\x7F\p{Zs}])';
95 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
96 // @codingStandardsIgnoreStart Generic.Files.LineLength
97 const EXT_IMAGE_REGEX
= '/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}]+)
98 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
99 // @codingStandardsIgnoreEnd
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>';
140 public $mTagHooks = [];
141 public $mTransparentTagHooks = [];
142 public $mFunctionHooks = [];
143 public $mFunctionSynonyms = [ 0 => [], 1 => [] ];
144 public $mFunctionTagHooks = [];
145 public $mStripList = [];
146 public $mDefaultStripList = [];
147 public $mVarCache = [];
148 public $mImageParams = [];
149 public $mImageParamsMagicArray = [];
150 public $mMarkerIndex = 0;
151 public $mFirstCall = true;
153 # Initialised by initialiseVariables()
156 * @var MagicWordArray
161 * @var MagicWordArray
164 # Initialised in constructor
165 public $mConf, $mExtLinkBracketedRegex, $mUrlProtocols;
167 # Initialized in getPreprocessor()
168 /** @var Preprocessor */
169 public $mPreprocessor;
171 # Cleared with clearState():
183 public $mIncludeCount;
185 * @var LinkHolderArray
187 public $mLinkHolders;
190 public $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth;
191 public $mDefaultSort;
192 public $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
193 public $mExpensiveFunctionCount; # number of expensive parser function calls
194 public $mShowToc, $mForceTocPosition;
199 public $mUser; # User object; only used when doing pre-save transform
202 # These are variables reset at least once per parse regardless of $clearState
212 public $mTitle; # Title context, used for self-link rendering and similar things
213 public $mOutputType; # Output type, one of the OT_xxx constants
214 public $ot; # Shortcut alias, see setOutputType()
215 public $mRevisionObject; # The revision object of the specified revision ID
216 public $mRevisionId; # ID to display in {{REVISIONID}} tags
217 public $mRevisionTimestamp; # The timestamp of the specified revision ID
218 public $mRevisionUser; # User to display in {{REVISIONUSER}} tag
219 public $mRevisionSize; # Size to display in {{REVISIONSIZE}} variable
220 public $mRevIdForTs; # The revision ID which was used to fetch the timestamp
221 public $mInputSize = false; # For {{PAGESIZE}} on current page.
224 * @var string Deprecated accessor for the strip marker prefix.
225 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
227 public $mUniqPrefix = Parser
::MARKER_PREFIX
;
230 * @var array Array with the language name of each language link (i.e. the
231 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
232 * duplicate language links to the ParserOutput.
234 public $mLangLinkLanguages;
237 * @var MapCacheLRU|null
240 * A cache of the current revisions of titles. Keys are $title->getPrefixedDbKey()
242 public $currentRevisionCache;
245 * @var bool Recursive call protection.
246 * This variable should be treated as if it were private.
248 public $mInParse = false;
250 /** @var SectionProfiler */
251 protected $mProfiler;
256 protected $mLinkRenderer;
261 public function __construct( $conf = [] ) {
262 $this->mConf
= $conf;
263 $this->mUrlProtocols
= wfUrlProtocols();
264 $this->mExtLinkBracketedRegex
= '/\[(((?i)' . $this->mUrlProtocols
. ')' .
265 self
::EXT_LINK_ADDR
.
266 self
::EXT_LINK_URL_CLASS
. '*)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F]*?)\]/Su';
267 if ( isset( $conf['preprocessorClass'] ) ) {
268 $this->mPreprocessorClass
= $conf['preprocessorClass'];
269 } elseif ( defined( 'HPHP_VERSION' ) ) {
270 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
271 $this->mPreprocessorClass
= 'Preprocessor_Hash';
272 } elseif ( extension_loaded( 'domxml' ) ) {
273 # PECL extension that conflicts with the core DOM extension (bug 13770)
274 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
275 $this->mPreprocessorClass
= 'Preprocessor_Hash';
276 } elseif ( extension_loaded( 'dom' ) ) {
277 $this->mPreprocessorClass
= 'Preprocessor_DOM';
279 $this->mPreprocessorClass
= 'Preprocessor_Hash';
281 wfDebug( __CLASS__
. ": using preprocessor: {$this->mPreprocessorClass}\n" );
285 * Reduce memory usage to reduce the impact of circular references
287 public function __destruct() {
288 if ( isset( $this->mLinkHolders
) ) {
289 unset( $this->mLinkHolders
);
291 foreach ( $this as $name => $value ) {
292 unset( $this->$name );
297 * Allow extensions to clean up when the parser is cloned
299 public function __clone() {
300 $this->mInParse
= false;
302 // Bug 56226: When you create a reference "to" an object field, that
303 // makes the object field itself be a reference too (until the other
304 // reference goes out of scope). When cloning, any field that's a
305 // reference is copied as a reference in the new object. Both of these
306 // are defined PHP5 behaviors, as inconvenient as it is for us when old
307 // hooks from PHP4 days are passing fields by reference.
308 foreach ( [ 'mStripState', 'mVarCache' ] as $k ) {
309 // Make a non-reference copy of the field, then rebind the field to
310 // reference the new copy.
316 Hooks
::run( 'ParserCloned', [ $this ] );
320 * Do various kinds of initialisation on the first call of the parser
322 public function firstCallInit() {
323 if ( !$this->mFirstCall
) {
326 $this->mFirstCall
= false;
328 CoreParserFunctions
::register( $this );
329 CoreTagHooks
::register( $this );
330 $this->initialiseVariables();
332 Hooks
::run( 'ParserFirstCallInit', [ &$this ] );
340 public function clearState() {
341 if ( $this->mFirstCall
) {
342 $this->firstCallInit();
344 $this->mOutput
= new ParserOutput
;
345 $this->mOptions
->registerWatcher( [ $this->mOutput
, 'recordOption' ] );
346 $this->mAutonumber
= 0;
347 $this->mIncludeCount
= [];
348 $this->mLinkHolders
= new LinkHolderArray( $this );
350 $this->mRevisionObject
= $this->mRevisionTimestamp
=
351 $this->mRevisionId
= $this->mRevisionUser
= $this->mRevisionSize
= null;
352 $this->mVarCache
= [];
354 $this->mLangLinkLanguages
= [];
355 $this->currentRevisionCache
= null;
357 $this->mStripState
= new StripState
;
359 # Clear these on every parse, bug 4549
360 $this->mTplRedirCache
= $this->mTplDomCache
= [];
362 $this->mShowToc
= true;
363 $this->mForceTocPosition
= false;
364 $this->mIncludeSizes
= [
368 $this->mPPNodeCount
= 0;
369 $this->mGeneratedPPNodeCount
= 0;
370 $this->mHighestExpansionDepth
= 0;
371 $this->mDefaultSort
= false;
372 $this->mHeadings
= [];
373 $this->mDoubleUnderscores
= [];
374 $this->mExpensiveFunctionCount
= 0;
377 if ( isset( $this->mPreprocessor
) && $this->mPreprocessor
->parser
!== $this ) {
378 $this->mPreprocessor
= null;
381 $this->mProfiler
= new SectionProfiler();
383 Hooks
::run( 'ParserClearState', [ &$this ] );
387 * Convert wikitext to HTML
388 * Do not call this function recursively.
390 * @param string $text Text we want to parse
391 * @param Title $title
392 * @param ParserOptions $options
393 * @param bool $linestart
394 * @param bool $clearState
395 * @param int $revid Number to pass in {{REVISIONID}}
396 * @return ParserOutput A ParserOutput
398 public function parse( $text, Title
$title, ParserOptions
$options,
399 $linestart = true, $clearState = true, $revid = null
402 * First pass--just handle <nowiki> sections, pass the rest off
403 * to internalParse() which does all the real work.
406 global $wgShowHostnames;
409 // We use U+007F DELETE to construct strip markers, so we have to make
410 // sure that this character does not occur in the input text.
411 $text = strtr( $text, "\x7f", "?" );
412 $magicScopeVariable = $this->lock();
415 $this->startParse( $title, $options, self
::OT_HTML
, $clearState );
417 $this->currentRevisionCache
= null;
418 $this->mInputSize
= strlen( $text );
419 if ( $this->mOptions
->getEnableLimitReport() ) {
420 $this->mOutput
->resetParseStartTime();
423 $oldRevisionId = $this->mRevisionId
;
424 $oldRevisionObject = $this->mRevisionObject
;
425 $oldRevisionTimestamp = $this->mRevisionTimestamp
;
426 $oldRevisionUser = $this->mRevisionUser
;
427 $oldRevisionSize = $this->mRevisionSize
;
428 if ( $revid !== null ) {
429 $this->mRevisionId
= $revid;
430 $this->mRevisionObject
= null;
431 $this->mRevisionTimestamp
= null;
432 $this->mRevisionUser
= null;
433 $this->mRevisionSize
= null;
436 Hooks
::run( 'ParserBeforeStrip', [ &$this, &$text, &$this->mStripState
] );
438 Hooks
::run( 'ParserAfterStrip', [ &$this, &$text, &$this->mStripState
] );
439 $text = $this->internalParse( $text );
440 Hooks
::run( 'ParserAfterParse', [ &$this, &$text, &$this->mStripState
] );
442 $text = $this->internalParseHalfParsed( $text, true, $linestart );
445 * A converted title will be provided in the output object if title and
446 * content conversion are enabled, the article text does not contain
447 * a conversion-suppressing double-underscore tag, and no
448 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
449 * automatic link conversion.
451 if ( !( $options->getDisableTitleConversion()
452 ||
isset( $this->mDoubleUnderscores
['nocontentconvert'] )
453 ||
isset( $this->mDoubleUnderscores
['notitleconvert'] )
454 ||
$this->mOutput
->getDisplayTitle() !== false )
456 $convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
457 if ( $convruletitle ) {
458 $this->mOutput
->setTitleText( $convruletitle );
460 $titleText = $this->getConverterLanguage()->convertTitle( $title );
461 $this->mOutput
->setTitleText( $titleText );
465 if ( $this->mExpensiveFunctionCount
> $this->mOptions
->getExpensiveParserFunctionLimit() ) {
466 $this->limitationWarn( 'expensive-parserfunction',
467 $this->mExpensiveFunctionCount
,
468 $this->mOptions
->getExpensiveParserFunctionLimit()
472 # Information on include size limits, for the benefit of users who try to skirt them
473 if ( $this->mOptions
->getEnableLimitReport() ) {
474 $max = $this->mOptions
->getMaxIncludeSize();
476 $cpuTime = $this->mOutput
->getTimeSinceStart( 'cpu' );
477 if ( $cpuTime !== null ) {
478 $this->mOutput
->setLimitReportData( 'limitreport-cputime',
479 sprintf( "%.3f", $cpuTime )
483 $wallTime = $this->mOutput
->getTimeSinceStart( 'wall' );
484 $this->mOutput
->setLimitReportData( 'limitreport-walltime',
485 sprintf( "%.3f", $wallTime )
488 $this->mOutput
->setLimitReportData( 'limitreport-ppvisitednodes',
489 [ $this->mPPNodeCount
, $this->mOptions
->getMaxPPNodeCount() ]
491 $this->mOutput
->setLimitReportData( 'limitreport-ppgeneratednodes',
492 [ $this->mGeneratedPPNodeCount
, $this->mOptions
->getMaxGeneratedPPNodeCount() ]
494 $this->mOutput
->setLimitReportData( 'limitreport-postexpandincludesize',
495 [ $this->mIncludeSizes
['post-expand'], $max ]
497 $this->mOutput
->setLimitReportData( 'limitreport-templateargumentsize',
498 [ $this->mIncludeSizes
['arg'], $max ]
500 $this->mOutput
->setLimitReportData( 'limitreport-expansiondepth',
501 [ $this->mHighestExpansionDepth
, $this->mOptions
->getMaxPPExpandDepth() ]
503 $this->mOutput
->setLimitReportData( 'limitreport-expensivefunctioncount',
504 [ $this->mExpensiveFunctionCount
, $this->mOptions
->getExpensiveParserFunctionLimit() ]
506 Hooks
::run( 'ParserLimitReportPrepare', [ $this, $this->mOutput
] );
508 $limitReport = "NewPP limit report\n";
509 if ( $wgShowHostnames ) {
510 $limitReport .= 'Parsed by ' . wfHostname() . "\n";
512 $limitReport .= 'Cached time: ' . $this->mOutput
->getCacheTime() . "\n";
513 $limitReport .= 'Cache expiry: ' . $this->mOutput
->getCacheExpiry() . "\n";
514 $limitReport .= 'Dynamic content: ' .
515 ( $this->mOutput
->hasDynamicContent() ?
'true' : 'false' ) .
518 foreach ( $this->mOutput
->getLimitReportData() as $key => $value ) {
519 if ( Hooks
::run( 'ParserLimitReportFormat',
520 [ $key, &$value, &$limitReport, false, false ]
522 $keyMsg = wfMessage( $key )->inLanguage( 'en' )->useDatabase( false );
523 $valueMsg = wfMessage( [ "$key-value-text", "$key-value" ] )
524 ->inLanguage( 'en' )->useDatabase( false );
525 if ( !$valueMsg->exists() ) {
526 $valueMsg = new RawMessage( '$1' );
528 if ( !$keyMsg->isDisabled() && !$valueMsg->isDisabled() ) {
529 $valueMsg->params( $value );
530 $limitReport .= "{$keyMsg->text()}: {$valueMsg->text()}\n";
534 // Since we're not really outputting HTML, decode the entities and
535 // then re-encode the things that need hiding inside HTML comments.
536 $limitReport = htmlspecialchars_decode( $limitReport );
537 Hooks
::run( 'ParserLimitReport', [ $this, &$limitReport ] );
539 // Sanitize for comment. Note '‐' in the replacement is U+2010,
540 // which looks much like the problematic '-'.
541 $limitReport = str_replace( [ '-', '&' ], [ '‐', '&' ], $limitReport );
542 $text .= "\n<!-- \n$limitReport-->\n";
544 // Add on template profiling data
545 $dataByFunc = $this->mProfiler
->getFunctionStats();
546 uasort( $dataByFunc, function ( $a, $b ) {
547 return $a['real'] < $b['real']; // descending order
549 $profileReport = "Transclusion expansion time report (%,ms,calls,template)\n";
550 foreach ( array_slice( $dataByFunc, 0, 10 ) as $item ) {
551 $profileReport .= sprintf( "%6.2f%% %8.3f %6d - %s\n",
552 $item['%real'], $item['real'], $item['calls'],
553 htmlspecialchars( $item['name'] ) );
555 $text .= "\n<!-- \n$profileReport-->\n";
557 if ( $this->mGeneratedPPNodeCount
> $this->mOptions
->getMaxGeneratedPPNodeCount() / 10 ) {
558 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount
. ' ' .
559 $this->mTitle
->getPrefixedDBkey() );
562 $this->mOutput
->setText( $text );
564 $this->mRevisionId
= $oldRevisionId;
565 $this->mRevisionObject
= $oldRevisionObject;
566 $this->mRevisionTimestamp
= $oldRevisionTimestamp;
567 $this->mRevisionUser
= $oldRevisionUser;
568 $this->mRevisionSize
= $oldRevisionSize;
569 $this->mInputSize
= false;
570 $this->currentRevisionCache
= null;
572 return $this->mOutput
;
576 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
577 * can be called from an extension tag hook.
579 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
580 * instead, which means that lists and links have not been fully parsed yet,
581 * and strip markers are still present.
583 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
585 * Use this function if you're a parser tag hook and you want to parse
586 * wikitext before or after applying additional transformations, and you
587 * intend to *return the result as hook output*, which will cause it to go
588 * through the rest of parsing process automatically.
590 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
591 * $text are not expanded
593 * @param string $text Text extension wants to have parsed
594 * @param bool|PPFrame $frame The frame to use for expanding any template variables
595 * @return string UNSAFE half-parsed HTML
597 public function recursiveTagParse( $text, $frame = false ) {
598 Hooks
::run( 'ParserBeforeStrip', [ &$this, &$text, &$this->mStripState
] );
599 Hooks
::run( 'ParserAfterStrip', [ &$this, &$text, &$this->mStripState
] );
600 $text = $this->internalParse( $text, false, $frame );
605 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
606 * point can be called from an extension tag hook.
608 * The output of this function is fully-parsed HTML that is safe for output.
609 * If you're a parser tag hook, you might want to use recursiveTagParse()
612 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
613 * $text are not expanded
617 * @param string $text Text extension wants to have parsed
618 * @param bool|PPFrame $frame The frame to use for expanding any template variables
619 * @return string Fully parsed HTML
621 public function recursiveTagParseFully( $text, $frame = false ) {
622 $text = $this->recursiveTagParse( $text, $frame );
623 $text = $this->internalParseHalfParsed( $text, false );
628 * Expand templates and variables in the text, producing valid, static wikitext.
629 * Also removes comments.
630 * Do not call this function recursively.
631 * @param string $text
632 * @param Title $title
633 * @param ParserOptions $options
634 * @param int|null $revid
635 * @param bool|PPFrame $frame
636 * @return mixed|string
638 public function preprocess( $text, Title
$title = null,
639 ParserOptions
$options, $revid = null, $frame = false
641 $magicScopeVariable = $this->lock();
642 $this->startParse( $title, $options, self
::OT_PREPROCESS
, true );
643 if ( $revid !== null ) {
644 $this->mRevisionId
= $revid;
646 Hooks
::run( 'ParserBeforeStrip', [ &$this, &$text, &$this->mStripState
] );
647 Hooks
::run( 'ParserAfterStrip', [ &$this, &$text, &$this->mStripState
] );
648 $text = $this->replaceVariables( $text, $frame );
649 $text = $this->mStripState
->unstripBoth( $text );
654 * Recursive parser entry point that can be called from an extension tag
657 * @param string $text Text to be expanded
658 * @param bool|PPFrame $frame The frame to use for expanding any template variables
662 public function recursivePreprocess( $text, $frame = false ) {
663 $text = $this->replaceVariables( $text, $frame );
664 $text = $this->mStripState
->unstripBoth( $text );
669 * Process the wikitext for the "?preload=" feature. (bug 5210)
671 * "<noinclude>", "<includeonly>" etc. are parsed as for template
672 * transclusion, comments, templates, arguments, tags hooks and parser
673 * functions are untouched.
675 * @param string $text
676 * @param Title $title
677 * @param ParserOptions $options
678 * @param array $params
681 public function getPreloadText( $text, Title
$title, ParserOptions
$options, $params = [] ) {
682 $msg = new RawMessage( $text );
683 $text = $msg->params( $params )->plain();
685 # Parser (re)initialisation
686 $magicScopeVariable = $this->lock();
687 $this->startParse( $title, $options, self
::OT_PLAIN
, true );
689 $flags = PPFrame
::NO_ARGS | PPFrame
::NO_TEMPLATES
;
690 $dom = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
691 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
692 $text = $this->mStripState
->unstripBoth( $text );
697 * Get a random string
700 * @deprecated since 1.26; use wfRandomString() instead.
702 public static function getRandomString() {
703 wfDeprecated( __METHOD__
, '1.26' );
704 return wfRandomString( 16 );
708 * Set the current user.
709 * Should only be used when doing pre-save transform.
711 * @param User|null $user User object or null (to reset)
713 public function setUser( $user ) {
714 $this->mUser
= $user;
718 * Accessor for mUniqPrefix.
721 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
723 public function uniqPrefix() {
724 wfDeprecated( __METHOD__
, '1.26' );
725 return self
::MARKER_PREFIX
;
729 * Set the context title
733 public function setTitle( $t ) {
735 $t = Title
::newFromText( 'NO TITLE' );
738 if ( $t->hasFragment() ) {
739 # Strip the fragment to avoid various odd effects
740 $this->mTitle
= $t->createFragmentTarget( '' );
747 * Accessor for the Title object
751 public function getTitle() {
752 return $this->mTitle
;
756 * Accessor/mutator for the Title object
758 * @param Title $x Title object or null to just get the current one
761 public function Title( $x = null ) {
762 return wfSetVar( $this->mTitle
, $x );
766 * Set the output type
768 * @param int $ot New value
770 public function setOutputType( $ot ) {
771 $this->mOutputType
= $ot;
774 'html' => $ot == self
::OT_HTML
,
775 'wiki' => $ot == self
::OT_WIKI
,
776 'pre' => $ot == self
::OT_PREPROCESS
,
777 'plain' => $ot == self
::OT_PLAIN
,
782 * Accessor/mutator for the output type
784 * @param int|null $x New value or null to just get the current one
787 public function OutputType( $x = null ) {
788 return wfSetVar( $this->mOutputType
, $x );
792 * Get the ParserOutput object
794 * @return ParserOutput
796 public function getOutput() {
797 return $this->mOutput
;
801 * Get the ParserOptions object
803 * @return ParserOptions
805 public function getOptions() {
806 return $this->mOptions
;
810 * Accessor/mutator for the ParserOptions object
812 * @param ParserOptions $x New value or null to just get the current one
813 * @return ParserOptions Current ParserOptions object
815 public function Options( $x = null ) {
816 return wfSetVar( $this->mOptions
, $x );
822 public function nextLinkID() {
823 return $this->mLinkID++
;
829 public function setLinkID( $id ) {
830 $this->mLinkID
= $id;
834 * Get a language object for use in parser functions such as {{FORMATNUM:}}
837 public function getFunctionLang() {
838 return $this->getTargetLanguage();
842 * Get the target language for the content being parsed. This is usually the
843 * language that the content is in.
847 * @throws MWException
850 public function getTargetLanguage() {
851 $target = $this->mOptions
->getTargetLanguage();
853 if ( $target !== null ) {
855 } elseif ( $this->mOptions
->getInterfaceMessage() ) {
856 return $this->mOptions
->getUserLangObj();
857 } elseif ( is_null( $this->mTitle
) ) {
858 throw new MWException( __METHOD__
. ': $this->mTitle is null' );
861 return $this->mTitle
->getPageLanguage();
865 * Get the language object for language conversion
866 * @return Language|null
868 public function getConverterLanguage() {
869 return $this->getTargetLanguage();
873 * Get a User object either from $this->mUser, if set, or from the
874 * ParserOptions object otherwise
878 public function getUser() {
879 if ( !is_null( $this->mUser
) ) {
882 return $this->mOptions
->getUser();
886 * Get a preprocessor object
888 * @return Preprocessor
890 public function getPreprocessor() {
891 if ( !isset( $this->mPreprocessor
) ) {
892 $class = $this->mPreprocessorClass
;
893 $this->mPreprocessor
= new $class( $this );
895 return $this->mPreprocessor
;
899 * Get a LinkRenderer instance to make links with
902 * @return LinkRenderer
904 public function getLinkRenderer() {
905 if ( !$this->mLinkRenderer
) {
906 $this->mLinkRenderer
= MediaWikiServices
::getInstance()
907 ->getLinkRendererFactory()->create();
908 $this->mLinkRenderer
->setStubThreshold(
909 $this->getOptions()->getStubThreshold()
913 return $this->mLinkRenderer
;
917 * Replaces all occurrences of HTML-style comments and the given tags
918 * in the text with a random marker and returns the next text. The output
919 * parameter $matches will be an associative array filled with data in
923 * 'UNIQ-xxxxx' => array(
926 * array( 'param' => 'x' ),
927 * '<element param="x">tag content</element>' ) )
930 * @param array $elements List of element names. Comments are always extracted.
931 * @param string $text Source text string.
932 * @param array $matches Out parameter, Array: extracted tags
933 * @param string|null $uniq_prefix
934 * @return string Stripped text
935 * @since 1.26 The uniq_prefix argument is deprecated.
937 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = null ) {
938 if ( $uniq_prefix !== null ) {
939 wfDeprecated( __METHOD__
. ' called with $prefix argument', '1.26' );
945 $taglist = implode( '|', $elements );
946 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
948 while ( $text != '' ) {
949 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE
);
951 if ( count( $p ) < 5 ) {
954 if ( count( $p ) > 5 ) {
968 $marker = self
::MARKER_PREFIX
. "-$element-" . sprintf( '%08X', $n++
) . self
::MARKER_SUFFIX
;
969 $stripped .= $marker;
971 if ( $close === '/>' ) {
972 # Empty element tag, <tag />
977 if ( $element === '!--' ) {
980 $end = "/(<\\/$element\\s*>)/i";
982 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE
);
984 if ( count( $q ) < 3 ) {
985 # No end tag -- let it run out to the end of the text.
994 $matches[$marker] = [ $element,
996 Sanitizer
::decodeTagAttributes( $attributes ),
997 "<$element$attributes$close$content$tail" ];
1003 * Get a list of strippable XML-like elements
1007 public function getStripList() {
1008 return $this->mStripList
;
1012 * Add an item to the strip state
1013 * Returns the unique tag which must be inserted into the stripped text
1014 * The tag will be replaced with the original text in unstrip()
1016 * @param string $text
1020 public function insertStripItem( $text ) {
1021 $marker = self
::MARKER_PREFIX
. "-item-{$this->mMarkerIndex}-" . self
::MARKER_SUFFIX
;
1022 $this->mMarkerIndex++
;
1023 $this->mStripState
->addGeneral( $marker, $text );
1028 * parse the wiki syntax used to render tables
1031 * @param string $text
1034 public function doTableStuff( $text ) {
1036 $lines = StringUtils
::explode( "\n", $text );
1038 $td_history = []; # Is currently a td tag open?
1039 $last_tag_history = []; # Save history of last lag activated (td, th or caption)
1040 $tr_history = []; # Is currently a tr tag open?
1041 $tr_attributes = []; # history of tr attributes
1042 $has_opened_tr = []; # Did this table open a <tr> element?
1043 $indent_level = 0; # indent level of the table
1045 foreach ( $lines as $outLine ) {
1046 $line = trim( $outLine );
1048 if ( $line === '' ) { # empty line, go to next line
1049 $out .= $outLine . "\n";
1053 $first_character = $line[0];
1054 $first_two = substr( $line, 0, 2 );
1057 if ( preg_match( '/^(:*)\s*\{\|(.*)$/', $line, $matches ) ) {
1058 # First check if we are starting a new table
1059 $indent_level = strlen( $matches[1] );
1061 $attributes = $this->mStripState
->unstripBoth( $matches[2] );
1062 $attributes = Sanitizer
::fixTagAttributes( $attributes, 'table' );
1064 $outLine = str_repeat( '<dl><dd>', $indent_level ) . "<table{$attributes}>";
1065 array_push( $td_history, false );
1066 array_push( $last_tag_history, '' );
1067 array_push( $tr_history, false );
1068 array_push( $tr_attributes, '' );
1069 array_push( $has_opened_tr, false );
1070 } elseif ( count( $td_history ) == 0 ) {
1071 # Don't do any of the following
1072 $out .= $outLine . "\n";
1074 } elseif ( $first_two === '|}' ) {
1075 # We are ending a table
1076 $line = '</table>' . substr( $line, 2 );
1077 $last_tag = array_pop( $last_tag_history );
1079 if ( !array_pop( $has_opened_tr ) ) {
1080 $line = "<tr><td></td></tr>{$line}";
1083 if ( array_pop( $tr_history ) ) {
1084 $line = "</tr>{$line}";
1087 if ( array_pop( $td_history ) ) {
1088 $line = "</{$last_tag}>{$line}";
1090 array_pop( $tr_attributes );
1091 $outLine = $line . str_repeat( '</dd></dl>', $indent_level );
1092 } elseif ( $first_two === '|-' ) {
1093 # Now we have a table row
1094 $line = preg_replace( '#^\|-+#', '', $line );
1096 # Whats after the tag is now only attributes
1097 $attributes = $this->mStripState
->unstripBoth( $line );
1098 $attributes = Sanitizer
::fixTagAttributes( $attributes, 'tr' );
1099 array_pop( $tr_attributes );
1100 array_push( $tr_attributes, $attributes );
1103 $last_tag = array_pop( $last_tag_history );
1104 array_pop( $has_opened_tr );
1105 array_push( $has_opened_tr, true );
1107 if ( array_pop( $tr_history ) ) {
1111 if ( array_pop( $td_history ) ) {
1112 $line = "</{$last_tag}>{$line}";
1116 array_push( $tr_history, false );
1117 array_push( $td_history, false );
1118 array_push( $last_tag_history, '' );
1119 } elseif ( $first_character === '|'
1120 ||
$first_character === '!'
1121 ||
$first_two === '|+'
1123 # This might be cell elements, td, th or captions
1124 if ( $first_two === '|+' ) {
1125 $first_character = '+';
1126 $line = substr( $line, 2 );
1128 $line = substr( $line, 1 );
1131 // Implies both are valid for table headings.
1132 if ( $first_character === '!' ) {
1133 $line = StringUtils
::replaceMarkup( '!!', '||', $line );
1136 # Split up multiple cells on the same line.
1137 # FIXME : This can result in improper nesting of tags processed
1138 # by earlier parser steps.
1139 $cells = explode( '||', $line );
1143 # Loop through each table cell
1144 foreach ( $cells as $cell ) {
1146 if ( $first_character !== '+' ) {
1147 $tr_after = array_pop( $tr_attributes );
1148 if ( !array_pop( $tr_history ) ) {
1149 $previous = "<tr{$tr_after}>\n";
1151 array_push( $tr_history, true );
1152 array_push( $tr_attributes, '' );
1153 array_pop( $has_opened_tr );
1154 array_push( $has_opened_tr, true );
1157 $last_tag = array_pop( $last_tag_history );
1159 if ( array_pop( $td_history ) ) {
1160 $previous = "</{$last_tag}>\n{$previous}";
1163 if ( $first_character === '|' ) {
1165 } elseif ( $first_character === '!' ) {
1167 } elseif ( $first_character === '+' ) {
1168 $last_tag = 'caption';
1173 array_push( $last_tag_history, $last_tag );
1175 # A cell could contain both parameters and data
1176 $cell_data = explode( '|', $cell, 2 );
1178 # Bug 553: Note that a '|' inside an invalid link should not
1179 # be mistaken as delimiting cell parameters
1180 if ( strpos( $cell_data[0], '[[' ) !== false ) {
1181 $cell = "{$previous}<{$last_tag}>{$cell}";
1182 } elseif ( count( $cell_data ) == 1 ) {
1183 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
1185 $attributes = $this->mStripState
->unstripBoth( $cell_data[0] );
1186 $attributes = Sanitizer
::fixTagAttributes( $attributes, $last_tag );
1187 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
1191 array_push( $td_history, true );
1194 $out .= $outLine . "\n";
1197 # Closing open td, tr && table
1198 while ( count( $td_history ) > 0 ) {
1199 if ( array_pop( $td_history ) ) {
1202 if ( array_pop( $tr_history ) ) {
1205 if ( !array_pop( $has_opened_tr ) ) {
1206 $out .= "<tr><td></td></tr>\n";
1209 $out .= "</table>\n";
1212 # Remove trailing line-ending (b/c)
1213 if ( substr( $out, -1 ) === "\n" ) {
1214 $out = substr( $out, 0, -1 );
1217 # special case: don't return empty table
1218 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1226 * Helper function for parse() that transforms wiki markup into half-parsed
1227 * HTML. Only called for $mOutputType == self::OT_HTML.
1231 * @param string $text The text to parse
1232 * @param bool $isMain Whether this is being called from the main parse() function
1233 * @param PPFrame|bool $frame A pre-processor frame
1237 public function internalParse( $text, $isMain = true, $frame = false ) {
1241 # Hook to suspend the parser in this state
1242 if ( !Hooks
::run( 'ParserBeforeInternalParse', [ &$this, &$text, &$this->mStripState
] ) ) {
1246 # if $frame is provided, then use $frame for replacing any variables
1248 # use frame depth to infer how include/noinclude tags should be handled
1249 # depth=0 means this is the top-level document; otherwise it's an included document
1250 if ( !$frame->depth
) {
1253 $flag = Parser
::PTD_FOR_INCLUSION
;
1255 $dom = $this->preprocessToDom( $text, $flag );
1256 $text = $frame->expand( $dom );
1258 # if $frame is not provided, then use old-style replaceVariables
1259 $text = $this->replaceVariables( $text );
1262 Hooks
::run( 'InternalParseBeforeSanitize', [ &$this, &$text, &$this->mStripState
] );
1263 $text = Sanitizer
::removeHTMLtags(
1265 [ &$this, 'attributeStripCallback' ],
1267 array_keys( $this->mTransparentTagHooks
)
1269 Hooks
::run( 'InternalParseBeforeLinks', [ &$this, &$text, &$this->mStripState
] );
1271 # Tables need to come after variable replacement for things to work
1272 # properly; putting them before other transformations should keep
1273 # exciting things like link expansions from showing up in surprising
1275 $text = $this->doTableStuff( $text );
1277 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1279 $text = $this->doDoubleUnderscore( $text );
1281 $text = $this->doHeadings( $text );
1282 $text = $this->replaceInternalLinks( $text );
1283 $text = $this->doAllQuotes( $text );
1284 $text = $this->replaceExternalLinks( $text );
1286 # replaceInternalLinks may sometimes leave behind
1287 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1288 $text = str_replace( self
::MARKER_PREFIX
. 'NOPARSE', '', $text );
1290 $text = $this->doMagicLinks( $text );
1291 $text = $this->formatHeadings( $text, $origText, $isMain );
1297 * Helper function for parse() that transforms half-parsed HTML into fully
1300 * @param string $text
1301 * @param bool $isMain
1302 * @param bool $linestart
1305 private function internalParseHalfParsed( $text, $isMain = true, $linestart = true ) {
1306 $text = $this->mStripState
->unstripGeneral( $text );
1309 Hooks
::run( 'ParserAfterUnstrip', [ &$this, &$text ] );
1312 # Clean up special characters, only run once, next-to-last before doBlockLevels
1314 # french spaces, last one Guillemet-left
1315 # only if there is something before the space
1316 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 ',
1317 # french spaces, Guillemet-right
1318 '/(\\302\\253) /' => '\\1 ',
1319 '/ (!\s*important)/' => ' \\1', # Beware of CSS magic word !important, bug #11874.
1321 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
1323 $text = $this->doBlockLevels( $text, $linestart );
1325 $this->replaceLinkHolders( $text );
1328 * The input doesn't get language converted if
1330 * b) Content isn't converted
1331 * c) It's a conversion table
1332 * d) it is an interface message (which is in the user language)
1334 if ( !( $this->mOptions
->getDisableContentConversion()
1335 ||
isset( $this->mDoubleUnderscores
['nocontentconvert'] ) )
1337 if ( !$this->mOptions
->getInterfaceMessage() ) {
1338 # The position of the convert() call should not be changed. it
1339 # assumes that the links are all replaced and the only thing left
1340 # is the <nowiki> mark.
1341 $text = $this->getConverterLanguage()->convert( $text );
1345 $text = $this->mStripState
->unstripNoWiki( $text );
1348 Hooks
::run( 'ParserBeforeTidy', [ &$this, &$text ] );
1351 $text = $this->replaceTransparentTags( $text );
1352 $text = $this->mStripState
->unstripGeneral( $text );
1354 $text = Sanitizer
::normalizeCharReferences( $text );
1356 if ( MWTidy
::isEnabled() && $this->mOptions
->getTidy() ) {
1357 $text = MWTidy
::tidy( $text );
1358 $this->mOutput
->addModuleStyles( MWTidy
::getModuleStyles() );
1360 # attempt to sanitize at least some nesting problems
1361 # (bug #2702 and quite a few others)
1363 # ''Something [http://www.cool.com cool''] -->
1364 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1365 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1366 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1367 # fix up an anchor inside another anchor, only
1368 # at least for a single single nested link (bug 3695)
1369 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1370 '\\1\\2</a>\\3</a>\\1\\4</a>',
1371 # fix div inside inline elements- doBlockLevels won't wrap a line which
1372 # contains a div, so fix it up here; replace
1373 # div with escaped text
1374 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1375 '\\1\\3<div\\5>\\6</div>\\8\\9',
1376 # remove empty italic or bold tag pairs, some
1377 # introduced by rules above
1378 '/<([bi])><\/\\1>/' => '',
1381 $text = preg_replace(
1382 array_keys( $tidyregs ),
1383 array_values( $tidyregs ),
1388 Hooks
::run( 'ParserAfterTidy', [ &$this, &$text ] );
1395 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1396 * magic external links.
1401 * @param string $text
1405 public function doMagicLinks( $text ) {
1406 $prots = wfUrlProtocolsWithoutProtRel();
1407 $urlChar = self
::EXT_LINK_URL_CLASS
;
1408 $addr = self
::EXT_LINK_ADDR
;
1409 $space = self
::SPACE_NOT_NL
; # non-newline space
1410 $spdash = "(?:-|$space)"; # a dash or a non-newline space
1411 $spaces = "$space++"; # possessive match of 1 or more spaces
1412 $text = preg_replace_callback(
1414 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1415 (<.*?>) | # m[2]: Skip stuff inside
1416 # HTML elements' . "
1417 (\b(?i:$prots)($addr$urlChar*)) | # m[3]: Free external links
1418 # m[4]: Post-protocol path
1419 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1421 \bISBN $spaces ( # m[6]: ISBN, capture number
1422 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1423 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1424 [0-9Xx] # check digit
1426 )!xu", [ &$this, 'magicLinkCallback' ], $text );
1431 * @throws MWException
1433 * @return HTML|string
1435 public function magicLinkCallback( $m ) {
1436 if ( isset( $m[1] ) && $m[1] !== '' ) {
1439 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1442 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1443 # Free external link
1444 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1445 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1447 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1450 $cssClass = 'mw-magiclink-rfc';
1452 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1454 $urlmsg = 'pubmedurl';
1455 $cssClass = 'mw-magiclink-pmid';
1458 throw new MWException( __METHOD__
. ': unrecognised match type "' .
1459 substr( $m[0], 0, 20 ) . '"' );
1461 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1462 return Linker
::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle
);
1463 } elseif ( isset( $m[6] ) && $m[6] !== '' ) {
1466 $space = self
::SPACE_NOT_NL
; # non-newline space
1467 $isbn = preg_replace( "/$space/", ' ', $isbn );
1468 $num = strtr( $isbn, [
1473 $titleObj = SpecialPage
::getTitleFor( 'Booksources', $num );
1474 return '<a href="' .
1475 htmlspecialchars( $titleObj->getLocalURL() ) .
1476 "\" class=\"internal mw-magiclink-isbn\">ISBN $isbn</a>";
1483 * Make a free external link, given a user-supplied URL
1485 * @param string $url
1486 * @param int $numPostProto
1487 * The number of characters after the protocol.
1488 * @return string HTML
1491 public function makeFreeExternalLink( $url, $numPostProto ) {
1494 # The characters '<' and '>' (which were escaped by
1495 # removeHTMLtags()) should not be included in
1496 # URLs, per RFC 2396.
1497 # Make terminate a URL as well (bug T84937)
1500 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1505 $trail = substr( $url, $m2[0][1] ) . $trail;
1506 $url = substr( $url, 0, $m2[0][1] );
1509 # Move trailing punctuation to $trail
1511 # If there is no left bracket, then consider right brackets fair game too
1512 if ( strpos( $url, '(' ) === false ) {
1516 $urlRev = strrev( $url );
1517 $numSepChars = strspn( $urlRev, $sep );
1518 # Don't break a trailing HTML entity by moving the ; into $trail
1519 # This is in hot code, so use substr_compare to avoid having to
1520 # create a new string object for the comparison
1521 if ( $numSepChars && substr_compare( $url, ";", -$numSepChars, 1 ) === 0 ) {
1522 # more optimization: instead of running preg_match with a $
1523 # anchor, which can be slow, do the match on the reversed
1524 # string starting at the desired offset.
1525 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1526 if ( preg_match( '/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1530 if ( $numSepChars ) {
1531 $trail = substr( $url, -$numSepChars ) . $trail;
1532 $url = substr( $url, 0, -$numSepChars );
1535 # Verify that we still have a real URL after trail removal, and
1536 # not just lone protocol
1537 if ( strlen( $trail ) >= $numPostProto ) {
1538 return $url . $trail;
1541 $url = Sanitizer
::cleanUrl( $url );
1543 # Is this an external image?
1544 $text = $this->maybeMakeExternalImage( $url );
1545 if ( $text === false ) {
1546 # Not an image, make a link
1547 $text = Linker
::makeExternalLink( $url,
1548 $this->getConverterLanguage()->markNoConversion( $url, true ),
1550 $this->getExternalLinkAttribs( $url ), $this->mTitle
);
1551 # Register it in the output object...
1552 # Replace unnecessary URL escape codes with their equivalent characters
1553 $pasteurized = self
::normalizeLinkUrl( $url );
1554 $this->mOutput
->addExternalLink( $pasteurized );
1556 return $text . $trail;
1560 * Parse headers and return html
1564 * @param string $text
1568 public function doHeadings( $text ) {
1569 for ( $i = 6; $i >= 1; --$i ) {
1570 $h = str_repeat( '=', $i );
1571 $text = preg_replace( "/^$h(.+)$h\\s*$/m", "<h$i>\\1</h$i>", $text );
1577 * Replace single quotes with HTML markup
1580 * @param string $text
1582 * @return string The altered text
1584 public function doAllQuotes( $text ) {
1586 $lines = StringUtils
::explode( "\n", $text );
1587 foreach ( $lines as $line ) {
1588 $outtext .= $this->doQuotes( $line ) . "\n";
1590 $outtext = substr( $outtext, 0, -1 );
1595 * Helper function for doAllQuotes()
1597 * @param string $text
1601 public function doQuotes( $text ) {
1602 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE
);
1603 $countarr = count( $arr );
1604 if ( $countarr == 1 ) {
1608 // First, do some preliminary work. This may shift some apostrophes from
1609 // being mark-up to being text. It also counts the number of occurrences
1610 // of bold and italics mark-ups.
1613 for ( $i = 1; $i < $countarr; $i +
= 2 ) {
1614 $thislen = strlen( $arr[$i] );
1615 // If there are ever four apostrophes, assume the first is supposed to
1616 // be text, and the remaining three constitute mark-up for bold text.
1617 // (bug 13227: ''''foo'''' turns into ' ''' foo ' ''')
1618 if ( $thislen == 4 ) {
1619 $arr[$i - 1] .= "'";
1622 } elseif ( $thislen > 5 ) {
1623 // If there are more than 5 apostrophes in a row, assume they're all
1624 // text except for the last 5.
1625 // (bug 13227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1626 $arr[$i - 1] .= str_repeat( "'", $thislen - 5 );
1630 // Count the number of occurrences of bold and italics mark-ups.
1631 if ( $thislen == 2 ) {
1633 } elseif ( $thislen == 3 ) {
1635 } elseif ( $thislen == 5 ) {
1641 // If there is an odd number of both bold and italics, it is likely
1642 // that one of the bold ones was meant to be an apostrophe followed
1643 // by italics. Which one we cannot know for certain, but it is more
1644 // likely to be one that has a single-letter word before it.
1645 if ( ( $numbold %
2 == 1 ) && ( $numitalics %
2 == 1 ) ) {
1646 $firstsingleletterword = -1;
1647 $firstmultiletterword = -1;
1649 for ( $i = 1; $i < $countarr; $i +
= 2 ) {
1650 if ( strlen( $arr[$i] ) == 3 ) {
1651 $x1 = substr( $arr[$i - 1], -1 );
1652 $x2 = substr( $arr[$i - 1], -2, 1 );
1653 if ( $x1 === ' ' ) {
1654 if ( $firstspace == -1 ) {
1657 } elseif ( $x2 === ' ' ) {
1658 $firstsingleletterword = $i;
1659 // if $firstsingleletterword is set, we don't
1660 // look at the other options, so we can bail early.
1663 if ( $firstmultiletterword == -1 ) {
1664 $firstmultiletterword = $i;
1670 // If there is a single-letter word, use it!
1671 if ( $firstsingleletterword > -1 ) {
1672 $arr[$firstsingleletterword] = "''";
1673 $arr[$firstsingleletterword - 1] .= "'";
1674 } elseif ( $firstmultiletterword > -1 ) {
1675 // If not, but there's a multi-letter word, use that one.
1676 $arr[$firstmultiletterword] = "''";
1677 $arr[$firstmultiletterword - 1] .= "'";
1678 } elseif ( $firstspace > -1 ) {
1679 // ... otherwise use the first one that has neither.
1680 // (notice that it is possible for all three to be -1 if, for example,
1681 // there is only one pentuple-apostrophe in the line)
1682 $arr[$firstspace] = "''";
1683 $arr[$firstspace - 1] .= "'";
1687 // Now let's actually convert our apostrophic mush to HTML!
1692 foreach ( $arr as $r ) {
1693 if ( ( $i %
2 ) == 0 ) {
1694 if ( $state === 'both' ) {
1700 $thislen = strlen( $r );
1701 if ( $thislen == 2 ) {
1702 if ( $state === 'i' ) {
1705 } elseif ( $state === 'bi' ) {
1708 } elseif ( $state === 'ib' ) {
1709 $output .= '</b></i><b>';
1711 } elseif ( $state === 'both' ) {
1712 $output .= '<b><i>' . $buffer . '</i>';
1714 } else { // $state can be 'b' or ''
1718 } elseif ( $thislen == 3 ) {
1719 if ( $state === 'b' ) {
1722 } elseif ( $state === 'bi' ) {
1723 $output .= '</i></b><i>';
1725 } elseif ( $state === 'ib' ) {
1728 } elseif ( $state === 'both' ) {
1729 $output .= '<i><b>' . $buffer . '</b>';
1731 } else { // $state can be 'i' or ''
1735 } elseif ( $thislen == 5 ) {
1736 if ( $state === 'b' ) {
1737 $output .= '</b><i>';
1739 } elseif ( $state === 'i' ) {
1740 $output .= '</i><b>';
1742 } elseif ( $state === 'bi' ) {
1743 $output .= '</i></b>';
1745 } elseif ( $state === 'ib' ) {
1746 $output .= '</b></i>';
1748 } elseif ( $state === 'both' ) {
1749 $output .= '<i><b>' . $buffer . '</b></i>';
1751 } else { // ($state == '')
1759 // Now close all remaining tags. Notice that the order is important.
1760 if ( $state === 'b' ||
$state === 'ib' ) {
1763 if ( $state === 'i' ||
$state === 'bi' ||
$state === 'ib' ) {
1766 if ( $state === 'bi' ) {
1769 // There might be lonely ''''', so make sure we have a buffer
1770 if ( $state === 'both' && $buffer ) {
1771 $output .= '<b><i>' . $buffer . '</i></b>';
1777 * Replace external links (REL)
1779 * Note: this is all very hackish and the order of execution matters a lot.
1780 * Make sure to run tests/parserTests.php if you change this code.
1784 * @param string $text
1786 * @throws MWException
1789 public function replaceExternalLinks( $text ) {
1791 $bits = preg_split( $this->mExtLinkBracketedRegex
, $text, -1, PREG_SPLIT_DELIM_CAPTURE
);
1792 if ( $bits === false ) {
1793 throw new MWException( "PCRE needs to be compiled with "
1794 . "--enable-unicode-properties in order for MediaWiki to function" );
1796 $s = array_shift( $bits );
1799 while ( $i < count( $bits ) ) {
1802 $text = $bits[$i++
];
1803 $trail = $bits[$i++
];
1805 # The characters '<' and '>' (which were escaped by
1806 # removeHTMLtags()) should not be included in
1807 # URLs, per RFC 2396.
1809 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE
) ) {
1810 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1811 $url = substr( $url, 0, $m2[0][1] );
1814 # If the link text is an image URL, replace it with an <img> tag
1815 # This happened by accident in the original parser, but some people used it extensively
1816 $img = $this->maybeMakeExternalImage( $text );
1817 if ( $img !== false ) {
1823 # Set linktype for CSS - if URL==text, link is essentially free
1824 $linktype = ( $text === $url ) ?
'free' : 'text';
1826 # No link text, e.g. [http://domain.tld/some.link]
1827 if ( $text == '' ) {
1829 $langObj = $this->getTargetLanguage();
1830 $text = '[' . $langObj->formatNum( ++
$this->mAutonumber
) . ']';
1831 $linktype = 'autonumber';
1833 # Have link text, e.g. [http://domain.tld/some.link text]s
1835 list( $dtrail, $trail ) = Linker
::splitTrail( $trail );
1838 $text = $this->getConverterLanguage()->markNoConversion( $text );
1840 $url = Sanitizer
::cleanUrl( $url );
1842 # Use the encoded URL
1843 # This means that users can paste URLs directly into the text
1844 # Funny characters like ö aren't valid in URLs anyway
1845 # This was changed in August 2004
1846 $s .= Linker
::makeExternalLink( $url, $text, false, $linktype,
1847 $this->getExternalLinkAttribs( $url ), $this->mTitle
) . $dtrail . $trail;
1849 # Register link in the output object.
1850 # Replace unnecessary URL escape codes with the referenced character
1851 # This prevents spammers from hiding links from the filters
1852 $pasteurized = self
::normalizeLinkUrl( $url );
1853 $this->mOutput
->addExternalLink( $pasteurized );
1860 * Get the rel attribute for a particular external link.
1863 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1864 * nofollow if appropriate
1865 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1866 * @return string|null Rel attribute for $url
1868 public static function getExternalLinkRel( $url = false, $title = null ) {
1869 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1870 $ns = $title ?
$title->getNamespace() : false;
1871 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions )
1872 && !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions )
1880 * Get an associative array of additional HTML attributes appropriate for a
1881 * particular external link. This currently may include rel => nofollow
1882 * (depending on configuration, namespace, and the URL's domain) and/or a
1883 * target attribute (depending on configuration).
1885 * @param string $url URL to extract the domain from for rel =>
1886 * nofollow if appropriate
1887 * @return array Associative array of HTML attributes
1889 public function getExternalLinkAttribs( $url ) {
1891 $rel = self
::getExternalLinkRel( $url, $this->mTitle
);
1893 $target = $this->mOptions
->getExternalLinkTarget();
1895 $attribs['target'] = $target;
1896 if ( !in_array( $target, [ '_self', '_parent', '_top' ] ) ) {
1897 // T133507. New windows can navigate parent cross-origin.
1898 // Including noreferrer due to lacking browser
1899 // support of noopener. Eventually noreferrer should be removed.
1900 if ( $rel !== '' ) {
1903 $rel .= 'noreferrer noopener';
1906 $attribs['rel'] = $rel;
1911 * Replace unusual escape codes in a URL with their equivalent characters
1913 * @deprecated since 1.24, use normalizeLinkUrl
1914 * @param string $url
1917 public static function replaceUnusualEscapes( $url ) {
1918 wfDeprecated( __METHOD__
, '1.24' );
1919 return self
::normalizeLinkUrl( $url );
1923 * Replace unusual escape codes in a URL with their equivalent characters
1925 * This generally follows the syntax defined in RFC 3986, with special
1926 * consideration for HTTP query strings.
1928 * @param string $url
1931 public static function normalizeLinkUrl( $url ) {
1932 # First, make sure unsafe characters are encoded
1933 $url = preg_replace_callback( '/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
1935 return rawurlencode( $m[0] );
1941 $end = strlen( $url );
1943 # Fragment part - 'fragment'
1944 $start = strpos( $url, '#' );
1945 if ( $start !== false && $start < $end ) {
1946 $ret = self
::normalizeUrlComponent(
1947 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}' ) . $ret;
1951 # Query part - 'query' minus &=+;
1952 $start = strpos( $url, '?' );
1953 if ( $start !== false && $start < $end ) {
1954 $ret = self
::normalizeUrlComponent(
1955 substr( $url, $start, $end - $start ), '"#%<>[\]^`{|}&=+;' ) . $ret;
1959 # Scheme and path part - 'pchar'
1960 # (we assume no userinfo or encoded colons in the host)
1961 $ret = self
::normalizeUrlComponent(
1962 substr( $url, 0, $end ), '"#%<>[\]^`{|}/?' ) . $ret;
1967 private static function normalizeUrlComponent( $component, $unsafe ) {
1968 $callback = function ( $matches ) use ( $unsafe ) {
1969 $char = urldecode( $matches[0] );
1970 $ord = ord( $char );
1971 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) === false ) {
1975 # Leave it escaped, but use uppercase for a-f
1976 return strtoupper( $matches[0] );
1979 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', $callback, $component );
1983 * make an image if it's allowed, either through the global
1984 * option, through the exception, or through the on-wiki whitelist
1986 * @param string $url
1990 private function maybeMakeExternalImage( $url ) {
1991 $imagesfrom = $this->mOptions
->getAllowExternalImagesFrom();
1992 $imagesexception = !empty( $imagesfrom );
1994 # $imagesfrom could be either a single string or an array of strings, parse out the latter
1995 if ( $imagesexception && is_array( $imagesfrom ) ) {
1996 $imagematch = false;
1997 foreach ( $imagesfrom as $match ) {
1998 if ( strpos( $url, $match ) === 0 ) {
2003 } elseif ( $imagesexception ) {
2004 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2006 $imagematch = false;
2009 if ( $this->mOptions
->getAllowExternalImages()
2010 ||
( $imagesexception && $imagematch )
2012 if ( preg_match( self
::EXT_IMAGE_REGEX
, $url ) ) {
2014 $text = Linker
::makeExternalImage( $url );
2017 if ( !$text && $this->mOptions
->getEnableImageWhitelist()
2018 && preg_match( self
::EXT_IMAGE_REGEX
, $url )
2020 $whitelist = explode(
2022 wfMessage( 'external_image_whitelist' )->inContentLanguage()->text()
2025 foreach ( $whitelist as $entry ) {
2026 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2027 if ( strpos( $entry, '#' ) === 0 ||
$entry === '' ) {
2030 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
2031 # Image matches a whitelist entry
2032 $text = Linker
::makeExternalImage( $url );
2041 * Process [[ ]] wikilinks
2045 * @return string Processed text
2049 public function replaceInternalLinks( $s ) {
2050 $this->mLinkHolders
->merge( $this->replaceInternalLinks2( $s ) );
2055 * Process [[ ]] wikilinks (RIL)
2057 * @throws MWException
2058 * @return LinkHolderArray
2062 public function replaceInternalLinks2( &$s ) {
2063 global $wgExtraInterlanguageLinkPrefixes;
2065 static $tc = false, $e1, $e1_img;
2066 # the % is needed to support urlencoded titles as well
2068 $tc = Title
::legalChars() . '#%';
2069 # Match a link having the form [[namespace:link|alternate]]trail
2070 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2071 # Match cases where there is no "]]", which might still be images
2072 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
2075 $holders = new LinkHolderArray( $this );
2077 # split the entire text string on occurrences of [[
2078 $a = StringUtils
::explode( '[[', ' ' . $s );
2079 # get the first element (all text up to first [[), and remove the space we added
2082 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
2083 $s = substr( $s, 1 );
2085 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2087 if ( $useLinkPrefixExtension ) {
2088 # Match the end of a line for a word that's not followed by whitespace,
2089 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2091 $charset = $wgContLang->linkPrefixCharset();
2092 $e2 = "/^((?>.*[^$charset]|))(.+)$/sDu";
2095 if ( is_null( $this->mTitle
) ) {
2096 throw new MWException( __METHOD__
. ": \$this->mTitle is null\n" );
2098 $nottalk = !$this->mTitle
->isTalkPage();
2100 if ( $useLinkPrefixExtension ) {
2102 if ( preg_match( $e2, $s, $m ) ) {
2103 $first_prefix = $m[2];
2105 $first_prefix = false;
2111 $useSubpages = $this->areSubpagesAllowed();
2113 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2114 # Loop for each link
2115 for ( ; $line !== false && $line !== null; $a->next(), $line = $a->current() ) {
2116 // @codingStandardsIgnoreEnd
2118 # Check for excessive memory usage
2119 if ( $holders->isBig() ) {
2121 # Do the existence check, replace the link holders and clear the array
2122 $holders->replace( $s );
2126 if ( $useLinkPrefixExtension ) {
2127 if ( preg_match( $e2, $s, $m ) ) {
2134 if ( $first_prefix ) {
2135 $prefix = $first_prefix;
2136 $first_prefix = false;
2140 $might_be_img = false;
2142 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
2144 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2145 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2146 # the real problem is with the $e1 regex
2148 # Still some problems for cases where the ] is meant to be outside punctuation,
2149 # and no image is in sight. See bug 2095.
2151 && substr( $m[3], 0, 1 ) === ']'
2152 && strpos( $text, '[' ) !== false
2154 $text .= ']'; # so that replaceExternalLinks($text) works later
2155 $m[3] = substr( $m[3], 1 );
2157 # fix up urlencoded title texts
2158 if ( strpos( $m[1], '%' ) !== false ) {
2159 # Should anchors '#' also be rejected?
2160 $m[1] = str_replace( [ '<', '>' ], [ '<', '>' ], rawurldecode( $m[1] ) );
2163 } elseif ( preg_match( $e1_img, $line, $m ) ) {
2164 # Invalid, but might be an image with a link in its caption
2165 $might_be_img = true;
2167 if ( strpos( $m[1], '%' ) !== false ) {
2168 $m[1] = rawurldecode( $m[1] );
2171 } else { # Invalid form; output directly
2172 $s .= $prefix . '[[' . $line;
2178 # Don't allow internal links to pages containing
2179 # PROTO: where PROTO is a valid URL protocol; these
2180 # should be external links.
2181 if ( preg_match( '/^(?i:' . $this->mUrlProtocols
. ')/', $origLink ) ) {
2182 $s .= $prefix . '[[' . $line;
2186 # Make subpage if necessary
2187 if ( $useSubpages ) {
2188 $link = $this->maybeDoSubpageLink( $origLink, $text );
2193 $noforce = ( substr( $origLink, 0, 1 ) !== ':' );
2195 # Strip off leading ':'
2196 $link = substr( $link, 1 );
2199 $unstrip = $this->mStripState
->unstripNoWiki( $link );
2200 $nt = is_string( $unstrip ) ? Title
::newFromText( $unstrip ) : null;
2201 if ( $nt === null ) {
2202 $s .= $prefix . '[[' . $line;
2206 $ns = $nt->getNamespace();
2207 $iw = $nt->getInterwiki();
2209 if ( $might_be_img ) { # if this is actually an invalid link
2210 if ( $ns == NS_FILE
&& $noforce ) { # but might be an image
2213 # look at the next 'line' to see if we can close it there
2215 $next_line = $a->current();
2216 if ( $next_line === false ||
$next_line === null ) {
2219 $m = explode( ']]', $next_line, 3 );
2220 if ( count( $m ) == 3 ) {
2221 # the first ]] closes the inner link, the second the image
2223 $text .= "[[{$m[0]}]]{$m[1]}";
2226 } elseif ( count( $m ) == 2 ) {
2227 # if there's exactly one ]] that's fine, we'll keep looking
2228 $text .= "[[{$m[0]}]]{$m[1]}";
2230 # if $next_line is invalid too, we need look no further
2231 $text .= '[[' . $next_line;
2236 # we couldn't find the end of this imageLink, so output it raw
2237 # but don't ignore what might be perfectly normal links in the text we've examined
2238 $holders->merge( $this->replaceInternalLinks2( $text ) );
2239 $s .= "{$prefix}[[$link|$text";
2240 # note: no $trail, because without an end, there *is* no trail
2243 } else { # it's not an image, so output it raw
2244 $s .= "{$prefix}[[$link|$text";
2245 # note: no $trail, because without an end, there *is* no trail
2250 $wasblank = ( $text == '' );
2254 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
2255 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2256 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2257 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2258 $text = $this->doQuotes( $text );
2261 # Link not escaped by : , create the various objects
2262 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2265 $iw && $this->mOptions
->getInterwikiMagic() && $nottalk && (
2266 Language
::fetchLanguageName( $iw, null, 'mw' ) ||
2267 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2270 # Bug 24502: filter duplicates
2271 if ( !isset( $this->mLangLinkLanguages
[$iw] ) ) {
2272 $this->mLangLinkLanguages
[$iw] = true;
2273 $this->mOutput
->addLanguageLink( $nt->getFullText() );
2276 $s = rtrim( $s . $prefix );
2277 $s .= trim( $trail, "\n" ) == '' ?
'': $prefix . $trail;
2281 if ( $ns == NS_FILE
) {
2282 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle
) ) {
2284 # if no parameters were passed, $text
2285 # becomes something like "File:Foo.png",
2286 # which we don't want to pass on to the
2290 # recursively parse links inside the image caption
2291 # actually, this will parse them in any other parameters, too,
2292 # but it might be hard to fix that, and it doesn't matter ATM
2293 $text = $this->replaceExternalLinks( $text );
2294 $holders->merge( $this->replaceInternalLinks2( $text ) );
2296 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2297 $s .= $prefix . $this->armorLinks(
2298 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2300 $s .= $prefix . $trail;
2305 if ( $ns == NS_CATEGORY
) {
2306 $s = rtrim( $s . "\n" ); # bug 87
2309 $sortkey = $this->getDefaultSort();
2313 $sortkey = Sanitizer
::decodeCharReferences( $sortkey );
2314 $sortkey = str_replace( "\n", '', $sortkey );
2315 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
2316 $this->mOutput
->addCategory( $nt->getDBkey(), $sortkey );
2319 * Strip the whitespace Category links produce, see bug 87
2321 $s .= trim( $prefix . $trail, "\n" ) == '' ?
'' : $prefix . $trail;
2327 # Self-link checking. For some languages, variants of the title are checked in
2328 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2329 # for linking to a different variant.
2330 if ( $ns != NS_SPECIAL
&& $nt->equals( $this->mTitle
) && !$nt->hasFragment() ) {
2331 $s .= $prefix . Linker
::makeSelfLinkObj( $nt, $text, '', $trail );
2335 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2336 # @todo FIXME: Should do batch file existence checks, see comment below
2337 if ( $ns == NS_MEDIA
) {
2338 # Give extensions a chance to select the file revision for us
2341 Hooks
::run( 'BeforeParserFetchFileAndTitle',
2342 [ $this, $nt, &$options, &$descQuery ] );
2343 # Fetch and register the file (file title may be different via hooks)
2344 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2345 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2346 $s .= $prefix . $this->armorLinks(
2347 Linker
::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2351 # Some titles, such as valid special pages or files in foreign repos, should
2352 # be shown as bluelinks even though they're not included in the page table
2353 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2354 # batch file existence checks for NS_FILE and NS_MEDIA
2355 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2356 $this->mOutput
->addLink( $nt );
2357 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2359 # Links will be added to the output link list after checking
2360 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2367 * Render a forced-blue link inline; protect against double expansion of
2368 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2369 * Since this little disaster has to split off the trail text to avoid
2370 * breaking URLs in the following text without breaking trails on the
2371 * wiki links, it's been made into a horrible function.
2374 * @param string $text
2375 * @param string $trail
2376 * @param string $prefix
2377 * @return string HTML-wikitext mix oh yuck
2379 protected function makeKnownLinkHolder( $nt, $text = '', $trail = '', $prefix = '' ) {
2380 list( $inside, $trail ) = Linker
::splitTrail( $trail );
2382 if ( $text == '' ) {
2383 $text = htmlspecialchars( $nt->getPrefixedText() );
2386 $link = $this->getLinkRenderer()->makeKnownLink(
2387 $nt, new HtmlArmor( "$prefix$text$inside" )
2390 return $this->armorLinks( $link ) . $trail;
2394 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2395 * going to go through further parsing steps before inline URL expansion.
2397 * Not needed quite as much as it used to be since free links are a bit
2398 * more sensible these days. But bracketed links are still an issue.
2400 * @param string $text More-or-less HTML
2401 * @return string Less-or-more HTML with NOPARSE bits
2403 public function armorLinks( $text ) {
2404 return preg_replace( '/\b((?i)' . $this->mUrlProtocols
. ')/',
2405 self
::MARKER_PREFIX
. "NOPARSE$1", $text );
2409 * Return true if subpage links should be expanded on this page.
2412 public function areSubpagesAllowed() {
2413 # Some namespaces don't allow subpages
2414 return MWNamespace
::hasSubpages( $this->mTitle
->getNamespace() );
2418 * Handle link to subpage if necessary
2420 * @param string $target The source of the link
2421 * @param string &$text The link text, modified as necessary
2422 * @return string The full name of the link
2425 public function maybeDoSubpageLink( $target, &$text ) {
2426 return Linker
::normalizeSubpageLink( $this->mTitle
, $target, $text );
2430 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2432 * @param string $text
2433 * @param bool $linestart Whether or not this is at the start of a line.
2435 * @return string The lists rendered as HTML
2437 public function doBlockLevels( $text, $linestart ) {
2438 return BlockLevelPass
::doBlockLevels( $text, $linestart );
2442 * Return value of a magic variable (like PAGENAME)
2447 * @param bool|PPFrame $frame
2449 * @throws MWException
2452 public function getVariableValue( $index, $frame = false ) {
2453 global $wgContLang, $wgSitename, $wgServer, $wgServerName;
2454 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2456 if ( is_null( $this->mTitle
) ) {
2457 // If no title set, bad things are going to happen
2458 // later. Title should always be set since this
2459 // should only be called in the middle of a parse
2460 // operation (but the unit-tests do funky stuff)
2461 throw new MWException( __METHOD__
. ' Should only be '
2462 . ' called while parsing (no title set)' );
2466 * Some of these require message or data lookups and can be
2467 * expensive to check many times.
2469 if ( Hooks
::run( 'ParserGetVariableValueVarCache', [ &$this, &$this->mVarCache
] ) ) {
2470 if ( isset( $this->mVarCache
[$index] ) ) {
2471 return $this->mVarCache
[$index];
2475 $ts = wfTimestamp( TS_UNIX
, $this->mOptions
->getTimestamp() );
2476 Hooks
::run( 'ParserGetVariableValueTs', [ &$this, &$ts ] );
2478 $pageLang = $this->getFunctionLang();
2484 case 'currentmonth':
2485 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'm' ) );
2487 case 'currentmonth1':
2488 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2490 case 'currentmonthname':
2491 $value = $pageLang->getMonthName( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2493 case 'currentmonthnamegen':
2494 $value = $pageLang->getMonthNameGen( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2496 case 'currentmonthabbrev':
2497 $value = $pageLang->getMonthAbbreviation( MWTimestamp
::getInstance( $ts )->format( 'n' ) );
2500 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'j' ) );
2503 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'd' ) );
2506 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'm' ) );
2509 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2511 case 'localmonthname':
2512 $value = $pageLang->getMonthName( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2514 case 'localmonthnamegen':
2515 $value = $pageLang->getMonthNameGen( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2517 case 'localmonthabbrev':
2518 $value = $pageLang->getMonthAbbreviation( MWTimestamp
::getLocalInstance( $ts )->format( 'n' ) );
2521 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'j' ) );
2524 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'd' ) );
2527 $value = wfEscapeWikiText( $this->mTitle
->getText() );
2530 $value = wfEscapeWikiText( $this->mTitle
->getPartialURL() );
2532 case 'fullpagename':
2533 $value = wfEscapeWikiText( $this->mTitle
->getPrefixedText() );
2535 case 'fullpagenamee':
2536 $value = wfEscapeWikiText( $this->mTitle
->getPrefixedURL() );
2539 $value = wfEscapeWikiText( $this->mTitle
->getSubpageText() );
2541 case 'subpagenamee':
2542 $value = wfEscapeWikiText( $this->mTitle
->getSubpageUrlForm() );
2544 case 'rootpagename':
2545 $value = wfEscapeWikiText( $this->mTitle
->getRootText() );
2547 case 'rootpagenamee':
2548 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2551 $this->mTitle
->getRootText()
2554 case 'basepagename':
2555 $value = wfEscapeWikiText( $this->mTitle
->getBaseText() );
2557 case 'basepagenamee':
2558 $value = wfEscapeWikiText( wfUrlencode( str_replace(
2561 $this->mTitle
->getBaseText()
2564 case 'talkpagename':
2565 if ( $this->mTitle
->canTalk() ) {
2566 $talkPage = $this->mTitle
->getTalkPage();
2567 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2572 case 'talkpagenamee':
2573 if ( $this->mTitle
->canTalk() ) {
2574 $talkPage = $this->mTitle
->getTalkPage();
2575 $value = wfEscapeWikiText( $talkPage->getPrefixedURL() );
2580 case 'subjectpagename':
2581 $subjPage = $this->mTitle
->getSubjectPage();
2582 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2584 case 'subjectpagenamee':
2585 $subjPage = $this->mTitle
->getSubjectPage();
2586 $value = wfEscapeWikiText( $subjPage->getPrefixedURL() );
2588 case 'pageid': // requested in bug 23427
2589 $pageid = $this->getTitle()->getArticleID();
2590 if ( $pageid == 0 ) {
2591 # 0 means the page doesn't exist in the database,
2592 # which means the user is previewing a new page.
2593 # The vary-revision flag must be set, because the magic word
2594 # will have a different value once the page is saved.
2595 $this->mOutput
->setFlag( 'vary-revision' );
2596 wfDebug( __METHOD__
. ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2598 $value = $pageid ?
$pageid : null;
2601 # Let the edit saving system know we should parse the page
2602 # *after* a revision ID has been assigned.
2603 $this->mOutput
->setFlag( 'vary-revision' );
2604 wfDebug( __METHOD__
. ": {{REVISIONID}} used, setting vary-revision...\n" );
2605 $value = $this->mRevisionId
;
2608 # Let the edit saving system know we should parse the page
2609 # *after* a revision ID has been assigned. This is for null edits.
2610 $this->mOutput
->setFlag( 'vary-revision' );
2611 wfDebug( __METHOD__
. ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2612 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2614 case 'revisionday2':
2615 # Let the edit saving system know we should parse the page
2616 # *after* a revision ID has been assigned. This is for null edits.
2617 $this->mOutput
->setFlag( 'vary-revision' );
2618 wfDebug( __METHOD__
. ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2619 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2621 case 'revisionmonth':
2622 # Let the edit saving system know we should parse the page
2623 # *after* a revision ID has been assigned. This is for null edits.
2624 $this->mOutput
->setFlag( 'vary-revision' );
2625 wfDebug( __METHOD__
. ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2626 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2628 case 'revisionmonth1':
2629 # Let the edit saving system know we should parse the page
2630 # *after* a revision ID has been assigned. This is for null edits.
2631 $this->mOutput
->setFlag( 'vary-revision' );
2632 wfDebug( __METHOD__
. ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2633 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2635 case 'revisionyear':
2636 # Let the edit saving system know we should parse the page
2637 # *after* a revision ID has been assigned. This is for null edits.
2638 $this->mOutput
->setFlag( 'vary-revision' );
2639 wfDebug( __METHOD__
. ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2640 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2642 case 'revisiontimestamp':
2643 # Let the edit saving system know we should parse the page
2644 # *after* a revision ID has been assigned. This is for null edits.
2645 $this->mOutput
->setFlag( 'vary-revision' );
2646 wfDebug( __METHOD__
. ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2647 $value = $this->getRevisionTimestamp();
2649 case 'revisionuser':
2650 # Let the edit saving system know we should parse the page
2651 # *after* a revision ID has been assigned. This is for null edits.
2652 $this->mOutput
->setFlag( 'vary-revision' );
2653 wfDebug( __METHOD__
. ": {{REVISIONUSER}} used, setting vary-revision...\n" );
2654 $value = $this->getRevisionUser();
2656 case 'revisionsize':
2657 # Let the edit saving system know we should parse the page
2658 # *after* a revision ID has been assigned. This is for null edits.
2659 $value = $this->getRevisionSize();
2662 $value = str_replace( '_', ' ', $wgContLang->getNsText( $this->mTitle
->getNamespace() ) );
2665 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle
->getNamespace() ) );
2667 case 'namespacenumber':
2668 $value = $this->mTitle
->getNamespace();
2671 $value = $this->mTitle
->canTalk()
2672 ?
str_replace( '_', ' ', $this->mTitle
->getTalkNsText() )
2676 $value = $this->mTitle
->canTalk() ?
wfUrlencode( $this->mTitle
->getTalkNsText() ) : '';
2678 case 'subjectspace':
2679 $value = str_replace( '_', ' ', $this->mTitle
->getSubjectNsText() );
2681 case 'subjectspacee':
2682 $value = ( wfUrlencode( $this->mTitle
->getSubjectNsText() ) );
2684 case 'currentdayname':
2685 $value = $pageLang->getWeekdayName( (int)MWTimestamp
::getInstance( $ts )->format( 'w' ) +
1 );
2688 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'Y' ), true );
2691 $value = $pageLang->time( wfTimestamp( TS_MW
, $ts ), false, false );
2694 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'H' ), true );
2697 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2698 # int to remove the padding
2699 $value = $pageLang->formatNum( (int)MWTimestamp
::getInstance( $ts )->format( 'W' ) );
2702 $value = $pageLang->formatNum( MWTimestamp
::getInstance( $ts )->format( 'w' ) );
2704 case 'localdayname':
2705 $value = $pageLang->getWeekdayName(
2706 (int)MWTimestamp
::getLocalInstance( $ts )->format( 'w' ) +
1
2710 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'Y' ), true );
2713 $value = $pageLang->time(
2714 MWTimestamp
::getLocalInstance( $ts )->format( 'YmdHis' ),
2720 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'H' ), true );
2723 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2724 # int to remove the padding
2725 $value = $pageLang->formatNum( (int)MWTimestamp
::getLocalInstance( $ts )->format( 'W' ) );
2728 $value = $pageLang->formatNum( MWTimestamp
::getLocalInstance( $ts )->format( 'w' ) );
2730 case 'numberofarticles':
2731 $value = $pageLang->formatNum( SiteStats
::articles() );
2733 case 'numberoffiles':
2734 $value = $pageLang->formatNum( SiteStats
::images() );
2736 case 'numberofusers':
2737 $value = $pageLang->formatNum( SiteStats
::users() );
2739 case 'numberofactiveusers':
2740 $value = $pageLang->formatNum( SiteStats
::activeUsers() );
2742 case 'numberofpages':
2743 $value = $pageLang->formatNum( SiteStats
::pages() );
2745 case 'numberofadmins':
2746 $value = $pageLang->formatNum( SiteStats
::numberingroup( 'sysop' ) );
2748 case 'numberofedits':
2749 $value = $pageLang->formatNum( SiteStats
::edits() );
2751 case 'currenttimestamp':
2752 $value = wfTimestamp( TS_MW
, $ts );
2754 case 'localtimestamp':
2755 $value = MWTimestamp
::getLocalInstance( $ts )->format( 'YmdHis' );
2757 case 'currentversion':
2758 $value = SpecialVersion
::getVersion();
2761 return $wgArticlePath;
2767 return $wgServerName;
2769 return $wgScriptPath;
2771 return $wgStylePath;
2772 case 'directionmark':
2773 return $pageLang->getDirMark();
2774 case 'contentlanguage':
2775 global $wgLanguageCode;
2776 return $wgLanguageCode;
2777 case 'cascadingsources':
2778 $value = CoreParserFunctions
::cascadingsources( $this );
2783 'ParserGetVariableValueSwitch',
2784 [ &$this, &$this->mVarCache
, &$index, &$ret, &$frame ]
2791 $this->mVarCache
[$index] = $value;
2798 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2802 public function initialiseVariables() {
2803 $variableIDs = MagicWord
::getVariableIDs();
2804 $substIDs = MagicWord
::getSubstIDs();
2806 $this->mVariables
= new MagicWordArray( $variableIDs );
2807 $this->mSubstWords
= new MagicWordArray( $substIDs );
2811 * Preprocess some wikitext and return the document tree.
2812 * This is the ghost of replace_variables().
2814 * @param string $text The text to parse
2815 * @param int $flags Bitwise combination of:
2816 * - self::PTD_FOR_INCLUSION: Handle "<noinclude>" and "<includeonly>" as if the text is being
2817 * included. Default is to assume a direct page view.
2819 * The generated DOM tree must depend only on the input text and the flags.
2820 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
2822 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2823 * change in the DOM tree for a given text, must be passed through the section identifier
2824 * in the section edit link and thus back to extractSections().
2826 * The output of this function is currently only cached in process memory, but a persistent
2827 * cache may be implemented at a later date which takes further advantage of these strict
2828 * dependency requirements.
2832 public function preprocessToDom( $text, $flags = 0 ) {
2833 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
2838 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2844 public static function splitWhitespace( $s ) {
2845 $ltrimmed = ltrim( $s );
2846 $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) );
2847 $trimmed = rtrim( $ltrimmed );
2848 $diff = strlen( $ltrimmed ) - strlen( $trimmed );
2850 $w2 = substr( $ltrimmed, -$diff );
2854 return [ $w1, $trimmed, $w2 ];
2858 * Replace magic variables, templates, and template arguments
2859 * with the appropriate text. Templates are substituted recursively,
2860 * taking care to avoid infinite loops.
2862 * Note that the substitution depends on value of $mOutputType:
2863 * self::OT_WIKI: only {{subst:}} templates
2864 * self::OT_PREPROCESS: templates but not extension tags
2865 * self::OT_HTML: all templates and extension tags
2867 * @param string $text The text to transform
2868 * @param bool|PPFrame $frame Object describing the arguments passed to the
2869 * template. Arguments may also be provided as an associative array, as
2870 * was the usual case before MW1.12. Providing arguments this way may be
2871 * useful for extensions wishing to perform variable replacement
2873 * @param bool $argsOnly Only do argument (triple-brace) expansion, not
2874 * double-brace expansion.
2877 public function replaceVariables( $text, $frame = false, $argsOnly = false ) {
2878 # Is there any text? Also, Prevent too big inclusions!
2879 $textSize = strlen( $text );
2880 if ( $textSize < 1 ||
$textSize > $this->mOptions
->getMaxIncludeSize() ) {
2884 if ( $frame === false ) {
2885 $frame = $this->getPreprocessor()->newFrame();
2886 } elseif ( !( $frame instanceof PPFrame
) ) {
2887 wfDebug( __METHOD__
. " called using plain parameters instead of "
2888 . "a PPFrame instance. Creating custom frame.\n" );
2889 $frame = $this->getPreprocessor()->newCustomFrame( $frame );
2892 $dom = $this->preprocessToDom( $text );
2893 $flags = $argsOnly ? PPFrame
::NO_TEMPLATES
: 0;
2894 $text = $frame->expand( $dom, $flags );
2900 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
2902 * @param array $args
2906 public static function createAssocArgs( $args ) {
2909 foreach ( $args as $arg ) {
2910 $eqpos = strpos( $arg, '=' );
2911 if ( $eqpos === false ) {
2912 $assocArgs[$index++
] = $arg;
2914 $name = trim( substr( $arg, 0, $eqpos ) );
2915 $value = trim( substr( $arg, $eqpos +
1 ) );
2916 if ( $value === false ) {
2919 if ( $name !== false ) {
2920 $assocArgs[$name] = $value;
2929 * Warn the user when a parser limitation is reached
2930 * Will warn at most once the user per limitation type
2932 * The results are shown during preview and run through the Parser (See EditPage.php)
2934 * @param string $limitationType Should be one of:
2935 * 'expensive-parserfunction' (corresponding messages:
2936 * 'expensive-parserfunction-warning',
2937 * 'expensive-parserfunction-category')
2938 * 'post-expand-template-argument' (corresponding messages:
2939 * 'post-expand-template-argument-warning',
2940 * 'post-expand-template-argument-category')
2941 * 'post-expand-template-inclusion' (corresponding messages:
2942 * 'post-expand-template-inclusion-warning',
2943 * 'post-expand-template-inclusion-category')
2944 * 'node-count-exceeded' (corresponding messages:
2945 * 'node-count-exceeded-warning',
2946 * 'node-count-exceeded-category')
2947 * 'expansion-depth-exceeded' (corresponding messages:
2948 * 'expansion-depth-exceeded-warning',
2949 * 'expansion-depth-exceeded-category')
2950 * @param string|int|null $current Current value
2951 * @param string|int|null $max Maximum allowed, when an explicit limit has been
2952 * exceeded, provide the values (optional)
2954 public function limitationWarn( $limitationType, $current = '', $max = '' ) {
2955 # does no harm if $current and $max are present but are unnecessary for the message
2956 # Not doing ->inLanguage( $this->mOptions->getUserLangObj() ), since this is shown
2957 # only during preview, and that would split the parser cache unnecessarily.
2958 $warning = wfMessage( "$limitationType-warning" )->numParams( $current, $max )
2960 $this->mOutput
->addWarning( $warning );
2961 $this->addTrackingCategory( "$limitationType-category" );
2965 * Return the text of a template, after recursively
2966 * replacing any variables or templates within the template.
2968 * @param array $piece The parts of the template
2969 * $piece['title']: the title, i.e. the part before the |
2970 * $piece['parts']: the parameter array
2971 * $piece['lineStart']: whether the brace was at the start of a line
2972 * @param PPFrame $frame The current frame, contains template arguments
2974 * @return string The text of the template
2976 public function braceSubstitution( $piece, $frame ) {
2980 // $text has been filled
2982 // wiki markup in $text should be escaped
2984 // $text is HTML, armour it against wikitext transformation
2986 // Force interwiki transclusion to be done in raw mode not rendered
2987 $forceRawInterwiki = false;
2988 // $text is a DOM node needing expansion in a child frame
2989 $isChildObj = false;
2990 // $text is a DOM node needing expansion in the current frame
2991 $isLocalObj = false;
2993 # Title object, where $text came from
2996 # $part1 is the bit before the first |, and must contain only title characters.
2997 # Various prefixes will be stripped from it later.
2998 $titleWithSpaces = $frame->expand( $piece['title'] );
2999 $part1 = trim( $titleWithSpaces );
3002 # Original title text preserved for various purposes
3003 $originalTitle = $part1;
3005 # $args is a list of argument nodes, starting from index 0, not including $part1
3006 # @todo FIXME: If piece['parts'] is null then the call to getLength()
3007 # below won't work b/c this $args isn't an object
3008 $args = ( null == $piece['parts'] ) ?
[] : $piece['parts'];
3010 $profileSection = null; // profile templates
3014 $substMatch = $this->mSubstWords
->matchStartAndRemove( $part1 );
3016 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3017 # Decide whether to expand template or keep wikitext as-is.
3018 if ( $this->ot
['wiki'] ) {
3019 if ( $substMatch === false ) {
3020 $literal = true; # literal when in PST with no prefix
3022 $literal = false; # expand when in PST with subst: or safesubst:
3025 if ( $substMatch == 'subst' ) {
3026 $literal = true; # literal when not in PST with plain subst:
3028 $literal = false; # expand when not in PST with safesubst: or no prefix
3032 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3039 if ( !$found && $args->getLength() == 0 ) {
3040 $id = $this->mVariables
->matchStartToEnd( $part1 );
3041 if ( $id !== false ) {
3042 $text = $this->getVariableValue( $id, $frame );
3043 if ( MagicWord
::getCacheTTL( $id ) > -1 ) {
3044 $this->mOutput
->updateCacheExpiry( MagicWord
::getCacheTTL( $id ) );
3050 # MSG, MSGNW and RAW
3053 $mwMsgnw = MagicWord
::get( 'msgnw' );
3054 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
3057 # Remove obsolete MSG:
3058 $mwMsg = MagicWord
::get( 'msg' );
3059 $mwMsg->matchStartAndRemove( $part1 );
3063 $mwRaw = MagicWord
::get( 'raw' );
3064 if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
3065 $forceRawInterwiki = true;
3071 $colonPos = strpos( $part1, ':' );
3072 if ( $colonPos !== false ) {
3073 $func = substr( $part1, 0, $colonPos );
3074 $funcArgs = [ trim( substr( $part1, $colonPos +
1 ) ) ];
3075 $argsLength = $args->getLength();
3076 for ( $i = 0; $i < $argsLength; $i++
) {
3077 $funcArgs[] = $args->item( $i );
3080 $result = $this->callParserFunction( $frame, $func, $funcArgs );
3081 } catch ( Exception
$ex ) {
3085 # The interface for parser functions allows for extracting
3086 # flags into the local scope. Extract any forwarded flags
3092 # Finish mangling title and then check for loops.
3093 # Set $title to a Title object and $titleText to the PDBK
3096 # Split the title into page and subpage
3098 $relative = $this->maybeDoSubpageLink( $part1, $subpage );
3099 if ( $part1 !== $relative ) {
3101 $ns = $this->mTitle
->getNamespace();
3103 $title = Title
::newFromText( $part1, $ns );
3105 $titleText = $title->getPrefixedText();
3106 # Check for language variants if the template is not found
3107 if ( $this->getConverterLanguage()->hasVariants() && $title->getArticleID() == 0 ) {
3108 $this->getConverterLanguage()->findVariantLink( $part1, $title, true );
3110 # Do recursion depth check
3111 $limit = $this->mOptions
->getMaxTemplateDepth();
3112 if ( $frame->depth
>= $limit ) {
3114 $text = '<span class="error">'
3115 . wfMessage( 'parser-template-recursion-depth-warning' )
3116 ->numParams( $limit )->inContentLanguage()->text()
3122 # Load from database
3123 if ( !$found && $title ) {
3124 $profileSection = $this->mProfiler
->scopedProfileIn( $title->getPrefixedDBkey() );
3125 if ( !$title->isExternal() ) {
3126 if ( $title->isSpecialPage()
3127 && $this->mOptions
->getAllowSpecialInclusion()
3128 && $this->ot
['html']
3130 // Pass the template arguments as URL parameters.
3131 // "uselang" will have no effect since the Language object
3132 // is forced to the one defined in ParserOptions.
3134 $argsLength = $args->getLength();
3135 for ( $i = 0; $i < $argsLength; $i++
) {
3136 $bits = $args->item( $i )->splitArg();
3137 if ( strval( $bits['index'] ) === '' ) {
3138 $name = trim( $frame->expand( $bits['name'], PPFrame
::STRIP_COMMENTS
) );
3139 $value = trim( $frame->expand( $bits['value'] ) );
3140 $pageArgs[$name] = $value;
3144 // Create a new context to execute the special page
3145 $context = new RequestContext
;
3146 $context->setTitle( $title );
3147 $context->setRequest( new FauxRequest( $pageArgs ) );
3148 $context->setUser( $this->getUser() );
3149 $context->setLanguage( $this->mOptions
->getUserLangObj() );
3150 $ret = SpecialPageFactory
::capturePath( $title, $context );
3152 $text = $context->getOutput()->getHTML();
3153 $this->mOutput
->addOutputPageMetadata( $context->getOutput() );
3156 $this->disableCache();
3158 } elseif ( MWNamespace
::isNonincludable( $title->getNamespace() ) ) {
3159 $found = false; # access denied
3160 wfDebug( __METHOD__
. ": template inclusion denied for " .
3161 $title->getPrefixedDBkey() . "\n" );
3163 list( $text, $title ) = $this->getTemplateDom( $title );
3164 if ( $text !== false ) {
3170 # If the title is valid but undisplayable, make a link to it
3171 if ( !$found && ( $this->ot
['html'] ||
$this->ot
['pre'] ) ) {
3172 $text = "[[:$titleText]]";
3175 } elseif ( $title->isTrans() ) {
3176 # Interwiki transclusion
3177 if ( $this->ot
['html'] && !$forceRawInterwiki ) {
3178 $text = $this->interwikiTransclude( $title, 'render' );
3181 $text = $this->interwikiTransclude( $title, 'raw' );
3182 # Preprocess it like a template
3183 $text = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
3189 # Do infinite loop check
3190 # This has to be done after redirect resolution to avoid infinite loops via redirects
3191 if ( !$frame->loopCheck( $title ) ) {
3193 $text = '<span class="error">'
3194 . wfMessage( 'parser-template-loop-warning', $titleText )->inContentLanguage()->text()
3196 wfDebug( __METHOD__
. ": template loop broken at '$titleText'\n" );
3200 # If we haven't found text to substitute by now, we're done
3201 # Recover the source wikitext and return it
3203 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3204 if ( $profileSection ) {
3205 $this->mProfiler
->scopedProfileOut( $profileSection );
3207 return [ 'object' => $text ];
3210 # Expand DOM-style return values in a child frame
3211 if ( $isChildObj ) {
3212 # Clean up argument array
3213 $newFrame = $frame->newChild( $args, $title );
3216 $text = $newFrame->expand( $text, PPFrame
::RECOVER_ORIG
);
3217 } elseif ( $titleText !== false && $newFrame->isEmpty() ) {
3218 # Expansion is eligible for the empty-frame cache
3219 $text = $newFrame->cachedExpand( $titleText, $text );
3221 # Uncached expansion
3222 $text = $newFrame->expand( $text );
3225 if ( $isLocalObj && $nowiki ) {
3226 $text = $frame->expand( $text, PPFrame
::RECOVER_ORIG
);
3227 $isLocalObj = false;
3230 if ( $profileSection ) {
3231 $this->mProfiler
->scopedProfileOut( $profileSection );
3234 # Replace raw HTML by a placeholder
3236 $text = $this->insertStripItem( $text );
3237 } elseif ( $nowiki && ( $this->ot
['html'] ||
$this->ot
['pre'] ) ) {
3238 # Escape nowiki-style return values
3239 $text = wfEscapeWikiText( $text );
3240 } elseif ( is_string( $text )
3241 && !$piece['lineStart']
3242 && preg_match( '/^(?:{\\||:|;|#|\*)/', $text )
3244 # Bug 529: if the template begins with a table or block-level
3245 # element, it should be treated as beginning a new line.
3246 # This behavior is somewhat controversial.
3247 $text = "\n" . $text;
3250 if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
3251 # Error, oversize inclusion
3252 if ( $titleText !== false ) {
3253 # Make a working, properly escaped link if possible (bug 23588)
3254 $text = "[[:$titleText]]";
3256 # This will probably not be a working link, but at least it may
3257 # provide some hint of where the problem is
3258 preg_replace( '/^:/', '', $originalTitle );
3259 $text = "[[:$originalTitle]]";
3261 $text .= $this->insertStripItem( '<!-- WARNING: template omitted, '
3262 . 'post-expand include size too large -->' );
3263 $this->limitationWarn( 'post-expand-template-inclusion' );
3266 if ( $isLocalObj ) {
3267 $ret = [ 'object' => $text ];
3269 $ret = [ 'text' => $text ];
3276 * Call a parser function and return an array with text and flags.
3278 * The returned array will always contain a boolean 'found', indicating
3279 * whether the parser function was found or not. It may also contain the
3281 * text: string|object, resulting wikitext or PP DOM object
3282 * isHTML: bool, $text is HTML, armour it against wikitext transformation
3283 * isChildObj: bool, $text is a DOM node needing expansion in a child frame
3284 * isLocalObj: bool, $text is a DOM node needing expansion in the current frame
3285 * nowiki: bool, wiki markup in $text should be escaped
3288 * @param PPFrame $frame The current frame, contains template arguments
3289 * @param string $function Function name
3290 * @param array $args Arguments to the function
3291 * @throws MWException
3294 public function callParserFunction( $frame, $function, array $args = [] ) {
3297 # Case sensitive functions
3298 if ( isset( $this->mFunctionSynonyms
[1][$function] ) ) {
3299 $function = $this->mFunctionSynonyms
[1][$function];
3301 # Case insensitive functions
3302 $function = $wgContLang->lc( $function );
3303 if ( isset( $this->mFunctionSynonyms
[0][$function] ) ) {
3304 $function = $this->mFunctionSynonyms
[0][$function];
3306 return [ 'found' => false ];
3310 list( $callback, $flags ) = $this->mFunctionHooks
[$function];
3312 # Workaround for PHP bug 35229 and similar
3313 if ( !is_callable( $callback ) ) {
3314 throw new MWException( "Tag hook for $function is not callable\n" );
3317 $allArgs = [ &$this ];
3318 if ( $flags & self
::SFH_OBJECT_ARGS
) {
3319 # Convert arguments to PPNodes and collect for appending to $allArgs
3321 foreach ( $args as $k => $v ) {
3322 if ( $v instanceof PPNode ||
$k === 0 ) {
3325 $funcArgs[] = $this->mPreprocessor
->newPartNodeArray( [ $k => $v ] )->item( 0 );
3329 # Add a frame parameter, and pass the arguments as an array
3330 $allArgs[] = $frame;
3331 $allArgs[] = $funcArgs;
3333 # Convert arguments to plain text and append to $allArgs
3334 foreach ( $args as $k => $v ) {
3335 if ( $v instanceof PPNode
) {
3336 $allArgs[] = trim( $frame->expand( $v ) );
3337 } elseif ( is_int( $k ) && $k >= 0 ) {
3338 $allArgs[] = trim( $v );
3340 $allArgs[] = trim( "$k=$v" );
3345 $result = call_user_func_array( $callback, $allArgs );
3347 # The interface for function hooks allows them to return a wikitext
3348 # string or an array containing the string and any flags. This mungs
3349 # things around to match what this method should return.
3350 if ( !is_array( $result ) ) {
3356 if ( isset( $result[0] ) && !isset( $result['text'] ) ) {
3357 $result['text'] = $result[0];
3359 unset( $result[0] );
3366 $preprocessFlags = 0;
3367 if ( isset( $result['noparse'] ) ) {
3368 $noparse = $result['noparse'];
3370 if ( isset( $result['preprocessFlags'] ) ) {
3371 $preprocessFlags = $result['preprocessFlags'];
3375 $result['text'] = $this->preprocessToDom( $result['text'], $preprocessFlags );
3376 $result['isChildObj'] = true;
3383 * Get the semi-parsed DOM representation of a template with a given title,
3384 * and its redirect destination title. Cached.
3386 * @param Title $title
3390 public function getTemplateDom( $title ) {
3391 $cacheTitle = $title;
3392 $titleText = $title->getPrefixedDBkey();
3394 if ( isset( $this->mTplRedirCache
[$titleText] ) ) {
3395 list( $ns, $dbk ) = $this->mTplRedirCache
[$titleText];
3396 $title = Title
::makeTitle( $ns, $dbk );
3397 $titleText = $title->getPrefixedDBkey();
3399 if ( isset( $this->mTplDomCache
[$titleText] ) ) {
3400 return [ $this->mTplDomCache
[$titleText], $title ];
3403 # Cache miss, go to the database
3404 list( $text, $title ) = $this->fetchTemplateAndTitle( $title );
3406 if ( $text === false ) {
3407 $this->mTplDomCache
[$titleText] = false;
3408 return [ false, $title ];
3411 $dom = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
3412 $this->mTplDomCache
[$titleText] = $dom;
3414 if ( !$title->equals( $cacheTitle ) ) {
3415 $this->mTplRedirCache
[$cacheTitle->getPrefixedDBkey()] =
3416 [ $title->getNamespace(), $cdb = $title->getDBkey() ];
3419 return [ $dom, $title ];
3423 * Fetch the current revision of a given title. Note that the revision
3424 * (and even the title) may not exist in the database, so everything
3425 * contributing to the output of the parser should use this method
3426 * where possible, rather than getting the revisions themselves. This
3427 * method also caches its results, so using it benefits performance.
3430 * @param Title $title
3433 public function fetchCurrentRevisionOfTitle( $title ) {
3434 $cacheKey = $title->getPrefixedDBkey();
3435 if ( !$this->currentRevisionCache
) {
3436 $this->currentRevisionCache
= new MapCacheLRU( 100 );
3438 if ( !$this->currentRevisionCache
->has( $cacheKey ) ) {
3439 $this->currentRevisionCache
->set( $cacheKey,
3440 // Defaults to Parser::statelessFetchRevision()
3441 call_user_func( $this->mOptions
->getCurrentRevisionCallback(), $title, $this )
3444 return $this->currentRevisionCache
->get( $cacheKey );
3448 * Wrapper around Revision::newFromTitle to allow passing additional parameters
3449 * without passing them on to it.
3452 * @param Title $title
3453 * @param Parser|bool $parser
3456 public static function statelessFetchRevision( $title, $parser = false ) {
3457 return Revision
::newFromTitle( $title );
3461 * Fetch the unparsed text of a template and register a reference to it.
3462 * @param Title $title
3463 * @return array ( string or false, Title )
3465 public function fetchTemplateAndTitle( $title ) {
3466 // Defaults to Parser::statelessFetchTemplate()
3467 $templateCb = $this->mOptions
->getTemplateCallback();
3468 $stuff = call_user_func( $templateCb, $title, $this );
3469 // We use U+007F DELETE to distinguish strip markers from regular text.
3470 $text = $stuff['text'];
3471 if ( is_string( $stuff['text'] ) ) {
3472 $text = strtr( $text, "\x7f", "?" );
3474 $finalTitle = isset( $stuff['finalTitle'] ) ?
$stuff['finalTitle'] : $title;
3475 if ( isset( $stuff['deps'] ) ) {
3476 foreach ( $stuff['deps'] as $dep ) {
3477 $this->mOutput
->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
3478 if ( $dep['title']->equals( $this->getTitle() ) ) {
3479 // If we transclude ourselves, the final result
3480 // will change based on the new version of the page
3481 $this->mOutput
->setFlag( 'vary-revision' );
3485 return [ $text, $finalTitle ];
3489 * Fetch the unparsed text of a template and register a reference to it.
3490 * @param Title $title
3491 * @return string|bool
3493 public function fetchTemplate( $title ) {
3494 return $this->fetchTemplateAndTitle( $title )[0];
3498 * Static function to get a template
3499 * Can be overridden via ParserOptions::setTemplateCallback().
3501 * @param Title $title
3502 * @param bool|Parser $parser
3506 public static function statelessFetchTemplate( $title, $parser = false ) {
3507 $text = $skip = false;
3508 $finalTitle = $title;
3511 # Loop to fetch the article, with up to 1 redirect
3512 // @codingStandardsIgnoreStart Generic.CodeAnalysis.ForLoopWithTestFunctionCall.NotAllowed
3513 for ( $i = 0; $i < 2 && is_object( $title ); $i++
) {
3514 // @codingStandardsIgnoreEnd
3515 # Give extensions a chance to select the revision instead
3516 $id = false; # Assume current
3517 Hooks
::run( 'BeforeParserFetchTemplateAndtitle',
3518 [ $parser, $title, &$skip, &$id ] );
3524 'page_id' => $title->getArticleID(),
3531 $rev = Revision
::newFromId( $id );
3532 } elseif ( $parser ) {
3533 $rev = $parser->fetchCurrentRevisionOfTitle( $title );
3535 $rev = Revision
::newFromTitle( $title );
3537 $rev_id = $rev ?
$rev->getId() : 0;
3538 # If there is no current revision, there is no page
3539 if ( $id === false && !$rev ) {
3540 $linkCache = LinkCache
::singleton();
3541 $linkCache->addBadLinkObj( $title );
3546 'page_id' => $title->getArticleID(),
3547 'rev_id' => $rev_id ];
3548 if ( $rev && !$title->equals( $rev->getTitle() ) ) {
3549 # We fetched a rev from a different title; register it too...
3551 'title' => $rev->getTitle(),
3552 'page_id' => $rev->getPage(),
3553 'rev_id' => $rev_id ];
3557 $content = $rev->getContent();
3558 $text = $content ?
$content->getWikitextForTransclusion() : null;
3560 if ( $text === false ||
$text === null ) {
3564 } elseif ( $title->getNamespace() == NS_MEDIAWIKI
) {
3566 $message = wfMessage( $wgContLang->lcfirst( $title->getText() ) )->inContentLanguage();
3567 if ( !$message->exists() ) {
3571 $content = $message->content();
3572 $text = $message->plain();
3580 $finalTitle = $title;
3581 $title = $content->getRedirectTarget();
3585 'finalTitle' => $finalTitle,
3590 * Fetch a file and its title and register a reference to it.
3591 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3592 * @param Title $title
3593 * @param array $options Array of options to RepoGroup::findFile
3596 public function fetchFile( $title, $options = [] ) {
3597 return $this->fetchFileAndTitle( $title, $options )[0];
3601 * Fetch a file and its title and register a reference to it.
3602 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3603 * @param Title $title
3604 * @param array $options Array of options to RepoGroup::findFile
3605 * @return array ( File or false, Title of file )
3607 public function fetchFileAndTitle( $title, $options = [] ) {
3608 $file = $this->fetchFileNoRegister( $title, $options );
3610 $time = $file ?
$file->getTimestamp() : false;
3611 $sha1 = $file ?
$file->getSha1() : false;
3612 # Register the file as a dependency...
3613 $this->mOutput
->addImage( $title->getDBkey(), $time, $sha1 );
3614 if ( $file && !$title->equals( $file->getTitle() ) ) {
3615 # Update fetched file title
3616 $title = $file->getTitle();
3617 $this->mOutput
->addImage( $title->getDBkey(), $time, $sha1 );
3619 return [ $file, $title ];
3623 * Helper function for fetchFileAndTitle.
3625 * Also useful if you need to fetch a file but not use it yet,
3626 * for example to get the file's handler.
3628 * @param Title $title
3629 * @param array $options Array of options to RepoGroup::findFile
3632 protected function fetchFileNoRegister( $title, $options = [] ) {
3633 if ( isset( $options['broken'] ) ) {
3634 $file = false; // broken thumbnail forced by hook
3635 } elseif ( isset( $options['sha1'] ) ) { // get by (sha1,timestamp)
3636 $file = RepoGroup
::singleton()->findFileFromKey( $options['sha1'], $options );
3637 } else { // get by (name,timestamp)
3638 $file = wfFindFile( $title, $options );
3644 * Transclude an interwiki link.
3646 * @param Title $title
3647 * @param string $action
3651 public function interwikiTransclude( $title, $action ) {
3652 global $wgEnableScaryTranscluding;
3654 if ( !$wgEnableScaryTranscluding ) {
3655 return wfMessage( 'scarytranscludedisabled' )->inContentLanguage()->text();
3658 $url = $title->getFullURL( [ 'action' => $action ] );
3660 if ( strlen( $url ) > 255 ) {
3661 return wfMessage( 'scarytranscludetoolong' )->inContentLanguage()->text();
3663 return $this->fetchScaryTemplateMaybeFromCache( $url );
3667 * @param string $url
3668 * @return mixed|string
3670 public function fetchScaryTemplateMaybeFromCache( $url ) {
3671 global $wgTranscludeCacheExpiry;
3672 $dbr = wfGetDB( DB_SLAVE
);
3673 $tsCond = $dbr->timestamp( time() - $wgTranscludeCacheExpiry );
3674 $obj = $dbr->selectRow( 'transcache', [ 'tc_time', 'tc_contents' ],
3675 [ 'tc_url' => $url, "tc_time >= " . $dbr->addQuotes( $tsCond ) ] );
3677 return $obj->tc_contents
;
3680 $req = MWHttpRequest
::factory( $url, [], __METHOD__
);
3681 $status = $req->execute(); // Status object
3682 if ( $status->isOK() ) {
3683 $text = $req->getContent();
3684 } elseif ( $req->getStatus() != 200 ) {
3685 // Though we failed to fetch the content, this status is useless.
3686 return wfMessage( 'scarytranscludefailed-httpstatus' )
3687 ->params( $url, $req->getStatus() /* HTTP status */ )->inContentLanguage()->text();
3689 return wfMessage( 'scarytranscludefailed', $url )->inContentLanguage()->text();
3692 $dbw = wfGetDB( DB_MASTER
);
3693 $dbw->replace( 'transcache', [ 'tc_url' ], [
3695 'tc_time' => $dbw->timestamp( time() ),
3696 'tc_contents' => $text
3702 * Triple brace replacement -- used for template arguments
3705 * @param array $piece
3706 * @param PPFrame $frame
3710 public function argSubstitution( $piece, $frame ) {
3713 $parts = $piece['parts'];
3714 $nameWithSpaces = $frame->expand( $piece['title'] );
3715 $argName = trim( $nameWithSpaces );
3717 $text = $frame->getArgument( $argName );
3718 if ( $text === false && $parts->getLength() > 0
3719 && ( $this->ot
['html']
3721 ||
( $this->ot
['wiki'] && $frame->isTemplate() )
3724 # No match in frame, use the supplied default
3725 $object = $parts->item( 0 )->getChildren();
3727 if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
3728 $error = '<!-- WARNING: argument omitted, expansion size too large -->';
3729 $this->limitationWarn( 'post-expand-template-argument' );
3732 if ( $text === false && $object === false ) {
3734 $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts );
3736 if ( $error !== false ) {
3739 if ( $object !== false ) {
3740 $ret = [ 'object' => $object ];
3742 $ret = [ 'text' => $text ];
3749 * Return the text to be used for a given extension tag.
3750 * This is the ghost of strip().
3752 * @param array $params Associative array of parameters:
3753 * name PPNode for the tag name
3754 * attr PPNode for unparsed text where tag attributes are thought to be
3755 * attributes Optional associative array of parsed attributes
3756 * inner Contents of extension element
3757 * noClose Original text did not have a close tag
3758 * @param PPFrame $frame
3760 * @throws MWException
3763 public function extensionSubstitution( $params, $frame ) {
3764 $name = $frame->expand( $params['name'] );
3765 $attrText = !isset( $params['attr'] ) ?
null : $frame->expand( $params['attr'] );
3766 $content = !isset( $params['inner'] ) ?
null : $frame->expand( $params['inner'] );
3767 $marker = self
::MARKER_PREFIX
. "-$name-"
3768 . sprintf( '%08X', $this->mMarkerIndex++
) . self
::MARKER_SUFFIX
;
3770 $isFunctionTag = isset( $this->mFunctionTagHooks
[strtolower( $name )] ) &&
3771 ( $this->ot
['html'] ||
$this->ot
['pre'] );
3772 if ( $isFunctionTag ) {
3773 $markerType = 'none';
3775 $markerType = 'general';
3777 if ( $this->ot
['html'] ||
$isFunctionTag ) {
3778 $name = strtolower( $name );
3779 $attributes = Sanitizer
::decodeTagAttributes( $attrText );
3780 if ( isset( $params['attributes'] ) ) {
3781 $attributes = $attributes +
$params['attributes'];
3784 if ( isset( $this->mTagHooks
[$name] ) ) {
3785 # Workaround for PHP bug 35229 and similar
3786 if ( !is_callable( $this->mTagHooks
[$name] ) ) {
3787 throw new MWException( "Tag hook for $name is not callable\n" );
3789 $output = call_user_func_array( $this->mTagHooks
[$name],
3790 [ $content, $attributes, $this, $frame ] );
3791 } elseif ( isset( $this->mFunctionTagHooks
[$name] ) ) {
3792 list( $callback, ) = $this->mFunctionTagHooks
[$name];
3793 if ( !is_callable( $callback ) ) {
3794 throw new MWException( "Tag hook for $name is not callable\n" );
3797 $output = call_user_func_array( $callback, [ &$this, $frame, $content, $attributes ] );
3799 $output = '<span class="error">Invalid tag extension name: ' .
3800 htmlspecialchars( $name ) . '</span>';
3803 if ( is_array( $output ) ) {
3804 # Extract flags to local scope (to override $markerType)
3806 $output = $flags[0];
3811 if ( is_null( $attrText ) ) {
3814 if ( isset( $params['attributes'] ) ) {
3815 foreach ( $params['attributes'] as $attrName => $attrValue ) {
3816 $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' .
3817 htmlspecialchars( $attrValue ) . '"';
3820 if ( $content === null ) {
3821 $output = "<$name$attrText/>";
3823 $close = is_null( $params['close'] ) ?
'' : $frame->expand( $params['close'] );
3824 $output = "<$name$attrText>$content$close";
3828 if ( $markerType === 'none' ) {
3830 } elseif ( $markerType === 'nowiki' ) {
3831 $this->mStripState
->addNoWiki( $marker, $output );
3832 } elseif ( $markerType === 'general' ) {
3833 $this->mStripState
->addGeneral( $marker, $output );
3835 throw new MWException( __METHOD__
. ': invalid marker type' );
3841 * Increment an include size counter
3843 * @param string $type The type of expansion
3844 * @param int $size The size of the text
3845 * @return bool False if this inclusion would take it over the maximum, true otherwise
3847 public function incrementIncludeSize( $type, $size ) {
3848 if ( $this->mIncludeSizes
[$type] +
$size > $this->mOptions
->getMaxIncludeSize() ) {
3851 $this->mIncludeSizes
[$type] +
= $size;
3857 * Increment the expensive function count
3859 * @return bool False if the limit has been exceeded
3861 public function incrementExpensiveFunctionCount() {
3862 $this->mExpensiveFunctionCount++
;
3863 return $this->mExpensiveFunctionCount
<= $this->mOptions
->getExpensiveParserFunctionLimit();
3867 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
3868 * Fills $this->mDoubleUnderscores, returns the modified text
3870 * @param string $text
3874 public function doDoubleUnderscore( $text ) {
3876 # The position of __TOC__ needs to be recorded
3877 $mw = MagicWord
::get( 'toc' );
3878 if ( $mw->match( $text ) ) {
3879 $this->mShowToc
= true;
3880 $this->mForceTocPosition
= true;
3882 # Set a placeholder. At the end we'll fill it in with the TOC.
3883 $text = $mw->replace( '<!--MWTOC-->', $text, 1 );
3885 # Only keep the first one.
3886 $text = $mw->replace( '', $text );
3889 # Now match and remove the rest of them
3890 $mwa = MagicWord
::getDoubleUnderscoreArray();
3891 $this->mDoubleUnderscores
= $mwa->matchAndRemove( $text );
3893 if ( isset( $this->mDoubleUnderscores
['nogallery'] ) ) {
3894 $this->mOutput
->mNoGallery
= true;
3896 if ( isset( $this->mDoubleUnderscores
['notoc'] ) && !$this->mForceTocPosition
) {
3897 $this->mShowToc
= false;
3899 if ( isset( $this->mDoubleUnderscores
['hiddencat'] )
3900 && $this->mTitle
->getNamespace() == NS_CATEGORY
3902 $this->addTrackingCategory( 'hidden-category-category' );
3904 # (bug 8068) Allow control over whether robots index a page.
3905 # @todo FIXME: Bug 14899: __INDEX__ always overrides __NOINDEX__ here! This
3906 # is not desirable, the last one on the page should win.
3907 if ( isset( $this->mDoubleUnderscores
['noindex'] ) && $this->mTitle
->canUseNoindex() ) {
3908 $this->mOutput
->setIndexPolicy( 'noindex' );
3909 $this->addTrackingCategory( 'noindex-category' );
3911 if ( isset( $this->mDoubleUnderscores
['index'] ) && $this->mTitle
->canUseNoindex() ) {
3912 $this->mOutput
->setIndexPolicy( 'index' );
3913 $this->addTrackingCategory( 'index-category' );
3916 # Cache all double underscores in the database
3917 foreach ( $this->mDoubleUnderscores
as $key => $val ) {
3918 $this->mOutput
->setProperty( $key, '' );
3925 * @see ParserOutput::addTrackingCategory()
3926 * @param string $msg Message key
3927 * @return bool Whether the addition was successful
3929 public function addTrackingCategory( $msg ) {
3930 return $this->mOutput
->addTrackingCategory( $msg, $this->mTitle
);
3934 * This function accomplishes several tasks:
3935 * 1) Auto-number headings if that option is enabled
3936 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
3937 * 3) Add a Table of contents on the top for users who have enabled the option
3938 * 4) Auto-anchor headings
3940 * It loops through all headlines, collects the necessary data, then splits up the
3941 * string and re-inserts the newly formatted headlines.
3943 * @param string $text
3944 * @param string $origText Original, untouched wikitext
3945 * @param bool $isMain
3946 * @return mixed|string
3949 public function formatHeadings( $text, $origText, $isMain = true ) {
3950 global $wgMaxTocLevel, $wgExperimentalHtmlIds;
3952 # Inhibit editsection links if requested in the page
3953 if ( isset( $this->mDoubleUnderscores
['noeditsection'] ) ) {
3954 $maybeShowEditLink = $showEditLink = false;
3956 $maybeShowEditLink = true; /* Actual presence will depend on ParserOptions option */
3957 $showEditLink = $this->mOptions
->getEditSection();
3959 if ( $showEditLink ) {
3960 $this->mOutput
->setEditSectionTokens( true );
3963 # Get all headlines for numbering them and adding funky stuff like [edit]
3964 # links - this is for later, but we need the number of headlines right now
3966 $numMatches = preg_match_all(
3967 '/<H(?P<level>[1-6])(?P<attrib>.*?>)\s*(?P<header>[\s\S]*?)\s*<\/H[1-6] *>/i',
3972 # if there are fewer than 4 headlines in the article, do not show TOC
3973 # unless it's been explicitly enabled.
3974 $enoughToc = $this->mShowToc
&&
3975 ( ( $numMatches >= 4 ) ||
$this->mForceTocPosition
);
3977 # Allow user to stipulate that a page should have a "new section"
3978 # link added via __NEWSECTIONLINK__
3979 if ( isset( $this->mDoubleUnderscores
['newsectionlink'] ) ) {
3980 $this->mOutput
->setNewSection( true );
3983 # Allow user to remove the "new section"
3984 # link via __NONEWSECTIONLINK__
3985 if ( isset( $this->mDoubleUnderscores
['nonewsectionlink'] ) ) {
3986 $this->mOutput
->hideNewSection( true );
3989 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
3990 # override above conditions and always show TOC above first header
3991 if ( isset( $this->mDoubleUnderscores
['forcetoc'] ) ) {
3992 $this->mShowToc
= true;
4000 # Ugh .. the TOC should have neat indentation levels which can be
4001 # passed to the skin functions. These are determined here
4005 $sublevelCount = [];
4011 $markerRegex = self
::MARKER_PREFIX
. "-h-(\d+)-" . self
::MARKER_SUFFIX
;
4012 $baseTitleText = $this->mTitle
->getPrefixedDBkey();
4013 $oldType = $this->mOutputType
;
4014 $this->setOutputType( self
::OT_WIKI
);
4015 $frame = $this->getPreprocessor()->newFrame();
4016 $root = $this->preprocessToDom( $origText );
4017 $node = $root->getFirstChild();
4022 $headlines = $numMatches !== false ?
$matches[3] : [];
4024 foreach ( $headlines as $headline ) {
4025 $isTemplate = false;
4027 $sectionIndex = false;
4029 $markerMatches = [];
4030 if ( preg_match( "/^$markerRegex/", $headline, $markerMatches ) ) {
4031 $serial = $markerMatches[1];
4032 list( $titleText, $sectionIndex ) = $this->mHeadings
[$serial];
4033 $isTemplate = ( $titleText != $baseTitleText );
4034 $headline = preg_replace( "/^$markerRegex\\s*/", "", $headline );
4038 $prevlevel = $level;
4040 $level = $matches[1][$headlineCount];
4042 if ( $level > $prevlevel ) {
4043 # Increase TOC level
4045 $sublevelCount[$toclevel] = 0;
4046 if ( $toclevel < $wgMaxTocLevel ) {
4047 $prevtoclevel = $toclevel;
4048 $toc .= Linker
::tocIndent();
4051 } elseif ( $level < $prevlevel && $toclevel > 1 ) {
4052 # Decrease TOC level, find level to jump to
4054 for ( $i = $toclevel; $i > 0; $i-- ) {
4055 if ( $levelCount[$i] == $level ) {
4056 # Found last matching level
4059 } elseif ( $levelCount[$i] < $level ) {
4060 # Found first matching level below current level
4068 if ( $toclevel < $wgMaxTocLevel ) {
4069 if ( $prevtoclevel < $wgMaxTocLevel ) {
4070 # Unindent only if the previous toc level was shown :p
4071 $toc .= Linker
::tocUnindent( $prevtoclevel - $toclevel );
4072 $prevtoclevel = $toclevel;
4074 $toc .= Linker
::tocLineEnd();
4078 # No change in level, end TOC line
4079 if ( $toclevel < $wgMaxTocLevel ) {
4080 $toc .= Linker
::tocLineEnd();
4084 $levelCount[$toclevel] = $level;
4086 # count number of headlines for each level
4087 $sublevelCount[$toclevel]++
;
4089 for ( $i = 1; $i <= $toclevel; $i++
) {
4090 if ( !empty( $sublevelCount[$i] ) ) {
4094 $numbering .= $this->getTargetLanguage()->formatNum( $sublevelCount[$i] );
4099 # The safe header is a version of the header text safe to use for links
4101 # Remove link placeholders by the link text.
4102 # <!--LINK number-->
4104 # link text with suffix
4105 # Do this before unstrip since link text can contain strip markers
4106 $safeHeadline = $this->replaceLinkHoldersText( $headline );
4108 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4109 $safeHeadline = $this->mStripState
->unstripBoth( $safeHeadline );
4111 # Strip out HTML (first regex removes any tag not allowed)
4113 # * <sup> and <sub> (bug 8393)
4116 # * <bdi> (bug 72884)
4117 # * <span dir="rtl"> and <span dir="ltr"> (bug 35167)
4118 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
4119 # to allow setting directionality in toc items.
4120 $tocline = preg_replace(
4122 '#<(?!/?(span|sup|sub|bdi|i|b)(?: [^>]*)?>).*?>#',
4123 '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|bdi|i|b))(?: .*?)?>#'
4129 # Strip '<span></span>', which is the result from the above if
4130 # <span id="foo"></span> is used to produce an additional anchor
4132 $tocline = str_replace( '<span></span>', '', $tocline );
4134 $tocline = trim( $tocline );
4136 # For the anchor, strip out HTML-y stuff period
4137 $safeHeadline = preg_replace( '/<.*?>/', '', $safeHeadline );
4138 $safeHeadline = Sanitizer
::normalizeSectionNameWhitespace( $safeHeadline );
4140 # Save headline for section edit hint before it's escaped
4141 $headlineHint = $safeHeadline;
4143 if ( $wgExperimentalHtmlIds ) {
4144 # For reverse compatibility, provide an id that's
4145 # HTML4-compatible, like we used to.
4146 # It may be worth noting, academically, that it's possible for
4147 # the legacy anchor to conflict with a non-legacy headline
4148 # anchor on the page. In this case likely the "correct" thing
4149 # would be to either drop the legacy anchors or make sure
4150 # they're numbered first. However, this would require people
4151 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
4152 # manually, so let's not bother worrying about it.
4153 $legacyHeadline = Sanitizer
::escapeId( $safeHeadline,
4154 [ 'noninitial', 'legacy' ] );
4155 $safeHeadline = Sanitizer
::escapeId( $safeHeadline );
4157 if ( $legacyHeadline == $safeHeadline ) {
4158 # No reason to have both (in fact, we can't)
4159 $legacyHeadline = false;
4162 $legacyHeadline = false;
4163 $safeHeadline = Sanitizer
::escapeId( $safeHeadline,
4167 # HTML names must be case-insensitively unique (bug 10721).
4168 # This does not apply to Unicode characters per
4169 # http://www.w3.org/TR/html5/infrastructure.html#case-sensitivity-and-string-comparison
4170 # @todo FIXME: We may be changing them depending on the current locale.
4171 $arrayKey = strtolower( $safeHeadline );
4172 if ( $legacyHeadline === false ) {
4173 $legacyArrayKey = false;
4175 $legacyArrayKey = strtolower( $legacyHeadline );
4178 # Create the anchor for linking from the TOC to the section
4179 $anchor = $safeHeadline;
4180 $legacyAnchor = $legacyHeadline;
4181 if ( isset( $refers[$arrayKey] ) ) {
4182 // @codingStandardsIgnoreStart
4183 for ( $i = 2; isset( $refers["${arrayKey}_$i"] ); ++
$i );
4184 // @codingStandardsIgnoreEnd
4186 $refers["${arrayKey}_$i"] = true;
4188 $refers[$arrayKey] = true;
4190 if ( $legacyHeadline !== false && isset( $refers[$legacyArrayKey] ) ) {
4191 // @codingStandardsIgnoreStart
4192 for ( $i = 2; isset( $refers["${legacyArrayKey}_$i"] ); ++
$i );
4193 // @codingStandardsIgnoreEnd
4194 $legacyAnchor .= "_$i";
4195 $refers["${legacyArrayKey}_$i"] = true;
4197 $refers[$legacyArrayKey] = true;
4200 # Don't number the heading if it is the only one (looks silly)
4201 if ( count( $matches[3] ) > 1 && $this->mOptions
->getNumberHeadings() ) {
4202 # the two are different if the line contains a link
4203 $headline = Html
::element(
4205 [ 'class' => 'mw-headline-number' ],
4207 ) . ' ' . $headline;
4210 if ( $enoughToc && ( !isset( $wgMaxTocLevel ) ||
$toclevel < $wgMaxTocLevel ) ) {
4211 $toc .= Linker
::tocLine( $anchor, $tocline,
4212 $numbering, $toclevel, ( $isTemplate ?
false : $sectionIndex ) );
4215 # Add the section to the section tree
4216 # Find the DOM node for this header
4217 $noOffset = ( $isTemplate ||
$sectionIndex === false );
4218 while ( $node && !$noOffset ) {
4219 if ( $node->getName() === 'h' ) {
4220 $bits = $node->splitHeading();
4221 if ( $bits['i'] == $sectionIndex ) {
4225 $byteOffset +
= mb_strlen( $this->mStripState
->unstripBoth(
4226 $frame->expand( $node, PPFrame
::RECOVER_ORIG
) ) );
4227 $node = $node->getNextSibling();
4230 'toclevel' => $toclevel,
4233 'number' => $numbering,
4234 'index' => ( $isTemplate ?
'T-' : '' ) . $sectionIndex,
4235 'fromtitle' => $titleText,
4236 'byteoffset' => ( $noOffset ?
null : $byteOffset ),
4237 'anchor' => $anchor,
4240 # give headline the correct <h#> tag
4241 if ( $maybeShowEditLink && $sectionIndex !== false ) {
4242 // Output edit section links as markers with styles that can be customized by skins
4243 if ( $isTemplate ) {
4244 # Put a T flag in the section identifier, to indicate to extractSections()
4245 # that sections inside <includeonly> should be counted.
4246 $editsectionPage = $titleText;
4247 $editsectionSection = "T-$sectionIndex";
4248 $editsectionContent = null;
4250 $editsectionPage = $this->mTitle
->getPrefixedText();
4251 $editsectionSection = $sectionIndex;
4252 $editsectionContent = $headlineHint;
4254 // We use a bit of pesudo-xml for editsection markers. The
4255 // language converter is run later on. Using a UNIQ style marker
4256 // leads to the converter screwing up the tokens when it
4257 // converts stuff. And trying to insert strip tags fails too. At
4258 // this point all real inputted tags have already been escaped,
4259 // so we don't have to worry about a user trying to input one of
4260 // these markers directly. We use a page and section attribute
4261 // to stop the language converter from converting these
4262 // important bits of data, but put the headline hint inside a
4263 // content block because the language converter is supposed to
4264 // be able to convert that piece of data.
4265 // Gets replaced with html in ParserOutput::getText
4266 $editlink = '<mw:editsection page="' . htmlspecialchars( $editsectionPage );
4267 $editlink .= '" section="' . htmlspecialchars( $editsectionSection ) . '"';
4268 if ( $editsectionContent !== null ) {
4269 $editlink .= '>' . $editsectionContent . '</mw:editsection>';
4276 $head[$headlineCount] = Linker
::makeHeadline( $level,
4277 $matches['attrib'][$headlineCount], $anchor, $headline,
4278 $editlink, $legacyAnchor );
4283 $this->setOutputType( $oldType );
4285 # Never ever show TOC if no headers
4286 if ( $numVisible < 1 ) {
4291 if ( $prevtoclevel > 0 && $prevtoclevel < $wgMaxTocLevel ) {
4292 $toc .= Linker
::tocUnindent( $prevtoclevel - 1 );
4294 $toc = Linker
::tocList( $toc, $this->mOptions
->getUserLangObj() );
4295 $this->mOutput
->setTOCHTML( $toc );
4296 $toc = self
::TOC_START
. $toc . self
::TOC_END
;
4297 $this->mOutput
->addModules( 'mediawiki.toc' );
4301 $this->mOutput
->setSections( $tocraw );
4304 # split up and insert constructed headlines
4305 $blocks = preg_split( '/<H[1-6].*?>[\s\S]*?<\/H[1-6]>/i', $text );
4308 // build an array of document sections
4310 foreach ( $blocks as $block ) {
4311 // $head is zero-based, sections aren't.
4312 if ( empty( $head[$i - 1] ) ) {
4313 $sections[$i] = $block;
4315 $sections[$i] = $head[$i - 1] . $block;
4319 * Send a hook, one per section.
4320 * The idea here is to be able to make section-level DIVs, but to do so in a
4321 * lower-impact, more correct way than r50769
4324 * $section : the section number
4325 * &$sectionContent : ref to the content of the section
4326 * $showEditLinks : boolean describing whether this section has an edit link
4328 Hooks
::run( 'ParserSectionCreate', [ $this, $i, &$sections[$i], $showEditLink ] );
4333 if ( $enoughToc && $isMain && !$this->mForceTocPosition
) {
4334 // append the TOC at the beginning
4335 // Top anchor now in skin
4336 $sections[0] = $sections[0] . $toc . "\n";
4339 $full .= implode( '', $sections );
4341 if ( $this->mForceTocPosition
) {
4342 return str_replace( '<!--MWTOC-->', $toc, $full );
4349 * Transform wiki markup when saving a page by doing "\r\n" -> "\n"
4350 * conversion, substituting signatures, {{subst:}} templates, etc.
4352 * @param string $text The text to transform
4353 * @param Title $title The Title object for the current article
4354 * @param User $user The User object describing the current user
4355 * @param ParserOptions $options Parsing options
4356 * @param bool $clearState Whether to clear the parser state first
4357 * @return string The altered wiki markup
4359 public function preSaveTransform( $text, Title
$title, User
$user,
4360 ParserOptions
$options, $clearState = true
4362 if ( $clearState ) {
4363 $magicScopeVariable = $this->lock();
4365 $this->startParse( $title, $options, self
::OT_WIKI
, $clearState );
4366 $this->setUser( $user );
4372 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
4373 if ( $options->getPreSaveTransform() ) {
4374 $text = $this->pstPass2( $text, $user );
4376 $text = $this->mStripState
->unstripBoth( $text );
4378 $this->setUser( null ); # Reset
4384 * Pre-save transform helper function
4386 * @param string $text
4391 private function pstPass2( $text, $user ) {
4394 # Note: This is the timestamp saved as hardcoded wikitext to
4395 # the database, we use $wgContLang here in order to give
4396 # everyone the same signature and use the default one rather
4397 # than the one selected in each user's preferences.
4398 # (see also bug 12815)
4399 $ts = $this->mOptions
->getTimestamp();
4400 $timestamp = MWTimestamp
::getLocalInstance( $ts );
4401 $ts = $timestamp->format( 'YmdHis' );
4402 $tzMsg = $timestamp->getTimezoneMessage()->inContentLanguage()->text();
4404 $d = $wgContLang->timeanddate( $ts, false, false ) . " ($tzMsg)";
4406 # Variable replacement
4407 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4408 $text = $this->replaceVariables( $text );
4410 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4411 # which may corrupt this parser instance via its wfMessage()->text() call-
4414 $sigText = $this->getUserSig( $user );
4415 $text = strtr( $text, [
4417 '~~~~' => "$sigText $d",
4421 # Context links ("pipe tricks"): [[|name]] and [[name (context)|]]
4422 $tc = '[' . Title
::legalChars() . ']';
4423 $nc = '[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can use non-ascii!
4425 // [[ns:page (context)|]]
4426 $p1 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\))\\|]]/";
4427 // [[ns:page(context)|]] (double-width brackets, added in r40257)
4428 $p4 = "/\[\[(:?$nc+:|:|)($tc+?)( ?($tc+))\\|]]/";
4429 // [[ns:page (context), context|]] (using either single or double-width comma)
4430 $p3 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\)|)((?:, |,)$tc+|)\\|]]/";
4431 // [[|page]] (reverse pipe trick: add context from page title)
4432 $p2 = "/\[\[\\|($tc+)]]/";
4434 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4435 $text = preg_replace( $p1, '[[\\1\\2\\3|\\2]]', $text );
4436 $text = preg_replace( $p4, '[[\\1\\2\\3|\\2]]', $text );
4437 $text = preg_replace( $p3, '[[\\1\\2\\3\\4|\\2]]', $text );
4439 $t = $this->mTitle
->getText();
4441 if ( preg_match( "/^($nc+:|)$tc+?( \\($tc+\\))$/", $t, $m ) ) {
4442 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4443 } elseif ( preg_match( "/^($nc+:|)$tc+?(, $tc+|)$/", $t, $m ) && "$m[1]$m[2]" != '' ) {
4444 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4446 # if there's no context, don't bother duplicating the title
4447 $text = preg_replace( $p2, '[[\\1]]', $text );
4450 # Trim trailing whitespace
4451 $text = rtrim( $text );
4457 * Fetch the user's signature text, if any, and normalize to
4458 * validated, ready-to-insert wikitext.
4459 * If you have pre-fetched the nickname or the fancySig option, you can
4460 * specify them here to save a database query.
4461 * Do not reuse this parser instance after calling getUserSig(),
4462 * as it may have changed if it's the $wgParser.
4465 * @param string|bool $nickname Nickname to use or false to use user's default nickname
4466 * @param bool|null $fancySig whether the nicknname is the complete signature
4467 * or null to use default value
4470 public function getUserSig( &$user, $nickname = false, $fancySig = null ) {
4471 global $wgMaxSigChars;
4473 $username = $user->getName();
4475 # If not given, retrieve from the user object.
4476 if ( $nickname === false ) {
4477 $nickname = $user->getOption( 'nickname' );
4480 if ( is_null( $fancySig ) ) {
4481 $fancySig = $user->getBoolOption( 'fancysig' );
4484 $nickname = $nickname == null ?
$username : $nickname;
4486 if ( mb_strlen( $nickname ) > $wgMaxSigChars ) {
4487 $nickname = $username;
4488 wfDebug( __METHOD__
. ": $username has overlong signature.\n" );
4489 } elseif ( $fancySig !== false ) {
4490 # Sig. might contain markup; validate this
4491 if ( $this->validateSig( $nickname ) !== false ) {
4492 # Validated; clean up (if needed) and return it
4493 return $this->cleanSig( $nickname, true );
4495 # Failed to validate; fall back to the default
4496 $nickname = $username;
4497 wfDebug( __METHOD__
. ": $username has bad XML tags in signature.\n" );
4501 # Make sure nickname doesnt get a sig in a sig
4502 $nickname = self
::cleanSigInSig( $nickname );
4504 # If we're still here, make it a link to the user page
4505 $userText = wfEscapeWikiText( $username );
4506 $nickText = wfEscapeWikiText( $nickname );
4507 $msgName = $user->isAnon() ?
'signature-anon' : 'signature';
4509 return wfMessage( $msgName, $userText, $nickText )->inContentLanguage()
4510 ->title( $this->getTitle() )->text();
4514 * Check that the user's signature contains no bad XML
4516 * @param string $text
4517 * @return string|bool An expanded string, or false if invalid.
4519 public function validateSig( $text ) {
4520 return Xml
::isWellFormedXmlFragment( $text ) ?
$text : false;
4524 * Clean up signature text
4526 * 1) Strip 3, 4 or 5 tildes out of signatures @see cleanSigInSig
4527 * 2) Substitute all transclusions
4529 * @param string $text
4530 * @param bool $parsing Whether we're cleaning (preferences save) or parsing
4531 * @return string Signature text
4533 public function cleanSig( $text, $parsing = false ) {
4536 $magicScopeVariable = $this->lock();
4537 $this->startParse( $wgTitle, new ParserOptions
, self
::OT_PREPROCESS
, true );
4540 # Option to disable this feature
4541 if ( !$this->mOptions
->getCleanSignatures() ) {
4545 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4546 # => Move this logic to braceSubstitution()
4547 $substWord = MagicWord
::get( 'subst' );
4548 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4549 $substText = '{{' . $substWord->getSynonym( 0 );
4551 $text = preg_replace( $substRegex, $substText, $text );
4552 $text = self
::cleanSigInSig( $text );
4553 $dom = $this->preprocessToDom( $text );
4554 $frame = $this->getPreprocessor()->newFrame();
4555 $text = $frame->expand( $dom );
4558 $text = $this->mStripState
->unstripBoth( $text );
4565 * Strip 3, 4 or 5 tildes out of signatures.
4567 * @param string $text
4568 * @return string Signature text with /~{3,5}/ removed
4570 public static function cleanSigInSig( $text ) {
4571 $text = preg_replace( '/~{3,5}/', '', $text );
4576 * Set up some variables which are usually set up in parse()
4577 * so that an external function can call some class members with confidence
4579 * @param Title|null $title
4580 * @param ParserOptions $options
4581 * @param int $outputType
4582 * @param bool $clearState
4584 public function startExternalParse( Title
$title = null, ParserOptions
$options,
4585 $outputType, $clearState = true
4587 $this->startParse( $title, $options, $outputType, $clearState );
4591 * @param Title|null $title
4592 * @param ParserOptions $options
4593 * @param int $outputType
4594 * @param bool $clearState
4596 private function startParse( Title
$title = null, ParserOptions
$options,
4597 $outputType, $clearState = true
4599 $this->setTitle( $title );
4600 $this->mOptions
= $options;
4601 $this->setOutputType( $outputType );
4602 if ( $clearState ) {
4603 $this->clearState();
4608 * Wrapper for preprocess()
4610 * @param string $text The text to preprocess
4611 * @param ParserOptions $options Options
4612 * @param Title|null $title Title object or null to use $wgTitle
4615 public function transformMsg( $text, $options, $title = null ) {
4616 static $executing = false;
4618 # Guard against infinite recursion
4629 $text = $this->preprocess( $text, $title, $options );
4636 * Create an HTML-style tag, e.g. "<yourtag>special text</yourtag>"
4637 * The callback should have the following form:
4638 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4640 * Transform and return $text. Use $parser for any required context, e.g. use
4641 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4643 * Hooks may return extended information by returning an array, of which the
4644 * first numbered element (index 0) must be the return string, and all other
4645 * entries are extracted into local variables within an internal function
4646 * in the Parser class.
4648 * This interface (introduced r61913) appears to be undocumented, but
4649 * 'markerType' is used by some core tag hooks to override which strip
4650 * array their results are placed in. **Use great caution if attempting
4651 * this interface, as it is not documented and injudicious use could smash
4652 * private variables.**
4654 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4655 * @param callable $callback The callback function (and object) to use for the tag
4656 * @throws MWException
4657 * @return callable|null The old value of the mTagHooks array associated with the hook
4659 public function setHook( $tag, $callback ) {
4660 $tag = strtolower( $tag );
4661 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4662 throw new MWException( "Invalid character {$m[0]} in setHook('$tag', ...) call" );
4664 $oldVal = isset( $this->mTagHooks
[$tag] ) ?
$this->mTagHooks
[$tag] : null;
4665 $this->mTagHooks
[$tag] = $callback;
4666 if ( !in_array( $tag, $this->mStripList
) ) {
4667 $this->mStripList
[] = $tag;
4674 * As setHook(), but letting the contents be parsed.
4676 * Transparent tag hooks are like regular XML-style tag hooks, except they
4677 * operate late in the transformation sequence, on HTML instead of wikitext.
4679 * This is probably obsoleted by things dealing with parser frames?
4680 * The only extension currently using it is geoserver.
4683 * @todo better document or deprecate this
4685 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4686 * @param callable $callback The callback function (and object) to use for the tag
4687 * @throws MWException
4688 * @return callable|null The old value of the mTagHooks array associated with the hook
4690 public function setTransparentTagHook( $tag, $callback ) {
4691 $tag = strtolower( $tag );
4692 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4693 throw new MWException( "Invalid character {$m[0]} in setTransparentHook('$tag', ...) call" );
4695 $oldVal = isset( $this->mTransparentTagHooks
[$tag] ) ?
$this->mTransparentTagHooks
[$tag] : null;
4696 $this->mTransparentTagHooks
[$tag] = $callback;
4702 * Remove all tag hooks
4704 public function clearTagHooks() {
4705 $this->mTagHooks
= [];
4706 $this->mFunctionTagHooks
= [];
4707 $this->mStripList
= $this->mDefaultStripList
;
4711 * Create a function, e.g. {{sum:1|2|3}}
4712 * The callback function should have the form:
4713 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4715 * Or with Parser::SFH_OBJECT_ARGS:
4716 * function myParserFunction( $parser, $frame, $args ) { ... }
4718 * The callback may either return the text result of the function, or an array with the text
4719 * in element 0, and a number of flags in the other elements. The names of the flags are
4720 * specified in the keys. Valid flags are:
4721 * found The text returned is valid, stop processing the template. This
4723 * nowiki Wiki markup in the return value should be escaped
4724 * isHTML The returned text is HTML, armour it against wikitext transformation
4726 * @param string $id The magic word ID
4727 * @param callable $callback The callback function (and object) to use
4728 * @param int $flags A combination of the following flags:
4729 * Parser::SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4731 * Parser::SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text.
4732 * This allows for conditional expansion of the parse tree, allowing you to eliminate dead
4733 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4734 * the arguments, and to control the way they are expanded.
4736 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4737 * arguments, for instance:
4738 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4740 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4741 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4742 * working if/when this is changed.
4744 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4747 * Please read the documentation in includes/parser/Preprocessor.php for more information
4748 * about the methods available in PPFrame and PPNode.
4750 * @throws MWException
4751 * @return string|callable The old callback function for this name, if any
4753 public function setFunctionHook( $id, $callback, $flags = 0 ) {
4756 $oldVal = isset( $this->mFunctionHooks
[$id] ) ?
$this->mFunctionHooks
[$id][0] : null;
4757 $this->mFunctionHooks
[$id] = [ $callback, $flags ];
4759 # Add to function cache
4760 $mw = MagicWord
::get( $id );
4762 throw new MWException( __METHOD__
. '() expecting a magic word identifier.' );
4765 $synonyms = $mw->getSynonyms();
4766 $sensitive = intval( $mw->isCaseSensitive() );
4768 foreach ( $synonyms as $syn ) {
4770 if ( !$sensitive ) {
4771 $syn = $wgContLang->lc( $syn );
4774 if ( !( $flags & self
::SFH_NO_HASH
) ) {
4777 # Remove trailing colon
4778 if ( substr( $syn, -1, 1 ) === ':' ) {
4779 $syn = substr( $syn, 0, -1 );
4781 $this->mFunctionSynonyms
[$sensitive][$syn] = $id;
4787 * Get all registered function hook identifiers
4791 public function getFunctionHooks() {
4792 return array_keys( $this->mFunctionHooks
);
4796 * Create a tag function, e.g. "<test>some stuff</test>".
4797 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4798 * Unlike parser functions, their content is not preprocessed.
4799 * @param string $tag
4800 * @param callable $callback
4802 * @throws MWException
4805 public function setFunctionTagHook( $tag, $callback, $flags ) {
4806 $tag = strtolower( $tag );
4807 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4808 throw new MWException( "Invalid character {$m[0]} in setFunctionTagHook('$tag', ...) call" );
4810 $old = isset( $this->mFunctionTagHooks
[$tag] ) ?
4811 $this->mFunctionTagHooks
[$tag] : null;
4812 $this->mFunctionTagHooks
[$tag] = [ $callback, $flags ];
4814 if ( !in_array( $tag, $this->mStripList
) ) {
4815 $this->mStripList
[] = $tag;
4822 * Replace "<!--LINK-->" link placeholders with actual links, in the buffer
4823 * Placeholders created in Linker::link()
4825 * @param string $text
4826 * @param int $options
4828 public function replaceLinkHolders( &$text, $options = 0 ) {
4829 $this->mLinkHolders
->replace( $text );
4833 * Replace "<!--LINK-->" link placeholders with plain text of links
4834 * (not HTML-formatted).
4836 * @param string $text
4839 public function replaceLinkHoldersText( $text ) {
4840 return $this->mLinkHolders
->replaceText( $text );
4844 * Renders an image gallery from a text with one line per image.
4845 * text labels may be given by using |-style alternative text. E.g.
4846 * Image:one.jpg|The number "1"
4847 * Image:tree.jpg|A tree
4848 * given as text will return the HTML of a gallery with two images,
4849 * labeled 'The number "1"' and
4852 * @param string $text
4853 * @param array $params
4854 * @return string HTML
4856 public function renderImageGallery( $text, $params ) {
4859 if ( isset( $params['mode'] ) ) {
4860 $mode = $params['mode'];
4864 $ig = ImageGalleryBase
::factory( $mode );
4865 } catch ( Exception
$e ) {
4866 // If invalid type set, fallback to default.
4867 $ig = ImageGalleryBase
::factory( false );
4870 $ig->setContextTitle( $this->mTitle
);
4871 $ig->setShowBytes( false );
4872 $ig->setShowFilename( false );
4873 $ig->setParser( $this );
4874 $ig->setHideBadImages();
4875 $ig->setAttributes( Sanitizer
::validateTagAttributes( $params, 'table' ) );
4877 if ( isset( $params['showfilename'] ) ) {
4878 $ig->setShowFilename( true );
4880 $ig->setShowFilename( false );
4882 if ( isset( $params['caption'] ) ) {
4883 $caption = $params['caption'];
4884 $caption = htmlspecialchars( $caption );
4885 $caption = $this->replaceInternalLinks( $caption );
4886 $ig->setCaptionHtml( $caption );
4888 if ( isset( $params['perrow'] ) ) {
4889 $ig->setPerRow( $params['perrow'] );
4891 if ( isset( $params['widths'] ) ) {
4892 $ig->setWidths( $params['widths'] );
4894 if ( isset( $params['heights'] ) ) {
4895 $ig->setHeights( $params['heights'] );
4897 $ig->setAdditionalOptions( $params );
4899 Hooks
::run( 'BeforeParserrenderImageGallery', [ &$this, &$ig ] );
4901 $lines = StringUtils
::explode( "\n", $text );
4902 foreach ( $lines as $line ) {
4903 # match lines like these:
4904 # Image:someimage.jpg|This is some image
4906 preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
4908 if ( count( $matches ) == 0 ) {
4912 if ( strpos( $matches[0], '%' ) !== false ) {
4913 $matches[1] = rawurldecode( $matches[1] );
4915 $title = Title
::newFromText( $matches[1], NS_FILE
);
4916 if ( is_null( $title ) ) {
4917 # Bogus title. Ignore these so we don't bomb out later.
4921 # We need to get what handler the file uses, to figure out parameters.
4922 # Note, a hook can overide the file name, and chose an entirely different
4923 # file (which potentially could be of a different type and have different handler).
4926 Hooks
::run( 'BeforeParserFetchFileAndTitle',
4927 [ $this, $title, &$options, &$descQuery ] );
4928 # Don't register it now, as ImageGallery does that later.
4929 $file = $this->fetchFileNoRegister( $title, $options );
4930 $handler = $file ?
$file->getHandler() : false;
4933 'img_alt' => 'gallery-internal-alt',
4934 'img_link' => 'gallery-internal-link',
4937 $paramMap = $paramMap +
$handler->getParamMap();
4938 // We don't want people to specify per-image widths.
4939 // Additionally the width parameter would need special casing anyhow.
4940 unset( $paramMap['img_width'] );
4943 $mwArray = new MagicWordArray( array_keys( $paramMap ) );
4948 $handlerOptions = [];
4949 if ( isset( $matches[3] ) ) {
4950 // look for an |alt= definition while trying not to break existing
4951 // captions with multiple pipes (|) in it, until a more sensible grammar
4952 // is defined for images in galleries
4954 // FIXME: Doing recursiveTagParse at this stage, and the trim before
4955 // splitting on '|' is a bit odd, and different from makeImage.
4956 $matches[3] = $this->recursiveTagParse( trim( $matches[3] ) );
4957 $parameterMatches = StringUtils
::explode( '|', $matches[3] );
4959 foreach ( $parameterMatches as $parameterMatch ) {
4960 list( $magicName, $match ) = $mwArray->matchVariableStartToEnd( $parameterMatch );
4962 $paramName = $paramMap[$magicName];
4964 switch ( $paramName ) {
4965 case 'gallery-internal-alt':
4966 $alt = $this->stripAltText( $match, false );
4968 case 'gallery-internal-link':
4969 $linkValue = strip_tags( $this->replaceLinkHoldersText( $match ) );
4970 $chars = self
::EXT_LINK_URL_CLASS
;
4971 $addr = self
::EXT_LINK_ADDR
;
4972 $prots = $this->mUrlProtocols
;
4973 // check to see if link matches an absolute url, if not then it must be a wiki link.
4974 if ( preg_match( "/^($prots)$addr$chars*$/u", $linkValue ) ) {
4977 $localLinkTitle = Title
::newFromText( $linkValue );
4978 if ( $localLinkTitle !== null ) {
4979 $link = $localLinkTitle->getLinkURL();
4984 // Must be a handler specific parameter.
4985 if ( $handler->validateParam( $paramName, $match ) ) {
4986 $handlerOptions[$paramName] = $match;
4988 // Guess not, consider it as caption.
4989 wfDebug( "$parameterMatch failed parameter validation\n" );
4990 $label = '|' . $parameterMatch;
4996 $label = '|' . $parameterMatch;
5000 $label = substr( $label, 1 );
5003 $ig->add( $title, $label, $alt, $link, $handlerOptions );
5005 $html = $ig->toHTML();
5006 Hooks
::run( 'AfterParserFetchFileAndTitle', [ $this, $ig, &$html ] );
5011 * @param MediaHandler $handler
5014 public function getImageParams( $handler ) {
5016 $handlerClass = get_class( $handler );
5020 if ( !isset( $this->mImageParams
[$handlerClass] ) ) {
5021 # Initialise static lists
5022 static $internalParamNames = [
5023 'horizAlign' => [ 'left', 'right', 'center', 'none' ],
5024 'vertAlign' => [ 'baseline', 'sub', 'super', 'top', 'text-top', 'middle',
5025 'bottom', 'text-bottom' ],
5026 'frame' => [ 'thumbnail', 'manualthumb', 'framed', 'frameless',
5027 'upright', 'border', 'link', 'alt', 'class' ],
5029 static $internalParamMap;
5030 if ( !$internalParamMap ) {
5031 $internalParamMap = [];
5032 foreach ( $internalParamNames as $type => $names ) {
5033 foreach ( $names as $name ) {
5034 $magicName = str_replace( '-', '_', "img_$name" );
5035 $internalParamMap[$magicName] = [ $type, $name ];
5040 # Add handler params
5041 $paramMap = $internalParamMap;
5043 $handlerParamMap = $handler->getParamMap();
5044 foreach ( $handlerParamMap as $magic => $paramName ) {
5045 $paramMap[$magic] = [ 'handler', $paramName ];
5048 $this->mImageParams
[$handlerClass] = $paramMap;
5049 $this->mImageParamsMagicArray
[$handlerClass] = new MagicWordArray( array_keys( $paramMap ) );
5051 return [ $this->mImageParams
[$handlerClass], $this->mImageParamsMagicArray
[$handlerClass] ];
5055 * Parse image options text and use it to make an image
5057 * @param Title $title
5058 * @param string $options
5059 * @param LinkHolderArray|bool $holders
5060 * @return string HTML
5062 public function makeImage( $title, $options, $holders = false ) {
5063 # Check if the options text is of the form "options|alt text"
5065 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5066 # * left no resizing, just left align. label is used for alt= only
5067 # * right same, but right aligned
5068 # * none same, but not aligned
5069 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5070 # * center center the image
5071 # * frame Keep original image size, no magnify-button.
5072 # * framed Same as "frame"
5073 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5074 # * upright reduce width for upright images, rounded to full __0 px
5075 # * border draw a 1px border around the image
5076 # * alt Text for HTML alt attribute (defaults to empty)
5077 # * class Set a class for img node
5078 # * link Set the target of the image link. Can be external, interwiki, or local
5079 # vertical-align values (no % or length right now):
5089 $parts = StringUtils
::explode( "|", $options );
5091 # Give extensions a chance to select the file revision for us
5094 Hooks
::run( 'BeforeParserFetchFileAndTitle',
5095 [ $this, $title, &$options, &$descQuery ] );
5096 # Fetch and register the file (file title may be different via hooks)
5097 list( $file, $title ) = $this->fetchFileAndTitle( $title, $options );
5100 $handler = $file ?
$file->getHandler() : false;
5102 list( $paramMap, $mwArray ) = $this->getImageParams( $handler );
5105 $this->addTrackingCategory( 'broken-file-category' );
5108 # Process the input parameters
5110 $params = [ 'frame' => [], 'handler' => [],
5111 'horizAlign' => [], 'vertAlign' => [] ];
5112 $seenformat = false;
5113 foreach ( $parts as $part ) {
5114 $part = trim( $part );
5115 list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part );
5117 if ( isset( $paramMap[$magicName] ) ) {
5118 list( $type, $paramName ) = $paramMap[$magicName];
5120 # Special case; width and height come in one variable together
5121 if ( $type === 'handler' && $paramName === 'width' ) {
5122 $parsedWidthParam = $this->parseWidthParam( $value );
5123 if ( isset( $parsedWidthParam['width'] ) ) {
5124 $width = $parsedWidthParam['width'];
5125 if ( $handler->validateParam( 'width', $width ) ) {
5126 $params[$type]['width'] = $width;
5130 if ( isset( $parsedWidthParam['height'] ) ) {
5131 $height = $parsedWidthParam['height'];
5132 if ( $handler->validateParam( 'height', $height ) ) {
5133 $params[$type]['height'] = $height;
5137 # else no validation -- bug 13436
5139 if ( $type === 'handler' ) {
5140 # Validate handler parameter
5141 $validated = $handler->validateParam( $paramName, $value );
5143 # Validate internal parameters
5144 switch ( $paramName ) {
5148 # @todo FIXME: Possibly check validity here for
5149 # manualthumb? downstream behavior seems odd with
5150 # missing manual thumbs.
5152 $value = $this->stripAltText( $value, $holders );
5155 $chars = self
::EXT_LINK_URL_CLASS
;
5156 $addr = self
::EXT_LINK_ADDR
;
5157 $prots = $this->mUrlProtocols
;
5158 if ( $value === '' ) {
5159 $paramName = 'no-link';
5162 } elseif ( preg_match( "/^((?i)$prots)/", $value ) ) {
5163 if ( preg_match( "/^((?i)$prots)$addr$chars*$/u", $value, $m ) ) {
5164 $paramName = 'link-url';
5165 $this->mOutput
->addExternalLink( $value );
5166 if ( $this->mOptions
->getExternalLinkTarget() ) {
5167 $params[$type]['link-target'] = $this->mOptions
->getExternalLinkTarget();
5172 $linkTitle = Title
::newFromText( $value );
5174 $paramName = 'link-title';
5175 $value = $linkTitle;
5176 $this->mOutput
->addLink( $linkTitle );
5184 // use first appearing option, discard others.
5185 $validated = ! $seenformat;
5189 # Most other things appear to be empty or numeric...
5190 $validated = ( $value === false ||
is_numeric( trim( $value ) ) );
5195 $params[$type][$paramName] = $value;
5199 if ( !$validated ) {
5204 # Process alignment parameters
5205 if ( $params['horizAlign'] ) {
5206 $params['frame']['align'] = key( $params['horizAlign'] );
5208 if ( $params['vertAlign'] ) {
5209 $params['frame']['valign'] = key( $params['vertAlign'] );
5212 $params['frame']['caption'] = $caption;
5214 # Will the image be presented in a frame, with the caption below?
5215 $imageIsFramed = isset( $params['frame']['frame'] )
5216 ||
isset( $params['frame']['framed'] )
5217 ||
isset( $params['frame']['thumbnail'] )
5218 ||
isset( $params['frame']['manualthumb'] );
5220 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5221 # came to also set the caption, ordinary text after the image -- which
5222 # makes no sense, because that just repeats the text multiple times in
5223 # screen readers. It *also* came to set the title attribute.
5224 # Now that we have an alt attribute, we should not set the alt text to
5225 # equal the caption: that's worse than useless, it just repeats the
5226 # text. This is the framed/thumbnail case. If there's no caption, we
5227 # use the unnamed parameter for alt text as well, just for the time be-
5228 # ing, if the unnamed param is set and the alt param is not.
5229 # For the future, we need to figure out if we want to tweak this more,
5230 # e.g., introducing a title= parameter for the title; ignoring the un-
5231 # named parameter entirely for images without a caption; adding an ex-
5232 # plicit caption= parameter and preserving the old magic unnamed para-
5234 if ( $imageIsFramed ) { # Framed image
5235 if ( $caption === '' && !isset( $params['frame']['alt'] ) ) {
5236 # No caption or alt text, add the filename as the alt text so
5237 # that screen readers at least get some description of the image
5238 $params['frame']['alt'] = $title->getText();
5240 # Do not set $params['frame']['title'] because tooltips don't make sense
5242 } else { # Inline image
5243 if ( !isset( $params['frame']['alt'] ) ) {
5244 # No alt text, use the "caption" for the alt text
5245 if ( $caption !== '' ) {
5246 $params['frame']['alt'] = $this->stripAltText( $caption, $holders );
5248 # No caption, fall back to using the filename for the
5250 $params['frame']['alt'] = $title->getText();
5253 # Use the "caption" for the tooltip text
5254 $params['frame']['title'] = $this->stripAltText( $caption, $holders );
5257 Hooks
::run( 'ParserMakeImageParams', [ $title, $file, &$params, $this ] );
5259 # Linker does the rest
5260 $time = isset( $options['time'] ) ?
$options['time'] : false;
5261 $ret = Linker
::makeImageLink( $this, $title, $file, $params['frame'], $params['handler'],
5262 $time, $descQuery, $this->mOptions
->getThumbSize() );
5264 # Give the handler a chance to modify the parser object
5266 $handler->parserTransformHook( $this, $file );
5273 * @param string $caption
5274 * @param LinkHolderArray|bool $holders
5275 * @return mixed|string
5277 protected function stripAltText( $caption, $holders ) {
5278 # Strip bad stuff out of the title (tooltip). We can't just use
5279 # replaceLinkHoldersText() here, because if this function is called
5280 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5282 $tooltip = $holders->replaceText( $caption );
5284 $tooltip = $this->replaceLinkHoldersText( $caption );
5287 # make sure there are no placeholders in thumbnail attributes
5288 # that are later expanded to html- so expand them now and
5290 $tooltip = $this->mStripState
->unstripBoth( $tooltip );
5291 $tooltip = Sanitizer
::stripAllTags( $tooltip );
5297 * Set a flag in the output object indicating that the content is dynamic and
5298 * shouldn't be cached.
5300 public function disableCache() {
5301 wfDebug( "Parser output marked as uncacheable.\n" );
5302 if ( !$this->mOutput
) {
5303 throw new MWException( __METHOD__
.
5304 " can only be called when actually parsing something" );
5306 $this->mOutput
->updateCacheExpiry( 0 ); // new style, for consistency
5310 * Callback from the Sanitizer for expanding items found in HTML attribute
5311 * values, so they can be safely tested and escaped.
5313 * @param string $text
5314 * @param bool|PPFrame $frame
5317 public function attributeStripCallback( &$text, $frame = false ) {
5318 $text = $this->replaceVariables( $text, $frame );
5319 $text = $this->mStripState
->unstripBoth( $text );
5328 public function getTags() {
5330 array_keys( $this->mTransparentTagHooks
),
5331 array_keys( $this->mTagHooks
),
5332 array_keys( $this->mFunctionTagHooks
)
5337 * Replace transparent tags in $text with the values given by the callbacks.
5339 * Transparent tag hooks are like regular XML-style tag hooks, except they
5340 * operate late in the transformation sequence, on HTML instead of wikitext.
5342 * @param string $text
5346 public function replaceTransparentTags( $text ) {
5348 $elements = array_keys( $this->mTransparentTagHooks
);
5349 $text = self
::extractTagsAndParams( $elements, $text, $matches );
5352 foreach ( $matches as $marker => $data ) {
5353 list( $element, $content, $params, $tag ) = $data;
5354 $tagName = strtolower( $element );
5355 if ( isset( $this->mTransparentTagHooks
[$tagName] ) ) {
5356 $output = call_user_func_array(
5357 $this->mTransparentTagHooks
[$tagName],
5358 [ $content, $params, $this ]
5363 $replacements[$marker] = $output;
5365 return strtr( $text, $replacements );
5369 * Break wikitext input into sections, and either pull or replace
5370 * some particular section's text.
5372 * External callers should use the getSection and replaceSection methods.
5374 * @param string $text Page wikitext
5375 * @param string|number $sectionId A section identifier string of the form:
5376 * "<flag1> - <flag2> - ... - <section number>"
5378 * Currently the only recognised flag is "T", which means the target section number
5379 * was derived during a template inclusion parse, in other words this is a template
5380 * section edit link. If no flags are given, it was an ordinary section edit link.
5381 * This flag is required to avoid a section numbering mismatch when a section is
5382 * enclosed by "<includeonly>" (bug 6563).
5384 * The section number 0 pulls the text before the first heading; other numbers will
5385 * pull the given section along with its lower-level subsections. If the section is
5386 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5388 * Section 0 is always considered to exist, even if it only contains the empty
5389 * string. If $text is the empty string and section 0 is replaced, $newText is
5392 * @param string $mode One of "get" or "replace"
5393 * @param string $newText Replacement text for section data.
5394 * @return string For "get", the extracted section text.
5395 * for "replace", the whole page with the section replaced.
5397 private function extractSections( $text, $sectionId, $mode, $newText = '' ) {
5398 global $wgTitle; # not generally used but removes an ugly failure mode
5400 $magicScopeVariable = $this->lock();
5401 $this->startParse( $wgTitle, new ParserOptions
, self
::OT_PLAIN
, true );
5403 $frame = $this->getPreprocessor()->newFrame();
5405 # Process section extraction flags
5407 $sectionParts = explode( '-', $sectionId );
5408 $sectionIndex = array_pop( $sectionParts );
5409 foreach ( $sectionParts as $part ) {
5410 if ( $part === 'T' ) {
5411 $flags |
= self
::PTD_FOR_INCLUSION
;
5415 # Check for empty input
5416 if ( strval( $text ) === '' ) {
5417 # Only sections 0 and T-0 exist in an empty document
5418 if ( $sectionIndex == 0 ) {
5419 if ( $mode === 'get' ) {
5425 if ( $mode === 'get' ) {
5433 # Preprocess the text
5434 $root = $this->preprocessToDom( $text, $flags );
5436 # <h> nodes indicate section breaks
5437 # They can only occur at the top level, so we can find them by iterating the root's children
5438 $node = $root->getFirstChild();
5440 # Find the target section
5441 if ( $sectionIndex == 0 ) {
5442 # Section zero doesn't nest, level=big
5443 $targetLevel = 1000;
5446 if ( $node->getName() === 'h' ) {
5447 $bits = $node->splitHeading();
5448 if ( $bits['i'] == $sectionIndex ) {
5449 $targetLevel = $bits['level'];
5453 if ( $mode === 'replace' ) {
5454 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5456 $node = $node->getNextSibling();
5462 if ( $mode === 'get' ) {
5469 # Find the end of the section, including nested sections
5471 if ( $node->getName() === 'h' ) {
5472 $bits = $node->splitHeading();
5473 $curLevel = $bits['level'];
5474 if ( $bits['i'] != $sectionIndex && $curLevel <= $targetLevel ) {
5478 if ( $mode === 'get' ) {
5479 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5481 $node = $node->getNextSibling();
5484 # Write out the remainder (in replace mode only)
5485 if ( $mode === 'replace' ) {
5486 # Output the replacement text
5487 # Add two newlines on -- trailing whitespace in $newText is conventionally
5488 # stripped by the editor, so we need both newlines to restore the paragraph gap
5489 # Only add trailing whitespace if there is newText
5490 if ( $newText != "" ) {
5491 $outText .= $newText . "\n\n";
5495 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5496 $node = $node->getNextSibling();
5500 if ( is_string( $outText ) ) {
5501 # Re-insert stripped tags
5502 $outText = rtrim( $this->mStripState
->unstripBoth( $outText ) );
5509 * This function returns the text of a section, specified by a number ($section).
5510 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5511 * the first section before any such heading (section 0).
5513 * If a section contains subsections, these are also returned.
5515 * @param string $text Text to look in
5516 * @param string|number $sectionId Section identifier as a number or string
5517 * (e.g. 0, 1 or 'T-1').
5518 * @param string $defaultText Default to return if section is not found
5520 * @return string Text of the requested section
5522 public function getSection( $text, $sectionId, $defaultText = '' ) {
5523 return $this->extractSections( $text, $sectionId, 'get', $defaultText );
5527 * This function returns $oldtext after the content of the section
5528 * specified by $section has been replaced with $text. If the target
5529 * section does not exist, $oldtext is returned unchanged.
5531 * @param string $oldText Former text of the article
5532 * @param string|number $sectionId Section identifier as a number or string
5533 * (e.g. 0, 1 or 'T-1').
5534 * @param string $newText Replacing text
5536 * @return string Modified text
5538 public function replaceSection( $oldText, $sectionId, $newText ) {
5539 return $this->extractSections( $oldText, $sectionId, 'replace', $newText );
5543 * Get the ID of the revision we are parsing
5547 public function getRevisionId() {
5548 return $this->mRevisionId
;
5552 * Get the revision object for $this->mRevisionId
5554 * @return Revision|null Either a Revision object or null
5555 * @since 1.23 (public since 1.23)
5557 public function getRevisionObject() {
5558 if ( !is_null( $this->mRevisionObject
) ) {
5559 return $this->mRevisionObject
;
5561 if ( is_null( $this->mRevisionId
) ) {
5565 $rev = call_user_func(
5566 $this->mOptions
->getCurrentRevisionCallback(), $this->getTitle(), $this
5569 # If the parse is for a new revision, then the callback should have
5570 # already been set to force the object and should match mRevisionId.
5571 # If not, try to fetch by mRevisionId for sanity.
5572 if ( $rev && $rev->getId() != $this->mRevisionId
) {
5573 $rev = Revision
::newFromId( $this->mRevisionId
);
5576 $this->mRevisionObject
= $rev;
5578 return $this->mRevisionObject
;
5582 * Get the timestamp associated with the current revision, adjusted for
5583 * the default server-local timestamp
5586 public function getRevisionTimestamp() {
5587 if ( is_null( $this->mRevisionTimestamp
) ) {
5590 $revObject = $this->getRevisionObject();
5591 $timestamp = $revObject ?
$revObject->getTimestamp() : wfTimestampNow();
5593 # The cryptic '' timezone parameter tells to use the site-default
5594 # timezone offset instead of the user settings.
5595 # Since this value will be saved into the parser cache, served
5596 # to other users, and potentially even used inside links and such,
5597 # it needs to be consistent for all visitors.
5598 $this->mRevisionTimestamp
= $wgContLang->userAdjust( $timestamp, '' );
5601 return $this->mRevisionTimestamp
;
5605 * Get the name of the user that edited the last revision
5607 * @return string User name
5609 public function getRevisionUser() {
5610 if ( is_null( $this->mRevisionUser
) ) {
5611 $revObject = $this->getRevisionObject();
5613 # if this template is subst: the revision id will be blank,
5614 # so just use the current user's name
5616 $this->mRevisionUser
= $revObject->getUserText();
5617 } elseif ( $this->ot
['wiki'] ||
$this->mOptions
->getIsPreview() ) {
5618 $this->mRevisionUser
= $this->getUser()->getName();
5621 return $this->mRevisionUser
;
5625 * Get the size of the revision
5627 * @return int|null Revision size
5629 public function getRevisionSize() {
5630 if ( is_null( $this->mRevisionSize
) ) {
5631 $revObject = $this->getRevisionObject();
5633 # if this variable is subst: the revision id will be blank,
5634 # so just use the parser input size, because the own substituation
5635 # will change the size.
5637 $this->mRevisionSize
= $revObject->getSize();
5639 $this->mRevisionSize
= $this->mInputSize
;
5642 return $this->mRevisionSize
;
5646 * Mutator for $mDefaultSort
5648 * @param string $sort New value
5650 public function setDefaultSort( $sort ) {
5651 $this->mDefaultSort
= $sort;
5652 $this->mOutput
->setProperty( 'defaultsort', $sort );
5656 * Accessor for $mDefaultSort
5657 * Will use the empty string if none is set.
5659 * This value is treated as a prefix, so the
5660 * empty string is equivalent to sorting by
5665 public function getDefaultSort() {
5666 if ( $this->mDefaultSort
!== false ) {
5667 return $this->mDefaultSort
;
5674 * Accessor for $mDefaultSort
5675 * Unlike getDefaultSort(), will return false if none is set
5677 * @return string|bool
5679 public function getCustomDefaultSort() {
5680 return $this->mDefaultSort
;
5684 * Try to guess the section anchor name based on a wikitext fragment
5685 * presumably extracted from a heading, for example "Header" from
5688 * @param string $text
5692 public function guessSectionNameFromWikiText( $text ) {
5693 # Strip out wikitext links(they break the anchor)
5694 $text = $this->stripSectionName( $text );
5695 $text = Sanitizer
::normalizeSectionNameWhitespace( $text );
5696 return '#' . Sanitizer
::escapeId( $text, 'noninitial' );
5700 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5701 * instead. For use in redirects, since IE6 interprets Redirect: headers
5702 * as something other than UTF-8 (apparently?), resulting in breakage.
5704 * @param string $text The section name
5705 * @return string An anchor
5707 public function guessLegacySectionNameFromWikiText( $text ) {
5708 # Strip out wikitext links(they break the anchor)
5709 $text = $this->stripSectionName( $text );
5710 $text = Sanitizer
::normalizeSectionNameWhitespace( $text );
5711 return '#' . Sanitizer
::escapeId( $text, [ 'noninitial', 'legacy' ] );
5715 * Strips a text string of wikitext for use in a section anchor
5717 * Accepts a text string and then removes all wikitext from the
5718 * string and leaves only the resultant text (i.e. the result of
5719 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5720 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5721 * to create valid section anchors by mimicing the output of the
5722 * parser when headings are parsed.
5724 * @param string $text Text string to be stripped of wikitext
5725 * for use in a Section anchor
5726 * @return string Filtered text string
5728 public function stripSectionName( $text ) {
5729 # Strip internal link markup
5730 $text = preg_replace( '/\[\[:?([^[|]+)\|([^[]+)\]\]/', '$2', $text );
5731 $text = preg_replace( '/\[\[:?([^[]+)\|?\]\]/', '$1', $text );
5733 # Strip external link markup
5734 # @todo FIXME: Not tolerant to blank link text
5735 # I.E. [https://www.mediawiki.org] will render as [1] or something depending
5736 # on how many empty links there are on the page - need to figure that out.
5737 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols
. ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5739 # Parse wikitext quotes (italics & bold)
5740 $text = $this->doQuotes( $text );
5743 $text = StringUtils
::delimiterReplace( '<', '>', '', $text );
5748 * strip/replaceVariables/unstrip for preprocessor regression testing
5750 * @param string $text
5751 * @param Title $title
5752 * @param ParserOptions $options
5753 * @param int $outputType
5757 public function testSrvus( $text, Title
$title, ParserOptions
$options,
5758 $outputType = self
::OT_HTML
5760 $magicScopeVariable = $this->lock();
5761 $this->startParse( $title, $options, $outputType, true );
5763 $text = $this->replaceVariables( $text );
5764 $text = $this->mStripState
->unstripBoth( $text );
5765 $text = Sanitizer
::removeHTMLtags( $text );
5770 * @param string $text
5771 * @param Title $title
5772 * @param ParserOptions $options
5775 public function testPst( $text, Title
$title, ParserOptions
$options ) {
5776 return $this->preSaveTransform( $text, $title, $options->getUser(), $options );
5780 * @param string $text
5781 * @param Title $title
5782 * @param ParserOptions $options
5785 public function testPreprocess( $text, Title
$title, ParserOptions
$options ) {
5786 return $this->testSrvus( $text, $title, $options, self
::OT_PREPROCESS
);
5790 * Call a callback function on all regions of the given text that are not
5791 * inside strip markers, and replace those regions with the return value
5792 * of the callback. For example, with input:
5796 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
5797 * two strings will be replaced with the value returned by the callback in
5801 * @param callable $callback
5805 public function markerSkipCallback( $s, $callback ) {
5808 while ( $i < strlen( $s ) ) {
5809 $markerStart = strpos( $s, self
::MARKER_PREFIX
, $i );
5810 if ( $markerStart === false ) {
5811 $out .= call_user_func( $callback, substr( $s, $i ) );
5814 $out .= call_user_func( $callback, substr( $s, $i, $markerStart - $i ) );
5815 $markerEnd = strpos( $s, self
::MARKER_SUFFIX
, $markerStart );
5816 if ( $markerEnd === false ) {
5817 $out .= substr( $s, $markerStart );
5820 $markerEnd +
= strlen( self
::MARKER_SUFFIX
);
5821 $out .= substr( $s, $markerStart, $markerEnd - $markerStart );
5830 * Remove any strip markers found in the given text.
5832 * @param string $text Input string
5835 public function killMarkers( $text ) {
5836 return $this->mStripState
->killMarkers( $text );
5840 * Save the parser state required to convert the given half-parsed text to
5841 * HTML. "Half-parsed" in this context means the output of
5842 * recursiveTagParse() or internalParse(). This output has strip markers
5843 * from replaceVariables (extensionSubstitution() etc.), and link
5844 * placeholders from replaceLinkHolders().
5846 * Returns an array which can be serialized and stored persistently. This
5847 * array can later be loaded into another parser instance with
5848 * unserializeHalfParsedText(). The text can then be safely incorporated into
5849 * the return value of a parser hook.
5851 * @param string $text
5855 public function serializeHalfParsedText( $text ) {
5858 'version' => self
::HALF_PARSED_VERSION
,
5859 'stripState' => $this->mStripState
->getSubState( $text ),
5860 'linkHolders' => $this->mLinkHolders
->getSubArray( $text )
5866 * Load the parser state given in the $data array, which is assumed to
5867 * have been generated by serializeHalfParsedText(). The text contents is
5868 * extracted from the array, and its markers are transformed into markers
5869 * appropriate for the current Parser instance. This transformed text is
5870 * returned, and can be safely included in the return value of a parser
5873 * If the $data array has been stored persistently, the caller should first
5874 * check whether it is still valid, by calling isValidHalfParsedText().
5876 * @param array $data Serialized data
5877 * @throws MWException
5880 public function unserializeHalfParsedText( $data ) {
5881 if ( !isset( $data['version'] ) ||
$data['version'] != self
::HALF_PARSED_VERSION
) {
5882 throw new MWException( __METHOD__
. ': invalid version' );
5885 # First, extract the strip state.
5886 $texts = [ $data['text'] ];
5887 $texts = $this->mStripState
->merge( $data['stripState'], $texts );
5889 # Now renumber links
5890 $texts = $this->mLinkHolders
->mergeForeign( $data['linkHolders'], $texts );
5892 # Should be good to go.
5897 * Returns true if the given array, presumed to be generated by
5898 * serializeHalfParsedText(), is compatible with the current version of the
5901 * @param array $data
5905 public function isValidHalfParsedText( $data ) {
5906 return isset( $data['version'] ) && $data['version'] == self
::HALF_PARSED_VERSION
;
5910 * Parsed a width param of imagelink like 300px or 200x300px
5912 * @param string $value
5917 public function parseWidthParam( $value ) {
5918 $parsedWidthParam = [];
5919 if ( $value === '' ) {
5920 return $parsedWidthParam;
5923 # (bug 13500) In both cases (width/height and width only),
5924 # permit trailing "px" for backward compatibility.
5925 if ( preg_match( '/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/', $value, $m ) ) {
5926 $width = intval( $m[1] );
5927 $height = intval( $m[2] );
5928 $parsedWidthParam['width'] = $width;
5929 $parsedWidthParam['height'] = $height;
5930 } elseif ( preg_match( '/^[0-9]*\s*(?:px)?\s*$/', $value ) ) {
5931 $width = intval( $value );
5932 $parsedWidthParam['width'] = $width;
5934 return $parsedWidthParam;
5938 * Lock the current instance of the parser.
5940 * This is meant to stop someone from calling the parser
5941 * recursively and messing up all the strip state.
5943 * @throws MWException If parser is in a parse
5944 * @return ScopedCallback The lock will be released once the return value goes out of scope.
5946 protected function lock() {
5947 if ( $this->mInParse
) {
5948 throw new MWException( "Parser state cleared while parsing. "
5949 . "Did you call Parser::parse recursively?" );
5951 $this->mInParse
= true;
5953 $recursiveCheck = new ScopedCallback( function() {
5954 $this->mInParse
= false;
5957 return $recursiveCheck;
5961 * Strip outer <p></p> tag from the HTML source of a single paragraph.
5963 * Returns original HTML if the <p/> tag has any attributes, if there's no wrapping <p/> tag,
5964 * or if there is more than one <p/> tag in the input HTML.
5966 * @param string $html
5970 public static function stripOuterParagraph( $html ) {
5972 if ( preg_match( '/^<p>(.*)\n?<\/p>\n?$/sU', $html, $m ) ) {
5973 if ( strpos( $m[1], '</p>' ) === false ) {
5982 * Return this parser if it is not doing anything, otherwise
5983 * get a fresh parser. You can use this method by doing
5984 * $myParser = $wgParser->getFreshParser(), or more simply
5985 * $wgParser->getFreshParser()->parse( ... );
5986 * if you're unsure if $wgParser is safe to use.
5989 * @return Parser A parser object that is not parsing anything
5991 public function getFreshParser() {
5992 global $wgParserConf;
5993 if ( $this->mInParse
) {
5994 return new $wgParserConf['class']( $wgParserConf );
6001 * Set's up the PHP implementation of OOUI for use in this request
6002 * and instructs OutputPage to enable OOUI for itself.
6006 public function enableOOUI() {
6007 OutputPage
::setupOOUI();
6008 $this->mOutput
->setEnableOOUI( true );