3 * @defgroup Parser Parser
7 * File for Parser and related classes
12 * PHP Parser - Processes wiki markup (which uses a more user-friendly
13 * syntax, such as "[[link]]" for making links), and provides a one-way
14 * transformation of that wiki markup it into XHTML output / markup
15 * (which in turn the browser understands, and can display).
18 * There are five main entry points into the Parser class:
20 * produces HTML output
22 * produces altered wiki markup.
24 * removes HTML comments and expands templates
25 * cleanSig() / cleanSigInSig()
26 * Cleans a signature before saving it to preferences
28 * Return the content of a section from an article for section editing
30 * Replaces a section by number inside an article
32 * Removes <noinclude> sections, and <includeonly> tags.
35 * objects: $wgLang, $wgContLang
37 * NOT $wgUser or $wgTitle. Keep them away!
40 * $wgUseDynamicDates*, $wgInterwikiMagic*,
41 * $wgNamespacesWithSubpages, $wgAllowExternalImages*,
42 * $wgLocaltimezone, $wgAllowSpecialInclusion*,
45 * * only within ParserOptions
52 * Update this version number when the ParserOutput format
53 * changes in an incompatible way, so the parser cache
54 * can automatically discard old data.
56 const VERSION
= '1.6.4';
59 * Update this version number when the output of serialiseHalfParsedText()
60 * changes in an incompatible way
62 const HALF_PARSED_VERSION
= 2;
64 # Flags for Parser::setFunctionHook
65 # Also available as global constants from Defines.php
66 const SFH_NO_HASH
= 1;
67 const SFH_OBJECT_ARGS
= 2;
69 # Constants needed for external link processing
70 # Everything except bracket, space, or control characters
71 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
72 # as well as U+3000 is IDEOGRAPHIC SPACE for bug 19052
73 const EXT_LINK_URL_CLASS
= '[^][<>"\\x00-\\x20\\x7F\p{Zs}]';
74 const EXT_IMAGE_REGEX
= '/^(http:\/\/|https:\/\/)([^][<>"\\x00-\\x20\\x7F\p{Zs}]+)
75 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
77 # State constants for the definition list colon extraction
78 const COLON_STATE_TEXT
= 0;
79 const COLON_STATE_TAG
= 1;
80 const COLON_STATE_TAGSTART
= 2;
81 const COLON_STATE_CLOSETAG
= 3;
82 const COLON_STATE_TAGSLASH
= 4;
83 const COLON_STATE_COMMENT
= 5;
84 const COLON_STATE_COMMENTDASH
= 6;
85 const COLON_STATE_COMMENTDASHDASH
= 7;
87 # Flags for preprocessToDom
88 const PTD_FOR_INCLUSION
= 1;
90 # Allowed values for $this->mOutputType
91 # Parameter to startExternalParse().
92 const OT_HTML
= 1; # like parse()
93 const OT_WIKI
= 2; # like preSaveTransform()
94 const OT_PREPROCESS
= 3; # like preprocess()
96 const OT_PLAIN
= 4; # like extractSections() - portions of the original are returned unchanged.
98 # Marker Suffix needs to be accessible staticly.
99 const MARKER_SUFFIX
= "-QINU\x7f";
102 var $mTagHooks = array();
103 var $mTransparentTagHooks = array();
104 var $mFunctionHooks = array();
105 var $mFunctionSynonyms = array( 0 => array(), 1 => array() );
106 var $mFunctionTagHooks = array();
107 var $mStripList = array();
108 var $mDefaultStripList = array();
109 var $mVarCache = array();
110 var $mImageParams = array();
111 var $mImageParamsMagicArray = array();
112 var $mMarkerIndex = 0;
113 var $mFirstCall = true;
115 # Initialised by initialiseVariables()
118 * @var MagicWordArray
123 * @var MagicWordArray
126 var $mConf, $mPreprocessor, $mExtLinkBracketedRegex, $mUrlProtocols; # Initialised in constructor
128 # Cleared with clearState():
133 var $mAutonumber, $mDTopen;
140 var $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
142 * @var LinkHolderArray
147 var $mIncludeSizes, $mPPNodeCount, $mDefaultSort;
148 var $mTplExpandCache; # empty-frame expansion cache
149 var $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
150 var $mExpensiveFunctionCount; # number of expensive parser function calls
151 var $mShowToc, $mForceTocPosition;
156 var $mUser; # User object; only used when doing pre-save transform
159 # These are variables reset at least once per parse regardless of $clearState
169 var $mTitle; # Title context, used for self-link rendering and similar things
170 var $mOutputType; # Output type, one of the OT_xxx constants
171 var $ot; # Shortcut alias, see setOutputType()
172 var $mRevisionObject; # The revision object of the specified revision ID
173 var $mRevisionId; # ID to display in {{REVISIONID}} tags
174 var $mRevisionTimestamp; # The timestamp of the specified revision ID
175 var $mRevisionUser; # User to display in {{REVISIONUSER}} tag
176 var $mRevIdForTs; # The revision ID which was used to fetch the timestamp
188 public function __construct( $conf = array() ) {
189 $this->mConf
= $conf;
190 $this->mUrlProtocols
= wfUrlProtocols();
191 $this->mExtLinkBracketedRegex
= '/\[((' . wfUrlProtocols() . ')'.
192 self
::EXT_LINK_URL_CLASS
.'+)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F]*?)\]/Su';
193 if ( isset( $conf['preprocessorClass'] ) ) {
194 $this->mPreprocessorClass
= $conf['preprocessorClass'];
195 } elseif ( defined( 'MW_COMPILED' ) ) {
196 # Preprocessor_Hash is much faster than Preprocessor_DOM in compiled mode
197 $this->mPreprocessorClass
= 'Preprocessor_Hash';
198 } elseif ( extension_loaded( 'domxml' ) ) {
199 # PECL extension that conflicts with the core DOM extension (bug 13770)
200 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
201 $this->mPreprocessorClass
= 'Preprocessor_Hash';
202 } elseif ( extension_loaded( 'dom' ) ) {
203 $this->mPreprocessorClass
= 'Preprocessor_DOM';
205 $this->mPreprocessorClass
= 'Preprocessor_Hash';
207 wfDebug( __CLASS__
. ": using preprocessor: {$this->mPreprocessorClass}\n" );
211 * Reduce memory usage to reduce the impact of circular references
213 function __destruct() {
214 if ( isset( $this->mLinkHolders
) ) {
215 unset( $this->mLinkHolders
);
217 foreach ( $this as $name => $value ) {
218 unset( $this->$name );
223 * Do various kinds of initialisation on the first call of the parser
225 function firstCallInit() {
226 if ( !$this->mFirstCall
) {
229 $this->mFirstCall
= false;
231 wfProfileIn( __METHOD__
);
233 CoreParserFunctions
::register( $this );
234 CoreTagHooks
::register( $this );
235 $this->initialiseVariables();
237 wfRunHooks( 'ParserFirstCallInit', array( &$this ) );
238 wfProfileOut( __METHOD__
);
246 function clearState() {
247 wfProfileIn( __METHOD__
);
248 if ( $this->mFirstCall
) {
249 $this->firstCallInit();
251 $this->mOutput
= new ParserOutput
;
252 $this->mOptions
->registerWatcher( array( $this->mOutput
, 'recordOption' ) );
253 $this->mAutonumber
= 0;
254 $this->mLastSection
= '';
255 $this->mDTopen
= false;
256 $this->mIncludeCount
= array();
257 $this->mArgStack
= false;
258 $this->mInPre
= false;
259 $this->mLinkHolders
= new LinkHolderArray( $this );
261 $this->mRevisionObject
= $this->mRevisionTimestamp
=
262 $this->mRevisionId
= $this->mRevisionUser
= null;
263 $this->mVarCache
= array();
267 * Prefix for temporary replacement strings for the multipass parser.
268 * \x07 should never appear in input as it's disallowed in XML.
269 * Using it at the front also gives us a little extra robustness
270 * since it shouldn't match when butted up against identifier-like
273 * Must not consist of all title characters, or else it will change
274 * the behaviour of <nowiki> in a link.
276 # $this->mUniqPrefix = "\x07UNIQ" . Parser::getRandomString();
277 # Changed to \x7f to allow XML double-parsing -- TS
278 $this->mUniqPrefix
= "\x7fUNIQ" . self
::getRandomString();
279 $this->mStripState
= new StripState( $this->mUniqPrefix
);
282 # Clear these on every parse, bug 4549
283 $this->mTplExpandCache
= $this->mTplRedirCache
= $this->mTplDomCache
= array();
285 $this->mShowToc
= true;
286 $this->mForceTocPosition
= false;
287 $this->mIncludeSizes
= array(
291 $this->mPPNodeCount
= 0;
292 $this->mDefaultSort
= false;
293 $this->mHeadings
= array();
294 $this->mDoubleUnderscores
= array();
295 $this->mExpensiveFunctionCount
= 0;
298 if ( isset( $this->mPreprocessor
) && $this->mPreprocessor
->parser
!== $this ) {
299 $this->mPreprocessor
= null;
302 wfRunHooks( 'ParserClearState', array( &$this ) );
303 wfProfileOut( __METHOD__
);
307 * Convert wikitext to HTML
308 * Do not call this function recursively.
310 * @param $text String: text we want to parse
311 * @param $title Title object
312 * @param $options ParserOptions
313 * @param $linestart boolean
314 * @param $clearState boolean
315 * @param $revid Int: number to pass in {{REVISIONID}}
316 * @return ParserOutput a ParserOutput
318 public function parse( $text, Title
$title, ParserOptions
$options, $linestart = true, $clearState = true, $revid = null ) {
320 * First pass--just handle <nowiki> sections, pass the rest off
321 * to internalParse() which does all the real work.
324 global $wgUseTidy, $wgAlwaysUseTidy, $wgDisableLangConversion, $wgDisableTitleConversion;
325 $fname = __METHOD__
.'-' . wfGetCaller();
326 wfProfileIn( __METHOD__
);
327 wfProfileIn( $fname );
329 $this->startParse( $title, $options, self
::OT_HTML
, $clearState );
331 $oldRevisionId = $this->mRevisionId
;
332 $oldRevisionObject = $this->mRevisionObject
;
333 $oldRevisionTimestamp = $this->mRevisionTimestamp
;
334 $oldRevisionUser = $this->mRevisionUser
;
335 if ( $revid !== null ) {
336 $this->mRevisionId
= $revid;
337 $this->mRevisionObject
= null;
338 $this->mRevisionTimestamp
= null;
339 $this->mRevisionUser
= null;
342 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState
) );
344 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState
) );
345 $text = $this->internalParse( $text );
347 $text = $this->mStripState
->unstripGeneral( $text );
349 # Clean up special characters, only run once, next-to-last before doBlockLevels
351 # french spaces, last one Guillemet-left
352 # only if there is something before the space
353 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 ',
354 # french spaces, Guillemet-right
355 '/(\\302\\253) /' => '\\1 ',
356 '/ (!\s*important)/' => ' \\1', # Beware of CSS magic word !important, bug #11874.
358 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
360 $text = $this->doBlockLevels( $text, $linestart );
362 $this->replaceLinkHolders( $text );
365 * The input doesn't get language converted if
367 * b) Content isn't converted
368 * c) It's a conversion table
369 * d) it is an interface message (which is in the user language)
371 if ( !( $wgDisableLangConversion
372 ||
isset( $this->mDoubleUnderscores
['nocontentconvert'] )
373 ||
$this->mTitle
->isConversionTable()
374 ||
$this->mOptions
->getInterfaceMessage() ) ) {
376 # The position of the convert() call should not be changed. it
377 # assumes that the links are all replaced and the only thing left
378 # is the <nowiki> mark.
379 $text = $this->getFunctionLang()->convert( $text );
383 * A converted title will be provided in the output object if title and
384 * content conversion are enabled, the article text does not contain
385 * a conversion-suppressing double-underscore tag, and no
386 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
387 * automatic link conversion.
389 if ( !( $wgDisableLangConversion
390 ||
$wgDisableTitleConversion
391 ||
isset( $this->mDoubleUnderscores
['nocontentconvert'] )
392 ||
isset( $this->mDoubleUnderscores
['notitleconvert'] )
393 ||
$this->mOutput
->getDisplayTitle() !== false ) )
395 $convruletitle = $this->getFunctionLang()->getConvRuleTitle();
396 if ( $convruletitle ) {
397 $this->mOutput
->setTitleText( $convruletitle );
399 $titleText = $this->getFunctionLang()->convertTitle( $title );
400 $this->mOutput
->setTitleText( $titleText );
404 $text = $this->mStripState
->unstripNoWiki( $text );
406 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) );
408 $text = $this->replaceTransparentTags( $text );
409 $text = $this->mStripState
->unstripGeneral( $text );
411 $text = Sanitizer
::normalizeCharReferences( $text );
413 if ( ( $wgUseTidy && $this->mOptions
->getTidy() ) ||
$wgAlwaysUseTidy ) {
414 $text = MWTidy
::tidy( $text );
416 # attempt to sanitize at least some nesting problems
417 # (bug #2702 and quite a few others)
419 # ''Something [http://www.cool.com cool''] -->
420 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
421 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
422 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
423 # fix up an anchor inside another anchor, only
424 # at least for a single single nested link (bug 3695)
425 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
426 '\\1\\2</a>\\3</a>\\1\\4</a>',
427 # fix div inside inline elements- doBlockLevels won't wrap a line which
428 # contains a div, so fix it up here; replace
429 # div with escaped text
430 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
431 '\\1\\3<div\\5>\\6</div>\\8\\9',
432 # remove empty italic or bold tag pairs, some
433 # introduced by rules above
434 '/<([bi])><\/\\1>/' => '',
437 $text = preg_replace(
438 array_keys( $tidyregs ),
439 array_values( $tidyregs ),
442 global $wgExpensiveParserFunctionLimit;
443 if ( $this->mExpensiveFunctionCount
> $wgExpensiveParserFunctionLimit ) {
444 $this->limitationWarn( 'expensive-parserfunction', $this->mExpensiveFunctionCount
, $wgExpensiveParserFunctionLimit );
447 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) );
449 # Information on include size limits, for the benefit of users who try to skirt them
450 if ( $this->mOptions
->getEnableLimitReport() ) {
451 $max = $this->mOptions
->getMaxIncludeSize();
452 $PFreport = "Expensive parser function count: {$this->mExpensiveFunctionCount}/$wgExpensiveParserFunctionLimit\n";
454 "NewPP limit report\n" .
455 "Preprocessor node count: {$this->mPPNodeCount}/{$this->mOptions->getMaxPPNodeCount()}\n" .
456 "Post-expand include size: {$this->mIncludeSizes['post-expand']}/$max bytes\n" .
457 "Template argument size: {$this->mIncludeSizes['arg']}/$max bytes\n".
459 wfRunHooks( 'ParserLimitReport', array( $this, &$limitReport ) );
460 $text .= "\n<!-- \n$limitReport-->\n";
462 $this->mOutput
->setText( $text );
464 $this->mRevisionId
= $oldRevisionId;
465 $this->mRevisionObject
= $oldRevisionObject;
466 $this->mRevisionTimestamp
= $oldRevisionTimestamp;
467 $this->mRevisionUser
= $oldRevisionUser;
468 wfProfileOut( $fname );
469 wfProfileOut( __METHOD__
);
471 return $this->mOutput
;
475 * Recursive parser entry point that can be called from an extension tag
478 * If $frame is not provided, then template variables (e.g., {{{1}}}) within $text are not expanded
480 * @param $text String: text extension wants to have parsed
481 * @param $frame PPFrame: The frame to use for expanding any template variables
485 function recursiveTagParse( $text, $frame=false ) {
486 wfProfileIn( __METHOD__
);
487 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState
) );
488 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState
) );
489 $text = $this->internalParse( $text, false, $frame );
490 wfProfileOut( __METHOD__
);
495 * Expand templates and variables in the text, producing valid, static wikitext.
496 * Also removes comments.
498 function preprocess( $text, Title
$title, ParserOptions
$options, $revid = null ) {
499 wfProfileIn( __METHOD__
);
500 $this->startParse( $title, $options, self
::OT_PREPROCESS
, true );
501 if ( $revid !== null ) {
502 $this->mRevisionId
= $revid;
504 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState
) );
505 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState
) );
506 $text = $this->replaceVariables( $text );
507 $text = $this->mStripState
->unstripBoth( $text );
508 wfProfileOut( __METHOD__
);
513 * Recursive parser entry point that can be called from an extension tag
516 * @param $text String: text to be expanded
517 * @param $frame PPFrame: The frame to use for expanding any template variables
520 public function recursivePreprocess( $text, $frame = false ) {
521 wfProfileIn( __METHOD__
);
522 $text = $this->replaceVariables( $text, $frame );
523 $text = $this->mStripState
->unstripBoth( $text );
524 wfProfileOut( __METHOD__
);
529 * Process the wikitext for the ?preload= feature. (bug 5210)
531 * <noinclude>, <includeonly> etc. are parsed as for template transclusion,
532 * comments, templates, arguments, tags hooks and parser functions are untouched.
534 * @param $text String
535 * @param $title Title
536 * @param $options ParserOptions
539 public function getPreloadText( $text, Title
$title, ParserOptions
$options ) {
540 # Parser (re)initialisation
541 $this->startParse( $title, $options, self
::OT_PLAIN
, true );
543 $flags = PPFrame
::NO_ARGS | PPFrame
::NO_TEMPLATES
;
544 $dom = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
545 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
546 $text = $this->mStripState
->unstripBoth( $text );
551 * Get a random string
555 static public function getRandomString() {
556 return dechex( mt_rand( 0, 0x7fffffff ) ) . dechex( mt_rand( 0, 0x7fffffff ) );
560 * Set the current user.
561 * Should only be used when doing pre-save transform.
563 * @param $user Mixed: User object or null (to reset)
565 function setUser( $user ) {
566 $this->mUser
= $user;
570 * Accessor for mUniqPrefix.
574 public function uniqPrefix() {
575 if ( !isset( $this->mUniqPrefix
) ) {
576 # @todo FIXME: This is probably *horribly wrong*
577 # LanguageConverter seems to want $wgParser's uniqPrefix, however
578 # if this is called for a parser cache hit, the parser may not
579 # have ever been initialized in the first place.
580 # Not really sure what the heck is supposed to be going on here.
582 # throw new MWException( "Accessing uninitialized mUniqPrefix" );
584 return $this->mUniqPrefix
;
588 * Set the context title
592 function setTitle( $t ) {
593 if ( !$t ||
$t instanceof FakeTitle
) {
594 $t = Title
::newFromText( 'NO TITLE' );
597 if ( strval( $t->getFragment() ) !== '' ) {
598 # Strip the fragment to avoid various odd effects
599 $this->mTitle
= clone $t;
600 $this->mTitle
->setFragment( '' );
607 * Accessor for the Title object
609 * @return Title object
611 function getTitle() {
612 return $this->mTitle
;
616 * Accessor/mutator for the Title object
618 * @param $x New Title object or null to just get the current one
619 * @return Title object
621 function Title( $x = null ) {
622 return wfSetVar( $this->mTitle
, $x );
626 * Set the output type
628 * @param $ot Integer: new value
630 function setOutputType( $ot ) {
631 $this->mOutputType
= $ot;
634 'html' => $ot == self
::OT_HTML
,
635 'wiki' => $ot == self
::OT_WIKI
,
636 'pre' => $ot == self
::OT_PREPROCESS
,
637 'plain' => $ot == self
::OT_PLAIN
,
642 * Accessor/mutator for the output type
644 * @param $x New value or null to just get the current one
647 function OutputType( $x = null ) {
648 return wfSetVar( $this->mOutputType
, $x );
652 * Get the ParserOutput object
654 * @return ParserOutput object
656 function getOutput() {
657 return $this->mOutput
;
661 * Get the ParserOptions object
663 * @return ParserOptions object
665 function getOptions() {
666 return $this->mOptions
;
670 * Accessor/mutator for the ParserOptions object
672 * @param $x New value or null to just get the current one
673 * @return Current ParserOptions object
675 function Options( $x = null ) {
676 return wfSetVar( $this->mOptions
, $x );
682 function nextLinkID() {
683 return $this->mLinkID++
;
689 function setLinkID( $id ) {
690 $this->mLinkID
= $id;
696 function getFunctionLang() {
697 $target = $this->mOptions
->getTargetLanguage();
698 if ( $target !== null ) {
700 } elseif( $this->mOptions
->getInterfaceMessage() ) {
703 } elseif( is_null( $this->mTitle
) ) {
704 throw new MWException( __METHOD__
.': $this->mTitle is null' );
706 return $this->mTitle
->getPageLanguage();
710 * Get a User object either from $this->mUser, if set, or from the
711 * ParserOptions object otherwise
713 * @return User object
716 if ( !is_null( $this->mUser
) ) {
719 return $this->mOptions
->getUser();
723 * Get a preprocessor object
725 * @return Preprocessor instance
727 function getPreprocessor() {
728 if ( !isset( $this->mPreprocessor
) ) {
729 $class = $this->mPreprocessorClass
;
730 $this->mPreprocessor
= new $class( $this );
732 return $this->mPreprocessor
;
736 * Replaces all occurrences of HTML-style comments and the given tags
737 * in the text with a random marker and returns the next text. The output
738 * parameter $matches will be an associative array filled with data in
740 * 'UNIQ-xxxxx' => array(
743 * array( 'param' => 'x' ),
744 * '<element param="x">tag content</element>' ) )
746 * @param $elements array list of element names. Comments are always extracted.
747 * @param $text string Source text string.
748 * @param $matches array Out parameter, Array: extracted tags
749 * @param $uniq_prefix string
750 * @return String: stripped text
752 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = '' ) {
757 $taglist = implode( '|', $elements );
758 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
760 while ( $text != '' ) {
761 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE
);
763 if ( count( $p ) < 5 ) {
766 if ( count( $p ) > 5 ) {
780 $marker = "$uniq_prefix-$element-" . sprintf( '%08X', $n++
) . self
::MARKER_SUFFIX
;
781 $stripped .= $marker;
783 if ( $close === '/>' ) {
784 # Empty element tag, <tag />
789 if ( $element === '!--' ) {
792 $end = "/(<\\/$element\\s*>)/i";
794 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE
);
796 if ( count( $q ) < 3 ) {
797 # No end tag -- let it run out to the end of the text.
806 $matches[$marker] = array( $element,
808 Sanitizer
::decodeTagAttributes( $attributes ),
809 "<$element$attributes$close$content$tail" );
815 * Get a list of strippable XML-like elements
819 function getStripList() {
820 return $this->mStripList
;
824 * Add an item to the strip state
825 * Returns the unique tag which must be inserted into the stripped text
826 * The tag will be replaced with the original text in unstrip()
828 * @param $text string
832 function insertStripItem( $text ) {
833 $rnd = "{$this->mUniqPrefix}-item-{$this->mMarkerIndex}-" . self
::MARKER_SUFFIX
;
834 $this->mMarkerIndex++
;
835 $this->mStripState
->addGeneral( $rnd, $text );
840 * parse the wiki syntax used to render tables
844 function doTableStuff( $text ) {
845 wfProfileIn( __METHOD__
);
847 $lines = StringUtils
::explode( "\n", $text );
849 $td_history = array(); # Is currently a td tag open?
850 $last_tag_history = array(); # Save history of last lag activated (td, th or caption)
851 $tr_history = array(); # Is currently a tr tag open?
852 $tr_attributes = array(); # history of tr attributes
853 $has_opened_tr = array(); # Did this table open a <tr> element?
854 $indent_level = 0; # indent level of the table
856 foreach ( $lines as $outLine ) {
857 $line = trim( $outLine );
859 if ( $line === '' ) { # empty line, go to next line
860 $out .= $outLine."\n";
864 $first_character = $line[0];
867 if ( preg_match( '/^(:*)\{\|(.*)$/', $line , $matches ) ) {
868 # First check if we are starting a new table
869 $indent_level = strlen( $matches[1] );
871 $attributes = $this->mStripState
->unstripBoth( $matches[2] );
872 $attributes = Sanitizer
::fixTagAttributes( $attributes , 'table' );
874 $outLine = str_repeat( '<dl><dd>' , $indent_level ) . "<table{$attributes}>";
875 array_push( $td_history , false );
876 array_push( $last_tag_history , '' );
877 array_push( $tr_history , false );
878 array_push( $tr_attributes , '' );
879 array_push( $has_opened_tr , false );
880 } elseif ( count( $td_history ) == 0 ) {
881 # Don't do any of the following
882 $out .= $outLine."\n";
884 } elseif ( substr( $line , 0 , 2 ) === '|}' ) {
885 # We are ending a table
886 $line = '</table>' . substr( $line , 2 );
887 $last_tag = array_pop( $last_tag_history );
889 if ( !array_pop( $has_opened_tr ) ) {
890 $line = "<tr><td></td></tr>{$line}";
893 if ( array_pop( $tr_history ) ) {
894 $line = "</tr>{$line}";
897 if ( array_pop( $td_history ) ) {
898 $line = "</{$last_tag}>{$line}";
900 array_pop( $tr_attributes );
901 $outLine = $line . str_repeat( '</dd></dl>' , $indent_level );
902 } elseif ( substr( $line , 0 , 2 ) === '|-' ) {
903 # Now we have a table row
904 $line = preg_replace( '#^\|-+#', '', $line );
906 # Whats after the tag is now only attributes
907 $attributes = $this->mStripState
->unstripBoth( $line );
908 $attributes = Sanitizer
::fixTagAttributes( $attributes, 'tr' );
909 array_pop( $tr_attributes );
910 array_push( $tr_attributes, $attributes );
913 $last_tag = array_pop( $last_tag_history );
914 array_pop( $has_opened_tr );
915 array_push( $has_opened_tr , true );
917 if ( array_pop( $tr_history ) ) {
921 if ( array_pop( $td_history ) ) {
922 $line = "</{$last_tag}>{$line}";
926 array_push( $tr_history , false );
927 array_push( $td_history , false );
928 array_push( $last_tag_history , '' );
929 } elseif ( $first_character === '|' ||
$first_character === '!' ||
substr( $line , 0 , 2 ) === '|+' ) {
930 # This might be cell elements, td, th or captions
931 if ( substr( $line , 0 , 2 ) === '|+' ) {
932 $first_character = '+';
933 $line = substr( $line , 1 );
936 $line = substr( $line , 1 );
938 if ( $first_character === '!' ) {
939 $line = str_replace( '!!' , '||' , $line );
942 # Split up multiple cells on the same line.
943 # FIXME : This can result in improper nesting of tags processed
944 # by earlier parser steps, but should avoid splitting up eg
945 # attribute values containing literal "||".
946 $cells = StringUtils
::explodeMarkup( '||' , $line );
950 # Loop through each table cell
951 foreach ( $cells as $cell ) {
953 if ( $first_character !== '+' ) {
954 $tr_after = array_pop( $tr_attributes );
955 if ( !array_pop( $tr_history ) ) {
956 $previous = "<tr{$tr_after}>\n";
958 array_push( $tr_history , true );
959 array_push( $tr_attributes , '' );
960 array_pop( $has_opened_tr );
961 array_push( $has_opened_tr , true );
964 $last_tag = array_pop( $last_tag_history );
966 if ( array_pop( $td_history ) ) {
967 $previous = "</{$last_tag}>\n{$previous}";
970 if ( $first_character === '|' ) {
972 } elseif ( $first_character === '!' ) {
974 } elseif ( $first_character === '+' ) {
975 $last_tag = 'caption';
980 array_push( $last_tag_history , $last_tag );
982 # A cell could contain both parameters and data
983 $cell_data = explode( '|' , $cell , 2 );
985 # Bug 553: Note that a '|' inside an invalid link should not
986 # be mistaken as delimiting cell parameters
987 if ( strpos( $cell_data[0], '[[' ) !== false ) {
988 $cell = "{$previous}<{$last_tag}>{$cell}";
989 } elseif ( count( $cell_data ) == 1 ) {
990 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
992 $attributes = $this->mStripState
->unstripBoth( $cell_data[0] );
993 $attributes = Sanitizer
::fixTagAttributes( $attributes , $last_tag );
994 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
998 array_push( $td_history , true );
1001 $out .= $outLine . "\n";
1004 # Closing open td, tr && table
1005 while ( count( $td_history ) > 0 ) {
1006 if ( array_pop( $td_history ) ) {
1009 if ( array_pop( $tr_history ) ) {
1012 if ( !array_pop( $has_opened_tr ) ) {
1013 $out .= "<tr><td></td></tr>\n" ;
1016 $out .= "</table>\n";
1019 # Remove trailing line-ending (b/c)
1020 if ( substr( $out, -1 ) === "\n" ) {
1021 $out = substr( $out, 0, -1 );
1024 # special case: don't return empty table
1025 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1029 wfProfileOut( __METHOD__
);
1035 * Helper function for parse() that transforms wiki markup into
1036 * HTML. Only called for $mOutputType == self::OT_HTML.
1040 * @param $text string
1041 * @param $isMain bool
1042 * @param $frame bool
1046 function internalParse( $text, $isMain = true, $frame = false ) {
1047 wfProfileIn( __METHOD__
);
1051 # Hook to suspend the parser in this state
1052 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$this->mStripState
) ) ) {
1053 wfProfileOut( __METHOD__
);
1057 # if $frame is provided, then use $frame for replacing any variables
1059 # use frame depth to infer how include/noinclude tags should be handled
1060 # depth=0 means this is the top-level document; otherwise it's an included document
1061 if ( !$frame->depth
) {
1064 $flag = Parser
::PTD_FOR_INCLUSION
;
1066 $dom = $this->preprocessToDom( $text, $flag );
1067 $text = $frame->expand( $dom );
1069 # if $frame is not provided, then use old-style replaceVariables
1070 $text = $this->replaceVariables( $text );
1073 $text = Sanitizer
::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ), false, array_keys( $this->mTransparentTagHooks
) );
1074 wfRunHooks( 'InternalParseBeforeLinks', array( &$this, &$text, &$this->mStripState
) );
1076 # Tables need to come after variable replacement for things to work
1077 # properly; putting them before other transformations should keep
1078 # exciting things like link expansions from showing up in surprising
1080 $text = $this->doTableStuff( $text );
1082 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1084 $text = $this->doDoubleUnderscore( $text );
1086 $text = $this->doHeadings( $text );
1087 if ( $this->mOptions
->getUseDynamicDates() ) {
1088 $df = DateFormatter
::getInstance();
1089 $text = $df->reformat( $this->mOptions
->getDateFormat(), $text );
1091 $text = $this->replaceInternalLinks( $text );
1092 $text = $this->doAllQuotes( $text );
1093 $text = $this->replaceExternalLinks( $text );
1095 # replaceInternalLinks may sometimes leave behind
1096 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1097 $text = str_replace( $this->mUniqPrefix
.'NOPARSE', '', $text );
1099 $text = $this->doMagicLinks( $text );
1100 $text = $this->formatHeadings( $text, $origText, $isMain );
1102 wfProfileOut( __METHOD__
);
1107 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1108 * magic external links.
1113 * @param $text string
1117 function doMagicLinks( $text ) {
1118 wfProfileIn( __METHOD__
);
1119 $prots = wfUrlProtocolsWithoutProtRel();
1120 $urlChar = self
::EXT_LINK_URL_CLASS
;
1121 $text = preg_replace_callback(
1123 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1124 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1125 (\\b(?:$prots)$urlChar+) | # m[3]: Free external links" . '
1126 (?:RFC|PMID)\s+([0-9]+) | # m[4]: RFC or PMID, capture number
1127 ISBN\s+(\b # m[5]: ISBN, capture number
1128 (?: 97[89] [\ \-]? )? # optional 13-digit ISBN prefix
1129 (?: [0-9] [\ \-]? ){9} # 9 digits with opt. delimiters
1130 [0-9Xx] # check digit
1132 )!xu', array( &$this, 'magicLinkCallback' ), $text );
1133 wfProfileOut( __METHOD__
);
1138 * @throws MWException
1140 * @return HTML|string
1142 function magicLinkCallback( $m ) {
1143 if ( isset( $m[1] ) && $m[1] !== '' ) {
1146 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1149 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1150 # Free external link
1151 return $this->makeFreeExternalLink( $m[0] );
1152 } elseif ( isset( $m[4] ) && $m[4] !== '' ) {
1154 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1157 $CssClass = 'mw-magiclink-rfc';
1159 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1161 $urlmsg = 'pubmedurl';
1162 $CssClass = 'mw-magiclink-pmid';
1165 throw new MWException( __METHOD__
.': unrecognised match type "' .
1166 substr( $m[0], 0, 20 ) . '"' );
1168 $url = wfMsgForContent( $urlmsg, $id );
1169 return Linker
::makeExternalLink( $url, "{$keyword} {$id}", true, $CssClass );
1170 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1173 $num = strtr( $isbn, array(
1178 $titleObj = SpecialPage
::getTitleFor( 'Booksources', $num );
1180 $titleObj->escapeLocalUrl() .
1181 "\" class=\"internal mw-magiclink-isbn\">ISBN $isbn</a>";
1188 * Make a free external link, given a user-supplied URL
1190 * @param $url string
1192 * @return string HTML
1195 function makeFreeExternalLink( $url ) {
1196 wfProfileIn( __METHOD__
);
1200 # The characters '<' and '>' (which were escaped by
1201 # removeHTMLtags()) should not be included in
1202 # URLs, per RFC 2396.
1204 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE
) ) {
1205 $trail = substr( $url, $m2[0][1] ) . $trail;
1206 $url = substr( $url, 0, $m2[0][1] );
1209 # Move trailing punctuation to $trail
1211 # If there is no left bracket, then consider right brackets fair game too
1212 if ( strpos( $url, '(' ) === false ) {
1216 $numSepChars = strspn( strrev( $url ), $sep );
1217 if ( $numSepChars ) {
1218 $trail = substr( $url, -$numSepChars ) . $trail;
1219 $url = substr( $url, 0, -$numSepChars );
1222 $url = Sanitizer
::cleanUrl( $url );
1224 # Is this an external image?
1225 $text = $this->maybeMakeExternalImage( $url );
1226 if ( $text === false ) {
1227 # Not an image, make a link
1228 $text = Linker
::makeExternalLink( $url, $this->getFunctionLang()->markNoConversion($url), true, 'free',
1229 $this->getExternalLinkAttribs( $url ) );
1230 # Register it in the output object...
1231 # Replace unnecessary URL escape codes with their equivalent characters
1232 $pasteurized = self
::replaceUnusualEscapes( $url );
1233 $this->mOutput
->addExternalLink( $pasteurized );
1235 wfProfileOut( __METHOD__
);
1236 return $text . $trail;
1241 * Parse headers and return html
1245 * @param $text string
1249 function doHeadings( $text ) {
1250 wfProfileIn( __METHOD__
);
1251 for ( $i = 6; $i >= 1; --$i ) {
1252 $h = str_repeat( '=', $i );
1253 $text = preg_replace( "/^$h(.+)$h\\s*$/m",
1254 "<h$i>\\1</h$i>", $text );
1256 wfProfileOut( __METHOD__
);
1261 * Replace single quotes with HTML markup
1264 * @param $text string
1266 * @return string the altered text
1268 function doAllQuotes( $text ) {
1269 wfProfileIn( __METHOD__
);
1271 $lines = StringUtils
::explode( "\n", $text );
1272 foreach ( $lines as $line ) {
1273 $outtext .= $this->doQuotes( $line ) . "\n";
1275 $outtext = substr( $outtext, 0,-1 );
1276 wfProfileOut( __METHOD__
);
1281 * Helper function for doAllQuotes()
1283 * @param $text string
1287 public function doQuotes( $text ) {
1288 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE
);
1289 if ( count( $arr ) == 1 ) {
1292 # First, do some preliminary work. This may shift some apostrophes from
1293 # being mark-up to being text. It also counts the number of occurrences
1294 # of bold and italics mark-ups.
1297 for ( $i = 0; $i < count( $arr ); $i++
) {
1298 if ( ( $i %
2 ) == 1 ) {
1299 # If there are ever four apostrophes, assume the first is supposed to
1300 # be text, and the remaining three constitute mark-up for bold text.
1301 if ( strlen( $arr[$i] ) == 4 ) {
1304 } elseif ( strlen( $arr[$i] ) > 5 ) {
1305 # If there are more than 5 apostrophes in a row, assume they're all
1306 # text except for the last 5.
1307 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
1310 # Count the number of occurrences of bold and italics mark-ups.
1311 # We are not counting sequences of five apostrophes.
1312 if ( strlen( $arr[$i] ) == 2 ) {
1314 } elseif ( strlen( $arr[$i] ) == 3 ) {
1316 } elseif ( strlen( $arr[$i] ) == 5 ) {
1323 # If there is an odd number of both bold and italics, it is likely
1324 # that one of the bold ones was meant to be an apostrophe followed
1325 # by italics. Which one we cannot know for certain, but it is more
1326 # likely to be one that has a single-letter word before it.
1327 if ( ( $numbold %
2 == 1 ) && ( $numitalics %
2 == 1 ) ) {
1329 $firstsingleletterword = -1;
1330 $firstmultiletterword = -1;
1332 foreach ( $arr as $r ) {
1333 if ( ( $i %
2 == 1 ) and ( strlen( $r ) == 3 ) ) {
1334 $x1 = substr( $arr[$i-1], -1 );
1335 $x2 = substr( $arr[$i-1], -2, 1 );
1336 if ( $x1 === ' ' ) {
1337 if ( $firstspace == -1 ) {
1340 } elseif ( $x2 === ' ') {
1341 if ( $firstsingleletterword == -1 ) {
1342 $firstsingleletterword = $i;
1345 if ( $firstmultiletterword == -1 ) {
1346 $firstmultiletterword = $i;
1353 # If there is a single-letter word, use it!
1354 if ( $firstsingleletterword > -1 ) {
1355 $arr[$firstsingleletterword] = "''";
1356 $arr[$firstsingleletterword-1] .= "'";
1357 } elseif ( $firstmultiletterword > -1 ) {
1358 # If not, but there's a multi-letter word, use that one.
1359 $arr[$firstmultiletterword] = "''";
1360 $arr[$firstmultiletterword-1] .= "'";
1361 } elseif ( $firstspace > -1 ) {
1362 # ... otherwise use the first one that has neither.
1363 # (notice that it is possible for all three to be -1 if, for example,
1364 # there is only one pentuple-apostrophe in the line)
1365 $arr[$firstspace] = "''";
1366 $arr[$firstspace-1] .= "'";
1370 # Now let's actually convert our apostrophic mush to HTML!
1375 foreach ( $arr as $r ) {
1376 if ( ( $i %
2 ) == 0 ) {
1377 if ( $state === 'both' ) {
1383 if ( strlen( $r ) == 2 ) {
1384 if ( $state === 'i' ) {
1385 $output .= '</i>'; $state = '';
1386 } elseif ( $state === 'bi' ) {
1387 $output .= '</i>'; $state = 'b';
1388 } elseif ( $state === 'ib' ) {
1389 $output .= '</b></i><b>'; $state = 'b';
1390 } elseif ( $state === 'both' ) {
1391 $output .= '<b><i>'.$buffer.'</i>'; $state = 'b';
1392 } else { # $state can be 'b' or ''
1393 $output .= '<i>'; $state .= 'i';
1395 } elseif ( strlen( $r ) == 3 ) {
1396 if ( $state === 'b' ) {
1397 $output .= '</b>'; $state = '';
1398 } elseif ( $state === 'bi' ) {
1399 $output .= '</i></b><i>'; $state = 'i';
1400 } elseif ( $state === 'ib' ) {
1401 $output .= '</b>'; $state = 'i';
1402 } elseif ( $state === 'both' ) {
1403 $output .= '<i><b>'.$buffer.'</b>'; $state = 'i';
1404 } else { # $state can be 'i' or ''
1405 $output .= '<b>'; $state .= 'b';
1407 } elseif ( strlen( $r ) == 5 ) {
1408 if ( $state === 'b' ) {
1409 $output .= '</b><i>'; $state = 'i';
1410 } elseif ( $state === 'i' ) {
1411 $output .= '</i><b>'; $state = 'b';
1412 } elseif ( $state === 'bi' ) {
1413 $output .= '</i></b>'; $state = '';
1414 } elseif ( $state === 'ib' ) {
1415 $output .= '</b></i>'; $state = '';
1416 } elseif ( $state === 'both' ) {
1417 $output .= '<i><b>'.$buffer.'</b></i>'; $state = '';
1418 } else { # ($state == '')
1419 $buffer = ''; $state = 'both';
1425 # Now close all remaining tags. Notice that the order is important.
1426 if ( $state === 'b' ||
$state === 'ib' ) {
1429 if ( $state === 'i' ||
$state === 'bi' ||
$state === 'ib' ) {
1432 if ( $state === 'bi' ) {
1435 # There might be lonely ''''', so make sure we have a buffer
1436 if ( $state === 'both' && $buffer ) {
1437 $output .= '<b><i>'.$buffer.'</i></b>';
1444 * Replace external links (REL)
1446 * Note: this is all very hackish and the order of execution matters a lot.
1447 * Make sure to run maintenance/parserTests.php if you change this code.
1451 * @param $text string
1455 function replaceExternalLinks( $text ) {
1456 wfProfileIn( __METHOD__
);
1458 $bits = preg_split( $this->mExtLinkBracketedRegex
, $text, -1, PREG_SPLIT_DELIM_CAPTURE
);
1459 $s = array_shift( $bits );
1462 while ( $i<count( $bits ) ) {
1464 $protocol = $bits[$i++
];
1465 $text = $bits[$i++
];
1466 $trail = $bits[$i++
];
1468 # The characters '<' and '>' (which were escaped by
1469 # removeHTMLtags()) should not be included in
1470 # URLs, per RFC 2396.
1472 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE
) ) {
1473 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1474 $url = substr( $url, 0, $m2[0][1] );
1477 # If the link text is an image URL, replace it with an <img> tag
1478 # This happened by accident in the original parser, but some people used it extensively
1479 $img = $this->maybeMakeExternalImage( $text );
1480 if ( $img !== false ) {
1486 # Set linktype for CSS - if URL==text, link is essentially free
1487 $linktype = ( $text === $url ) ?
'free' : 'text';
1489 # No link text, e.g. [http://domain.tld/some.link]
1490 if ( $text == '' ) {
1492 $langObj = $this->getFunctionLang();
1493 $text = '[' . $langObj->formatNum( ++
$this->mAutonumber
) . ']';
1494 $linktype = 'autonumber';
1496 # Have link text, e.g. [http://domain.tld/some.link text]s
1498 list( $dtrail, $trail ) = Linker
::splitTrail( $trail );
1501 $text = $this->getFunctionLang()->markNoConversion( $text );
1503 $url = Sanitizer
::cleanUrl( $url );
1505 # Use the encoded URL
1506 # This means that users can paste URLs directly into the text
1507 # Funny characters like ö aren't valid in URLs anyway
1508 # This was changed in August 2004
1509 $s .= Linker
::makeExternalLink( $url, $text, false, $linktype,
1510 $this->getExternalLinkAttribs( $url ) ) . $dtrail . $trail;
1512 # Register link in the output object.
1513 # Replace unnecessary URL escape codes with the referenced character
1514 # This prevents spammers from hiding links from the filters
1515 $pasteurized = self
::replaceUnusualEscapes( $url );
1516 $this->mOutput
->addExternalLink( $pasteurized );
1519 wfProfileOut( __METHOD__
);
1524 * Get an associative array of additional HTML attributes appropriate for a
1525 * particular external link. This currently may include rel => nofollow
1526 * (depending on configuration, namespace, and the URL's domain) and/or a
1527 * target attribute (depending on configuration).
1529 * @param $url String|bool optional URL, to extract the domain from for rel =>
1530 * nofollow if appropriate
1531 * @return Array associative array of HTML attributes
1533 function getExternalLinkAttribs( $url = false ) {
1535 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1536 $ns = $this->mTitle
->getNamespace();
1537 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions ) &&
1538 !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions ) )
1540 $attribs['rel'] = 'nofollow';
1542 if ( $this->mOptions
->getExternalLinkTarget() ) {
1543 $attribs['target'] = $this->mOptions
->getExternalLinkTarget();
1549 * Replace unusual URL escape codes with their equivalent characters
1551 * @param $url String
1554 * @todo This can merge genuinely required bits in the path or query string,
1555 * breaking legit URLs. A proper fix would treat the various parts of
1556 * the URL differently; as a workaround, just use the output for
1557 * statistical records, not for actual linking/output.
1559 static function replaceUnusualEscapes( $url ) {
1560 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/',
1561 array( __CLASS__
, 'replaceUnusualEscapesCallback' ), $url );
1565 * Callback function used in replaceUnusualEscapes().
1566 * Replaces unusual URL escape codes with their equivalent character
1568 * @param $matches array
1572 private static function replaceUnusualEscapesCallback( $matches ) {
1573 $char = urldecode( $matches[0] );
1574 $ord = ord( $char );
1575 # Is it an unsafe or HTTP reserved character according to RFC 1738?
1576 if ( $ord > 32 && $ord < 127 && strpos( '<>"#{}|\^~[]`;/?', $char ) === false ) {
1577 # No, shouldn't be escaped
1580 # Yes, leave it escaped
1586 * make an image if it's allowed, either through the global
1587 * option, through the exception, or through the on-wiki whitelist
1590 * $param $url string
1594 function maybeMakeExternalImage( $url ) {
1595 $imagesfrom = $this->mOptions
->getAllowExternalImagesFrom();
1596 $imagesexception = !empty( $imagesfrom );
1598 # $imagesfrom could be either a single string or an array of strings, parse out the latter
1599 if ( $imagesexception && is_array( $imagesfrom ) ) {
1600 $imagematch = false;
1601 foreach ( $imagesfrom as $match ) {
1602 if ( strpos( $url, $match ) === 0 ) {
1607 } elseif ( $imagesexception ) {
1608 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
1610 $imagematch = false;
1612 if ( $this->mOptions
->getAllowExternalImages()
1613 ||
( $imagesexception && $imagematch ) ) {
1614 if ( preg_match( self
::EXT_IMAGE_REGEX
, $url ) ) {
1616 $text = Linker
::makeExternalImage( $url );
1619 if ( !$text && $this->mOptions
->getEnableImageWhitelist()
1620 && preg_match( self
::EXT_IMAGE_REGEX
, $url ) ) {
1621 $whitelist = explode( "\n", wfMsgForContent( 'external_image_whitelist' ) );
1622 foreach ( $whitelist as $entry ) {
1623 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
1624 if ( strpos( $entry, '#' ) === 0 ||
$entry === '' ) {
1627 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
1628 # Image matches a whitelist entry
1629 $text = Linker
::makeExternalImage( $url );
1638 * Process [[ ]] wikilinks
1642 * @return String: processed text
1646 function replaceInternalLinks( $s ) {
1647 $this->mLinkHolders
->merge( $this->replaceInternalLinks2( $s ) );
1652 * Process [[ ]] wikilinks (RIL)
1653 * @return LinkHolderArray
1657 function replaceInternalLinks2( &$s ) {
1658 wfProfileIn( __METHOD__
);
1660 wfProfileIn( __METHOD__
.'-setup' );
1661 static $tc = FALSE, $e1, $e1_img;
1662 # the % is needed to support urlencoded titles as well
1664 $tc = Title
::legalChars() . '#%';
1665 # Match a link having the form [[namespace:link|alternate]]trail
1666 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
1667 # Match cases where there is no "]]", which might still be images
1668 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
1671 $holders = new LinkHolderArray( $this );
1673 # split the entire text string on occurences of [[
1674 $a = StringUtils
::explode( '[[', ' ' . $s );
1675 # get the first element (all text up to first [[), and remove the space we added
1678 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
1679 $s = substr( $s, 1 );
1681 $useLinkPrefixExtension = $this->getFunctionLang()->linkPrefixExtension();
1683 if ( $useLinkPrefixExtension ) {
1684 # Match the end of a line for a word that's not followed by whitespace,
1685 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1686 $e2 = wfMsgForContent( 'linkprefix' );
1689 if ( is_null( $this->mTitle
) ) {
1690 wfProfileOut( __METHOD__
.'-setup' );
1691 wfProfileOut( __METHOD__
);
1692 throw new MWException( __METHOD__
.": \$this->mTitle is null\n" );
1694 $nottalk = !$this->mTitle
->isTalkPage();
1696 if ( $useLinkPrefixExtension ) {
1698 if ( preg_match( $e2, $s, $m ) ) {
1699 $first_prefix = $m[2];
1701 $first_prefix = false;
1707 if ( $this->getFunctionLang()->hasVariants() ) {
1708 $selflink = $this->getFunctionLang()->autoConvertToAllVariants( $this->mTitle
->getPrefixedText() );
1710 $selflink = array( $this->mTitle
->getPrefixedText() );
1712 $useSubpages = $this->areSubpagesAllowed();
1713 wfProfileOut( __METHOD__
.'-setup' );
1715 # Loop for each link
1716 for ( ; $line !== false && $line !== null ; $a->next(), $line = $a->current() ) {
1717 # Check for excessive memory usage
1718 if ( $holders->isBig() ) {
1720 # Do the existence check, replace the link holders and clear the array
1721 $holders->replace( $s );
1725 if ( $useLinkPrefixExtension ) {
1726 wfProfileIn( __METHOD__
.'-prefixhandling' );
1727 if ( preg_match( $e2, $s, $m ) ) {
1734 if ( $first_prefix ) {
1735 $prefix = $first_prefix;
1736 $first_prefix = false;
1738 wfProfileOut( __METHOD__
.'-prefixhandling' );
1741 $might_be_img = false;
1743 wfProfileIn( __METHOD__
."-e1" );
1744 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1746 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1747 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1748 # the real problem is with the $e1 regex
1751 # Still some problems for cases where the ] is meant to be outside punctuation,
1752 # and no image is in sight. See bug 2095.
1754 if ( $text !== '' &&
1755 substr( $m[3], 0, 1 ) === ']' &&
1756 strpos( $text, '[' ) !== false
1759 $text .= ']'; # so that replaceExternalLinks($text) works later
1760 $m[3] = substr( $m[3], 1 );
1762 # fix up urlencoded title texts
1763 if ( strpos( $m[1], '%' ) !== false ) {
1764 # Should anchors '#' also be rejected?
1765 $m[1] = str_replace( array('<', '>'), array('<', '>'), rawurldecode( $m[1] ) );
1768 } elseif ( preg_match( $e1_img, $line, $m ) ) { # Invalid, but might be an image with a link in its caption
1769 $might_be_img = true;
1771 if ( strpos( $m[1], '%' ) !== false ) {
1772 $m[1] = rawurldecode( $m[1] );
1775 } else { # Invalid form; output directly
1776 $s .= $prefix . '[[' . $line ;
1777 wfProfileOut( __METHOD__
."-e1" );
1780 wfProfileOut( __METHOD__
."-e1" );
1781 wfProfileIn( __METHOD__
."-misc" );
1783 # Don't allow internal links to pages containing
1784 # PROTO: where PROTO is a valid URL protocol; these
1785 # should be external links.
1786 if ( preg_match( '/^(?:' . wfUrlProtocols() . ')/', $m[1] ) ) {
1787 $s .= $prefix . '[[' . $line ;
1788 wfProfileOut( __METHOD__
."-misc" );
1792 # Make subpage if necessary
1793 if ( $useSubpages ) {
1794 $link = $this->maybeDoSubpageLink( $m[1], $text );
1799 $noforce = ( substr( $m[1], 0, 1 ) !== ':' );
1801 # Strip off leading ':'
1802 $link = substr( $link, 1 );
1805 wfProfileOut( __METHOD__
."-misc" );
1806 wfProfileIn( __METHOD__
."-title" );
1807 $nt = Title
::newFromText( $this->mStripState
->unstripNoWiki( $link ) );
1808 if ( $nt === null ) {
1809 $s .= $prefix . '[[' . $line;
1810 wfProfileOut( __METHOD__
."-title" );
1814 $ns = $nt->getNamespace();
1815 $iw = $nt->getInterWiki();
1816 wfProfileOut( __METHOD__
."-title" );
1818 if ( $might_be_img ) { # if this is actually an invalid link
1819 wfProfileIn( __METHOD__
."-might_be_img" );
1820 if ( $ns == NS_FILE
&& $noforce ) { # but might be an image
1823 # look at the next 'line' to see if we can close it there
1825 $next_line = $a->current();
1826 if ( $next_line === false ||
$next_line === null ) {
1829 $m = explode( ']]', $next_line, 3 );
1830 if ( count( $m ) == 3 ) {
1831 # the first ]] closes the inner link, the second the image
1833 $text .= "[[{$m[0]}]]{$m[1]}";
1836 } elseif ( count( $m ) == 2 ) {
1837 # if there's exactly one ]] that's fine, we'll keep looking
1838 $text .= "[[{$m[0]}]]{$m[1]}";
1840 # if $next_line is invalid too, we need look no further
1841 $text .= '[[' . $next_line;
1846 # we couldn't find the end of this imageLink, so output it raw
1847 # but don't ignore what might be perfectly normal links in the text we've examined
1848 $holders->merge( $this->replaceInternalLinks2( $text ) );
1849 $s .= "{$prefix}[[$link|$text";
1850 # note: no $trail, because without an end, there *is* no trail
1851 wfProfileOut( __METHOD__
."-might_be_img" );
1854 } else { # it's not an image, so output it raw
1855 $s .= "{$prefix}[[$link|$text";
1856 # note: no $trail, because without an end, there *is* no trail
1857 wfProfileOut( __METHOD__
."-might_be_img" );
1860 wfProfileOut( __METHOD__
."-might_be_img" );
1863 $wasblank = ( $text == '' );
1867 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
1868 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
1869 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
1870 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
1871 $text = $this->doQuotes( $text );
1874 # Link not escaped by : , create the various objects
1879 wfProfileIn( __METHOD__
."-interwiki" );
1880 if ( $iw && $this->mOptions
->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1881 $this->mOutput
->addLanguageLink( $nt->getFullText() );
1882 $s = rtrim( $s . $prefix );
1883 $s .= trim( $trail, "\n" ) == '' ?
'': $prefix . $trail;
1884 wfProfileOut( __METHOD__
."-interwiki" );
1887 wfProfileOut( __METHOD__
."-interwiki" );
1889 if ( $ns == NS_FILE
) {
1890 wfProfileIn( __METHOD__
."-image" );
1891 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle
) ) {
1893 # if no parameters were passed, $text
1894 # becomes something like "File:Foo.png",
1895 # which we don't want to pass on to the
1899 # recursively parse links inside the image caption
1900 # actually, this will parse them in any other parameters, too,
1901 # but it might be hard to fix that, and it doesn't matter ATM
1902 $text = $this->replaceExternalLinks( $text );
1903 $holders->merge( $this->replaceInternalLinks2( $text ) );
1905 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1906 $s .= $prefix . $this->armorLinks(
1907 $this->makeImage( $nt, $text, $holders ) ) . $trail;
1909 $s .= $prefix . $trail;
1911 wfProfileOut( __METHOD__
."-image" );
1915 if ( $ns == NS_CATEGORY
) {
1916 wfProfileIn( __METHOD__
."-category" );
1917 $s = rtrim( $s . "\n" ); # bug 87
1920 $sortkey = $this->getDefaultSort();
1924 $sortkey = Sanitizer
::decodeCharReferences( $sortkey );
1925 $sortkey = str_replace( "\n", '', $sortkey );
1926 $sortkey = $this->getFunctionLang()->convertCategoryKey( $sortkey );
1927 $this->mOutput
->addCategory( $nt->getDBkey(), $sortkey );
1930 * Strip the whitespace Category links produce, see bug 87
1931 * @todo We might want to use trim($tmp, "\n") here.
1933 $s .= trim( $prefix . $trail, "\n" ) == '' ?
'' : $prefix . $trail;
1935 wfProfileOut( __METHOD__
."-category" );
1940 # Self-link checking
1941 if ( $nt->getFragment() === '' && $ns != NS_SPECIAL
) {
1942 if ( in_array( $nt->getPrefixedText(), $selflink, true ) ) {
1943 $s .= $prefix . Linker
::makeSelfLinkObj( $nt, $text, '', $trail );
1948 # NS_MEDIA is a pseudo-namespace for linking directly to a file
1949 # @todo FIXME: Should do batch file existence checks, see comment below
1950 if ( $ns == NS_MEDIA
) {
1951 wfProfileIn( __METHOD__
."-media" );
1952 # Give extensions a chance to select the file revision for us
1955 wfRunHooks( 'BeforeParserFetchFileAndTitle',
1956 array( $this, $nt, &$options, &$descQuery ) );
1957 # Fetch and register the file (file title may be different via hooks)
1958 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
1959 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
1960 $s .= $prefix . $this->armorLinks(
1961 Linker
::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
1962 wfProfileOut( __METHOD__
."-media" );
1966 wfProfileIn( __METHOD__
."-always_known" );
1967 # Some titles, such as valid special pages or files in foreign repos, should
1968 # be shown as bluelinks even though they're not included in the page table
1970 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
1971 # batch file existence checks for NS_FILE and NS_MEDIA
1972 if ( $iw == '' && $nt->isAlwaysKnown() ) {
1973 $this->mOutput
->addLink( $nt );
1974 $s .= $this->makeKnownLinkHolder( $nt, $text, array(), $trail, $prefix );
1976 # Links will be added to the output link list after checking
1977 $s .= $holders->makeHolder( $nt, $text, array(), $trail, $prefix );
1979 wfProfileOut( __METHOD__
."-always_known" );
1981 wfProfileOut( __METHOD__
);
1986 * Render a forced-blue link inline; protect against double expansion of
1987 * URLs if we're in a mode that prepends full URL prefixes to internal links.
1988 * Since this little disaster has to split off the trail text to avoid
1989 * breaking URLs in the following text without breaking trails on the
1990 * wiki links, it's been made into a horrible function.
1993 * @param $text String
1994 * @param $query Array or String
1995 * @param $trail String
1996 * @param $prefix String
1997 * @return String: HTML-wikitext mix oh yuck
1999 function makeKnownLinkHolder( $nt, $text = '', $query = array(), $trail = '', $prefix = '' ) {
2000 list( $inside, $trail ) = Linker
::splitTrail( $trail );
2002 if ( is_string( $query ) ) {
2003 $query = wfCgiToArray( $query );
2005 if ( $text == '' ) {
2006 $text = htmlspecialchars( $nt->getPrefixedText() );
2009 $link = Linker
::linkKnown( $nt, "$prefix$text$inside", array(), $query );
2011 return $this->armorLinks( $link ) . $trail;
2015 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2016 * going to go through further parsing steps before inline URL expansion.
2018 * Not needed quite as much as it used to be since free links are a bit
2019 * more sensible these days. But bracketed links are still an issue.
2021 * @param $text String: more-or-less HTML
2022 * @return String: less-or-more HTML with NOPARSE bits
2024 function armorLinks( $text ) {
2025 return preg_replace( '/\b(' . wfUrlProtocols() . ')/',
2026 "{$this->mUniqPrefix}NOPARSE$1", $text );
2030 * Return true if subpage links should be expanded on this page.
2033 function areSubpagesAllowed() {
2034 # Some namespaces don't allow subpages
2035 return MWNamespace
::hasSubpages( $this->mTitle
->getNamespace() );
2039 * Handle link to subpage if necessary
2041 * @param $target String: the source of the link
2042 * @param &$text String: the link text, modified as necessary
2043 * @return string the full name of the link
2046 function maybeDoSubpageLink( $target, &$text ) {
2047 return Linker
::normalizeSubpageLink( $this->mTitle
, $target, $text );
2051 * Used by doBlockLevels()
2056 function closeParagraph() {
2058 if ( $this->mLastSection
!= '' ) {
2059 $result = '</' . $this->mLastSection
. ">\n";
2061 $this->mInPre
= false;
2062 $this->mLastSection
= '';
2067 * getCommon() returns the length of the longest common substring
2068 * of both arguments, starting at the beginning of both.
2071 * @param $st1 string
2072 * @param $st2 string
2076 function getCommon( $st1, $st2 ) {
2077 $fl = strlen( $st1 );
2078 $shorter = strlen( $st2 );
2079 if ( $fl < $shorter ) {
2083 for ( $i = 0; $i < $shorter; ++
$i ) {
2084 if ( $st1[$i] != $st2[$i] ) {
2092 * These next three functions open, continue, and close the list
2093 * element appropriate to the prefix character passed into them.
2100 function openList( $char ) {
2101 $result = $this->closeParagraph();
2103 if ( '*' === $char ) {
2104 $result .= '<ul><li>';
2105 } elseif ( '#' === $char ) {
2106 $result .= '<ol><li>';
2107 } elseif ( ':' === $char ) {
2108 $result .= '<dl><dd>';
2109 } elseif ( ';' === $char ) {
2110 $result .= '<dl><dt>';
2111 $this->mDTopen
= true;
2113 $result = '<!-- ERR 1 -->';
2121 * @param $char String
2126 function nextItem( $char ) {
2127 if ( '*' === $char ||
'#' === $char ) {
2129 } elseif ( ':' === $char ||
';' === $char ) {
2131 if ( $this->mDTopen
) {
2134 if ( ';' === $char ) {
2135 $this->mDTopen
= true;
2136 return $close . '<dt>';
2138 $this->mDTopen
= false;
2139 return $close . '<dd>';
2142 return '<!-- ERR 2 -->';
2147 * @param $char String
2152 function closeList( $char ) {
2153 if ( '*' === $char ) {
2154 $text = '</li></ul>';
2155 } elseif ( '#' === $char ) {
2156 $text = '</li></ol>';
2157 } elseif ( ':' === $char ) {
2158 if ( $this->mDTopen
) {
2159 $this->mDTopen
= false;
2160 $text = '</dt></dl>';
2162 $text = '</dd></dl>';
2165 return '<!-- ERR 3 -->';
2172 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2174 * @param $text String
2175 * @param $linestart Boolean: whether or not this is at the start of a line.
2177 * @return string the lists rendered as HTML
2179 function doBlockLevels( $text, $linestart ) {
2180 wfProfileIn( __METHOD__
);
2182 # Parsing through the text line by line. The main thing
2183 # happening here is handling of block-level elements p, pre,
2184 # and making lists from lines starting with * # : etc.
2186 $textLines = StringUtils
::explode( "\n", $text );
2188 $lastPrefix = $output = '';
2189 $this->mDTopen
= $inBlockElem = false;
2191 $paragraphStack = false;
2193 foreach ( $textLines as $oLine ) {
2195 if ( !$linestart ) {
2205 $lastPrefixLength = strlen( $lastPrefix );
2206 $preCloseMatch = preg_match( '/<\\/pre/i', $oLine );
2207 $preOpenMatch = preg_match( '/<pre/i', $oLine );
2208 # If not in a <pre> element, scan for and figure out what prefixes are there.
2209 if ( !$this->mInPre
) {
2210 # Multiple prefixes may abut each other for nested lists.
2211 $prefixLength = strspn( $oLine, '*#:;' );
2212 $prefix = substr( $oLine, 0, $prefixLength );
2215 # ; and : are both from definition-lists, so they're equivalent
2216 # for the purposes of determining whether or not we need to open/close
2218 $prefix2 = str_replace( ';', ':', $prefix );
2219 $t = substr( $oLine, $prefixLength );
2220 $this->mInPre
= (bool)$preOpenMatch;
2222 # Don't interpret any other prefixes in preformatted text
2224 $prefix = $prefix2 = '';
2229 if ( $prefixLength && $lastPrefix === $prefix2 ) {
2230 # Same as the last item, so no need to deal with nesting or opening stuff
2231 $output .= $this->nextItem( substr( $prefix, -1 ) );
2232 $paragraphStack = false;
2234 if ( substr( $prefix, -1 ) === ';') {
2235 # The one nasty exception: definition lists work like this:
2236 # ; title : definition text
2237 # So we check for : in the remainder text to split up the
2238 # title and definition, without b0rking links.
2240 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
2242 $output .= $term . $this->nextItem( ':' );
2245 } elseif ( $prefixLength ||
$lastPrefixLength ) {
2246 # We need to open or close prefixes, or both.
2248 # Either open or close a level...
2249 $commonPrefixLength = $this->getCommon( $prefix, $lastPrefix );
2250 $paragraphStack = false;
2252 # Close all the prefixes which aren't shared.
2253 while ( $commonPrefixLength < $lastPrefixLength ) {
2254 $output .= $this->closeList( $lastPrefix[$lastPrefixLength-1] );
2255 --$lastPrefixLength;
2258 # Continue the current prefix if appropriate.
2259 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
2260 $output .= $this->nextItem( $prefix[$commonPrefixLength-1] );
2263 # Open prefixes where appropriate.
2264 while ( $prefixLength > $commonPrefixLength ) {
2265 $char = substr( $prefix, $commonPrefixLength, 1 );
2266 $output .= $this->openList( $char );
2268 if ( ';' === $char ) {
2269 # @todo FIXME: This is dupe of code above
2270 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
2272 $output .= $term . $this->nextItem( ':' );
2275 ++
$commonPrefixLength;
2277 $lastPrefix = $prefix2;
2280 # If we have no prefixes, go to paragraph mode.
2281 if ( 0 == $prefixLength ) {
2282 wfProfileIn( __METHOD__
."-paragraph" );
2283 # No prefix (not in list)--go to paragraph mode
2284 # XXX: use a stack for nestable elements like span, table and div
2285 $openmatch = preg_match('/(?:<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<ol|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
2286 $closematch = preg_match(
2287 '/(?:<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
2288 '<td|<th|<\\/?div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix
.'-pre|<\\/li|<\\/ul|<\\/ol|<\\/?center)/iS', $t );
2289 if ( $openmatch or $closematch ) {
2290 $paragraphStack = false;
2291 # TODO bug 5718: paragraph closed
2292 $output .= $this->closeParagraph();
2293 if ( $preOpenMatch and !$preCloseMatch ) {
2294 $this->mInPre
= true;
2296 $inBlockElem = !$closematch;
2297 } elseif ( !$inBlockElem && !$this->mInPre
) {
2298 if ( ' ' == substr( $t, 0, 1 ) and ( $this->mLastSection
=== 'pre' ||
trim( $t ) != '' ) ) {
2300 if ( $this->mLastSection
!== 'pre' ) {
2301 $paragraphStack = false;
2302 $output .= $this->closeParagraph().'<pre>';
2303 $this->mLastSection
= 'pre';
2305 $t = substr( $t, 1 );
2308 if ( trim( $t ) === '' ) {
2309 if ( $paragraphStack ) {
2310 $output .= $paragraphStack.'<br />';
2311 $paragraphStack = false;
2312 $this->mLastSection
= 'p';
2314 if ( $this->mLastSection
!== 'p' ) {
2315 $output .= $this->closeParagraph();
2316 $this->mLastSection
= '';
2317 $paragraphStack = '<p>';
2319 $paragraphStack = '</p><p>';
2323 if ( $paragraphStack ) {
2324 $output .= $paragraphStack;
2325 $paragraphStack = false;
2326 $this->mLastSection
= 'p';
2327 } elseif ( $this->mLastSection
!== 'p' ) {
2328 $output .= $this->closeParagraph().'<p>';
2329 $this->mLastSection
= 'p';
2334 wfProfileOut( __METHOD__
."-paragraph" );
2336 # somewhere above we forget to get out of pre block (bug 785)
2337 if ( $preCloseMatch && $this->mInPre
) {
2338 $this->mInPre
= false;
2340 if ( $paragraphStack === false ) {
2344 while ( $prefixLength ) {
2345 $output .= $this->closeList( $prefix2[$prefixLength-1] );
2348 if ( $this->mLastSection
!= '' ) {
2349 $output .= '</' . $this->mLastSection
. '>';
2350 $this->mLastSection
= '';
2353 wfProfileOut( __METHOD__
);
2358 * Split up a string on ':', ignoring any occurences inside tags
2359 * to prevent illegal overlapping.
2361 * @param $str String the string to split
2362 * @param &$before String set to everything before the ':'
2363 * @param &$after String set to everything after the ':'
2364 * @return String the position of the ':', or false if none found
2366 function findColonNoLinks( $str, &$before, &$after ) {
2367 wfProfileIn( __METHOD__
);
2369 $pos = strpos( $str, ':' );
2370 if ( $pos === false ) {
2372 wfProfileOut( __METHOD__
);
2376 $lt = strpos( $str, '<' );
2377 if ( $lt === false ||
$lt > $pos ) {
2378 # Easy; no tag nesting to worry about
2379 $before = substr( $str, 0, $pos );
2380 $after = substr( $str, $pos+
1 );
2381 wfProfileOut( __METHOD__
);
2385 # Ugly state machine to walk through avoiding tags.
2386 $state = self
::COLON_STATE_TEXT
;
2388 $len = strlen( $str );
2389 for( $i = 0; $i < $len; $i++
) {
2393 # (Using the number is a performance hack for common cases)
2394 case 0: # self::COLON_STATE_TEXT:
2397 # Could be either a <start> tag or an </end> tag
2398 $state = self
::COLON_STATE_TAGSTART
;
2401 if ( $stack == 0 ) {
2403 $before = substr( $str, 0, $i );
2404 $after = substr( $str, $i +
1 );
2405 wfProfileOut( __METHOD__
);
2408 # Embedded in a tag; don't break it.
2411 # Skip ahead looking for something interesting
2412 $colon = strpos( $str, ':', $i );
2413 if ( $colon === false ) {
2414 # Nothing else interesting
2415 wfProfileOut( __METHOD__
);
2418 $lt = strpos( $str, '<', $i );
2419 if ( $stack === 0 ) {
2420 if ( $lt === false ||
$colon < $lt ) {
2422 $before = substr( $str, 0, $colon );
2423 $after = substr( $str, $colon +
1 );
2424 wfProfileOut( __METHOD__
);
2428 if ( $lt === false ) {
2429 # Nothing else interesting to find; abort!
2430 # We're nested, but there's no close tags left. Abort!
2433 # Skip ahead to next tag start
2435 $state = self
::COLON_STATE_TAGSTART
;
2438 case 1: # self::COLON_STATE_TAG:
2443 $state = self
::COLON_STATE_TEXT
;
2446 # Slash may be followed by >?
2447 $state = self
::COLON_STATE_TAGSLASH
;
2453 case 2: # self::COLON_STATE_TAGSTART:
2456 $state = self
::COLON_STATE_CLOSETAG
;
2459 $state = self
::COLON_STATE_COMMENT
;
2462 # Illegal early close? This shouldn't happen D:
2463 $state = self
::COLON_STATE_TEXT
;
2466 $state = self
::COLON_STATE_TAG
;
2469 case 3: # self::COLON_STATE_CLOSETAG:
2474 wfDebug( __METHOD__
.": Invalid input; too many close tags\n" );
2475 wfProfileOut( __METHOD__
);
2478 $state = self
::COLON_STATE_TEXT
;
2481 case self
::COLON_STATE_TAGSLASH
:
2483 # Yes, a self-closed tag <blah/>
2484 $state = self
::COLON_STATE_TEXT
;
2486 # Probably we're jumping the gun, and this is an attribute
2487 $state = self
::COLON_STATE_TAG
;
2490 case 5: # self::COLON_STATE_COMMENT:
2492 $state = self
::COLON_STATE_COMMENTDASH
;
2495 case self
::COLON_STATE_COMMENTDASH
:
2497 $state = self
::COLON_STATE_COMMENTDASHDASH
;
2499 $state = self
::COLON_STATE_COMMENT
;
2502 case self
::COLON_STATE_COMMENTDASHDASH
:
2504 $state = self
::COLON_STATE_TEXT
;
2506 $state = self
::COLON_STATE_COMMENT
;
2510 throw new MWException( "State machine error in " . __METHOD__
);
2514 wfDebug( __METHOD__
.": Invalid input; not enough close tags (stack $stack, state $state)\n" );
2515 wfProfileOut( __METHOD__
);
2518 wfProfileOut( __METHOD__
);
2523 * Return value of a magic variable (like PAGENAME)
2527 * @param $index integer
2528 * @param $frame PPFrame
2532 function getVariableValue( $index, $frame = false ) {
2533 global $wgContLang, $wgSitename, $wgServer;
2534 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2536 if ( is_null( $this->mTitle
) ) {
2537 // If no title set, bad things are going to happen
2538 // later. Title should always be set since this
2539 // should only be called in the middle of a parse
2540 // operation (but the unit-tests do funky stuff)
2541 throw new MWException( __METHOD__
. ' Should only be '
2542 . ' called while parsing (no title set)' );
2546 * Some of these require message or data lookups and can be
2547 * expensive to check many times.
2549 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$this->mVarCache
) ) ) {
2550 if ( isset( $this->mVarCache
[$index] ) ) {
2551 return $this->mVarCache
[$index];
2555 $ts = wfTimestamp( TS_UNIX
, $this->mOptions
->getTimestamp() );
2556 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) );
2559 global $wgLocaltimezone;
2560 if ( isset( $wgLocaltimezone ) ) {
2561 $oldtz = date_default_timezone_get();
2562 date_default_timezone_set( $wgLocaltimezone );
2565 $localTimestamp = date( 'YmdHis', $ts );
2566 $localMonth = date( 'm', $ts );
2567 $localMonth1 = date( 'n', $ts );
2568 $localMonthName = date( 'n', $ts );
2569 $localDay = date( 'j', $ts );
2570 $localDay2 = date( 'd', $ts );
2571 $localDayOfWeek = date( 'w', $ts );
2572 $localWeek = date( 'W', $ts );
2573 $localYear = date( 'Y', $ts );
2574 $localHour = date( 'H', $ts );
2575 if ( isset( $wgLocaltimezone ) ) {
2576 date_default_timezone_set( $oldtz );
2579 $pageLang = $this->getFunctionLang();
2582 case 'currentmonth':
2583 $value = $pageLang->formatNum( gmdate( 'm', $ts ) );
2585 case 'currentmonth1':
2586 $value = $pageLang->formatNum( gmdate( 'n', $ts ) );
2588 case 'currentmonthname':
2589 $value = $pageLang->getMonthName( gmdate( 'n', $ts ) );
2591 case 'currentmonthnamegen':
2592 $value = $pageLang->getMonthNameGen( gmdate( 'n', $ts ) );
2594 case 'currentmonthabbrev':
2595 $value = $pageLang->getMonthAbbreviation( gmdate( 'n', $ts ) );
2598 $value = $pageLang->formatNum( gmdate( 'j', $ts ) );
2601 $value = $pageLang->formatNum( gmdate( 'd', $ts ) );
2604 $value = $pageLang->formatNum( $localMonth );
2607 $value = $pageLang->formatNum( $localMonth1 );
2609 case 'localmonthname':
2610 $value = $pageLang->getMonthName( $localMonthName );
2612 case 'localmonthnamegen':
2613 $value = $pageLang->getMonthNameGen( $localMonthName );
2615 case 'localmonthabbrev':
2616 $value = $pageLang->getMonthAbbreviation( $localMonthName );
2619 $value = $pageLang->formatNum( $localDay );
2622 $value = $pageLang->formatNum( $localDay2 );
2625 $value = wfEscapeWikiText( $this->mTitle
->getText() );
2628 $value = wfEscapeWikiText( $this->mTitle
->getPartialURL() );
2630 case 'fullpagename':
2631 $value = wfEscapeWikiText( $this->mTitle
->getPrefixedText() );
2633 case 'fullpagenamee':
2634 $value = wfEscapeWikiText( $this->mTitle
->getPrefixedURL() );
2637 $value = wfEscapeWikiText( $this->mTitle
->getSubpageText() );
2639 case 'subpagenamee':
2640 $value = wfEscapeWikiText( $this->mTitle
->getSubpageUrlForm() );
2642 case 'basepagename':
2643 $value = wfEscapeWikiText( $this->mTitle
->getBaseText() );
2645 case 'basepagenamee':
2646 $value = wfEscapeWikiText( wfUrlEncode( str_replace( ' ', '_', $this->mTitle
->getBaseText() ) ) );
2648 case 'talkpagename':
2649 if ( $this->mTitle
->canTalk() ) {
2650 $talkPage = $this->mTitle
->getTalkPage();
2651 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2656 case 'talkpagenamee':
2657 if ( $this->mTitle
->canTalk() ) {
2658 $talkPage = $this->mTitle
->getTalkPage();
2659 $value = wfEscapeWikiText( $talkPage->getPrefixedUrl() );
2664 case 'subjectpagename':
2665 $subjPage = $this->mTitle
->getSubjectPage();
2666 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2668 case 'subjectpagenamee':
2669 $subjPage = $this->mTitle
->getSubjectPage();
2670 $value = wfEscapeWikiText( $subjPage->getPrefixedUrl() );
2673 # Let the edit saving system know we should parse the page
2674 # *after* a revision ID has been assigned.
2675 $this->mOutput
->setFlag( 'vary-revision' );
2676 wfDebug( __METHOD__
. ": {{REVISIONID}} used, setting vary-revision...\n" );
2677 $value = $this->mRevisionId
;
2680 # Let the edit saving system know we should parse the page
2681 # *after* a revision ID has been assigned. This is for null edits.
2682 $this->mOutput
->setFlag( 'vary-revision' );
2683 wfDebug( __METHOD__
. ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2684 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2686 case 'revisionday2':
2687 # Let the edit saving system know we should parse the page
2688 # *after* a revision ID has been assigned. This is for null edits.
2689 $this->mOutput
->setFlag( 'vary-revision' );
2690 wfDebug( __METHOD__
. ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2691 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2693 case 'revisionmonth':
2694 # Let the edit saving system know we should parse the page
2695 # *after* a revision ID has been assigned. This is for null edits.
2696 $this->mOutput
->setFlag( 'vary-revision' );
2697 wfDebug( __METHOD__
. ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2698 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2700 case 'revisionmonth1':
2701 # Let the edit saving system know we should parse the page
2702 # *after* a revision ID has been assigned. This is for null edits.
2703 $this->mOutput
->setFlag( 'vary-revision' );
2704 wfDebug( __METHOD__
. ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2705 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2707 case 'revisionyear':
2708 # Let the edit saving system know we should parse the page
2709 # *after* a revision ID has been assigned. This is for null edits.
2710 $this->mOutput
->setFlag( 'vary-revision' );
2711 wfDebug( __METHOD__
. ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2712 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2714 case 'revisiontimestamp':
2715 # Let the edit saving system know we should parse the page
2716 # *after* a revision ID has been assigned. This is for null edits.
2717 $this->mOutput
->setFlag( 'vary-revision' );
2718 wfDebug( __METHOD__
. ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2719 $value = $this->getRevisionTimestamp();
2721 case 'revisionuser':
2722 # Let the edit saving system know we should parse the page
2723 # *after* a revision ID has been assigned. This is for null edits.
2724 $this->mOutput
->setFlag( 'vary-revision' );
2725 wfDebug( __METHOD__
. ": {{REVISIONUSER}} used, setting vary-revision...\n" );
2726 $value = $this->getRevisionUser();
2729 $value = str_replace( '_',' ',$wgContLang->getNsText( $this->mTitle
->getNamespace() ) );
2732 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle
->getNamespace() ) );
2735 $value = $this->mTitle
->canTalk() ?
str_replace( '_',' ',$this->mTitle
->getTalkNsText() ) : '';
2738 $value = $this->mTitle
->canTalk() ?
wfUrlencode( $this->mTitle
->getTalkNsText() ) : '';
2740 case 'subjectspace':
2741 $value = $this->mTitle
->getSubjectNsText();
2743 case 'subjectspacee':
2744 $value = ( wfUrlencode( $this->mTitle
->getSubjectNsText() ) );
2746 case 'currentdayname':
2747 $value = $pageLang->getWeekdayName( gmdate( 'w', $ts ) +
1 );
2750 $value = $pageLang->formatNum( gmdate( 'Y', $ts ), true );
2753 $value = $pageLang->time( wfTimestamp( TS_MW
, $ts ), false, false );
2756 $value = $pageLang->formatNum( gmdate( 'H', $ts ), true );
2759 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2760 # int to remove the padding
2761 $value = $pageLang->formatNum( (int)gmdate( 'W', $ts ) );
2764 $value = $pageLang->formatNum( gmdate( 'w', $ts ) );
2766 case 'localdayname':
2767 $value = $pageLang->getWeekdayName( $localDayOfWeek +
1 );
2770 $value = $pageLang->formatNum( $localYear, true );
2773 $value = $pageLang->time( $localTimestamp, false, false );
2776 $value = $pageLang->formatNum( $localHour, true );
2779 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2780 # int to remove the padding
2781 $value = $pageLang->formatNum( (int)$localWeek );
2784 $value = $pageLang->formatNum( $localDayOfWeek );
2786 case 'numberofarticles':
2787 $value = $pageLang->formatNum( SiteStats
::articles() );
2789 case 'numberoffiles':
2790 $value = $pageLang->formatNum( SiteStats
::images() );
2792 case 'numberofusers':
2793 $value = $pageLang->formatNum( SiteStats
::users() );
2795 case 'numberofactiveusers':
2796 $value = $pageLang->formatNum( SiteStats
::activeUsers() );
2798 case 'numberofpages':
2799 $value = $pageLang->formatNum( SiteStats
::pages() );
2801 case 'numberofadmins':
2802 $value = $pageLang->formatNum( SiteStats
::numberingroup( 'sysop' ) );
2804 case 'numberofedits':
2805 $value = $pageLang->formatNum( SiteStats
::edits() );
2807 case 'numberofviews':
2808 $value = $pageLang->formatNum( SiteStats
::views() );
2810 case 'currenttimestamp':
2811 $value = wfTimestamp( TS_MW
, $ts );
2813 case 'localtimestamp':
2814 $value = $localTimestamp;
2816 case 'currentversion':
2817 $value = SpecialVersion
::getVersion();
2820 return $wgArticlePath;
2826 wfSuppressWarnings(); # May give an E_WARNING in PHP < 5.3.3
2827 $serverName = parse_url( $wgServer, PHP_URL_HOST
);
2828 wfRestoreWarnings();
2829 return $serverName ?
$serverName : $wgServer;
2831 return $wgScriptPath;
2833 return $wgStylePath;
2834 case 'directionmark':
2835 return $pageLang->getDirMark();
2836 case 'contentlanguage':
2837 global $wgLanguageCode;
2838 return $wgLanguageCode;
2841 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$this->mVarCache
, &$index, &$ret, &$frame ) ) ) {
2849 $this->mVarCache
[$index] = $value;
2856 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2860 function initialiseVariables() {
2861 wfProfileIn( __METHOD__
);
2862 $variableIDs = MagicWord
::getVariableIDs();
2863 $substIDs = MagicWord
::getSubstIDs();
2865 $this->mVariables
= new MagicWordArray( $variableIDs );
2866 $this->mSubstWords
= new MagicWordArray( $substIDs );
2867 wfProfileOut( __METHOD__
);
2871 * Preprocess some wikitext and return the document tree.
2872 * This is the ghost of replace_variables().
2874 * @param $text String: The text to parse
2875 * @param $flags Integer: bitwise combination of:
2876 * self::PTD_FOR_INCLUSION Handle <noinclude>/<includeonly> as if the text is being
2877 * included. Default is to assume a direct page view.
2879 * The generated DOM tree must depend only on the input text and the flags.
2880 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
2882 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2883 * change in the DOM tree for a given text, must be passed through the section identifier
2884 * in the section edit link and thus back to extractSections().
2886 * The output of this function is currently only cached in process memory, but a persistent
2887 * cache may be implemented at a later date which takes further advantage of these strict
2888 * dependency requirements.
2894 function preprocessToDom( $text, $flags = 0 ) {
2895 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
2900 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2906 public static function splitWhitespace( $s ) {
2907 $ltrimmed = ltrim( $s );
2908 $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) );
2909 $trimmed = rtrim( $ltrimmed );
2910 $diff = strlen( $ltrimmed ) - strlen( $trimmed );
2912 $w2 = substr( $ltrimmed, -$diff );
2916 return array( $w1, $trimmed, $w2 );
2920 * Replace magic variables, templates, and template arguments
2921 * with the appropriate text. Templates are substituted recursively,
2922 * taking care to avoid infinite loops.
2924 * Note that the substitution depends on value of $mOutputType:
2925 * self::OT_WIKI: only {{subst:}} templates
2926 * self::OT_PREPROCESS: templates but not extension tags
2927 * self::OT_HTML: all templates and extension tags
2929 * @param $text String the text to transform
2930 * @param $frame PPFrame Object describing the arguments passed to the template.
2931 * Arguments may also be provided as an associative array, as was the usual case before MW1.12.
2932 * Providing arguments this way may be useful for extensions wishing to perform variable replacement explicitly.
2933 * @param $argsOnly Boolean only do argument (triple-brace) expansion, not double-brace expansion
2938 function replaceVariables( $text, $frame = false, $argsOnly = false ) {
2939 # Is there any text? Also, Prevent too big inclusions!
2940 if ( strlen( $text ) < 1 ||
strlen( $text ) > $this->mOptions
->getMaxIncludeSize() ) {
2943 wfProfileIn( __METHOD__
);
2945 if ( $frame === false ) {
2946 $frame = $this->getPreprocessor()->newFrame();
2947 } elseif ( !( $frame instanceof PPFrame
) ) {
2948 wfDebug( __METHOD__
." called using plain parameters instead of a PPFrame instance. Creating custom frame.\n" );
2949 $frame = $this->getPreprocessor()->newCustomFrame( $frame );
2952 $dom = $this->preprocessToDom( $text );
2953 $flags = $argsOnly ? PPFrame
::NO_TEMPLATES
: 0;
2954 $text = $frame->expand( $dom, $flags );
2956 wfProfileOut( __METHOD__
);
2961 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
2963 * @param $args array
2967 static function createAssocArgs( $args ) {
2968 $assocArgs = array();
2970 foreach ( $args as $arg ) {
2971 $eqpos = strpos( $arg, '=' );
2972 if ( $eqpos === false ) {
2973 $assocArgs[$index++
] = $arg;
2975 $name = trim( substr( $arg, 0, $eqpos ) );
2976 $value = trim( substr( $arg, $eqpos+
1 ) );
2977 if ( $value === false ) {
2980 if ( $name !== false ) {
2981 $assocArgs[$name] = $value;
2990 * Warn the user when a parser limitation is reached
2991 * Will warn at most once the user per limitation type
2993 * @param $limitationType String: should be one of:
2994 * 'expensive-parserfunction' (corresponding messages:
2995 * 'expensive-parserfunction-warning',
2996 * 'expensive-parserfunction-category')
2997 * 'post-expand-template-argument' (corresponding messages:
2998 * 'post-expand-template-argument-warning',
2999 * 'post-expand-template-argument-category')
3000 * 'post-expand-template-inclusion' (corresponding messages:
3001 * 'post-expand-template-inclusion-warning',
3002 * 'post-expand-template-inclusion-category')
3003 * @param $current Current value
3004 * @param $max Maximum allowed, when an explicit limit has been
3005 * exceeded, provide the values (optional)
3007 function limitationWarn( $limitationType, $current=null, $max=null) {
3008 # does no harm if $current and $max are present but are unnecessary for the message
3009 $warning = wfMsgExt( "$limitationType-warning", array( 'parsemag', 'escape' ), $current, $max );
3010 $this->mOutput
->addWarning( $warning );
3011 $this->addTrackingCategory( "$limitationType-category" );
3015 * Return the text of a template, after recursively
3016 * replacing any variables or templates within the template.
3018 * @param $piece Array: the parts of the template
3019 * $piece['title']: the title, i.e. the part before the |
3020 * $piece['parts']: the parameter array
3021 * $piece['lineStart']: whether the brace was at the start of a line
3022 * @param $frame PPFrame The current frame, contains template arguments
3023 * @return String: the text of the template
3026 function braceSubstitution( $piece, $frame ) {
3027 global $wgNonincludableNamespaces, $wgEnableInterwikiTranscluding, $wgEnableInterwikiTemplatesTracking;
3028 wfProfileIn( __METHOD__
);
3029 wfProfileIn( __METHOD__
.'-setup' );
3032 $found = false; # $text has been filled
3033 $nowiki = false; # wiki markup in $text should be escaped
3034 $isHTML = false; # $text is HTML, armour it against wikitext transformation
3035 $isChildObj = false; # $text is a DOM node needing expansion in a child frame
3036 $isLocalObj = false; # $text is a DOM node needing expansion in the current frame
3038 # Title object, where $text came from
3041 # $part1 is the bit before the first |, and must contain only title characters.
3042 # Various prefixes will be stripped from it later.
3043 $titleWithSpaces = $frame->expand( $piece['title'] );
3044 $part1 = trim( $titleWithSpaces );
3047 # Original title text preserved for various purposes
3048 $originalTitle = $part1;
3050 # $args is a list of argument nodes, starting from index 0, not including $part1
3051 # @todo FIXME: If piece['parts'] is null then the call to getLength() below won't work b/c this $args isn't an object
3052 $args = ( null == $piece['parts'] ) ?
array() : $piece['parts'];
3053 wfProfileOut( __METHOD__
.'-setup' );
3054 wfProfileIn( __METHOD__
."-title-$originalTitle" );
3057 wfProfileIn( __METHOD__
.'-modifiers' );
3060 $substMatch = $this->mSubstWords
->matchStartAndRemove( $part1 );
3062 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3063 # Decide whether to expand template or keep wikitext as-is.
3064 if ( $this->ot
['wiki'] ) {
3065 if ( $substMatch === false ) {
3066 $literal = true; # literal when in PST with no prefix
3068 $literal = false; # expand when in PST with subst: or safesubst:
3071 if ( $substMatch == 'subst' ) {
3072 $literal = true; # literal when not in PST with plain subst:
3074 $literal = false; # expand when not in PST with safesubst: or no prefix
3078 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3085 if ( !$found && $args->getLength() == 0 ) {
3086 $id = $this->mVariables
->matchStartToEnd( $part1 );
3087 if ( $id !== false ) {
3088 $text = $this->getVariableValue( $id, $frame );
3089 if ( MagicWord
::getCacheTTL( $id ) > -1 ) {
3090 $this->mOutput
->updateCacheExpiry( MagicWord
::getCacheTTL( $id ) );
3096 # MSG, MSGNW and RAW
3099 $mwMsgnw = MagicWord
::get( 'msgnw' );
3100 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
3103 # Remove obsolete MSG:
3104 $mwMsg = MagicWord
::get( 'msg' );
3105 $mwMsg->matchStartAndRemove( $part1 );
3109 $mwRaw = MagicWord
::get( 'raw' );
3110 if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
3111 $forceRawInterwiki = true;
3114 wfProfileOut( __METHOD__
.'-modifiers' );
3118 wfProfileIn( __METHOD__
. '-pfunc' );
3120 $colonPos = strpos( $part1, ':' );
3121 if ( $colonPos !== false ) {
3122 # Case sensitive functions
3123 $function = substr( $part1, 0, $colonPos );
3124 if ( isset( $this->mFunctionSynonyms
[1][$function] ) ) {
3125 $function = $this->mFunctionSynonyms
[1][$function];
3127 # Case insensitive functions
3128 $function = $this->getFunctionLang()->lc( $function );
3129 if ( isset( $this->mFunctionSynonyms
[0][$function] ) ) {
3130 $function = $this->mFunctionSynonyms
[0][$function];
3136 list( $callback, $flags ) = $this->mFunctionHooks
[$function];
3137 $initialArgs = array( &$this );
3138 $funcArgs = array( trim( substr( $part1, $colonPos +
1 ) ) );
3139 if ( $flags & SFH_OBJECT_ARGS
) {
3140 # Add a frame parameter, and pass the arguments as an array
3141 $allArgs = $initialArgs;
3142 $allArgs[] = $frame;
3143 for ( $i = 0; $i < $args->getLength(); $i++
) {
3144 $funcArgs[] = $args->item( $i );
3146 $allArgs[] = $funcArgs;
3148 # Convert arguments to plain text
3149 for ( $i = 0; $i < $args->getLength(); $i++
) {
3150 $funcArgs[] = trim( $frame->expand( $args->item( $i ) ) );
3152 $allArgs = array_merge( $initialArgs, $funcArgs );
3155 # Workaround for PHP bug 35229 and similar
3156 if ( !is_callable( $callback ) ) {
3157 wfProfileOut( __METHOD__
. '-pfunc' );
3158 wfProfileOut( __METHOD__
);
3159 throw new MWException( "Tag hook for $function is not callable\n" );
3161 $result = call_user_func_array( $callback, $allArgs );
3164 $preprocessFlags = 0;
3166 if ( is_array( $result ) ) {
3167 if ( isset( $result[0] ) ) {
3169 unset( $result[0] );
3172 # Extract flags into the local scope
3173 # This allows callers to set flags such as nowiki, found, etc.
3179 $text = $this->preprocessToDom( $text, $preprocessFlags );
3184 wfProfileOut( __METHOD__
. '-pfunc' );
3187 # Finish mangling title and then check for loops.
3188 # Set $title to a Title object and $titleText to the PDBK
3191 # Split the title into page and subpage
3193 $part1 = $this->maybeDoSubpageLink( $part1, $subpage );
3194 if ( $subpage !== '' ) {
3195 $ns = $this->mTitle
->getNamespace();
3197 $title = Title
::newFromText( $part1, $ns );
3199 if ( !$title->isExternal() && $piece['interwiki'] !== '' ) {
3200 $title->setInterwiki( $piece['interwiki'] );
3202 $titleText = $title->getPrefixedText();
3203 # Check for language variants if the template is not found
3204 if ( $this->getFunctionLang()->hasVariants() && $title->getArticleID() == 0 ) {
3205 $this->getFunctionLang()->findVariantLink( $part1, $title, true );
3207 # Do recursion depth check
3208 $limit = $this->mOptions
->getMaxTemplateDepth();
3209 if ( $frame->depth
>= $limit ) {
3211 $text = '<span class="error">'
3212 . wfMsgForContent( 'parser-template-recursion-depth-warning', $limit )
3218 # Load from database
3219 if ( !$found && $title ) {
3220 wfProfileIn( __METHOD__
. '-loadtpl' );
3221 if ( !$title->isExternal() ) {
3222 if ( $title->getNamespace() == NS_SPECIAL
3223 && $this->mOptions
->getAllowSpecialInclusion()
3224 && $this->ot
['html'] )
3226 $pageArgs = array();
3227 for ( $i = 0; $i < $args->getLength(); $i++
) {
3228 $bits = $args->item( $i )->splitArg();
3229 if ( strval( $bits['index'] ) === '' ) {
3230 $name = trim( $frame->expand( $bits['name'], PPFrame
::STRIP_COMMENTS
) );
3231 $value = trim( $frame->expand( $bits['value'] ) );
3232 $pageArgs[$name] = $value;
3235 $context = new RequestContext
;
3236 $context->setTitle( $title );
3237 $context->setRequest( new FauxRequest( $pageArgs ) );
3238 $context->setUser( $this->getUser() );
3239 $context->setLang( Language
::factory( $this->mOptions
->getUserLang() ) );
3240 $ret = SpecialPageFactory
::capturePath( $title, $context );
3242 $text = $context->getOutput()->getHTML();
3243 $this->mOutput
->addOutputPageMetadata( $context->getOutput() );
3246 $this->disableCache();
3248 } elseif ( $wgNonincludableNamespaces && in_array( $title->getNamespace(), $wgNonincludableNamespaces ) ) {
3249 $found = false; # access denied
3250 wfDebug( __METHOD__
.": template inclusion denied for " . $title->getPrefixedDBkey() );
3252 list( $text, $title ) = $this->getTemplateDom( $title );
3253 if ( $text !== false ) {
3259 # If the title is valid but undisplayable, make a link to it
3260 if ( !$found && ( $this->ot
['html'] ||
$this->ot
['pre'] ) ) {
3261 $text = "[[:$titleText]]";
3264 } elseif ( $wgEnableInterwikiTranscluding && $title->isTrans() ) {
3266 $text = Interwiki
::interwikiTransclude( $title );
3267 $this->registerDistantTemplate( $title );
3269 if ( $wgEnableInterwikiTemplatesTracking ) {
3270 $this->registerDistantTemplate( $title );
3273 if ( $text !== false ) {
3274 # Preprocess it like a template
3275 $text = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
3282 # Do infinite loop check
3283 # This has to be done after redirect resolution to avoid infinite loops via redirects
3284 if ( !$frame->loopCheck( $title ) ) {
3286 $text = '<span class="error">' . wfMsgForContent( 'parser-template-loop-warning', $titleText ) . '</span>';
3287 wfDebug( __METHOD__
.": template loop broken at '$titleText'\n" );
3289 wfProfileOut( __METHOD__
. '-loadtpl' );
3292 # If we haven't found text to substitute by now, we're done
3293 # Recover the source wikitext and return it
3295 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3296 wfProfileOut( __METHOD__
."-title-$originalTitle" );
3297 wfProfileOut( __METHOD__
);
3298 return array( 'object' => $text );
3301 # Expand DOM-style return values in a child frame
3302 if ( $isChildObj ) {
3303 # Clean up argument array
3304 $newFrame = $frame->newChild( $args, $title );
3307 $text = $newFrame->expand( $text, PPFrame
::RECOVER_ORIG
);
3308 } elseif ( $titleText !== false && $newFrame->isEmpty() ) {
3309 # Expansion is eligible for the empty-frame cache
3310 if ( isset( $this->mTplExpandCache
[$titleText] ) ) {
3311 $text = $this->mTplExpandCache
[$titleText];
3313 $text = $newFrame->expand( $text );
3314 $this->mTplExpandCache
[$titleText] = $text;
3317 # Uncached expansion
3318 $text = $newFrame->expand( $text );
3321 if ( $isLocalObj && $nowiki ) {
3322 $text = $frame->expand( $text, PPFrame
::RECOVER_ORIG
);
3323 $isLocalObj = false;
3326 # Replace raw HTML by a placeholder
3327 # Add a blank line preceding, to prevent it from mucking up
3328 # immediately preceding headings
3330 $text = "\n\n" . $this->insertStripItem( $text );
3331 } elseif ( $nowiki && ( $this->ot
['html'] ||
$this->ot
['pre'] ) ) {
3332 # Escape nowiki-style return values
3333 $text = wfEscapeWikiText( $text );
3334 } elseif ( is_string( $text )
3335 && !$piece['lineStart']
3336 && preg_match( '/^(?:{\\||:|;|#|\*)/', $text ) )
3338 # Bug 529: if the template begins with a table or block-level
3339 # element, it should be treated as beginning a new line.
3340 # This behaviour is somewhat controversial.
3341 $text = "\n" . $text;
3344 if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
3345 # Error, oversize inclusion
3346 if ( $titleText !== false ) {
3347 # Make a working, properly escaped link if possible (bug 23588)
3348 $text = "[[:$titleText]]";
3350 # This will probably not be a working link, but at least it may
3351 # provide some hint of where the problem is
3352 preg_replace( '/^:/', '', $originalTitle );
3353 $text = "[[:$originalTitle]]";
3355 $text .= $this->insertStripItem( '<!-- WARNING: template omitted, post-expand include size too large -->' );
3356 $this->limitationWarn( 'post-expand-template-inclusion' );
3359 if ( $isLocalObj ) {
3360 $ret = array( 'object' => $text );
3362 $ret = array( 'text' => $text );
3365 wfProfileOut( __METHOD__
."-title-$originalTitle" );
3366 wfProfileOut( __METHOD__
);
3371 * Get the semi-parsed DOM representation of a template with a given title,
3372 * and its redirect destination title. Cached.
3374 * @param $title Title
3378 function getTemplateDom( $title ) {
3379 $cacheTitle = $title;
3380 $titleText = $title->getPrefixedDBkey();
3382 if ( isset( $this->mTplRedirCache
[$titleText] ) ) {
3383 list( $ns, $dbk ) = $this->mTplRedirCache
[$titleText];
3384 $title = Title
::makeTitle( $ns, $dbk );
3385 $titleText = $title->getPrefixedDBkey();
3387 if ( isset( $this->mTplDomCache
[$titleText] ) ) {
3388 return array( $this->mTplDomCache
[$titleText], $title );
3391 # Cache miss, go to the database
3392 list( $text, $title ) = $this->fetchTemplateAndTitle( $title );
3394 if ( $text === false ) {
3395 $this->mTplDomCache
[$titleText] = false;
3396 return array( false, $title );
3399 $dom = $this->preprocessToDom( $text, self
::PTD_FOR_INCLUSION
);
3400 $this->mTplDomCache
[ $titleText ] = $dom;
3402 if ( !$title->equals( $cacheTitle ) ) {
3403 $this->mTplRedirCache
[$cacheTitle->getPrefixedDBkey()] =
3404 array( $title->getNamespace(), $cdb = $title->getDBkey() );
3407 return array( $dom, $title );
3411 * Fetch the unparsed text of a template and register a reference to it.
3412 * @param Title $title
3413 * @return Array ( string or false, Title )
3415 function fetchTemplateAndTitle( $title ) {
3416 $templateCb = $this->mOptions
->getTemplateCallback(); # Defaults to Parser::statelessFetchTemplate()
3417 $stuff = call_user_func( $templateCb, $title, $this );
3418 $text = $stuff['text'];
3419 $finalTitle = isset( $stuff['finalTitle'] ) ?
$stuff['finalTitle'] : $title;
3420 if ( isset( $stuff['deps'] ) ) {
3421 foreach ( $stuff['deps'] as $dep ) {
3422 $this->mOutput
->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
3425 return array( $text, $finalTitle );
3429 * Register a distant template as used
3431 function registerDistantTemplate( $title ) {
3432 $stuff = Parser
::distantTemplateCallback( $title, $this );
3433 $text = $stuff['text'];
3434 $finalTitle = isset( $stuff['finalTitle'] ) ?
$stuff['finalTitle'] : $title;
3435 if ( isset( $stuff['deps'] ) ) {
3436 foreach ( $stuff['deps'] as $dep ) {
3437 $this->mOutput
->addDistantTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
3442 function fetchTemplate( $title ) {
3443 $rv = $this->fetchTemplateAndTitle( $title );
3448 * Static function to get a template
3449 * Can be overridden via ParserOptions::setTemplateCallback().
3451 * @parma $title Title
3452 * @param $parser Parser
3456 static function statelessFetchTemplate( $title, $parser = false ) {
3457 $text = $skip = false;
3458 $finalTitle = $title;
3461 # Loop to fetch the article, with up to 1 redirect
3462 for ( $i = 0; $i < 2 && is_object( $title ); $i++
) {
3463 # Give extensions a chance to select the revision instead
3464 $id = false; # Assume current
3465 wfRunHooks( 'BeforeParserFetchTemplateAndtitle',
3466 array( $parser, $title, &$skip, &$id ) );
3472 'page_id' => $title->getArticleID(),
3479 ? Revision
::newFromId( $id )
3480 : Revision
::newFromTitle( $title );
3481 $rev_id = $rev ?
$rev->getId() : 0;
3482 # If there is no current revision, there is no page
3483 if ( $id === false && !$rev ) {
3484 $linkCache = LinkCache
::singleton();
3485 $linkCache->addBadLinkObj( $title );
3490 'page_id' => $title->getArticleID(),
3491 'rev_id' => $rev_id );
3492 if ( $rev && !$title->equals( $rev->getTitle() ) ) {
3493 # We fetched a rev from a different title; register it too...
3495 'title' => $rev->getTitle(),
3496 'page_id' => $rev->getPage(),
3497 'rev_id' => $rev_id );
3501 $text = $rev->getText();
3502 } elseif ( $title->getNamespace() == NS_MEDIAWIKI
) {
3504 $message = wfMessage( $wgContLang->lcfirst( $title->getText() ) )->inContentLanguage();
3505 if ( !$message->exists() ) {
3509 $text = $message->plain();
3513 if ( $text === false ) {
3517 $finalTitle = $title;
3518 $title = Title
::newFromRedirect( $text );
3522 'finalTitle' => $finalTitle,
3527 * Fetch a file and its title and register a reference to it.
3528 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3529 * @param Title $title
3530 * @param Array $options Array of options to RepoGroup::findFile
3531 * @return File|false
3533 function fetchFile( $title, $options = array() ) {
3534 $res = $this->fetchFileAndTitle( $title, $options );
3539 * Fetch a file and its title and register a reference to it.
3540 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3541 * @param Title $title
3542 * @param Array $options Array of options to RepoGroup::findFile
3543 * @return Array ( File or false, Title of file )
3545 function fetchFileAndTitle( $title, $options = array() ) {
3546 if ( isset( $options['broken'] ) ) {
3547 $file = false; // broken thumbnail forced by hook
3548 } elseif ( isset( $options['sha1'] ) ) { // get by (sha1,timestamp)
3549 $file = RepoGroup
::singleton()->findFileFromKey( $options['sha1'], $options );
3550 } else { // get by (name,timestamp)
3551 $file = wfFindFile( $title, $options );
3553 $time = $file ?
$file->getTimestamp() : false;
3554 $sha1 = $file ?
$file->getSha1() : false;
3555 # Register the file as a dependency...
3556 $this->mOutput
->addImage( $title->getDBkey(), $time, $sha1 );
3557 if ( $file && !$title->equals( $file->getTitle() ) ) {
3558 # Update fetched file title
3559 $title = $file->getTitle();
3560 if ( is_null( $file->getRedirectedTitle() ) ) {
3561 # This file was not a redirect, but the title does not match.
3562 # Register under the new name because otherwise the link will
3564 $this->mOutput
->addImage( $title->getDBkey(), $time, $sha1 );
3567 return array( $file, $title );
3570 static function distantTemplateCallback( $title, $parser=false ) {
3575 'page_id' => $title->getArticleID(),
3576 'rev_id' => $rev_id );
3578 $finalTitle = $title;
3582 'finalTitle' => $finalTitle,
3587 * Triple brace replacement -- used for template arguments
3590 * @param $peice array
3591 * @param $frame PPFrame
3595 function argSubstitution( $piece, $frame ) {
3596 wfProfileIn( __METHOD__
);
3599 $parts = $piece['parts'];
3600 $nameWithSpaces = $frame->expand( $piece['title'] );
3601 $argName = trim( $nameWithSpaces );
3603 $text = $frame->getArgument( $argName );
3604 if ( $text === false && $parts->getLength() > 0
3608 ||
( $this->ot
['wiki'] && $frame->isTemplate() )
3611 # No match in frame, use the supplied default
3612 $object = $parts->item( 0 )->getChildren();
3614 if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
3615 $error = '<!-- WARNING: argument omitted, expansion size too large -->';
3616 $this->limitationWarn( 'post-expand-template-argument' );
3619 if ( $text === false && $object === false ) {
3621 $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts );
3623 if ( $error !== false ) {
3626 if ( $object !== false ) {
3627 $ret = array( 'object' => $object );
3629 $ret = array( 'text' => $text );
3632 wfProfileOut( __METHOD__
);
3637 * Return the text to be used for a given extension tag.
3638 * This is the ghost of strip().
3640 * @param $params Associative array of parameters:
3641 * name PPNode for the tag name
3642 * attr PPNode for unparsed text where tag attributes are thought to be
3643 * attributes Optional associative array of parsed attributes
3644 * inner Contents of extension element
3645 * noClose Original text did not have a close tag
3646 * @param $frame PPFrame
3650 function extensionSubstitution( $params, $frame ) {
3651 $name = $frame->expand( $params['name'] );
3652 $attrText = !isset( $params['attr'] ) ?
null : $frame->expand( $params['attr'] );
3653 $content = !isset( $params['inner'] ) ?
null : $frame->expand( $params['inner'] );
3654 $marker = "{$this->mUniqPrefix}-$name-" . sprintf( '%08X', $this->mMarkerIndex++
) . self
::MARKER_SUFFIX
;
3656 $isFunctionTag = isset( $this->mFunctionTagHooks
[strtolower($name)] ) &&
3657 ( $this->ot
['html'] ||
$this->ot
['pre'] );
3658 if ( $isFunctionTag ) {
3659 $markerType = 'none';
3661 $markerType = 'general';
3663 if ( $this->ot
['html'] ||
$isFunctionTag ) {
3664 $name = strtolower( $name );
3665 $attributes = Sanitizer
::decodeTagAttributes( $attrText );
3666 if ( isset( $params['attributes'] ) ) {
3667 $attributes = $attributes +
$params['attributes'];
3670 if ( isset( $this->mTagHooks
[$name] ) ) {
3671 # Workaround for PHP bug 35229 and similar
3672 if ( !is_callable( $this->mTagHooks
[$name] ) ) {
3673 throw new MWException( "Tag hook for $name is not callable\n" );
3675 $output = call_user_func_array( $this->mTagHooks
[$name],
3676 array( $content, $attributes, $this, $frame ) );
3677 } elseif ( isset( $this->mFunctionTagHooks
[$name] ) ) {
3678 list( $callback, $flags ) = $this->mFunctionTagHooks
[$name];
3679 if ( !is_callable( $callback ) ) {
3680 throw new MWException( "Tag hook for $name is not callable\n" );
3683 $output = call_user_func_array( $callback, array( &$this, $frame, $content, $attributes ) );
3685 $output = '<span class="error">Invalid tag extension name: ' .
3686 htmlspecialchars( $name ) . '</span>';
3689 if ( is_array( $output ) ) {
3690 # Extract flags to local scope (to override $markerType)
3692 $output = $flags[0];
3697 if ( is_null( $attrText ) ) {
3700 if ( isset( $params['attributes'] ) ) {
3701 foreach ( $params['attributes'] as $attrName => $attrValue ) {
3702 $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' .
3703 htmlspecialchars( $attrValue ) . '"';
3706 if ( $content === null ) {
3707 $output = "<$name$attrText/>";
3709 $close = is_null( $params['close'] ) ?
'' : $frame->expand( $params['close'] );
3710 $output = "<$name$attrText>$content$close";
3714 if ( $markerType === 'none' ) {
3716 } elseif ( $markerType === 'nowiki' ) {
3717 $this->mStripState
->addNoWiki( $marker, $output );
3718 } elseif ( $markerType === 'general' ) {
3719 $this->mStripState
->addGeneral( $marker, $output );
3721 throw new MWException( __METHOD__
.': invalid marker type' );
3727 * Increment an include size counter
3729 * @param $type String: the type of expansion
3730 * @param $size Integer: the size of the text
3731 * @return Boolean: false if this inclusion would take it over the maximum, true otherwise
3733 function incrementIncludeSize( $type, $size ) {
3734 if ( $this->mIncludeSizes
[$type] +
$size > $this->mOptions
->getMaxIncludeSize() ) {
3737 $this->mIncludeSizes
[$type] +
= $size;
3743 * Increment the expensive function count
3745 * @return Boolean: false if the limit has been exceeded
3747 function incrementExpensiveFunctionCount() {
3748 global $wgExpensiveParserFunctionLimit;
3749 $this->mExpensiveFunctionCount++
;
3750 if ( $this->mExpensiveFunctionCount
<= $wgExpensiveParserFunctionLimit ) {
3757 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
3758 * Fills $this->mDoubleUnderscores, returns the modified text
3760 * @param $text string
3764 function doDoubleUnderscore( $text ) {
3765 wfProfileIn( __METHOD__
);
3767 # The position of __TOC__ needs to be recorded
3768 $mw = MagicWord
::get( 'toc' );
3769 if ( $mw->match( $text ) ) {
3770 $this->mShowToc
= true;
3771 $this->mForceTocPosition
= true;
3773 # Set a placeholder. At the end we'll fill it in with the TOC.
3774 $text = $mw->replace( '<!--MWTOC-->', $text, 1 );
3776 # Only keep the first one.
3777 $text = $mw->replace( '', $text );
3780 # Now match and remove the rest of them
3781 $mwa = MagicWord
::getDoubleUnderscoreArray();
3782 $this->mDoubleUnderscores
= $mwa->matchAndRemove( $text );
3784 if ( isset( $this->mDoubleUnderscores
['nogallery'] ) ) {
3785 $this->mOutput
->mNoGallery
= true;
3787 if ( isset( $this->mDoubleUnderscores
['notoc'] ) && !$this->mForceTocPosition
) {
3788 $this->mShowToc
= false;
3790 if ( isset( $this->mDoubleUnderscores
['hiddencat'] ) && $this->mTitle
->getNamespace() == NS_CATEGORY
) {
3791 $this->addTrackingCategory( 'hidden-category-category' );
3793 # (bug 8068) Allow control over whether robots index a page.
3795 # @todo FIXME: Bug 14899: __INDEX__ always overrides __NOINDEX__ here! This
3796 # is not desirable, the last one on the page should win.
3797 if ( isset( $this->mDoubleUnderscores
['noindex'] ) && $this->mTitle
->canUseNoindex() ) {
3798 $this->mOutput
->setIndexPolicy( 'noindex' );
3799 $this->addTrackingCategory( 'noindex-category' );
3801 if ( isset( $this->mDoubleUnderscores
['index'] ) && $this->mTitle
->canUseNoindex() ) {
3802 $this->mOutput
->setIndexPolicy( 'index' );
3803 $this->addTrackingCategory( 'index-category' );
3806 # Cache all double underscores in the database
3807 foreach ( $this->mDoubleUnderscores
as $key => $val ) {
3808 $this->mOutput
->setProperty( $key, '' );
3811 wfProfileOut( __METHOD__
);
3816 * Add a tracking category, getting the title from a system message,
3817 * or print a debug message if the title is invalid.
3819 * @param $msg String: message key
3820 * @return Boolean: whether the addition was successful
3822 protected function addTrackingCategory( $msg ) {
3823 if ( $this->mTitle
->getNamespace() === NS_SPECIAL
) {
3824 wfDebug( __METHOD__
.": Not adding tracking category $msg to special page!\n" );
3827 $cat = wfMsgForContent( $msg );
3829 # Allow tracking categories to be disabled by setting them to "-"
3830 if ( $cat === '-' ) {
3834 $containerCategory = Title
::makeTitleSafe( NS_CATEGORY
, $cat );
3835 if ( $containerCategory ) {
3836 $this->mOutput
->addCategory( $containerCategory->getDBkey(), $this->getDefaultSort() );
3839 wfDebug( __METHOD__
.": [[MediaWiki:$msg]] is not a valid title!\n" );
3845 * This function accomplishes several tasks:
3846 * 1) Auto-number headings if that option is enabled
3847 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
3848 * 3) Add a Table of contents on the top for users who have enabled the option
3849 * 4) Auto-anchor headings
3851 * It loops through all headlines, collects the necessary data, then splits up the
3852 * string and re-inserts the newly formatted headlines.
3854 * @param $text String
3855 * @param $origText String: original, untouched wikitext
3856 * @param $isMain Boolean
3859 function formatHeadings( $text, $origText, $isMain=true ) {
3860 global $wgMaxTocLevel, $wgHtml5, $wgExperimentalHtmlIds;
3862 # Inhibit editsection links if requested in the page
3863 if ( isset( $this->mDoubleUnderscores
['noeditsection'] ) ) {
3866 $showEditLink = $this->mOptions
->getEditSection();
3868 if ( $showEditLink ) {
3869 $this->mOutput
->setEditSectionTokens( true );
3872 # Get all headlines for numbering them and adding funky stuff like [edit]
3873 # links - this is for later, but we need the number of headlines right now
3875 $numMatches = preg_match_all( '/<H(?P<level>[1-6])(?P<attrib>.*?'.'>)(?P<header>.*?)<\/H[1-6] *>/i', $text, $matches );
3877 # if there are fewer than 4 headlines in the article, do not show TOC
3878 # unless it's been explicitly enabled.
3879 $enoughToc = $this->mShowToc
&&
3880 ( ( $numMatches >= 4 ) ||
$this->mForceTocPosition
);
3882 # Allow user to stipulate that a page should have a "new section"
3883 # link added via __NEWSECTIONLINK__
3884 if ( isset( $this->mDoubleUnderscores
['newsectionlink'] ) ) {
3885 $this->mOutput
->setNewSection( true );
3888 # Allow user to remove the "new section"
3889 # link via __NONEWSECTIONLINK__
3890 if ( isset( $this->mDoubleUnderscores
['nonewsectionlink'] ) ) {
3891 $this->mOutput
->hideNewSection( true );
3894 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
3895 # override above conditions and always show TOC above first header
3896 if ( isset( $this->mDoubleUnderscores
['forcetoc'] ) ) {
3897 $this->mShowToc
= true;
3905 # Ugh .. the TOC should have neat indentation levels which can be
3906 # passed to the skin functions. These are determined here
3910 $sublevelCount = array();
3911 $levelCount = array();
3916 $markerRegex = "{$this->mUniqPrefix}-h-(\d+)-" . self
::MARKER_SUFFIX
;
3917 $baseTitleText = $this->mTitle
->getPrefixedDBkey();
3918 $oldType = $this->mOutputType
;
3919 $this->setOutputType( self
::OT_WIKI
);
3920 $frame = $this->getPreprocessor()->newFrame();
3921 $root = $this->preprocessToDom( $origText );
3922 $node = $root->getFirstChild();
3927 foreach ( $matches[3] as $headline ) {
3928 $isTemplate = false;
3930 $sectionIndex = false;
3932 $markerMatches = array();
3933 if ( preg_match("/^$markerRegex/", $headline, $markerMatches ) ) {
3934 $serial = $markerMatches[1];
3935 list( $titleText, $sectionIndex ) = $this->mHeadings
[$serial];
3936 $isTemplate = ( $titleText != $baseTitleText );
3937 $headline = preg_replace( "/^$markerRegex/", "", $headline );
3941 $prevlevel = $level;
3943 $level = $matches[1][$headlineCount];
3945 if ( $level > $prevlevel ) {
3946 # Increase TOC level
3948 $sublevelCount[$toclevel] = 0;
3949 if ( $toclevel<$wgMaxTocLevel ) {
3950 $prevtoclevel = $toclevel;
3951 $toc .= Linker
::tocIndent();
3954 } elseif ( $level < $prevlevel && $toclevel > 1 ) {
3955 # Decrease TOC level, find level to jump to
3957 for ( $i = $toclevel; $i > 0; $i-- ) {
3958 if ( $levelCount[$i] == $level ) {
3959 # Found last matching level
3962 } elseif ( $levelCount[$i] < $level ) {
3963 # Found first matching level below current level
3971 if ( $toclevel<$wgMaxTocLevel ) {
3972 if ( $prevtoclevel < $wgMaxTocLevel ) {
3973 # Unindent only if the previous toc level was shown :p
3974 $toc .= Linker
::tocUnindent( $prevtoclevel - $toclevel );
3975 $prevtoclevel = $toclevel;
3977 $toc .= Linker
::tocLineEnd();
3981 # No change in level, end TOC line
3982 if ( $toclevel<$wgMaxTocLevel ) {
3983 $toc .= Linker
::tocLineEnd();
3987 $levelCount[$toclevel] = $level;
3989 # count number of headlines for each level
3990 @$sublevelCount[$toclevel]++
;
3992 for( $i = 1; $i <= $toclevel; $i++
) {
3993 if ( !empty( $sublevelCount[$i] ) ) {
3997 $numbering .= $this->getFunctionLang()->formatNum( $sublevelCount[$i] );
4002 # The safe header is a version of the header text safe to use for links
4003 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4004 $safeHeadline = $this->mStripState
->unstripBoth( $headline );
4006 # Remove link placeholders by the link text.
4007 # <!--LINK number-->
4009 # link text with suffix
4010 $safeHeadline = $this->replaceLinkHoldersText( $safeHeadline );
4012 # Strip out HTML (other than plain <sup> and <sub>: bug 8393)
4013 $tocline = preg_replace(
4014 array( '#<(?!/?(sup|sub)).*?'.'>#', '#<(/?(sup|sub)).*?'.'>#' ),
4015 array( '', '<$1>' ),
4018 $tocline = trim( $tocline );
4020 # For the anchor, strip out HTML-y stuff period
4021 $safeHeadline = preg_replace( '/<.*?'.'>/', '', $safeHeadline );
4022 $safeHeadline = Sanitizer
::normalizeSectionNameWhitespace( $safeHeadline );
4024 # Save headline for section edit hint before it's escaped
4025 $headlineHint = $safeHeadline;
4027 if ( $wgHtml5 && $wgExperimentalHtmlIds ) {
4028 # For reverse compatibility, provide an id that's
4029 # HTML4-compatible, like we used to.
4031 # It may be worth noting, academically, that it's possible for
4032 # the legacy anchor to conflict with a non-legacy headline
4033 # anchor on the page. In this case likely the "correct" thing
4034 # would be to either drop the legacy anchors or make sure
4035 # they're numbered first. However, this would require people
4036 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
4037 # manually, so let's not bother worrying about it.
4038 $legacyHeadline = Sanitizer
::escapeId( $safeHeadline,
4039 array( 'noninitial', 'legacy' ) );
4040 $safeHeadline = Sanitizer
::escapeId( $safeHeadline );
4042 if ( $legacyHeadline == $safeHeadline ) {
4043 # No reason to have both (in fact, we can't)
4044 $legacyHeadline = false;
4047 $legacyHeadline = false;
4048 $safeHeadline = Sanitizer
::escapeId( $safeHeadline,
4052 # HTML names must be case-insensitively unique (bug 10721).
4053 # This does not apply to Unicode characters per
4054 # http://dev.w3.org/html5/spec/infrastructure.html#case-sensitivity-and-string-comparison
4055 # @todo FIXME: We may be changing them depending on the current locale.
4056 $arrayKey = strtolower( $safeHeadline );
4057 if ( $legacyHeadline === false ) {
4058 $legacyArrayKey = false;
4060 $legacyArrayKey = strtolower( $legacyHeadline );
4063 # count how many in assoc. array so we can track dupes in anchors
4064 if ( isset( $refers[$arrayKey] ) ) {
4065 $refers[$arrayKey]++
;
4067 $refers[$arrayKey] = 1;
4069 if ( isset( $refers[$legacyArrayKey] ) ) {
4070 $refers[$legacyArrayKey]++
;
4072 $refers[$legacyArrayKey] = 1;
4075 # Don't number the heading if it is the only one (looks silly)
4076 if ( count( $matches[3] ) > 1 && $this->mOptions
->getNumberHeadings() ) {
4077 # the two are different if the line contains a link
4078 $headline = $numbering . ' ' . $headline;
4081 # Create the anchor for linking from the TOC to the section
4082 $anchor = $safeHeadline;
4083 $legacyAnchor = $legacyHeadline;
4084 if ( $refers[$arrayKey] > 1 ) {
4085 $anchor .= '_' . $refers[$arrayKey];
4087 if ( $legacyHeadline !== false && $refers[$legacyArrayKey] > 1 ) {
4088 $legacyAnchor .= '_' . $refers[$legacyArrayKey];
4090 if ( $enoughToc && ( !isset( $wgMaxTocLevel ) ||
$toclevel < $wgMaxTocLevel ) ) {
4091 $toc .= Linker
::tocLine( $anchor, $tocline,
4092 $numbering, $toclevel, ( $isTemplate ?
false : $sectionIndex ) );
4095 # Add the section to the section tree
4096 # Find the DOM node for this header
4097 while ( $node && !$isTemplate ) {
4098 if ( $node->getName() === 'h' ) {
4099 $bits = $node->splitHeading();
4100 if ( $bits['i'] == $sectionIndex ) {
4104 $byteOffset +
= mb_strlen( $this->mStripState
->unstripBoth(
4105 $frame->expand( $node, PPFrame
::RECOVER_ORIG
) ) );
4106 $node = $node->getNextSibling();
4109 'toclevel' => $toclevel,
4112 'number' => $numbering,
4113 'index' => ( $isTemplate ?
'T-' : '' ) . $sectionIndex,
4114 'fromtitle' => $titleText,
4115 'byteoffset' => ( $isTemplate ?
null : $byteOffset ),
4116 'anchor' => $anchor,
4119 # give headline the correct <h#> tag
4120 if ( $showEditLink && $sectionIndex !== false ) {
4121 // Output edit section links as markers with styles that can be customized by skins
4122 if ( $isTemplate ) {
4123 # Put a T flag in the section identifier, to indicate to extractSections()
4124 # that sections inside <includeonly> should be counted.
4125 $editlinkArgs = array( $titleText, "T-$sectionIndex"/*, null */ );
4127 $editlinkArgs = array( $this->mTitle
->getPrefixedText(), $sectionIndex, $headlineHint );
4129 // We use a bit of pesudo-xml for editsection markers. The language converter is run later on
4130 // Using a UNIQ style marker leads to the converter screwing up the tokens when it converts stuff
4131 // And trying to insert strip tags fails too. At this point all real inputted tags have already been escaped
4132 // so we don't have to worry about a user trying to input one of these markers directly.
4133 // We use a page and section attribute to stop the language converter from converting these important bits
4134 // of data, but put the headline hint inside a content block because the language converter is supposed to
4135 // be able to convert that piece of data.
4136 $editlink = '<mw:editsection page="' . htmlspecialchars($editlinkArgs[0]);
4137 $editlink .= '" section="' . htmlspecialchars($editlinkArgs[1]) .'"';
4138 if ( isset($editlinkArgs[2]) ) {
4139 $editlink .= '>' . $editlinkArgs[2] . '</mw:editsection>';
4146 $head[$headlineCount] = Linker
::makeHeadline( $level,
4147 $matches['attrib'][$headlineCount], $anchor, $headline,
4148 $editlink, $legacyAnchor );
4153 $this->setOutputType( $oldType );
4155 # Never ever show TOC if no headers
4156 if ( $numVisible < 1 ) {
4161 if ( $prevtoclevel > 0 && $prevtoclevel < $wgMaxTocLevel ) {
4162 $toc .= Linker
::tocUnindent( $prevtoclevel - 1 );
4164 $toc = Linker
::tocList( $toc, $this->mOptions
->getUserLang() );
4165 $this->mOutput
->setTOCHTML( $toc );
4169 $this->mOutput
->setSections( $tocraw );
4172 # split up and insert constructed headlines
4173 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
4176 // build an array of document sections
4177 $sections = array();
4178 foreach ( $blocks as $block ) {
4179 // $head is zero-based, sections aren't.
4180 if ( empty( $head[$i - 1] ) ) {
4181 $sections[$i] = $block;
4183 $sections[$i] = $head[$i - 1] . $block;
4187 * Send a hook, one per section.
4188 * The idea here is to be able to make section-level DIVs, but to do so in a
4189 * lower-impact, more correct way than r50769
4192 * $section : the section number
4193 * &$sectionContent : ref to the content of the section
4194 * $showEditLinks : boolean describing whether this section has an edit link
4196 wfRunHooks( 'ParserSectionCreate', array( $this, $i, &$sections[$i], $showEditLink ) );
4201 if ( $enoughToc && $isMain && !$this->mForceTocPosition
) {
4202 // append the TOC at the beginning
4203 // Top anchor now in skin
4204 $sections[0] = $sections[0] . $toc . "\n";
4207 $full .= join( '', $sections );
4209 if ( $this->mForceTocPosition
) {
4210 return str_replace( '<!--MWTOC-->', $toc, $full );
4217 * Transform wiki markup when saving a page by doing \r\n -> \n
4218 * conversion, substitting signatures, {{subst:}} templates, etc.
4220 * @param $text String: the text to transform
4221 * @param $title Title: the Title object for the current article
4222 * @param $user User: the User object describing the current user
4223 * @param $options ParserOptions: parsing options
4224 * @param $clearState Boolean: whether to clear the parser state first
4225 * @return String: the altered wiki markup
4227 public function preSaveTransform( $text, Title
$title, User
$user, ParserOptions
$options, $clearState = true ) {
4228 $this->startParse( $title, $options, self
::OT_WIKI
, $clearState );
4229 $this->setUser( $user );
4234 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
4235 if( $options->getPreSaveTransform() ) {
4236 $text = $this->pstPass2( $text, $user );
4238 $text = $this->mStripState
->unstripBoth( $text );
4240 $this->setUser( null ); #Reset
4246 * Pre-save transform helper function
4249 * @param $text string
4254 function pstPass2( $text, $user ) {
4255 global $wgContLang, $wgLocaltimezone;
4257 # Note: This is the timestamp saved as hardcoded wikitext to
4258 # the database, we use $wgContLang here in order to give
4259 # everyone the same signature and use the default one rather
4260 # than the one selected in each user's preferences.
4261 # (see also bug 12815)
4262 $ts = $this->mOptions
->getTimestamp();
4263 if ( isset( $wgLocaltimezone ) ) {
4264 $tz = $wgLocaltimezone;
4266 $tz = date_default_timezone_get();
4269 $unixts = wfTimestamp( TS_UNIX
, $ts );
4270 $oldtz = date_default_timezone_get();
4271 date_default_timezone_set( $tz );
4272 $ts = date( 'YmdHis', $unixts );
4273 $tzMsg = date( 'T', $unixts ); # might vary on DST changeover!
4275 # Allow translation of timezones through wiki. date() can return
4276 # whatever crap the system uses, localised or not, so we cannot
4277 # ship premade translations.
4278 $key = 'timezone-' . strtolower( trim( $tzMsg ) );
4279 $msg = wfMessage( $key )->inContentLanguage();
4280 if ( $msg->exists() ) {
4281 $tzMsg = $msg->text();
4284 date_default_timezone_set( $oldtz );
4286 $d = $wgContLang->timeanddate( $ts, false, false ) . " ($tzMsg)";
4288 # Variable replacement
4289 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4290 $text = $this->replaceVariables( $text );
4292 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4293 # which may corrupt this parser instance via its wfMsgExt( parsemag ) call-
4296 $sigText = $this->getUserSig( $user );
4297 $text = strtr( $text, array(
4299 '~~~~' => "$sigText $d",
4303 # Context links: [[|name]] and [[name (context)|]]
4304 global $wgLegalTitleChars;
4305 $tc = "[$wgLegalTitleChars]";
4306 $nc = '[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can use non-ascii!
4308 $p1 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\))\\|]]/"; # [[ns:page (context)|]]
4309 $p4 = "/\[\[(:?$nc+:|:|)($tc+?)( ?($tc+))\\|]]/"; # [[ns:page(context)|]]
4310 $p3 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\)|)(, $tc+|)\\|]]/"; # [[ns:page (context), context|]]
4311 $p2 = "/\[\[\\|($tc+)]]/"; # [[|page]]
4313 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4314 $text = preg_replace( $p1, '[[\\1\\2\\3|\\2]]', $text );
4315 $text = preg_replace( $p4, '[[\\1\\2\\3|\\2]]', $text );
4316 $text = preg_replace( $p3, '[[\\1\\2\\3\\4|\\2]]', $text );
4318 $t = $this->mTitle
->getText();
4320 if ( preg_match( "/^($nc+:|)$tc+?( \\($tc+\\))$/", $t, $m ) ) {
4321 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4322 } elseif ( preg_match( "/^($nc+:|)$tc+?(, $tc+|)$/", $t, $m ) && "$m[1]$m[2]" != '' ) {
4323 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4325 # if there's no context, don't bother duplicating the title
4326 $text = preg_replace( $p2, '[[\\1]]', $text );
4329 # Trim trailing whitespace
4330 $text = rtrim( $text );
4336 * Fetch the user's signature text, if any, and normalize to
4337 * validated, ready-to-insert wikitext.
4338 * If you have pre-fetched the nickname or the fancySig option, you can
4339 * specify them here to save a database query.
4340 * Do not reuse this parser instance after calling getUserSig(),
4341 * as it may have changed if it's the $wgParser.
4344 * @param $nickname String|bool nickname to use or false to use user's default nickname
4345 * @param $fancySig Boolean|null whether the nicknname is the complete signature
4346 * or null to use default value
4349 function getUserSig( &$user, $nickname = false, $fancySig = null ) {
4350 global $wgMaxSigChars;
4352 $username = $user->getName();
4354 # If not given, retrieve from the user object.
4355 if ( $nickname === false )
4356 $nickname = $user->getOption( 'nickname' );
4358 if ( is_null( $fancySig ) ) {
4359 $fancySig = $user->getBoolOption( 'fancysig' );
4362 $nickname = $nickname == null ?
$username : $nickname;
4364 if ( mb_strlen( $nickname ) > $wgMaxSigChars ) {
4365 $nickname = $username;
4366 wfDebug( __METHOD__
. ": $username has overlong signature.\n" );
4367 } elseif ( $fancySig !== false ) {
4368 # Sig. might contain markup; validate this
4369 if ( $this->validateSig( $nickname ) !== false ) {
4370 # Validated; clean up (if needed) and return it
4371 return $this->cleanSig( $nickname, true );
4373 # Failed to validate; fall back to the default
4374 $nickname = $username;
4375 wfDebug( __METHOD__
.": $username has bad XML tags in signature.\n" );
4379 # Make sure nickname doesnt get a sig in a sig
4380 $nickname = $this->cleanSigInSig( $nickname );
4382 # If we're still here, make it a link to the user page
4383 $userText = wfEscapeWikiText( $username );
4384 $nickText = wfEscapeWikiText( $nickname );
4385 $msgName = $user->isAnon() ?
'signature-anon' : 'signature';
4387 return wfMessage( $msgName, $userText, $nickText )->inContentLanguage()->title( $this->getTitle() )->text();
4391 * Check that the user's signature contains no bad XML
4393 * @param $text String
4394 * @return mixed An expanded string, or false if invalid.
4396 function validateSig( $text ) {
4397 return( Xml
::isWellFormedXmlFragment( $text ) ?
$text : false );
4401 * Clean up signature text
4403 * 1) Strip ~~~, ~~~~ and ~~~~~ out of signatures @see cleanSigInSig
4404 * 2) Substitute all transclusions
4406 * @param $text String
4407 * @param $parsing bool Whether we're cleaning (preferences save) or parsing
4408 * @return String: signature text
4410 function cleanSig( $text, $parsing = false ) {
4413 $this->startParse( $wgTitle, new ParserOptions
, self
::OT_PREPROCESS
, true );
4416 # Option to disable this feature
4417 if ( !$this->mOptions
->getCleanSignatures() ) {
4421 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4422 # => Move this logic to braceSubstitution()
4423 $substWord = MagicWord
::get( 'subst' );
4424 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4425 $substText = '{{' . $substWord->getSynonym( 0 );
4427 $text = preg_replace( $substRegex, $substText, $text );
4428 $text = $this->cleanSigInSig( $text );
4429 $dom = $this->preprocessToDom( $text );
4430 $frame = $this->getPreprocessor()->newFrame();
4431 $text = $frame->expand( $dom );
4434 $text = $this->mStripState
->unstripBoth( $text );
4441 * Strip ~~~, ~~~~ and ~~~~~ out of signatures
4443 * @param $text String
4444 * @return String: signature text with /~{3,5}/ removed
4446 function cleanSigInSig( $text ) {
4447 $text = preg_replace( '/~{3,5}/', '', $text );
4452 * Set up some variables which are usually set up in parse()
4453 * so that an external function can call some class members with confidence
4455 * @param $title Title|null
4456 * @param $options ParserOptions
4457 * @param $outputType
4458 * @param $clearState bool
4460 public function startExternalParse( Title
$title = null, ParserOptions
$options, $outputType, $clearState = true ) {
4461 $this->startParse( $title, $options, $outputType, $clearState );
4465 * @param $title Title|null
4466 * @param $options ParserOptions
4467 * @param $outputType
4468 * @param $clearState bool
4470 private function startParse( Title
$title = null, ParserOptions
$options, $outputType, $clearState = true ) {
4471 $this->setTitle( $title );
4472 $this->mOptions
= $options;
4473 $this->setOutputType( $outputType );
4474 if ( $clearState ) {
4475 $this->clearState();
4480 * Wrapper for preprocess()
4482 * @param $text String: the text to preprocess
4483 * @param $options ParserOptions: options
4484 * @param $title Title object or null to use $wgTitle
4487 public function transformMsg( $text, $options, $title = null ) {
4488 static $executing = false;
4490 # Guard against infinite recursion
4496 wfProfileIn( __METHOD__
);
4502 # It's not uncommon having a null $wgTitle in scripts. See r80898
4503 # Create a ghost title in such case
4504 $title = Title
::newFromText( 'Dwimmerlaik' );
4506 $text = $this->preprocess( $text, $title, $options );
4509 wfProfileOut( __METHOD__
);
4514 * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
4515 * The callback should have the following form:
4516 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4518 * Transform and return $text. Use $parser for any required context, e.g. use
4519 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4521 * Hooks may return extended information by returning an array, of which the
4522 * first numbered element (index 0) must be the return string, and all other
4523 * entries are extracted into local variables within an internal function
4524 * in the Parser class.
4526 * This interface (introduced r61913) appears to be undocumented, but
4527 * 'markerName' is used by some core tag hooks to override which strip
4528 * array their results are placed in. **Use great caution if attempting
4529 * this interface, as it is not documented and injudicious use could smash
4530 * private variables.**
4532 * @param $tag Mixed: the tag to use, e.g. 'hook' for <hook>
4533 * @param $callback Mixed: the callback function (and object) to use for the tag
4534 * @return The old value of the mTagHooks array associated with the hook
4536 public function setHook( $tag, $callback ) {
4537 $tag = strtolower( $tag );
4538 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) throw new MWException( "Invalid character {$m[0]} in setHook('$tag', ...) call" );
4539 $oldVal = isset( $this->mTagHooks
[$tag] ) ?
$this->mTagHooks
[$tag] : null;
4540 $this->mTagHooks
[$tag] = $callback;
4541 if ( !in_array( $tag, $this->mStripList
) ) {
4542 $this->mStripList
[] = $tag;
4549 * As setHook(), but letting the contents be parsed.
4551 * Transparent tag hooks are like regular XML-style tag hooks, except they
4552 * operate late in the transformation sequence, on HTML instead of wikitext.
4554 * This is probably obsoleted by things dealing with parser frames?
4555 * The only extension currently using it is geoserver.
4558 * @todo better document or deprecate this
4560 * @param $tag Mixed: the tag to use, e.g. 'hook' for <hook>
4561 * @param $callback Mixed: the callback function (and object) to use for the tag
4562 * @return The old value of the mTagHooks array associated with the hook
4564 function setTransparentTagHook( $tag, $callback ) {
4565 $tag = strtolower( $tag );
4566 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) throw new MWException( "Invalid character {$m[0]} in setTransparentHook('$tag', ...) call" );
4567 $oldVal = isset( $this->mTransparentTagHooks
[$tag] ) ?
$this->mTransparentTagHooks
[$tag] : null;
4568 $this->mTransparentTagHooks
[$tag] = $callback;
4574 * Remove all tag hooks
4576 function clearTagHooks() {
4577 $this->mTagHooks
= array();
4578 $this->mStripList
= $this->mDefaultStripList
;
4582 * Create a function, e.g. {{sum:1|2|3}}
4583 * The callback function should have the form:
4584 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4586 * Or with SFH_OBJECT_ARGS:
4587 * function myParserFunction( $parser, $frame, $args ) { ... }
4589 * The callback may either return the text result of the function, or an array with the text
4590 * in element 0, and a number of flags in the other elements. The names of the flags are
4591 * specified in the keys. Valid flags are:
4592 * found The text returned is valid, stop processing the template. This
4594 * nowiki Wiki markup in the return value should be escaped
4595 * isHTML The returned text is HTML, armour it against wikitext transformation
4597 * @param $id String: The magic word ID
4598 * @param $callback Mixed: the callback function (and object) to use
4599 * @param $flags Integer: a combination of the following flags:
4600 * SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4602 * SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text. This
4603 * allows for conditional expansion of the parse tree, allowing you to eliminate dead
4604 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4605 * the arguments, and to control the way they are expanded.
4607 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4608 * arguments, for instance:
4609 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4611 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4612 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4613 * working if/when this is changed.
4615 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4618 * Please read the documentation in includes/parser/Preprocessor.php for more information
4619 * about the methods available in PPFrame and PPNode.
4621 * @return The old callback function for this name, if any
4623 public function setFunctionHook( $id, $callback, $flags = 0 ) {
4626 $oldVal = isset( $this->mFunctionHooks
[$id] ) ?
$this->mFunctionHooks
[$id][0] : null;
4627 $this->mFunctionHooks
[$id] = array( $callback, $flags );
4629 # Add to function cache
4630 $mw = MagicWord
::get( $id );
4632 throw new MWException( __METHOD__
.'() expecting a magic word identifier.' );
4634 $synonyms = $mw->getSynonyms();
4635 $sensitive = intval( $mw->isCaseSensitive() );
4637 foreach ( $synonyms as $syn ) {
4639 if ( !$sensitive ) {
4640 $syn = $wgContLang->lc( $syn );
4643 if ( !( $flags & SFH_NO_HASH
) ) {
4646 # Remove trailing colon
4647 if ( substr( $syn, -1, 1 ) === ':' ) {
4648 $syn = substr( $syn, 0, -1 );
4650 $this->mFunctionSynonyms
[$sensitive][$syn] = $id;
4656 * Get all registered function hook identifiers
4660 function getFunctionHooks() {
4661 return array_keys( $this->mFunctionHooks
);
4665 * Create a tag function, e.g. <test>some stuff</test>.
4666 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4667 * Unlike parser functions, their content is not preprocessed.
4669 function setFunctionTagHook( $tag, $callback, $flags ) {
4670 $tag = strtolower( $tag );
4671 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) throw new MWException( "Invalid character {$m[0]} in setFunctionTagHook('$tag', ...) call" );
4672 $old = isset( $this->mFunctionTagHooks
[$tag] ) ?
4673 $this->mFunctionTagHooks
[$tag] : null;
4674 $this->mFunctionTagHooks
[$tag] = array( $callback, $flags );
4676 if ( !in_array( $tag, $this->mStripList
) ) {
4677 $this->mStripList
[] = $tag;
4684 * @todo FIXME: Update documentation. makeLinkObj() is deprecated.
4685 * Replace <!--LINK--> link placeholders with actual links, in the buffer
4686 * Placeholders created in Skin::makeLinkObj()
4688 * @param $text string
4689 * @param $options int
4691 * @return array of link CSS classes, indexed by PDBK.
4693 function replaceLinkHolders( &$text, $options = 0 ) {
4694 return $this->mLinkHolders
->replace( $text );
4698 * Replace <!--LINK--> link placeholders with plain text of links
4699 * (not HTML-formatted).
4701 * @param $text String
4704 function replaceLinkHoldersText( $text ) {
4705 return $this->mLinkHolders
->replaceText( $text );
4709 * Renders an image gallery from a text with one line per image.
4710 * text labels may be given by using |-style alternative text. E.g.
4711 * Image:one.jpg|The number "1"
4712 * Image:tree.jpg|A tree
4713 * given as text will return the HTML of a gallery with two images,
4714 * labeled 'The number "1"' and
4717 * @param string $text
4718 * @param array $params
4719 * @return string HTML
4721 function renderImageGallery( $text, $params ) {
4722 $ig = new ImageGallery();
4723 $ig->setContextTitle( $this->mTitle
);
4724 $ig->setShowBytes( false );
4725 $ig->setShowFilename( false );
4726 $ig->setParser( $this );
4727 $ig->setHideBadImages();
4728 $ig->setAttributes( Sanitizer
::validateTagAttributes( $params, 'table' ) );
4730 if ( isset( $params['showfilename'] ) ) {
4731 $ig->setShowFilename( true );
4733 $ig->setShowFilename( false );
4735 if ( isset( $params['caption'] ) ) {
4736 $caption = $params['caption'];
4737 $caption = htmlspecialchars( $caption );
4738 $caption = $this->replaceInternalLinks( $caption );
4739 $ig->setCaptionHtml( $caption );
4741 if ( isset( $params['perrow'] ) ) {
4742 $ig->setPerRow( $params['perrow'] );
4744 if ( isset( $params['widths'] ) ) {
4745 $ig->setWidths( $params['widths'] );
4747 if ( isset( $params['heights'] ) ) {
4748 $ig->setHeights( $params['heights'] );
4751 wfRunHooks( 'BeforeParserrenderImageGallery', array( &$this, &$ig ) );
4753 $lines = StringUtils
::explode( "\n", $text );
4754 foreach ( $lines as $line ) {
4755 # match lines like these:
4756 # Image:someimage.jpg|This is some image
4758 preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
4760 if ( count( $matches ) == 0 ) {
4764 if ( strpos( $matches[0], '%' ) !== false ) {
4765 $matches[1] = rawurldecode( $matches[1] );
4767 $title = Title
::newFromText( $matches[1], NS_FILE
);
4768 if ( is_null( $title ) ) {
4769 # Bogus title. Ignore these so we don't bomb out later.
4775 if ( isset( $matches[3] ) ) {
4776 // look for an |alt= definition while trying not to break existing
4777 // captions with multiple pipes (|) in it, until a more sensible grammar
4778 // is defined for images in galleries
4780 $matches[3] = $this->recursiveTagParse( trim( $matches[3] ) );
4781 $altmatches = StringUtils
::explode('|', $matches[3]);
4782 $magicWordAlt = MagicWord
::get( 'img_alt' );
4784 foreach ( $altmatches as $altmatch ) {
4785 $match = $magicWordAlt->matchVariableStartToEnd( $altmatch );
4787 $alt = $this->stripAltText( $match, false );
4790 // concatenate all other pipes
4791 $label .= '|' . $altmatch;
4794 // remove the first pipe
4795 $label = substr( $label, 1 );
4798 $ig->add( $title, $label, $alt );
4800 return $ig->toHTML();
4807 function getImageParams( $handler ) {
4809 $handlerClass = get_class( $handler );
4813 if ( !isset( $this->mImageParams
[$handlerClass] ) ) {
4814 # Initialise static lists
4815 static $internalParamNames = array(
4816 'horizAlign' => array( 'left', 'right', 'center', 'none' ),
4817 'vertAlign' => array( 'baseline', 'sub', 'super', 'top', 'text-top', 'middle',
4818 'bottom', 'text-bottom' ),
4819 'frame' => array( 'thumbnail', 'manualthumb', 'framed', 'frameless',
4820 'upright', 'border', 'link', 'alt' ),
4822 static $internalParamMap;
4823 if ( !$internalParamMap ) {
4824 $internalParamMap = array();
4825 foreach ( $internalParamNames as $type => $names ) {
4826 foreach ( $names as $name ) {
4827 $magicName = str_replace( '-', '_', "img_$name" );
4828 $internalParamMap[$magicName] = array( $type, $name );
4833 # Add handler params
4834 $paramMap = $internalParamMap;
4836 $handlerParamMap = $handler->getParamMap();
4837 foreach ( $handlerParamMap as $magic => $paramName ) {
4838 $paramMap[$magic] = array( 'handler', $paramName );
4841 $this->mImageParams
[$handlerClass] = $paramMap;
4842 $this->mImageParamsMagicArray
[$handlerClass] = new MagicWordArray( array_keys( $paramMap ) );
4844 return array( $this->mImageParams
[$handlerClass], $this->mImageParamsMagicArray
[$handlerClass] );
4848 * Parse image options text and use it to make an image
4850 * @param $title Title
4851 * @param $options String
4852 * @param $holders LinkHolderArray|false
4853 * @return string HTML
4855 function makeImage( $title, $options, $holders = false ) {
4856 # Check if the options text is of the form "options|alt text"
4858 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
4859 # * left no resizing, just left align. label is used for alt= only
4860 # * right same, but right aligned
4861 # * none same, but not aligned
4862 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
4863 # * center center the image
4864 # * frame Keep original image size, no magnify-button.
4865 # * framed Same as "frame"
4866 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
4867 # * upright reduce width for upright images, rounded to full __0 px
4868 # * border draw a 1px border around the image
4869 # * alt Text for HTML alt attribute (defaults to empty)
4870 # * link Set the target of the image link. Can be external, interwiki, or local
4871 # vertical-align values (no % or length right now):
4881 $parts = StringUtils
::explode( "|", $options );
4883 # Give extensions a chance to select the file revision for us
4886 wfRunHooks( 'BeforeParserFetchFileAndTitle',
4887 array( $this, $title, &$options, &$descQuery ) );
4888 # Fetch and register the file (file title may be different via hooks)
4889 list( $file, $title ) = $this->fetchFileAndTitle( $title, $options );
4892 $handler = $file ?
$file->getHandler() : false;
4894 list( $paramMap, $mwArray ) = $this->getImageParams( $handler );
4897 $this->addTrackingCategory( 'broken-file-category' );
4900 # Process the input parameters
4902 $params = array( 'frame' => array(), 'handler' => array(),
4903 'horizAlign' => array(), 'vertAlign' => array() );
4904 foreach ( $parts as $part ) {
4905 $part = trim( $part );
4906 list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part );
4908 if ( isset( $paramMap[$magicName] ) ) {
4909 list( $type, $paramName ) = $paramMap[$magicName];
4911 # Special case; width and height come in one variable together
4912 if ( $type === 'handler' && $paramName === 'width' ) {
4914 # (bug 13500) In both cases (width/height and width only),
4915 # permit trailing "px" for backward compatibility.
4916 if ( preg_match( '/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/', $value, $m ) ) {
4917 $width = intval( $m[1] );
4918 $height = intval( $m[2] );
4919 if ( $handler->validateParam( 'width', $width ) ) {
4920 $params[$type]['width'] = $width;
4923 if ( $handler->validateParam( 'height', $height ) ) {
4924 $params[$type]['height'] = $height;
4927 } elseif ( preg_match( '/^[0-9]*\s*(?:px)?\s*$/', $value ) ) {
4928 $width = intval( $value );
4929 if ( $handler->validateParam( 'width', $width ) ) {
4930 $params[$type]['width'] = $width;
4933 } # else no validation -- bug 13436
4935 if ( $type === 'handler' ) {
4936 # Validate handler parameter
4937 $validated = $handler->validateParam( $paramName, $value );
4939 # Validate internal parameters
4940 switch( $paramName ) {
4943 # @todo FIXME: Possibly check validity here for
4944 # manualthumb? downstream behavior seems odd with
4945 # missing manual thumbs.
4947 $value = $this->stripAltText( $value, $holders );
4950 $chars = self
::EXT_LINK_URL_CLASS
;
4951 $prots = $this->mUrlProtocols
;
4952 if ( $value === '' ) {
4953 $paramName = 'no-link';
4956 } elseif ( preg_match( "/^$prots/", $value ) ) {
4957 if ( preg_match( "/^($prots)$chars+$/u", $value, $m ) ) {
4958 $paramName = 'link-url';
4959 $this->mOutput
->addExternalLink( $value );
4960 if ( $this->mOptions
->getExternalLinkTarget() ) {
4961 $params[$type]['link-target'] = $this->mOptions
->getExternalLinkTarget();
4966 $linkTitle = Title
::newFromText( $value );
4968 $paramName = 'link-title';
4969 $value = $linkTitle;
4970 $this->mOutput
->addLink( $linkTitle );
4976 # Most other things appear to be empty or numeric...
4977 $validated = ( $value === false ||
is_numeric( trim( $value ) ) );
4982 $params[$type][$paramName] = $value;
4986 if ( !$validated ) {
4991 # Process alignment parameters
4992 if ( $params['horizAlign'] ) {
4993 $params['frame']['align'] = key( $params['horizAlign'] );
4995 if ( $params['vertAlign'] ) {
4996 $params['frame']['valign'] = key( $params['vertAlign'] );
4999 $params['frame']['caption'] = $caption;
5001 # Will the image be presented in a frame, with the caption below?
5002 $imageIsFramed = isset( $params['frame']['frame'] ) ||
5003 isset( $params['frame']['framed'] ) ||
5004 isset( $params['frame']['thumbnail'] ) ||
5005 isset( $params['frame']['manualthumb'] );
5007 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5008 # came to also set the caption, ordinary text after the image -- which
5009 # makes no sense, because that just repeats the text multiple times in
5010 # screen readers. It *also* came to set the title attribute.
5012 # Now that we have an alt attribute, we should not set the alt text to
5013 # equal the caption: that's worse than useless, it just repeats the
5014 # text. This is the framed/thumbnail case. If there's no caption, we
5015 # use the unnamed parameter for alt text as well, just for the time be-
5016 # ing, if the unnamed param is set and the alt param is not.
5018 # For the future, we need to figure out if we want to tweak this more,
5019 # e.g., introducing a title= parameter for the title; ignoring the un-
5020 # named parameter entirely for images without a caption; adding an ex-
5021 # plicit caption= parameter and preserving the old magic unnamed para-
5023 if ( $imageIsFramed ) { # Framed image
5024 if ( $caption === '' && !isset( $params['frame']['alt'] ) ) {
5025 # No caption or alt text, add the filename as the alt text so
5026 # that screen readers at least get some description of the image
5027 $params['frame']['alt'] = $title->getText();
5029 # Do not set $params['frame']['title'] because tooltips don't make sense
5031 } else { # Inline image
5032 if ( !isset( $params['frame']['alt'] ) ) {
5033 # No alt text, use the "caption" for the alt text
5034 if ( $caption !== '') {
5035 $params['frame']['alt'] = $this->stripAltText( $caption, $holders );
5037 # No caption, fall back to using the filename for the
5039 $params['frame']['alt'] = $title->getText();
5042 # Use the "caption" for the tooltip text
5043 $params['frame']['title'] = $this->stripAltText( $caption, $holders );
5046 wfRunHooks( 'ParserMakeImageParams', array( $title, $file, &$params ) );
5048 # Linker does the rest
5049 $time = isset( $options['time'] ) ?
$options['time'] : false;
5050 $ret = Linker
::makeImageLink2( $title, $file, $params['frame'], $params['handler'],
5051 $time, $descQuery, $this->mOptions
->getThumbSize() );
5053 # Give the handler a chance to modify the parser object
5055 $handler->parserTransformHook( $this, $file );
5063 * @param $holders LinkHolderArray
5064 * @return mixed|String
5066 protected function stripAltText( $caption, $holders ) {
5067 # Strip bad stuff out of the title (tooltip). We can't just use
5068 # replaceLinkHoldersText() here, because if this function is called
5069 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5071 $tooltip = $holders->replaceText( $caption );
5073 $tooltip = $this->replaceLinkHoldersText( $caption );
5076 # make sure there are no placeholders in thumbnail attributes
5077 # that are later expanded to html- so expand them now and
5079 $tooltip = $this->mStripState
->unstripBoth( $tooltip );
5080 $tooltip = Sanitizer
::stripAllTags( $tooltip );
5086 * Set a flag in the output object indicating that the content is dynamic and
5087 * shouldn't be cached.
5089 function disableCache() {
5090 wfDebug( "Parser output marked as uncacheable.\n" );
5091 $this->mOutput
->setCacheTime( -1 ); // old style, for compatibility
5092 $this->mOutput
->updateCacheExpiry( 0 ); // new style, for consistency
5096 * Callback from the Sanitizer for expanding items found in HTML attribute
5097 * values, so they can be safely tested and escaped.
5099 * @param $text String
5100 * @param $frame PPFrame
5103 function attributeStripCallback( &$text, $frame = false ) {
5104 $text = $this->replaceVariables( $text, $frame );
5105 $text = $this->mStripState
->unstripBoth( $text );
5114 function getTags() {
5115 return array_merge( array_keys( $this->mTransparentTagHooks
), array_keys( $this->mTagHooks
) );
5119 * Replace transparent tags in $text with the values given by the callbacks.
5121 * Transparent tag hooks are like regular XML-style tag hooks, except they
5122 * operate late in the transformation sequence, on HTML instead of wikitext.
5124 * @param $text string
5128 function replaceTransparentTags( $text ) {
5130 $elements = array_keys( $this->mTransparentTagHooks
);
5131 $text = self
::extractTagsAndParams( $elements, $text, $matches, $this->mUniqPrefix
);
5132 $replacements = array();
5134 foreach ( $matches as $marker => $data ) {
5135 list( $element, $content, $params, $tag ) = $data;
5136 $tagName = strtolower( $element );
5137 if ( isset( $this->mTransparentTagHooks
[$tagName] ) ) {
5138 $output = call_user_func_array( $this->mTransparentTagHooks
[$tagName], array( $content, $params, $this ) );
5142 $replacements[$marker] = $output;
5144 return strtr( $text, $replacements );
5148 * Break wikitext input into sections, and either pull or replace
5149 * some particular section's text.
5151 * External callers should use the getSection and replaceSection methods.
5153 * @param $text String: Page wikitext
5154 * @param $section String: a section identifier string of the form:
5155 * <flag1> - <flag2> - ... - <section number>
5157 * Currently the only recognised flag is "T", which means the target section number
5158 * was derived during a template inclusion parse, in other words this is a template
5159 * section edit link. If no flags are given, it was an ordinary section edit link.
5160 * This flag is required to avoid a section numbering mismatch when a section is
5161 * enclosed by <includeonly> (bug 6563).
5163 * The section number 0 pulls the text before the first heading; other numbers will
5164 * pull the given section along with its lower-level subsections. If the section is
5165 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5167 * Section 0 is always considered to exist, even if it only contains the empty
5168 * string. If $text is the empty string and section 0 is replaced, $newText is
5171 * @param $mode String: one of "get" or "replace"
5172 * @param $newText String: replacement text for section data.
5173 * @return String: for "get", the extracted section text.
5174 * for "replace", the whole page with the section replaced.
5176 private function extractSections( $text, $section, $mode, $newText='' ) {
5177 global $wgTitle; # not generally used but removes an ugly failure mode
5178 $this->startParse( $wgTitle, new ParserOptions
, self
::OT_PLAIN
, true );
5180 $frame = $this->getPreprocessor()->newFrame();
5182 # Process section extraction flags
5184 $sectionParts = explode( '-', $section );
5185 $sectionIndex = array_pop( $sectionParts );
5186 foreach ( $sectionParts as $part ) {
5187 if ( $part === 'T' ) {
5188 $flags |
= self
::PTD_FOR_INCLUSION
;
5192 # Check for empty input
5193 if ( strval( $text ) === '' ) {
5194 # Only sections 0 and T-0 exist in an empty document
5195 if ( $sectionIndex == 0 ) {
5196 if ( $mode === 'get' ) {
5202 if ( $mode === 'get' ) {
5210 # Preprocess the text
5211 $root = $this->preprocessToDom( $text, $flags );
5213 # <h> nodes indicate section breaks
5214 # They can only occur at the top level, so we can find them by iterating the root's children
5215 $node = $root->getFirstChild();
5217 # Find the target section
5218 if ( $sectionIndex == 0 ) {
5219 # Section zero doesn't nest, level=big
5220 $targetLevel = 1000;
5223 if ( $node->getName() === 'h' ) {
5224 $bits = $node->splitHeading();
5225 if ( $bits['i'] == $sectionIndex ) {
5226 $targetLevel = $bits['level'];
5230 if ( $mode === 'replace' ) {
5231 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5233 $node = $node->getNextSibling();
5239 if ( $mode === 'get' ) {
5246 # Find the end of the section, including nested sections
5248 if ( $node->getName() === 'h' ) {
5249 $bits = $node->splitHeading();
5250 $curLevel = $bits['level'];
5251 if ( $bits['i'] != $sectionIndex && $curLevel <= $targetLevel ) {
5255 if ( $mode === 'get' ) {
5256 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5258 $node = $node->getNextSibling();
5261 # Write out the remainder (in replace mode only)
5262 if ( $mode === 'replace' ) {
5263 # Output the replacement text
5264 # Add two newlines on -- trailing whitespace in $newText is conventionally
5265 # stripped by the editor, so we need both newlines to restore the paragraph gap
5266 # Only add trailing whitespace if there is newText
5267 if ( $newText != "" ) {
5268 $outText .= $newText . "\n\n";
5272 $outText .= $frame->expand( $node, PPFrame
::RECOVER_ORIG
);
5273 $node = $node->getNextSibling();
5277 if ( is_string( $outText ) ) {
5278 # Re-insert stripped tags
5279 $outText = rtrim( $this->mStripState
->unstripBoth( $outText ) );
5286 * This function returns the text of a section, specified by a number ($section).
5287 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5288 * the first section before any such heading (section 0).
5290 * If a section contains subsections, these are also returned.
5292 * @param $text String: text to look in
5293 * @param $section String: section identifier
5294 * @param $deftext String: default to return if section is not found
5295 * @return string text of the requested section
5297 public function getSection( $text, $section, $deftext='' ) {
5298 return $this->extractSections( $text, $section, "get", $deftext );
5302 * This function returns $oldtext after the content of the section
5303 * specified by $section has been replaced with $text. If the target
5304 * section does not exist, $oldtext is returned unchanged.
5306 * @param $oldtext String: former text of the article
5307 * @param $section Numeric: section identifier
5308 * @param $text String: replacing text
5309 * @return String: modified text
5311 public function replaceSection( $oldtext, $section, $text ) {
5312 return $this->extractSections( $oldtext, $section, "replace", $text );
5316 * Get the ID of the revision we are parsing
5318 * @return Mixed: integer or null
5320 function getRevisionId() {
5321 return $this->mRevisionId
;
5325 * Get the revision object for $this->mRevisionId
5327 * @return Revision|null either a Revision object or null
5329 protected function getRevisionObject() {
5330 if ( !is_null( $this->mRevisionObject
) ) {
5331 return $this->mRevisionObject
;
5333 if ( is_null( $this->mRevisionId
) ) {
5337 $this->mRevisionObject
= Revision
::newFromId( $this->mRevisionId
);
5338 return $this->mRevisionObject
;
5342 * Get the timestamp associated with the current revision, adjusted for
5343 * the default server-local timestamp
5345 function getRevisionTimestamp() {
5346 if ( is_null( $this->mRevisionTimestamp
) ) {
5347 wfProfileIn( __METHOD__
);
5349 $revObject = $this->getRevisionObject();
5350 $timestamp = $revObject ?
$revObject->getTimestamp() : false;
5352 if( $timestamp !== false ) {
5355 # The cryptic '' timezone parameter tells to use the site-default
5356 # timezone offset instead of the user settings.
5358 # Since this value will be saved into the parser cache, served
5359 # to other users, and potentially even used inside links and such,
5360 # it needs to be consistent for all visitors.
5361 $this->mRevisionTimestamp
= $wgContLang->userAdjust( $timestamp, '' );
5364 wfProfileOut( __METHOD__
);
5366 return $this->mRevisionTimestamp
;
5370 * Get the name of the user that edited the last revision
5372 * @return String: user name
5374 function getRevisionUser() {
5375 if( is_null( $this->mRevisionUser
) ) {
5376 $revObject = $this->getRevisionObject();
5378 # if this template is subst: the revision id will be blank,
5379 # so just use the current user's name
5381 $this->mRevisionUser
= $revObject->getUserText();
5382 } elseif( $this->ot
['wiki'] ||
$this->mOptions
->getIsPreview() ) {
5383 $this->mRevisionUser
= $this->getUser()->getName();
5386 return $this->mRevisionUser
;
5390 * Mutator for $mDefaultSort
5392 * @param $sort New value
5394 public function setDefaultSort( $sort ) {
5395 $this->mDefaultSort
= $sort;
5396 $this->mOutput
->setProperty( 'defaultsort', $sort );
5400 * Accessor for $mDefaultSort
5401 * Will use the empty string if none is set.
5403 * This value is treated as a prefix, so the
5404 * empty string is equivalent to sorting by
5409 public function getDefaultSort() {
5410 if ( $this->mDefaultSort
!== false ) {
5411 return $this->mDefaultSort
;
5418 * Accessor for $mDefaultSort
5419 * Unlike getDefaultSort(), will return false if none is set
5421 * @return string or false
5423 public function getCustomDefaultSort() {
5424 return $this->mDefaultSort
;
5428 * Try to guess the section anchor name based on a wikitext fragment
5429 * presumably extracted from a heading, for example "Header" from
5432 * @param $text string
5436 public function guessSectionNameFromWikiText( $text ) {
5437 # Strip out wikitext links(they break the anchor)
5438 $text = $this->stripSectionName( $text );
5439 $text = Sanitizer
::normalizeSectionNameWhitespace( $text );
5440 return '#' . Sanitizer
::escapeId( $text, 'noninitial' );
5444 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5445 * instead. For use in redirects, since IE6 interprets Redirect: headers
5446 * as something other than UTF-8 (apparently?), resulting in breakage.
5448 * @param $text String: The section name
5449 * @return string An anchor
5451 public function guessLegacySectionNameFromWikiText( $text ) {
5452 # Strip out wikitext links(they break the anchor)
5453 $text = $this->stripSectionName( $text );
5454 $text = Sanitizer
::normalizeSectionNameWhitespace( $text );
5455 return '#' . Sanitizer
::escapeId( $text, array( 'noninitial', 'legacy' ) );
5459 * Strips a text string of wikitext for use in a section anchor
5461 * Accepts a text string and then removes all wikitext from the
5462 * string and leaves only the resultant text (i.e. the result of
5463 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5464 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5465 * to create valid section anchors by mimicing the output of the
5466 * parser when headings are parsed.
5468 * @param $text String: text string to be stripped of wikitext
5469 * for use in a Section anchor
5470 * @return Filtered text string
5472 public function stripSectionName( $text ) {
5473 # Strip internal link markup
5474 $text = preg_replace( '/\[\[:?([^[|]+)\|([^[]+)\]\]/', '$2', $text );
5475 $text = preg_replace( '/\[\[:?([^[]+)\|?\]\]/', '$1', $text );
5477 # Strip external link markup
5478 # @todo FIXME: Not tolerant to blank link text
5479 # I.E. [http://www.mediawiki.org] will render as [1] or something depending
5480 # on how many empty links there are on the page - need to figure that out.
5481 $text = preg_replace( '/\[(?:' . wfUrlProtocols() . ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5483 # Parse wikitext quotes (italics & bold)
5484 $text = $this->doQuotes( $text );
5487 $text = StringUtils
::delimiterReplace( '<', '>', '', $text );
5492 * strip/replaceVariables/unstrip for preprocessor regression testing
5494 * @param $text string
5495 * @param $title Title
5496 * @param $options ParserOptions
5497 * @param $outputType int
5501 function testSrvus( $text, Title
$title, ParserOptions
$options, $outputType = self
::OT_HTML
) {
5502 $this->startParse( $title, $options, $outputType, true );
5504 $text = $this->replaceVariables( $text );
5505 $text = $this->mStripState
->unstripBoth( $text );
5506 $text = Sanitizer
::removeHTMLtags( $text );
5511 * @param $text string
5512 * @param $title Title
5513 * @param $options ParserOptions
5516 function testPst( $text, Title
$title, ParserOptions
$options ) {
5517 return $this->preSaveTransform( $text, $title, $options->getUser(), $options );
5522 * @param $title Title
5523 * @param $options ParserOptions
5526 function testPreprocess( $text, Title
$title, ParserOptions
$options ) {
5527 return $this->testSrvus( $text, $title, $options, self
::OT_PREPROCESS
);
5531 * Call a callback function on all regions of the given text that are not
5532 * inside strip markers, and replace those regions with the return value
5533 * of the callback. For example, with input:
5537 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
5538 * two strings will be replaced with the value returned by the callback in
5546 function markerSkipCallback( $s, $callback ) {
5549 while ( $i < strlen( $s ) ) {
5550 $markerStart = strpos( $s, $this->mUniqPrefix
, $i );
5551 if ( $markerStart === false ) {
5552 $out .= call_user_func( $callback, substr( $s, $i ) );
5555 $out .= call_user_func( $callback, substr( $s, $i, $markerStart - $i ) );
5556 $markerEnd = strpos( $s, self
::MARKER_SUFFIX
, $markerStart );
5557 if ( $markerEnd === false ) {
5558 $out .= substr( $s, $markerStart );
5561 $markerEnd +
= strlen( self
::MARKER_SUFFIX
);
5562 $out .= substr( $s, $markerStart, $markerEnd - $markerStart );
5571 * Save the parser state required to convert the given half-parsed text to
5572 * HTML. "Half-parsed" in this context means the output of
5573 * recursiveTagParse() or internalParse(). This output has strip markers
5574 * from replaceVariables (extensionSubstitution() etc.), and link
5575 * placeholders from replaceLinkHolders().
5577 * Returns an array which can be serialized and stored persistently. This
5578 * array can later be loaded into another parser instance with
5579 * unserializeHalfParsedText(). The text can then be safely incorporated into
5580 * the return value of a parser hook.
5582 * @param $text string
5586 function serializeHalfParsedText( $text ) {
5587 wfProfileIn( __METHOD__
);
5590 'version' => self
::HALF_PARSED_VERSION
,
5591 'stripState' => $this->mStripState
->getSubState( $text ),
5592 'linkHolders' => $this->mLinkHolders
->getSubArray( $text )
5594 wfProfileOut( __METHOD__
);
5599 * Load the parser state given in the $data array, which is assumed to
5600 * have been generated by serializeHalfParsedText(). The text contents is
5601 * extracted from the array, and its markers are transformed into markers
5602 * appropriate for the current Parser instance. This transformed text is
5603 * returned, and can be safely included in the return value of a parser
5606 * If the $data array has been stored persistently, the caller should first
5607 * check whether it is still valid, by calling isValidHalfParsedText().
5609 * @param $data Serialized data
5612 function unserializeHalfParsedText( $data ) {
5613 if ( !isset( $data['version'] ) ||
$data['version'] != self
::HALF_PARSED_VERSION
) {
5614 throw new MWException( __METHOD__
.': invalid version' );
5617 # First, extract the strip state.
5618 $texts = array( $data['text'] );
5619 $texts = $this->mStripState
->merge( $data['stripState'], $texts );
5621 # Now renumber links
5622 $texts = $this->mLinkHolders
->mergeForeign( $data['linkHolders'], $texts );
5624 # Should be good to go.
5629 * Returns true if the given array, presumed to be generated by
5630 * serializeHalfParsedText(), is compatible with the current version of the
5633 * @param $data Array
5637 function isValidHalfParsedText( $data ) {
5638 return isset( $data['version'] ) && $data['version'] == self
::HALF_PARSED_VERSION
;