26 use Wikimedia\ScopedCallback;
74 const VERSION =
'1.6.4';
80 const HALF_PARSED_VERSION = 2;
82 # Flags for Parser::setFunctionHook
86 # Constants needed for external link processing
87 # Everything except bracket, space, or control characters
88 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
89 # as well as U+3000 is IDEOGRAPHIC SPACE for T21052
90 # \x{FFFD} is the Unicode replacement character, which Preprocessor_DOM
91 # uses to replace invalid HTML characters.
92 const EXT_LINK_URL_CLASS =
'[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]';
93 # Simplified expression to match an IPv4 or IPv6 address, or
94 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
95 const EXT_LINK_ADDR =
'(?:[0-9.]+|\\[(?i:[0-9a-f:.]+)\\]|[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}])';
96 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
98 const EXT_IMAGE_REGEX =
'/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}\x{FFFD}]+)
99 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
101 # Regular expression for a non-newline space
102 const SPACE_NOT_NL =
'(?:\t| |&\#0*160;|&\#[Xx]0*[Aa]0;|\p{Zs})';
104 # Flags for preprocessToDom
105 const PTD_FOR_INCLUSION = 1;
107 # Allowed values for $this->mOutputType
108 # Parameter to startExternalParse().
110 const
OT_WIKI = 2;
# like preSaveTransform()
113 const
OT_PLAIN = 4;
# like extractSections() - portions of the original are returned unchanged.
132 const MARKER_SUFFIX =
"-QINU`\"'\x7f";
133 const MARKER_PREFIX =
"\x7f'\"`UNIQ-";
135 # Markers used for wrapping the table of contents
136 const TOC_START =
'<mw:toc>';
137 const TOC_END =
'</mw:toc>';
143 public $mTagHooks = [];
144 public $mTransparentTagHooks = [];
145 public $mFunctionHooks = [];
146 public $mFunctionSynonyms = [ 0 => [], 1 => [] ];
147 public $mFunctionTagHooks = [];
148 public $mStripList = [];
149 public $mDefaultStripList = [];
150 public $mVarCache = [];
151 public $mImageParams = [];
152 public $mImageParamsMagicArray = [];
153 public $mMarkerIndex = 0;
157 public $mFirstCall =
true;
159 # Initialised by initialiseVariables()
170 # Initialised in constructor
171 public $mConf, $mExtLinkBracketedRegex, $mUrlProtocols;
173 # Initialized in getPreprocessor()
175 public $mPreprocessor;
177 # Cleared with clearState():
189 public $mIncludeCount;
193 public $mLinkHolders;
196 public $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth;
197 public $mDefaultSort;
198 public $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
199 public $mExpensiveFunctionCount; # number
of expensive
parser function calls
200 public $mShowToc, $mForceTocPosition;
208 # These are variables reset at least once per parse regardless of $clearState
219 public $mOutputType; # Output
type, one
of the OT_xxx constants
220 public $ot; # Shortcut alias,
see setOutputType()
221 public $mRevisionObject;
# The revision object of the specified revision ID
222 public $mRevisionId; # ID to
display in {{REVISIONID}} tags
225 public $mRevisionSize; # Size to
display in {{REVISIONSIZE}} variable
226 public $mRevIdForTs; # The
revision ID which was
used to fetch the timestamp
227 public $mInputSize =
false; # For {{PAGESIZE}}
on current
page.
233 public $mUniqPrefix = self::MARKER_PREFIX;
240 public $mLangLinkLanguages;
248 public $currentRevisionCache;
254 public $mInParse =
false;
257 protected $mProfiler;
262 protected $mLinkRenderer;
265 private $magicWordFactory;
274 private $specialPageFactory;
284 public function __construct(
288 $this->mConf = $conf;
290 $this->mExtLinkBracketedRegex =
'/\[(((?i)' . $this->mUrlProtocols .
')' .
291 self::EXT_LINK_ADDR .
292 self::EXT_LINK_URL_CLASS .
'*)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F\\x{FFFD}]*?)\]/Su';
293 if ( isset( $conf[
'preprocessorClass'] ) ) {
294 $this->mPreprocessorClass = $conf[
'preprocessorClass'];
296 # Under HHVM Preprocessor_Hash is much faster than Preprocessor_DOM
298 } elseif ( extension_loaded(
'domxml' ) ) {
299 # PECL extension that conflicts with the core DOM extension (T15770)
300 wfDebug(
"Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
302 } elseif ( extension_loaded(
'dom' ) ) {
307 wfDebug( __CLASS__ .
": using preprocessor: {$this->mPreprocessorClass}\n" );
309 $services = MediaWikiServices::getInstance();
310 $this->magicWordFactory = $magicWordFactory ??
313 $this->contLang = $contLang ??
$services->getContentLanguage();
315 $this->factory = $factory ??
$services->getParserFactory();
316 $this->specialPageFactory = $spFactory ??
$services->getSpecialPageFactory();
322 public function __destruct() {
323 if ( isset( $this->mLinkHolders ) ) {
324 unset( $this->mLinkHolders );
327 unset( $this->$name );
334 public function __clone() {
335 $this->mInParse =
false;
343 foreach ( [
'mStripState',
'mVarCache' ]
as $k ) {
357 public function firstCallInit() {
358 if ( !$this->mFirstCall ) {
361 $this->mFirstCall =
false;
365 $this->initialiseVariables();
377 public function clearState() {
378 $this->firstCallInit();
380 $this->mOptions->registerWatcher( [ $this->mOutput,
'recordOption' ] );
381 $this->mAutonumber = 0;
382 $this->mIncludeCount = [];
385 $this->mRevisionObject = $this->mRevisionTimestamp =
386 $this->mRevisionId = $this->mRevisionUser = $this->mRevisionSize =
null;
387 $this->mVarCache = [];
389 $this->mLangLinkLanguages = [];
390 $this->currentRevisionCache =
null;
394 # Clear these on every parse, T6549
395 $this->mTplRedirCache = $this->mTplDomCache = [];
397 $this->mShowToc =
true;
398 $this->mForceTocPosition =
false;
399 $this->mIncludeSizes = [
403 $this->mPPNodeCount = 0;
404 $this->mGeneratedPPNodeCount = 0;
405 $this->mHighestExpansionDepth = 0;
406 $this->mDefaultSort =
false;
407 $this->mHeadings = [];
408 $this->mDoubleUnderscores = [];
409 $this->mExpensiveFunctionCount = 0;
412 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) {
413 $this->mPreprocessor =
null;
437 public function parse(
439 $linestart =
true, $clearState =
true, $revid =
null
444 $text = strtr( $text,
"\x7f",
"?" );
445 $magicScopeVariable = $this->lock();
448 $text = str_replace(
"\000",
'', $text );
452 $this->currentRevisionCache =
null;
453 $this->mInputSize = strlen( $text );
454 if ( $this->mOptions->getEnableLimitReport() ) {
455 $this->mOutput->resetParseStartTime();
458 $oldRevisionId = $this->mRevisionId;
459 $oldRevisionObject = $this->mRevisionObject;
460 $oldRevisionTimestamp = $this->mRevisionTimestamp;
461 $oldRevisionUser = $this->mRevisionUser;
462 $oldRevisionSize = $this->mRevisionSize;
463 if ( $revid !==
null ) {
464 $this->mRevisionId = $revid;
465 $this->mRevisionObject =
null;
466 $this->mRevisionTimestamp =
null;
467 $this->mRevisionUser =
null;
468 $this->mRevisionSize =
null;
476 $text = $this->internalParse( $text );
479 $text = $this->internalParseHalfParsed( $text,
true, $linestart );
488 if ( !(
$options->getDisableTitleConversion()
489 || isset( $this->mDoubleUnderscores[
'nocontentconvert'] )
490 || isset( $this->mDoubleUnderscores[
'notitleconvert'] )
491 || $this->mOutput->getDisplayTitle() !==
false )
493 $convruletitle = $this->getTargetLanguage()->getConvRuleTitle();
494 if ( $convruletitle ) {
495 $this->mOutput->setTitleText( $convruletitle );
497 $titleText = $this->getTargetLanguage()->convertTitle(
$title );
498 $this->mOutput->setTitleText( $titleText );
502 # Compute runtime adaptive expiry if set
503 $this->mOutput->finalizeAdaptiveCacheExpiry();
505 # Warn if too many heavyweight parser functions were used
506 if ( $this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit() ) {
507 $this->limitationWarn(
'expensive-parserfunction',
508 $this->mExpensiveFunctionCount,
509 $this->mOptions->getExpensiveParserFunctionLimit()
513 # Information on limits, for the benefit of users who try to skirt them
514 if ( $this->mOptions->getEnableLimitReport() ) {
515 $text .= $this->makeLimitReport();
518 # Wrap non-interface parser output in a <div> so it can be targeted
520 $class = $this->mOptions->getWrapOutputClass();
521 if ( $class !==
false && !$this->mOptions->getInterfaceMessage() ) {
522 $this->mOutput->addWrapperDivClass( $class );
525 $this->mOutput->setText( $text );
527 $this->mRevisionId = $oldRevisionId;
528 $this->mRevisionObject = $oldRevisionObject;
529 $this->mRevisionTimestamp = $oldRevisionTimestamp;
530 $this->mRevisionUser = $oldRevisionUser;
531 $this->mRevisionSize = $oldRevisionSize;
532 $this->mInputSize =
false;
533 $this->currentRevisionCache =
null;
535 return $this->mOutput;
544 protected function makeLimitReport() {
547 $maxIncludeSize = $this->mOptions->getMaxIncludeSize();
549 $cpuTime = $this->mOutput->getTimeSinceStart(
'cpu' );
550 if ( $cpuTime !==
null ) {
551 $this->mOutput->setLimitReportData(
'limitreport-cputime',
552 sprintf(
"%.3f", $cpuTime )
556 $wallTime = $this->mOutput->getTimeSinceStart(
'wall' );
557 $this->mOutput->setLimitReportData(
'limitreport-walltime',
558 sprintf(
"%.3f", $wallTime )
561 $this->mOutput->setLimitReportData(
'limitreport-ppvisitednodes',
562 [ $this->mPPNodeCount, $this->mOptions->getMaxPPNodeCount() ]
564 $this->mOutput->setLimitReportData(
'limitreport-ppgeneratednodes',
565 [ $this->mGeneratedPPNodeCount, $this->mOptions->getMaxGeneratedPPNodeCount() ]
567 $this->mOutput->setLimitReportData(
'limitreport-postexpandincludesize',
568 [ $this->mIncludeSizes[
'post-expand'], $maxIncludeSize ]
570 $this->mOutput->setLimitReportData(
'limitreport-templateargumentsize',
571 [ $this->mIncludeSizes[
'arg'], $maxIncludeSize ]
573 $this->mOutput->setLimitReportData(
'limitreport-expansiondepth',
574 [ $this->mHighestExpansionDepth, $this->mOptions->getMaxPPExpandDepth() ]
576 $this->mOutput->setLimitReportData(
'limitreport-expensivefunctioncount',
577 [ $this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit() ]
580 foreach ( $this->mStripState->getLimitReport()
as list( $key,
$value ) ) {
581 $this->mOutput->setLimitReportData( $key,
$value );
584 Hooks::run(
'ParserLimitReportPrepare', [ $this, $this->mOutput ] );
586 $limitReport =
"NewPP limit report\n";
588 $limitReport .=
'Parsed by ' .
wfHostname() .
"\n";
590 $limitReport .=
'Cached time: ' . $this->mOutput->getCacheTime() .
"\n";
591 $limitReport .=
'Cache expiry: ' . $this->mOutput->getCacheExpiry() .
"\n";
592 $limitReport .=
'Dynamic content: ' .
593 ( $this->mOutput->hasDynamicContent() ?
'true' :
'false' ) .
596 foreach ( $this->mOutput->getLimitReportData()
as $key =>
$value ) {
598 [ $key, &
$value, &$limitReport,
false,
false ]
600 $keyMsg =
wfMessage( $key )->inLanguage(
'en' )->useDatabase(
false );
601 $valueMsg =
wfMessage( [
"$key-value-text",
"$key-value" ] )
602 ->inLanguage(
'en' )->useDatabase(
false );
603 if ( !$valueMsg->exists() ) {
606 if ( !$keyMsg->isDisabled() && !$valueMsg->isDisabled() ) {
607 $valueMsg->params(
$value );
608 $limitReport .=
"{$keyMsg->text()}: {$valueMsg->text()}\n";
614 $limitReport = htmlspecialchars_decode( $limitReport );
618 $limitReport = str_replace( [
'-',
'&' ], [
'‐',
'&' ], $limitReport );
619 $text =
"\n<!-- \n$limitReport-->\n";
622 $dataByFunc = $this->mProfiler->getFunctionStats();
623 uasort( $dataByFunc,
function ( $a, $b ) {
624 return $b[
'real'] <=> $a[
'real'];
627 foreach ( array_slice( $dataByFunc, 0, 10 )
as $item ) {
628 $profileReport[] = sprintf(
"%6.2f%% %8.3f %6d %s",
629 $item[
'%real'], $item[
'real'], $item[
'calls'],
630 htmlspecialchars( $item[
'name'] ) );
632 $text .=
"<!--\nTransclusion expansion time report (%,ms,calls,template)\n";
633 $text .= implode(
"\n", $profileReport ) .
"\n-->\n";
635 $this->mOutput->setLimitReportData(
'limitreport-timingprofile', $profileReport );
639 $this->mOutput->setLimitReportData(
'cachereport-origin',
wfHostname() );
641 $this->mOutput->setLimitReportData(
'cachereport-timestamp',
642 $this->mOutput->getCacheTime() );
643 $this->mOutput->setLimitReportData(
'cachereport-ttl',
644 $this->mOutput->getCacheExpiry() );
645 $this->mOutput->setLimitReportData(
'cachereport-transientcontent',
646 $this->mOutput->hasDynamicContent() );
648 if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) {
649 wfDebugLog(
'generated-pp-node-count', $this->mGeneratedPPNodeCount .
' ' .
650 $this->mTitle->getPrefixedDBkey() );
679 public function recursiveTagParse( $text, $frame =
false ) {
684 $text = $this->internalParse( $text,
false, $frame );
707 public function recursiveTagParseFully( $text, $frame =
false ) {
708 $text = $this->recursiveTagParse( $text, $frame );
709 $text = $this->internalParseHalfParsed( $text,
false );
724 public function preprocess( $text,
Title $title =
null,
727 $magicScopeVariable = $this->lock();
729 if ( $revid !==
null ) {
730 $this->mRevisionId = $revid;
736 $text = $this->replaceVariables( $text, $frame );
737 $text = $this->mStripState->unstripBoth( $text );
750 public function recursivePreprocess( $text, $frame =
false ) {
751 $text = $this->replaceVariables( $text, $frame );
752 $text = $this->mStripState->unstripBoth( $text );
771 $text = $msg->params(
$params )->plain();
773 # Parser (re)initialisation
774 $magicScopeVariable = $this->lock();
778 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
779 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
780 $text = $this->mStripState->unstripBoth( $text );
790 public function setUser(
$user ) {
791 $this->mUser =
$user;
799 public function setTitle(
$t ) {
804 if (
$t->hasFragment() ) {
805 # Strip the fragment to avoid various odd effects
806 $this->mTitle =
$t->createFragmentTarget(
'' );
817 public function getTitle() {
818 return $this->mTitle;
827 public function Title( $x =
null ) {
828 return wfSetVar( $this->mTitle, $x );
836 public function setOutputType( $ot ) {
837 $this->mOutputType = $ot;
853 public function OutputType( $x =
null ) {
854 return wfSetVar( $this->mOutputType, $x );
862 public function getOutput() {
863 return $this->mOutput;
871 public function getOptions() {
872 return $this->mOptions;
881 public function Options( $x =
null ) {
882 return wfSetVar( $this->mOptions, $x );
888 public function nextLinkID() {
889 return $this->mLinkID++;
895 public function setLinkID( $id ) {
896 $this->mLinkID = $id;
903 public function getFunctionLang() {
904 return $this->getTargetLanguage();
916 public function getTargetLanguage() {
917 $target = $this->mOptions->getTargetLanguage();
919 if ( $target !==
null ) {
921 } elseif ( $this->mOptions->getInterfaceMessage() ) {
922 return $this->mOptions->getUserLangObj();
923 } elseif ( is_null( $this->mTitle ) ) {
924 throw new MWException( __METHOD__ .
': $this->mTitle is null' );
927 return $this->mTitle->getPageLanguage();
935 public function getConverterLanguage() {
936 return $this->getTargetLanguage();
945 public function getUser() {
946 if ( !is_null( $this->mUser ) ) {
949 return $this->mOptions->getUser();
957 public function getPreprocessor() {
958 if ( !isset( $this->mPreprocessor ) ) {
959 $class = $this->mPreprocessorClass;
960 $this->mPreprocessor =
new $class( $this );
962 return $this->mPreprocessor;
971 public function getLinkRenderer() {
972 if ( !$this->mLinkRenderer ) {
973 $this->mLinkRenderer = MediaWikiServices::getInstance()
974 ->getLinkRendererFactory()->create();
975 $this->mLinkRenderer->setStubThreshold(
976 $this->getOptions()->getStubThreshold()
980 return $this->mLinkRenderer;
989 public function getMagicWordFactory() {
990 return $this->magicWordFactory;
999 public function getContentLanguage() {
1000 return $this->contLang;
1022 public static function extractTagsAndParams( $elements, $text, &
$matches ) {
1027 $taglist = implode(
'|', $elements );
1028 $start =
"/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" .
">)|<(!--)/i";
1030 while ( $text !=
'' ) {
1031 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
1033 if (
count( $p ) < 5 ) {
1036 if (
count( $p ) > 5 ) {
1045 $attributes = $p[2];
1050 $marker = self::MARKER_PREFIX .
"-$element-" . sprintf(
'%08X', $n++ ) . self::MARKER_SUFFIX;
1051 $stripped .= $marker;
1053 if ( $close ===
'/>' ) {
1054 # Empty element tag, <tag />
1059 if ( $element ===
'!--' ) {
1062 $end =
"/(<\\/$element\\s*>)/i";
1064 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
1066 if (
count( $q ) < 3 ) {
1067 # No end tag -- let it run out to the end of the text.
1078 Sanitizer::decodeTagAttributes( $attributes ),
1079 "<$element$attributes$close$content$tail" ];
1089 public function getStripList() {
1090 return $this->mStripList;
1102 public function insertStripItem( $text ) {
1103 $marker = self::MARKER_PREFIX .
"-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
1104 $this->mMarkerIndex++;
1105 $this->mStripState->addGeneral( $marker, $text );
1116 public function doTableStuff( $text ) {
1119 $td_history = []; # Is currently
a td
tag open?
1120 $last_tag_history = []; # Save
history of last lag activated (td, th
or caption)
1121 $tr_history = []; # Is currently
a tr
tag open?
1122 $tr_attributes = []; #
history of tr attributes
1123 $has_opened_tr = []; # Did
this table open a <tr> element?
1124 $indent_level = 0; # indent level
of the
table
1127 $line = trim( $outLine );
1129 if (
$line ===
'' ) { # empty line, go to next line
1130 $out .= $outLine .
"\n";
1134 $first_character =
$line[0];
1135 $first_two = substr(
$line, 0, 2 );
1138 if ( preg_match(
'/^(:*)\s*\{\|(.*)$/',
$line,
$matches ) ) {
1139 # First check if we are starting a new table
1140 $indent_level = strlen(
$matches[1] );
1142 $attributes = $this->mStripState->unstripBoth(
$matches[2] );
1143 $attributes = Sanitizer::fixTagAttributes( $attributes,
'table' );
1145 $outLine = str_repeat(
'<dl><dd>', $indent_level ) .
"<table{$attributes}>";
1146 array_push( $td_history,
false );
1147 array_push( $last_tag_history,
'' );
1148 array_push( $tr_history,
false );
1149 array_push( $tr_attributes,
'' );
1150 array_push( $has_opened_tr,
false );
1151 } elseif (
count( $td_history ) == 0 ) {
1152 # Don't do any of the following
1153 $out .= $outLine .
"\n";
1155 } elseif ( $first_two ===
'|}' ) {
1156 # We are ending a table
1158 $last_tag = array_pop( $last_tag_history );
1160 if ( !array_pop( $has_opened_tr ) ) {
1161 $line =
"<tr><td></td></tr>{$line}";
1164 if ( array_pop( $tr_history ) ) {
1165 $line =
"</tr>{$line}";
1168 if ( array_pop( $td_history ) ) {
1169 $line =
"</{$last_tag}>{$line}";
1171 array_pop( $tr_attributes );
1172 if ( $indent_level > 0 ) {
1173 $outLine = rtrim(
$line ) . str_repeat(
'</dd></dl>', $indent_level );
1177 } elseif ( $first_two ===
'|-' ) {
1178 # Now we have a table row
1179 $line = preg_replace(
'#^\|-+#',
'',
$line );
1181 # Whats after the tag is now only attributes
1182 $attributes = $this->mStripState->unstripBoth(
$line );
1183 $attributes = Sanitizer::fixTagAttributes( $attributes,
'tr' );
1184 array_pop( $tr_attributes );
1185 array_push( $tr_attributes, $attributes );
1188 $last_tag = array_pop( $last_tag_history );
1189 array_pop( $has_opened_tr );
1190 array_push( $has_opened_tr,
true );
1192 if ( array_pop( $tr_history ) ) {
1196 if ( array_pop( $td_history ) ) {
1197 $line =
"</{$last_tag}>{$line}";
1201 array_push( $tr_history,
false );
1202 array_push( $td_history,
false );
1203 array_push( $last_tag_history,
'' );
1204 } elseif ( $first_character ===
'|'
1205 || $first_character ===
'!'
1206 || $first_two ===
'|+'
1208 # This might be cell elements, td, th or captions
1209 if ( $first_two ===
'|+' ) {
1210 $first_character =
'+';
1217 if ( $first_character ===
'!' ) {
1221 # Split up multiple cells on the same line.
1222 # FIXME : This can result in improper nesting of tags processed
1223 # by earlier parser steps.
1224 $cells = explode(
'||',
$line );
1228 # Loop through each table cell
1229 foreach ( $cells
as $cell ) {
1231 if ( $first_character !==
'+' ) {
1232 $tr_after = array_pop( $tr_attributes );
1233 if ( !array_pop( $tr_history ) ) {
1234 $previous =
"<tr{$tr_after}>\n";
1236 array_push( $tr_history,
true );
1237 array_push( $tr_attributes,
'' );
1238 array_pop( $has_opened_tr );
1239 array_push( $has_opened_tr,
true );
1242 $last_tag = array_pop( $last_tag_history );
1244 if ( array_pop( $td_history ) ) {
1245 $previous =
"</{$last_tag}>\n{$previous}";
1248 if ( $first_character ===
'|' ) {
1250 } elseif ( $first_character ===
'!' ) {
1252 } elseif ( $first_character ===
'+' ) {
1253 $last_tag =
'caption';
1258 array_push( $last_tag_history, $last_tag );
1260 # A cell could contain both parameters and data
1261 $cell_data = explode(
'|', $cell, 2 );
1263 # T2553: Note that a '|' inside an invalid link should not
1264 # be mistaken as delimiting cell parameters
1265 # Bug T153140: Neither should language converter markup.
1266 if ( preg_match(
'/\[\[|-\{/', $cell_data[0] ) === 1 ) {
1267 $cell =
"{$previous}<{$last_tag}>" . trim( $cell );
1268 } elseif (
count( $cell_data ) == 1 ) {
1270 $cell =
"{$previous}<{$last_tag}>" . trim( $cell_data[0] );
1272 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1273 $attributes = Sanitizer::fixTagAttributes( $attributes, $last_tag );
1275 $cell =
"{$previous}<{$last_tag}{$attributes}>" . trim( $cell_data[1] );
1279 array_push( $td_history,
true );
1282 $out .= $outLine .
"\n";
1285 # Closing open td, tr && table
1286 while (
count( $td_history ) > 0 ) {
1287 if ( array_pop( $td_history ) ) {
1290 if ( array_pop( $tr_history ) ) {
1293 if ( !array_pop( $has_opened_tr ) ) {
1294 $out .=
"<tr><td></td></tr>\n";
1297 $out .=
"</table>\n";
1300 # Remove trailing line-ending (b/c)
1301 if ( substr(
$out, -1 ) ===
"\n" ) {
1305 # special case: don't return empty table
1306 if (
$out ===
"<table>\n<tr><td></td></tr>\n</table>" ) {
1326 public function internalParse( $text, $isMain =
true, $frame =
false ) {
1332 # Hook to suspend the parser in this state
1333 if ( !
Hooks::run(
'ParserBeforeInternalParse', [ &
$parser, &$text, &$this->mStripState ] ) ) {
1337 # if $frame is provided, then use $frame for replacing any variables
1339 # use frame depth to infer how include/noinclude tags should be handled
1340 # depth=0 means this is the top-level document; otherwise it's an included document
1341 if ( !$frame->depth ) {
1344 $flag = self::PTD_FOR_INCLUSION;
1346 $dom = $this->preprocessToDom( $text, $flag );
1347 $text = $frame->expand( $dom );
1349 # if $frame is not provided, then use old-style replaceVariables
1350 $text = $this->replaceVariables( $text );
1353 Hooks::run(
'InternalParseBeforeSanitize', [ &
$parser, &$text, &$this->mStripState ] );
1354 $text = Sanitizer::removeHTMLtags(
1356 [ $this,
'attributeStripCallback' ],
1358 array_keys( $this->mTransparentTagHooks ),
1360 [ $this,
'addTrackingCategory' ]
1362 Hooks::run(
'InternalParseBeforeLinks', [ &
$parser, &$text, &$this->mStripState ] );
1364 # Tables need to come after variable replacement for things to work
1365 # properly; putting them before other transformations should keep
1366 # exciting things like link expansions from showing up in surprising
1368 $text = $this->doTableStuff( $text );
1370 $text = preg_replace(
'/(^|\n)-----*/',
'\\1<hr />', $text );
1372 $text = $this->doDoubleUnderscore( $text );
1374 $text = $this->doHeadings( $text );
1375 $text = $this->replaceInternalLinks( $text );
1376 $text = $this->doAllQuotes( $text );
1377 $text = $this->replaceExternalLinks( $text );
1379 # replaceInternalLinks may sometimes leave behind
1380 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1381 $text = str_replace( self::MARKER_PREFIX .
'NOPARSE',
'', $text );
1383 $text = $this->doMagicLinks( $text );
1384 $text = $this->formatHeadings( $text, $origText, $isMain );
1398 private function internalParseHalfParsed( $text, $isMain =
true, $linestart =
true ) {
1399 $text = $this->mStripState->unstripGeneral( $text );
1408 # Clean up special characters, only run once, next-to-last before doBlockLevels
1409 $text = Sanitizer::armorFrenchSpaces( $text );
1411 $text = $this->doBlockLevels( $text, $linestart );
1413 $this->replaceLinkHolders( $text );
1422 if ( !( $this->mOptions->getDisableContentConversion()
1423 || isset( $this->mDoubleUnderscores[
'nocontentconvert'] ) )
1425 if ( !$this->mOptions->getInterfaceMessage() ) {
1426 # The position of the convert() call should not be changed. it
1427 # assumes that the links are all replaced and the only thing left
1428 # is the <nowiki> mark.
1429 $text = $this->getTargetLanguage()->convert( $text );
1433 $text = $this->mStripState->unstripNoWiki( $text );
1439 $text = $this->replaceTransparentTags( $text );
1440 $text = $this->mStripState->unstripGeneral( $text );
1442 $text = Sanitizer::normalizeCharReferences( $text );
1445 if ( $this->mOptions->getTidy() ) {
1449 # attempt to sanitize at least some nesting problems
1450 # (T4702 and quite a few others)
1452 # ''Something [http://www.cool.com cool''] -->
1453 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1454 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
1455 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
1456 # fix up an anchor inside another anchor, only
1457 # at least for a single single nested link (T5695)
1458 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
1459 '\\1\\2</a>\\3</a>\\1\\4</a>',
1460 # fix div inside inline elements- doBlockLevels won't wrap a line which
1461 # contains a div, so fix it up here; replace
1462 # div with escaped text
1463 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
1464 '\\1\\3<div\\5>\\6</div>\\8\\9',
1465 # remove empty italic or bold tag pairs, some
1466 # introduced by rules above
1467 '/<([bi])><\/\\1>/' =>
'',
1470 $text = preg_replace(
1471 array_keys( $tidyregs ),
1472 array_values( $tidyregs ),
1494 public function doMagicLinks( $text ) {
1496 $urlChar = self::EXT_LINK_URL_CLASS;
1497 $addr = self::EXT_LINK_ADDR;
1498 $space = self::SPACE_NOT_NL; # non-newline space
1499 $spdash =
"(?:-|$space)"; #
a dash
or a non-newline space
1500 $spaces =
"$space++"; # possessive match
of 1
or more spaces
1501 $text = preg_replace_callback(
1503 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1504 (<.*?>) | # m[2]: Skip stuff inside HTML elements' .
"
1505 (\b # m[3]: Free external links
1507 ($addr$urlChar*) # m[4]: Post-protocol path
1509 \b(?:RFC|PMID) $spaces # m[5]: RFC or PMID, capture number
1511 \bISBN $spaces ( # m[6]: ISBN, capture number
1512 (?: 97[89] $spdash? )? # optional 13-digit ISBN prefix
1513 (?: [0-9] $spdash? ){9} # 9 digits with opt. delimiters
1514 [0-9Xx] # check digit
1516 )!xu", [ $this,
'magicLinkCallback' ], $text );
1525 public function magicLinkCallback( $m ) {
1526 if ( isset( $m[1] ) && $m[1] !==
'' ) {
1529 } elseif ( isset( $m[2] ) && $m[2] !==
'' ) {
1532 } elseif ( isset( $m[3] ) && $m[3] !==
'' ) {
1533 # Free external link
1534 return $this->makeFreeExternalLink( $m[0], strlen( $m[4] ) );
1535 } elseif ( isset( $m[5] ) && $m[5] !==
'' ) {
1537 if ( substr( $m[0], 0, 3 ) ===
'RFC' ) {
1538 if ( !$this->mOptions->getMagicRFCLinks() ) {
1543 $cssClass =
'mw-magiclink-rfc';
1544 $trackingCat =
'magiclink-tracking-rfc';
1546 } elseif ( substr( $m[0], 0, 4 ) ===
'PMID' ) {
1547 if ( !$this->mOptions->getMagicPMIDLinks() ) {
1551 $urlmsg =
'pubmedurl';
1552 $cssClass =
'mw-magiclink-pmid';
1553 $trackingCat =
'magiclink-tracking-pmid';
1556 throw new MWException( __METHOD__ .
': unrecognised match type "' .
1557 substr( $m[0], 0, 20 ) .
'"' );
1559 $url =
wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1560 $this->addTrackingCategory( $trackingCat );
1562 } elseif ( isset( $m[6] ) && $m[6] !==
''
1563 && $this->mOptions->getMagicISBNLinks()
1567 $space = self::SPACE_NOT_NL; # non-newline space
1568 $isbn = preg_replace(
"/$space/",
' ', $isbn );
1569 $num = strtr( $isbn, [
1574 $this->addTrackingCategory(
'magiclink-tracking-isbn' );
1575 return $this->getLinkRenderer()->makeKnownLink(
1579 'class' =>
'internal mw-magiclink-isbn',
1597 public function makeFreeExternalLink( $url, $numPostProto ) {
1600 # The characters '<' and '>' (which were escaped by
1601 # removeHTMLtags()) should not be included in
1602 # URLs, per RFC 2396.
1603 # Make terminate a URL as well (bug T84937)
1606 '/&(lt|gt|nbsp|#x0*(3[CcEe]|[Aa]0)|#0*(60|62|160));/',
1611 $trail = substr( $url, $m2[0][1] ) . $trail;
1612 $url = substr( $url, 0, $m2[0][1] );
1615 # Move trailing punctuation to $trail
1617 # If there is no left bracket, then consider right brackets fair game too
1618 if ( strpos( $url,
'(' ) ===
false ) {
1622 $urlRev = strrev( $url );
1623 $numSepChars = strspn( $urlRev, $sep );
1624 # Don't break a trailing HTML entity by moving the ; into $trail
1625 # This is in hot code, so use substr_compare to avoid having to
1626 # create a new string object for the comparison
1627 if ( $numSepChars && substr_compare( $url,
";", -$numSepChars, 1 ) === 0 ) {
1628 # more optimization: instead of running preg_match with a $
1629 # anchor, which can be slow, do the match on the reversed
1630 # string starting at the desired offset.
1631 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1632 if ( preg_match(
'/\G([a-z]+|[\da-f]+x#|\d+#)&/i', $urlRev, $m2, 0, $numSepChars ) ) {
1636 if ( $numSepChars ) {
1637 $trail = substr( $url, -$numSepChars ) . $trail;
1638 $url = substr( $url, 0, -$numSepChars );
1641 # Verify that we still have a real URL after trail removal, and
1642 # not just lone protocol
1643 if ( strlen( $trail ) >= $numPostProto ) {
1644 return $url . $trail;
1647 $url = Sanitizer::cleanUrl( $url );
1649 # Is this an external image?
1650 $text = $this->maybeMakeExternalImage( $url );
1651 if ( $text ===
false ) {
1652 # Not an image, make a link
1654 $this->getTargetLanguage()->getConverter()->markNoConversion( $url ),
1656 $this->getExternalLinkAttribs( $url ), $this->mTitle );
1657 # Register it in the output object...
1658 $this->mOutput->addExternalLink( $url );
1660 return $text . $trail;
1672 public function doHeadings( $text ) {
1673 for ( $i = 6; $i >= 1; --$i ) {
1674 $h = str_repeat(
'=', $i );
1677 $text = preg_replace(
"/^(?:$h)[ \\t]*(.+?)[ \\t]*(?:$h)\\s*$/m",
"<h$i>\\1</h$i>", $text );
1690 public function doAllQuotes( $text ) {
1694 $outtext .= $this->doQuotes(
$line ) .
"\n";
1696 $outtext = substr( $outtext, 0, -1 );
1707 public function doQuotes( $text ) {
1708 $arr = preg_split(
"/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1709 $countarr =
count( $arr );
1710 if ( $countarr == 1 ) {
1719 for ( $i = 1; $i < $countarr; $i += 2 ) {
1720 $thislen = strlen( $arr[$i] );
1724 if ( $thislen == 4 ) {
1725 $arr[$i - 1] .=
"'";
1728 } elseif ( $thislen > 5 ) {
1732 $arr[$i - 1] .= str_repeat(
"'", $thislen - 5 );
1737 if ( $thislen == 2 ) {
1739 } elseif ( $thislen == 3 ) {
1741 } elseif ( $thislen == 5 ) {
1751 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1752 $firstsingleletterword = -1;
1753 $firstmultiletterword = -1;
1755 for ( $i = 1; $i < $countarr; $i += 2 ) {
1756 if ( strlen( $arr[$i] ) == 3 ) {
1757 $x1 = substr( $arr[$i - 1], -1 );
1758 $x2 = substr( $arr[$i - 1], -2, 1 );
1759 if ( $x1 ===
' ' ) {
1760 if ( $firstspace == -1 ) {
1763 } elseif ( $x2 ===
' ' ) {
1764 $firstsingleletterword = $i;
1769 if ( $firstmultiletterword == -1 ) {
1770 $firstmultiletterword = $i;
1777 if ( $firstsingleletterword > -1 ) {
1778 $arr[$firstsingleletterword] =
"''";
1779 $arr[$firstsingleletterword - 1] .=
"'";
1780 } elseif ( $firstmultiletterword > -1 ) {
1782 $arr[$firstmultiletterword] =
"''";
1783 $arr[$firstmultiletterword - 1] .=
"'";
1784 } elseif ( $firstspace > -1 ) {
1788 $arr[$firstspace] =
"''";
1789 $arr[$firstspace - 1] .=
"'";
1798 foreach ( $arr
as $r ) {
1799 if ( ( $i % 2 ) == 0 ) {
1800 if ( $state ===
'both' ) {
1806 $thislen = strlen( $r );
1807 if ( $thislen == 2 ) {
1808 if ( $state ===
'i' ) {
1811 } elseif ( $state ===
'bi' ) {
1814 } elseif ( $state ===
'ib' ) {
1817 } elseif ( $state ===
'both' ) {
1824 } elseif ( $thislen == 3 ) {
1825 if ( $state ===
'b' ) {
1828 } elseif ( $state ===
'bi' ) {
1831 } elseif ( $state ===
'ib' ) {
1834 } elseif ( $state ===
'both' ) {
1841 } elseif ( $thislen == 5 ) {
1842 if ( $state ===
'b' ) {
1845 } elseif ( $state ===
'i' ) {
1848 } elseif ( $state ===
'bi' ) {
1851 } elseif ( $state ===
'ib' ) {
1854 } elseif ( $state ===
'both' ) {
1866 if ( $state ===
'b' || $state ===
'ib' ) {
1869 if ( $state ===
'i' || $state ===
'bi' || $state ===
'ib' ) {
1872 if ( $state ===
'bi' ) {
1876 if ( $state ===
'both' &&
$buffer ) {
1895 public function replaceExternalLinks( $text ) {
1896 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1897 if ( $bits ===
false ) {
1898 throw new MWException(
"PCRE needs to be compiled with "
1899 .
"--enable-unicode-properties in order for MediaWiki to function" );
1901 $s = array_shift( $bits );
1904 while ( $i <
count( $bits ) ) {
1907 $text = $bits[$i++];
1908 $trail = $bits[$i++];
1910 # The characters '<' and '>' (which were escaped by
1911 # removeHTMLtags()) should not be included in
1912 # URLs, per RFC 2396.
1914 if ( preg_match(
'/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1915 $text = substr( $url, $m2[0][1] ) .
' ' . $text;
1916 $url = substr( $url, 0, $m2[0][1] );
1919 # If the link text is an image URL, replace it with an <img> tag
1920 # This happened by accident in the original parser, but some people used it extensively
1921 $img = $this->maybeMakeExternalImage( $text );
1922 if ( $img !==
false ) {
1928 # Set linktype for CSS
1931 # No link text, e.g. [http://domain.tld/some.link]
1932 if ( $text ==
'' ) {
1934 $langObj = $this->getTargetLanguage();
1935 $text =
'[' . $langObj->formatNum( ++$this->mAutonumber ) .
']';
1936 $linktype =
'autonumber';
1938 # Have link text, e.g. [http://domain.tld/some.link text]s
1945 $text = $this->getTargetLanguage()->getConverter()->markNoConversion( $text );
1948 $url = Sanitizer::cleanUrl( $url );
1950 # Use the encoded URL
1951 # This means that users can paste URLs directly into the text
1952 # Funny characters like ö aren't valid in URLs anyway
1953 # This was changed in August 2004
1955 $this->getExternalLinkAttribs( $url ), $this->mTitle ) . $dtrail . $trail;
1957 # Register link in the output object.
1958 $this->mOutput->addExternalLink( $url );
1973 public static function getExternalLinkRel( $url =
false,
$title =
null ) {
1994 public function getExternalLinkAttribs( $url ) {
1996 $rel = self::getExternalLinkRel( $url, $this->mTitle );
1998 $target = $this->mOptions->getExternalLinkTarget();
2001 if ( !in_array( $target, [
'_self',
'_parent',
'_top' ] ) ) {
2005 if ( $rel !==
'' ) {
2008 $rel .=
'noreferrer noopener';
2024 public static function normalizeLinkUrl( $url ) {
2025 # First, make sure unsafe characters are encoded
2026 $url = preg_replace_callback(
'/[\x00-\x20"<>\[\\\\\]^`{|}\x7F-\xFF]/',
2028 return rawurlencode( $m[0] );
2034 $end = strlen( $url );
2036 # Fragment part - 'fragment'
2037 $start = strpos( $url,
'#' );
2038 if ( $start !==
false && $start < $end ) {
2039 $ret = self::normalizeUrlComponent(
2040 substr( $url, $start, $end - $start ),
'"#%<>[\]^`{|}' ) .
$ret;
2044 # Query part - 'query' minus &=+;
2045 $start = strpos( $url,
'?' );
2046 if ( $start !==
false && $start < $end ) {
2047 $ret = self::normalizeUrlComponent(
2048 substr( $url, $start, $end - $start ),
'"#%<>[\]^`{|}&=+;' ) .
$ret;
2052 # Scheme and path part - 'pchar'
2053 # (we assume no userinfo or encoded colons in the host)
2054 $ret = self::normalizeUrlComponent(
2055 substr( $url, 0, $end ),
'"#%<>[\]^`{|}/?' ) .
$ret;
2060 private static function normalizeUrlComponent( $component, $unsafe ) {
2061 $callback =
function (
$matches )
use ( $unsafe ) {
2063 $ord = ord( $char );
2064 if ( $ord > 32 && $ord < 127 && strpos( $unsafe, $char ) ===
false ) {
2068 # Leave it escaped, but use uppercase for a-f
2072 return preg_replace_callback(
'/%[0-9A-Fa-f]{2}/', $callback, $component );
2083 private function maybeMakeExternalImage( $url ) {
2084 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
2085 $imagesexception = !empty( $imagesfrom );
2087 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2088 if ( $imagesexception && is_array( $imagesfrom ) ) {
2089 $imagematch =
false;
2090 foreach ( $imagesfrom
as $match ) {
2091 if ( strpos( $url, $match ) === 0 ) {
2096 } elseif ( $imagesexception ) {
2097 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
2099 $imagematch =
false;
2102 if ( $this->mOptions->getAllowExternalImages()
2103 || ( $imagesexception && $imagematch )
2105 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
2110 if ( !$text && $this->mOptions->getEnableImageWhitelist()
2111 && preg_match( self::EXT_IMAGE_REGEX, $url )
2113 $whitelist = explode(
2115 wfMessage(
'external_image_whitelist' )->inContentLanguage()->
text()
2118 foreach ( $whitelist
as $entry ) {
2119 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2120 if ( strpos( $entry,
'#' ) === 0 || $entry ===
'' ) {
2123 if ( preg_match(
'/' . str_replace(
'/',
'\\/', $entry ) .
'/i', $url ) ) {
2124 # Image matches a whitelist entry
2142 public function replaceInternalLinks(
$s ) {
2143 $this->mLinkHolders->merge( $this->replaceInternalLinks2(
$s ) );
2155 public function replaceInternalLinks2( &
$s ) {
2158 static $tc =
false, $e1, $e1_img;
2159 # the % is needed to support urlencoded titles as well
2162 # Match a link having the form [[namespace:link|alternate]]trail
2163 $e1 =
"/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
2164 # Match cases where there is no "]]", which might still be images
2165 $e1_img =
"/^([{$tc}]+)\\|(.*)\$/sD";
2170 # split the entire text string on occurrences of [[
2172 # get the first element (all text up to first [[), and remove the space we added
2175 $line = $a->current(); # Workaround
for broken ArrayIterator::next()
that returns "
void"
2176 $s = substr(
$s, 1 );
2178 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
2180 if ( $useLinkPrefixExtension ) {
2181 # Match the end of a line for a word that's not followed by whitespace,
2182 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2183 $charset = $this->contLang->linkPrefixCharset();
2184 $e2 =
"/^((?>.*[^$charset]|))(.+)$/sDu";
2187 if ( is_null( $this->mTitle ) ) {
2188 throw new MWException( __METHOD__ .
": \$this->mTitle is null\n" );
2190 $nottalk = !$this->mTitle->isTalkPage();
2192 if ( $useLinkPrefixExtension ) {
2194 if ( preg_match( $e2,
$s, $m ) ) {
2195 $first_prefix = $m[2];
2197 $first_prefix =
false;
2203 $useSubpages = $this->areSubpagesAllowed();
2205 # Loop for each link
2206 for ( ;
$line !==
false &&
$line !==
null; $a->next(),
$line = $a->current() ) {
2207 # Check for excessive memory usage
2208 if ( $holders->isBig() ) {
2210 # Do the existence check, replace the link holders and clear the array
2211 $holders->replace(
$s );
2215 if ( $useLinkPrefixExtension ) {
2216 if ( preg_match( $e2,
$s, $m ) ) {
2223 if ( $first_prefix ) {
2224 $prefix = $first_prefix;
2225 $first_prefix =
false;
2229 $might_be_img =
false;
2233 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2234 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2235 # the real problem is with the $e1 regex
2237 # Still some problems for cases where the ] is meant to be outside punctuation,
2238 # and no image is in sight. See T4095.
2240 && substr( $m[3], 0, 1 ) ===
']'
2241 && strpos( $text,
'[' ) !==
false
2243 $text .=
']'; #
so that replaceExternalLinks($text) works
later
2244 $m[3] = substr( $m[3], 1 );
2246 # fix up urlencoded title texts
2247 if ( strpos( $m[1],
'%' ) !==
false ) {
2248 # Should anchors '#' also be rejected?
2249 $m[1] = str_replace( [
'<',
'>' ], [
'<',
'>' ], rawurldecode( $m[1] ) );
2252 } elseif ( preg_match( $e1_img,
$line, $m ) ) {
2253 # Invalid, but might be an image with a link in its caption
2254 $might_be_img =
true;
2256 if ( strpos( $m[1],
'%' ) !==
false ) {
2257 $m[1] = str_replace( [
'<',
'>' ], [
'<',
'>' ], rawurldecode( $m[1] ) );
2265 $origLink = ltrim( $m[1],
' ' );
2267 # Don't allow internal links to pages containing
2268 # PROTO: where PROTO is a valid URL protocol; these
2269 # should be external links.
2270 if ( preg_match(
'/^(?i:' . $this->mUrlProtocols .
')/', $origLink ) ) {
2275 # Make subpage if necessary
2276 if ( $useSubpages ) {
2277 $link = $this->maybeDoSubpageLink( $origLink, $text );
2286 $unstrip = $this->mStripState->killMarkers(
$link );
2287 $noMarkers = ( $unstrip ===
$link );
2290 if ( $nt ===
null ) {
2295 $ns = $nt->getNamespace();
2296 $iw = $nt->getInterwiki();
2298 $noforce = ( substr( $origLink, 0, 1 ) !==
':' );
2300 if ( $might_be_img ) { #
if this is actually an invalid
link
2301 if ( $ns ==
NS_FILE && $noforce ) { # but might be an image
2304 # look at the next 'line' to see if we can close it there
2306 $next_line = $a->current();
2307 if ( $next_line ===
false || $next_line ===
null ) {
2310 $m = explode(
']]', $next_line, 3 );
2311 if (
count( $m ) == 3 ) {
2312 # the first ]] closes the inner link, the second the image
2314 $text .=
"[[{$m[0]}]]{$m[1]}";
2317 } elseif (
count( $m ) == 2 ) {
2318 # if there's exactly one ]] that's fine, we'll keep looking
2319 $text .=
"[[{$m[0]}]]{$m[1]}";
2321 # if $next_line is invalid too, we need look no further
2322 $text .=
'[[' . $next_line;
2327 # we couldn't find the end of this imageLink, so output it raw
2328 # but don't ignore what might be perfectly normal links in the text we've examined
2329 $holders->merge( $this->replaceInternalLinks2( $text ) );
2330 $s .=
"{$prefix}[[$link|$text";
2331 # note: no $trail, because without an end, there *is* no trail
2334 }
else { #
it's not an image, so output it raw
2335 $s .= "{$prefix}[[$link|$text";
2336 # note: no $trail, because without an end, there *is* no trail
2341 $wasblank = ( $text == '' );
2345 # Strip off leading ':
'
2346 $text = substr( $text, 1 );
2349 # T6598 madness. Handle the quotes only if they come from the alternate part
2350 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2351 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2352 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2353 $text = $this->doQuotes( $text );
2356 # Link not escaped by : , create the various objects
2357 if ( $noforce && !$nt->wasLocalInterwiki() ) {
2360 $iw && $this->mOptions->getInterwikiMagic() && $nottalk && (
2361 Language::fetchLanguageName( $iw, null, 'mw
' ) ||
2362 in_array( $iw, $wgExtraInterlanguageLinkPrefixes )
2365 # T26502: filter duplicates
2366 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
2367 $this->mLangLinkLanguages[$iw] = true;
2368 $this->mOutput->addLanguageLink( $nt->getFullText() );
2374 $s = rtrim( $s . $prefix ) . $trail; # T175416
2378 if ( $ns == NS_FILE ) {
2379 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
2381 # if no parameters were passed, $text
2382 # becomes something like "File:Foo.png",
2383 # which we don't want to pass
on to the
2387 # recursively parse links inside the image caption
2388 # actually, this will parse them in any other parameters, too,
2389 # but it might be hard to fix that, and it doesn't matter ATM
2390 $text = $this->replaceExternalLinks( $text );
2391 $holders->merge( $this->replaceInternalLinks2( $text ) );
2393 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2394 $s .= $prefix . $this->armorLinks(
2395 $this->makeImage( $nt, $text, $holders ) ) . $trail;
2402 $s = rtrim(
$s . $prefix ) . $trail; # T2087, T87753
2405 $sortkey = $this->getDefaultSort();
2409 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
2410 $sortkey = str_replace(
"\n",
'', $sortkey );
2411 $sortkey = $this->getTargetLanguage()->convertCategoryKey( $sortkey );
2412 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
2418 # Self-link checking. For some languages, variants of the title are checked in
2419 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2420 # for linking to a different variant.
2421 if ( $ns !=
NS_SPECIAL && $nt->equals( $this->mTitle ) && !$nt->hasFragment() ) {
2426 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2427 # @todo FIXME: Should do batch file existence checks, see comment below
2429 # Give extensions a chance to select the file revision for us
2433 [ $this, $nt, &
$options, &$descQuery ] );
2434 # Fetch and register the file (file title may be different via hooks)
2435 list( $file, $nt ) = $this->fetchFileAndTitle( $nt,
$options );
2436 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2437 $s .= $prefix . $this->armorLinks(
2442 # Some titles, such as valid special pages or files in foreign repos, should
2443 # be shown as bluelinks even though they're not included in the page table
2444 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2445 # batch file existence checks for NS_FILE and NS_MEDIA
2446 if ( $iw ==
'' && $nt->isAlwaysKnown() ) {
2447 $this->mOutput->addLink( $nt );
2448 $s .= $this->makeKnownLinkHolder( $nt, $text, $trail, $prefix );
2450 # Links will be added to the output link list after checking
2451 $s .= $holders->makeHolder( $nt, $text, [], $trail, $prefix );
2470 protected function makeKnownLinkHolder( $nt, $text =
'', $trail =
'', $prefix =
'' ) {
2473 if ( $text ==
'' ) {
2474 $text = htmlspecialchars( $nt->getPrefixedText() );
2477 $link = $this->getLinkRenderer()->makeKnownLink(
2478 $nt,
new HtmlArmor(
"$prefix$text$inside" )
2481 return $this->armorLinks(
$link ) . $trail;
2494 public function armorLinks( $text ) {
2495 return preg_replace(
'/\b((?i)' . $this->mUrlProtocols .
')/',
2496 self::MARKER_PREFIX .
"NOPARSE$1", $text );
2503 public function areSubpagesAllowed() {
2504 # Some namespaces don't allow subpages
2516 public function maybeDoSubpageLink( $target, &$text ) {
2528 public function doBlockLevels( $text, $linestart ) {
2543 public function getVariableValue( $index, $frame =
false ) {
2547 if ( is_null( $this->mTitle ) ) {
2552 throw new MWException( __METHOD__ .
' Should only be '
2553 .
' called while parsing (no title set)' );
2563 if (
Hooks::run(
'ParserGetVariableValueVarCache', [ &
$parser, &$this->mVarCache ] ) ) {
2564 if ( isset( $this->mVarCache[$index] ) ) {
2565 return $this->mVarCache[$index];
2569 $ts =
wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2572 $pageLang = $this->getFunctionLang();
2578 case 'currentmonth':
2581 case 'currentmonth1':
2584 case 'currentmonthname':
2587 case 'currentmonthnamegen':
2590 case 'currentmonthabbrev':
2605 case 'localmonthname':
2608 case 'localmonthnamegen':
2611 case 'localmonthabbrev':
2626 case 'fullpagename':
2629 case 'fullpagenamee':
2635 case 'subpagenamee':
2638 case 'rootpagename':
2641 case 'rootpagenamee':
2645 $this->mTitle->getRootText()
2648 case 'basepagename':
2651 case 'basepagenamee':
2655 $this->mTitle->getBaseText()
2658 case 'talkpagename':
2659 if ( $this->mTitle->canHaveTalkPage() ) {
2660 $talkPage = $this->mTitle->getTalkPage();
2666 case 'talkpagenamee':
2667 if ( $this->mTitle->canHaveTalkPage() ) {
2668 $talkPage = $this->mTitle->getTalkPage();
2674 case 'subjectpagename':
2675 $subjPage = $this->mTitle->getSubjectPage();
2678 case 'subjectpagenamee':
2679 $subjPage = $this->mTitle->getSubjectPage();
2683 $pageid = $this->getTitle()->getArticleID();
2684 if ( $pageid == 0 ) {
2685 # 0 means the page doesn't exist in the database,
2686 # which means the user is previewing a new page.
2687 # The vary-revision flag must be set, because the magic word
2688 # will have a different value once the page is saved.
2689 $this->mOutput->setFlag(
'vary-revision' );
2690 wfDebug( __METHOD__ .
": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2692 $value = $pageid ?:
null;
2695 # Let the edit saving system know we should parse the page
2696 # *after* a revision ID has been assigned.
2697 $this->mOutput->setFlag(
'vary-revision-id' );
2698 wfDebug( __METHOD__ .
": {{REVISIONID}} used, setting vary-revision-id...\n" );
2699 $value = $this->mRevisionId;
2702 $rev = $this->getRevisionObject();
2709 $value = $this->mOptions->getSpeculativeRevId();
2711 $this->mOutput->setSpeculativeRevIdUsed(
$value );
2716 $value = (int)$this->getRevisionTimestampSubstring( 6, 2, self::MAX_TTS, $index );
2718 case 'revisionday2':
2719 $value = $this->getRevisionTimestampSubstring( 6, 2, self::MAX_TTS, $index );
2721 case 'revisionmonth':
2722 $value = $this->getRevisionTimestampSubstring( 4, 2, self::MAX_TTS, $index );
2724 case 'revisionmonth1':
2725 $value = (int)$this->getRevisionTimestampSubstring( 4, 2, self::MAX_TTS, $index );
2727 case 'revisionyear':
2728 $value = $this->getRevisionTimestampSubstring( 0, 4, self::MAX_TTS, $index );
2730 case 'revisiontimestamp':
2731 # Let the edit saving system know we should parse the page
2732 # *after* a revision ID has been assigned. This is for null edits.
2733 $this->mOutput->setFlag(
'vary-revision' );
2734 wfDebug( __METHOD__ .
": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2735 $value = $this->getRevisionTimestamp();
2737 case 'revisionuser':
2738 # Let the edit saving system know we should parse the page
2739 # *after* a revision ID has been assigned for null edits.
2740 $this->mOutput->setFlag(
'vary-user' );
2741 wfDebug( __METHOD__ .
": {{REVISIONUSER}} used, setting vary-user...\n" );
2742 $value = $this->getRevisionUser();
2744 case 'revisionsize':
2745 $value = $this->getRevisionSize();
2748 $value = str_replace(
'_',
' ',
2749 $this->contLang->getNsText( $this->mTitle->getNamespace() ) );
2752 $value =
wfUrlencode( $this->contLang->getNsText( $this->mTitle->getNamespace() ) );
2754 case 'namespacenumber':
2755 $value = $this->mTitle->getNamespace();
2758 $value = $this->mTitle->canHaveTalkPage()
2759 ? str_replace(
'_',
' ', $this->mTitle->getTalkNsText() )
2763 $value = $this->mTitle->canHaveTalkPage() ?
wfUrlencode( $this->mTitle->getTalkNsText() ) :
'';
2765 case 'subjectspace':
2766 $value = str_replace(
'_',
' ', $this->mTitle->getSubjectNsText() );
2768 case 'subjectspacee':
2771 case 'currentdayname':
2784 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2785 # int to remove the padding
2791 case 'localdayname':
2792 $value = $pageLang->getWeekdayName(
2800 $value = $pageLang->time(
2810 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2811 # int to remove the padding
2817 case 'numberofarticles':
2820 case 'numberoffiles':
2823 case 'numberofusers':
2826 case 'numberofactiveusers':
2829 case 'numberofpages':
2832 case 'numberofadmins':
2835 case 'numberofedits':
2838 case 'currenttimestamp':
2841 case 'localtimestamp':
2844 case 'currentversion':
2859 case 'directionmark':
2860 return $pageLang->getDirMark();
2861 case 'contentlanguage':
2864 case 'pagelanguage':
2865 $value = $pageLang->getCode();
2867 case 'cascadingsources':
2873 'ParserGetVariableValueSwitch',
2874 [ &
$parser, &$this->mVarCache, &$index, &
$ret, &$frame ]
2881 $this->mVarCache[$index] =
$value;
2894 private function getRevisionTimestampSubstring( $start, $len, $mtts, $variable ) {
2895 # Get the timezone-adjusted timestamp to be used for this revision
2896 $resNow = substr( $this->getRevisionTimestamp(), $start, $len );
2897 # Possibly set vary-revision if there is not yet an associated revision
2898 if ( !$this->getRevisionObject() ) {
2899 # Get the timezone-adjusted timestamp $mtts seconds in the future
2901 $this->contLang->userAdjust(
wfTimestamp( TS_MW, time() + $mtts ),
'' ),
2906 if ( $resNow !== $resThen ) {
2907 # Let the edit saving system know we should parse the page
2908 # *after* a revision ID has been assigned. This is for null edits.
2909 $this->mOutput->setFlag(
'vary-revision' );
2910 wfDebug( __METHOD__ .
": $variable used, setting vary-revision...\n" );
2922 public function initialiseVariables() {
2923 $variableIDs = $this->magicWordFactory->getVariableIDs();
2924 $substIDs = $this->magicWordFactory->getSubstIDs();
2926 $this->mVariables = $this->magicWordFactory->newArray( $variableIDs );
2927 $this->mSubstWords = $this->magicWordFactory->newArray( $substIDs );
2952 public function preprocessToDom( $text, $flags = 0 ) {
2953 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
2964 public static function splitWhitespace(
$s ) {
2965 $ltrimmed = ltrim(
$s );
2966 $w1 = substr(
$s, 0, strlen(
$s ) - strlen( $ltrimmed ) );
2967 $trimmed = rtrim( $ltrimmed );
2968 $diff = strlen( $ltrimmed ) - strlen( $trimmed );
2970 $w2 = substr( $ltrimmed, -$diff );
2974 return [ $w1, $trimmed, $w2 ];
2997 public function replaceVariables( $text, $frame =
false, $argsOnly =
false ) {
2998 # Is there any text? Also, Prevent too big inclusions!
2999 $textSize = strlen( $text );
3000 if ( $textSize < 1 || $textSize > $this->mOptions->getMaxIncludeSize() ) {
3004 if ( $frame ===
false ) {
3005 $frame = $this->getPreprocessor()->newFrame();
3006 } elseif ( !( $frame instanceof
PPFrame ) ) {
3007 wfDebug( __METHOD__ .
" called using plain parameters instead of "
3008 .
"a PPFrame instance. Creating custom frame.\n" );
3009 $frame = $this->getPreprocessor()->newCustomFrame( $frame );
3012 $dom = $this->preprocessToDom( $text );
3014 $text = $frame->expand( $dom, $flags );
3026 public static function createAssocArgs(
$args ) {
3030 $eqpos = strpos( $arg,
'=' );
3031 if ( $eqpos ===
false ) {
3032 $assocArgs[$index++] = $arg;
3034 $name = trim( substr( $arg, 0, $eqpos ) );
3035 $value = trim( substr( $arg, $eqpos + 1 ) );
3036 if (
$value ===
false ) {
3039 if (
$name !==
false ) {
3074 public function limitationWarn( $limitationType, $current =
'', $max =
'' ) {
3075 # does no harm if $current and $max are present but are unnecessary for the message
3076 # Not doing ->inLanguage( $this->mOptions->getUserLangObj() ), since this is shown
3077 # only during preview, and that would split the parser cache unnecessarily.
3078 $warning =
wfMessage(
"$limitationType-warning" )->numParams( $current, $max )
3080 $this->mOutput->addWarning( $warning );
3081 $this->addTrackingCategory(
"$limitationType-category" );
3096 public function braceSubstitution( $piece, $frame ) {
3106 $forceRawInterwiki =
false;
3108 $isChildObj =
false;
3110 $isLocalObj =
false;
3112 # Title object, where $text came from
3115 # $part1 is the bit before the first |, and must contain only title characters.
3116 # Various prefixes will be stripped from it later.
3117 $titleWithSpaces = $frame->expand( $piece[
'title'] );
3118 $part1 = trim( $titleWithSpaces );
3121 # Original title text preserved for various purposes
3122 $originalTitle = $part1;
3124 # $args is a list of argument nodes, starting from index 0, not including $part1
3125 # @todo FIXME: If piece['parts'] is null then the call to getLength()
3126 # below won't work b/c this $args isn't an object
3127 $args = (
null == $piece[
'parts'] ) ? [] : $piece[
'parts'];
3129 $profileSection =
null;
3133 $substMatch = $this->mSubstWords->matchStartAndRemove( $part1 );
3135 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3136 # Decide whether to expand template or keep wikitext as-is.
3137 if ( $this->ot[
'wiki'] ) {
3138 if ( $substMatch ===
false ) {
3139 $literal =
true; # literal when
in PST with no prefix
3141 $literal =
false; # expand when
in PST with subst:
or safesubst:
3144 if ( $substMatch ==
'subst' ) {
3145 $literal =
true; # literal when
not in PST with
plain subst:
3147 $literal =
false; # expand when
not in PST with safesubst:
or no prefix
3151 $text = $frame->virtualBracketedImplode(
'{{',
'|',
'}}', $titleWithSpaces,
$args );
3158 if ( !$found &&
$args->getLength() == 0 ) {
3159 $id = $this->mVariables->matchStartToEnd( $part1 );
3160 if ( $id !==
false ) {
3161 $text = $this->getVariableValue( $id, $frame );
3162 if ( $this->magicWordFactory->getCacheTTL( $id ) > -1 ) {
3163 $this->mOutput->updateCacheExpiry(
3164 $this->magicWordFactory->getCacheTTL( $id ) );
3170 # MSG, MSGNW and RAW
3173 $mwMsgnw = $this->magicWordFactory->get(
'msgnw' );
3174 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
3177 # Remove obsolete MSG:
3178 $mwMsg = $this->magicWordFactory->get(
'msg' );
3179 $mwMsg->matchStartAndRemove( $part1 );
3183 $mwRaw = $this->magicWordFactory->get(
'raw' );
3184 if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
3185 $forceRawInterwiki =
true;
3191 $colonPos = strpos( $part1,
':' );
3192 if ( $colonPos !==
false ) {
3193 $func = substr( $part1, 0, $colonPos );
3194 $funcArgs = [ trim( substr( $part1, $colonPos + 1 ) ) ];
3195 $argsLength =
$args->getLength();
3196 for ( $i = 0; $i < $argsLength; $i++ ) {
3197 $funcArgs[] =
$args->item( $i );
3200 $result = $this->callParserFunction( $frame, $func, $funcArgs );
3203 if ( isset(
$result[
'title'] ) ) {
3206 if ( isset(
$result[
'found'] ) ) {
3209 if ( array_key_exists(
'text',
$result ) ) {
3213 if ( isset(
$result[
'nowiki'] ) ) {
3216 if ( isset(
$result[
'isHTML'] ) ) {
3219 if ( isset(
$result[
'forceRawInterwiki'] ) ) {
3220 $forceRawInterwiki =
$result[
'forceRawInterwiki'];
3222 if ( isset(
$result[
'isChildObj'] ) ) {
3223 $isChildObj =
$result[
'isChildObj'];
3225 if ( isset(
$result[
'isLocalObj'] ) ) {
3226 $isLocalObj =
$result[
'isLocalObj'];
3231 # Finish mangling title and then check for loops.
3232 # Set $title to a Title object and $titleText to the PDBK
3235 # Split the title into page and subpage
3237 $relative = $this->maybeDoSubpageLink( $part1, $subpage );
3238 if ( $part1 !== $relative ) {
3240 $ns = $this->mTitle->getNamespace();
3244 $titleText =
$title->getPrefixedText();
3245 # Check for language variants if the template is not found
3246 if ( $this->getTargetLanguage()->hasVariants() &&
$title->getArticleID() == 0 ) {
3247 $this->getTargetLanguage()->findVariantLink( $part1,
$title,
true );
3249 # Do recursion depth check
3250 $limit = $this->mOptions->getMaxTemplateDepth();
3251 if ( $frame->depth >= $limit ) {
3253 $text =
'<span class="error">'
3254 .
wfMessage(
'parser-template-recursion-depth-warning' )
3255 ->numParams( $limit )->inContentLanguage()->text()
3261 # Load from database
3262 if ( !$found &&
$title ) {
3263 $profileSection = $this->mProfiler->scopedProfileIn(
$title->getPrefixedDBkey() );
3264 if ( !
$title->isExternal() ) {
3265 if (
$title->isSpecialPage()
3266 && $this->mOptions->getAllowSpecialInclusion()
3267 && $this->ot[
'html']
3269 $specialPage = $this->specialPageFactory->getPage(
$title->getDBkey() );
3274 $argsLength =
$args->getLength();
3275 for ( $i = 0; $i < $argsLength; $i++ ) {
3276 $bits =
$args->item( $i )->splitArg();
3277 if ( strval( $bits[
'index'] ) ===
'' ) {
3279 $value = trim( $frame->expand( $bits[
'value'] ) );
3288 if ( $specialPage && $specialPage->maxIncludeCacheTime() === 0 ) {
3289 $context->setUser( $this->getUser() );
3294 $context->setLanguage( $this->mOptions->getUserLangObj() );
3295 $ret = $this->specialPageFactory->capturePath(
$title,
$context, $this->getLinkRenderer() );
3297 $text =
$context->getOutput()->getHTML();
3298 $this->mOutput->addOutputPageMetadata(
$context->getOutput() );
3301 if ( $specialPage && $specialPage->maxIncludeCacheTime() !==
false ) {
3302 $this->mOutput->updateRuntimeAdaptiveExpiry(
3303 $specialPage->maxIncludeCacheTime()
3308 $found =
false; # access denied
3309 wfDebug( __METHOD__ .
": template inclusion denied for " .
3310 $title->getPrefixedDBkey() .
"\n" );
3313 if ( $text !==
false ) {
3319 # If the title is valid but undisplayable, make a link to it
3320 if ( !$found && ( $this->ot[
'html'] || $this->ot[
'pre'] ) ) {
3321 $text =
"[[:$titleText]]";
3324 } elseif (
$title->isTrans() ) {
3325 # Interwiki transclusion
3326 if ( $this->ot[
'html'] && !$forceRawInterwiki ) {
3327 $text = $this->interwikiTransclude(
$title,
'render' );
3330 $text = $this->interwikiTransclude(
$title,
'raw' );
3331 # Preprocess it like a template
3332 $text = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
3338 # Do infinite loop check
3339 # This has to be done after redirect resolution to avoid infinite loops via redirects
3340 if ( !$frame->loopCheck(
$title ) ) {
3342 $text =
'<span class="error">'
3343 .
wfMessage(
'parser-template-loop-warning', $titleText )->inContentLanguage()->text()
3345 $this->addTrackingCategory(
'template-loop-category' );
3346 $this->mOutput->addWarning(
wfMessage(
'template-loop-warning',
3348 wfDebug( __METHOD__ .
": template loop broken at '$titleText'\n" );
3352 # If we haven't found text to substitute by now, we're done
3353 # Recover the source wikitext and return it
3355 $text = $frame->virtualBracketedImplode(
'{{',
'|',
'}}', $titleWithSpaces,
$args );
3356 if ( $profileSection ) {
3357 $this->mProfiler->scopedProfileOut( $profileSection );
3359 return [
'object' => $text ];
3362 # Expand DOM-style return values in a child frame
3363 if ( $isChildObj ) {
3364 # Clean up argument array
3369 } elseif ( $titleText !==
false && $newFrame->isEmpty() ) {
3370 # Expansion is eligible for the empty-frame cache
3371 $text = $newFrame->cachedExpand( $titleText, $text );
3373 # Uncached expansion
3374 $text = $newFrame->expand( $text );
3377 if ( $isLocalObj && $nowiki ) {
3379 $isLocalObj =
false;
3382 if ( $profileSection ) {
3383 $this->mProfiler->scopedProfileOut( $profileSection );
3386 # Replace raw HTML by a placeholder
3388 $text = $this->insertStripItem( $text );
3389 } elseif ( $nowiki && ( $this->ot[
'html'] || $this->ot[
'pre'] ) ) {
3390 # Escape nowiki-style return values
3392 } elseif ( is_string( $text )
3393 && !$piece[
'lineStart']
3394 && preg_match(
'/^(?:{\\||:|;|#|\*)/', $text )
3396 # T2529: if the template begins with a table or block-level
3397 # element, it should be treated as beginning a new line.
3398 # This behavior is somewhat controversial.
3399 $text =
"\n" . $text;
3402 if ( is_string( $text ) && !$this->incrementIncludeSize(
'post-expand', strlen( $text ) ) ) {
3403 # Error, oversize inclusion
3404 if ( $titleText !==
false ) {
3405 # Make a working, properly escaped link if possible (T25588)
3406 $text =
"[[:$titleText]]";
3408 # This will probably not be a working link, but at least it may
3409 # provide some hint of where the problem is
3410 preg_replace(
'/^:/',
'', $originalTitle );
3411 $text =
"[[:$originalTitle]]";
3413 $text .= $this->insertStripItem(
'<!-- WARNING: template omitted, '
3414 .
'post-expand include size too large -->' );
3415 $this->limitationWarn(
'post-expand-template-inclusion' );
3418 if ( $isLocalObj ) {
3419 $ret = [
'object' => $text ];
3421 $ret = [
'text' => $text ];
3446 public function callParserFunction( $frame, $function,
array $args = [] ) {
3447 # Case sensitive functions
3448 if ( isset( $this->mFunctionSynonyms[1][$function] ) ) {
3449 $function = $this->mFunctionSynonyms[1][$function];
3451 # Case insensitive functions
3452 $function = $this->contLang->lc( $function );
3453 if ( isset( $this->mFunctionSynonyms[0][$function] ) ) {
3454 $function = $this->mFunctionSynonyms[0][$function];
3456 return [
'found' =>
false ];
3460 list( $callback, $flags ) = $this->mFunctionHooks[$function];
3467 # Convert arguments to PPNodes and collect for appending to $allArgs
3469 foreach (
$args as $k => $v ) {
3470 if ( $v instanceof
PPNode || $k === 0 ) {
3473 $funcArgs[] = $this->mPreprocessor->newPartNodeArray( [ $k => $v ] )->item( 0 );
3477 # Add a frame parameter, and pass the arguments as an array
3478 $allArgs[] = $frame;
3479 $allArgs[] = $funcArgs;
3481 # Convert arguments to plain text and append to $allArgs
3482 foreach (
$args as $k => $v ) {
3483 if ( $v instanceof
PPNode ) {
3484 $allArgs[] = trim( $frame->expand( $v ) );
3485 } elseif ( is_int( $k ) && $k >= 0 ) {
3486 $allArgs[] = trim( $v );
3488 $allArgs[] = trim(
"$k=$v" );
3493 $result = $callback( ...$allArgs );
3495 # The interface for function hooks allows them to return a wikitext
3496 # string or an array containing the string and any flags. This mungs
3497 # things around to match what this method should return.
3514 $preprocessFlags = 0;
3515 if ( isset(
$result[
'noparse'] ) ) {
3516 $noparse =
$result[
'noparse'];
3518 if ( isset(
$result[
'preprocessFlags'] ) ) {
3519 $preprocessFlags =
$result[
'preprocessFlags'];
3523 $result[
'text'] = $this->preprocessToDom(
$result[
'text'], $preprocessFlags );
3538 public function getTemplateDom(
$title ) {
3540 $titleText =
$title->getPrefixedDBkey();
3542 if ( isset( $this->mTplRedirCache[$titleText] ) ) {
3543 list( $ns, $dbk ) = $this->mTplRedirCache[$titleText];
3545 $titleText =
$title->getPrefixedDBkey();
3547 if ( isset( $this->mTplDomCache[$titleText] ) ) {
3548 return [ $this->mTplDomCache[$titleText],
$title ];
3551 # Cache miss, go to the database
3554 if ( $text ===
false ) {
3555 $this->mTplDomCache[$titleText] =
false;
3556 return [
false,
$title ];
3559 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
3560 $this->mTplDomCache[$titleText] = $dom;
3562 if ( !
$title->equals( $cacheTitle ) ) {
3563 $this->mTplRedirCache[$cacheTitle->getPrefixedDBkey()] =
3581 public function fetchCurrentRevisionOfTitle(
$title ) {
3582 $cacheKey =
$title->getPrefixedDBkey();
3583 if ( !$this->currentRevisionCache ) {
3584 $this->currentRevisionCache =
new MapCacheLRU( 100 );
3586 if ( !$this->currentRevisionCache->has( $cacheKey ) ) {
3587 $this->currentRevisionCache->set( $cacheKey,
3589 call_user_func( $this->mOptions->getCurrentRevisionCallback(),
$title, $this )
3592 return $this->currentRevisionCache->get( $cacheKey );
3615 public function fetchTemplateAndTitle(
$title ) {
3617 $templateCb = $this->mOptions->getTemplateCallback();
3618 $stuff = call_user_func( $templateCb,
$title, $this );
3620 $text = $stuff[
'text'];
3621 if ( is_string( $stuff[
'text'] ) ) {
3622 $text = strtr( $text,
"\x7f",
"?" );
3624 $finalTitle = $stuff[
'finalTitle'] ??
$title;
3625 if ( isset( $stuff[
'deps'] ) ) {
3626 foreach ( $stuff[
'deps']
as $dep ) {
3627 $this->mOutput->addTemplate( $dep[
'title'], $dep[
'page_id'], $dep[
'rev_id'] );
3628 if ( $dep[
'title']->equals( $this->getTitle() ) ) {
3631 $this->mOutput->setFlag(
'vary-revision' );
3635 return [ $text, $finalTitle ];
3643 public function fetchTemplate(
$title ) {
3644 return $this->fetchTemplateAndTitle(
$title )[0];
3656 public static function statelessFetchTemplate(
$title,
$parser =
false ) {
3657 $text = $skip =
false;
3661 # Loop to fetch the article, with up to 1 redirect
3663 for ( $i = 0; $i < 2 && is_object(
$title ); $i++ ) {
3664 # Give extensions a chance to select the revision instead
3665 $id =
false; # Assume current
3666 Hooks::run(
'BeforeParserFetchTemplateAndtitle',
3673 'page_id' =>
$title->getArticleID(),
3686 $rev_id =
$rev ?
$rev->getId() : 0;
3687 # If there is no current revision, there is no page
3688 if ( $id ===
false && !
$rev ) {
3689 $linkCache = MediaWikiServices::getInstance()->getLinkCache();
3690 $linkCache->addBadLinkObj(
$title );
3695 'page_id' =>
$title->getArticleID(),
3696 'rev_id' => $rev_id ];
3698 # We fetched a rev from a different title; register it too...
3700 'title' =>
$rev->getTitle(),
3701 'page_id' =>
$rev->getPage(),
3702 'rev_id' => $rev_id ];
3712 if ( $text ===
false || $text ===
null ) {
3717 $message =
wfMessage( MediaWikiServices::getInstance()->getContentLanguage()->
3718 lcfirst(
$title->getText() ) )->inContentLanguage();
3719 if ( !$message->exists() ) {
3724 $text = $message->plain();
3737 'finalTitle' => $finalTitle,
3764 $time = $file ? $file->getTimestamp() :
false;
3765 $sha1 = $file ? $file->getSha1() :
false;
3766 # Register the file as a dependency...
3767 $this->mOutput->addImage(
$title->getDBkey(),
$time, $sha1 );
3768 if ( $file && !
$title->equals( $file->getTitle() ) ) {
3769 # Update fetched file title
3770 $title = $file->getTitle();
3771 $this->mOutput->addImage(
$title->getDBkey(),
$time, $sha1 );
3773 return [ $file,
$title ];
3786 protected function fetchFileNoRegister(
$title,
$options = [] ) {
3787 if ( isset(
$options[
'broken'] ) ) {
3789 } elseif ( isset(
$options[
'sha1'] ) ) {
3805 public function interwikiTransclude(
$title, $action ) {
3809 return wfMessage(
'scarytranscludedisabled' )->inContentLanguage()->text();
3812 $url =
$title->getFullURL( [
'action' => $action ] );
3813 if ( strlen( $url ) > 1024 ) {
3814 return wfMessage(
'scarytranscludetoolong' )->inContentLanguage()->text();
3817 $wikiId =
$title->getTransWikiID();
3820 $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
3822 $data =
$cache->getWithSetCallback(
3824 'interwiki-transclude',
3825 ( $wikiId !==
false ) ? $wikiId :
'external',
3834 $ttl = $cache::TTL_UNCACHEABLE;
3835 } elseif (
$req->getResponseHeader(
'X-Database-Lagged' ) !== null ) {
3836 $ttl = min( $cache::TTL_LAGGED, $ttl );
3840 'text' =>
$status->isOK() ?
$req->getContent() :
null,
3841 'code' =>
$req->getStatus()
3845 'checkKeys' => ( $wikiId !==
false )
3846 ? [
$cache->makeGlobalKey(
'interwiki-page', $wikiId,
$title->getDBkey() ) ]
3848 'pcGroup' =>
'interwiki-transclude:5',
3849 'pcTTL' => $cache::TTL_PROC_LONG
3853 if ( is_string( $data[
'text'] ) ) {
3854 $text = $data[
'text'];
3855 } elseif ( $data[
'code'] != 200 ) {
3857 $text =
wfMessage(
'scarytranscludefailed-httpstatus' )
3858 ->params( $url, $data[
'code'] )->inContentLanguage()->text();
3860 $text =
wfMessage(
'scarytranscludefailed', $url )->inContentLanguage()->text();
3875 public function argSubstitution( $piece, $frame ) {
3877 $parts = $piece[
'parts'];
3878 $nameWithSpaces = $frame->expand( $piece[
'title'] );
3879 $argName = trim( $nameWithSpaces );
3881 $text = $frame->getArgument( $argName );
3882 if ( $text ===
false && $parts->getLength() > 0
3883 && ( $this->ot[
'html']
3885 || ( $this->ot[
'wiki'] && $frame->isTemplate() )
3888 # No match in frame, use the supplied default
3889 $object = $parts->item( 0 )->getChildren();
3891 if ( !$this->incrementIncludeSize(
'arg', strlen( $text ) ) ) {
3892 $error =
'<!-- WARNING: argument omitted, expansion size too large -->';
3893 $this->limitationWarn(
'post-expand-template-argument' );
3896 if ( $text ===
false && $object ===
false ) {
3898 $object = $frame->virtualBracketedImplode(
'{{{',
'|',
'}}}', $nameWithSpaces, $parts );
3900 if ( $error !==
false ) {
3903 if ( $object !==
false ) {
3904 $ret = [
'object' => $object ];
3906 $ret = [
'text' => $text ];
3927 public function extensionSubstitution(
$params, $frame ) {
3928 static $errorStr =
'<span class="error">';
3929 static $errorLen = 20;
3932 if ( substr(
$name, 0, $errorLen ) === $errorStr ) {
3938 $attrText = !isset(
$params[
'attr'] ) ? null : $frame->expand(
$params[
'attr'] );
3939 if ( substr( $attrText, 0, $errorLen ) === $errorStr ) {
3949 $marker = self::MARKER_PREFIX .
"-$name-"
3950 . sprintf(
'%08X', $this->mMarkerIndex++ ) . self::MARKER_SUFFIX;
3952 $isFunctionTag = isset( $this->mFunctionTagHooks[strtolower(
$name )] ) &&
3953 ( $this->ot[
'html'] || $this->ot[
'pre'] );
3954 if ( $isFunctionTag ) {
3955 $markerType =
'none';
3957 $markerType =
'general';
3959 if ( $this->ot[
'html'] || $isFunctionTag ) {
3961 $attributes = Sanitizer::decodeTagAttributes( $attrText );
3962 if ( isset(
$params[
'attributes'] ) ) {
3963 $attributes = $attributes +
$params[
'attributes'];
3966 if ( isset( $this->mTagHooks[
$name] ) ) {
3967 $output = call_user_func_array( $this->mTagHooks[
$name],
3968 [
$content, $attributes, $this, $frame ] );
3969 } elseif ( isset( $this->mFunctionTagHooks[
$name] ) ) {
3970 list( $callback, ) = $this->mFunctionTagHooks[
$name];
3976 $output =
'<span class="error">Invalid tag extension name: ' .
3977 htmlspecialchars(
$name ) .
'</span>';
3984 if ( isset( $flags[
'markerType'] ) ) {
3985 $markerType = $flags[
'markerType'];
3989 if ( is_null( $attrText ) ) {
3992 if ( isset(
$params[
'attributes'] ) ) {
3993 foreach (
$params[
'attributes']
as $attrName => $attrValue ) {
3994 $attrText .=
' ' . htmlspecialchars( $attrName ) .
'="' .
3995 htmlspecialchars( $attrValue ) .
'"';
3999 $output =
"<$name$attrText/>";
4001 $close = is_null(
$params[
'close'] ) ?
'' : $frame->expand(
$params[
'close'] );
4002 if ( substr( $close, 0, $errorLen ) === $errorStr ) {
4006 $output =
"<$name$attrText>$content$close";
4010 if ( $markerType ===
'none' ) {
4012 } elseif ( $markerType ===
'nowiki' ) {
4013 $this->mStripState->addNoWiki( $marker,
$output );
4014 } elseif ( $markerType ===
'general' ) {
4015 $this->mStripState->addGeneral( $marker,
$output );
4017 throw new MWException( __METHOD__ .
': invalid marker type' );
4029 public function incrementIncludeSize(
$type, $size ) {
4030 if ( $this->mIncludeSizes[
$type] + $size > $this->mOptions->getMaxIncludeSize() ) {
4033 $this->mIncludeSizes[
$type] += $size;
4043 public function incrementExpensiveFunctionCount() {
4044 $this->mExpensiveFunctionCount++;
4045 return $this->mExpensiveFunctionCount <= $this->mOptions->getExpensiveParserFunctionLimit();
4056 public function doDoubleUnderscore( $text ) {
4057 # The position of __TOC__ needs to be recorded
4058 $mw = $this->magicWordFactory->get(
'toc' );
4059 if ( $mw->match( $text ) ) {
4060 $this->mShowToc =
true;
4061 $this->mForceTocPosition =
true;
4063 # Set a placeholder. At the end we'll fill it in with the TOC.
4064 $text = $mw->replace(
'<!--MWTOC\'"-->', $text, 1 );
4066 # Only keep the first one.
4067 $text = $mw->replace(
'', $text );
4070 # Now match and remove the rest of them
4071 $mwa = $this->magicWordFactory->getDoubleUnderscoreArray();
4072 $this->mDoubleUnderscores = $mwa->matchAndRemove( $text );
4074 if ( isset( $this->mDoubleUnderscores[
'nogallery'] ) ) {
4075 $this->mOutput->mNoGallery =
true;
4077 if ( isset( $this->mDoubleUnderscores[
'notoc'] ) && !$this->mForceTocPosition ) {
4078 $this->mShowToc =
false;
4080 if ( isset( $this->mDoubleUnderscores[
'hiddencat'] )
4083 $this->addTrackingCategory(
'hidden-category-category' );
4085 # (T10068) Allow control over whether robots index a page.
4086 # __INDEX__ always overrides __NOINDEX__, see T16899
4087 if ( isset( $this->mDoubleUnderscores[
'noindex'] ) && $this->mTitle->canUseNoindex() ) {
4088 $this->mOutput->setIndexPolicy(
'noindex' );
4089 $this->addTrackingCategory(
'noindex-category' );
4091 if ( isset( $this->mDoubleUnderscores[
'index'] ) && $this->mTitle->canUseNoindex() ) {
4092 $this->mOutput->setIndexPolicy(
'index' );
4093 $this->addTrackingCategory(
'index-category' );
4096 # Cache all double underscores in the database
4097 foreach ( $this->mDoubleUnderscores
as $key => $val ) {
4098 $this->mOutput->setProperty( $key,
'' );
4109 public function addTrackingCategory( $msg ) {
4110 return $this->mOutput->addTrackingCategory( $msg, $this->mTitle );
4129 public function formatHeadings( $text, $origText, $isMain =
true ) {
4132 # Inhibit editsection links if requested in the page
4133 if ( isset( $this->mDoubleUnderscores[
'noeditsection'] ) ) {
4134 $maybeShowEditLink =
false;
4136 $maybeShowEditLink =
true;
4139 # Get all headlines for numbering them and adding funky stuff like [edit]
4140 # links - this is for later, but we need the number of headlines right now
4141 # NOTE: white space in headings have been trimmed in doHeadings. They shouldn't
4142 # be trimmed here since whitespace in HTML headings is significant.
4144 $numMatches = preg_match_all(
4145 '/<H(?P<level>[1-6])(?P<attrib>.*?>)(?P<header>[\s\S]*?)<\/H[1-6] *>/i',
4150 # if there are fewer than 4 headlines in the article, do not show TOC
4151 # unless it's been explicitly enabled.
4152 $enoughToc = $this->mShowToc &&
4153 ( ( $numMatches >= 4 ) || $this->mForceTocPosition );
4155 # Allow user to stipulate that a page should have a "new section"
4156 # link added via __NEWSECTIONLINK__
4157 if ( isset( $this->mDoubleUnderscores[
'newsectionlink'] ) ) {
4158 $this->mOutput->setNewSection(
true );
4161 # Allow user to remove the "new section"
4162 # link via __NONEWSECTIONLINK__
4163 if ( isset( $this->mDoubleUnderscores[
'nonewsectionlink'] ) ) {
4164 $this->mOutput->hideNewSection(
true );
4167 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
4168 # override above conditions and always show TOC above first header
4169 if ( isset( $this->mDoubleUnderscores[
'forcetoc'] ) ) {
4170 $this->mShowToc =
true;
4178 # Ugh .. the TOC should have neat indentation levels which can be
4179 # passed to the skin functions. These are determined here
4183 $sublevelCount = [];
4189 $markerRegex = self::MARKER_PREFIX .
"-h-(\d+)-" . self::MARKER_SUFFIX;
4190 $baseTitleText = $this->mTitle->getPrefixedDBkey();
4191 $oldType = $this->mOutputType;
4193 $frame = $this->getPreprocessor()->newFrame();
4194 $root = $this->preprocessToDom( $origText );
4195 $node = $root->getFirstChild();
4200 $headlines = $numMatches !==
false ?
$matches[3] : [];
4202 foreach ( $headlines
as $headline ) {
4203 $isTemplate =
false;
4205 $sectionIndex =
false;
4207 $markerMatches = [];
4208 if ( preg_match(
"/^$markerRegex/", $headline, $markerMatches ) ) {
4209 $serial = $markerMatches[1];
4210 list( $titleText, $sectionIndex ) = $this->mHeadings[$serial];
4211 $isTemplate = ( $titleText != $baseTitleText );
4212 $headline = preg_replace(
"/^$markerRegex\\s*/",
"", $headline );
4216 $prevlevel = $level;
4218 $level =
$matches[1][$headlineCount];
4220 if ( $level > $prevlevel ) {
4221 # Increase TOC level
4223 $sublevelCount[$toclevel] = 0;
4225 $prevtoclevel = $toclevel;
4229 } elseif ( $level < $prevlevel && $toclevel > 1 ) {
4230 # Decrease TOC level, find level to jump to
4232 for ( $i = $toclevel; $i > 0; $i-- ) {
4233 if ( $levelCount[$i] == $level ) {
4234 # Found last matching level
4237 } elseif ( $levelCount[$i] < $level ) {
4238 # Found first matching level below current level
4248 # Unindent only if the previous toc level was shown :p
4250 $prevtoclevel = $toclevel;
4256 # No change in level, end TOC line
4262 $levelCount[$toclevel] = $level;
4264 # count number of headlines for each level
4265 $sublevelCount[$toclevel]++;
4267 for ( $i = 1; $i <= $toclevel; $i++ ) {
4268 if ( !empty( $sublevelCount[$i] ) ) {
4272 $numbering .= $this->getTargetLanguage()->formatNum( $sublevelCount[$i] );
4277 # The safe header is a version of the header text safe to use for links
4279 # Remove link placeholders by the link text.
4280 # <!--LINK number-->
4282 # link text with suffix
4283 # Do this before unstrip since link text can contain strip markers
4284 $safeHeadline = $this->replaceLinkHoldersText( $headline );
4286 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4287 $safeHeadline = $this->mStripState->unstripBoth( $safeHeadline );
4289 # Remove any <style> or <script> tags (T198618)
4290 $safeHeadline = preg_replace(
4291 '#<(style|script)(?: [^>]*[^>/])?>.*?</\1>#is',
4296 # Strip out HTML (first regex removes any tag not allowed)
4298 # * <sup> and <sub> (T10393)
4302 # * <span dir="rtl"> and <span dir="ltr"> (T37167)
4303 # * <s> and <strike> (T35715)
4304 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
4305 # to allow setting directionality in toc items.
4306 $tocline = preg_replace(
4308 '#<(?!/?(span|sup|sub|bdi|i|b|s|strike)(?: [^>]*)?>).*?>#',
4309 '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|bdi|i|b|s|strike))(?: .*?)?>#'
4315 # Strip '<span></span>', which is the result from the above if
4316 # <span id="foo"></span> is used to produce an additional anchor
4318 $tocline = str_replace(
'<span></span>',
'', $tocline );
4320 $tocline = trim( $tocline );
4322 # For the anchor, strip out HTML-y stuff period
4323 $safeHeadline = preg_replace(
'/<.*?>/',
'', $safeHeadline );
4324 $safeHeadline = Sanitizer::normalizeSectionNameWhitespace( $safeHeadline );
4326 # Save headline for section edit hint before it's escaped
4327 $headlineHint = $safeHeadline;
4329 # Decode HTML entities
4330 $safeHeadline = Sanitizer::decodeCharReferences( $safeHeadline );
4332 $safeHeadline = self::normalizeSectionName( $safeHeadline );
4334 $fallbackHeadline = Sanitizer::escapeIdForAttribute( $safeHeadline, Sanitizer::ID_FALLBACK );
4335 $linkAnchor = Sanitizer::escapeIdForLink( $safeHeadline );
4336 $safeHeadline = Sanitizer::escapeIdForAttribute( $safeHeadline, Sanitizer::ID_PRIMARY );
4337 if ( $fallbackHeadline === $safeHeadline ) {
4338 # No reason to have both (in fact, we can't)
4339 $fallbackHeadline =
false;
4342 # HTML IDs must be case-insensitively unique for IE compatibility (T12721).
4343 # @todo FIXME: We may be changing them depending on the current locale.
4344 $arrayKey = strtolower( $safeHeadline );
4345 if ( $fallbackHeadline ===
false ) {
4346 $fallbackArrayKey =
false;
4348 $fallbackArrayKey = strtolower( $fallbackHeadline );
4351 # Create the anchor for linking from the TOC to the section
4352 $anchor = $safeHeadline;
4353 $fallbackAnchor = $fallbackHeadline;
4354 if ( isset( $refers[$arrayKey] ) ) {
4356 for ( $i = 2; isset( $refers[
"${arrayKey}_$i"] ); ++$i );
4358 $linkAnchor .=
"_$i";
4359 $refers[
"${arrayKey}_$i"] =
true;
4361 $refers[$arrayKey] =
true;
4363 if ( $fallbackHeadline !==
false && isset( $refers[$fallbackArrayKey] ) ) {
4365 for ( $i = 2; isset( $refers[
"${fallbackArrayKey}_$i"] ); ++$i );
4366 $fallbackAnchor .=
"_$i";
4367 $refers[
"${fallbackArrayKey}_$i"] =
true;
4369 $refers[$fallbackArrayKey] =
true;
4372 # Don't number the heading if it is the only one (looks silly)
4373 if (
count(
$matches[3] ) > 1 && $this->mOptions->getNumberHeadings() ) {
4374 # the two are different if the line contains a link
4377 [
'class' =>
'mw-headline-number' ],
4379 ) .
' ' . $headline;
4384 $numbering, $toclevel, ( $isTemplate ?
false : $sectionIndex ) );
4387 # Add the section to the section tree
4388 # Find the DOM node for this header
4389 $noOffset = ( $isTemplate || $sectionIndex ===
false );
4390 while ( $node && !$noOffset ) {
4391 if ( $node->getName() ===
'h' ) {
4392 $bits = $node->splitHeading();
4393 if ( $bits[
'i'] == $sectionIndex ) {
4397 $byteOffset += mb_strlen( $this->mStripState->unstripBoth(
4399 $node = $node->getNextSibling();
4402 'toclevel' => $toclevel,
4405 'number' => $numbering,
4406 'index' => ( $isTemplate ?
'T-' :
'' ) . $sectionIndex,
4407 'fromtitle' => $titleText,
4408 'byteoffset' => ( $noOffset ?
null : $byteOffset ),
4409 'anchor' => $anchor,
4412 # give headline the correct <h#> tag
4413 if ( $maybeShowEditLink && $sectionIndex !==
false ) {
4415 if ( $isTemplate ) {
4416 # Put a T flag in the section identifier, to indicate to extractSections()
4417 # that sections inside <includeonly> should be counted.
4418 $editsectionPage = $titleText;
4419 $editsectionSection =
"T-$sectionIndex";
4420 $editsectionContent =
null;
4422 $editsectionPage = $this->mTitle->getPrefixedText();
4423 $editsectionSection = $sectionIndex;
4424 $editsectionContent = $headlineHint;
4438 $editlink =
'<mw:editsection page="' . htmlspecialchars( $editsectionPage );
4439 $editlink .=
'" section="' . htmlspecialchars( $editsectionSection ) .
'"';
4440 if ( $editsectionContent !==
null ) {
4441 $editlink .=
'>' . $editsectionContent .
'</mw:editsection>';
4449 $matches[
'attrib'][$headlineCount], $anchor, $headline,
4450 $editlink, $fallbackAnchor );
4455 $this->setOutputType( $oldType );
4457 # Never ever show TOC if no headers
4458 if ( $numVisible < 1 ) {
4467 $this->mOutput->setTOCHTML( $toc );
4468 $toc = self::TOC_START . $toc . self::TOC_END;
4472 $this->mOutput->setSections( $tocraw );
4475 # split up and insert constructed headlines
4476 $blocks = preg_split(
'/<H[1-6].*?>[\s\S]*?<\/H[1-6]>/i', $text );
4481 foreach ( $blocks
as $block ) {
4483 if ( empty( $head[$i - 1] ) ) {
4484 $sections[$i] = $block;
4486 $sections[$i] = $head[$i - 1] . $block;
4499 Hooks::run(
'ParserSectionCreate', [ $this, $i, &$sections[$i], $maybeShowEditLink ] );
4504 if ( $enoughToc && $isMain && !$this->mForceTocPosition ) {
4507 $sections[0] = $sections[0] . $toc .
"\n";
4510 $full .= implode(
'', $sections );
4512 if ( $this->mForceTocPosition ) {
4513 return str_replace(
'<!--MWTOC\'"-->', $toc, $full );
4533 if ( $clearState ) {
4534 $magicScopeVariable = $this->lock();
4537 $this->setUser(
$user );
4540 $text = str_replace(
"\000",
'', $text );
4547 if (
$options->getPreSaveTransform() ) {
4548 $text = $this->pstPass2( $text,
$user );
4550 $text = $this->mStripState->unstripBoth( $text );
4552 $this->setUser(
null ); # Reset
4565 private function pstPass2( $text,
$user ) {
4566 # Note: This is the timestamp saved as hardcoded wikitext to the database, we use
4567 # $this->contLang here in order to give everyone the same signature and use the default one
4568 # rather than the one selected in each user's preferences. (see also T14815)
4569 $ts = $this->mOptions->getTimestamp();
4571 $ts = $timestamp->format(
'YmdHis' );
4572 $tzMsg = $timestamp->getTimezoneMessage()->inContentLanguage()->text();
4574 $d = $this->contLang->timeanddate( $ts,
false,
false ) .
" ($tzMsg)";
4576 # Variable replacement
4577 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4578 $text = $this->replaceVariables( $text );
4580 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4581 # which may corrupt this parser instance via its wfMessage()->text() call-
4584 if ( strpos( $text,
'~~~' ) !==
false ) {
4585 $sigText = $this->getUserSig(
$user );
4586 $text = strtr( $text, [
4588 '~~~~' =>
"$sigText $d",
4591 # The main two signature forms used above are time-sensitive
4592 $this->mOutput->setFlag(
'user-signature' );
4595 # Context links ("pipe tricks"): [[|name]] and [[name (context)|]]
4597 $nc =
'[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can
use non-ascii!
4600 $p1 =
"/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\))\\|]]/";
4602 $p4 =
"/\[\[(:?$nc+:|:|)($tc+?)( ?($tc+))\\|]]/";
4604 $p3 =
"/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\)|)((?:, |,)$tc+|)\\|]]/";
4606 $p2 =
"/\[\[\\|($tc+)]]/";
4608 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4609 $text = preg_replace( $p1,
'[[\\1\\2\\3|\\2]]', $text );
4610 $text = preg_replace( $p4,
'[[\\1\\2\\3|\\2]]', $text );
4611 $text = preg_replace( $p3,
'[[\\1\\2\\3\\4|\\2]]', $text );
4613 $t = $this->mTitle->getText();
4615 if ( preg_match(
"/^($nc+:|)$tc+?( \\($tc+\\))$/",
$t, $m ) ) {
4616 $text = preg_replace( $p2,
"[[$m[1]\\1$m[2]|\\1]]", $text );
4617 } elseif ( preg_match(
"/^($nc+:|)$tc+?(, $tc+|)$/",
$t, $m ) &&
"$m[1]$m[2]" !=
'' ) {
4618 $text = preg_replace( $p2,
"[[$m[1]\\1$m[2]|\\1]]", $text );
4620 # if there's no context, don't bother duplicating the title
4621 $text = preg_replace( $p2,
'[[\\1]]', $text );
4641 public function getUserSig( &
$user, $nickname =
false, $fancySig =
null ) {
4646 # If not given, retrieve from the user object.
4647 if ( $nickname ===
false ) {
4648 $nickname =
$user->getOption(
'nickname' );
4651 if ( is_null( $fancySig ) ) {
4652 $fancySig =
$user->getBoolOption(
'fancysig' );
4655 $nickname = $nickname ==
null ?
$username : $nickname;
4659 wfDebug( __METHOD__ .
": $username has overlong signature.\n" );
4660 } elseif ( $fancySig !==
false ) {
4661 # Sig. might contain markup; validate this
4662 if ( $this->validateSig( $nickname ) !==
false ) {
4663 # Validated; clean up (if needed) and return it
4664 return $this->cleanSig( $nickname,
true );
4666 # Failed to validate; fall back to the default
4668 wfDebug( __METHOD__ .
": $username has bad XML tags in signature.\n" );
4672 # Make sure nickname doesnt get a sig in a sig
4673 $nickname = self::cleanSigInSig( $nickname );
4675 # If we're still here, make it a link to the user page
4678 $msgName =
$user->isAnon() ?
'signature-anon' :
'signature';
4680 return wfMessage( $msgName, $userText, $nickText )->inContentLanguage()
4681 ->title( $this->getTitle() )->text();
4690 public function validateSig( $text ) {
4704 public function cleanSig( $text, $parsing =
false ) {
4707 $magicScopeVariable = $this->lock();
4711 # Option to disable this feature
4712 if ( !$this->mOptions->getCleanSignatures() ) {
4716 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4717 # => Move this logic to braceSubstitution()
4718 $substWord = $this->magicWordFactory->get(
'subst' );
4719 $substRegex =
'/\{\{(?!(?:' . $substWord->getBaseRegex() .
'))/x' . $substWord->getRegexCase();
4720 $substText =
'{{' . $substWord->getSynonym( 0 );
4722 $text = preg_replace( $substRegex, $substText, $text );
4723 $text = self::cleanSigInSig( $text );
4724 $dom = $this->preprocessToDom( $text );
4725 $frame = $this->getPreprocessor()->newFrame();
4726 $text = $frame->expand( $dom );
4729 $text = $this->mStripState->unstripBoth( $text );
4741 public static function cleanSigInSig( $text ) {
4742 $text = preg_replace(
'/~{3,5}/',
'', $text );
4756 $outputType, $clearState =
true
4768 $outputType, $clearState =
true
4770 $this->setTitle(
$title );
4772 $this->setOutputType( $outputType );
4773 if ( $clearState ) {
4774 $this->clearState();
4787 static $executing =
false;
4789 # Guard against infinite recursion
4830 public function setHook( $tag, callable $callback ) {
4831 $tag = strtolower( $tag );
4832 if ( preg_match(
'/[<>\r\n]/', $tag, $m ) ) {
4833 throw new MWException(
"Invalid character {$m[0]} in setHook('$tag', ...) call" );
4835 $oldVal = $this->mTagHooks[$tag] ??
null;
4836 $this->mTagHooks[$tag] = $callback;
4837 if ( !in_array( $tag, $this->mStripList ) ) {
4838 $this->mStripList[] = $tag;
4861 public function setTransparentTagHook( $tag, callable $callback ) {
4862 $tag = strtolower( $tag );
4863 if ( preg_match(
'/[<>\r\n]/', $tag, $m ) ) {
4864 throw new MWException(
"Invalid character {$m[0]} in setTransparentHook('$tag', ...) call" );
4866 $oldVal = $this->mTransparentTagHooks[$tag] ??
null;
4867 $this->mTransparentTagHooks[$tag] = $callback;
4875 public function clearTagHooks() {
4876 $this->mTagHooks = [];
4877 $this->mFunctionTagHooks = [];
4878 $this->mStripList = $this->mDefaultStripList;
4924 public function setFunctionHook( $id, callable $callback, $flags = 0 ) {
4925 $oldVal = isset( $this->mFunctionHooks[$id] ) ? $this->mFunctionHooks[$id][0] :
null;
4926 $this->mFunctionHooks[$id] = [ $callback, $flags ];
4928 # Add to function cache
4929 $mw = $this->magicWordFactory->get( $id );
4931 throw new MWException( __METHOD__ .
'() expecting a magic word identifier.' );
4934 $synonyms = $mw->getSynonyms();
4935 $sensitive = intval( $mw->isCaseSensitive() );
4937 foreach ( $synonyms
as $syn ) {
4939 if ( !$sensitive ) {
4940 $syn = $this->contLang->lc( $syn );
4946 # Remove trailing colon
4947 if ( substr( $syn, -1, 1 ) ===
':' ) {
4948 $syn = substr( $syn, 0, -1 );
4950 $this->mFunctionSynonyms[$sensitive][$syn] = $id;
4960 public function getFunctionHooks() {
4961 $this->firstCallInit();
4962 return array_keys( $this->mFunctionHooks );
4975 public function setFunctionTagHook( $tag, callable $callback, $flags ) {
4976 $tag = strtolower( $tag );
4977 if ( preg_match(
'/[<>\r\n]/', $tag, $m ) ) {
4978 throw new MWException(
"Invalid character {$m[0]} in setFunctionTagHook('$tag', ...) call" );
4980 $old = $this->mFunctionTagHooks[$tag] ??
null;
4981 $this->mFunctionTagHooks[$tag] = [ $callback, $flags ];
4983 if ( !in_array( $tag, $this->mStripList ) ) {
4984 $this->mStripList[] = $tag;
4997 public function replaceLinkHolders( &$text,
$options = 0 ) {
4998 $this->mLinkHolders->replace( $text );
5008 public function replaceLinkHoldersText( $text ) {
5009 return $this->mLinkHolders->replaceText( $text );
5025 public function renderImageGallery( $text,
$params ) {
5027 if ( isset(
$params[
'mode'] ) ) {
5033 }
catch ( Exception
$e ) {
5038 $ig->setContextTitle( $this->mTitle );
5039 $ig->setShowBytes(
false );
5040 $ig->setShowDimensions(
false );
5041 $ig->setShowFilename(
false );
5042 $ig->setParser( $this );
5043 $ig->setHideBadImages();
5044 $ig->setAttributes( Sanitizer::validateTagAttributes(
$params,
'ul' ) );
5046 if ( isset(
$params[
'showfilename'] ) ) {
5047 $ig->setShowFilename(
true );
5049 $ig->setShowFilename(
false );
5051 if ( isset(
$params[
'caption'] ) ) {
5052 $caption =
$params[
'caption'];
5053 $caption = htmlspecialchars( $caption );
5054 $caption = $this->replaceInternalLinks( $caption );
5055 $ig->setCaptionHtml( $caption );
5057 if ( isset(
$params[
'perrow'] ) ) {
5058 $ig->setPerRow(
$params[
'perrow'] );
5060 if ( isset(
$params[
'widths'] ) ) {
5061 $ig->setWidths(
$params[
'widths'] );
5063 if ( isset(
$params[
'heights'] ) ) {
5064 $ig->setHeights(
$params[
'heights'] );
5066 $ig->setAdditionalOptions(
$params );
5074 # match lines like these:
5075 # Image:someimage.jpg|This is some image
5083 if ( strpos(
$matches[0],
'%' ) !==
false ) {
5087 if ( is_null(
$title ) ) {
5088 # Bogus title. Ignore these so we don't bomb out later.
5092 # We need to get what handler the file uses, to figure out parameters.
5093 # Note, a hook can overide the file name, and chose an entirely different
5094 # file (which potentially could be of a different type and have different handler).
5099 # Don't register it now, as TraditionalImageGallery does that later.
5101 $handler = $file ? $file->getHandler() :
false;
5104 'img_alt' =>
'gallery-internal-alt',
5105 'img_link' =>
'gallery-internal-link',
5108 $paramMap = $paramMap +
$handler->getParamMap();
5111 unset( $paramMap[
'img_width'] );
5114 $mwArray = $this->magicWordFactory->newArray( array_keys( $paramMap ) );
5119 $handlerOptions = [];
5133 foreach ( $parameterMatches
as $parameterMatch ) {
5134 list( $magicName, $match ) = $mwArray->matchVariableStartToEnd( $parameterMatch );
5136 $paramName = $paramMap[$magicName];
5138 switch ( $paramName ) {
5139 case 'gallery-internal-alt':
5140 $alt = $this->stripAltText( $match,
false );
5142 case 'gallery-internal-link':
5143 $linkValue = strip_tags( $this->replaceLinkHoldersText( $match ) );
5144 if ( preg_match(
'/^-{R|(.*)}-$/', $linkValue ) ) {
5147 $linkValue = substr( $linkValue, 4, -2 );
5149 list(
$type, $target ) = $this->parseLinkParameter( $linkValue );
5150 if (
$type ===
'link-url' ) {
5152 $this->mOutput->addExternalLink( $target );
5153 } elseif (
$type ===
'link-title' ) {
5154 $link = $target->getLinkURL();
5155 $this->mOutput->addLink( $target );
5160 if (
$handler->validateParam( $paramName, $match ) ) {
5161 $handlerOptions[$paramName] = $match;
5164 wfDebug(
"$parameterMatch failed parameter validation\n" );
5165 $label = $parameterMatch;
5171 $label = $parameterMatch;
5176 $ig->add(
$title, $label, $alt,
$link, $handlerOptions );
5178 $html = $ig->toHTML();
5187 public function getImageParams(
$handler ) {
5189 $handlerClass = get_class(
$handler );
5193 if ( !isset( $this->mImageParams[$handlerClass] ) ) {
5194 # Initialise static lists
5195 static $internalParamNames = [
5196 'horizAlign' => [
'left',
'right',
'center',
'none' ],
5197 'vertAlign' => [
'baseline',
'sub',
'super',
'top',
'text-top',
'middle',
5198 'bottom',
'text-bottom' ],
5199 'frame' => [
'thumbnail',
'manualthumb',
'framed',
'frameless',
5200 'upright',
'border',
'link',
'alt',
'class' ],
5202 static $internalParamMap;
5203 if ( !$internalParamMap ) {
5204 $internalParamMap = [];
5205 foreach ( $internalParamNames
as $type => $names ) {
5212 $magicName = str_replace(
'-',
'_',
"img_$name" );
5213 $internalParamMap[$magicName] = [
$type,
$name ];
5218 # Add handler params
5219 $paramMap = $internalParamMap;
5221 $handlerParamMap =
$handler->getParamMap();
5222 foreach ( $handlerParamMap
as $magic => $paramName ) {
5223 $paramMap[$magic] = [
'handler', $paramName ];
5226 $this->mImageParams[$handlerClass] = $paramMap;
5227 $this->mImageParamsMagicArray[$handlerClass] =
5228 $this->magicWordFactory->newArray( array_keys( $paramMap ) );
5230 return [ $this->mImageParams[$handlerClass], $this->mImageParamsMagicArray[$handlerClass] ];
5241 public function makeImage(
$title,
$options, $holders =
false ) {
5242 # Check if the options text is of the form "options|alt text"
5244 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5245 # * left no resizing, just left align. label is used for alt= only
5246 # * right same, but right aligned
5247 # * none same, but not aligned
5248 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5249 # * center center the image
5250 # * frame Keep original image size, no magnify-button.
5251 # * framed Same as "frame"
5252 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5253 # * upright reduce width for upright images, rounded to full __0 px
5254 # * border draw a 1px border around the image
5255 # * alt Text for HTML alt attribute (defaults to empty)
5256 # * class Set a class for img node
5257 # * link Set the target of the image link. Can be external, interwiki, or local
5258 # vertical-align values (no % or length right now):
5268 # Protect LanguageConverter markup when splitting into parts
5273 # Give extensions a chance to select the file revision for us
5278 # Fetch and register the file (file title may be different via hooks)
5282 $handler = $file ? $file->getHandler() :
false;
5284 list( $paramMap, $mwArray ) = $this->getImageParams(
$handler );
5287 $this->addTrackingCategory(
'broken-file-category' );
5290 # Process the input parameters
5292 $params = [
'frame' => [],
'handler' => [],
5293 'horizAlign' => [],
'vertAlign' => [] ];
5294 $seenformat =
false;
5295 foreach ( $parts
as $part ) {
5296 $part = trim( $part );
5297 list( $magicName,
$value ) = $mwArray->matchVariableStartToEnd( $part );
5299 if ( isset( $paramMap[$magicName] ) ) {
5300 list(
$type, $paramName ) = $paramMap[$magicName];
5302 # Special case; width and height come in one variable together
5303 if (
$type ===
'handler' && $paramName ===
'width' ) {
5304 $parsedWidthParam = self::parseWidthParam(
$value );
5305 if ( isset( $parsedWidthParam[
'width'] ) ) {
5306 $width = $parsedWidthParam[
'width'];
5307 if (
$handler->validateParam(
'width', $width ) ) {
5312 if ( isset( $parsedWidthParam[
'height'] ) ) {
5313 $height = $parsedWidthParam[
'height'];
5314 if (
$handler->validateParam(
'height', $height ) ) {
5319 # else no validation -- T15436
5321 if (
$type ===
'handler' ) {
5322 # Validate handler parameter
5325 # Validate internal parameters
5326 switch ( $paramName ) {
5330 # @todo FIXME: Possibly check validity here for
5331 # manualthumb? downstream behavior seems odd with
5332 # missing manual thumbs.
5340 if ( $paramName ===
'no-link' ) {
5343 if ( $paramName ===
'link-url' ) {
5344 if ( $this->mOptions->getExternalLinkTarget() ) {
5345 $params[
$type][
'link-target'] = $this->mOptions->getExternalLinkTarget();
5354 $validated = !$seenformat;
5358 # Most other things appear to be empty or numeric...
5359 $validated = (
$value ===
false || is_numeric( trim(
$value ) ) );
5368 if ( !$validated ) {
5373 # Process alignment parameters
5374 if (
$params[
'horizAlign'] ) {
5381 $params[
'frame'][
'caption'] = $caption;
5383 # Will the image be presented in a frame, with the caption below?
5384 $imageIsFramed = isset(
$params[
'frame'][
'frame'] )
5385 || isset(
$params[
'frame'][
'framed'] )
5386 || isset(
$params[
'frame'][
'thumbnail'] )
5387 || isset(
$params[
'frame'][
'manualthumb'] );
5389 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5390 # came to also set the caption, ordinary text after the image -- which
5391 # makes no sense, because that just repeats the text multiple times in
5392 # screen readers. It *also* came to set the title attribute.
5393 # Now that we have an alt attribute, we should not set the alt text to
5394 # equal the caption: that's worse than useless, it just repeats the
5395 # text. This is the framed/thumbnail case. If there's no caption, we
5396 # use the unnamed parameter for alt text as well, just for the time be-
5397 # ing, if the unnamed param is set and the alt param is not.
5398 # For the future, we need to figure out if we want to tweak this more,
5399 # e.g., introducing a title= parameter for the title; ignoring the un-
5400 # named parameter entirely for images without a caption; adding an ex-
5401 # plicit caption= parameter and preserving the old magic unnamed para-
5403 if ( $imageIsFramed ) { # Framed image
5404 if ( $caption ===
'' && !isset(
$params[
'frame'][
'alt'] ) ) {
5405 # No caption or alt text, add the filename as the alt text so
5406 # that screen readers at least get some description of the image
5409 # Do not set $params['frame']['title'] because tooltips don't make sense
5411 }
else { # Inline image
5412 if ( !isset(
$params[
'frame'][
'alt'] ) ) {
5413 # No alt text, use the "caption" for the alt text
5414 if ( $caption !==
'' ) {
5415 $params[
'frame'][
'alt'] = $this->stripAltText( $caption, $holders );
5417 # No caption, fall back to using the filename for the
5422 # Use the "caption" for the tooltip text
5423 $params[
'frame'][
'title'] = $this->stripAltText( $caption, $holders );
5428 # Linker does the rest
5431 $time, $descQuery, $this->mOptions->getThumbSize() );
5433 # Give the handler a chance to modify the parser object
5435 $handler->parserTransformHook( $this, $file );
5458 public function parseLinkParameter(
$value ) {
5459 $chars = self::EXT_LINK_URL_CLASS;
5460 $addr = self::EXT_LINK_ADDR;
5461 $prots = $this->mUrlProtocols;
5466 } elseif ( preg_match(
"/^((?i)$prots)/",
$value ) ) {
5467 if ( preg_match(
"/^((?i)$prots)$addr$chars*$/u",
$value, $m ) ) {
5468 $this->mOutput->addExternalLink(
$value );
5475 $this->mOutput->addLink( $linkTitle );
5476 $type =
'link-title';
5477 $target = $linkTitle;
5480 return [
$type, $target ];
5488 protected function stripAltText( $caption, $holders ) {
5489 # Strip bad stuff out of the title (tooltip). We can't just use
5490 # replaceLinkHoldersText() here, because if this function is called
5491 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5493 $tooltip = $holders->replaceText( $caption );
5495 $tooltip = $this->replaceLinkHoldersText( $caption );
5498 # make sure there are no placeholders in thumbnail attributes
5499 # that are later expanded to html- so expand them now and
5501 $tooltip = $this->mStripState->unstripBoth( $tooltip );
5502 $tooltip = Sanitizer::stripAllTags( $tooltip );
5512 public function disableCache() {
5513 wfDebug(
"Parser output marked as uncacheable.\n" );
5514 if ( !$this->mOutput ) {
5516 " can only be called when actually parsing something" );
5518 $this->mOutput->updateCacheExpiry( 0 );
5529 public function attributeStripCallback( &$text, $frame =
false ) {
5530 $text = $this->replaceVariables( $text, $frame );
5531 $text = $this->mStripState->unstripBoth( $text );
5540 public function getTags() {
5541 $this->firstCallInit();
5543 array_keys( $this->mTransparentTagHooks ),
5544 array_keys( $this->mTagHooks ),
5545 array_keys( $this->mFunctionTagHooks )
5553 public function getFunctionSynonyms() {
5554 $this->firstCallInit();
5555 return $this->mFunctionSynonyms;
5562 public function getUrlProtocols() {
5563 return $this->mUrlProtocols;
5576 public function replaceTransparentTags( $text ) {
5578 $elements = array_keys( $this->mTransparentTagHooks );
5579 $text = self::extractTagsAndParams( $elements, $text,
$matches );
5584 $tagName = strtolower( $element );
5585 if ( isset( $this->mTransparentTagHooks[$tagName] ) ) {
5586 $output = call_user_func_array(
5587 $this->mTransparentTagHooks[$tagName],
5593 $replacements[$marker] =
$output;
5595 return strtr( $text, $replacements );
5627 private function extractSections( $text, $sectionId, $mode, $newText =
'' ) {
5630 $magicScopeVariable = $this->lock();
5633 $frame = $this->getPreprocessor()->newFrame();
5635 # Process section extraction flags
5637 $sectionParts = explode(
'-', $sectionId );
5638 $sectionIndex = array_pop( $sectionParts );
5639 foreach ( $sectionParts
as $part ) {
5640 if ( $part ===
'T' ) {
5641 $flags |= self::PTD_FOR_INCLUSION;
5645 # Check for empty input
5646 if ( strval( $text ) ===
'' ) {
5647 # Only sections 0 and T-0 exist in an empty document
5648 if ( $sectionIndex == 0 ) {
5649 if ( $mode ===
'get' ) {
5655 if ( $mode ===
'get' ) {
5663 # Preprocess the text
5664 $root = $this->preprocessToDom( $text, $flags );
5666 # <h> nodes indicate section breaks
5667 # They can only occur at the top level, so we can find them by iterating the root's children
5668 $node = $root->getFirstChild();
5670 # Find the target section
5671 if ( $sectionIndex == 0 ) {
5672 # Section zero doesn't nest, level=big
5673 $targetLevel = 1000;
5676 if ( $node->getName() ===
'h' ) {
5677 $bits = $node->splitHeading();
5678 if ( $bits[
'i'] == $sectionIndex ) {
5679 $targetLevel = $bits[
'level'];
5683 if ( $mode ===
'replace' ) {
5686 $node = $node->getNextSibling();
5692 if ( $mode ===
'get' ) {
5699 # Find the end of the section, including nested sections
5701 if ( $node->getName() ===
'h' ) {
5702 $bits = $node->splitHeading();
5703 $curLevel = $bits[
'level'];
5704 if ( $bits[
'i'] != $sectionIndex && $curLevel <= $targetLevel ) {
5708 if ( $mode ===
'get' ) {
5711 $node = $node->getNextSibling();
5714 # Write out the remainder (in replace mode only)
5715 if ( $mode ===
'replace' ) {
5716 # Output the replacement text
5717 # Add two newlines on -- trailing whitespace in $newText is conventionally
5718 # stripped by the editor, so we need both newlines to restore the paragraph gap
5719 # Only add trailing whitespace if there is newText
5720 if ( $newText !=
"" ) {
5721 $outText .= $newText .
"\n\n";
5726 $node = $node->getNextSibling();
5730 if ( is_string( $outText ) ) {
5731 # Re-insert stripped tags
5732 $outText = rtrim( $this->mStripState->unstripBoth( $outText ) );
5752 public function getSection( $text, $sectionId, $defaultText =
'' ) {
5753 return $this->extractSections( $text, $sectionId,
'get', $defaultText );
5768 public function replaceSection( $oldText, $sectionId, $newText ) {
5769 return $this->extractSections( $oldText, $sectionId,
'replace', $newText );
5777 public function getRevisionId() {
5778 return $this->mRevisionId;
5787 public function getRevisionObject() {
5788 if ( !is_null( $this->mRevisionObject ) ) {
5789 return $this->mRevisionObject;
5799 $rev = call_user_func(
5800 $this->mOptions->getCurrentRevisionCallback(), $this->getTitle(), $this
5803 if ( $this->mRevisionId ===
null &&
$rev &&
$rev->getId() ) {
5815 if ( $this->mRevisionId &&
$rev &&
$rev->getId() != $this->mRevisionId ) {
5819 $this->mRevisionObject =
$rev;
5821 return $this->mRevisionObject;
5829 public function getRevisionTimestamp() {
5830 if ( is_null( $this->mRevisionTimestamp ) ) {
5831 $revObject = $this->getRevisionObject();
5832 $timestamp = $revObject ? $revObject->getTimestamp() :
wfTimestampNow();
5834 # The cryptic '' timezone parameter tells to use the site-default
5835 # timezone offset instead of the user settings.
5836 # Since this value will be saved into the parser cache, served
5837 # to other users, and potentially even used inside links and such,
5838 # it needs to be consistent for all visitors.
5839 $this->mRevisionTimestamp = $this->contLang->userAdjust( $timestamp,
'' );
5841 return $this->mRevisionTimestamp;
5849 public function getRevisionUser() {
5850 if ( is_null( $this->mRevisionUser ) ) {
5851 $revObject = $this->getRevisionObject();
5853 # if this template is subst: the revision id will be blank,
5854 # so just use the current user's name
5856 $this->mRevisionUser = $revObject->getUserText();
5857 } elseif ( $this->ot[
'wiki'] || $this->mOptions->getIsPreview() ) {
5858 $this->mRevisionUser = $this->getUser()->getName();
5861 return $this->mRevisionUser;
5869 public function getRevisionSize() {
5870 if ( is_null( $this->mRevisionSize ) ) {
5871 $revObject = $this->getRevisionObject();
5873 # if this variable is subst: the revision id will be blank,
5874 # so just use the parser input size, because the own substituation
5875 # will change the size.
5877 $this->mRevisionSize = $revObject->getSize();
5879 $this->mRevisionSize = $this->mInputSize;
5882 return $this->mRevisionSize;
5890 public function setDefaultSort(
$sort ) {
5891 $this->mDefaultSort =
$sort;
5892 $this->mOutput->setProperty(
'defaultsort',
$sort );
5905 public function getDefaultSort() {
5906 if ( $this->mDefaultSort !==
false ) {
5907 return $this->mDefaultSort;
5919 public function getCustomDefaultSort() {
5920 return $this->mDefaultSort;
5923 private static function getSectionNameFromStrippedText( $text ) {
5924 $text = Sanitizer::normalizeSectionNameWhitespace( $text );
5925 $text = Sanitizer::decodeCharReferences( $text );
5926 $text = self::normalizeSectionName( $text );
5930 private static function makeAnchor( $sectionName ) {
5931 return '#' . Sanitizer::escapeIdForLink( $sectionName );
5934 private static function makeLegacyAnchor( $sectionName ) {
5938 $id = Sanitizer::escapeIdForAttribute( $sectionName, Sanitizer::ID_FALLBACK );
5940 $id = Sanitizer::escapeIdForLink( $sectionName );
5954 public function guessSectionNameFromWikiText( $text ) {
5955 # Strip out wikitext links(they break the anchor)
5956 $text = $this->stripSectionName( $text );
5957 $sectionName = self::getSectionNameFromStrippedText( $text );
5958 return self::makeAnchor( $sectionName );
5970 public function guessLegacySectionNameFromWikiText( $text ) {
5971 # Strip out wikitext links(they break the anchor)
5972 $text = $this->stripSectionName( $text );
5973 $sectionName = self::getSectionNameFromStrippedText( $text );
5974 return self::makeLegacyAnchor( $sectionName );
5982 public static function guessSectionNameFromStrippedText( $text ) {
5983 $sectionName = self::getSectionNameFromStrippedText( $text );
5984 return self::makeAnchor( $sectionName );
5993 private static function normalizeSectionName( $text ) {
5994 # T90902: ensure the same normalization is applied for IDs as to links
5995 $titleParser = MediaWikiServices::getInstance()->getTitleParser();
5998 $parts = $titleParser->splitTitleString(
"#$text" );
6002 return $parts[
'fragment'];
6019 public function stripSectionName( $text ) {
6020 # Strip internal link markup
6021 $text = preg_replace(
'/\[\[:?([^[|]+)\|([^[]+)\]\]/',
'$2', $text );
6022 $text = preg_replace(
'/\[\[:?([^[]+)\|?\]\]/',
'$1', $text );
6024 # Strip external link markup
6025 # @todo FIXME: Not tolerant to blank link text
6026 # I.E. [https://www.mediawiki.org] will render as [1] or something depending
6027 # on how many empty links there are on the page - need to figure that out.
6028 $text = preg_replace(
'/\[(?i:' . $this->mUrlProtocols .
')([^ ]+?) ([^[]+)\]/',
'$2', $text );
6030 # Parse wikitext quotes (italics & bold)
6031 $text = $this->doQuotes( $text );
6051 $magicScopeVariable = $this->lock();
6054 $text = $this->replaceVariables( $text );
6055 $text = $this->mStripState->unstripBoth( $text );
6056 $text = Sanitizer::removeHTMLtags( $text );
6096 public function markerSkipCallback(
$s, $callback ) {
6099 while ( $i < strlen(
$s ) ) {
6100 $markerStart = strpos(
$s, self::MARKER_PREFIX, $i );
6101 if ( $markerStart ===
false ) {
6102 $out .= call_user_func( $callback, substr(
$s, $i ) );
6105 $out .= call_user_func( $callback, substr(
$s, $i, $markerStart - $i ) );
6106 $markerEnd = strpos(
$s, self::MARKER_SUFFIX, $markerStart );
6107 if ( $markerEnd ===
false ) {
6108 $out .= substr(
$s, $markerStart );
6111 $markerEnd += strlen( self::MARKER_SUFFIX );
6112 $out .= substr(
$s, $markerStart, $markerEnd - $markerStart );
6126 public function killMarkers( $text ) {
6127 return $this->mStripState->killMarkers( $text );
6147 public function serializeHalfParsedText( $text ) {
6151 'version' => self::HALF_PARSED_VERSION,
6152 'stripState' => $this->mStripState->getSubState( $text ),
6153 'linkHolders' => $this->mLinkHolders->getSubArray( $text )
6174 public function unserializeHalfParsedText( $data ) {
6176 if ( !isset( $data[
'version'] ) || $data[
'version'] != self::HALF_PARSED_VERSION ) {
6177 throw new MWException( __METHOD__ .
': invalid version' );
6180 # First, extract the strip state.
6181 $texts = [ $data[
'text'] ];
6182 $texts = $this->mStripState->merge( $data[
'stripState'], $texts );
6184 # Now renumber links
6185 $texts = $this->mLinkHolders->mergeForeign( $data[
'linkHolders'], $texts );
6187 # Should be good to go.
6201 public function isValidHalfParsedText( $data ) {
6203 return isset( $data[
'version'] ) && $data[
'version'] == self::HALF_PARSED_VERSION;
6215 public static function parseWidthParam(
$value, $parseHeight =
true ) {
6216 $parsedWidthParam = [];
6218 return $parsedWidthParam;
6221 # (T15500) In both cases (width/height and width only),
6222 # permit trailing "px" for backward compatibility.
6223 if ( $parseHeight && preg_match(
'/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/',
$value, $m ) ) {
6224 $width = intval( $m[1] );
6225 $height = intval( $m[2] );
6226 $parsedWidthParam[
'width'] = $width;
6227 $parsedWidthParam[
'height'] = $height;
6228 } elseif ( preg_match(
'/^[0-9]*\s*(?:px)?\s*$/',
$value ) ) {
6229 $width = intval(
$value );
6230 $parsedWidthParam[
'width'] = $width;
6232 return $parsedWidthParam;
6244 protected function lock() {
6245 if ( $this->mInParse ) {
6246 throw new MWException(
"Parser state cleared while parsing. "
6247 .
"Did you call Parser::parse recursively? Lock is held by: " . $this->mInParse );
6253 $this->mInParse =
$e->getTraceAsString();
6255 $recursiveCheck =
new ScopedCallback(
function () {
6256 $this->mInParse =
false;
6259 return $recursiveCheck;
6272 public static function stripOuterParagraph(
$html ) {
6274 if ( preg_match(
'/^<p>(.*)\n?<\/p>\n?$/sU',
$html, $m ) ) {
6275 if ( strpos( $m[1],
'</p>' ) ===
false ) {
6293 public function getFreshParser() {
6294 if ( $this->mInParse ) {
6295 return $this->factory->create();
6307 public function enableOOUI() {
6308 OutputPage::setupOOUI();
6309 $this->mOutput->setEnableOOUI(
true );