MediaWiki
master
|
00001 <?php 00069 class Parser { 00075 const VERSION = '1.6.4'; 00076 00081 const HALF_PARSED_VERSION = 2; 00082 00083 # Flags for Parser::setFunctionHook 00084 # Also available as global constants from Defines.php 00085 const SFH_NO_HASH = 1; 00086 const SFH_OBJECT_ARGS = 2; 00087 00088 # Constants needed for external link processing 00089 # Everything except bracket, space, or control characters 00090 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20 00091 # as well as U+3000 is IDEOGRAPHIC SPACE for bug 19052 00092 const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F\p{Zs}]'; 00093 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)([^][<>"\\x00-\\x20\\x7F\p{Zs}]+) 00094 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu'; 00095 00096 # State constants for the definition list colon extraction 00097 const COLON_STATE_TEXT = 0; 00098 const COLON_STATE_TAG = 1; 00099 const COLON_STATE_TAGSTART = 2; 00100 const COLON_STATE_CLOSETAG = 3; 00101 const COLON_STATE_TAGSLASH = 4; 00102 const COLON_STATE_COMMENT = 5; 00103 const COLON_STATE_COMMENTDASH = 6; 00104 const COLON_STATE_COMMENTDASHDASH = 7; 00105 00106 # Flags for preprocessToDom 00107 const PTD_FOR_INCLUSION = 1; 00108 00109 # Allowed values for $this->mOutputType 00110 # Parameter to startExternalParse(). 00111 const OT_HTML = 1; # like parse() 00112 const OT_WIKI = 2; # like preSaveTransform() 00113 const OT_PREPROCESS = 3; # like preprocess() 00114 const OT_MSG = 3; 00115 const OT_PLAIN = 4; # like extractSections() - portions of the original are returned unchanged. 00116 00117 # Marker Suffix needs to be accessible staticly. 00118 const MARKER_SUFFIX = "-QINU\x7f"; 00119 00120 # Persistent: 00121 var $mTagHooks = array(); 00122 var $mTransparentTagHooks = array(); 00123 var $mFunctionHooks = array(); 00124 var $mFunctionSynonyms = array( 0 => array(), 1 => array() ); 00125 var $mFunctionTagHooks = array(); 00126 var $mStripList = array(); 00127 var $mDefaultStripList = array(); 00128 var $mVarCache = array(); 00129 var $mImageParams = array(); 00130 var $mImageParamsMagicArray = array(); 00131 var $mMarkerIndex = 0; 00132 var $mFirstCall = true; 00133 00134 # Initialised by initialiseVariables() 00135 00139 var $mVariables; 00140 00144 var $mSubstWords; 00145 var $mConf, $mPreprocessor, $mExtLinkBracketedRegex, $mUrlProtocols; # Initialised in constructor 00146 00147 # Cleared with clearState(): 00148 00151 var $mOutput; 00152 var $mAutonumber, $mDTopen; 00153 00157 var $mStripState; 00158 00159 var $mIncludeCount, $mArgStack, $mLastSection, $mInPre; 00163 var $mLinkHolders; 00164 00165 var $mLinkID; 00166 var $mIncludeSizes, $mPPNodeCount, $mGeneratedPPNodeCount, $mHighestExpansionDepth; 00167 var $mDefaultSort; 00168 var $mTplExpandCache; # empty-frame expansion cache 00169 var $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores; 00170 var $mExpensiveFunctionCount; # number of expensive parser function calls 00171 var $mShowToc, $mForceTocPosition; 00172 00176 var $mUser; # User object; only used when doing pre-save transform 00177 00178 # Temporary 00179 # These are variables reset at least once per parse regardless of $clearState 00180 00184 var $mOptions; 00185 00189 var $mTitle; # Title context, used for self-link rendering and similar things 00190 var $mOutputType; # Output type, one of the OT_xxx constants 00191 var $ot; # Shortcut alias, see setOutputType() 00192 var $mRevisionObject; # The revision object of the specified revision ID 00193 var $mRevisionId; # ID to display in {{REVISIONID}} tags 00194 var $mRevisionTimestamp; # The timestamp of the specified revision ID 00195 var $mRevisionUser; # User to display in {{REVISIONUSER}} tag 00196 var $mRevIdForTs; # The revision ID which was used to fetch the timestamp 00197 00201 var $mUniqPrefix; 00202 00208 var $mLangLinkLanguages; 00209 00215 public function __construct( $conf = array() ) { 00216 $this->mConf = $conf; 00217 $this->mUrlProtocols = wfUrlProtocols(); 00218 $this->mExtLinkBracketedRegex = '/\[(((?i)' . $this->mUrlProtocols . ')'. 00219 self::EXT_LINK_URL_CLASS.'+)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F]*?)\]/Su'; 00220 if ( isset( $conf['preprocessorClass'] ) ) { 00221 $this->mPreprocessorClass = $conf['preprocessorClass']; 00222 } elseif ( defined( 'MW_COMPILED' ) ) { 00223 # Preprocessor_Hash is much faster than Preprocessor_DOM in compiled mode 00224 $this->mPreprocessorClass = 'Preprocessor_Hash'; 00225 } elseif ( extension_loaded( 'domxml' ) ) { 00226 # PECL extension that conflicts with the core DOM extension (bug 13770) 00227 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" ); 00228 $this->mPreprocessorClass = 'Preprocessor_Hash'; 00229 } elseif ( extension_loaded( 'dom' ) ) { 00230 $this->mPreprocessorClass = 'Preprocessor_DOM'; 00231 } else { 00232 $this->mPreprocessorClass = 'Preprocessor_Hash'; 00233 } 00234 wfDebug( __CLASS__ . ": using preprocessor: {$this->mPreprocessorClass}\n" ); 00235 } 00236 00240 function __destruct() { 00241 if ( isset( $this->mLinkHolders ) ) { 00242 unset( $this->mLinkHolders ); 00243 } 00244 foreach ( $this as $name => $value ) { 00245 unset( $this->$name ); 00246 } 00247 } 00248 00252 function firstCallInit() { 00253 if ( !$this->mFirstCall ) { 00254 return; 00255 } 00256 $this->mFirstCall = false; 00257 00258 wfProfileIn( __METHOD__ ); 00259 00260 CoreParserFunctions::register( $this ); 00261 CoreTagHooks::register( $this ); 00262 $this->initialiseVariables(); 00263 00264 wfRunHooks( 'ParserFirstCallInit', array( &$this ) ); 00265 wfProfileOut( __METHOD__ ); 00266 } 00267 00273 function clearState() { 00274 wfProfileIn( __METHOD__ ); 00275 if ( $this->mFirstCall ) { 00276 $this->firstCallInit(); 00277 } 00278 $this->mOutput = new ParserOutput; 00279 $this->mOptions->registerWatcher( array( $this->mOutput, 'recordOption' ) ); 00280 $this->mAutonumber = 0; 00281 $this->mLastSection = ''; 00282 $this->mDTopen = false; 00283 $this->mIncludeCount = array(); 00284 $this->mArgStack = false; 00285 $this->mInPre = false; 00286 $this->mLinkHolders = new LinkHolderArray( $this ); 00287 $this->mLinkID = 0; 00288 $this->mRevisionObject = $this->mRevisionTimestamp = 00289 $this->mRevisionId = $this->mRevisionUser = null; 00290 $this->mVarCache = array(); 00291 $this->mUser = null; 00292 $this->mLangLinkLanguages = array(); 00293 00304 $this->mUniqPrefix = "\x7fUNIQ" . self::getRandomString(); 00305 $this->mStripState = new StripState( $this->mUniqPrefix ); 00306 00307 00308 # Clear these on every parse, bug 4549 00309 $this->mTplExpandCache = $this->mTplRedirCache = $this->mTplDomCache = array(); 00310 00311 $this->mShowToc = true; 00312 $this->mForceTocPosition = false; 00313 $this->mIncludeSizes = array( 00314 'post-expand' => 0, 00315 'arg' => 0, 00316 ); 00317 $this->mPPNodeCount = 0; 00318 $this->mGeneratedPPNodeCount = 0; 00319 $this->mHighestExpansionDepth = 0; 00320 $this->mDefaultSort = false; 00321 $this->mHeadings = array(); 00322 $this->mDoubleUnderscores = array(); 00323 $this->mExpensiveFunctionCount = 0; 00324 00325 # Fix cloning 00326 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) { 00327 $this->mPreprocessor = null; 00328 } 00329 00330 wfRunHooks( 'ParserClearState', array( &$this ) ); 00331 wfProfileOut( __METHOD__ ); 00332 } 00333 00346 public function parse( $text, Title $title, ParserOptions $options, $linestart = true, $clearState = true, $revid = null ) { 00352 global $wgUseTidy, $wgAlwaysUseTidy; 00353 $fname = __METHOD__.'-' . wfGetCaller(); 00354 wfProfileIn( __METHOD__ ); 00355 wfProfileIn( $fname ); 00356 00357 $this->startParse( $title, $options, self::OT_HTML, $clearState ); 00358 00359 # Remove the strip marker tag prefix from the input, if present. 00360 if ( $clearState ) { 00361 $text = str_replace( $this->mUniqPrefix, '', $text ); 00362 } 00363 00364 $oldRevisionId = $this->mRevisionId; 00365 $oldRevisionObject = $this->mRevisionObject; 00366 $oldRevisionTimestamp = $this->mRevisionTimestamp; 00367 $oldRevisionUser = $this->mRevisionUser; 00368 if ( $revid !== null ) { 00369 $this->mRevisionId = $revid; 00370 $this->mRevisionObject = null; 00371 $this->mRevisionTimestamp = null; 00372 $this->mRevisionUser = null; 00373 } 00374 00375 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) ); 00376 # No more strip! 00377 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) ); 00378 $text = $this->internalParse( $text ); 00379 wfRunHooks( 'ParserAfterParse', array( &$this, &$text, &$this->mStripState ) ); 00380 00381 $text = $this->mStripState->unstripGeneral( $text ); 00382 00383 # Clean up special characters, only run once, next-to-last before doBlockLevels 00384 $fixtags = array( 00385 # french spaces, last one Guillemet-left 00386 # only if there is something before the space 00387 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 ', 00388 # french spaces, Guillemet-right 00389 '/(\\302\\253) /' => '\\1 ', 00390 '/ (!\s*important)/' => ' \\1', # Beware of CSS magic word !important, bug #11874. 00391 ); 00392 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text ); 00393 00394 $text = $this->doBlockLevels( $text, $linestart ); 00395 00396 $this->replaceLinkHolders( $text ); 00397 00405 if ( !( $options->getDisableContentConversion() 00406 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) ) ) 00407 { 00408 # Run convert unconditionally in 1.18-compatible mode 00409 global $wgBug34832TransitionalRollback; 00410 if ( $wgBug34832TransitionalRollback || !$this->mOptions->getInterfaceMessage() ) { 00411 # The position of the convert() call should not be changed. it 00412 # assumes that the links are all replaced and the only thing left 00413 # is the <nowiki> mark. 00414 $text = $this->getConverterLanguage()->convert( $text ); 00415 } 00416 } 00417 00425 if ( !( $options->getDisableTitleConversion() 00426 || isset( $this->mDoubleUnderscores['nocontentconvert'] ) 00427 || isset( $this->mDoubleUnderscores['notitleconvert'] ) 00428 || $this->mOutput->getDisplayTitle() !== false ) ) 00429 { 00430 $convruletitle = $this->getConverterLanguage()->getConvRuleTitle(); 00431 if ( $convruletitle ) { 00432 $this->mOutput->setTitleText( $convruletitle ); 00433 } else { 00434 $titleText = $this->getConverterLanguage()->convertTitle( $title ); 00435 $this->mOutput->setTitleText( $titleText ); 00436 } 00437 } 00438 00439 $text = $this->mStripState->unstripNoWiki( $text ); 00440 00441 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) ); 00442 00443 $text = $this->replaceTransparentTags( $text ); 00444 $text = $this->mStripState->unstripGeneral( $text ); 00445 00446 $text = Sanitizer::normalizeCharReferences( $text ); 00447 00448 if ( ( $wgUseTidy && $this->mOptions->getTidy() ) || $wgAlwaysUseTidy ) { 00449 $text = MWTidy::tidy( $text ); 00450 } else { 00451 # attempt to sanitize at least some nesting problems 00452 # (bug #2702 and quite a few others) 00453 $tidyregs = array( 00454 # ''Something [http://www.cool.com cool''] --> 00455 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a> 00456 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' => 00457 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9', 00458 # fix up an anchor inside another anchor, only 00459 # at least for a single single nested link (bug 3695) 00460 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' => 00461 '\\1\\2</a>\\3</a>\\1\\4</a>', 00462 # fix div inside inline elements- doBlockLevels won't wrap a line which 00463 # contains a div, so fix it up here; replace 00464 # div with escaped text 00465 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' => 00466 '\\1\\3<div\\5>\\6</div>\\8\\9', 00467 # remove empty italic or bold tag pairs, some 00468 # introduced by rules above 00469 '/<([bi])><\/\\1>/' => '', 00470 ); 00471 00472 $text = preg_replace( 00473 array_keys( $tidyregs ), 00474 array_values( $tidyregs ), 00475 $text ); 00476 } 00477 00478 if ( $this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit() ) { 00479 $this->limitationWarn( 'expensive-parserfunction', 00480 $this->mExpensiveFunctionCount, 00481 $this->mOptions->getExpensiveParserFunctionLimit() 00482 ); 00483 } 00484 00485 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) ); 00486 00487 # Information on include size limits, for the benefit of users who try to skirt them 00488 if ( $this->mOptions->getEnableLimitReport() ) { 00489 $max = $this->mOptions->getMaxIncludeSize(); 00490 $PFreport = "Expensive parser function count: {$this->mExpensiveFunctionCount}/{$this->mOptions->getExpensiveParserFunctionLimit()}\n"; 00491 $limitReport = 00492 "NewPP limit report\n" . 00493 "Preprocessor visited node count: {$this->mPPNodeCount}/{$this->mOptions->getMaxPPNodeCount()}\n" . 00494 "Preprocessor generated node count: " . 00495 "{$this->mGeneratedPPNodeCount}/{$this->mOptions->getMaxGeneratedPPNodeCount()}\n" . 00496 "Post-expand include size: {$this->mIncludeSizes['post-expand']}/$max bytes\n" . 00497 "Template argument size: {$this->mIncludeSizes['arg']}/$max bytes\n". 00498 "Highest expansion depth: {$this->mHighestExpansionDepth}/{$this->mOptions->getMaxPPExpandDepth()}\n". 00499 $PFreport; 00500 wfRunHooks( 'ParserLimitReport', array( $this, &$limitReport ) ); 00501 $text .= "\n<!-- \n$limitReport-->\n"; 00502 00503 if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) { 00504 wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' . 00505 $this->mTitle->getPrefixedDBkey() ); 00506 } 00507 } 00508 $this->mOutput->setText( $text ); 00509 00510 $this->mRevisionId = $oldRevisionId; 00511 $this->mRevisionObject = $oldRevisionObject; 00512 $this->mRevisionTimestamp = $oldRevisionTimestamp; 00513 $this->mRevisionUser = $oldRevisionUser; 00514 wfProfileOut( $fname ); 00515 wfProfileOut( __METHOD__ ); 00516 00517 return $this->mOutput; 00518 } 00519 00531 function recursiveTagParse( $text, $frame=false ) { 00532 wfProfileIn( __METHOD__ ); 00533 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) ); 00534 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) ); 00535 $text = $this->internalParse( $text, false, $frame ); 00536 wfProfileOut( __METHOD__ ); 00537 return $text; 00538 } 00539 00545 function preprocess( $text, Title $title, ParserOptions $options, $revid = null ) { 00546 wfProfileIn( __METHOD__ ); 00547 $this->startParse( $title, $options, self::OT_PREPROCESS, true ); 00548 if ( $revid !== null ) { 00549 $this->mRevisionId = $revid; 00550 } 00551 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) ); 00552 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) ); 00553 $text = $this->replaceVariables( $text ); 00554 $text = $this->mStripState->unstripBoth( $text ); 00555 wfProfileOut( __METHOD__ ); 00556 return $text; 00557 } 00558 00568 public function recursivePreprocess( $text, $frame = false ) { 00569 wfProfileIn( __METHOD__ ); 00570 $text = $this->replaceVariables( $text, $frame ); 00571 $text = $this->mStripState->unstripBoth( $text ); 00572 wfProfileOut( __METHOD__ ); 00573 return $text; 00574 } 00575 00588 public function getPreloadText( $text, Title $title, ParserOptions $options ) { 00589 # Parser (re)initialisation 00590 $this->startParse( $title, $options, self::OT_PLAIN, true ); 00591 00592 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES; 00593 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION ); 00594 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags ); 00595 $text = $this->mStripState->unstripBoth( $text ); 00596 return $text; 00597 } 00598 00604 static public function getRandomString() { 00605 return wfRandomString( 16 ); 00606 } 00607 00614 function setUser( $user ) { 00615 $this->mUser = $user; 00616 } 00617 00623 public function uniqPrefix() { 00624 if ( !isset( $this->mUniqPrefix ) ) { 00625 # @todo FIXME: This is probably *horribly wrong* 00626 # LanguageConverter seems to want $wgParser's uniqPrefix, however 00627 # if this is called for a parser cache hit, the parser may not 00628 # have ever been initialized in the first place. 00629 # Not really sure what the heck is supposed to be going on here. 00630 return ''; 00631 # throw new MWException( "Accessing uninitialized mUniqPrefix" ); 00632 } 00633 return $this->mUniqPrefix; 00634 } 00635 00641 function setTitle( $t ) { 00642 if ( !$t || $t instanceof FakeTitle ) { 00643 $t = Title::newFromText( 'NO TITLE' ); 00644 } 00645 00646 if ( strval( $t->getFragment() ) !== '' ) { 00647 # Strip the fragment to avoid various odd effects 00648 $this->mTitle = clone $t; 00649 $this->mTitle->setFragment( '' ); 00650 } else { 00651 $this->mTitle = $t; 00652 } 00653 } 00654 00660 function getTitle() { 00661 return $this->mTitle; 00662 } 00663 00670 function Title( $x = null ) { 00671 return wfSetVar( $this->mTitle, $x ); 00672 } 00673 00679 function setOutputType( $ot ) { 00680 $this->mOutputType = $ot; 00681 # Shortcut alias 00682 $this->ot = array( 00683 'html' => $ot == self::OT_HTML, 00684 'wiki' => $ot == self::OT_WIKI, 00685 'pre' => $ot == self::OT_PREPROCESS, 00686 'plain' => $ot == self::OT_PLAIN, 00687 ); 00688 } 00689 00696 function OutputType( $x = null ) { 00697 return wfSetVar( $this->mOutputType, $x ); 00698 } 00699 00705 function getOutput() { 00706 return $this->mOutput; 00707 } 00708 00714 function getOptions() { 00715 return $this->mOptions; 00716 } 00717 00724 function Options( $x = null ) { 00725 return wfSetVar( $this->mOptions, $x ); 00726 } 00727 00731 function nextLinkID() { 00732 return $this->mLinkID++; 00733 } 00734 00738 function setLinkID( $id ) { 00739 $this->mLinkID = $id; 00740 } 00741 00746 function getFunctionLang() { 00747 return $this->getTargetLanguage(); 00748 } 00749 00759 public function getTargetLanguage() { 00760 $target = $this->mOptions->getTargetLanguage(); 00761 00762 if ( $target !== null ) { 00763 return $target; 00764 } elseif( $this->mOptions->getInterfaceMessage() ) { 00765 return $this->mOptions->getUserLangObj(); 00766 } elseif( is_null( $this->mTitle ) ) { 00767 throw new MWException( __METHOD__ . ': $this->mTitle is null' ); 00768 } 00769 00770 return $this->mTitle->getPageLanguage(); 00771 } 00772 00776 function getConverterLanguage() { 00777 global $wgBug34832TransitionalRollback, $wgContLang; 00778 if ( $wgBug34832TransitionalRollback ) { 00779 return $wgContLang; 00780 } else { 00781 return $this->getTargetLanguage(); 00782 } 00783 } 00784 00791 function getUser() { 00792 if ( !is_null( $this->mUser ) ) { 00793 return $this->mUser; 00794 } 00795 return $this->mOptions->getUser(); 00796 } 00797 00803 function getPreprocessor() { 00804 if ( !isset( $this->mPreprocessor ) ) { 00805 $class = $this->mPreprocessorClass; 00806 $this->mPreprocessor = new $class( $this ); 00807 } 00808 return $this->mPreprocessor; 00809 } 00810 00831 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = '' ) { 00832 static $n = 1; 00833 $stripped = ''; 00834 $matches = array(); 00835 00836 $taglist = implode( '|', $elements ); 00837 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i"; 00838 00839 while ( $text != '' ) { 00840 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE ); 00841 $stripped .= $p[0]; 00842 if ( count( $p ) < 5 ) { 00843 break; 00844 } 00845 if ( count( $p ) > 5 ) { 00846 # comment 00847 $element = $p[4]; 00848 $attributes = ''; 00849 $close = ''; 00850 $inside = $p[5]; 00851 } else { 00852 # tag 00853 $element = $p[1]; 00854 $attributes = $p[2]; 00855 $close = $p[3]; 00856 $inside = $p[4]; 00857 } 00858 00859 $marker = "$uniq_prefix-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX; 00860 $stripped .= $marker; 00861 00862 if ( $close === '/>' ) { 00863 # Empty element tag, <tag /> 00864 $content = null; 00865 $text = $inside; 00866 $tail = null; 00867 } else { 00868 if ( $element === '!--' ) { 00869 $end = '/(-->)/'; 00870 } else { 00871 $end = "/(<\\/$element\\s*>)/i"; 00872 } 00873 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE ); 00874 $content = $q[0]; 00875 if ( count( $q ) < 3 ) { 00876 # No end tag -- let it run out to the end of the text. 00877 $tail = ''; 00878 $text = ''; 00879 } else { 00880 $tail = $q[1]; 00881 $text = $q[2]; 00882 } 00883 } 00884 00885 $matches[$marker] = array( $element, 00886 $content, 00887 Sanitizer::decodeTagAttributes( $attributes ), 00888 "<$element$attributes$close$content$tail" ); 00889 } 00890 return $stripped; 00891 } 00892 00898 function getStripList() { 00899 return $this->mStripList; 00900 } 00901 00911 function insertStripItem( $text ) { 00912 $rnd = "{$this->mUniqPrefix}-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX; 00913 $this->mMarkerIndex++; 00914 $this->mStripState->addGeneral( $rnd, $text ); 00915 return $rnd; 00916 } 00917 00924 function doTableStuff( $text ) { 00925 wfProfileIn( __METHOD__ ); 00926 00927 $lines = StringUtils::explode( "\n", $text ); 00928 $out = ''; 00929 $td_history = array(); # Is currently a td tag open? 00930 $last_tag_history = array(); # Save history of last lag activated (td, th or caption) 00931 $tr_history = array(); # Is currently a tr tag open? 00932 $tr_attributes = array(); # history of tr attributes 00933 $has_opened_tr = array(); # Did this table open a <tr> element? 00934 $indent_level = 0; # indent level of the table 00935 00936 foreach ( $lines as $outLine ) { 00937 $line = trim( $outLine ); 00938 00939 if ( $line === '' ) { # empty line, go to next line 00940 $out .= $outLine."\n"; 00941 continue; 00942 } 00943 00944 $first_character = $line[0]; 00945 $matches = array(); 00946 00947 if ( preg_match( '/^(:*)\{\|(.*)$/', $line , $matches ) ) { 00948 # First check if we are starting a new table 00949 $indent_level = strlen( $matches[1] ); 00950 00951 $attributes = $this->mStripState->unstripBoth( $matches[2] ); 00952 $attributes = Sanitizer::fixTagAttributes( $attributes , 'table' ); 00953 00954 $outLine = str_repeat( '<dl><dd>' , $indent_level ) . "<table{$attributes}>"; 00955 array_push( $td_history , false ); 00956 array_push( $last_tag_history , '' ); 00957 array_push( $tr_history , false ); 00958 array_push( $tr_attributes , '' ); 00959 array_push( $has_opened_tr , false ); 00960 } elseif ( count( $td_history ) == 0 ) { 00961 # Don't do any of the following 00962 $out .= $outLine."\n"; 00963 continue; 00964 } elseif ( substr( $line , 0 , 2 ) === '|}' ) { 00965 # We are ending a table 00966 $line = '</table>' . substr( $line , 2 ); 00967 $last_tag = array_pop( $last_tag_history ); 00968 00969 if ( !array_pop( $has_opened_tr ) ) { 00970 $line = "<tr><td></td></tr>{$line}"; 00971 } 00972 00973 if ( array_pop( $tr_history ) ) { 00974 $line = "</tr>{$line}"; 00975 } 00976 00977 if ( array_pop( $td_history ) ) { 00978 $line = "</{$last_tag}>{$line}"; 00979 } 00980 array_pop( $tr_attributes ); 00981 $outLine = $line . str_repeat( '</dd></dl>' , $indent_level ); 00982 } elseif ( substr( $line , 0 , 2 ) === '|-' ) { 00983 # Now we have a table row 00984 $line = preg_replace( '#^\|-+#', '', $line ); 00985 00986 # Whats after the tag is now only attributes 00987 $attributes = $this->mStripState->unstripBoth( $line ); 00988 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' ); 00989 array_pop( $tr_attributes ); 00990 array_push( $tr_attributes, $attributes ); 00991 00992 $line = ''; 00993 $last_tag = array_pop( $last_tag_history ); 00994 array_pop( $has_opened_tr ); 00995 array_push( $has_opened_tr , true ); 00996 00997 if ( array_pop( $tr_history ) ) { 00998 $line = '</tr>'; 00999 } 01000 01001 if ( array_pop( $td_history ) ) { 01002 $line = "</{$last_tag}>{$line}"; 01003 } 01004 01005 $outLine = $line; 01006 array_push( $tr_history , false ); 01007 array_push( $td_history , false ); 01008 array_push( $last_tag_history , '' ); 01009 } elseif ( $first_character === '|' || $first_character === '!' || substr( $line , 0 , 2 ) === '|+' ) { 01010 # This might be cell elements, td, th or captions 01011 if ( substr( $line , 0 , 2 ) === '|+' ) { 01012 $first_character = '+'; 01013 $line = substr( $line , 1 ); 01014 } 01015 01016 $line = substr( $line , 1 ); 01017 01018 if ( $first_character === '!' ) { 01019 $line = str_replace( '!!' , '||' , $line ); 01020 } 01021 01022 # Split up multiple cells on the same line. 01023 # FIXME : This can result in improper nesting of tags processed 01024 # by earlier parser steps, but should avoid splitting up eg 01025 # attribute values containing literal "||". 01026 $cells = StringUtils::explodeMarkup( '||' , $line ); 01027 01028 $outLine = ''; 01029 01030 # Loop through each table cell 01031 foreach ( $cells as $cell ) { 01032 $previous = ''; 01033 if ( $first_character !== '+' ) { 01034 $tr_after = array_pop( $tr_attributes ); 01035 if ( !array_pop( $tr_history ) ) { 01036 $previous = "<tr{$tr_after}>\n"; 01037 } 01038 array_push( $tr_history , true ); 01039 array_push( $tr_attributes , '' ); 01040 array_pop( $has_opened_tr ); 01041 array_push( $has_opened_tr , true ); 01042 } 01043 01044 $last_tag = array_pop( $last_tag_history ); 01045 01046 if ( array_pop( $td_history ) ) { 01047 $previous = "</{$last_tag}>\n{$previous}"; 01048 } 01049 01050 if ( $first_character === '|' ) { 01051 $last_tag = 'td'; 01052 } elseif ( $first_character === '!' ) { 01053 $last_tag = 'th'; 01054 } elseif ( $first_character === '+' ) { 01055 $last_tag = 'caption'; 01056 } else { 01057 $last_tag = ''; 01058 } 01059 01060 array_push( $last_tag_history , $last_tag ); 01061 01062 # A cell could contain both parameters and data 01063 $cell_data = explode( '|' , $cell , 2 ); 01064 01065 # Bug 553: Note that a '|' inside an invalid link should not 01066 # be mistaken as delimiting cell parameters 01067 if ( strpos( $cell_data[0], '[[' ) !== false ) { 01068 $cell = "{$previous}<{$last_tag}>{$cell}"; 01069 } elseif ( count( $cell_data ) == 1 ) { 01070 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}"; 01071 } else { 01072 $attributes = $this->mStripState->unstripBoth( $cell_data[0] ); 01073 $attributes = Sanitizer::fixTagAttributes( $attributes , $last_tag ); 01074 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}"; 01075 } 01076 01077 $outLine .= $cell; 01078 array_push( $td_history , true ); 01079 } 01080 } 01081 $out .= $outLine . "\n"; 01082 } 01083 01084 # Closing open td, tr && table 01085 while ( count( $td_history ) > 0 ) { 01086 if ( array_pop( $td_history ) ) { 01087 $out .= "</td>\n"; 01088 } 01089 if ( array_pop( $tr_history ) ) { 01090 $out .= "</tr>\n"; 01091 } 01092 if ( !array_pop( $has_opened_tr ) ) { 01093 $out .= "<tr><td></td></tr>\n" ; 01094 } 01095 01096 $out .= "</table>\n"; 01097 } 01098 01099 # Remove trailing line-ending (b/c) 01100 if ( substr( $out, -1 ) === "\n" ) { 01101 $out = substr( $out, 0, -1 ); 01102 } 01103 01104 # special case: don't return empty table 01105 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) { 01106 $out = ''; 01107 } 01108 01109 wfProfileOut( __METHOD__ ); 01110 01111 return $out; 01112 } 01113 01126 function internalParse( $text, $isMain = true, $frame = false ) { 01127 wfProfileIn( __METHOD__ ); 01128 01129 $origText = $text; 01130 01131 # Hook to suspend the parser in this state 01132 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$this->mStripState ) ) ) { 01133 wfProfileOut( __METHOD__ ); 01134 return $text ; 01135 } 01136 01137 # if $frame is provided, then use $frame for replacing any variables 01138 if ( $frame ) { 01139 # use frame depth to infer how include/noinclude tags should be handled 01140 # depth=0 means this is the top-level document; otherwise it's an included document 01141 if ( !$frame->depth ) { 01142 $flag = 0; 01143 } else { 01144 $flag = Parser::PTD_FOR_INCLUSION; 01145 } 01146 $dom = $this->preprocessToDom( $text, $flag ); 01147 $text = $frame->expand( $dom ); 01148 } else { 01149 # if $frame is not provided, then use old-style replaceVariables 01150 $text = $this->replaceVariables( $text ); 01151 } 01152 01153 wfRunHooks( 'InternalParseBeforeSanitize', array( &$this, &$text, &$this->mStripState ) ); 01154 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ), false, array_keys( $this->mTransparentTagHooks ) ); 01155 wfRunHooks( 'InternalParseBeforeLinks', array( &$this, &$text, &$this->mStripState ) ); 01156 01157 # Tables need to come after variable replacement for things to work 01158 # properly; putting them before other transformations should keep 01159 # exciting things like link expansions from showing up in surprising 01160 # places. 01161 $text = $this->doTableStuff( $text ); 01162 01163 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text ); 01164 01165 $text = $this->doDoubleUnderscore( $text ); 01166 01167 $text = $this->doHeadings( $text ); 01168 if ( $this->mOptions->getUseDynamicDates() ) { 01169 $df = DateFormatter::getInstance(); 01170 $text = $df->reformat( $this->mOptions->getDateFormat(), $text ); 01171 } 01172 $text = $this->replaceInternalLinks( $text ); 01173 $text = $this->doAllQuotes( $text ); 01174 $text = $this->replaceExternalLinks( $text ); 01175 01176 # replaceInternalLinks may sometimes leave behind 01177 # absolute URLs, which have to be masked to hide them from replaceExternalLinks 01178 $text = str_replace( $this->mUniqPrefix.'NOPARSE', '', $text ); 01179 01180 $text = $this->doMagicLinks( $text ); 01181 $text = $this->formatHeadings( $text, $origText, $isMain ); 01182 01183 wfProfileOut( __METHOD__ ); 01184 return $text; 01185 } 01186 01198 function doMagicLinks( $text ) { 01199 wfProfileIn( __METHOD__ ); 01200 $prots = wfUrlProtocolsWithoutProtRel(); 01201 $urlChar = self::EXT_LINK_URL_CLASS; 01202 $text = preg_replace_callback( 01203 '!(?: # Start cases 01204 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text 01205 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . " 01206 (\\b(?i:$prots)$urlChar+) | # m[3]: Free external links" . ' 01207 (?:RFC|PMID)\s+([0-9]+) | # m[4]: RFC or PMID, capture number 01208 ISBN\s+(\b # m[5]: ISBN, capture number 01209 (?: 97[89] [\ \-]? )? # optional 13-digit ISBN prefix 01210 (?: [0-9] [\ \-]? ){9} # 9 digits with opt. delimiters 01211 [0-9Xx] # check digit 01212 \b) 01213 )!xu', array( &$this, 'magicLinkCallback' ), $text ); 01214 wfProfileOut( __METHOD__ ); 01215 return $text; 01216 } 01217 01223 function magicLinkCallback( $m ) { 01224 if ( isset( $m[1] ) && $m[1] !== '' ) { 01225 # Skip anchor 01226 return $m[0]; 01227 } elseif ( isset( $m[2] ) && $m[2] !== '' ) { 01228 # Skip HTML element 01229 return $m[0]; 01230 } elseif ( isset( $m[3] ) && $m[3] !== '' ) { 01231 # Free external link 01232 return $this->makeFreeExternalLink( $m[0] ); 01233 } elseif ( isset( $m[4] ) && $m[4] !== '' ) { 01234 # RFC or PMID 01235 if ( substr( $m[0], 0, 3 ) === 'RFC' ) { 01236 $keyword = 'RFC'; 01237 $urlmsg = 'rfcurl'; 01238 $CssClass = 'mw-magiclink-rfc'; 01239 $id = $m[4]; 01240 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) { 01241 $keyword = 'PMID'; 01242 $urlmsg = 'pubmedurl'; 01243 $CssClass = 'mw-magiclink-pmid'; 01244 $id = $m[4]; 01245 } else { 01246 throw new MWException( __METHOD__.': unrecognised match type "' . 01247 substr( $m[0], 0, 20 ) . '"' ); 01248 } 01249 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text(); 01250 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $CssClass ); 01251 } elseif ( isset( $m[5] ) && $m[5] !== '' ) { 01252 # ISBN 01253 $isbn = $m[5]; 01254 $num = strtr( $isbn, array( 01255 '-' => '', 01256 ' ' => '', 01257 'x' => 'X', 01258 )); 01259 $titleObj = SpecialPage::getTitleFor( 'Booksources', $num ); 01260 return'<a href="' . 01261 htmlspecialchars( $titleObj->getLocalUrl() ) . 01262 "\" class=\"internal mw-magiclink-isbn\">ISBN $isbn</a>"; 01263 } else { 01264 return $m[0]; 01265 } 01266 } 01267 01276 function makeFreeExternalLink( $url ) { 01277 wfProfileIn( __METHOD__ ); 01278 01279 $trail = ''; 01280 01281 # The characters '<' and '>' (which were escaped by 01282 # removeHTMLtags()) should not be included in 01283 # URLs, per RFC 2396. 01284 $m2 = array(); 01285 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) { 01286 $trail = substr( $url, $m2[0][1] ) . $trail; 01287 $url = substr( $url, 0, $m2[0][1] ); 01288 } 01289 01290 # Move trailing punctuation to $trail 01291 $sep = ',;\.:!?'; 01292 # If there is no left bracket, then consider right brackets fair game too 01293 if ( strpos( $url, '(' ) === false ) { 01294 $sep .= ')'; 01295 } 01296 01297 $numSepChars = strspn( strrev( $url ), $sep ); 01298 if ( $numSepChars ) { 01299 $trail = substr( $url, -$numSepChars ) . $trail; 01300 $url = substr( $url, 0, -$numSepChars ); 01301 } 01302 01303 $url = Sanitizer::cleanUrl( $url ); 01304 01305 # Is this an external image? 01306 $text = $this->maybeMakeExternalImage( $url ); 01307 if ( $text === false ) { 01308 # Not an image, make a link 01309 $text = Linker::makeExternalLink( $url, 01310 $this->getConverterLanguage()->markNoConversion($url), true, 'free', 01311 $this->getExternalLinkAttribs( $url ) ); 01312 # Register it in the output object... 01313 # Replace unnecessary URL escape codes with their equivalent characters 01314 $pasteurized = self::replaceUnusualEscapes( $url ); 01315 $this->mOutput->addExternalLink( $pasteurized ); 01316 } 01317 wfProfileOut( __METHOD__ ); 01318 return $text . $trail; 01319 } 01320 01321 01331 function doHeadings( $text ) { 01332 wfProfileIn( __METHOD__ ); 01333 for ( $i = 6; $i >= 1; --$i ) { 01334 $h = str_repeat( '=', $i ); 01335 $text = preg_replace( "/^$h(.+)$h\\s*$/m", 01336 "<h$i>\\1</h$i>", $text ); 01337 } 01338 wfProfileOut( __METHOD__ ); 01339 return $text; 01340 } 01341 01350 function doAllQuotes( $text ) { 01351 wfProfileIn( __METHOD__ ); 01352 $outtext = ''; 01353 $lines = StringUtils::explode( "\n", $text ); 01354 foreach ( $lines as $line ) { 01355 $outtext .= $this->doQuotes( $line ) . "\n"; 01356 } 01357 $outtext = substr( $outtext, 0,-1 ); 01358 wfProfileOut( __METHOD__ ); 01359 return $outtext; 01360 } 01361 01369 public function doQuotes( $text ) { 01370 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE ); 01371 if ( count( $arr ) == 1 ) { 01372 return $text; 01373 } else { 01374 # First, do some preliminary work. This may shift some apostrophes from 01375 # being mark-up to being text. It also counts the number of occurrences 01376 # of bold and italics mark-ups. 01377 $numbold = 0; 01378 $numitalics = 0; 01379 for ( $i = 0; $i < count( $arr ); $i++ ) { 01380 if ( ( $i % 2 ) == 1 ) { 01381 # If there are ever four apostrophes, assume the first is supposed to 01382 # be text, and the remaining three constitute mark-up for bold text. 01383 if ( strlen( $arr[$i] ) == 4 ) { 01384 $arr[$i-1] .= "'"; 01385 $arr[$i] = "'''"; 01386 } elseif ( strlen( $arr[$i] ) > 5 ) { 01387 # If there are more than 5 apostrophes in a row, assume they're all 01388 # text except for the last 5. 01389 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 ); 01390 $arr[$i] = "'''''"; 01391 } 01392 # Count the number of occurrences of bold and italics mark-ups. 01393 # We are not counting sequences of five apostrophes. 01394 if ( strlen( $arr[$i] ) == 2 ) { 01395 $numitalics++; 01396 } elseif ( strlen( $arr[$i] ) == 3 ) { 01397 $numbold++; 01398 } elseif ( strlen( $arr[$i] ) == 5 ) { 01399 $numitalics++; 01400 $numbold++; 01401 } 01402 } 01403 } 01404 01405 # If there is an odd number of both bold and italics, it is likely 01406 # that one of the bold ones was meant to be an apostrophe followed 01407 # by italics. Which one we cannot know for certain, but it is more 01408 # likely to be one that has a single-letter word before it. 01409 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) { 01410 $i = 0; 01411 $firstsingleletterword = -1; 01412 $firstmultiletterword = -1; 01413 $firstspace = -1; 01414 foreach ( $arr as $r ) { 01415 if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) ) { 01416 $x1 = substr( $arr[$i-1], -1 ); 01417 $x2 = substr( $arr[$i-1], -2, 1 ); 01418 if ( $x1 === ' ' ) { 01419 if ( $firstspace == -1 ) { 01420 $firstspace = $i; 01421 } 01422 } elseif ( $x2 === ' ') { 01423 if ( $firstsingleletterword == -1 ) { 01424 $firstsingleletterword = $i; 01425 } 01426 } else { 01427 if ( $firstmultiletterword == -1 ) { 01428 $firstmultiletterword = $i; 01429 } 01430 } 01431 } 01432 $i++; 01433 } 01434 01435 # If there is a single-letter word, use it! 01436 if ( $firstsingleletterword > -1 ) { 01437 $arr[$firstsingleletterword] = "''"; 01438 $arr[$firstsingleletterword-1] .= "'"; 01439 } elseif ( $firstmultiletterword > -1 ) { 01440 # If not, but there's a multi-letter word, use that one. 01441 $arr[$firstmultiletterword] = "''"; 01442 $arr[$firstmultiletterword-1] .= "'"; 01443 } elseif ( $firstspace > -1 ) { 01444 # ... otherwise use the first one that has neither. 01445 # (notice that it is possible for all three to be -1 if, for example, 01446 # there is only one pentuple-apostrophe in the line) 01447 $arr[$firstspace] = "''"; 01448 $arr[$firstspace-1] .= "'"; 01449 } 01450 } 01451 01452 # Now let's actually convert our apostrophic mush to HTML! 01453 $output = ''; 01454 $buffer = ''; 01455 $state = ''; 01456 $i = 0; 01457 foreach ( $arr as $r ) { 01458 if ( ( $i % 2 ) == 0 ) { 01459 if ( $state === 'both' ) { 01460 $buffer .= $r; 01461 } else { 01462 $output .= $r; 01463 } 01464 } else { 01465 if ( strlen( $r ) == 2 ) { 01466 if ( $state === 'i' ) { 01467 $output .= '</i>'; $state = ''; 01468 } elseif ( $state === 'bi' ) { 01469 $output .= '</i>'; $state = 'b'; 01470 } elseif ( $state === 'ib' ) { 01471 $output .= '</b></i><b>'; $state = 'b'; 01472 } elseif ( $state === 'both' ) { 01473 $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; 01474 } else { # $state can be 'b' or '' 01475 $output .= '<i>'; $state .= 'i'; 01476 } 01477 } elseif ( strlen( $r ) == 3 ) { 01478 if ( $state === 'b' ) { 01479 $output .= '</b>'; $state = ''; 01480 } elseif ( $state === 'bi' ) { 01481 $output .= '</i></b><i>'; $state = 'i'; 01482 } elseif ( $state === 'ib' ) { 01483 $output .= '</b>'; $state = 'i'; 01484 } elseif ( $state === 'both' ) { 01485 $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; 01486 } else { # $state can be 'i' or '' 01487 $output .= '<b>'; $state .= 'b'; 01488 } 01489 } elseif ( strlen( $r ) == 5 ) { 01490 if ( $state === 'b' ) { 01491 $output .= '</b><i>'; $state = 'i'; 01492 } elseif ( $state === 'i' ) { 01493 $output .= '</i><b>'; $state = 'b'; 01494 } elseif ( $state === 'bi' ) { 01495 $output .= '</i></b>'; $state = ''; 01496 } elseif ( $state === 'ib' ) { 01497 $output .= '</b></i>'; $state = ''; 01498 } elseif ( $state === 'both' ) { 01499 $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; 01500 } else { # ($state == '') 01501 $buffer = ''; $state = 'both'; 01502 } 01503 } 01504 } 01505 $i++; 01506 } 01507 # Now close all remaining tags. Notice that the order is important. 01508 if ( $state === 'b' || $state === 'ib' ) { 01509 $output .= '</b>'; 01510 } 01511 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) { 01512 $output .= '</i>'; 01513 } 01514 if ( $state === 'bi' ) { 01515 $output .= '</b>'; 01516 } 01517 # There might be lonely ''''', so make sure we have a buffer 01518 if ( $state === 'both' && $buffer ) { 01519 $output .= '<b><i>'.$buffer.'</i></b>'; 01520 } 01521 return $output; 01522 } 01523 } 01524 01538 function replaceExternalLinks( $text ) { 01539 wfProfileIn( __METHOD__ ); 01540 01541 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE ); 01542 if ( $bits === false ) { 01543 throw new MWException( "PCRE needs to be compiled with --enable-unicode-properties in order for MediaWiki to function" ); 01544 } 01545 $s = array_shift( $bits ); 01546 01547 $i = 0; 01548 while ( $i<count( $bits ) ) { 01549 $url = $bits[$i++]; 01550 // @todo FIXME: Unused variable. 01551 $protocol = $bits[$i++]; 01552 $text = $bits[$i++]; 01553 $trail = $bits[$i++]; 01554 01555 # The characters '<' and '>' (which were escaped by 01556 # removeHTMLtags()) should not be included in 01557 # URLs, per RFC 2396. 01558 $m2 = array(); 01559 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) { 01560 $text = substr( $url, $m2[0][1] ) . ' ' . $text; 01561 $url = substr( $url, 0, $m2[0][1] ); 01562 } 01563 01564 # If the link text is an image URL, replace it with an <img> tag 01565 # This happened by accident in the original parser, but some people used it extensively 01566 $img = $this->maybeMakeExternalImage( $text ); 01567 if ( $img !== false ) { 01568 $text = $img; 01569 } 01570 01571 $dtrail = ''; 01572 01573 # Set linktype for CSS - if URL==text, link is essentially free 01574 $linktype = ( $text === $url ) ? 'free' : 'text'; 01575 01576 # No link text, e.g. [http://domain.tld/some.link] 01577 if ( $text == '' ) { 01578 # Autonumber 01579 $langObj = $this->getTargetLanguage(); 01580 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']'; 01581 $linktype = 'autonumber'; 01582 } else { 01583 # Have link text, e.g. [http://domain.tld/some.link text]s 01584 # Check for trail 01585 list( $dtrail, $trail ) = Linker::splitTrail( $trail ); 01586 } 01587 01588 $text = $this->getConverterLanguage()->markNoConversion( $text ); 01589 01590 $url = Sanitizer::cleanUrl( $url ); 01591 01592 # Use the encoded URL 01593 # This means that users can paste URLs directly into the text 01594 # Funny characters like ö aren't valid in URLs anyway 01595 # This was changed in August 2004 01596 $s .= Linker::makeExternalLink( $url, $text, false, $linktype, 01597 $this->getExternalLinkAttribs( $url ) ) . $dtrail . $trail; 01598 01599 # Register link in the output object. 01600 # Replace unnecessary URL escape codes with the referenced character 01601 # This prevents spammers from hiding links from the filters 01602 $pasteurized = self::replaceUnusualEscapes( $url ); 01603 $this->mOutput->addExternalLink( $pasteurized ); 01604 } 01605 01606 wfProfileOut( __METHOD__ ); 01607 return $s; 01608 } 01609 01620 function getExternalLinkAttribs( $url = false ) { 01621 $attribs = array(); 01622 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions; 01623 $ns = $this->mTitle->getNamespace(); 01624 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions ) && 01625 !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions ) ) 01626 { 01627 $attribs['rel'] = 'nofollow'; 01628 } 01629 if ( $this->mOptions->getExternalLinkTarget() ) { 01630 $attribs['target'] = $this->mOptions->getExternalLinkTarget(); 01631 } 01632 return $attribs; 01633 } 01634 01646 static function replaceUnusualEscapes( $url ) { 01647 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/', 01648 array( __CLASS__, 'replaceUnusualEscapesCallback' ), $url ); 01649 } 01650 01659 private static function replaceUnusualEscapesCallback( $matches ) { 01660 $char = urldecode( $matches[0] ); 01661 $ord = ord( $char ); 01662 # Is it an unsafe or HTTP reserved character according to RFC 1738? 01663 if ( $ord > 32 && $ord < 127 && strpos( '<>"#{}|\^~[]`;/?', $char ) === false ) { 01664 # No, shouldn't be escaped 01665 return $char; 01666 } else { 01667 # Yes, leave it escaped 01668 return $matches[0]; 01669 } 01670 } 01671 01681 function maybeMakeExternalImage( $url ) { 01682 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom(); 01683 $imagesexception = !empty( $imagesfrom ); 01684 $text = false; 01685 # $imagesfrom could be either a single string or an array of strings, parse out the latter 01686 if ( $imagesexception && is_array( $imagesfrom ) ) { 01687 $imagematch = false; 01688 foreach ( $imagesfrom as $match ) { 01689 if ( strpos( $url, $match ) === 0 ) { 01690 $imagematch = true; 01691 break; 01692 } 01693 } 01694 } elseif ( $imagesexception ) { 01695 $imagematch = ( strpos( $url, $imagesfrom ) === 0 ); 01696 } else { 01697 $imagematch = false; 01698 } 01699 if ( $this->mOptions->getAllowExternalImages() 01700 || ( $imagesexception && $imagematch ) ) { 01701 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) { 01702 # Image found 01703 $text = Linker::makeExternalImage( $url ); 01704 } 01705 } 01706 if ( !$text && $this->mOptions->getEnableImageWhitelist() 01707 && preg_match( self::EXT_IMAGE_REGEX, $url ) ) { 01708 $whitelist = explode( "\n", wfMessage( 'external_image_whitelist' )->inContentLanguage()->text() ); 01709 foreach ( $whitelist as $entry ) { 01710 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments 01711 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) { 01712 continue; 01713 } 01714 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) { 01715 # Image matches a whitelist entry 01716 $text = Linker::makeExternalImage( $url ); 01717 break; 01718 } 01719 } 01720 } 01721 return $text; 01722 } 01723 01733 function replaceInternalLinks( $s ) { 01734 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) ); 01735 return $s; 01736 } 01737 01746 function replaceInternalLinks2( &$s ) { 01747 wfProfileIn( __METHOD__ ); 01748 01749 wfProfileIn( __METHOD__.'-setup' ); 01750 static $tc = FALSE, $e1, $e1_img; 01751 # the % is needed to support urlencoded titles as well 01752 if ( !$tc ) { 01753 $tc = Title::legalChars() . '#%'; 01754 # Match a link having the form [[namespace:link|alternate]]trail 01755 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; 01756 # Match cases where there is no "]]", which might still be images 01757 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; 01758 } 01759 01760 $holders = new LinkHolderArray( $this ); 01761 01762 # split the entire text string on occurrences of [[ 01763 $a = StringUtils::explode( '[[', ' ' . $s ); 01764 # get the first element (all text up to first [[), and remove the space we added 01765 $s = $a->current(); 01766 $a->next(); 01767 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void" 01768 $s = substr( $s, 1 ); 01769 01770 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension(); 01771 $e2 = null; 01772 if ( $useLinkPrefixExtension ) { 01773 # Match the end of a line for a word that's not followed by whitespace, 01774 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched 01775 $e2 = wfMessage( 'linkprefix' )->inContentLanguage()->text(); 01776 } 01777 01778 if ( is_null( $this->mTitle ) ) { 01779 wfProfileOut( __METHOD__.'-setup' ); 01780 wfProfileOut( __METHOD__ ); 01781 throw new MWException( __METHOD__.": \$this->mTitle is null\n" ); 01782 } 01783 $nottalk = !$this->mTitle->isTalkPage(); 01784 01785 if ( $useLinkPrefixExtension ) { 01786 $m = array(); 01787 if ( preg_match( $e2, $s, $m ) ) { 01788 $first_prefix = $m[2]; 01789 } else { 01790 $first_prefix = false; 01791 } 01792 } else { 01793 $prefix = ''; 01794 } 01795 01796 if ( $this->getConverterLanguage()->hasVariants() ) { 01797 $selflink = $this->getConverterLanguage()->autoConvertToAllVariants( 01798 $this->mTitle->getPrefixedText() ); 01799 } else { 01800 $selflink = array( $this->mTitle->getPrefixedText() ); 01801 } 01802 $useSubpages = $this->areSubpagesAllowed(); 01803 wfProfileOut( __METHOD__.'-setup' ); 01804 01805 # Loop for each link 01806 for ( ; $line !== false && $line !== null ; $a->next(), $line = $a->current() ) { 01807 # Check for excessive memory usage 01808 if ( $holders->isBig() ) { 01809 # Too big 01810 # Do the existence check, replace the link holders and clear the array 01811 $holders->replace( $s ); 01812 $holders->clear(); 01813 } 01814 01815 if ( $useLinkPrefixExtension ) { 01816 wfProfileIn( __METHOD__.'-prefixhandling' ); 01817 if ( preg_match( $e2, $s, $m ) ) { 01818 $prefix = $m[2]; 01819 $s = $m[1]; 01820 } else { 01821 $prefix=''; 01822 } 01823 # first link 01824 if ( $first_prefix ) { 01825 $prefix = $first_prefix; 01826 $first_prefix = false; 01827 } 01828 wfProfileOut( __METHOD__.'-prefixhandling' ); 01829 } 01830 01831 $might_be_img = false; 01832 01833 wfProfileIn( __METHOD__."-e1" ); 01834 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt 01835 $text = $m[2]; 01836 # If we get a ] at the beginning of $m[3] that means we have a link that's something like: 01837 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up, 01838 # the real problem is with the $e1 regex 01839 # See bug 1300. 01840 # 01841 # Still some problems for cases where the ] is meant to be outside punctuation, 01842 # and no image is in sight. See bug 2095. 01843 # 01844 if ( $text !== '' && 01845 substr( $m[3], 0, 1 ) === ']' && 01846 strpos( $text, '[' ) !== false 01847 ) 01848 { 01849 $text .= ']'; # so that replaceExternalLinks($text) works later 01850 $m[3] = substr( $m[3], 1 ); 01851 } 01852 # fix up urlencoded title texts 01853 if ( strpos( $m[1], '%' ) !== false ) { 01854 # Should anchors '#' also be rejected? 01855 $m[1] = str_replace( array('<', '>'), array('<', '>'), rawurldecode( $m[1] ) ); 01856 } 01857 $trail = $m[3]; 01858 } elseif ( preg_match( $e1_img, $line, $m ) ) { # Invalid, but might be an image with a link in its caption 01859 $might_be_img = true; 01860 $text = $m[2]; 01861 if ( strpos( $m[1], '%' ) !== false ) { 01862 $m[1] = rawurldecode( $m[1] ); 01863 } 01864 $trail = ""; 01865 } else { # Invalid form; output directly 01866 $s .= $prefix . '[[' . $line ; 01867 wfProfileOut( __METHOD__."-e1" ); 01868 continue; 01869 } 01870 wfProfileOut( __METHOD__."-e1" ); 01871 wfProfileIn( __METHOD__."-misc" ); 01872 01873 # Don't allow internal links to pages containing 01874 # PROTO: where PROTO is a valid URL protocol; these 01875 # should be external links. 01876 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $m[1] ) ) { 01877 $s .= $prefix . '[[' . $line ; 01878 wfProfileOut( __METHOD__."-misc" ); 01879 continue; 01880 } 01881 01882 # Make subpage if necessary 01883 if ( $useSubpages ) { 01884 $link = $this->maybeDoSubpageLink( $m[1], $text ); 01885 } else { 01886 $link = $m[1]; 01887 } 01888 01889 $noforce = ( substr( $m[1], 0, 1 ) !== ':' ); 01890 if ( !$noforce ) { 01891 # Strip off leading ':' 01892 $link = substr( $link, 1 ); 01893 } 01894 01895 wfProfileOut( __METHOD__."-misc" ); 01896 wfProfileIn( __METHOD__."-title" ); 01897 $nt = Title::newFromText( $this->mStripState->unstripNoWiki( $link ) ); 01898 if ( $nt === null ) { 01899 $s .= $prefix . '[[' . $line; 01900 wfProfileOut( __METHOD__."-title" ); 01901 continue; 01902 } 01903 01904 $ns = $nt->getNamespace(); 01905 $iw = $nt->getInterWiki(); 01906 wfProfileOut( __METHOD__."-title" ); 01907 01908 if ( $might_be_img ) { # if this is actually an invalid link 01909 wfProfileIn( __METHOD__."-might_be_img" ); 01910 if ( $ns == NS_FILE && $noforce ) { # but might be an image 01911 $found = false; 01912 while ( true ) { 01913 # look at the next 'line' to see if we can close it there 01914 $a->next(); 01915 $next_line = $a->current(); 01916 if ( $next_line === false || $next_line === null ) { 01917 break; 01918 } 01919 $m = explode( ']]', $next_line, 3 ); 01920 if ( count( $m ) == 3 ) { 01921 # the first ]] closes the inner link, the second the image 01922 $found = true; 01923 $text .= "[[{$m[0]}]]{$m[1]}"; 01924 $trail = $m[2]; 01925 break; 01926 } elseif ( count( $m ) == 2 ) { 01927 # if there's exactly one ]] that's fine, we'll keep looking 01928 $text .= "[[{$m[0]}]]{$m[1]}"; 01929 } else { 01930 # if $next_line is invalid too, we need look no further 01931 $text .= '[[' . $next_line; 01932 break; 01933 } 01934 } 01935 if ( !$found ) { 01936 # we couldn't find the end of this imageLink, so output it raw 01937 # but don't ignore what might be perfectly normal links in the text we've examined 01938 $holders->merge( $this->replaceInternalLinks2( $text ) ); 01939 $s .= "{$prefix}[[$link|$text"; 01940 # note: no $trail, because without an end, there *is* no trail 01941 wfProfileOut( __METHOD__."-might_be_img" ); 01942 continue; 01943 } 01944 } else { # it's not an image, so output it raw 01945 $s .= "{$prefix}[[$link|$text"; 01946 # note: no $trail, because without an end, there *is* no trail 01947 wfProfileOut( __METHOD__."-might_be_img" ); 01948 continue; 01949 } 01950 wfProfileOut( __METHOD__."-might_be_img" ); 01951 } 01952 01953 $wasblank = ( $text == '' ); 01954 if ( $wasblank ) { 01955 $text = $link; 01956 } else { 01957 # Bug 4598 madness. Handle the quotes only if they come from the alternate part 01958 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a> 01959 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']] 01960 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a> 01961 $text = $this->doQuotes( $text ); 01962 } 01963 01964 # Link not escaped by : , create the various objects 01965 if ( $noforce ) { 01966 # Interwikis 01967 wfProfileIn( __METHOD__."-interwiki" ); 01968 if ( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && Language::fetchLanguageName( $iw, null, 'mw' ) ) { 01969 // XXX: the above check prevents links to sites with identifiers that are not language codes 01970 01971 # Bug 24502: filter duplicates 01972 if ( !isset( $this->mLangLinkLanguages[$iw] ) ) { 01973 $this->mLangLinkLanguages[$iw] = true; 01974 $this->mOutput->addLanguageLink( $nt->getFullText() ); 01975 } 01976 01977 $s = rtrim( $s . $prefix ); 01978 $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail; 01979 wfProfileOut( __METHOD__."-interwiki" ); 01980 continue; 01981 } 01982 wfProfileOut( __METHOD__."-interwiki" ); 01983 01984 if ( $ns == NS_FILE ) { 01985 wfProfileIn( __METHOD__."-image" ); 01986 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) { 01987 if ( $wasblank ) { 01988 # if no parameters were passed, $text 01989 # becomes something like "File:Foo.png", 01990 # which we don't want to pass on to the 01991 # image generator 01992 $text = ''; 01993 } else { 01994 # recursively parse links inside the image caption 01995 # actually, this will parse them in any other parameters, too, 01996 # but it might be hard to fix that, and it doesn't matter ATM 01997 $text = $this->replaceExternalLinks( $text ); 01998 $holders->merge( $this->replaceInternalLinks2( $text ) ); 01999 } 02000 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them 02001 $s .= $prefix . $this->armorLinks( 02002 $this->makeImage( $nt, $text, $holders ) ) . $trail; 02003 } else { 02004 $s .= $prefix . $trail; 02005 } 02006 wfProfileOut( __METHOD__."-image" ); 02007 continue; 02008 } 02009 02010 if ( $ns == NS_CATEGORY ) { 02011 wfProfileIn( __METHOD__."-category" ); 02012 $s = rtrim( $s . "\n" ); # bug 87 02013 02014 if ( $wasblank ) { 02015 $sortkey = $this->getDefaultSort(); 02016 } else { 02017 $sortkey = $text; 02018 } 02019 $sortkey = Sanitizer::decodeCharReferences( $sortkey ); 02020 $sortkey = str_replace( "\n", '', $sortkey ); 02021 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey ); 02022 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey ); 02023 02028 $s .= trim( $prefix . $trail, "\n" ) == '' ? '' : $prefix . $trail; 02029 02030 wfProfileOut( __METHOD__."-category" ); 02031 continue; 02032 } 02033 } 02034 02035 # Self-link checking 02036 if ( $nt->getFragment() === '' && $ns != NS_SPECIAL ) { 02037 if ( in_array( $nt->getPrefixedText(), $selflink, true ) ) { 02038 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail ); 02039 continue; 02040 } 02041 } 02042 02043 # NS_MEDIA is a pseudo-namespace for linking directly to a file 02044 # @todo FIXME: Should do batch file existence checks, see comment below 02045 if ( $ns == NS_MEDIA ) { 02046 wfProfileIn( __METHOD__."-media" ); 02047 # Give extensions a chance to select the file revision for us 02048 $options = array(); 02049 $descQuery = false; 02050 wfRunHooks( 'BeforeParserFetchFileAndTitle', 02051 array( $this, $nt, &$options, &$descQuery ) ); 02052 # Fetch and register the file (file title may be different via hooks) 02053 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options ); 02054 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks 02055 $s .= $prefix . $this->armorLinks( 02056 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail; 02057 wfProfileOut( __METHOD__."-media" ); 02058 continue; 02059 } 02060 02061 wfProfileIn( __METHOD__."-always_known" ); 02062 # Some titles, such as valid special pages or files in foreign repos, should 02063 # be shown as bluelinks even though they're not included in the page table 02064 # 02065 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do 02066 # batch file existence checks for NS_FILE and NS_MEDIA 02067 if ( $iw == '' && $nt->isAlwaysKnown() ) { 02068 $this->mOutput->addLink( $nt ); 02069 $s .= $this->makeKnownLinkHolder( $nt, $text, array(), $trail, $prefix ); 02070 } else { 02071 # Links will be added to the output link list after checking 02072 $s .= $holders->makeHolder( $nt, $text, array(), $trail, $prefix ); 02073 } 02074 wfProfileOut( __METHOD__."-always_known" ); 02075 } 02076 wfProfileOut( __METHOD__ ); 02077 return $holders; 02078 } 02079 02094 function makeKnownLinkHolder( $nt, $text = '', $query = array(), $trail = '', $prefix = '' ) { 02095 list( $inside, $trail ) = Linker::splitTrail( $trail ); 02096 02097 if ( is_string( $query ) ) { 02098 $query = wfCgiToArray( $query ); 02099 } 02100 if ( $text == '' ) { 02101 $text = htmlspecialchars( $nt->getPrefixedText() ); 02102 } 02103 02104 $link = Linker::linkKnown( $nt, "$prefix$text$inside", array(), $query ); 02105 02106 return $this->armorLinks( $link ) . $trail; 02107 } 02108 02119 function armorLinks( $text ) { 02120 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/', 02121 "{$this->mUniqPrefix}NOPARSE$1", $text ); 02122 } 02123 02128 function areSubpagesAllowed() { 02129 # Some namespaces don't allow subpages 02130 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() ); 02131 } 02132 02141 function maybeDoSubpageLink( $target, &$text ) { 02142 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text ); 02143 } 02144 02151 function closeParagraph() { 02152 $result = ''; 02153 if ( $this->mLastSection != '' ) { 02154 $result = '</' . $this->mLastSection . ">\n"; 02155 } 02156 $this->mInPre = false; 02157 $this->mLastSection = ''; 02158 return $result; 02159 } 02160 02171 function getCommon( $st1, $st2 ) { 02172 $fl = strlen( $st1 ); 02173 $shorter = strlen( $st2 ); 02174 if ( $fl < $shorter ) { 02175 $shorter = $fl; 02176 } 02177 02178 for ( $i = 0; $i < $shorter; ++$i ) { 02179 if ( $st1[$i] != $st2[$i] ) { 02180 break; 02181 } 02182 } 02183 return $i; 02184 } 02185 02195 function openList( $char ) { 02196 $result = $this->closeParagraph(); 02197 02198 if ( '*' === $char ) { 02199 $result .= '<ul><li>'; 02200 } elseif ( '#' === $char ) { 02201 $result .= '<ol><li>'; 02202 } elseif ( ':' === $char ) { 02203 $result .= '<dl><dd>'; 02204 } elseif ( ';' === $char ) { 02205 $result .= '<dl><dt>'; 02206 $this->mDTopen = true; 02207 } else { 02208 $result = '<!-- ERR 1 -->'; 02209 } 02210 02211 return $result; 02212 } 02213 02221 function nextItem( $char ) { 02222 if ( '*' === $char || '#' === $char ) { 02223 return '</li><li>'; 02224 } elseif ( ':' === $char || ';' === $char ) { 02225 $close = '</dd>'; 02226 if ( $this->mDTopen ) { 02227 $close = '</dt>'; 02228 } 02229 if ( ';' === $char ) { 02230 $this->mDTopen = true; 02231 return $close . '<dt>'; 02232 } else { 02233 $this->mDTopen = false; 02234 return $close . '<dd>'; 02235 } 02236 } 02237 return '<!-- ERR 2 -->'; 02238 } 02239 02247 function closeList( $char ) { 02248 if ( '*' === $char ) { 02249 $text = '</li></ul>'; 02250 } elseif ( '#' === $char ) { 02251 $text = '</li></ol>'; 02252 } elseif ( ':' === $char ) { 02253 if ( $this->mDTopen ) { 02254 $this->mDTopen = false; 02255 $text = '</dt></dl>'; 02256 } else { 02257 $text = '</dd></dl>'; 02258 } 02259 } else { 02260 return '<!-- ERR 3 -->'; 02261 } 02262 return $text."\n"; 02263 } 02274 function doBlockLevels( $text, $linestart ) { 02275 wfProfileIn( __METHOD__ ); 02276 02277 # Parsing through the text line by line. The main thing 02278 # happening here is handling of block-level elements p, pre, 02279 # and making lists from lines starting with * # : etc. 02280 # 02281 $textLines = StringUtils::explode( "\n", $text ); 02282 02283 $lastPrefix = $output = ''; 02284 $this->mDTopen = $inBlockElem = false; 02285 $prefixLength = 0; 02286 $paragraphStack = false; 02287 02288 foreach ( $textLines as $oLine ) { 02289 # Fix up $linestart 02290 if ( !$linestart ) { 02291 $output .= $oLine; 02292 $linestart = true; 02293 continue; 02294 } 02295 # * = ul 02296 # # = ol 02297 # ; = dt 02298 # : = dd 02299 02300 $lastPrefixLength = strlen( $lastPrefix ); 02301 $preCloseMatch = preg_match( '/<\\/pre/i', $oLine ); 02302 $preOpenMatch = preg_match( '/<pre/i', $oLine ); 02303 # If not in a <pre> element, scan for and figure out what prefixes are there. 02304 if ( !$this->mInPre ) { 02305 # Multiple prefixes may abut each other for nested lists. 02306 $prefixLength = strspn( $oLine, '*#:;' ); 02307 $prefix = substr( $oLine, 0, $prefixLength ); 02308 02309 # eh? 02310 # ; and : are both from definition-lists, so they're equivalent 02311 # for the purposes of determining whether or not we need to open/close 02312 # elements. 02313 $prefix2 = str_replace( ';', ':', $prefix ); 02314 $t = substr( $oLine, $prefixLength ); 02315 $this->mInPre = (bool)$preOpenMatch; 02316 } else { 02317 # Don't interpret any other prefixes in preformatted text 02318 $prefixLength = 0; 02319 $prefix = $prefix2 = ''; 02320 $t = $oLine; 02321 } 02322 02323 # List generation 02324 if ( $prefixLength && $lastPrefix === $prefix2 ) { 02325 # Same as the last item, so no need to deal with nesting or opening stuff 02326 $output .= $this->nextItem( substr( $prefix, -1 ) ); 02327 $paragraphStack = false; 02328 02329 if ( substr( $prefix, -1 ) === ';') { 02330 # The one nasty exception: definition lists work like this: 02331 # ; title : definition text 02332 # So we check for : in the remainder text to split up the 02333 # title and definition, without b0rking links. 02334 $term = $t2 = ''; 02335 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) { 02336 $t = $t2; 02337 $output .= $term . $this->nextItem( ':' ); 02338 } 02339 } 02340 } elseif ( $prefixLength || $lastPrefixLength ) { 02341 # We need to open or close prefixes, or both. 02342 02343 # Either open or close a level... 02344 $commonPrefixLength = $this->getCommon( $prefix, $lastPrefix ); 02345 $paragraphStack = false; 02346 02347 # Close all the prefixes which aren't shared. 02348 while ( $commonPrefixLength < $lastPrefixLength ) { 02349 $output .= $this->closeList( $lastPrefix[$lastPrefixLength-1] ); 02350 --$lastPrefixLength; 02351 } 02352 02353 # Continue the current prefix if appropriate. 02354 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) { 02355 $output .= $this->nextItem( $prefix[$commonPrefixLength-1] ); 02356 } 02357 02358 # Open prefixes where appropriate. 02359 while ( $prefixLength > $commonPrefixLength ) { 02360 $char = substr( $prefix, $commonPrefixLength, 1 ); 02361 $output .= $this->openList( $char ); 02362 02363 if ( ';' === $char ) { 02364 # @todo FIXME: This is dupe of code above 02365 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) { 02366 $t = $t2; 02367 $output .= $term . $this->nextItem( ':' ); 02368 } 02369 } 02370 ++$commonPrefixLength; 02371 } 02372 $lastPrefix = $prefix2; 02373 } 02374 02375 # If we have no prefixes, go to paragraph mode. 02376 if ( 0 == $prefixLength ) { 02377 wfProfileIn( __METHOD__."-paragraph" ); 02378 # No prefix (not in list)--go to paragraph mode 02379 # XXX: use a stack for nestable elements like span, table and div 02380 $openmatch = preg_match('/(?:<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<ol|<dl|<li|<\\/tr|<\\/td|<\\/th)/iS', $t ); 02381 $closematch = preg_match( 02382 '/(?:<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'. 02383 '<td|<th|<\\/?div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul|<\\/ol|<\\/dl|<\\/?center)/iS', $t ); 02384 if ( $openmatch or $closematch ) { 02385 $paragraphStack = false; 02386 # TODO bug 5718: paragraph closed 02387 $output .= $this->closeParagraph(); 02388 if ( $preOpenMatch and !$preCloseMatch ) { 02389 $this->mInPre = true; 02390 } 02391 $inBlockElem = !$closematch; 02392 } elseif ( !$inBlockElem && !$this->mInPre ) { 02393 if ( ' ' == substr( $t, 0, 1 ) and ( $this->mLastSection === 'pre' || trim( $t ) != '' ) ) { 02394 # pre 02395 if ( $this->mLastSection !== 'pre' ) { 02396 $paragraphStack = false; 02397 $output .= $this->closeParagraph().'<pre>'; 02398 $this->mLastSection = 'pre'; 02399 } 02400 $t = substr( $t, 1 ); 02401 } else { 02402 # paragraph 02403 if ( trim( $t ) === '' ) { 02404 if ( $paragraphStack ) { 02405 $output .= $paragraphStack.'<br />'; 02406 $paragraphStack = false; 02407 $this->mLastSection = 'p'; 02408 } else { 02409 if ( $this->mLastSection !== 'p' ) { 02410 $output .= $this->closeParagraph(); 02411 $this->mLastSection = ''; 02412 $paragraphStack = '<p>'; 02413 } else { 02414 $paragraphStack = '</p><p>'; 02415 } 02416 } 02417 } else { 02418 if ( $paragraphStack ) { 02419 $output .= $paragraphStack; 02420 $paragraphStack = false; 02421 $this->mLastSection = 'p'; 02422 } elseif ( $this->mLastSection !== 'p' ) { 02423 $output .= $this->closeParagraph().'<p>'; 02424 $this->mLastSection = 'p'; 02425 } 02426 } 02427 } 02428 } 02429 wfProfileOut( __METHOD__."-paragraph" ); 02430 } 02431 # somewhere above we forget to get out of pre block (bug 785) 02432 if ( $preCloseMatch && $this->mInPre ) { 02433 $this->mInPre = false; 02434 } 02435 if ( $paragraphStack === false ) { 02436 $output .= $t."\n"; 02437 } 02438 } 02439 while ( $prefixLength ) { 02440 $output .= $this->closeList( $prefix2[$prefixLength-1] ); 02441 --$prefixLength; 02442 } 02443 if ( $this->mLastSection != '' ) { 02444 $output .= '</' . $this->mLastSection . '>'; 02445 $this->mLastSection = ''; 02446 } 02447 02448 wfProfileOut( __METHOD__ ); 02449 return $output; 02450 } 02451 02462 function findColonNoLinks( $str, &$before, &$after ) { 02463 wfProfileIn( __METHOD__ ); 02464 02465 $pos = strpos( $str, ':' ); 02466 if ( $pos === false ) { 02467 # Nothing to find! 02468 wfProfileOut( __METHOD__ ); 02469 return false; 02470 } 02471 02472 $lt = strpos( $str, '<' ); 02473 if ( $lt === false || $lt > $pos ) { 02474 # Easy; no tag nesting to worry about 02475 $before = substr( $str, 0, $pos ); 02476 $after = substr( $str, $pos+1 ); 02477 wfProfileOut( __METHOD__ ); 02478 return $pos; 02479 } 02480 02481 # Ugly state machine to walk through avoiding tags. 02482 $state = self::COLON_STATE_TEXT; 02483 $stack = 0; 02484 $len = strlen( $str ); 02485 for( $i = 0; $i < $len; $i++ ) { 02486 $c = $str[$i]; 02487 02488 switch( $state ) { 02489 # (Using the number is a performance hack for common cases) 02490 case 0: # self::COLON_STATE_TEXT: 02491 switch( $c ) { 02492 case "<": 02493 # Could be either a <start> tag or an </end> tag 02494 $state = self::COLON_STATE_TAGSTART; 02495 break; 02496 case ":": 02497 if ( $stack == 0 ) { 02498 # We found it! 02499 $before = substr( $str, 0, $i ); 02500 $after = substr( $str, $i + 1 ); 02501 wfProfileOut( __METHOD__ ); 02502 return $i; 02503 } 02504 # Embedded in a tag; don't break it. 02505 break; 02506 default: 02507 # Skip ahead looking for something interesting 02508 $colon = strpos( $str, ':', $i ); 02509 if ( $colon === false ) { 02510 # Nothing else interesting 02511 wfProfileOut( __METHOD__ ); 02512 return false; 02513 } 02514 $lt = strpos( $str, '<', $i ); 02515 if ( $stack === 0 ) { 02516 if ( $lt === false || $colon < $lt ) { 02517 # We found it! 02518 $before = substr( $str, 0, $colon ); 02519 $after = substr( $str, $colon + 1 ); 02520 wfProfileOut( __METHOD__ ); 02521 return $i; 02522 } 02523 } 02524 if ( $lt === false ) { 02525 # Nothing else interesting to find; abort! 02526 # We're nested, but there's no close tags left. Abort! 02527 break 2; 02528 } 02529 # Skip ahead to next tag start 02530 $i = $lt; 02531 $state = self::COLON_STATE_TAGSTART; 02532 } 02533 break; 02534 case 1: # self::COLON_STATE_TAG: 02535 # In a <tag> 02536 switch( $c ) { 02537 case ">": 02538 $stack++; 02539 $state = self::COLON_STATE_TEXT; 02540 break; 02541 case "/": 02542 # Slash may be followed by >? 02543 $state = self::COLON_STATE_TAGSLASH; 02544 break; 02545 default: 02546 # ignore 02547 } 02548 break; 02549 case 2: # self::COLON_STATE_TAGSTART: 02550 switch( $c ) { 02551 case "/": 02552 $state = self::COLON_STATE_CLOSETAG; 02553 break; 02554 case "!": 02555 $state = self::COLON_STATE_COMMENT; 02556 break; 02557 case ">": 02558 # Illegal early close? This shouldn't happen D: 02559 $state = self::COLON_STATE_TEXT; 02560 break; 02561 default: 02562 $state = self::COLON_STATE_TAG; 02563 } 02564 break; 02565 case 3: # self::COLON_STATE_CLOSETAG: 02566 # In a </tag> 02567 if ( $c === ">" ) { 02568 $stack--; 02569 if ( $stack < 0 ) { 02570 wfDebug( __METHOD__.": Invalid input; too many close tags\n" ); 02571 wfProfileOut( __METHOD__ ); 02572 return false; 02573 } 02574 $state = self::COLON_STATE_TEXT; 02575 } 02576 break; 02577 case self::COLON_STATE_TAGSLASH: 02578 if ( $c === ">" ) { 02579 # Yes, a self-closed tag <blah/> 02580 $state = self::COLON_STATE_TEXT; 02581 } else { 02582 # Probably we're jumping the gun, and this is an attribute 02583 $state = self::COLON_STATE_TAG; 02584 } 02585 break; 02586 case 5: # self::COLON_STATE_COMMENT: 02587 if ( $c === "-" ) { 02588 $state = self::COLON_STATE_COMMENTDASH; 02589 } 02590 break; 02591 case self::COLON_STATE_COMMENTDASH: 02592 if ( $c === "-" ) { 02593 $state = self::COLON_STATE_COMMENTDASHDASH; 02594 } else { 02595 $state = self::COLON_STATE_COMMENT; 02596 } 02597 break; 02598 case self::COLON_STATE_COMMENTDASHDASH: 02599 if ( $c === ">" ) { 02600 $state = self::COLON_STATE_TEXT; 02601 } else { 02602 $state = self::COLON_STATE_COMMENT; 02603 } 02604 break; 02605 default: 02606 throw new MWException( "State machine error in " . __METHOD__ ); 02607 } 02608 } 02609 if ( $stack > 0 ) { 02610 wfDebug( __METHOD__.": Invalid input; not enough close tags (stack $stack, state $state)\n" ); 02611 wfProfileOut( __METHOD__ ); 02612 return false; 02613 } 02614 wfProfileOut( __METHOD__ ); 02615 return false; 02616 } 02617 02629 function getVariableValue( $index, $frame = false ) { 02630 global $wgContLang, $wgSitename, $wgServer; 02631 global $wgArticlePath, $wgScriptPath, $wgStylePath; 02632 02633 if ( is_null( $this->mTitle ) ) { 02634 // If no title set, bad things are going to happen 02635 // later. Title should always be set since this 02636 // should only be called in the middle of a parse 02637 // operation (but the unit-tests do funky stuff) 02638 throw new MWException( __METHOD__ . ' Should only be ' 02639 . ' called while parsing (no title set)' ); 02640 } 02641 02646 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$this->mVarCache ) ) ) { 02647 if ( isset( $this->mVarCache[$index] ) ) { 02648 return $this->mVarCache[$index]; 02649 } 02650 } 02651 02652 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() ); 02653 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) ); 02654 02655 # Use the time zone 02656 global $wgLocaltimezone; 02657 if ( isset( $wgLocaltimezone ) ) { 02658 $oldtz = date_default_timezone_get(); 02659 date_default_timezone_set( $wgLocaltimezone ); 02660 } 02661 02662 $localTimestamp = date( 'YmdHis', $ts ); 02663 $localMonth = date( 'm', $ts ); 02664 $localMonth1 = date( 'n', $ts ); 02665 $localMonthName = date( 'n', $ts ); 02666 $localDay = date( 'j', $ts ); 02667 $localDay2 = date( 'd', $ts ); 02668 $localDayOfWeek = date( 'w', $ts ); 02669 $localWeek = date( 'W', $ts ); 02670 $localYear = date( 'Y', $ts ); 02671 $localHour = date( 'H', $ts ); 02672 if ( isset( $wgLocaltimezone ) ) { 02673 date_default_timezone_set( $oldtz ); 02674 } 02675 02676 $pageLang = $this->getFunctionLang(); 02677 02678 switch ( $index ) { 02679 case 'currentmonth': 02680 $value = $pageLang->formatNum( gmdate( 'm', $ts ) ); 02681 break; 02682 case 'currentmonth1': 02683 $value = $pageLang->formatNum( gmdate( 'n', $ts ) ); 02684 break; 02685 case 'currentmonthname': 02686 $value = $pageLang->getMonthName( gmdate( 'n', $ts ) ); 02687 break; 02688 case 'currentmonthnamegen': 02689 $value = $pageLang->getMonthNameGen( gmdate( 'n', $ts ) ); 02690 break; 02691 case 'currentmonthabbrev': 02692 $value = $pageLang->getMonthAbbreviation( gmdate( 'n', $ts ) ); 02693 break; 02694 case 'currentday': 02695 $value = $pageLang->formatNum( gmdate( 'j', $ts ) ); 02696 break; 02697 case 'currentday2': 02698 $value = $pageLang->formatNum( gmdate( 'd', $ts ) ); 02699 break; 02700 case 'localmonth': 02701 $value = $pageLang->formatNum( $localMonth ); 02702 break; 02703 case 'localmonth1': 02704 $value = $pageLang->formatNum( $localMonth1 ); 02705 break; 02706 case 'localmonthname': 02707 $value = $pageLang->getMonthName( $localMonthName ); 02708 break; 02709 case 'localmonthnamegen': 02710 $value = $pageLang->getMonthNameGen( $localMonthName ); 02711 break; 02712 case 'localmonthabbrev': 02713 $value = $pageLang->getMonthAbbreviation( $localMonthName ); 02714 break; 02715 case 'localday': 02716 $value = $pageLang->formatNum( $localDay ); 02717 break; 02718 case 'localday2': 02719 $value = $pageLang->formatNum( $localDay2 ); 02720 break; 02721 case 'pagename': 02722 $value = wfEscapeWikiText( $this->mTitle->getText() ); 02723 break; 02724 case 'pagenamee': 02725 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() ); 02726 break; 02727 case 'fullpagename': 02728 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() ); 02729 break; 02730 case 'fullpagenamee': 02731 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() ); 02732 break; 02733 case 'subpagename': 02734 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() ); 02735 break; 02736 case 'subpagenamee': 02737 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() ); 02738 break; 02739 case 'basepagename': 02740 $value = wfEscapeWikiText( $this->mTitle->getBaseText() ); 02741 break; 02742 case 'basepagenamee': 02743 $value = wfEscapeWikiText( wfUrlEncode( str_replace( ' ', '_', $this->mTitle->getBaseText() ) ) ); 02744 break; 02745 case 'talkpagename': 02746 if ( $this->mTitle->canTalk() ) { 02747 $talkPage = $this->mTitle->getTalkPage(); 02748 $value = wfEscapeWikiText( $talkPage->getPrefixedText() ); 02749 } else { 02750 $value = ''; 02751 } 02752 break; 02753 case 'talkpagenamee': 02754 if ( $this->mTitle->canTalk() ) { 02755 $talkPage = $this->mTitle->getTalkPage(); 02756 $value = wfEscapeWikiText( $talkPage->getPrefixedUrl() ); 02757 } else { 02758 $value = ''; 02759 } 02760 break; 02761 case 'subjectpagename': 02762 $subjPage = $this->mTitle->getSubjectPage(); 02763 $value = wfEscapeWikiText( $subjPage->getPrefixedText() ); 02764 break; 02765 case 'subjectpagenamee': 02766 $subjPage = $this->mTitle->getSubjectPage(); 02767 $value = wfEscapeWikiText( $subjPage->getPrefixedUrl() ); 02768 break; 02769 case 'pageid': // requested in bug 23427 02770 $pageid = $this->getTitle()->getArticleId(); 02771 if( $pageid == 0 ) { 02772 # 0 means the page doesn't exist in the database, 02773 # which means the user is previewing a new page. 02774 # The vary-revision flag must be set, because the magic word 02775 # will have a different value once the page is saved. 02776 $this->mOutput->setFlag( 'vary-revision' ); 02777 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" ); 02778 } 02779 $value = $pageid ? $pageid : null; 02780 break; 02781 case 'revisionid': 02782 # Let the edit saving system know we should parse the page 02783 # *after* a revision ID has been assigned. 02784 $this->mOutput->setFlag( 'vary-revision' ); 02785 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision...\n" ); 02786 $value = $this->mRevisionId; 02787 break; 02788 case 'revisionday': 02789 # Let the edit saving system know we should parse the page 02790 # *after* a revision ID has been assigned. This is for null edits. 02791 $this->mOutput->setFlag( 'vary-revision' ); 02792 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" ); 02793 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) ); 02794 break; 02795 case 'revisionday2': 02796 # Let the edit saving system know we should parse the page 02797 # *after* a revision ID has been assigned. This is for null edits. 02798 $this->mOutput->setFlag( 'vary-revision' ); 02799 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" ); 02800 $value = substr( $this->getRevisionTimestamp(), 6, 2 ); 02801 break; 02802 case 'revisionmonth': 02803 # Let the edit saving system know we should parse the page 02804 # *after* a revision ID has been assigned. This is for null edits. 02805 $this->mOutput->setFlag( 'vary-revision' ); 02806 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" ); 02807 $value = substr( $this->getRevisionTimestamp(), 4, 2 ); 02808 break; 02809 case 'revisionmonth1': 02810 # Let the edit saving system know we should parse the page 02811 # *after* a revision ID has been assigned. This is for null edits. 02812 $this->mOutput->setFlag( 'vary-revision' ); 02813 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" ); 02814 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) ); 02815 break; 02816 case 'revisionyear': 02817 # Let the edit saving system know we should parse the page 02818 # *after* a revision ID has been assigned. This is for null edits. 02819 $this->mOutput->setFlag( 'vary-revision' ); 02820 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" ); 02821 $value = substr( $this->getRevisionTimestamp(), 0, 4 ); 02822 break; 02823 case 'revisiontimestamp': 02824 # Let the edit saving system know we should parse the page 02825 # *after* a revision ID has been assigned. This is for null edits. 02826 $this->mOutput->setFlag( 'vary-revision' ); 02827 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" ); 02828 $value = $this->getRevisionTimestamp(); 02829 break; 02830 case 'revisionuser': 02831 # Let the edit saving system know we should parse the page 02832 # *after* a revision ID has been assigned. This is for null edits. 02833 $this->mOutput->setFlag( 'vary-revision' ); 02834 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-revision...\n" ); 02835 $value = $this->getRevisionUser(); 02836 break; 02837 case 'namespace': 02838 $value = str_replace( '_',' ',$wgContLang->getNsText( $this->mTitle->getNamespace() ) ); 02839 break; 02840 case 'namespacee': 02841 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) ); 02842 break; 02843 case 'namespacenumber': 02844 $value = $this->mTitle->getNamespace(); 02845 break; 02846 case 'talkspace': 02847 $value = $this->mTitle->canTalk() ? str_replace( '_',' ',$this->mTitle->getTalkNsText() ) : ''; 02848 break; 02849 case 'talkspacee': 02850 $value = $this->mTitle->canTalk() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : ''; 02851 break; 02852 case 'subjectspace': 02853 $value = $this->mTitle->getSubjectNsText(); 02854 break; 02855 case 'subjectspacee': 02856 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) ); 02857 break; 02858 case 'currentdayname': 02859 $value = $pageLang->getWeekdayName( gmdate( 'w', $ts ) + 1 ); 02860 break; 02861 case 'currentyear': 02862 $value = $pageLang->formatNum( gmdate( 'Y', $ts ), true ); 02863 break; 02864 case 'currenttime': 02865 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false ); 02866 break; 02867 case 'currenthour': 02868 $value = $pageLang->formatNum( gmdate( 'H', $ts ), true ); 02869 break; 02870 case 'currentweek': 02871 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to 02872 # int to remove the padding 02873 $value = $pageLang->formatNum( (int)gmdate( 'W', $ts ) ); 02874 break; 02875 case 'currentdow': 02876 $value = $pageLang->formatNum( gmdate( 'w', $ts ) ); 02877 break; 02878 case 'localdayname': 02879 $value = $pageLang->getWeekdayName( $localDayOfWeek + 1 ); 02880 break; 02881 case 'localyear': 02882 $value = $pageLang->formatNum( $localYear, true ); 02883 break; 02884 case 'localtime': 02885 $value = $pageLang->time( $localTimestamp, false, false ); 02886 break; 02887 case 'localhour': 02888 $value = $pageLang->formatNum( $localHour, true ); 02889 break; 02890 case 'localweek': 02891 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to 02892 # int to remove the padding 02893 $value = $pageLang->formatNum( (int)$localWeek ); 02894 break; 02895 case 'localdow': 02896 $value = $pageLang->formatNum( $localDayOfWeek ); 02897 break; 02898 case 'numberofarticles': 02899 $value = $pageLang->formatNum( SiteStats::articles() ); 02900 break; 02901 case 'numberoffiles': 02902 $value = $pageLang->formatNum( SiteStats::images() ); 02903 break; 02904 case 'numberofusers': 02905 $value = $pageLang->formatNum( SiteStats::users() ); 02906 break; 02907 case 'numberofactiveusers': 02908 $value = $pageLang->formatNum( SiteStats::activeUsers() ); 02909 break; 02910 case 'numberofpages': 02911 $value = $pageLang->formatNum( SiteStats::pages() ); 02912 break; 02913 case 'numberofadmins': 02914 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) ); 02915 break; 02916 case 'numberofedits': 02917 $value = $pageLang->formatNum( SiteStats::edits() ); 02918 break; 02919 case 'numberofviews': 02920 global $wgDisableCounters; 02921 $value = !$wgDisableCounters ? $pageLang->formatNum( SiteStats::views() ) : ''; 02922 break; 02923 case 'currenttimestamp': 02924 $value = wfTimestamp( TS_MW, $ts ); 02925 break; 02926 case 'localtimestamp': 02927 $value = $localTimestamp; 02928 break; 02929 case 'currentversion': 02930 $value = SpecialVersion::getVersion(); 02931 break; 02932 case 'articlepath': 02933 return $wgArticlePath; 02934 case 'sitename': 02935 return $wgSitename; 02936 case 'server': 02937 return $wgServer; 02938 case 'servername': 02939 $serverParts = wfParseUrl( $wgServer ); 02940 return $serverParts && isset( $serverParts['host'] ) ? $serverParts['host'] : $wgServer; 02941 case 'scriptpath': 02942 return $wgScriptPath; 02943 case 'stylepath': 02944 return $wgStylePath; 02945 case 'directionmark': 02946 return $pageLang->getDirMark(); 02947 case 'contentlanguage': 02948 global $wgLanguageCode; 02949 return $wgLanguageCode; 02950 default: 02951 $ret = null; 02952 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$this->mVarCache, &$index, &$ret, &$frame ) ) ) { 02953 return $ret; 02954 } else { 02955 return null; 02956 } 02957 } 02958 02959 if ( $index ) { 02960 $this->mVarCache[$index] = $value; 02961 } 02962 02963 return $value; 02964 } 02965 02971 function initialiseVariables() { 02972 wfProfileIn( __METHOD__ ); 02973 $variableIDs = MagicWord::getVariableIDs(); 02974 $substIDs = MagicWord::getSubstIDs(); 02975 02976 $this->mVariables = new MagicWordArray( $variableIDs ); 02977 $this->mSubstWords = new MagicWordArray( $substIDs ); 02978 wfProfileOut( __METHOD__ ); 02979 } 02980 03005 function preprocessToDom( $text, $flags = 0 ) { 03006 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags ); 03007 return $dom; 03008 } 03009 03017 public static function splitWhitespace( $s ) { 03018 $ltrimmed = ltrim( $s ); 03019 $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) ); 03020 $trimmed = rtrim( $ltrimmed ); 03021 $diff = strlen( $ltrimmed ) - strlen( $trimmed ); 03022 if ( $diff > 0 ) { 03023 $w2 = substr( $ltrimmed, -$diff ); 03024 } else { 03025 $w2 = ''; 03026 } 03027 return array( $w1, $trimmed, $w2 ); 03028 } 03029 03049 function replaceVariables( $text, $frame = false, $argsOnly = false ) { 03050 # Is there any text? Also, Prevent too big inclusions! 03051 if ( strlen( $text ) < 1 || strlen( $text ) > $this->mOptions->getMaxIncludeSize() ) { 03052 return $text; 03053 } 03054 wfProfileIn( __METHOD__ ); 03055 03056 if ( $frame === false ) { 03057 $frame = $this->getPreprocessor()->newFrame(); 03058 } elseif ( !( $frame instanceof PPFrame ) ) { 03059 wfDebug( __METHOD__." called using plain parameters instead of a PPFrame instance. Creating custom frame.\n" ); 03060 $frame = $this->getPreprocessor()->newCustomFrame( $frame ); 03061 } 03062 03063 $dom = $this->preprocessToDom( $text ); 03064 $flags = $argsOnly ? PPFrame::NO_TEMPLATES : 0; 03065 $text = $frame->expand( $dom, $flags ); 03066 03067 wfProfileOut( __METHOD__ ); 03068 return $text; 03069 } 03070 03078 static function createAssocArgs( $args ) { 03079 $assocArgs = array(); 03080 $index = 1; 03081 foreach ( $args as $arg ) { 03082 $eqpos = strpos( $arg, '=' ); 03083 if ( $eqpos === false ) { 03084 $assocArgs[$index++] = $arg; 03085 } else { 03086 $name = trim( substr( $arg, 0, $eqpos ) ); 03087 $value = trim( substr( $arg, $eqpos+1 ) ); 03088 if ( $value === false ) { 03089 $value = ''; 03090 } 03091 if ( $name !== false ) { 03092 $assocArgs[$name] = $value; 03093 } 03094 } 03095 } 03096 03097 return $assocArgs; 03098 } 03099 03118 function limitationWarn( $limitationType, $current = '', $max = '' ) { 03119 # does no harm if $current and $max are present but are unnecessary for the message 03120 $warning = wfMessage( "$limitationType-warning" )->numParams( $current, $max ) 03121 ->inContentLanguage()->escaped(); 03122 $this->mOutput->addWarning( $warning ); 03123 $this->addTrackingCategory( "$limitationType-category" ); 03124 } 03125 03139 function braceSubstitution( $piece, $frame ) { 03140 global $wgContLang; 03141 wfProfileIn( __METHOD__ ); 03142 wfProfileIn( __METHOD__.'-setup' ); 03143 03144 # Flags 03145 $found = false; # $text has been filled 03146 $nowiki = false; # wiki markup in $text should be escaped 03147 $isHTML = false; # $text is HTML, armour it against wikitext transformation 03148 $forceRawInterwiki = false; # Force interwiki transclusion to be done in raw mode not rendered 03149 $isChildObj = false; # $text is a DOM node needing expansion in a child frame 03150 $isLocalObj = false; # $text is a DOM node needing expansion in the current frame 03151 03152 # Title object, where $text came from 03153 $title = false; 03154 03155 # $part1 is the bit before the first |, and must contain only title characters. 03156 # Various prefixes will be stripped from it later. 03157 $titleWithSpaces = $frame->expand( $piece['title'] ); 03158 $part1 = trim( $titleWithSpaces ); 03159 $titleText = false; 03160 03161 # Original title text preserved for various purposes 03162 $originalTitle = $part1; 03163 03164 # $args is a list of argument nodes, starting from index 0, not including $part1 03165 # @todo FIXME: If piece['parts'] is null then the call to getLength() below won't work b/c this $args isn't an object 03166 $args = ( null == $piece['parts'] ) ? array() : $piece['parts']; 03167 wfProfileOut( __METHOD__.'-setup' ); 03168 03169 $titleProfileIn = null; // profile templates 03170 03171 # SUBST 03172 wfProfileIn( __METHOD__.'-modifiers' ); 03173 if ( !$found ) { 03174 03175 $substMatch = $this->mSubstWords->matchStartAndRemove( $part1 ); 03176 03177 # Possibilities for substMatch: "subst", "safesubst" or FALSE 03178 # Decide whether to expand template or keep wikitext as-is. 03179 if ( $this->ot['wiki'] ) { 03180 if ( $substMatch === false ) { 03181 $literal = true; # literal when in PST with no prefix 03182 } else { 03183 $literal = false; # expand when in PST with subst: or safesubst: 03184 } 03185 } else { 03186 if ( $substMatch == 'subst' ) { 03187 $literal = true; # literal when not in PST with plain subst: 03188 } else { 03189 $literal = false; # expand when not in PST with safesubst: or no prefix 03190 } 03191 } 03192 if ( $literal ) { 03193 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args ); 03194 $isLocalObj = true; 03195 $found = true; 03196 } 03197 } 03198 03199 # Variables 03200 if ( !$found && $args->getLength() == 0 ) { 03201 $id = $this->mVariables->matchStartToEnd( $part1 ); 03202 if ( $id !== false ) { 03203 $text = $this->getVariableValue( $id, $frame ); 03204 if ( MagicWord::getCacheTTL( $id ) > -1 ) { 03205 $this->mOutput->updateCacheExpiry( MagicWord::getCacheTTL( $id ) ); 03206 } 03207 $found = true; 03208 } 03209 } 03210 03211 # MSG, MSGNW and RAW 03212 if ( !$found ) { 03213 # Check for MSGNW: 03214 $mwMsgnw = MagicWord::get( 'msgnw' ); 03215 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) { 03216 $nowiki = true; 03217 } else { 03218 # Remove obsolete MSG: 03219 $mwMsg = MagicWord::get( 'msg' ); 03220 $mwMsg->matchStartAndRemove( $part1 ); 03221 } 03222 03223 # Check for RAW: 03224 $mwRaw = MagicWord::get( 'raw' ); 03225 if ( $mwRaw->matchStartAndRemove( $part1 ) ) { 03226 $forceRawInterwiki = true; 03227 } 03228 } 03229 wfProfileOut( __METHOD__.'-modifiers' ); 03230 03231 # Parser functions 03232 if ( !$found ) { 03233 wfProfileIn( __METHOD__ . '-pfunc' ); 03234 03235 $colonPos = strpos( $part1, ':' ); 03236 if ( $colonPos !== false ) { 03237 # Case sensitive functions 03238 $function = substr( $part1, 0, $colonPos ); 03239 if ( isset( $this->mFunctionSynonyms[1][$function] ) ) { 03240 $function = $this->mFunctionSynonyms[1][$function]; 03241 } else { 03242 # Case insensitive functions 03243 $function = $wgContLang->lc( $function ); 03244 if ( isset( $this->mFunctionSynonyms[0][$function] ) ) { 03245 $function = $this->mFunctionSynonyms[0][$function]; 03246 } else { 03247 $function = false; 03248 } 03249 } 03250 if ( $function ) { 03251 wfProfileIn( __METHOD__ . '-pfunc-' . $function ); 03252 list( $callback, $flags ) = $this->mFunctionHooks[$function]; 03253 $initialArgs = array( &$this ); 03254 $funcArgs = array( trim( substr( $part1, $colonPos + 1 ) ) ); 03255 if ( $flags & SFH_OBJECT_ARGS ) { 03256 # Add a frame parameter, and pass the arguments as an array 03257 $allArgs = $initialArgs; 03258 $allArgs[] = $frame; 03259 for ( $i = 0; $i < $args->getLength(); $i++ ) { 03260 $funcArgs[] = $args->item( $i ); 03261 } 03262 $allArgs[] = $funcArgs; 03263 } else { 03264 # Convert arguments to plain text 03265 for ( $i = 0; $i < $args->getLength(); $i++ ) { 03266 $funcArgs[] = trim( $frame->expand( $args->item( $i ) ) ); 03267 } 03268 $allArgs = array_merge( $initialArgs, $funcArgs ); 03269 } 03270 03271 # Workaround for PHP bug 35229 and similar 03272 if ( !is_callable( $callback ) ) { 03273 wfProfileOut( __METHOD__ . '-pfunc-' . $function ); 03274 wfProfileOut( __METHOD__ . '-pfunc' ); 03275 wfProfileOut( __METHOD__ ); 03276 throw new MWException( "Tag hook for $function is not callable\n" ); 03277 } 03278 $result = call_user_func_array( $callback, $allArgs ); 03279 $found = true; 03280 $noparse = true; 03281 $preprocessFlags = 0; 03282 03283 if ( is_array( $result ) ) { 03284 if ( isset( $result[0] ) ) { 03285 $text = $result[0]; 03286 unset( $result[0] ); 03287 } 03288 03289 # Extract flags into the local scope 03290 # This allows callers to set flags such as nowiki, found, etc. 03291 extract( $result ); 03292 } else { 03293 $text = $result; 03294 } 03295 if ( !$noparse ) { 03296 $text = $this->preprocessToDom( $text, $preprocessFlags ); 03297 $isChildObj = true; 03298 } 03299 wfProfileOut( __METHOD__ . '-pfunc-' . $function ); 03300 } 03301 } 03302 wfProfileOut( __METHOD__ . '-pfunc' ); 03303 } 03304 03305 # Finish mangling title and then check for loops. 03306 # Set $title to a Title object and $titleText to the PDBK 03307 if ( !$found ) { 03308 $ns = NS_TEMPLATE; 03309 # Split the title into page and subpage 03310 $subpage = ''; 03311 $part1 = $this->maybeDoSubpageLink( $part1, $subpage ); 03312 if ( $subpage !== '' ) { 03313 $ns = $this->mTitle->getNamespace(); 03314 } 03315 $title = Title::newFromText( $part1, $ns ); 03316 if ( $title ) { 03317 $titleText = $title->getPrefixedText(); 03318 # Check for language variants if the template is not found 03319 if ( $this->getConverterLanguage()->hasVariants() && $title->getArticleID() == 0 ) { 03320 $this->getConverterLanguage()->findVariantLink( $part1, $title, true ); 03321 } 03322 # Do recursion depth check 03323 $limit = $this->mOptions->getMaxTemplateDepth(); 03324 if ( $frame->depth >= $limit ) { 03325 $found = true; 03326 $text = '<span class="error">' 03327 . wfMessage( 'parser-template-recursion-depth-warning' ) 03328 ->numParams( $limit )->inContentLanguage()->text() 03329 . '</span>'; 03330 } 03331 } 03332 } 03333 03334 # Load from database 03335 if ( !$found && $title ) { 03336 if ( !Profiler::instance()->isPersistent() ) { 03337 # Too many unique items can kill profiling DBs/collectors 03338 $titleProfileIn = __METHOD__ . "-title-" . $title->getDBKey(); 03339 wfProfileIn( $titleProfileIn ); // template in 03340 } 03341 wfProfileIn( __METHOD__ . '-loadtpl' ); 03342 if ( !$title->isExternal() ) { 03343 if ( $title->isSpecialPage() 03344 && $this->mOptions->getAllowSpecialInclusion() 03345 && $this->ot['html'] ) 03346 { 03347 // Pass the template arguments as URL parameters. 03348 // "uselang" will have no effect since the Language object 03349 // is forced to the one defined in ParserOptions. 03350 $pageArgs = array(); 03351 for ( $i = 0; $i < $args->getLength(); $i++ ) { 03352 $bits = $args->item( $i )->splitArg(); 03353 if ( strval( $bits['index'] ) === '' ) { 03354 $name = trim( $frame->expand( $bits['name'], PPFrame::STRIP_COMMENTS ) ); 03355 $value = trim( $frame->expand( $bits['value'] ) ); 03356 $pageArgs[$name] = $value; 03357 } 03358 } 03359 03360 // Create a new context to execute the special page 03361 $context = new RequestContext; 03362 $context->setTitle( $title ); 03363 $context->setRequest( new FauxRequest( $pageArgs ) ); 03364 $context->setUser( $this->getUser() ); 03365 $context->setLanguage( $this->mOptions->getUserLangObj() ); 03366 $ret = SpecialPageFactory::capturePath( $title, $context ); 03367 if ( $ret ) { 03368 $text = $context->getOutput()->getHTML(); 03369 $this->mOutput->addOutputPageMetadata( $context->getOutput() ); 03370 $found = true; 03371 $isHTML = true; 03372 $this->disableCache(); 03373 } 03374 } elseif ( MWNamespace::isNonincludable( $title->getNamespace() ) ) { 03375 $found = false; # access denied 03376 wfDebug( __METHOD__.": template inclusion denied for " . $title->getPrefixedDBkey() ); 03377 } else { 03378 list( $text, $title ) = $this->getTemplateDom( $title ); 03379 if ( $text !== false ) { 03380 $found = true; 03381 $isChildObj = true; 03382 } 03383 } 03384 03385 # If the title is valid but undisplayable, make a link to it 03386 if ( !$found && ( $this->ot['html'] || $this->ot['pre'] ) ) { 03387 $text = "[[:$titleText]]"; 03388 $found = true; 03389 } 03390 } elseif ( $title->isTrans() ) { 03391 # Interwiki transclusion 03392 if ( $this->ot['html'] && !$forceRawInterwiki ) { 03393 $text = $this->interwikiTransclude( $title, 'render' ); 03394 $isHTML = true; 03395 } else { 03396 $text = $this->interwikiTransclude( $title, 'raw' ); 03397 # Preprocess it like a template 03398 $text = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION ); 03399 $isChildObj = true; 03400 } 03401 $found = true; 03402 } 03403 03404 # Do infinite loop check 03405 # This has to be done after redirect resolution to avoid infinite loops via redirects 03406 if ( !$frame->loopCheck( $title ) ) { 03407 $found = true; 03408 $text = '<span class="error">' 03409 . wfMessage( 'parser-template-loop-warning', $titleText )->inContentLanguage()->text() 03410 . '</span>'; 03411 wfDebug( __METHOD__.": template loop broken at '$titleText'\n" ); 03412 } 03413 wfProfileOut( __METHOD__ . '-loadtpl' ); 03414 } 03415 03416 # If we haven't found text to substitute by now, we're done 03417 # Recover the source wikitext and return it 03418 if ( !$found ) { 03419 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args ); 03420 if ( $titleProfileIn ) { 03421 wfProfileOut( $titleProfileIn ); // template out 03422 } 03423 wfProfileOut( __METHOD__ ); 03424 return array( 'object' => $text ); 03425 } 03426 03427 # Expand DOM-style return values in a child frame 03428 if ( $isChildObj ) { 03429 # Clean up argument array 03430 $newFrame = $frame->newChild( $args, $title ); 03431 03432 if ( $nowiki ) { 03433 $text = $newFrame->expand( $text, PPFrame::RECOVER_ORIG ); 03434 } elseif ( $titleText !== false && $newFrame->isEmpty() ) { 03435 # Expansion is eligible for the empty-frame cache 03436 if ( isset( $this->mTplExpandCache[$titleText] ) ) { 03437 $text = $this->mTplExpandCache[$titleText]; 03438 } else { 03439 $text = $newFrame->expand( $text ); 03440 $this->mTplExpandCache[$titleText] = $text; 03441 } 03442 } else { 03443 # Uncached expansion 03444 $text = $newFrame->expand( $text ); 03445 } 03446 } 03447 if ( $isLocalObj && $nowiki ) { 03448 $text = $frame->expand( $text, PPFrame::RECOVER_ORIG ); 03449 $isLocalObj = false; 03450 } 03451 03452 if ( $titleProfileIn ) { 03453 wfProfileOut( $titleProfileIn ); // template out 03454 } 03455 03456 # Replace raw HTML by a placeholder 03457 if ( $isHTML ) { 03458 $text = $this->insertStripItem( $text ); 03459 } elseif ( $nowiki && ( $this->ot['html'] || $this->ot['pre'] ) ) { 03460 # Escape nowiki-style return values 03461 $text = wfEscapeWikiText( $text ); 03462 } elseif ( is_string( $text ) 03463 && !$piece['lineStart'] 03464 && preg_match( '/^(?:{\\||:|;|#|\*)/', $text ) ) 03465 { 03466 # Bug 529: if the template begins with a table or block-level 03467 # element, it should be treated as beginning a new line. 03468 # This behaviour is somewhat controversial. 03469 $text = "\n" . $text; 03470 } 03471 03472 if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) { 03473 # Error, oversize inclusion 03474 if ( $titleText !== false ) { 03475 # Make a working, properly escaped link if possible (bug 23588) 03476 $text = "[[:$titleText]]"; 03477 } else { 03478 # This will probably not be a working link, but at least it may 03479 # provide some hint of where the problem is 03480 preg_replace( '/^:/', '', $originalTitle ); 03481 $text = "[[:$originalTitle]]"; 03482 } 03483 $text .= $this->insertStripItem( '<!-- WARNING: template omitted, post-expand include size too large -->' ); 03484 $this->limitationWarn( 'post-expand-template-inclusion' ); 03485 } 03486 03487 if ( $isLocalObj ) { 03488 $ret = array( 'object' => $text ); 03489 } else { 03490 $ret = array( 'text' => $text ); 03491 } 03492 03493 wfProfileOut( __METHOD__ ); 03494 return $ret; 03495 } 03496 03505 function getTemplateDom( $title ) { 03506 $cacheTitle = $title; 03507 $titleText = $title->getPrefixedDBkey(); 03508 03509 if ( isset( $this->mTplRedirCache[$titleText] ) ) { 03510 list( $ns, $dbk ) = $this->mTplRedirCache[$titleText]; 03511 $title = Title::makeTitle( $ns, $dbk ); 03512 $titleText = $title->getPrefixedDBkey(); 03513 } 03514 if ( isset( $this->mTplDomCache[$titleText] ) ) { 03515 return array( $this->mTplDomCache[$titleText], $title ); 03516 } 03517 03518 # Cache miss, go to the database 03519 list( $text, $title ) = $this->fetchTemplateAndTitle( $title ); 03520 03521 if ( $text === false ) { 03522 $this->mTplDomCache[$titleText] = false; 03523 return array( false, $title ); 03524 } 03525 03526 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION ); 03527 $this->mTplDomCache[ $titleText ] = $dom; 03528 03529 if ( !$title->equals( $cacheTitle ) ) { 03530 $this->mTplRedirCache[$cacheTitle->getPrefixedDBkey()] = 03531 array( $title->getNamespace(), $cdb = $title->getDBkey() ); 03532 } 03533 03534 return array( $dom, $title ); 03535 } 03536 03542 function fetchTemplateAndTitle( $title ) { 03543 $templateCb = $this->mOptions->getTemplateCallback(); # Defaults to Parser::statelessFetchTemplate() 03544 $stuff = call_user_func( $templateCb, $title, $this ); 03545 $text = $stuff['text']; 03546 $finalTitle = isset( $stuff['finalTitle'] ) ? $stuff['finalTitle'] : $title; 03547 if ( isset( $stuff['deps'] ) ) { 03548 foreach ( $stuff['deps'] as $dep ) { 03549 $this->mOutput->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] ); 03550 } 03551 } 03552 return array( $text, $finalTitle ); 03553 } 03554 03560 function fetchTemplate( $title ) { 03561 $rv = $this->fetchTemplateAndTitle( $title ); 03562 return $rv[0]; 03563 } 03564 03574 static function statelessFetchTemplate( $title, $parser = false ) { 03575 $text = $skip = false; 03576 $finalTitle = $title; 03577 $deps = array(); 03578 03579 # Loop to fetch the article, with up to 1 redirect 03580 for ( $i = 0; $i < 2 && is_object( $title ); $i++ ) { 03581 # Give extensions a chance to select the revision instead 03582 $id = false; # Assume current 03583 wfRunHooks( 'BeforeParserFetchTemplateAndtitle', 03584 array( $parser, $title, &$skip, &$id ) ); 03585 03586 if ( $skip ) { 03587 $text = false; 03588 $deps[] = array( 03589 'title' => $title, 03590 'page_id' => $title->getArticleID(), 03591 'rev_id' => null 03592 ); 03593 break; 03594 } 03595 # Get the revision 03596 $rev = $id 03597 ? Revision::newFromId( $id ) 03598 : Revision::newFromTitle( $title, false, Revision::READ_NORMAL ); 03599 $rev_id = $rev ? $rev->getId() : 0; 03600 # If there is no current revision, there is no page 03601 if ( $id === false && !$rev ) { 03602 $linkCache = LinkCache::singleton(); 03603 $linkCache->addBadLinkObj( $title ); 03604 } 03605 03606 $deps[] = array( 03607 'title' => $title, 03608 'page_id' => $title->getArticleID(), 03609 'rev_id' => $rev_id ); 03610 if ( $rev && !$title->equals( $rev->getTitle() ) ) { 03611 # We fetched a rev from a different title; register it too... 03612 $deps[] = array( 03613 'title' => $rev->getTitle(), 03614 'page_id' => $rev->getPage(), 03615 'rev_id' => $rev_id ); 03616 } 03617 03618 if ( $rev ) { 03619 $content = $rev->getContent(); 03620 $text = $content ? $content->getWikitextForTransclusion() : null; 03621 03622 if ( $text === false || $text === null ) { 03623 $text = false; 03624 break; 03625 } 03626 } elseif ( $title->getNamespace() == NS_MEDIAWIKI ) { 03627 global $wgContLang; 03628 $message = wfMessage( $wgContLang->lcfirst( $title->getText() ) )->inContentLanguage(); 03629 if ( !$message->exists() ) { 03630 $text = false; 03631 break; 03632 } 03633 $content = $message->content(); 03634 $text = $message->plain(); 03635 } else { 03636 break; 03637 } 03638 if ( !$content ) { 03639 break; 03640 } 03641 # Redirect? 03642 $finalTitle = $title; 03643 $title = $content->getRedirectTarget(); 03644 } 03645 return array( 03646 'text' => $text, 03647 'finalTitle' => $finalTitle, 03648 'deps' => $deps ); 03649 } 03650 03658 function fetchFile( $title, $options = array() ) { 03659 $res = $this->fetchFileAndTitle( $title, $options ); 03660 return $res[0]; 03661 } 03662 03670 function fetchFileAndTitle( $title, $options = array() ) { 03671 if ( isset( $options['broken'] ) ) { 03672 $file = false; // broken thumbnail forced by hook 03673 } elseif ( isset( $options['sha1'] ) ) { // get by (sha1,timestamp) 03674 $file = RepoGroup::singleton()->findFileFromKey( $options['sha1'], $options ); 03675 } else { // get by (name,timestamp) 03676 $file = wfFindFile( $title, $options ); 03677 } 03678 $time = $file ? $file->getTimestamp() : false; 03679 $sha1 = $file ? $file->getSha1() : false; 03680 # Register the file as a dependency... 03681 $this->mOutput->addImage( $title->getDBkey(), $time, $sha1 ); 03682 if ( $file && !$title->equals( $file->getTitle() ) ) { 03683 # Update fetched file title 03684 $title = $file->getTitle(); 03685 if ( is_null( $file->getRedirectedTitle() ) ) { 03686 # This file was not a redirect, but the title does not match. 03687 # Register under the new name because otherwise the link will 03688 # get lost. 03689 $this->mOutput->addImage( $title->getDBkey(), $time, $sha1 ); 03690 } 03691 } 03692 return array( $file, $title ); 03693 } 03694 03703 function interwikiTransclude( $title, $action ) { 03704 global $wgEnableScaryTranscluding; 03705 03706 if ( !$wgEnableScaryTranscluding ) { 03707 return wfMessage('scarytranscludedisabled')->inContentLanguage()->text(); 03708 } 03709 03710 $url = $title->getFullUrl( "action=$action" ); 03711 03712 if ( strlen( $url ) > 255 ) { 03713 return wfMessage( 'scarytranscludetoolong' )->inContentLanguage()->text(); 03714 } 03715 return $this->fetchScaryTemplateMaybeFromCache( $url ); 03716 } 03717 03722 function fetchScaryTemplateMaybeFromCache( $url ) { 03723 global $wgTranscludeCacheExpiry; 03724 $dbr = wfGetDB( DB_SLAVE ); 03725 $tsCond = $dbr->timestamp( time() - $wgTranscludeCacheExpiry ); 03726 $obj = $dbr->selectRow( 'transcache', array('tc_time', 'tc_contents' ), 03727 array( 'tc_url' => $url, "tc_time >= " . $dbr->addQuotes( $tsCond ) ) ); 03728 if ( $obj ) { 03729 return $obj->tc_contents; 03730 } 03731 03732 $req = MWHttpRequest::factory( $url ); 03733 $status = $req->execute(); // Status object 03734 if ( $status->isOK() ) { 03735 $text = $req->getContent(); 03736 } elseif ( $req->getStatus() != 200 ) { // Though we failed to fetch the content, this status is useless. 03737 return wfMessage( 'scarytranscludefailed-httpstatus', $url, $req->getStatus() /* HTTP status */ )->inContentLanguage()->text(); 03738 } else { 03739 return wfMessage( 'scarytranscludefailed', $url )->inContentLanguage()->text(); 03740 } 03741 03742 $dbw = wfGetDB( DB_MASTER ); 03743 $dbw->replace( 'transcache', array('tc_url'), array( 03744 'tc_url' => $url, 03745 'tc_time' => $dbw->timestamp( time() ), 03746 'tc_contents' => $text) 03747 ); 03748 return $text; 03749 } 03750 03760 function argSubstitution( $piece, $frame ) { 03761 wfProfileIn( __METHOD__ ); 03762 03763 $error = false; 03764 $parts = $piece['parts']; 03765 $nameWithSpaces = $frame->expand( $piece['title'] ); 03766 $argName = trim( $nameWithSpaces ); 03767 $object = false; 03768 $text = $frame->getArgument( $argName ); 03769 if ( $text === false && $parts->getLength() > 0 03770 && ( 03771 $this->ot['html'] 03772 || $this->ot['pre'] 03773 || ( $this->ot['wiki'] && $frame->isTemplate() ) 03774 ) 03775 ) { 03776 # No match in frame, use the supplied default 03777 $object = $parts->item( 0 )->getChildren(); 03778 } 03779 if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) { 03780 $error = '<!-- WARNING: argument omitted, expansion size too large -->'; 03781 $this->limitationWarn( 'post-expand-template-argument' ); 03782 } 03783 03784 if ( $text === false && $object === false ) { 03785 # No match anywhere 03786 $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts ); 03787 } 03788 if ( $error !== false ) { 03789 $text .= $error; 03790 } 03791 if ( $object !== false ) { 03792 $ret = array( 'object' => $object ); 03793 } else { 03794 $ret = array( 'text' => $text ); 03795 } 03796 03797 wfProfileOut( __METHOD__ ); 03798 return $ret; 03799 } 03800 03816 function extensionSubstitution( $params, $frame ) { 03817 $name = $frame->expand( $params['name'] ); 03818 $attrText = !isset( $params['attr'] ) ? null : $frame->expand( $params['attr'] ); 03819 $content = !isset( $params['inner'] ) ? null : $frame->expand( $params['inner'] ); 03820 $marker = "{$this->mUniqPrefix}-$name-" . sprintf( '%08X', $this->mMarkerIndex++ ) . self::MARKER_SUFFIX; 03821 03822 $isFunctionTag = isset( $this->mFunctionTagHooks[strtolower($name)] ) && 03823 ( $this->ot['html'] || $this->ot['pre'] ); 03824 if ( $isFunctionTag ) { 03825 $markerType = 'none'; 03826 } else { 03827 $markerType = 'general'; 03828 } 03829 if ( $this->ot['html'] || $isFunctionTag ) { 03830 $name = strtolower( $name ); 03831 $attributes = Sanitizer::decodeTagAttributes( $attrText ); 03832 if ( isset( $params['attributes'] ) ) { 03833 $attributes = $attributes + $params['attributes']; 03834 } 03835 03836 if ( isset( $this->mTagHooks[$name] ) ) { 03837 # Workaround for PHP bug 35229 and similar 03838 if ( !is_callable( $this->mTagHooks[$name] ) ) { 03839 throw new MWException( "Tag hook for $name is not callable\n" ); 03840 } 03841 $output = call_user_func_array( $this->mTagHooks[$name], 03842 array( $content, $attributes, $this, $frame ) ); 03843 } elseif ( isset( $this->mFunctionTagHooks[$name] ) ) { 03844 list( $callback, $flags ) = $this->mFunctionTagHooks[$name]; 03845 if ( !is_callable( $callback ) ) { 03846 throw new MWException( "Tag hook for $name is not callable\n" ); 03847 } 03848 03849 $output = call_user_func_array( $callback, array( &$this, $frame, $content, $attributes ) ); 03850 } else { 03851 $output = '<span class="error">Invalid tag extension name: ' . 03852 htmlspecialchars( $name ) . '</span>'; 03853 } 03854 03855 if ( is_array( $output ) ) { 03856 # Extract flags to local scope (to override $markerType) 03857 $flags = $output; 03858 $output = $flags[0]; 03859 unset( $flags[0] ); 03860 extract( $flags ); 03861 } 03862 } else { 03863 if ( is_null( $attrText ) ) { 03864 $attrText = ''; 03865 } 03866 if ( isset( $params['attributes'] ) ) { 03867 foreach ( $params['attributes'] as $attrName => $attrValue ) { 03868 $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' . 03869 htmlspecialchars( $attrValue ) . '"'; 03870 } 03871 } 03872 if ( $content === null ) { 03873 $output = "<$name$attrText/>"; 03874 } else { 03875 $close = is_null( $params['close'] ) ? '' : $frame->expand( $params['close'] ); 03876 $output = "<$name$attrText>$content$close"; 03877 } 03878 } 03879 03880 if ( $markerType === 'none' ) { 03881 return $output; 03882 } elseif ( $markerType === 'nowiki' ) { 03883 $this->mStripState->addNoWiki( $marker, $output ); 03884 } elseif ( $markerType === 'general' ) { 03885 $this->mStripState->addGeneral( $marker, $output ); 03886 } else { 03887 throw new MWException( __METHOD__.': invalid marker type' ); 03888 } 03889 return $marker; 03890 } 03891 03899 function incrementIncludeSize( $type, $size ) { 03900 if ( $this->mIncludeSizes[$type] + $size > $this->mOptions->getMaxIncludeSize() ) { 03901 return false; 03902 } else { 03903 $this->mIncludeSizes[$type] += $size; 03904 return true; 03905 } 03906 } 03907 03913 function incrementExpensiveFunctionCount() { 03914 $this->mExpensiveFunctionCount++; 03915 return $this->mExpensiveFunctionCount <= $this->mOptions->getExpensiveParserFunctionLimit(); 03916 } 03917 03926 function doDoubleUnderscore( $text ) { 03927 wfProfileIn( __METHOD__ ); 03928 03929 # The position of __TOC__ needs to be recorded 03930 $mw = MagicWord::get( 'toc' ); 03931 if ( $mw->match( $text ) ) { 03932 $this->mShowToc = true; 03933 $this->mForceTocPosition = true; 03934 03935 # Set a placeholder. At the end we'll fill it in with the TOC. 03936 $text = $mw->replace( '<!--MWTOC-->', $text, 1 ); 03937 03938 # Only keep the first one. 03939 $text = $mw->replace( '', $text ); 03940 } 03941 03942 # Now match and remove the rest of them 03943 $mwa = MagicWord::getDoubleUnderscoreArray(); 03944 $this->mDoubleUnderscores = $mwa->matchAndRemove( $text ); 03945 03946 if ( isset( $this->mDoubleUnderscores['nogallery'] ) ) { 03947 $this->mOutput->mNoGallery = true; 03948 } 03949 if ( isset( $this->mDoubleUnderscores['notoc'] ) && !$this->mForceTocPosition ) { 03950 $this->mShowToc = false; 03951 } 03952 if ( isset( $this->mDoubleUnderscores['hiddencat'] ) && $this->mTitle->getNamespace() == NS_CATEGORY ) { 03953 $this->addTrackingCategory( 'hidden-category-category' ); 03954 } 03955 # (bug 8068) Allow control over whether robots index a page. 03956 # 03957 # @todo FIXME: Bug 14899: __INDEX__ always overrides __NOINDEX__ here! This 03958 # is not desirable, the last one on the page should win. 03959 if ( isset( $this->mDoubleUnderscores['noindex'] ) && $this->mTitle->canUseNoindex() ) { 03960 $this->mOutput->setIndexPolicy( 'noindex' ); 03961 $this->addTrackingCategory( 'noindex-category' ); 03962 } 03963 if ( isset( $this->mDoubleUnderscores['index'] ) && $this->mTitle->canUseNoindex() ) { 03964 $this->mOutput->setIndexPolicy( 'index' ); 03965 $this->addTrackingCategory( 'index-category' ); 03966 } 03967 03968 # Cache all double underscores in the database 03969 foreach ( $this->mDoubleUnderscores as $key => $val ) { 03970 $this->mOutput->setProperty( $key, '' ); 03971 } 03972 03973 wfProfileOut( __METHOD__ ); 03974 return $text; 03975 } 03976 03984 public function addTrackingCategory( $msg ) { 03985 if ( $this->mTitle->getNamespace() === NS_SPECIAL ) { 03986 wfDebug( __METHOD__.": Not adding tracking category $msg to special page!\n" ); 03987 return false; 03988 } 03989 // Important to parse with correct title (bug 31469) 03990 $cat = wfMessage( $msg ) 03991 ->title( $this->getTitle() ) 03992 ->inContentLanguage() 03993 ->text(); 03994 03995 # Allow tracking categories to be disabled by setting them to "-" 03996 if ( $cat === '-' ) { 03997 return false; 03998 } 03999 04000 $containerCategory = Title::makeTitleSafe( NS_CATEGORY, $cat ); 04001 if ( $containerCategory ) { 04002 $this->mOutput->addCategory( $containerCategory->getDBkey(), $this->getDefaultSort() ); 04003 return true; 04004 } else { 04005 wfDebug( __METHOD__.": [[MediaWiki:$msg]] is not a valid title!\n" ); 04006 return false; 04007 } 04008 } 04009 04026 function formatHeadings( $text, $origText, $isMain=true ) { 04027 global $wgMaxTocLevel, $wgHtml5, $wgExperimentalHtmlIds; 04028 04029 # Inhibit editsection links if requested in the page 04030 if ( isset( $this->mDoubleUnderscores['noeditsection'] ) ) { 04031 $maybeShowEditLink = $showEditLink = false; 04032 } else { 04033 $maybeShowEditLink = true; /* Actual presence will depend on ParserOptions option */ 04034 $showEditLink = $this->mOptions->getEditSection(); 04035 } 04036 if ( $showEditLink ) { 04037 $this->mOutput->setEditSectionTokens( true ); 04038 } 04039 04040 # Get all headlines for numbering them and adding funky stuff like [edit] 04041 # links - this is for later, but we need the number of headlines right now 04042 $matches = array(); 04043 $numMatches = preg_match_all( '/<H(?P<level>[1-6])(?P<attrib>.*?'.'>)(?P<header>.*?)<\/H[1-6] *>/i', $text, $matches ); 04044 04045 # if there are fewer than 4 headlines in the article, do not show TOC 04046 # unless it's been explicitly enabled. 04047 $enoughToc = $this->mShowToc && 04048 ( ( $numMatches >= 4 ) || $this->mForceTocPosition ); 04049 04050 # Allow user to stipulate that a page should have a "new section" 04051 # link added via __NEWSECTIONLINK__ 04052 if ( isset( $this->mDoubleUnderscores['newsectionlink'] ) ) { 04053 $this->mOutput->setNewSection( true ); 04054 } 04055 04056 # Allow user to remove the "new section" 04057 # link via __NONEWSECTIONLINK__ 04058 if ( isset( $this->mDoubleUnderscores['nonewsectionlink'] ) ) { 04059 $this->mOutput->hideNewSection( true ); 04060 } 04061 04062 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML, 04063 # override above conditions and always show TOC above first header 04064 if ( isset( $this->mDoubleUnderscores['forcetoc'] ) ) { 04065 $this->mShowToc = true; 04066 $enoughToc = true; 04067 } 04068 04069 # headline counter 04070 $headlineCount = 0; 04071 $numVisible = 0; 04072 04073 # Ugh .. the TOC should have neat indentation levels which can be 04074 # passed to the skin functions. These are determined here 04075 $toc = ''; 04076 $full = ''; 04077 $head = array(); 04078 $sublevelCount = array(); 04079 $levelCount = array(); 04080 $level = 0; 04081 $prevlevel = 0; 04082 $toclevel = 0; 04083 $prevtoclevel = 0; 04084 $markerRegex = "{$this->mUniqPrefix}-h-(\d+)-" . self::MARKER_SUFFIX; 04085 $baseTitleText = $this->mTitle->getPrefixedDBkey(); 04086 $oldType = $this->mOutputType; 04087 $this->setOutputType( self::OT_WIKI ); 04088 $frame = $this->getPreprocessor()->newFrame(); 04089 $root = $this->preprocessToDom( $origText ); 04090 $node = $root->getFirstChild(); 04091 $byteOffset = 0; 04092 $tocraw = array(); 04093 $refers = array(); 04094 04095 foreach ( $matches[3] as $headline ) { 04096 $isTemplate = false; 04097 $titleText = false; 04098 $sectionIndex = false; 04099 $numbering = ''; 04100 $markerMatches = array(); 04101 if ( preg_match("/^$markerRegex/", $headline, $markerMatches ) ) { 04102 $serial = $markerMatches[1]; 04103 list( $titleText, $sectionIndex ) = $this->mHeadings[$serial]; 04104 $isTemplate = ( $titleText != $baseTitleText ); 04105 $headline = preg_replace( "/^$markerRegex/", "", $headline ); 04106 } 04107 04108 if ( $toclevel ) { 04109 $prevlevel = $level; 04110 } 04111 $level = $matches[1][$headlineCount]; 04112 04113 if ( $level > $prevlevel ) { 04114 # Increase TOC level 04115 $toclevel++; 04116 $sublevelCount[$toclevel] = 0; 04117 if ( $toclevel<$wgMaxTocLevel ) { 04118 $prevtoclevel = $toclevel; 04119 $toc .= Linker::tocIndent(); 04120 $numVisible++; 04121 } 04122 } elseif ( $level < $prevlevel && $toclevel > 1 ) { 04123 # Decrease TOC level, find level to jump to 04124 04125 for ( $i = $toclevel; $i > 0; $i-- ) { 04126 if ( $levelCount[$i] == $level ) { 04127 # Found last matching level 04128 $toclevel = $i; 04129 break; 04130 } elseif ( $levelCount[$i] < $level ) { 04131 # Found first matching level below current level 04132 $toclevel = $i + 1; 04133 break; 04134 } 04135 } 04136 if ( $i == 0 ) { 04137 $toclevel = 1; 04138 } 04139 if ( $toclevel<$wgMaxTocLevel ) { 04140 if ( $prevtoclevel < $wgMaxTocLevel ) { 04141 # Unindent only if the previous toc level was shown :p 04142 $toc .= Linker::tocUnindent( $prevtoclevel - $toclevel ); 04143 $prevtoclevel = $toclevel; 04144 } else { 04145 $toc .= Linker::tocLineEnd(); 04146 } 04147 } 04148 } else { 04149 # No change in level, end TOC line 04150 if ( $toclevel<$wgMaxTocLevel ) { 04151 $toc .= Linker::tocLineEnd(); 04152 } 04153 } 04154 04155 $levelCount[$toclevel] = $level; 04156 04157 # count number of headlines for each level 04158 @$sublevelCount[$toclevel]++; 04159 $dot = 0; 04160 for( $i = 1; $i <= $toclevel; $i++ ) { 04161 if ( !empty( $sublevelCount[$i] ) ) { 04162 if ( $dot ) { 04163 $numbering .= '.'; 04164 } 04165 $numbering .= $this->getTargetLanguage()->formatNum( $sublevelCount[$i] ); 04166 $dot = 1; 04167 } 04168 } 04169 04170 # The safe header is a version of the header text safe to use for links 04171 04172 # Remove link placeholders by the link text. 04173 # <!--LINK number--> 04174 # turns into 04175 # link text with suffix 04176 # Do this before unstrip since link text can contain strip markers 04177 $safeHeadline = $this->replaceLinkHoldersText( $headline ); 04178 04179 # Avoid insertion of weird stuff like <math> by expanding the relevant sections 04180 $safeHeadline = $this->mStripState->unstripBoth( $safeHeadline ); 04181 04182 # Strip out HTML (first regex removes any tag not allowed) 04183 # Allowed tags are: 04184 # * <sup> and <sub> (bug 8393) 04185 # * <i> (bug 26375) 04186 # * <b> (r105284) 04187 # * <span dir="rtl"> and <span dir="ltr"> (bug 35167) 04188 # 04189 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>, 04190 # to allow setting directionality in toc items. 04191 $tocline = preg_replace( 04192 array( '#<(?!/?(span|sup|sub|i|b)(?: [^>]*)?>).*?'.'>#', '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|i|b))(?: .*?)?'.'>#' ), 04193 array( '', '<$1>' ), 04194 $safeHeadline 04195 ); 04196 $tocline = trim( $tocline ); 04197 04198 # For the anchor, strip out HTML-y stuff period 04199 $safeHeadline = preg_replace( '/<.*?'.'>/', '', $safeHeadline ); 04200 $safeHeadline = Sanitizer::normalizeSectionNameWhitespace( $safeHeadline ); 04201 04202 # Save headline for section edit hint before it's escaped 04203 $headlineHint = $safeHeadline; 04204 04205 if ( $wgHtml5 && $wgExperimentalHtmlIds ) { 04206 # For reverse compatibility, provide an id that's 04207 # HTML4-compatible, like we used to. 04208 # 04209 # It may be worth noting, academically, that it's possible for 04210 # the legacy anchor to conflict with a non-legacy headline 04211 # anchor on the page. In this case likely the "correct" thing 04212 # would be to either drop the legacy anchors or make sure 04213 # they're numbered first. However, this would require people 04214 # to type in section names like "abc_.D7.93.D7.90.D7.A4" 04215 # manually, so let's not bother worrying about it. 04216 $legacyHeadline = Sanitizer::escapeId( $safeHeadline, 04217 array( 'noninitial', 'legacy' ) ); 04218 $safeHeadline = Sanitizer::escapeId( $safeHeadline ); 04219 04220 if ( $legacyHeadline == $safeHeadline ) { 04221 # No reason to have both (in fact, we can't) 04222 $legacyHeadline = false; 04223 } 04224 } else { 04225 $legacyHeadline = false; 04226 $safeHeadline = Sanitizer::escapeId( $safeHeadline, 04227 'noninitial' ); 04228 } 04229 04230 # HTML names must be case-insensitively unique (bug 10721). 04231 # This does not apply to Unicode characters per 04232 # http://dev.w3.org/html5/spec/infrastructure.html#case-sensitivity-and-string-comparison 04233 # @todo FIXME: We may be changing them depending on the current locale. 04234 $arrayKey = strtolower( $safeHeadline ); 04235 if ( $legacyHeadline === false ) { 04236 $legacyArrayKey = false; 04237 } else { 04238 $legacyArrayKey = strtolower( $legacyHeadline ); 04239 } 04240 04241 # count how many in assoc. array so we can track dupes in anchors 04242 if ( isset( $refers[$arrayKey] ) ) { 04243 $refers[$arrayKey]++; 04244 } else { 04245 $refers[$arrayKey] = 1; 04246 } 04247 if ( isset( $refers[$legacyArrayKey] ) ) { 04248 $refers[$legacyArrayKey]++; 04249 } else { 04250 $refers[$legacyArrayKey] = 1; 04251 } 04252 04253 # Don't number the heading if it is the only one (looks silly) 04254 if ( count( $matches[3] ) > 1 && $this->mOptions->getNumberHeadings() ) { 04255 # the two are different if the line contains a link 04256 $headline = Html::element( 'span', array( 'class' => 'mw-headline-number' ), $numbering ) . ' ' . $headline; 04257 } 04258 04259 # Create the anchor for linking from the TOC to the section 04260 $anchor = $safeHeadline; 04261 $legacyAnchor = $legacyHeadline; 04262 if ( $refers[$arrayKey] > 1 ) { 04263 $anchor .= '_' . $refers[$arrayKey]; 04264 } 04265 if ( $legacyHeadline !== false && $refers[$legacyArrayKey] > 1 ) { 04266 $legacyAnchor .= '_' . $refers[$legacyArrayKey]; 04267 } 04268 if ( $enoughToc && ( !isset( $wgMaxTocLevel ) || $toclevel < $wgMaxTocLevel ) ) { 04269 $toc .= Linker::tocLine( $anchor, $tocline, 04270 $numbering, $toclevel, ( $isTemplate ? false : $sectionIndex ) ); 04271 } 04272 04273 # Add the section to the section tree 04274 # Find the DOM node for this header 04275 while ( $node && !$isTemplate ) { 04276 if ( $node->getName() === 'h' ) { 04277 $bits = $node->splitHeading(); 04278 if ( $bits['i'] == $sectionIndex ) { 04279 break; 04280 } 04281 } 04282 $byteOffset += mb_strlen( $this->mStripState->unstripBoth( 04283 $frame->expand( $node, PPFrame::RECOVER_ORIG ) ) ); 04284 $node = $node->getNextSibling(); 04285 } 04286 $tocraw[] = array( 04287 'toclevel' => $toclevel, 04288 'level' => $level, 04289 'line' => $tocline, 04290 'number' => $numbering, 04291 'index' => ( $isTemplate ? 'T-' : '' ) . $sectionIndex, 04292 'fromtitle' => $titleText, 04293 'byteoffset' => ( $isTemplate ? null : $byteOffset ), 04294 'anchor' => $anchor, 04295 ); 04296 04297 # give headline the correct <h#> tag 04298 if ( $maybeShowEditLink && $sectionIndex !== false ) { 04299 // Output edit section links as markers with styles that can be customized by skins 04300 if ( $isTemplate ) { 04301 # Put a T flag in the section identifier, to indicate to extractSections() 04302 # that sections inside <includeonly> should be counted. 04303 $editlinkArgs = array( $titleText, "T-$sectionIndex"/*, null */ ); 04304 } else { 04305 $editlinkArgs = array( $this->mTitle->getPrefixedText(), $sectionIndex, $headlineHint ); 04306 } 04307 // We use a bit of pesudo-xml for editsection markers. The language converter is run later on 04308 // Using a UNIQ style marker leads to the converter screwing up the tokens when it converts stuff 04309 // And trying to insert strip tags fails too. At this point all real inputted tags have already been escaped 04310 // so we don't have to worry about a user trying to input one of these markers directly. 04311 // We use a page and section attribute to stop the language converter from converting these important bits 04312 // of data, but put the headline hint inside a content block because the language converter is supposed to 04313 // be able to convert that piece of data. 04314 $editlink = '<mw:editsection page="' . htmlspecialchars($editlinkArgs[0]); 04315 $editlink .= '" section="' . htmlspecialchars($editlinkArgs[1]) .'"'; 04316 if ( isset($editlinkArgs[2]) ) { 04317 $editlink .= '>' . $editlinkArgs[2] . '</mw:editsection>'; 04318 } else { 04319 $editlink .= '/>'; 04320 } 04321 } else { 04322 $editlink = ''; 04323 } 04324 $head[$headlineCount] = Linker::makeHeadline( $level, 04325 $matches['attrib'][$headlineCount], $anchor, $headline, 04326 $editlink, $legacyAnchor ); 04327 04328 $headlineCount++; 04329 } 04330 04331 $this->setOutputType( $oldType ); 04332 04333 # Never ever show TOC if no headers 04334 if ( $numVisible < 1 ) { 04335 $enoughToc = false; 04336 } 04337 04338 if ( $enoughToc ) { 04339 if ( $prevtoclevel > 0 && $prevtoclevel < $wgMaxTocLevel ) { 04340 $toc .= Linker::tocUnindent( $prevtoclevel - 1 ); 04341 } 04342 $toc = Linker::tocList( $toc, $this->mOptions->getUserLangObj() ); 04343 $this->mOutput->setTOCHTML( $toc ); 04344 } 04345 04346 if ( $isMain ) { 04347 $this->mOutput->setSections( $tocraw ); 04348 } 04349 04350 # split up and insert constructed headlines 04351 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text ); 04352 $i = 0; 04353 04354 // build an array of document sections 04355 $sections = array(); 04356 foreach ( $blocks as $block ) { 04357 // $head is zero-based, sections aren't. 04358 if ( empty( $head[$i - 1] ) ) { 04359 $sections[$i] = $block; 04360 } else { 04361 $sections[$i] = $head[$i - 1] . $block; 04362 } 04363 04374 wfRunHooks( 'ParserSectionCreate', array( $this, $i, &$sections[$i], $showEditLink ) ); 04375 04376 $i++; 04377 } 04378 04379 if ( $enoughToc && $isMain && !$this->mForceTocPosition ) { 04380 // append the TOC at the beginning 04381 // Top anchor now in skin 04382 $sections[0] = $sections[0] . $toc . "\n"; 04383 } 04384 04385 $full .= join( '', $sections ); 04386 04387 if ( $this->mForceTocPosition ) { 04388 return str_replace( '<!--MWTOC-->', $toc, $full ); 04389 } else { 04390 return $full; 04391 } 04392 } 04393 04405 public function preSaveTransform( $text, Title $title, User $user, ParserOptions $options, $clearState = true ) { 04406 $this->startParse( $title, $options, self::OT_WIKI, $clearState ); 04407 $this->setUser( $user ); 04408 04409 $pairs = array( 04410 "\r\n" => "\n", 04411 ); 04412 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text ); 04413 if( $options->getPreSaveTransform() ) { 04414 $text = $this->pstPass2( $text, $user ); 04415 } 04416 $text = $this->mStripState->unstripBoth( $text ); 04417 04418 $this->setUser( null ); #Reset 04419 04420 return $text; 04421 } 04422 04432 function pstPass2( $text, $user ) { 04433 global $wgContLang, $wgLocaltimezone; 04434 04435 # Note: This is the timestamp saved as hardcoded wikitext to 04436 # the database, we use $wgContLang here in order to give 04437 # everyone the same signature and use the default one rather 04438 # than the one selected in each user's preferences. 04439 # (see also bug 12815) 04440 $ts = $this->mOptions->getTimestamp(); 04441 if ( isset( $wgLocaltimezone ) ) { 04442 $tz = $wgLocaltimezone; 04443 } else { 04444 $tz = date_default_timezone_get(); 04445 } 04446 04447 $unixts = wfTimestamp( TS_UNIX, $ts ); 04448 $oldtz = date_default_timezone_get(); 04449 date_default_timezone_set( $tz ); 04450 $ts = date( 'YmdHis', $unixts ); 04451 $tzMsg = date( 'T', $unixts ); # might vary on DST changeover! 04452 04453 # Allow translation of timezones through wiki. date() can return 04454 # whatever crap the system uses, localised or not, so we cannot 04455 # ship premade translations. 04456 $key = 'timezone-' . strtolower( trim( $tzMsg ) ); 04457 $msg = wfMessage( $key )->inContentLanguage(); 04458 if ( $msg->exists() ) { 04459 $tzMsg = $msg->text(); 04460 } 04461 04462 date_default_timezone_set( $oldtz ); 04463 04464 $d = $wgContLang->timeanddate( $ts, false, false ) . " ($tzMsg)"; 04465 04466 # Variable replacement 04467 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags 04468 $text = $this->replaceVariables( $text ); 04469 04470 # This works almost by chance, as the replaceVariables are done before the getUserSig(), 04471 # which may corrupt this parser instance via its wfMessage()->text() call- 04472 04473 # Signatures 04474 $sigText = $this->getUserSig( $user ); 04475 $text = strtr( $text, array( 04476 '~~~~~' => $d, 04477 '~~~~' => "$sigText $d", 04478 '~~~' => $sigText 04479 ) ); 04480 04481 # Context links: [[|name]] and [[name (context)|]] 04482 $tc = '[' . Title::legalChars() . ']'; 04483 $nc = '[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can use non-ascii! 04484 04485 $p1 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\))\\|]]/"; # [[ns:page (context)|]] 04486 $p4 = "/\[\[(:?$nc+:|:|)($tc+?)( ?($tc+))\\|]]/"; # [[ns:page(context)|]] 04487 $p3 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\)|)((?:, |,)$tc+|)\\|]]/"; # [[ns:page (context), context|]] 04488 $p2 = "/\[\[\\|($tc+)]]/"; # [[|page]] 04489 04490 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]" 04491 $text = preg_replace( $p1, '[[\\1\\2\\3|\\2]]', $text ); 04492 $text = preg_replace( $p4, '[[\\1\\2\\3|\\2]]', $text ); 04493 $text = preg_replace( $p3, '[[\\1\\2\\3\\4|\\2]]', $text ); 04494 04495 $t = $this->mTitle->getText(); 04496 $m = array(); 04497 if ( preg_match( "/^($nc+:|)$tc+?( \\($tc+\\))$/", $t, $m ) ) { 04498 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text ); 04499 } elseif ( preg_match( "/^($nc+:|)$tc+?(, $tc+|)$/", $t, $m ) && "$m[1]$m[2]" != '' ) { 04500 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text ); 04501 } else { 04502 # if there's no context, don't bother duplicating the title 04503 $text = preg_replace( $p2, '[[\\1]]', $text ); 04504 } 04505 04506 # Trim trailing whitespace 04507 $text = rtrim( $text ); 04508 04509 return $text; 04510 } 04511 04526 function getUserSig( &$user, $nickname = false, $fancySig = null ) { 04527 global $wgMaxSigChars; 04528 04529 $username = $user->getName(); 04530 04531 # If not given, retrieve from the user object. 04532 if ( $nickname === false ) 04533 $nickname = $user->getOption( 'nickname' ); 04534 04535 if ( is_null( $fancySig ) ) { 04536 $fancySig = $user->getBoolOption( 'fancysig' ); 04537 } 04538 04539 $nickname = $nickname == null ? $username : $nickname; 04540 04541 if ( mb_strlen( $nickname ) > $wgMaxSigChars ) { 04542 $nickname = $username; 04543 wfDebug( __METHOD__ . ": $username has overlong signature.\n" ); 04544 } elseif ( $fancySig !== false ) { 04545 # Sig. might contain markup; validate this 04546 if ( $this->validateSig( $nickname ) !== false ) { 04547 # Validated; clean up (if needed) and return it 04548 return $this->cleanSig( $nickname, true ); 04549 } else { 04550 # Failed to validate; fall back to the default 04551 $nickname = $username; 04552 wfDebug( __METHOD__.": $username has bad XML tags in signature.\n" ); 04553 } 04554 } 04555 04556 # Make sure nickname doesnt get a sig in a sig 04557 $nickname = self::cleanSigInSig( $nickname ); 04558 04559 # If we're still here, make it a link to the user page 04560 $userText = wfEscapeWikiText( $username ); 04561 $nickText = wfEscapeWikiText( $nickname ); 04562 $msgName = $user->isAnon() ? 'signature-anon' : 'signature'; 04563 04564 return wfMessage( $msgName, $userText, $nickText )->inContentLanguage()->title( $this->getTitle() )->text(); 04565 } 04566 04573 function validateSig( $text ) { 04574 return( Xml::isWellFormedXmlFragment( $text ) ? $text : false ); 04575 } 04576 04587 public function cleanSig( $text, $parsing = false ) { 04588 if ( !$parsing ) { 04589 global $wgTitle; 04590 $this->startParse( $wgTitle, new ParserOptions, self::OT_PREPROCESS, true ); 04591 } 04592 04593 # Option to disable this feature 04594 if ( !$this->mOptions->getCleanSignatures() ) { 04595 return $text; 04596 } 04597 04598 # @todo FIXME: Regex doesn't respect extension tags or nowiki 04599 # => Move this logic to braceSubstitution() 04600 $substWord = MagicWord::get( 'subst' ); 04601 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase(); 04602 $substText = '{{' . $substWord->getSynonym( 0 ); 04603 04604 $text = preg_replace( $substRegex, $substText, $text ); 04605 $text = self::cleanSigInSig( $text ); 04606 $dom = $this->preprocessToDom( $text ); 04607 $frame = $this->getPreprocessor()->newFrame(); 04608 $text = $frame->expand( $dom ); 04609 04610 if ( !$parsing ) { 04611 $text = $this->mStripState->unstripBoth( $text ); 04612 } 04613 04614 return $text; 04615 } 04616 04623 public static function cleanSigInSig( $text ) { 04624 $text = preg_replace( '/~{3,5}/', '', $text ); 04625 return $text; 04626 } 04627 04637 public function startExternalParse( Title $title = null, ParserOptions $options, $outputType, $clearState = true ) { 04638 $this->startParse( $title, $options, $outputType, $clearState ); 04639 } 04640 04647 private function startParse( Title $title = null, ParserOptions $options, $outputType, $clearState = true ) { 04648 $this->setTitle( $title ); 04649 $this->mOptions = $options; 04650 $this->setOutputType( $outputType ); 04651 if ( $clearState ) { 04652 $this->clearState(); 04653 } 04654 } 04655 04664 public function transformMsg( $text, $options, $title = null ) { 04665 static $executing = false; 04666 04667 # Guard against infinite recursion 04668 if ( $executing ) { 04669 return $text; 04670 } 04671 $executing = true; 04672 04673 wfProfileIn( __METHOD__ ); 04674 if ( !$title ) { 04675 global $wgTitle; 04676 $title = $wgTitle; 04677 } 04678 if ( !$title ) { 04679 # It's not uncommon having a null $wgTitle in scripts. See r80898 04680 # Create a ghost title in such case 04681 $title = Title::newFromText( 'Dwimmerlaik' ); 04682 } 04683 $text = $this->preprocess( $text, $title, $options ); 04684 04685 $executing = false; 04686 wfProfileOut( __METHOD__ ); 04687 return $text; 04688 } 04689 04714 public function setHook( $tag, $callback ) { 04715 $tag = strtolower( $tag ); 04716 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) { 04717 throw new MWException( "Invalid character {$m[0]} in setHook('$tag', ...) call" ); 04718 } 04719 $oldVal = isset( $this->mTagHooks[$tag] ) ? $this->mTagHooks[$tag] : null; 04720 $this->mTagHooks[$tag] = $callback; 04721 if ( !in_array( $tag, $this->mStripList ) ) { 04722 $this->mStripList[] = $tag; 04723 } 04724 04725 return $oldVal; 04726 } 04727 04745 function setTransparentTagHook( $tag, $callback ) { 04746 $tag = strtolower( $tag ); 04747 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) { 04748 throw new MWException( "Invalid character {$m[0]} in setTransparentHook('$tag', ...) call" ); 04749 } 04750 $oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null; 04751 $this->mTransparentTagHooks[$tag] = $callback; 04752 04753 return $oldVal; 04754 } 04755 04759 function clearTagHooks() { 04760 $this->mTagHooks = array(); 04761 $this->mFunctionTagHooks = array(); 04762 $this->mStripList = $this->mDefaultStripList; 04763 } 04764 04808 public function setFunctionHook( $id, $callback, $flags = 0 ) { 04809 global $wgContLang; 04810 04811 $oldVal = isset( $this->mFunctionHooks[$id] ) ? $this->mFunctionHooks[$id][0] : null; 04812 $this->mFunctionHooks[$id] = array( $callback, $flags ); 04813 04814 # Add to function cache 04815 $mw = MagicWord::get( $id ); 04816 if ( !$mw ) 04817 throw new MWException( __METHOD__.'() expecting a magic word identifier.' ); 04818 04819 $synonyms = $mw->getSynonyms(); 04820 $sensitive = intval( $mw->isCaseSensitive() ); 04821 04822 foreach ( $synonyms as $syn ) { 04823 # Case 04824 if ( !$sensitive ) { 04825 $syn = $wgContLang->lc( $syn ); 04826 } 04827 # Add leading hash 04828 if ( !( $flags & SFH_NO_HASH ) ) { 04829 $syn = '#' . $syn; 04830 } 04831 # Remove trailing colon 04832 if ( substr( $syn, -1, 1 ) === ':' ) { 04833 $syn = substr( $syn, 0, -1 ); 04834 } 04835 $this->mFunctionSynonyms[$sensitive][$syn] = $id; 04836 } 04837 return $oldVal; 04838 } 04839 04845 function getFunctionHooks() { 04846 return array_keys( $this->mFunctionHooks ); 04847 } 04848 04859 function setFunctionTagHook( $tag, $callback, $flags ) { 04860 $tag = strtolower( $tag ); 04861 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) throw new MWException( "Invalid character {$m[0]} in setFunctionTagHook('$tag', ...) call" ); 04862 $old = isset( $this->mFunctionTagHooks[$tag] ) ? 04863 $this->mFunctionTagHooks[$tag] : null; 04864 $this->mFunctionTagHooks[$tag] = array( $callback, $flags ); 04865 04866 if ( !in_array( $tag, $this->mStripList ) ) { 04867 $this->mStripList[] = $tag; 04868 } 04869 04870 return $old; 04871 } 04872 04883 function replaceLinkHolders( &$text, $options = 0 ) { 04884 return $this->mLinkHolders->replace( $text ); 04885 } 04886 04894 function replaceLinkHoldersText( $text ) { 04895 return $this->mLinkHolders->replaceText( $text ); 04896 } 04897 04911 function renderImageGallery( $text, $params ) { 04912 $ig = new ImageGallery(); 04913 $ig->setContextTitle( $this->mTitle ); 04914 $ig->setShowBytes( false ); 04915 $ig->setShowFilename( false ); 04916 $ig->setParser( $this ); 04917 $ig->setHideBadImages(); 04918 $ig->setAttributes( Sanitizer::validateTagAttributes( $params, 'table' ) ); 04919 04920 if ( isset( $params['showfilename'] ) ) { 04921 $ig->setShowFilename( true ); 04922 } else { 04923 $ig->setShowFilename( false ); 04924 } 04925 if ( isset( $params['caption'] ) ) { 04926 $caption = $params['caption']; 04927 $caption = htmlspecialchars( $caption ); 04928 $caption = $this->replaceInternalLinks( $caption ); 04929 $ig->setCaptionHtml( $caption ); 04930 } 04931 if ( isset( $params['perrow'] ) ) { 04932 $ig->setPerRow( $params['perrow'] ); 04933 } 04934 if ( isset( $params['widths'] ) ) { 04935 $ig->setWidths( $params['widths'] ); 04936 } 04937 if ( isset( $params['heights'] ) ) { 04938 $ig->setHeights( $params['heights'] ); 04939 } 04940 04941 wfRunHooks( 'BeforeParserrenderImageGallery', array( &$this, &$ig ) ); 04942 04943 $lines = StringUtils::explode( "\n", $text ); 04944 foreach ( $lines as $line ) { 04945 # match lines like these: 04946 # Image:someimage.jpg|This is some image 04947 $matches = array(); 04948 preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches ); 04949 # Skip empty lines 04950 if ( count( $matches ) == 0 ) { 04951 continue; 04952 } 04953 04954 if ( strpos( $matches[0], '%' ) !== false ) { 04955 $matches[1] = rawurldecode( $matches[1] ); 04956 } 04957 $title = Title::newFromText( $matches[1], NS_FILE ); 04958 if ( is_null( $title ) ) { 04959 # Bogus title. Ignore these so we don't bomb out later. 04960 continue; 04961 } 04962 04963 $label = ''; 04964 $alt = ''; 04965 $link = ''; 04966 if ( isset( $matches[3] ) ) { 04967 // look for an |alt= definition while trying not to break existing 04968 // captions with multiple pipes (|) in it, until a more sensible grammar 04969 // is defined for images in galleries 04970 04971 $matches[3] = $this->recursiveTagParse( trim( $matches[3] ) ); 04972 $parameterMatches = StringUtils::explode('|', $matches[3]); 04973 $magicWordAlt = MagicWord::get( 'img_alt' ); 04974 $magicWordLink = MagicWord::get( 'img_link' ); 04975 04976 foreach ( $parameterMatches as $parameterMatch ) { 04977 if ( $match = $magicWordAlt->matchVariableStartToEnd( $parameterMatch ) ) { 04978 $alt = $this->stripAltText( $match, false ); 04979 } 04980 elseif( $match = $magicWordLink->matchVariableStartToEnd( $parameterMatch ) ){ 04981 $link = strip_tags($this->replaceLinkHoldersText($match)); 04982 $chars = self::EXT_LINK_URL_CLASS; 04983 $prots = $this->mUrlProtocols; 04984 //check to see if link matches an absolute url, if not then it must be a wiki link. 04985 if(!preg_match( "/^($prots)$chars+$/u", $link)){ 04986 $localLinkTitle = Title::newFromText($link); 04987 $link = $localLinkTitle->getLocalURL(); 04988 } 04989 } 04990 else { 04991 // concatenate all other pipes 04992 $label .= '|' . $parameterMatch; 04993 } 04994 } 04995 // remove the first pipe 04996 $label = substr( $label, 1 ); 04997 } 04998 04999 $ig->add( $title, $label, $alt ,$link); 05000 } 05001 return $ig->toHTML(); 05002 } 05003 05008 function getImageParams( $handler ) { 05009 if ( $handler ) { 05010 $handlerClass = get_class( $handler ); 05011 } else { 05012 $handlerClass = ''; 05013 } 05014 if ( !isset( $this->mImageParams[$handlerClass] ) ) { 05015 # Initialise static lists 05016 static $internalParamNames = array( 05017 'horizAlign' => array( 'left', 'right', 'center', 'none' ), 05018 'vertAlign' => array( 'baseline', 'sub', 'super', 'top', 'text-top', 'middle', 05019 'bottom', 'text-bottom' ), 05020 'frame' => array( 'thumbnail', 'manualthumb', 'framed', 'frameless', 05021 'upright', 'border', 'link', 'alt', 'class' ), 05022 ); 05023 static $internalParamMap; 05024 if ( !$internalParamMap ) { 05025 $internalParamMap = array(); 05026 foreach ( $internalParamNames as $type => $names ) { 05027 foreach ( $names as $name ) { 05028 $magicName = str_replace( '-', '_', "img_$name" ); 05029 $internalParamMap[$magicName] = array( $type, $name ); 05030 } 05031 } 05032 } 05033 05034 # Add handler params 05035 $paramMap = $internalParamMap; 05036 if ( $handler ) { 05037 $handlerParamMap = $handler->getParamMap(); 05038 foreach ( $handlerParamMap as $magic => $paramName ) { 05039 $paramMap[$magic] = array( 'handler', $paramName ); 05040 } 05041 } 05042 $this->mImageParams[$handlerClass] = $paramMap; 05043 $this->mImageParamsMagicArray[$handlerClass] = new MagicWordArray( array_keys( $paramMap ) ); 05044 } 05045 return array( $this->mImageParams[$handlerClass], $this->mImageParamsMagicArray[$handlerClass] ); 05046 } 05047 05056 function makeImage( $title, $options, $holders = false ) { 05057 # Check if the options text is of the form "options|alt text" 05058 # Options are: 05059 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang 05060 # * left no resizing, just left align. label is used for alt= only 05061 # * right same, but right aligned 05062 # * none same, but not aligned 05063 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox 05064 # * center center the image 05065 # * frame Keep original image size, no magnify-button. 05066 # * framed Same as "frame" 05067 # * frameless like 'thumb' but without a frame. Keeps user preferences for width 05068 # * upright reduce width for upright images, rounded to full __0 px 05069 # * border draw a 1px border around the image 05070 # * alt Text for HTML alt attribute (defaults to empty) 05071 # * class Set a class for img node 05072 # * link Set the target of the image link. Can be external, interwiki, or local 05073 # vertical-align values (no % or length right now): 05074 # * baseline 05075 # * sub 05076 # * super 05077 # * top 05078 # * text-top 05079 # * middle 05080 # * bottom 05081 # * text-bottom 05082 05083 $parts = StringUtils::explode( "|", $options ); 05084 05085 # Give extensions a chance to select the file revision for us 05086 $options = array(); 05087 $descQuery = false; 05088 wfRunHooks( 'BeforeParserFetchFileAndTitle', 05089 array( $this, $title, &$options, &$descQuery ) ); 05090 # Fetch and register the file (file title may be different via hooks) 05091 list( $file, $title ) = $this->fetchFileAndTitle( $title, $options ); 05092 05093 # Get parameter map 05094 $handler = $file ? $file->getHandler() : false; 05095 05096 list( $paramMap, $mwArray ) = $this->getImageParams( $handler ); 05097 05098 if ( !$file ) { 05099 $this->addTrackingCategory( 'broken-file-category' ); 05100 } 05101 05102 # Process the input parameters 05103 $caption = ''; 05104 $params = array( 'frame' => array(), 'handler' => array(), 05105 'horizAlign' => array(), 'vertAlign' => array() ); 05106 foreach ( $parts as $part ) { 05107 $part = trim( $part ); 05108 list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part ); 05109 $validated = false; 05110 if ( isset( $paramMap[$magicName] ) ) { 05111 list( $type, $paramName ) = $paramMap[$magicName]; 05112 05113 # Special case; width and height come in one variable together 05114 if ( $type === 'handler' && $paramName === 'width' ) { 05115 $parsedWidthParam = $this->parseWidthParam( $value ); 05116 if( isset( $parsedWidthParam['width'] ) ) { 05117 $width = $parsedWidthParam['width']; 05118 if ( $handler->validateParam( 'width', $width ) ) { 05119 $params[$type]['width'] = $width; 05120 $validated = true; 05121 } 05122 } 05123 if( isset( $parsedWidthParam['height'] ) ) { 05124 $height = $parsedWidthParam['height']; 05125 if ( $handler->validateParam( 'height', $height ) ) { 05126 $params[$type]['height'] = $height; 05127 $validated = true; 05128 } 05129 } 05130 # else no validation -- bug 13436 05131 } else { 05132 if ( $type === 'handler' ) { 05133 # Validate handler parameter 05134 $validated = $handler->validateParam( $paramName, $value ); 05135 } else { 05136 # Validate internal parameters 05137 switch( $paramName ) { 05138 case 'manualthumb': 05139 case 'alt': 05140 case 'class': 05141 # @todo FIXME: Possibly check validity here for 05142 # manualthumb? downstream behavior seems odd with 05143 # missing manual thumbs. 05144 $validated = true; 05145 $value = $this->stripAltText( $value, $holders ); 05146 break; 05147 case 'link': 05148 $chars = self::EXT_LINK_URL_CLASS; 05149 $prots = $this->mUrlProtocols; 05150 if ( $value === '' ) { 05151 $paramName = 'no-link'; 05152 $value = true; 05153 $validated = true; 05154 } elseif ( preg_match( "/^(?i)$prots/", $value ) ) { 05155 if ( preg_match( "/^((?i)$prots)$chars+$/u", $value, $m ) ) { 05156 $paramName = 'link-url'; 05157 $this->mOutput->addExternalLink( $value ); 05158 if ( $this->mOptions->getExternalLinkTarget() ) { 05159 $params[$type]['link-target'] = $this->mOptions->getExternalLinkTarget(); 05160 } 05161 $validated = true; 05162 } 05163 } else { 05164 $linkTitle = Title::newFromText( $value ); 05165 if ( $linkTitle ) { 05166 $paramName = 'link-title'; 05167 $value = $linkTitle; 05168 $this->mOutput->addLink( $linkTitle ); 05169 $validated = true; 05170 } 05171 } 05172 break; 05173 default: 05174 # Most other things appear to be empty or numeric... 05175 $validated = ( $value === false || is_numeric( trim( $value ) ) ); 05176 } 05177 } 05178 05179 if ( $validated ) { 05180 $params[$type][$paramName] = $value; 05181 } 05182 } 05183 } 05184 if ( !$validated ) { 05185 $caption = $part; 05186 } 05187 } 05188 05189 # Process alignment parameters 05190 if ( $params['horizAlign'] ) { 05191 $params['frame']['align'] = key( $params['horizAlign'] ); 05192 } 05193 if ( $params['vertAlign'] ) { 05194 $params['frame']['valign'] = key( $params['vertAlign'] ); 05195 } 05196 05197 $params['frame']['caption'] = $caption; 05198 05199 # Will the image be presented in a frame, with the caption below? 05200 $imageIsFramed = isset( $params['frame']['frame'] ) || 05201 isset( $params['frame']['framed'] ) || 05202 isset( $params['frame']['thumbnail'] ) || 05203 isset( $params['frame']['manualthumb'] ); 05204 05205 # In the old days, [[Image:Foo|text...]] would set alt text. Later it 05206 # came to also set the caption, ordinary text after the image -- which 05207 # makes no sense, because that just repeats the text multiple times in 05208 # screen readers. It *also* came to set the title attribute. 05209 # 05210 # Now that we have an alt attribute, we should not set the alt text to 05211 # equal the caption: that's worse than useless, it just repeats the 05212 # text. This is the framed/thumbnail case. If there's no caption, we 05213 # use the unnamed parameter for alt text as well, just for the time be- 05214 # ing, if the unnamed param is set and the alt param is not. 05215 # 05216 # For the future, we need to figure out if we want to tweak this more, 05217 # e.g., introducing a title= parameter for the title; ignoring the un- 05218 # named parameter entirely for images without a caption; adding an ex- 05219 # plicit caption= parameter and preserving the old magic unnamed para- 05220 # meter for BC; ... 05221 if ( $imageIsFramed ) { # Framed image 05222 if ( $caption === '' && !isset( $params['frame']['alt'] ) ) { 05223 # No caption or alt text, add the filename as the alt text so 05224 # that screen readers at least get some description of the image 05225 $params['frame']['alt'] = $title->getText(); 05226 } 05227 # Do not set $params['frame']['title'] because tooltips don't make sense 05228 # for framed images 05229 } else { # Inline image 05230 if ( !isset( $params['frame']['alt'] ) ) { 05231 # No alt text, use the "caption" for the alt text 05232 if ( $caption !== '') { 05233 $params['frame']['alt'] = $this->stripAltText( $caption, $holders ); 05234 } else { 05235 # No caption, fall back to using the filename for the 05236 # alt text 05237 $params['frame']['alt'] = $title->getText(); 05238 } 05239 } 05240 # Use the "caption" for the tooltip text 05241 $params['frame']['title'] = $this->stripAltText( $caption, $holders ); 05242 } 05243 05244 wfRunHooks( 'ParserMakeImageParams', array( $title, $file, &$params, $this ) ); 05245 05246 # Linker does the rest 05247 $time = isset( $options['time'] ) ? $options['time'] : false; 05248 $ret = Linker::makeImageLink( $this, $title, $file, $params['frame'], $params['handler'], 05249 $time, $descQuery, $this->mOptions->getThumbSize() ); 05250 05251 # Give the handler a chance to modify the parser object 05252 if ( $handler ) { 05253 $handler->parserTransformHook( $this, $file ); 05254 } 05255 05256 return $ret; 05257 } 05258 05264 protected function stripAltText( $caption, $holders ) { 05265 # Strip bad stuff out of the title (tooltip). We can't just use 05266 # replaceLinkHoldersText() here, because if this function is called 05267 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date. 05268 if ( $holders ) { 05269 $tooltip = $holders->replaceText( $caption ); 05270 } else { 05271 $tooltip = $this->replaceLinkHoldersText( $caption ); 05272 } 05273 05274 # make sure there are no placeholders in thumbnail attributes 05275 # that are later expanded to html- so expand them now and 05276 # remove the tags 05277 $tooltip = $this->mStripState->unstripBoth( $tooltip ); 05278 $tooltip = Sanitizer::stripAllTags( $tooltip ); 05279 05280 return $tooltip; 05281 } 05282 05287 function disableCache() { 05288 wfDebug( "Parser output marked as uncacheable.\n" ); 05289 if ( !$this->mOutput ) { 05290 throw new MWException( __METHOD__ . 05291 " can only be called when actually parsing something" ); 05292 } 05293 $this->mOutput->setCacheTime( -1 ); // old style, for compatibility 05294 $this->mOutput->updateCacheExpiry( 0 ); // new style, for consistency 05295 } 05296 05305 function attributeStripCallback( &$text, $frame = false ) { 05306 $text = $this->replaceVariables( $text, $frame ); 05307 $text = $this->mStripState->unstripBoth( $text ); 05308 return $text; 05309 } 05310 05316 function getTags() { 05317 return array_merge( array_keys( $this->mTransparentTagHooks ), array_keys( $this->mTagHooks ), array_keys( $this->mFunctionTagHooks ) ); 05318 } 05319 05330 function replaceTransparentTags( $text ) { 05331 $matches = array(); 05332 $elements = array_keys( $this->mTransparentTagHooks ); 05333 $text = self::extractTagsAndParams( $elements, $text, $matches, $this->mUniqPrefix ); 05334 $replacements = array(); 05335 05336 foreach ( $matches as $marker => $data ) { 05337 list( $element, $content, $params, $tag ) = $data; 05338 $tagName = strtolower( $element ); 05339 if ( isset( $this->mTransparentTagHooks[$tagName] ) ) { 05340 $output = call_user_func_array( $this->mTransparentTagHooks[$tagName], array( $content, $params, $this ) ); 05341 } else { 05342 $output = $tag; 05343 } 05344 $replacements[$marker] = $output; 05345 } 05346 return strtr( $text, $replacements ); 05347 } 05348 05378 private function extractSections( $text, $section, $mode, $newText='' ) { 05379 global $wgTitle; # not generally used but removes an ugly failure mode 05380 $this->startParse( $wgTitle, new ParserOptions, self::OT_PLAIN, true ); 05381 $outText = ''; 05382 $frame = $this->getPreprocessor()->newFrame(); 05383 05384 # Process section extraction flags 05385 $flags = 0; 05386 $sectionParts = explode( '-', $section ); 05387 $sectionIndex = array_pop( $sectionParts ); 05388 foreach ( $sectionParts as $part ) { 05389 if ( $part === 'T' ) { 05390 $flags |= self::PTD_FOR_INCLUSION; 05391 } 05392 } 05393 05394 # Check for empty input 05395 if ( strval( $text ) === '' ) { 05396 # Only sections 0 and T-0 exist in an empty document 05397 if ( $sectionIndex == 0 ) { 05398 if ( $mode === 'get' ) { 05399 return ''; 05400 } else { 05401 return $newText; 05402 } 05403 } else { 05404 if ( $mode === 'get' ) { 05405 return $newText; 05406 } else { 05407 return $text; 05408 } 05409 } 05410 } 05411 05412 # Preprocess the text 05413 $root = $this->preprocessToDom( $text, $flags ); 05414 05415 # <h> nodes indicate section breaks 05416 # They can only occur at the top level, so we can find them by iterating the root's children 05417 $node = $root->getFirstChild(); 05418 05419 # Find the target section 05420 if ( $sectionIndex == 0 ) { 05421 # Section zero doesn't nest, level=big 05422 $targetLevel = 1000; 05423 } else { 05424 while ( $node ) { 05425 if ( $node->getName() === 'h' ) { 05426 $bits = $node->splitHeading(); 05427 if ( $bits['i'] == $sectionIndex ) { 05428 $targetLevel = $bits['level']; 05429 break; 05430 } 05431 } 05432 if ( $mode === 'replace' ) { 05433 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG ); 05434 } 05435 $node = $node->getNextSibling(); 05436 } 05437 } 05438 05439 if ( !$node ) { 05440 # Not found 05441 if ( $mode === 'get' ) { 05442 return $newText; 05443 } else { 05444 return $text; 05445 } 05446 } 05447 05448 # Find the end of the section, including nested sections 05449 do { 05450 if ( $node->getName() === 'h' ) { 05451 $bits = $node->splitHeading(); 05452 $curLevel = $bits['level']; 05453 if ( $bits['i'] != $sectionIndex && $curLevel <= $targetLevel ) { 05454 break; 05455 } 05456 } 05457 if ( $mode === 'get' ) { 05458 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG ); 05459 } 05460 $node = $node->getNextSibling(); 05461 } while ( $node ); 05462 05463 # Write out the remainder (in replace mode only) 05464 if ( $mode === 'replace' ) { 05465 # Output the replacement text 05466 # Add two newlines on -- trailing whitespace in $newText is conventionally 05467 # stripped by the editor, so we need both newlines to restore the paragraph gap 05468 # Only add trailing whitespace if there is newText 05469 if ( $newText != "" ) { 05470 $outText .= $newText . "\n\n"; 05471 } 05472 05473 while ( $node ) { 05474 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG ); 05475 $node = $node->getNextSibling(); 05476 } 05477 } 05478 05479 if ( is_string( $outText ) ) { 05480 # Re-insert stripped tags 05481 $outText = rtrim( $this->mStripState->unstripBoth( $outText ) ); 05482 } 05483 05484 return $outText; 05485 } 05486 05499 public function getSection( $text, $section, $deftext='' ) { 05500 return $this->extractSections( $text, $section, "get", $deftext ); 05501 } 05502 05513 public function replaceSection( $oldtext, $section, $text ) { 05514 return $this->extractSections( $oldtext, $section, "replace", $text ); 05515 } 05516 05522 function getRevisionId() { 05523 return $this->mRevisionId; 05524 } 05525 05531 protected function getRevisionObject() { 05532 if ( !is_null( $this->mRevisionObject ) ) { 05533 return $this->mRevisionObject; 05534 } 05535 if ( is_null( $this->mRevisionId ) ) { 05536 return null; 05537 } 05538 05539 $this->mRevisionObject = Revision::newFromId( $this->mRevisionId ); 05540 return $this->mRevisionObject; 05541 } 05542 05547 function getRevisionTimestamp() { 05548 if ( is_null( $this->mRevisionTimestamp ) ) { 05549 wfProfileIn( __METHOD__ ); 05550 05551 global $wgContLang; 05552 05553 $revObject = $this->getRevisionObject(); 05554 $timestamp = $revObject ? $revObject->getTimestamp() : wfTimestampNow(); 05555 05556 # The cryptic '' timezone parameter tells to use the site-default 05557 # timezone offset instead of the user settings. 05558 # 05559 # Since this value will be saved into the parser cache, served 05560 # to other users, and potentially even used inside links and such, 05561 # it needs to be consistent for all visitors. 05562 $this->mRevisionTimestamp = $wgContLang->userAdjust( $timestamp, '' ); 05563 05564 wfProfileOut( __METHOD__ ); 05565 } 05566 return $this->mRevisionTimestamp; 05567 } 05568 05574 function getRevisionUser() { 05575 if( is_null( $this->mRevisionUser ) ) { 05576 $revObject = $this->getRevisionObject(); 05577 05578 # if this template is subst: the revision id will be blank, 05579 # so just use the current user's name 05580 if( $revObject ) { 05581 $this->mRevisionUser = $revObject->getUserText(); 05582 } elseif( $this->ot['wiki'] || $this->mOptions->getIsPreview() ) { 05583 $this->mRevisionUser = $this->getUser()->getName(); 05584 } 05585 } 05586 return $this->mRevisionUser; 05587 } 05588 05594 public function setDefaultSort( $sort ) { 05595 $this->mDefaultSort = $sort; 05596 $this->mOutput->setProperty( 'defaultsort', $sort ); 05597 } 05598 05609 public function getDefaultSort() { 05610 if ( $this->mDefaultSort !== false ) { 05611 return $this->mDefaultSort; 05612 } else { 05613 return ''; 05614 } 05615 } 05616 05623 public function getCustomDefaultSort() { 05624 return $this->mDefaultSort; 05625 } 05626 05636 public function guessSectionNameFromWikiText( $text ) { 05637 # Strip out wikitext links(they break the anchor) 05638 $text = $this->stripSectionName( $text ); 05639 $text = Sanitizer::normalizeSectionNameWhitespace( $text ); 05640 return '#' . Sanitizer::escapeId( $text, 'noninitial' ); 05641 } 05642 05651 public function guessLegacySectionNameFromWikiText( $text ) { 05652 # Strip out wikitext links(they break the anchor) 05653 $text = $this->stripSectionName( $text ); 05654 $text = Sanitizer::normalizeSectionNameWhitespace( $text ); 05655 return '#' . Sanitizer::escapeId( $text, array( 'noninitial', 'legacy' ) ); 05656 } 05657 05672 public function stripSectionName( $text ) { 05673 # Strip internal link markup 05674 $text = preg_replace( '/\[\[:?([^[|]+)\|([^[]+)\]\]/', '$2', $text ); 05675 $text = preg_replace( '/\[\[:?([^[]+)\|?\]\]/', '$1', $text ); 05676 05677 # Strip external link markup 05678 # @todo FIXME: Not tolerant to blank link text 05679 # I.E. [http://www.mediawiki.org] will render as [1] or something depending 05680 # on how many empty links there are on the page - need to figure that out. 05681 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols . ')([^ ]+?) ([^[]+)\]/', '$2', $text ); 05682 05683 # Parse wikitext quotes (italics & bold) 05684 $text = $this->doQuotes( $text ); 05685 05686 # Strip HTML tags 05687 $text = StringUtils::delimiterReplace( '<', '>', '', $text ); 05688 return $text; 05689 } 05690 05701 function testSrvus( $text, Title $title, ParserOptions $options, $outputType = self::OT_HTML ) { 05702 $this->startParse( $title, $options, $outputType, true ); 05703 05704 $text = $this->replaceVariables( $text ); 05705 $text = $this->mStripState->unstripBoth( $text ); 05706 $text = Sanitizer::removeHTMLtags( $text ); 05707 return $text; 05708 } 05709 05716 function testPst( $text, Title $title, ParserOptions $options ) { 05717 return $this->preSaveTransform( $text, $title, $options->getUser(), $options ); 05718 } 05719 05726 function testPreprocess( $text, Title $title, ParserOptions $options ) { 05727 return $this->testSrvus( $text, $title, $options, self::OT_PREPROCESS ); 05728 } 05729 05746 function markerSkipCallback( $s, $callback ) { 05747 $i = 0; 05748 $out = ''; 05749 while ( $i < strlen( $s ) ) { 05750 $markerStart = strpos( $s, $this->mUniqPrefix, $i ); 05751 if ( $markerStart === false ) { 05752 $out .= call_user_func( $callback, substr( $s, $i ) ); 05753 break; 05754 } else { 05755 $out .= call_user_func( $callback, substr( $s, $i, $markerStart - $i ) ); 05756 $markerEnd = strpos( $s, self::MARKER_SUFFIX, $markerStart ); 05757 if ( $markerEnd === false ) { 05758 $out .= substr( $s, $markerStart ); 05759 break; 05760 } else { 05761 $markerEnd += strlen( self::MARKER_SUFFIX ); 05762 $out .= substr( $s, $markerStart, $markerEnd - $markerStart ); 05763 $i = $markerEnd; 05764 } 05765 } 05766 } 05767 return $out; 05768 } 05769 05776 function killMarkers( $text ) { 05777 return $this->mStripState->killMarkers( $text ); 05778 } 05779 05796 function serializeHalfParsedText( $text ) { 05797 wfProfileIn( __METHOD__ ); 05798 $data = array( 05799 'text' => $text, 05800 'version' => self::HALF_PARSED_VERSION, 05801 'stripState' => $this->mStripState->getSubState( $text ), 05802 'linkHolders' => $this->mLinkHolders->getSubArray( $text ) 05803 ); 05804 wfProfileOut( __METHOD__ ); 05805 return $data; 05806 } 05807 05823 function unserializeHalfParsedText( $data ) { 05824 if ( !isset( $data['version'] ) || $data['version'] != self::HALF_PARSED_VERSION ) { 05825 throw new MWException( __METHOD__.': invalid version' ); 05826 } 05827 05828 # First, extract the strip state. 05829 $texts = array( $data['text'] ); 05830 $texts = $this->mStripState->merge( $data['stripState'], $texts ); 05831 05832 # Now renumber links 05833 $texts = $this->mLinkHolders->mergeForeign( $data['linkHolders'], $texts ); 05834 05835 # Should be good to go. 05836 return $texts[0]; 05837 } 05838 05848 function isValidHalfParsedText( $data ) { 05849 return isset( $data['version'] ) && $data['version'] == self::HALF_PARSED_VERSION; 05850 } 05851 05860 public function parseWidthParam( $value ) { 05861 $parsedWidthParam = array(); 05862 if( $value === '' ) { 05863 return $parsedWidthParam; 05864 } 05865 $m = array(); 05866 # (bug 13500) In both cases (width/height and width only), 05867 # permit trailing "px" for backward compatibility. 05868 if ( preg_match( '/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/', $value, $m ) ) { 05869 $width = intval( $m[1] ); 05870 $height = intval( $m[2] ); 05871 $parsedWidthParam['width'] = $width; 05872 $parsedWidthParam['height'] = $height; 05873 } elseif ( preg_match( '/^[0-9]*\s*(?:px)?\s*$/', $value ) ) { 05874 $width = intval( $value ); 05875 $parsedWidthParam['width'] = $width; 05876 } 05877 return $parsedWidthParam; 05878 } 05879 }