X-Git-Url: https://git.heureux-cyclage.org/?a=blobdiff_plain;f=includes%2FParser.php;h=ea02a6e6199d53464f06d7cf57bd02bc6ec55613;hb=e71a915538d99506098253bbc187bb65f9001904;hp=740dbd25187173f6f0b9a9182d5fdfe898727fec;hpb=0be61ee338f4f34f8c67cee8bec3456f4f7d5383;p=lhc%2Fweb%2Fwiklou.git diff --git a/includes/Parser.php b/includes/Parser.php index 740dbd2518..ea02a6e619 100644 --- a/includes/Parser.php +++ b/includes/Parser.php @@ -7,55 +7,6 @@ * @addtogroup Parser */ -/** - * Update this version number when the ParserOutput format - * changes in an incompatible way, so the parser cache - * can automatically discard old data. - */ -define( 'MW_PARSER_VERSION', '1.6.1' ); - -define( 'RLH_FOR_UPDATE', 1 ); - -# Allowed values for $mOutputType -define( 'OT_HTML', 1 ); -define( 'OT_WIKI', 2 ); -define( 'OT_MSG' , 3 ); -define( 'OT_PREPROCESS', 4 ); - -# Flags for setFunctionHook -define( 'SFH_NO_HASH', 1 ); - -# string parameter for extractTags which will cause it -# to strip HTML comments in addition to regular -# -style tags. This should not be anything we -# may want to use in wikisyntax -define( 'STRIP_COMMENTS', 'HTMLCommentStrip' ); - -# Constants needed for external link processing -define( 'HTTP_PROTOCOLS', 'http:\/\/|https:\/\/' ); -# Everything except bracket, space, or control characters -define( 'EXT_LINK_URL_CLASS', '[^][<>"\\x00-\\x20\\x7F]' ); -# Including space, but excluding newlines -define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x0a\\x0d]' ); -define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' ); -define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' ); -define( 'EXT_LINK_BRACKETED', '/\[(\b(' . wfUrlProtocols() . ')'. - EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' ); -define( 'EXT_IMAGE_REGEX', - '/^('.HTTP_PROTOCOLS.')'. # Protocol - '('.EXT_LINK_URL_CLASS.'+)\\/'. # Hostname and path - '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename -); - -// State constants for the definition list colon extraction -define( 'MW_COLON_STATE_TEXT', 0 ); -define( 'MW_COLON_STATE_TAG', 1 ); -define( 'MW_COLON_STATE_TAGSTART', 2 ); -define( 'MW_COLON_STATE_CLOSETAG', 3 ); -define( 'MW_COLON_STATE_TAGSLASH', 4 ); -define( 'MW_COLON_STATE_COMMENT', 5 ); -define( 'MW_COLON_STATE_COMMENTDASH', 6 ); -define( 'MW_COLON_STATE_COMMENTDASHDASH', 7 ); /** * PHP Parser - Processes wiki markup (which uses a more user-friendly @@ -92,22 +43,54 @@ define( 'MW_COLON_STATE_COMMENTDASHDASH', 7 ); */ class Parser { - const VERSION = MW_PARSER_VERSION; + /** + * Update this version number when the ParserOutput format + * changes in an incompatible way, so the parser cache + * can automatically discard old data. + */ + const VERSION = '1.6.4'; + + # Flags for Parser::setFunctionHook + # Also available as global constants from Defines.php + const SFH_NO_HASH = 1; + const SFH_OBJECT_ARGS = 2; + + # Constants needed for external link processing + # Everything except bracket, space, or control characters + const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F]'; + const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)([^][<>"\\x00-\\x20\\x7F]+) + \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sx'; + + // State constants for the definition list colon extraction + const COLON_STATE_TEXT = 0; + const COLON_STATE_TAG = 1; + const COLON_STATE_TAGSTART = 2; + const COLON_STATE_CLOSETAG = 3; + const COLON_STATE_TAGSLASH = 4; + const COLON_STATE_COMMENT = 5; + const COLON_STATE_COMMENTDASH = 6; + const COLON_STATE_COMMENTDASHDASH = 7; + + // Flags for preprocessToDom + const PTD_FOR_INCLUSION = 1; + /**#@+ * @private */ # Persistent: - var $mTagHooks, $mFunctionHooks, $mFunctionSynonyms, $mVariables; + var $mTagHooks, $mTransparentTagHooks, $mFunctionHooks, $mFunctionSynonyms, $mVariables, + $mImageParams, $mImageParamsMagicArray, $mStripList, $mMarkerSuffix, + $mExtLinkBracketedRegex; # Cleared with clearState(): var $mOutput, $mAutonumber, $mDTopen, $mStripState; var $mIncludeCount, $mArgStack, $mLastSection, $mInPre; - var $mInterwikiLinkHolders, $mLinkHolders, $mUniqPrefix; - var $mIncludeSizes, $mDefaultSort; - var $mTemplates, // cache of already loaded templates, avoids - // multiple SQL queries for the same string + var $mInterwikiLinkHolders, $mLinkHolders; + var $mIncludeSizes, $mPPNodeCount, $mDefaultSort; + var $mTplExpandCache,// empty-frame expansion cache $mTemplatePath; // stores an unsorted hash of all the templates already loaded // in this path. Used for loop detection. + var $mTplRedirCache, $mTplDomCache, $mHeadings; # Temporary # These are variables reset at least once per parse regardless of $clearState @@ -126,10 +109,15 @@ class Parser * * @public */ - function Parser() { + function __construct( $conf = array() ) { $this->mTagHooks = array(); + $this->mTransparentTagHooks = array(); $this->mFunctionHooks = array(); $this->mFunctionSynonyms = array( 0 => array(), 1 => array() ); + $this->mStripList = array( 'nowiki', 'gallery' ); + $this->mMarkerSuffix = "-QINU\x7f"; + $this->mExtLinkBracketedRegex = '/\[(\b(' . wfUrlProtocols() . ')'. + '[^][<>"\\x00-\\x20\\x7F]+) *([^\]\\x0a\\x0d]*?)\]/S'; $this->mFirstCall = true; } @@ -143,35 +131,41 @@ class Parser wfProfileIn( __METHOD__ ); global $wgAllowDisplayTitle, $wgAllowSlowParserFunctions; - + $this->setHook( 'pre', array( $this, 'renderPreTag' ) ); - - $this->setFunctionHook( 'int', array( 'CoreParserFunctions', 'intFunction' ), SFH_NO_HASH ); - $this->setFunctionHook( 'ns', array( 'CoreParserFunctions', 'ns' ), SFH_NO_HASH ); - $this->setFunctionHook( 'urlencode', array( 'CoreParserFunctions', 'urlencode' ), SFH_NO_HASH ); - $this->setFunctionHook( 'lcfirst', array( 'CoreParserFunctions', 'lcfirst' ), SFH_NO_HASH ); - $this->setFunctionHook( 'ucfirst', array( 'CoreParserFunctions', 'ucfirst' ), SFH_NO_HASH ); - $this->setFunctionHook( 'lc', array( 'CoreParserFunctions', 'lc' ), SFH_NO_HASH ); - $this->setFunctionHook( 'uc', array( 'CoreParserFunctions', 'uc' ), SFH_NO_HASH ); - $this->setFunctionHook( 'localurl', array( 'CoreParserFunctions', 'localurl' ), SFH_NO_HASH ); - $this->setFunctionHook( 'localurle', array( 'CoreParserFunctions', 'localurle' ), SFH_NO_HASH ); - $this->setFunctionHook( 'fullurl', array( 'CoreParserFunctions', 'fullurl' ), SFH_NO_HASH ); - $this->setFunctionHook( 'fullurle', array( 'CoreParserFunctions', 'fullurle' ), SFH_NO_HASH ); - $this->setFunctionHook( 'formatnum', array( 'CoreParserFunctions', 'formatnum' ), SFH_NO_HASH ); - $this->setFunctionHook( 'grammar', array( 'CoreParserFunctions', 'grammar' ), SFH_NO_HASH ); - $this->setFunctionHook( 'plural', array( 'CoreParserFunctions', 'plural' ), SFH_NO_HASH ); - $this->setFunctionHook( 'numberofpages', array( 'CoreParserFunctions', 'numberofpages' ), SFH_NO_HASH ); - $this->setFunctionHook( 'numberofusers', array( 'CoreParserFunctions', 'numberofusers' ), SFH_NO_HASH ); + + # Syntax for arguments (see self::setFunctionHook): + # "name for lookup in localized magic words array", + # function callback, + # optional SFH_NO_HASH to omit the hash from calls (e.g. {{int:...} + # instead of {{#int:...}}) + $this->setFunctionHook( 'int', array( 'CoreParserFunctions', 'intFunction' ), SFH_NO_HASH ); + $this->setFunctionHook( 'ns', array( 'CoreParserFunctions', 'ns' ), SFH_NO_HASH ); + $this->setFunctionHook( 'urlencode', array( 'CoreParserFunctions', 'urlencode' ), SFH_NO_HASH ); + $this->setFunctionHook( 'lcfirst', array( 'CoreParserFunctions', 'lcfirst' ), SFH_NO_HASH ); + $this->setFunctionHook( 'ucfirst', array( 'CoreParserFunctions', 'ucfirst' ), SFH_NO_HASH ); + $this->setFunctionHook( 'lc', array( 'CoreParserFunctions', 'lc' ), SFH_NO_HASH ); + $this->setFunctionHook( 'uc', array( 'CoreParserFunctions', 'uc' ), SFH_NO_HASH ); + $this->setFunctionHook( 'localurl', array( 'CoreParserFunctions', 'localurl' ), SFH_NO_HASH ); + $this->setFunctionHook( 'localurle', array( 'CoreParserFunctions', 'localurle' ), SFH_NO_HASH ); + $this->setFunctionHook( 'fullurl', array( 'CoreParserFunctions', 'fullurl' ), SFH_NO_HASH ); + $this->setFunctionHook( 'fullurle', array( 'CoreParserFunctions', 'fullurle' ), SFH_NO_HASH ); + $this->setFunctionHook( 'formatnum', array( 'CoreParserFunctions', 'formatnum' ), SFH_NO_HASH ); + $this->setFunctionHook( 'grammar', array( 'CoreParserFunctions', 'grammar' ), SFH_NO_HASH ); + $this->setFunctionHook( 'plural', array( 'CoreParserFunctions', 'plural' ), SFH_NO_HASH ); + $this->setFunctionHook( 'numberofpages', array( 'CoreParserFunctions', 'numberofpages' ), SFH_NO_HASH ); + $this->setFunctionHook( 'numberofusers', array( 'CoreParserFunctions', 'numberofusers' ), SFH_NO_HASH ); $this->setFunctionHook( 'numberofarticles', array( 'CoreParserFunctions', 'numberofarticles' ), SFH_NO_HASH ); - $this->setFunctionHook( 'numberoffiles', array( 'CoreParserFunctions', 'numberoffiles' ), SFH_NO_HASH ); - $this->setFunctionHook( 'numberofadmins', array( 'CoreParserFunctions', 'numberofadmins' ), SFH_NO_HASH ); - $this->setFunctionHook( 'numberofedits', array( 'CoreParserFunctions', 'numberofedits' ), SFH_NO_HASH ); - $this->setFunctionHook( 'language', array( 'CoreParserFunctions', 'language' ), SFH_NO_HASH ); - $this->setFunctionHook( 'padleft', array( 'CoreParserFunctions', 'padleft' ), SFH_NO_HASH ); - $this->setFunctionHook( 'padright', array( 'CoreParserFunctions', 'padright' ), SFH_NO_HASH ); - $this->setFunctionHook( 'anchorencode', array( 'CoreParserFunctions', 'anchorencode' ), SFH_NO_HASH ); - $this->setFunctionHook( 'special', array( 'CoreParserFunctions', 'special' ) ); - $this->setFunctionHook( 'defaultsort', array( 'CoreParserFunctions', 'defaultsort' ), SFH_NO_HASH ); + $this->setFunctionHook( 'numberoffiles', array( 'CoreParserFunctions', 'numberoffiles' ), SFH_NO_HASH ); + $this->setFunctionHook( 'numberofadmins', array( 'CoreParserFunctions', 'numberofadmins' ), SFH_NO_HASH ); + $this->setFunctionHook( 'numberofedits', array( 'CoreParserFunctions', 'numberofedits' ), SFH_NO_HASH ); + $this->setFunctionHook( 'language', array( 'CoreParserFunctions', 'language' ), SFH_NO_HASH ); + $this->setFunctionHook( 'padleft', array( 'CoreParserFunctions', 'padleft' ), SFH_NO_HASH ); + $this->setFunctionHook( 'padright', array( 'CoreParserFunctions', 'padright' ), SFH_NO_HASH ); + $this->setFunctionHook( 'anchorencode', array( 'CoreParserFunctions', 'anchorencode' ), SFH_NO_HASH ); + $this->setFunctionHook( 'special', array( 'CoreParserFunctions', 'special' ) ); + $this->setFunctionHook( 'defaultsort', array( 'CoreParserFunctions', 'defaultsort' ), SFH_NO_HASH ); + $this->setFunctionHook( 'filepath', array( 'CoreParserFunctions', 'filepath' ), SFH_NO_HASH ); if ( $wgAllowDisplayTitle ) { $this->setFunctionHook( 'displaytitle', array( 'CoreParserFunctions', 'displaytitle' ), SFH_NO_HASH ); @@ -201,7 +195,7 @@ class Parser $this->mDTopen = false; $this->mIncludeCount = array(); $this->mStripState = new StripState; - $this->mArgStack = array(); + $this->mArgStack = false; $this->mInPre = false; $this->mInterwikiLinkHolders = array( 'texts' => array(), @@ -222,21 +216,26 @@ class Parser * Using it at the front also gives us a little extra robustness * since it shouldn't match when butted up against identifier-like * string constructs. + * + * Must not consist of all title characters, or else it will change + * the behaviour of in a link. */ - $this->mUniqPrefix = "\x07UNIQ" . Parser::getRandomString(); + #$this->mUniqPrefix = "\x07UNIQ" . Parser::getRandomString(); + $this->mUniqPrefix = "\x7fUNIQ" . Parser::getRandomString(); # Clear these on every parse, bug 4549 - $this->mTemplates = array(); $this->mTemplatePath = array(); + $this->mTplExpandCache = $this->mTplRedirCache = $this->mTplDomCache = array(); $this->mShowToc = true; $this->mForceTocPosition = false; $this->mIncludeSizes = array( - 'pre-expand' => 0, 'post-expand' => 0, - 'arg' => 0 + 'arg' => 0, ); + $this->mPPNodeCount = 0; $this->mDefaultSort = false; + $this->mHeadings = array(); wfRunHooks( 'ParserClearState', array( &$this ) ); wfProfileOut( __METHOD__ ); @@ -259,6 +258,15 @@ class Parser * @public */ function uniqPrefix() { + if( !isset( $this->mUniqPrefix ) ) { + // @fixme this is probably *horribly wrong* + // LanguageConverter seems to want $wgParser's uniqPrefix, however + // if this is called for a parser cache hit, the parser may not + // have ever been initialized in the first place. + // Not really sure what the heck is supposed to be going on here. + return ''; + //throw new MWException( "Accessing uninitialized mUniqPrefix" ); + } return $this->mUniqPrefix; } @@ -299,7 +307,7 @@ class Parser } $this->setOutputType( OT_HTML ); wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) ); - $text = $this->strip( $text, $this->mStripState ); + # No more strip! wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) ); $text = $this->internalParse( $text ); $text = $this->mStripState->unstripGeneral( $text ); @@ -308,7 +316,7 @@ class Parser $fixtags = array( # french spaces, last one Guillemet-left # only if there is something before the space - '/(.) (?=\\?|:|;|!|\\302\\273)/' => '\\1 \\2', + '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 \\2', # french spaces, Guillemet-right '/(\\302\\253) /' => '\\1 ', ); @@ -329,6 +337,26 @@ class Parser wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) ); +//!JF Move to its own function + + $uniq_prefix = $this->mUniqPrefix; + $matches = array(); + $elements = array_keys( $this->mTransparentTagHooks ); + $text = Parser::extractTagsAndParams( $elements, $text, $matches, $uniq_prefix ); + + foreach( $matches as $marker => $data ) { + list( $element, $content, $params, $tag ) = $data; + $tagName = strtolower( $element ); + if( isset( $this->mTransparentTagHooks[$tagName] ) ) { + $output = call_user_func_array( $this->mTransparentTagHooks[$tagName], + array( $content, $params, $this ) ); + } else { + $output = $tag; + } + $this->mStripState->general->setPair( $marker, $output ); + } + $text = $this->mStripState->unstripGeneral( $text ); + $text = Sanitizer::normalizeCharReferences( $text ); if (($wgUseTidy and $this->mOptions->mTidy) or $wgAlwaysUseTidy) { @@ -364,14 +392,14 @@ class Parser wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) ); # Information on include size limits, for the benefit of users who try to skirt them - if ( max( $this->mIncludeSizes ) > 1000 ) { + if ( $this->mOptions->getEnableLimitReport() ) { $max = $this->mOptions->getMaxIncludeSize(); - $text .= "\n"; + $limitReport = + "Preprocessor node count: {$this->mPPNodeCount}/{$this->mOptions->mMaxPPNodeCount}\n" . + "Post-expand include size: {$this->mIncludeSizes['post-expand']}/$max bytes\n" . + "Template argument size: {$this->mIncludeSizes['arg']}/$max bytes\n"; + wfRunHooks( 'ParserLimitReport', array( $this, &$limitReport ) ); + $text .= "\n\n"; } $this->mOutput->setText( $text ); $this->mRevisionId = $oldRevisionId; @@ -389,7 +417,6 @@ class Parser function recursiveTagParse( $text ) { wfProfileIn( __METHOD__ ); wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) ); - $text = $this->strip( $text, $this->mStripState ); wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) ); $text = $this->internalParse( $text ); wfProfileOut( __METHOD__ ); @@ -400,19 +427,21 @@ class Parser * Expand templates and variables in the text, producing valid, static wikitext. * Also removes comments. */ - function preprocess( $text, $title, $options ) { + function preprocess( $text, $title, $options, $revid = null ) { wfProfileIn( __METHOD__ ); $this->clearState(); $this->setOutputType( OT_PREPROCESS ); $this->mOptions = $options; $this->mTitle = $title; + if( $revid !== null ) { + $this->mRevisionId = $revid; + } wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) ); - $text = $this->strip( $text, $this->mStripState ); wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) ); + $text = $this->replaceVariables( $text ); if ( $this->mOptions->getRemoveComments() ) { $text = Sanitizer::removeHTMLcomments( $text ); } - $text = $this->replaceVariables( $text ); $text = $this->mStripState->unstripBoth( $text ); wfProfileOut( __METHOD__ ); return $text; @@ -451,7 +480,7 @@ class Parser * @param $text Source text string. * @param $uniq_prefix * - * @private + * @public * @static */ function extractTagsAndParams($elements, $text, &$matches, $uniq_prefix = ''){ @@ -482,7 +511,7 @@ class Parser $inside = $p[4]; } - $marker = "$uniq_prefix-$element-" . sprintf('%08X', $n++) . '-QINU'; + $marker = "$uniq_prefix-$element-" . sprintf('%08X', $n++) . $this->mMarkerSuffix; $stripped .= $marker; if ( $close === '/>' ) { @@ -517,124 +546,24 @@ class Parser } /** - * Strips and renders nowiki, pre, math, hiero - * If $render is set, performs necessary rendering operations on plugins - * Returns the text, and fills an array with data needed in unstrip() - * - * @param StripState $state - * - * @param bool $stripcomments when set, HTML comments - * will be stripped in addition to other tags. This is important - * for section editing, where these comments cause confusion when - * counting the sections in the wikisource - * - * @param array dontstrip contains tags which should not be stripped; - * used to prevent stipping of when saving (fixes bug 2700) - * - * @private + * Get a list of strippable XML-like elements */ - function strip( $text, $state, $stripcomments = false , $dontstrip = array () ) { - global $wgContLang; - wfProfileIn( __METHOD__ ); - $render = ($this->mOutputType == OT_HTML); - - $uniq_prefix = $this->mUniqPrefix; - $commentState = new ReplacementArray; - $nowikiItems = array(); - $generalItems = array(); - - $elements = array_merge( - array( 'nowiki', 'gallery' ), - array_keys( $this->mTagHooks ) ); + function getStripList() { global $wgRawHtml; + $elements = $this->mStripList; if( $wgRawHtml ) { $elements[] = 'html'; } if( $this->mOptions->getUseTeX() ) { $elements[] = 'math'; } + return $elements; + } - # Removing $dontstrip tags from $elements list (currently only 'gallery', fixing bug 2700) - foreach ( $elements AS $k => $v ) { - if ( !in_array ( $v , $dontstrip ) ) continue; - unset ( $elements[$k] ); - } - - $matches = array(); - $text = Parser::extractTagsAndParams( $elements, $text, $matches, $uniq_prefix ); - - foreach( $matches as $marker => $data ) { - list( $element, $content, $params, $tag ) = $data; - if( $render ) { - $tagName = strtolower( $element ); - wfProfileIn( __METHOD__."-render-$tagName" ); - switch( $tagName ) { - case '!--': - // Comment - if( substr( $tag, -3 ) == '-->' ) { - $output = $tag; - } else { - // Unclosed comment in input. - // Close it so later stripping can remove it - $output = "$tag-->"; - } - break; - case 'html': - if( $wgRawHtml ) { - $output = $content; - break; - } - // Shouldn't happen otherwise. :) - case 'nowiki': - $output = Xml::escapeTagsOnly( $content ); - break; - case 'math': - $output = $wgContLang->armourMath( MathRenderer::renderMath( $content ) ); - break; - case 'gallery': - $output = $this->renderImageGallery( $content, $params ); - break; - default: - if( isset( $this->mTagHooks[$tagName] ) ) { - $output = call_user_func_array( $this->mTagHooks[$tagName], - array( $content, $params, $this ) ); - } else { - throw new MWException( "Invalid call hook $element" ); - } - } - wfProfileOut( __METHOD__."-render-$tagName" ); - } else { - // Just stripping tags; keep the source - $output = $tag; - } - - // Unstrip the output, to support recursive strip() calls - $output = $state->unstripBoth( $output ); - - if( !$stripcomments && $element == '!--' ) { - $commentState->setPair( $marker, $output ); - } elseif ( $element == 'html' || $element == 'nowiki' ) { - $nowikiItems[$marker] = $output; - } else { - $generalItems[$marker] = $output; - } - } - # Add the new items to the state - # We do this after the loop instead of during it to avoid slowing - # down the recursive unstrip - $state->nowiki->mergeArray( $nowikiItems ); - $state->general->mergeArray( $generalItems ); - - # Unstrip comments unless explicitly told otherwise. - # (The comments are always stripped prior to this point, so as to - # not invoke any extension tags / parser hooks contained within - # a comment.) - if ( !$stripcomments ) { - // Put them all back and forget them - $text = $commentState->replace( $text ); - } - - wfProfileOut( __METHOD__ ); + /** + * @deprecated use replaceVariables + */ + function strip( $text, $state, $stripcomments = false , $dontstrip = array () ) { return $text; } @@ -673,9 +602,11 @@ class Parser * * @private */ - function insertStripItem( $text, &$state ) { - $rnd = $this->mUniqPrefix . '-item' . Parser::getRandomString(); - $state->general->setPair( $rnd, $text ); + function insertStripItem( $text ) { + static $n = 0; + $rnd = "{$this->mUniqPrefix}-item-$n-{$this->mMarkerSuffix}"; + ++$n; + $this->mStripState->general->setPair( $rnd, $text ); return $rnd; } @@ -727,7 +658,7 @@ class Parser $descriptorspec = array( 0 => array('pipe', 'r'), 1 => array('pipe', 'w'), - 2 => array('file', '/dev/null', 'a') // FIXME: this line in UNIX-specific, it generates a warning on Windows, because /dev/null is not a valid Windows file. + 2 => array('file', wfGetNull(), 'a') ); $pipes = array(); $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes); @@ -759,8 +690,7 @@ class Parser /** * Use the HTML tidy PECL extension to use the tidy library in-process, - * saving the overhead of spawning a new process. Currently written to - * the PHP 4.3.x version of the extension, may not work on PHP 5. + * saving the overhead of spawning a new process. * * 'pear install tidy' should be able to compile the extension module. * @@ -768,21 +698,26 @@ class Parser * @static */ function internalTidy( $text ) { - global $wgTidyConf; + global $wgTidyConf, $IP, $wgDebugTidy; $fname = 'Parser::internalTidy'; wfProfileIn( $fname ); - tidy_load_config( $wgTidyConf ); - tidy_set_encoding( 'utf8' ); - tidy_parse_string( $text ); - tidy_clean_repair(); - if( tidy_get_status() == 2 ) { + $tidy = new tidy; + $tidy->parseString( $text, $wgTidyConf, 'utf8' ); + $tidy->cleanRepair(); + if( $tidy->getStatus() == 2 ) { // 2 is magic number for fatal error // http://www.php.net/manual/en/function.tidy-get-status.php $cleansource = null; } else { - $cleansource = tidy_get_output(); + $cleansource = tidy_get_output( $tidy ); + } + if ( $wgDebugTidy && $tidy->getStatus() > 0 ) { + $cleansource .= "', '-->', $tidy->errorBuffer ) . + "\n-->"; } + wfProfileOut( $fname ); return $cleansource; } @@ -986,7 +921,6 @@ class Parser * @private */ function internalParse( $text ) { - $args = array(); $isMain = true; $fname = 'Parser::internalParse'; wfProfileIn( $fname ); @@ -997,14 +931,8 @@ class Parser return $text ; } - # Remove tags and sections - $text = strtr( $text, array( '' => '' , '' => '' ) ); - $text = strtr( $text, array( '' => '', '' => '') ); - $text = StringUtils::delimiterReplace( '', '', '', $text ); - - $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ) ); - - $text = $this->replaceVariables( $text, $args ); + $text = $this->replaceVariables( $text ); + $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ), false, array_keys( $this->mTransparentTagHooks ) ); wfRunHooks( 'InternalParseBeforeLinks', array( &$this, &$text, &$this->mStripState ) ); // Tables need to come after variable replacement for things to work @@ -1043,7 +971,7 @@ class Parser * * @private */ - function &doMagicLinks( &$text ) { + function doMagicLinks( $text ) { wfProfileIn( __METHOD__ ); $text = preg_replace_callback( '!(?: # Start cases @@ -1107,8 +1035,8 @@ class Parser wfProfileIn( $fname ); for ( $i = 6; $i >= 1; --$i ) { $h = str_repeat( '=', $i ); - $text = preg_replace( "/^{$h}(.+){$h}\\s*$/m", - "\\1\\2", $text ); + $text = preg_replace( "/^$h(.+)$h\\s*$/m", + "\\1", $text ); } wfProfileOut( $fname ); return $text; @@ -1134,9 +1062,8 @@ class Parser /** * Helper function for doAllQuotes() - * @private */ - function doQuotes( $text ) { + public function doQuotes( $text ) { $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE ); if ( count( $arr ) == 1 ) return $text; @@ -1313,7 +1240,7 @@ class Parser $sk = $this->mOptions->getSkin(); - $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE ); + $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE ); $s = $this->replaceFreeExternalLinks( array_shift( $bits ) ); @@ -1407,7 +1334,7 @@ class Parser $remainder = $bits[$i++]; $m = array(); - if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) { + if ( preg_match( '/^('.self::EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) { # Found some characters after the protocol that look promising $url = $protocol . $m[1]; $trail = $m[2]; @@ -1417,7 +1344,7 @@ class Parser if(strlen($trail) == 0 && isset($bits[$i]) && preg_match('/^'. wfUrlProtocols() . '$/S', $bits[$i]) && - preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $bits[$i + 1], $m )) + preg_match( '/^('.self::EXT_LINK_URL_CLASS.'+)(.*)$/s', $bits[$i + 1], $m )) { # add protocol, arg $url .= $bits[$i] . $m[1]; # protocol, url as arg to previous link @@ -1514,7 +1441,7 @@ class Parser $text = false; if ( $this->mOptions->getAllowExternalImages() || ( $imagesexception && strpos( $url, $imagesfrom ) === 0 ) ) { - if ( preg_match( EXT_IMAGE_REGEX, $url ) ) { + if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) { # Image found $text = $sk->makeExternalImage( htmlspecialchars( $url ) ); } @@ -1552,11 +1479,15 @@ class Parser # Match cases where there is no "]]", which might still be images static $e1_img = FALSE; if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; } - # Match the end of a line for a word that's not followed by whitespace, - # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched - $e2 = wfMsgForContent( 'linkprefix' ); $useLinkPrefixExtension = $wgContLang->linkPrefixExtension(); + $e2 = null; + if ( $useLinkPrefixExtension ) { + # Match the end of a line for a word that's not followed by whitespace, + # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched + $e2 = wfMsgForContent( 'linkprefix' ); + } + if( is_null( $this->mTitle ) ) { throw new MWException( __METHOD__.": \$this->mTitle is null\n" ); } @@ -1799,11 +1730,15 @@ class Parser $this->mOutput->addImage( $nt->getDBkey() ); continue; } elseif( $ns == NS_SPECIAL ) { - $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix ); + if( SpecialPage::exists( $nt->getDBkey() ) ) { + $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix ); + } else { + $s .= $this->makeLinkHolder( $nt, $text, '', $trail, $prefix ); + } continue; } elseif( $ns == NS_IMAGE ) { - $img = new Image( $nt ); - if( $img->exists() ) { + $img = wfFindFile( $nt ); + if( $img ) { // Force a blue link if the file exists; may be a remote // upload on the shared repository, and we want to see its // auto-generated page. @@ -1920,12 +1855,18 @@ class Parser wfProfileIn( $fname ); $ret = $target; # default return value is no change - # bug 7425 - $target = trim( $target ); - # Some namespaces don't allow subpages, # so only perform processing if subpages are allowed if( $this->areSubpagesAllowed() ) { + $hash = strpos( $target, '#' ); + if( $hash !== false ) { + $suffix = substr( $target, $hash ); + $target = substr( $target, 0, $hash ); + } else { + $suffix = ''; + } + # bug 7425 + $target = trim( $target ); # Look at the first character if( $target != '' && $target{0} == '/' ) { # / at end means we don't want the slash to be shown @@ -1937,9 +1878,9 @@ class Parser $noslash = substr( $target, 1 ); } - $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash); + $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash) . $suffix; if( '' === $text ) { - $text = $target; + $text = $target . $suffix; } # this might be changed for ugliness reasons } else { # check for .. subpage backlinks @@ -1957,13 +1898,14 @@ class Parser if( substr( $nodotdot, -1, 1 ) == '/' ) { $nodotdot = substr( $nodotdot, 0, -1 ); if( '' === $text ) { - $text = $nodotdot; + $text = $nodotdot . $suffix; } } $nodotdot = trim( $nodotdot ); if( $nodotdot != '' ) { $ret .= '/' . $nodotdot; } + $ret .= $suffix; } } } @@ -2246,7 +2188,7 @@ class Parser } // Ugly state machine to walk through avoiding tags. - $state = MW_COLON_STATE_TEXT; + $state = self::COLON_STATE_TEXT; $stack = 0; $len = strlen( $str ); for( $i = 0; $i < $len; $i++ ) { @@ -2254,11 +2196,11 @@ class Parser switch( $state ) { // (Using the number is a performance hack for common cases) - case 0: // MW_COLON_STATE_TEXT: + case 0: // self::COLON_STATE_TEXT: switch( $c ) { case "<": // Could be either a tag or an tag - $state = MW_COLON_STATE_TAGSTART; + $state = self::COLON_STATE_TAGSTART; break; case ":": if( $stack == 0 ) { @@ -2295,41 +2237,41 @@ class Parser } // Skip ahead to next tag start $i = $lt; - $state = MW_COLON_STATE_TAGSTART; + $state = self::COLON_STATE_TAGSTART; } break; - case 1: // MW_COLON_STATE_TAG: + case 1: // self::COLON_STATE_TAG: // In a switch( $c ) { case ">": $stack++; - $state = MW_COLON_STATE_TEXT; + $state = self::COLON_STATE_TEXT; break; case "/": // Slash may be followed by >? - $state = MW_COLON_STATE_TAGSLASH; + $state = self::COLON_STATE_TAGSLASH; break; default: // ignore } break; - case 2: // MW_COLON_STATE_TAGSTART: + case 2: // self::COLON_STATE_TAGSTART: switch( $c ) { case "/": - $state = MW_COLON_STATE_CLOSETAG; + $state = self::COLON_STATE_CLOSETAG; break; case "!": - $state = MW_COLON_STATE_COMMENT; + $state = self::COLON_STATE_COMMENT; break; case ">": // Illegal early close? This shouldn't happen D: - $state = MW_COLON_STATE_TEXT; + $state = self::COLON_STATE_TEXT; break; default: - $state = MW_COLON_STATE_TAG; + $state = self::COLON_STATE_TAG; } break; - case 3: // MW_COLON_STATE_CLOSETAG: + case 3: // self::COLON_STATE_CLOSETAG: // In a if( $c == ">" ) { $stack--; @@ -2338,35 +2280,35 @@ class Parser wfProfileOut( $fname ); return false; } - $state = MW_COLON_STATE_TEXT; + $state = self::COLON_STATE_TEXT; } break; - case MW_COLON_STATE_TAGSLASH: + case self::COLON_STATE_TAGSLASH: if( $c == ">" ) { // Yes, a self-closed tag - $state = MW_COLON_STATE_TEXT; + $state = self::COLON_STATE_TEXT; } else { // Probably we're jumping the gun, and this is an attribute - $state = MW_COLON_STATE_TAG; + $state = self::COLON_STATE_TAG; } break; - case 5: // MW_COLON_STATE_COMMENT: + case 5: // self::COLON_STATE_COMMENT: if( $c == "-" ) { - $state = MW_COLON_STATE_COMMENTDASH; + $state = self::COLON_STATE_COMMENTDASH; } break; - case MW_COLON_STATE_COMMENTDASH: + case self::COLON_STATE_COMMENTDASH: if( $c == "-" ) { - $state = MW_COLON_STATE_COMMENTDASHDASH; + $state = self::COLON_STATE_COMMENTDASHDASH; } else { - $state = MW_COLON_STATE_COMMENT; + $state = self::COLON_STATE_COMMENT; } break; - case MW_COLON_STATE_COMMENTDASHDASH: + case self::COLON_STATE_COMMENTDASHDASH: if( $c == ">" ) { - $state = MW_COLON_STATE_TEXT; + $state = self::COLON_STATE_TEXT; } else { - $state = MW_COLON_STATE_COMMENT; + $state = self::COLON_STATE_COMMENT; } break; default: @@ -2409,6 +2351,8 @@ class Parser $oldtz = getenv( 'TZ' ); putenv( 'TZ='.$wgLocaltimezone ); } + + wfSuppressWarnings(); // E_STRICT system time bitching $localTimestamp = date( 'YmdHis', $ts ); $localMonth = date( 'm', $ts ); $localMonthName = date( 'n', $ts ); @@ -2421,20 +2365,21 @@ class Parser if ( isset( $wgLocaltimezone ) ) { putenv( 'TZ='.$oldtz ); } + wfRestoreWarnings(); switch ( $index ) { case 'currentmonth': - return $varCache[$index] = $wgContLang->formatNum( date( 'm', $ts ) ); + return $varCache[$index] = $wgContLang->formatNum( gmdate( 'm', $ts ) ); case 'currentmonthname': - return $varCache[$index] = $wgContLang->getMonthName( date( 'n', $ts ) ); + return $varCache[$index] = $wgContLang->getMonthName( gmdate( 'n', $ts ) ); case 'currentmonthnamegen': - return $varCache[$index] = $wgContLang->getMonthNameGen( date( 'n', $ts ) ); + return $varCache[$index] = $wgContLang->getMonthNameGen( gmdate( 'n', $ts ) ); case 'currentmonthabbrev': - return $varCache[$index] = $wgContLang->getMonthAbbreviation( date( 'n', $ts ) ); + return $varCache[$index] = $wgContLang->getMonthAbbreviation( gmdate( 'n', $ts ) ); case 'currentday': - return $varCache[$index] = $wgContLang->formatNum( date( 'j', $ts ) ); + return $varCache[$index] = $wgContLang->formatNum( gmdate( 'j', $ts ) ); case 'currentday2': - return $varCache[$index] = $wgContLang->formatNum( date( 'd', $ts ) ); + return $varCache[$index] = $wgContLang->formatNum( gmdate( 'd', $ts ) ); case 'localmonth': return $varCache[$index] = $wgContLang->formatNum( $localMonth ); case 'localmonthname': @@ -2448,25 +2393,25 @@ class Parser case 'localday2': return $varCache[$index] = $wgContLang->formatNum( $localDay2 ); case 'pagename': - return $this->mTitle->getText(); + return wfEscapeWikiText( $this->mTitle->getText() ); case 'pagenamee': return $this->mTitle->getPartialURL(); case 'fullpagename': - return $this->mTitle->getPrefixedText(); + return wfEscapeWikiText( $this->mTitle->getPrefixedText() ); case 'fullpagenamee': return $this->mTitle->getPrefixedURL(); case 'subpagename': - return $this->mTitle->getSubpageText(); + return wfEscapeWikiText( $this->mTitle->getSubpageText() ); case 'subpagenamee': return $this->mTitle->getSubpageUrlForm(); case 'basepagename': - return $this->mTitle->getBaseText(); + return wfEscapeWikiText( $this->mTitle->getBaseText() ); case 'basepagenamee': return wfUrlEncode( str_replace( ' ', '_', $this->mTitle->getBaseText() ) ); case 'talkpagename': if( $this->mTitle->canTalk() ) { $talkPage = $this->mTitle->getTalkPage(); - return $talkPage->getPrefixedText(); + return wfEscapeWikiText( $talkPage->getPrefixedText() ); } else { return ''; } @@ -2479,21 +2424,45 @@ class Parser } case 'subjectpagename': $subjPage = $this->mTitle->getSubjectPage(); - return $subjPage->getPrefixedText(); + return wfEscapeWikiText( $subjPage->getPrefixedText() ); case 'subjectpagenamee': $subjPage = $this->mTitle->getSubjectPage(); return $subjPage->getPrefixedUrl(); case 'revisionid': + // Let the edit saving system know we should parse the page + // *after* a revision ID has been assigned. + $this->mOutput->setFlag( 'vary-revision' ); + wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision...\n" ); return $this->mRevisionId; case 'revisionday': + // Let the edit saving system know we should parse the page + // *after* a revision ID has been assigned. This is for null edits. + $this->mOutput->setFlag( 'vary-revision' ); + wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" ); return intval( substr( $this->getRevisionTimestamp(), 6, 2 ) ); case 'revisionday2': + // Let the edit saving system know we should parse the page + // *after* a revision ID has been assigned. This is for null edits. + $this->mOutput->setFlag( 'vary-revision' ); + wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" ); return substr( $this->getRevisionTimestamp(), 6, 2 ); case 'revisionmonth': + // Let the edit saving system know we should parse the page + // *after* a revision ID has been assigned. This is for null edits. + $this->mOutput->setFlag( 'vary-revision' ); + wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" ); return intval( substr( $this->getRevisionTimestamp(), 4, 2 ) ); case 'revisionyear': + // Let the edit saving system know we should parse the page + // *after* a revision ID has been assigned. This is for null edits. + $this->mOutput->setFlag( 'vary-revision' ); + wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" ); return substr( $this->getRevisionTimestamp(), 0, 4 ); case 'revisiontimestamp': + // Let the edit saving system know we should parse the page + // *after* a revision ID has been assigned. This is for null edits. + $this->mOutput->setFlag( 'vary-revision' ); + wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" ); return $this->getRevisionTimestamp(); case 'namespace': return str_replace('_',' ',$wgContLang->getNsText( $this->mTitle->getNamespace() ) ); @@ -2508,19 +2477,19 @@ class Parser case 'subjectspacee': return( wfUrlencode( $this->mTitle->getSubjectNsText() ) ); case 'currentdayname': - return $varCache[$index] = $wgContLang->getWeekdayName( date( 'w', $ts ) + 1 ); + return $varCache[$index] = $wgContLang->getWeekdayName( gmdate( 'w', $ts ) + 1 ); case 'currentyear': - return $varCache[$index] = $wgContLang->formatNum( date( 'Y', $ts ), true ); + return $varCache[$index] = $wgContLang->formatNum( gmdate( 'Y', $ts ), true ); case 'currenttime': return $varCache[$index] = $wgContLang->time( wfTimestamp( TS_MW, $ts ), false, false ); case 'currenthour': - return $varCache[$index] = $wgContLang->formatNum( date( 'H', $ts ), true ); + return $varCache[$index] = $wgContLang->formatNum( gmdate( 'H', $ts ), true ); case 'currentweek': // @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to // int to remove the padding - return $varCache[$index] = $wgContLang->formatNum( (int)date( 'W', $ts ) ); + return $varCache[$index] = $wgContLang->formatNum( (int)gmdate( 'W', $ts ) ); case 'currentdow': - return $varCache[$index] = $wgContLang->formatNum( date( 'w', $ts ) ); + return $varCache[$index] = $wgContLang->formatNum( gmdate( 'w', $ts ) ); case 'localdayname': return $varCache[$index] = $wgContLang->getWeekdayName( $localDayOfWeek + 1 ); case 'localyear': @@ -2544,9 +2513,9 @@ class Parser case 'numberofpages': return $varCache[$index] = $wgContLang->formatNum( SiteStats::pages() ); case 'numberofadmins': - return $varCache[$index] = $wgContLang->formatNum( SiteStats::admins() ); + return $varCache[$index] = $wgContLang->formatNum( SiteStats::admins() ); case 'numberofedits': - return $varCache[$index] = $wgContLang->formatNum( SiteStats::edits() ); + return $varCache[$index] = $wgContLang->formatNum( SiteStats::edits() ); case 'currenttimestamp': return $varCache[$index] = wfTimestampNow(); case 'localtimestamp': @@ -2585,187 +2554,579 @@ class Parser wfProfileIn( $fname ); $variableIDs = MagicWord::getVariableIDs(); - $this->mVariables = array(); - foreach ( $variableIDs as $id ) { - $mw =& MagicWord::get( $id ); - $mw->addToArray( $this->mVariables, $id ); - } + $this->mVariables = new MagicWordArray( $variableIDs ); wfProfileOut( $fname ); } /** - * parse any parentheses in format ((title|part|part)) - * and call callbacks to get a replacement text for any found piece + * Preprocess some wikitext and return the document tree. + * This is the ghost of replace_variables(). * * @param string $text The text to parse - * @param array $callbacks rules in form: - * '{' => array( # opening parentheses - * 'end' => '}', # closing parentheses - * 'cb' => array(2 => callback, # replacement callback to call if {{..}} is found - * 3 => callback # replacement callback to call if {{{..}}} is found - * ) - * ) - * 'min' => 2, # Minimum parenthesis count in cb - * 'max' => 3, # Maximum parenthesis count in cb + * @param integer flags Bitwise combination of: + * self::PTD_FOR_INCLUSION Handle / as if the text is being + * included. Default is to assume a direct page view. + * + * The generated DOM tree must depend only on the input text, the flags, and $this->ot['msg']. + * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899. + * + * Any flag added to the $flags parameter here, or any other parameter liable to cause a + * change in the DOM tree for a given text, must be passed through the section identifier + * in the section edit link and thus back to extractSections(). + * + * The output of this function is currently only cached in process memory, but a persistent + * cache may be implemented at a later date which takes further advantage of these strict + * dependency requirements. + * * @private */ - function replace_callback ($text, $callbacks) { + function preprocessToDom ( $text, $flags = 0 ) { wfProfileIn( __METHOD__ ); - $openingBraceStack = array(); # this array will hold a stack of parentheses which are not closed yet - $lastOpeningBrace = -1; # last not closed parentheses + wfProfileIn( __METHOD__.'-makexml' ); + + static $msgRules, $normalRules, $inclusionSupertags, $nonInclusionSupertags; + if ( !$msgRules ) { + $msgRules = array( + '{' => array( + 'end' => '}', + 'names' => array( + 2 => 'template', + ), + 'min' => 2, + 'max' => 2, + ), + '[' => array( + 'end' => ']', + 'names' => array( 2 => null ), + 'min' => 2, + 'max' => 2, + ) + ); + $normalRules = array( + '{' => array( + 'end' => '}', + 'names' => array( + 2 => 'template', + 3 => 'tplarg', + ), + 'min' => 2, + 'max' => 3, + ), + '[' => array( + 'end' => ']', + 'names' => array( 2 => null ), + 'min' => 2, + 'max' => 2, + ) + ); + } + if ( $this->ot['msg'] ) { + $rules = $msgRules; + } else { + $rules = $normalRules; + } + $forInclusion = $flags & self::PTD_FOR_INCLUSION; + + $xmlishElements = $this->getStripList(); + $enableOnlyinclude = false; + if ( $forInclusion ) { + $ignoredTags = array( 'includeonly', '/includeonly' ); + $ignoredElements = array( 'noinclude' ); + $xmlishElements[] = 'noinclude'; + if ( strpos( $text, '' ) !== false && strpos( $text, '' ) !== false ) { + $enableOnlyinclude = true; + } + } else { + $ignoredTags = array( 'noinclude', '/noinclude', 'onlyinclude', '/onlyinclude' ); + $ignoredElements = array( 'includeonly' ); + $xmlishElements[] = 'includeonly'; + } + $xmlishRegex = implode( '|', array_merge( $xmlishElements, $ignoredTags ) ); + + // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset + $elementsRegex = "~($xmlishRegex)(?:\s|\/>|>)|(!--)~iA"; + + $stack = array(); # Stack of unclosed parentheses + $stackIndex = -1; # Stack read pointer + + $searchBase = implode( '', array_keys( $rules ) ) . '<'; + $revText = strrev( $text ); // For fast reverse searches + + $i = -1; # Input pointer, starts out pointing to a pseudo-newline before the start + $topAccum = ''; # Top level text accumulator + $accum =& $topAccum; # Current text accumulator + $findEquals = false; # True to find equals signs in arguments + $findHeading = false; # True to look at LF characters for possible headings + $findPipe = false; # True to take notice of pipe characters + $headingIndex = 1; + $noMoreGT = false; # True if there are no more greater-than (>) signs right of $i + $findOnlyinclude = $enableOnlyinclude; # True to ignore all input up to the next + + if ( $enableOnlyinclude ) { + $i = 0; + } - $validOpeningBraces = implode( '', array_keys( $callbacks ) ); + while ( true ) { + if ( $findOnlyinclude ) { + // Ignore all input up to the next + $startPos = strpos( $text, '', $i ); + if ( $startPos === false ) { + // Ignored section runs to the end + $accum .= '' . htmlspecialchars( substr( $text, $i ) ) . ''; + break; + } + $tagEndPos = $startPos + strlen( '' ); // past-the-end + $accum .= '' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i ) ) . ''; + $i = $tagEndPos; + $findOnlyinclude = false; + } - $i = 0; - while ( $i < strlen( $text ) ) { - # Find next opening brace, closing brace or pipe - if ( $lastOpeningBrace == -1 ) { - $currentClosing = ''; - $search = $validOpeningBraces; + if ( $i == -1 ) { + $found = 'line-start'; + $curChar = ''; } else { - $currentClosing = $openingBraceStack[$lastOpeningBrace]['braceEnd']; - $search = $validOpeningBraces . '|' . $currentClosing; - } - $rule = null; - $i += strcspn( $text, $search, $i ); - if ( $i < strlen( $text ) ) { - if ( $text[$i] == '|' ) { - $found = 'pipe'; - } elseif ( $text[$i] == $currentClosing ) { - $found = 'close'; - } elseif ( isset( $callbacks[$text[$i]] ) ) { - $found = 'open'; - $rule = $callbacks[$text[$i]]; + # Find next opening brace, closing brace or pipe + $search = $searchBase; + if ( $stackIndex == -1 ) { + $currentClosing = ''; + // Look for headings only at the top stack level + // Among other things, this resolves the ambiguity between = + // for headings and = for template arguments + $search .= "\n"; + } else { + $currentClosing = $stack[$stackIndex]['close']; + $search .= $currentClosing; + } + if ( $findPipe ) { + $search .= '|'; + } + if ( $findEquals ) { + $search .= '='; + } + $rule = null; + # Output literal section, advance input counter + $literalLength = strcspn( $text, $search, $i ); + if ( $literalLength > 0 ) { + $accum .= htmlspecialchars( substr( $text, $i, $literalLength ) ); + $i += $literalLength; + } + if ( $i >= strlen( $text ) ) { + if ( $currentClosing == "\n" ) { + // Do a past-the-end run to finish off the heading + $curChar = ''; + $found = 'line-end'; + } else { + # All done + break; + } } else { - # Some versions of PHP have a strcspn which stops on null characters - # Ignore and continue + $curChar = $text[$i]; + if ( $curChar == '|' ) { + $found = 'pipe'; + } elseif ( $curChar == '=' ) { + $found = 'equals'; + } elseif ( $curChar == '<' ) { + $found = 'angle'; + } elseif ( $curChar == "\n" ) { + if ( $stackIndex == -1 ) { + $found = 'line-start'; + } else { + $found = 'line-end'; + } + } elseif ( $curChar == $currentClosing ) { + $found = 'close'; + } elseif ( isset( $rules[$curChar] ) ) { + $found = 'open'; + $rule = $rules[$curChar]; + } else { + # Some versions of PHP have a strcspn which stops on null characters + # Ignore and continue + ++$i; + continue; + } + } + } + + if ( $found == 'angle' ) { + $matches = false; + // Handle + if ( $enableOnlyinclude && substr( $text, $i, strlen( '' ) ) == '' ) { + $findOnlyinclude = true; + continue; + } + + // Determine element name + if ( !preg_match( $elementsRegex, $text, $matches, 0, $i + 1 ) ) { + // Element name missing or not listed + $accum .= '<'; ++$i; continue; } - } else { - # All done - break; + // Handle comments + if ( isset( $matches[2] ) && $matches[2] == '!--' ) { + // To avoid leaving blank lines, when a comment is both preceded + // and followed by a newline (ignoring spaces), trim leading and + // trailing spaces and one of the newlines. + + // Find the end + $endPos = strpos( $text, '-->', $i + 4 ); + if ( $endPos === false ) { + // Unclosed comment in input, runs to end + $inner = substr( $text, $i ); + $accum .= '' . htmlspecialchars( $inner ) . ''; + $i = strlen( $text ); + } else { + // Search backwards for leading whitespace + $wsStart = $i ? ( $i - strspn( $revText, ' ', strlen( $text ) - $i ) ) : 0; + // Search forwards for trailing whitespace + // $wsEnd will be the position of the last space + $wsEnd = $endPos + 2 + strspn( $text, ' ', $endPos + 3 ); + // Eat the line if possible + if ( $wsStart > 0 && substr( $text, $wsStart - 1, 1 ) == "\n" + && substr( $text, $wsEnd + 1, 1 ) == "\n" ) + { + $startPos = $wsStart; + $endPos = $wsEnd + 1; + // Remove leading whitespace from the end of the accumulator + // Sanity check first though + $wsLength = $i - $wsStart; + if ( $wsLength > 0 && substr( $accum, -$wsLength ) === str_repeat( ' ', $wsLength ) ) { + $accum = substr( $accum, 0, -$wsLength ); + } + } else { + // No line to eat, just take the comment itself + $startPos = $i; + $endPos += 2; + } + + $inner = substr( $text, $startPos, $endPos - $startPos + 1 ); + $accum .= '' . htmlspecialchars( $inner ) . ''; + $i = $endPos + 1; + } + continue; + } + $name = $matches[1]; + $attrStart = $i + strlen( $name ) + 1; + + // Find end of tag + $tagEndPos = $noMoreGT ? false : strpos( $text, '>', $attrStart ); + if ( $tagEndPos === false ) { + // Infinite backtrack + // Disable tag search to prevent worst-case O(N^2) performance + $noMoreGT = true; + $accum .= '<'; + ++$i; + continue; + } + + // Handle ignored tags + if ( in_array( $name, $ignoredTags ) ) { + $accum .= '' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i + 1 ) ) . ''; + $i = $tagEndPos + 1; + continue; + } + + $tagStartPos = $i; + if ( $text[$tagEndPos-1] == '/' ) { + $attrEnd = $tagEndPos - 1; + $inner = null; + $i = $tagEndPos + 1; + $close = ''; + } else { + $attrEnd = $tagEndPos; + // Find closing tag + if ( preg_match( "/<\/$name\s*>/i", $text, $matches, PREG_OFFSET_CAPTURE, $tagEndPos + 1 ) ) { + $inner = substr( $text, $tagEndPos + 1, $matches[0][1] - $tagEndPos - 1 ); + $i = $matches[0][1] + strlen( $matches[0][0] ); + $close = '' . htmlspecialchars( $matches[0][0] ) . ''; + } else { + // No end tag -- let it run out to the end of the text. + $inner = substr( $text, $tagEndPos + 1 ); + $i = strlen( $text ); + $close = ''; + } + } + // and just become tags + if ( in_array( $name, $ignoredElements ) ) { + $accum .= '' . htmlspecialchars( substr( $text, $tagStartPos, $i - $tagStartPos ) ) + . ''; + continue; + } + + $accum .= ''; + if ( $attrEnd <= $attrStart ) { + $attr = ''; + } else { + $attr = substr( $text, $attrStart, $attrEnd - $attrStart ); + } + $accum .= '' . htmlspecialchars( $name ) . '' . + // Note that the attr element contains the whitespace between name and attribute, + // this is necessary for precise reconstruction during pre-save transform. + '' . htmlspecialchars( $attr ) . ''; + if ( $inner !== null ) { + $accum .= '' . htmlspecialchars( $inner ) . ''; + } + $accum .= $close . ''; } - if ( $found == 'open' ) { - # found opening brace, let's add it to parentheses stack - $piece = array('brace' => $text[$i], - 'braceEnd' => $rule['end'], - 'title' => '', - 'parts' => null); + elseif ( $found == 'line-start' ) { + // Is this the start of a heading? + // Line break belongs before the heading element in any case + $accum .= $curChar; + $i++; + + $count = strspn( $text, '=', $i, 6 ); + if ( $count > 0 ) { + $piece = array( + 'open' => "\n", + 'close' => "\n", + 'parts' => array( str_repeat( '=', $count ) ), + 'count' => $count ); + $stack[++$stackIndex] = $piece; + $i += $count; + $accum =& $stack[$stackIndex]['parts'][0]; + $findPipe = false; + } + } + elseif ( $found == 'line-end' ) { + $piece = $stack[$stackIndex]; + // A heading must be open, otherwise \n wouldn't have been in the search list + assert( $piece['open'] == "\n" ); + assert( $stackIndex == 0 ); + // Search back through the input to see if it has a proper close + // Do this using the reversed string since the other solutions (end anchor, etc.) are inefficient + $m = false; + $count = $piece['count']; + if ( preg_match( "/\s*(={{$count}})/A", $revText, $m, 0, strlen( $text ) - $i ) ) { + // Found match, output + $count = min( strlen( $m[1] ), $count ); + $element = "$accum"; + $headingIndex++; + } else { + // No match, no , just pass down the inner text + $element = $accum; + } + // Unwind the stack + // Headings can only occur on the top level, so this is a bit simpler than the + // generic stack unwind operation in the close case + unset( $stack[$stackIndex--] ); + $accum =& $topAccum; + $findEquals = false; + $findPipe = false; + + // Append the result to the enclosing accumulator + $accum .= $element; + // Note that we do NOT increment the input pointer. + // This is because the closing linebreak could be the opening linebreak of + // another heading. Infinite loops are avoided because the next iteration MUST + // hit the heading open case above, which unconditionally increments the + // input pointer. + } + + elseif ( $found == 'open' ) { # count opening brace characters - $piece['count'] = strspn( $text, $piece['brace'], $i ); - $piece['startAt'] = $piece['partStart'] = $i + $piece['count']; - $i += $piece['count']; + $count = strspn( $text, $curChar, $i ); # we need to add to stack only if opening brace count is enough for one of the rules - if ( $piece['count'] >= $rule['min'] ) { - $lastOpeningBrace ++; - $openingBraceStack[$lastOpeningBrace] = $piece; + if ( $count >= $rule['min'] ) { + # Add it to the stack + $piece = array( + 'open' => $curChar, + 'close' => $rule['end'], + 'count' => $count, + 'parts' => array( '' ), + 'eqpos' => array(), + 'lineStart' => ($i > 0 && $text[$i-1] == "\n"), + ); + + $stackIndex ++; + $stack[$stackIndex] = $piece; + $accum =& $stack[$stackIndex]['parts'][0]; + $findEquals = false; + $findPipe = true; + } else { + # Add literal brace(s) + $accum .= htmlspecialchars( str_repeat( $curChar, $count ) ); } - } elseif ( $found == 'close' ) { - # lets check if it is enough characters for closing brace - $maxCount = $openingBraceStack[$lastOpeningBrace]['count']; - $count = strspn( $text, $text[$i], $i, $maxCount ); + $i += $count; + } + + elseif ( $found == 'close' ) { + $piece = $stack[$stackIndex]; + # lets check if there are enough characters for closing brace + $maxCount = $piece['count']; + $count = strspn( $text, $curChar, $i, $maxCount ); # check for maximum matching characters (if there are 5 closing # characters, we will probably need only 3 - depending on the rules) $matchingCount = 0; - $matchingCallback = null; - $cbType = $callbacks[$openingBraceStack[$lastOpeningBrace]['brace']]; - if ( $count > $cbType['max'] ) { + $rule = $rules[$piece['open']]; + if ( $count > $rule['max'] ) { # The specified maximum exists in the callback array, unless the caller # has made an error - $matchingCount = $cbType['max']; + $matchingCount = $rule['max']; } else { # Count is less than the maximum # Skip any gaps in the callback array to find the true largest match # Need to use array_key_exists not isset because the callback can be null $matchingCount = $count; - while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $cbType['cb'] ) ) { + while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) { --$matchingCount; } } if ($matchingCount <= 0) { + # No matching element found in callback array + # Output a literal closing brace and continue + $accum .= htmlspecialchars( str_repeat( $curChar, $count ) ); $i += $count; continue; } - $matchingCallback = $cbType['cb'][$matchingCount]; - - # let's set a title or last part (if '|' was found) - if (null === $openingBraceStack[$lastOpeningBrace]['parts']) { - $openingBraceStack[$lastOpeningBrace]['title'] = - substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], - $i - $openingBraceStack[$lastOpeningBrace]['partStart']); + $name = $rule['names'][$matchingCount]; + if ( $name === null ) { + // No element, just literal text + $element = str_repeat( $piece['open'], $matchingCount ) . + implode( '|', $piece['parts'] ) . + str_repeat( $rule['end'], $matchingCount ); } else { - $openingBraceStack[$lastOpeningBrace]['parts'][] = - substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], - $i - $openingBraceStack[$lastOpeningBrace]['partStart']); + # Create XML element + # Note: $parts is already XML, does not need to be encoded further + $parts = $piece['parts']; + $title = $parts[0]; + unset( $parts[0] ); + + # The invocation is at the start of the line if lineStart is set in + # the stack, and all opening brackets are used up. + if ( $maxCount == $matchingCount && !empty( $piece['lineStart'] ) ) { + $attr = ' lineStart="1"'; + } else { + $attr = ''; + } + + $element = "<$name$attr>"; + $element .= "$title"; + $argIndex = 1; + foreach ( $parts as $partIndex => $part ) { + if ( isset( $piece['eqpos'][$partIndex] ) ) { + $eqpos = $piece['eqpos'][$partIndex]; + list( $ws1, $argName, $ws2 ) = self::splitWhitespace( substr( $part, 0, $eqpos ) ); + list( $ws3, $argValue, $ws4 ) = self::splitWhitespace( substr( $part, $eqpos + 1 ) ); + $element .= "$ws1$argName$ws2=$ws3$argValue$ws4"; + } else { + list( $ws1, $value, $ws2 ) = self::splitWhitespace( $part ); + $element .= "$ws1$value$ws2"; + $argIndex++; + } + } + $element .= ""; } - $pieceStart = $openingBraceStack[$lastOpeningBrace]['startAt'] - $matchingCount; - $pieceEnd = $i + $matchingCount; - - if( is_callable( $matchingCallback ) ) { - $cbArgs = array ( - 'text' => substr($text, $pieceStart, $pieceEnd - $pieceStart), - 'title' => trim($openingBraceStack[$lastOpeningBrace]['title']), - 'parts' => $openingBraceStack[$lastOpeningBrace]['parts'], - 'lineStart' => (($pieceStart > 0) && ($text[$pieceStart-1] == "\n")), - ); - # finally we can call a user callback and replace piece of text - $replaceWith = call_user_func( $matchingCallback, $cbArgs ); - $text = substr($text, 0, $pieceStart) . $replaceWith . substr($text, $pieceEnd); - $i = $pieceStart + strlen($replaceWith); + # Advance input pointer + $i += $matchingCount; + + # Unwind the stack + unset( $stack[$stackIndex--] ); + if ( $stackIndex == -1 ) { + $accum =& $topAccum; + $findEquals = false; + $findPipe = false; } else { - # null value for callback means that parentheses should be parsed, but not replaced - $i += $matchingCount; + $partCount = count( $stack[$stackIndex]['parts'] ); + $accum =& $stack[$stackIndex]['parts'][$partCount - 1]; + $findPipe = $stack[$stackIndex]['open'] != "\n"; + $findEquals = $findPipe && $partCount > 1 + && !isset( $stack[$stackIndex]['eqpos'][$partCount - 1] ); } - # reset last opening parentheses, but keep it in case there are unused characters - $piece = array('brace' => $openingBraceStack[$lastOpeningBrace]['brace'], - 'braceEnd' => $openingBraceStack[$lastOpeningBrace]['braceEnd'], - 'count' => $openingBraceStack[$lastOpeningBrace]['count'], - 'title' => '', - 'parts' => null, - 'startAt' => $openingBraceStack[$lastOpeningBrace]['startAt']); - $openingBraceStack[$lastOpeningBrace--] = null; - + # Re-add the old stack element if it still has unmatched opening characters remaining if ($matchingCount < $piece['count']) { + $piece['parts'] = array( '' ); $piece['count'] -= $matchingCount; - $piece['startAt'] -= $matchingCount; - $piece['partStart'] = $piece['startAt']; + $piece['eqpos'] = array(); # do we still qualify for any callback with remaining count? - $currentCbList = $callbacks[$piece['brace']]['cb']; + $names = $rules[$piece['open']]['names']; + $skippedBraces = 0; + $enclosingAccum =& $accum; while ( $piece['count'] ) { - if ( array_key_exists( $piece['count'], $currentCbList ) ) { - $lastOpeningBrace++; - $openingBraceStack[$lastOpeningBrace] = $piece; + if ( array_key_exists( $piece['count'], $names ) ) { + $stackIndex++; + $stack[$stackIndex] = $piece; + $accum =& $stack[$stackIndex]['parts'][0]; + $findEquals = true; + $findPipe = true; break; } --$piece['count']; + $skippedBraces ++; } + $enclosingAccum .= str_repeat( $piece['open'], $skippedBraces ); } - } elseif ( $found == 'pipe' ) { - # lets set a title if it is a first separator, or next part otherwise - if (null === $openingBraceStack[$lastOpeningBrace]['parts']) { - $openingBraceStack[$lastOpeningBrace]['title'] = - substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], - $i - $openingBraceStack[$lastOpeningBrace]['partStart']); - $openingBraceStack[$lastOpeningBrace]['parts'] = array(); - } else { - $openingBraceStack[$lastOpeningBrace]['parts'][] = - substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], - $i - $openingBraceStack[$lastOpeningBrace]['partStart']); - } - $openingBraceStack[$lastOpeningBrace]['partStart'] = ++$i; + + # Add XML element to the enclosing accumulator + $accum .= $element; + } + + elseif ( $found == 'pipe' ) { + $stack[$stackIndex]['parts'][] = ''; + $partsCount = count( $stack[$stackIndex]['parts'] ); + $accum =& $stack[$stackIndex]['parts'][$partsCount - 1]; + $findEquals = true; + ++$i; + } + + elseif ( $found == 'equals' ) { + $findEquals = false; + $partsCount = count( $stack[$stackIndex]['parts'] ); + $stack[$stackIndex]['eqpos'][$partsCount - 1] = strlen( $accum ); + $accum .= '='; + ++$i; } } + # Output any remaining unclosed brackets + foreach ( $stack as $piece ) { + if ( $piece['open'] == "\n" ) { + $topAccum .= $piece['parts'][0]; + } else { + $topAccum .= str_repeat( $piece['open'], $piece['count'] ) . implode( '|', $piece['parts'] ); + } + } + $topAccum .= ''; + + wfProfileOut( __METHOD__.'-makexml' ); + wfProfileIn( __METHOD__.'-loadXML' ); + $dom = new DOMDocument; + wfSuppressWarnings(); + $result = $dom->loadXML( $topAccum ); + wfRestoreWarnings(); + if ( !$result ) { + // Try running the XML through UtfNormal to get rid of invalid characters + $topAccum = UtfNormal::cleanUp( $topAccum ); + $result = $dom->loadXML( $topAccum ); + if ( !$result ) { + throw new MWException( __METHOD__.' generated invalid XML' ); + } + } + wfProfileOut( __METHOD__.'-loadXML' ); wfProfileOut( __METHOD__ ); - return $text; + return $dom; + } + + /* + * Return a three-element array: leading whitespace, string contents, trailing whitespace + */ + public static function splitWhitespace( $s ) { + $ltrimmed = ltrim( $s ); + $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) ); + $trimmed = rtrim( $ltrimmed ); + $diff = strlen( $ltrimmed ) - strlen( $trimmed ); + if ( $diff > 0 ) { + $w2 = substr( $ltrimmed, -$diff ); + } else { + $w2 = ''; + } + return array( $w1, $trimmed, $w2 ); } /** @@ -2779,89 +3140,33 @@ class Parser * OT_HTML: all templates and magic variables * * @param string $tex The text to transform - * @param array $args Key-value pairs representing template parameters to substitute + * @param PPFrame $frame Object describing the arguments passed to the template * @param bool $argsOnly Only do argument (triple-brace) expansion, not double-brace expansion * @private */ - function replaceVariables( $text, $args = array(), $argsOnly = false ) { + function replaceVariables( $text, $frame = false, $argsOnly = false ) { # Prevent too big inclusions if( strlen( $text ) > $this->mOptions->getMaxIncludeSize() ) { return $text; } - $fname = __METHOD__ /*. '-L' . count( $this->mArgStack )*/; + $fname = __METHOD__; wfProfileIn( $fname ); - # This function is called recursively. To keep track of arguments we need a stack: - array_push( $this->mArgStack, $args ); - - $braceCallbacks = array(); - if ( !$argsOnly ) { - $braceCallbacks[2] = array( &$this, 'braceSubstitution' ); - } - if ( $this->mOutputType != OT_MSG ) { - $braceCallbacks[3] = array( &$this, 'argSubstitution' ); + if ( $frame === false ) { + $frame = new PPFrame( $this ); + } elseif ( !( $frame instanceof PPFrame ) ) { + throw new MWException( __METHOD__ . ' called using the old argument format' ); } - if ( $braceCallbacks ) { - $callbacks = array( - '{' => array( - 'end' => '}', - 'cb' => $braceCallbacks, - 'min' => $argsOnly ? 3 : 2, - 'max' => isset( $braceCallbacks[3] ) ? 3 : 2, - ), - '[' => array( - 'end' => ']', - 'cb' => array(2=>null), - 'min' => 2, - 'max' => 2, - ) - ); - $text = $this->replace_callback ($text, $callbacks); - array_pop( $this->mArgStack ); - } - wfProfileOut( $fname ); - return $text; - } + $dom = $this->preprocessToDom( $text ); + $flags = $argsOnly ? PPFrame::NO_TEMPLATES : 0; + $text = $frame->expand( $dom, $flags ); - /** - * Replace magic variables - * @private - */ - function variableSubstitution( $matches ) { - global $wgContLang; - $fname = 'Parser::variableSubstitution'; - $varname = $wgContLang->lc($matches[1]); - wfProfileIn( $fname ); - $skip = false; - if ( $this->mOutputType == OT_WIKI ) { - # Do only magic variables prefixed by SUBST - $mwSubst =& MagicWord::get( 'subst' ); - if (!$mwSubst->matchStartAndRemove( $varname )) - $skip = true; - # Note that if we don't substitute the variable below, - # we don't remove the {{subst:}} magic word, in case - # it is a template rather than a magic variable. - } - if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) { - $id = $this->mVariables[$varname]; - # Now check if we did really match, case sensitive or not - $mw =& MagicWord::get( $id ); - if ($mw->match($matches[1])) { - $text = $this->getVariableValue( $id ); - $this->mOutput->mContainsOldMagic = true; - } else { - $text = $matches[0]; - } - } else { - $text = $matches[0]; - } wfProfileOut( $fname ); return $text; } - /// Clean up argument array - refactored in 1.9 so parserfunctions can use it, too. static function createAssocArgs( $args ) { $assocArgs = array(); @@ -2893,49 +3198,38 @@ class Parser * $piece['text']: matched text * $piece['title']: the title, i.e. the part before the | * $piece['parts']: the parameter array + * @param PPFrame The current frame, contains template arguments * @return string the text of the template * @private */ - function braceSubstitution( $piece ) { + function braceSubstitution( $piece, $frame ) { global $wgContLang, $wgLang, $wgAllowDisplayTitle, $wgNonincludableNamespaces; - $fname = __METHOD__ /*. '-L' . count( $this->mArgStack )*/; + $fname = __METHOD__; wfProfileIn( $fname ); wfProfileIn( __METHOD__.'-setup' ); # Flags $found = false; # $text has been filled $nowiki = false; # wiki markup in $text should be escaped - $noparse = false; # Unsafe HTML tags should not be stripped, etc. - $noargs = false; # Don't replace triple-brace arguments in $text - $replaceHeadings = false; # Make the edit section links go to the template not the article - $headingOffset = 0; # Skip headings when number, to account for those that weren't transcluded. $isHTML = false; # $text is HTML, armour it against wikitext transformation $forceRawInterwiki = false; # Force interwiki transclusion to be done in raw mode not rendered + $isDOM = false; # $text is a DOM node needing expansion # Title object, where $text came from $title = NULL; - $linestart = ''; + # $part1 is the bit before the first |, and must contain only title characters. + # Various prefixes will be stripped from it later. + $titleWithSpaces = $frame->expand( $piece['title'] ); + $part1 = trim( $titleWithSpaces ); + $titleText = false; + # Original title text preserved for various purposes + $originalTitle = $part1; - # $part1 is the bit before the first |, and must contain only title characters - # $args is a list of arguments, starting from index 0, not including $part1 - - $titleText = $part1 = $piece['title']; - # If the third subpattern matched anything, it will start with | - - if (null == $piece['parts']) { - $replaceWith = $this->variableSubstitution (array ($piece['text'], $piece['title'])); - if ($replaceWith != $piece['text']) { - $text = $replaceWith; - $found = true; - $noparse = true; - $noargs = true; - } - } - - $args = (null == $piece['parts']) ? array() : $piece['parts']; - wfProfileOut( __METHOD__.'-setup' ); + # $args is a list of argument nodes, starting from index 0, not including $part1 + $args = (null == $piece['parts']) ? array() : $piece['parts']; + wfProfileOut( __METHOD__.'-setup' ); # SUBST wfProfileIn( __METHOD__.'-modifiers' ); @@ -2946,10 +3240,18 @@ class Parser # 1) Found SUBST but not in the PST phase # 2) Didn't find SUBST and in the PST phase # In either case, return without further processing - $text = $piece['text']; + $text = '{{' . $frame->implode( '|', $titleWithSpaces, $args ) . '}}'; + $found = true; + } + } + + # Variables + if ( !$found && $args->length == 0 ) { + $id = $this->mVariables->matchStartToEnd( $part1 ); + if ( $id !== false ) { + $text = $this->getVariableValue( $id ); + $this->mOutput->mContainsOldMagic = true; $found = true; - $noparse = true; - $noargs = true; } } @@ -2973,7 +3275,7 @@ class Parser } wfProfileOut( __METHOD__.'-modifiers' ); - //save path level before recursing into functions & templates. + # Save path level before recursing into functions & templates. $lastPathLevel = $this->mTemplatePath; # Parser functions @@ -2996,277 +3298,263 @@ class Parser } } if ( $function ) { - $funcArgs = array_map( 'trim', $args ); - $funcArgs = array_merge( array( &$this, trim( substr( $part1, $colonPos + 1 ) ) ), $funcArgs ); - $result = call_user_func_array( $this->mFunctionHooks[$function], $funcArgs ); - $found = true; + list( $callback, $flags ) = $this->mFunctionHooks[$function]; + $initialArgs = array( &$this ); + $funcArgs = array( trim( substr( $part1, $colonPos + 1 ) ) ); + if ( $flags & SFH_OBJECT_ARGS ) { + # Add a frame parameter, and pass the arguments as an array + $allArgs = $initialArgs; + $allArgs[] = $frame; + foreach ( $args as $arg ) { + $funcArgs[] = $arg; + } + $allArgs[] = $funcArgs; + } else { + # Convert arguments to plain text + foreach ( $args as $arg ) { + $funcArgs[] = trim( $frame->expand( $arg ) ); + } + $allArgs = array_merge( $initialArgs, $funcArgs ); + } - // The text is usually already parsed, doesn't need triple-brace tags expanded, etc. - //$noargs = true; - //$noparse = true; + $result = call_user_func_array( $callback, $allArgs ); + $found = true; if ( is_array( $result ) ) { if ( isset( $result[0] ) ) { - $text = $linestart . $result[0]; + $text = $result[0]; unset( $result[0] ); } // Extract flags into the local scope - // This allows callers to set flags such as nowiki, noparse, found, etc. + // This allows callers to set flags such as nowiki, found, etc. extract( $result ); } else { - $text = $linestart . $result; + $text = $result; } } } wfProfileOut( __METHOD__ . '-pfunc' ); } - # Template table test - - # Did we encounter this template already? If yes, it is in the cache - # and we need to check for loops. - if ( !$found && isset( $this->mTemplates[$piece['title']] ) ) { - $found = true; - - # Infinite loop test - if ( isset( $this->mTemplatePath[$part1] ) ) { - $noparse = true; - $noargs = true; - $found = true; - $text = $linestart . - "[[$part1]]"; - wfDebug( __METHOD__.": template loop broken at '$part1'\n" ); - } else { - # set $text to cached message. - $text = $linestart . $this->mTemplates[$piece['title']]; - #treat title for cached page the same as others - $ns = NS_TEMPLATE; - $subpage = ''; - $part1 = $this->maybeDoSubpageLink( $part1, $subpage ); - if ($subpage !== '') { - $ns = $this->mTitle->getNamespace(); - } - $title = Title::newFromText( $part1, $ns ); - //used by include size checking - $titleText = $title->getPrefixedText(); - //used by edit section links - $replaceHeadings = true; - - } - } - - # Load from database + # Finish mangling title and then check for loops. + # Set $title to a Title object and $titleText to the PDBK if ( !$found ) { - wfProfileIn( __METHOD__ . '-loadtpl' ); $ns = NS_TEMPLATE; - # declaring $subpage directly in the function call - # does not work correctly with references and breaks - # {{/subpage}}-style inclusions + # Split the title into page and subpage $subpage = ''; $part1 = $this->maybeDoSubpageLink( $part1, $subpage ); if ($subpage !== '') { $ns = $this->mTitle->getNamespace(); } $title = Title::newFromText( $part1, $ns ); - - - if ( !is_null( $title ) ) { + if ( $title ) { $titleText = $title->getPrefixedText(); # Check for language variants if the template is not found if($wgContLang->hasVariants() && $title->getArticleID() == 0){ $wgContLang->findVariantLink($part1, $title); } + # Do infinite loop check + if ( isset( $this->mTemplatePath[$titleText] ) ) { + $found = true; + $text = "[[$part1]]" . $this->insertStripItem( '' ); + wfDebug( __METHOD__.": template loop broken at '$titleText'\n" ); + } + } + } - if ( !$title->isExternal() ) { - if ( $title->getNamespace() == NS_SPECIAL && $this->mOptions->getAllowSpecialInclusion() && $this->ot['html'] ) { - $text = SpecialPage::capturePath( $title ); - if ( is_string( $text ) ) { - $found = true; - $noparse = true; - $noargs = true; - $isHTML = true; - $this->disableCache(); - } - } else if ( $wgNonincludableNamespaces && in_array( $title->getNamespace(), $wgNonincludableNamespaces ) ) { - $found = false; //access denied - wfDebug( "$fname: template inclusion denied for " . $title->getPrefixedDBkey() ); - } else { - list($articleContent,$title) = $this->fetchTemplateAndtitle( $title ); - if ( $articleContent !== false ) { - $found = true; - $text = $articleContent; - $replaceHeadings = true; - } - } - - # If the title is valid but undisplayable, make a link to it - if ( !$found && ( $this->ot['html'] || $this->ot['pre'] ) ) { - $text = "[[:$titleText]]"; + # Load from database + if ( !$found && $title ) { + wfProfileIn( __METHOD__ . '-loadtpl' ); + if ( !$title->isExternal() ) { + if ( $title->getNamespace() == NS_SPECIAL && $this->mOptions->getAllowSpecialInclusion() && $this->ot['html'] ) { + $text = SpecialPage::capturePath( $title ); + if ( is_string( $text ) ) { $found = true; - } - } elseif ( $title->isTrans() ) { - // Interwiki transclusion - if ( $this->ot['html'] && !$forceRawInterwiki ) { - $text = $this->interwikiTransclude( $title, 'render' ); $isHTML = true; - $noparse = true; - } else { - $text = $this->interwikiTransclude( $title, 'raw' ); - $replaceHeadings = true; + $this->disableCache(); + } + } else if ( $wgNonincludableNamespaces && in_array( $title->getNamespace(), $wgNonincludableNamespaces ) ) { + $found = false; //access denied + wfDebug( "$fname: template inclusion denied for " . $title->getPrefixedDBkey() ); + } else { + list( $text, $title ) = $this->getTemplateDom( $title ); + if ( $text !== false ) { + $found = true; + $isDOM = true; } - $found = true; } - # Template cache array insertion - # Use the original $piece['title'] not the mangled $part1, so that - # modifiers such as RAW: produce separate cache entries - if( $found ) { - if( $isHTML ) { - // A special page; don't store it in the template cache. - } else { - $this->mTemplates[$piece['title']] = $text; - } - $text = $linestart . $text; + # If the title is valid but undisplayable, make a link to it + if ( !$found && ( $this->ot['html'] || $this->ot['pre'] ) ) { + $text = "[[:$titleText]]"; + $found = true; + } + } elseif ( $title->isTrans() ) { + // Interwiki transclusion + if ( $this->ot['html'] && !$forceRawInterwiki ) { + $text = $this->interwikiTransclude( $title, 'render' ); + $isHTML = true; + } else { + $text = $this->interwikiTransclude( $title, 'raw' ); + // Preprocess it like a template + $text = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION ); + $isDOM = true; } + $found = true; } wfProfileOut( __METHOD__ . '-loadtpl' ); } - if ( $found && !$this->incrementIncludeSize( 'pre-expand', strlen( $text ) ) ) { - # Error, oversize inclusion - $text = $linestart . - "[[$titleText]]"; - $noparse = true; - $noargs = true; + # If we haven't found text to substitute by now, we're done + # Recover the source wikitext and return it + if ( !$found ) { + $text = '{{' . $frame->implode( '|', $titleWithSpaces, $args ) . '}}'; + # Prune lower levels off the recursion check path + $this->mTemplatePath = $lastPathLevel; + wfProfileOut( $fname ); + return $text; } - # Recursive parsing, escaping and link table handling - # Only for HTML output - if ( $nowiki && $found && ( $this->ot['html'] || $this->ot['pre'] ) ) { - $text = wfEscapeWikiText( $text ); - } elseif ( !$this->ot['msg'] && $found ) { - if ( $noargs ) { - $assocArgs = array(); + # Expand DOM-style return values in a child frame + if ( $isDOM ) { + # Clean up argument array + $newFrame = $frame->newChild( $args, $title ); + # Add a new element to the templace recursion path + $this->mTemplatePath[$titleText] = 1; + + if ( $titleText !== false && count( $newFrame->args ) == 0 ) { + # Expansion is eligible for the empty-frame cache + if ( isset( $this->mTplExpandCache[$titleText] ) ) { + $text = $this->mTplExpandCache[$titleText]; + } else { + $text = $newFrame->expand( $text ); + $this->mTplExpandCache[$titleText] = $text; + } } else { - # Clean up argument array - $assocArgs = self::createAssocArgs($args); - # Add a new element to the templace recursion path - $this->mTemplatePath[$part1] = 1; + $text = $newFrame->expand( $text ); } + } - if ( !$noparse ) { - # If there are any tags, only include them - if ( in_string( '', $text ) && in_string( '', $text ) ) { - $replacer = new OnlyIncludeReplacer; - StringUtils::delimiterReplaceCallback( '', '', - array( &$replacer, 'replace' ), $text ); - $text = $replacer->output; - } - # Remove sections and tags - $text = StringUtils::delimiterReplace( '', '', '', $text ); - $text = strtr( $text, array( '' => '' , '' => '' ) ); - - if( $this->ot['html'] || $this->ot['pre'] ) { - # Strip , 
, etc.
-					$text = $this->strip( $text, $this->mStripState );
-					if ( $this->ot['html'] ) {
-						$text = Sanitizer::removeHTMLtags( $text, array( &$this, 'replaceVariables' ), $assocArgs );
-					} elseif ( $this->ot['pre'] && $this->mOptions->getRemoveComments() ) {
-						$text = Sanitizer::removeHTMLcomments( $text );
-					}
-				}
-				$text = $this->replaceVariables( $text, $assocArgs );
-
-				# If the template begins with a table or block-level
-				# element, it should be treated as beginning a new line.
-				if (!$piece['lineStart'] && preg_match('/^(?:{\\||:|;|#|\*)/', $text)) /*}*/{
-					$text = "\n" . $text;
-				}
-			} elseif ( !$noargs ) {
-				# $noparse and !$noargs
-				# Just replace the arguments, not any double-brace items
-				# This is used for rendered interwiki transclusion
-				$text = $this->replaceVariables( $text, $assocArgs, true );
-			}
+		# Replace raw HTML by a placeholder
+		# Add a blank line preceding, to prevent it from mucking up
+		# immediately preceding headings
+		if ( $isHTML ) {
+			$text = "\n\n" . $this->insertStripItem( $text );
+		}
+		# Escape nowiki-style return values
+		elseif ( $nowiki && ( $this->ot['html'] || $this->ot['pre'] ) ) {
+			$text = wfEscapeWikiText( $text );
 		}
+		# Bug 529: if the template begins with a table or block-level
+		# element, it should be treated as beginning a new line.
+		# This behaviour is somewhat controversial.
+		elseif ( !$piece['lineStart'] && preg_match('/^(?:{\\||:|;|#|\*)/', $text)) /*}*/{
+			$text = "\n" . $text;
+		}
+		
 		# Prune lower levels off the recursion check path
 		$this->mTemplatePath = $lastPathLevel;
 
-		if ( $found && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
+		if ( !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
 			# Error, oversize inclusion
-			$text = $linestart .
-				"[[$titleText]]";
-			$noparse = true;
-			$noargs = true;
+			$text = "[[$originalTitle]]" . 
+				$this->insertStripItem( '' );
 		}
 
-		if ( !$found ) {
-			wfProfileOut( $fname );
-			return $piece['text'];
-		} else {
-			wfProfileIn( __METHOD__ . '-placeholders' );
-			if ( $isHTML ) {
-				# Replace raw HTML by a placeholder
-				# Add a blank line preceding, to prevent it from mucking up
-				# immediately preceding headings
-				$text = "\n\n" . $this->insertStripItem( $text, $this->mStripState );
-			} else {
-				# replace ==section headers==
-				# XXX this needs to go away once we have a better parser.
-				if ( !$this->ot['wiki'] && !$this->ot['pre'] && $replaceHeadings ) {
-					if( !is_null( $title ) )
-						$encodedname = base64_encode($title->getPrefixedDBkey());
-					else
-						$encodedname = base64_encode("");
-					$m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
-						PREG_SPLIT_DELIM_CAPTURE);
-					$text = '';
-					$nsec = $headingOffset;
-
-					for( $i = 0; $i < count($m); $i += 2 ) {
-						$text .= $m[$i];
-						if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
-						$hl = $m[$i + 1];
-						if( strstr($hl, "" . $m2[3];
+		wfProfileOut( $fname );
+		return $text;
+	}
 
-						$nsec++;
-					}
-				}
-			}
-			wfProfileOut( __METHOD__ . '-placeholders' );
+	/**
+	 * Get the semi-parsed DOM representation of a template with a given title,
+	 * and its redirect destination title. Cached.
+	 */
+	function getTemplateDom( $title ) {
+		$cacheTitle = $title;
+		$titleText = $title->getPrefixedDBkey();
+		
+		if ( isset( $this->mTplRedirCache[$titleText] ) ) {
+			list( $ns, $dbk ) = $this->mTplRedirCache[$titleText];
+			$title = Title::makeTitle( $ns, $dbk );
+			$titleText = $title->getPrefixedDBkey();
+		}
+		if ( isset( $this->mTplDomCache[$titleText] ) ) {
+			return array( $this->mTplDomCache[$titleText], $title );
 		}
 
-		# Prune lower levels off the recursion check path
-		$this->mTemplatePath = $lastPathLevel;
+		// Cache miss, go to the database
+		list( $text, $title ) = $this->fetchTemplateAndTitle( $title );
 
-		if ( !$found ) {
-			wfProfileOut( $fname );
-			return $piece['text'];
-		} else {
-			wfProfileOut( $fname );
-			return $text;
+		if ( $text === false ) {
+			$this->mTplDomCache[$titleText] = false;
+			return array( false, $title );
+		}
+
+		$dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
+		$this->mTplDomCache[ $titleText ] = $dom;
+
+		if (! $title->equals($cacheTitle)) {
+			$this->mTplRedirCache[$cacheTitle->getPrefixedDBkey()] = 
+				array( $title->getNamespace(),$cdb = $title->getDBkey() );
 		}
+
+		return array( $dom, $title );
 	}
 
 	/**
 	 * Fetch the unparsed text of a template and register a reference to it.
 	 */
-	function fetchTemplateAndtitle( $title ) {
-		$text = false;
+	function fetchTemplateAndTitle( $title ) {
+		$templateCb = $this->mOptions->getTemplateCallback();
+		$stuff = call_user_func( $templateCb, $title );
+		$text = $stuff['text'];
+		$finalTitle = isset( $stuff['finalTitle'] ) ? $stuff['finalTitle'] : $title;
+		if ( isset( $stuff['deps'] ) ) {
+			foreach ( $stuff['deps'] as $dep ) {
+				$this->mOutput->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
+			}
+		}
+		return array($text,$finalTitle);
+	}
+
+	function fetchTemplate( $title ) {
+		$rv = $this->fetchTemplateAndTitle($title);
+		return $rv[0];
+	}
+
+	/**
+	 * Static function to get a template
+	 * Can be overridden via ParserOptions::setTemplateCallback().
+	 */
+	static function statelessFetchTemplate( $title ) {
+		$text = $skip = false;
 		$finalTitle = $title;
+		$deps = array();
+		
 		// Loop to fetch the article, with up to 1 redirect
 		for ( $i = 0; $i < 2 && is_object( $title ); $i++ ) {
-			$rev = Revision::newFromTitle( $title );
-			$this->mOutput->addTemplate( $title, $title->getArticleID() );
-			if ( $rev ) {
+			# Give extensions a chance to select the revision instead
+			$id = false; // Assume current
+			wfRunHooks( 'BeforeParserFetchTemplateAndtitle', array( false, &$title, &$skip, &$id ) );
+			
+			if( $skip ) {
+				$text = false;
+				$deps[] = array(
+					'title' => $title,
+					'page_id' => $title->getArticleID(),
+					'rev_id' => null );
+				break;
+			}
+			$rev = $id ? Revision::newFromId( $id ) : Revision::newFromTitle( $title );
+			$rev_id = $rev ? $rev->getId() : 0;
+
+			$deps[] = array( 
+				'title' => $title, 
+				'page_id' => $title->getArticleID(), 
+				'rev_id' => $rev_id );
+
+			if( $rev ) {
 				$text = $rev->getText();
 			} elseif( $title->getNamespace() == NS_MEDIAWIKI ) {
 				global $wgLang;
@@ -3286,12 +3574,10 @@ class Parser
 			$finalTitle = $title;
 			$title = Title::newFromRedirect( $text );
 		}
-		return array($text,$finalTitle);
-	}
-
-	function fetchTemplate( $title ) {
-		$rv = $this->fetchTemplateAndtitle($title);
-		return $rv[0];
+		return array(
+			'text' => $text,
+			'finalTitle' => $finalTitle,
+			'deps' => $deps );
 	}
 
 	/**
@@ -3340,25 +3626,103 @@ class Parser
 	 * Triple brace replacement -- used for template arguments
 	 * @private
 	 */
-	function argSubstitution( $matches ) {
-		$arg = trim( $matches['title'] );
-		$text = $matches['text'];
-		$inputArgs = end( $this->mArgStack );
+	function argSubstitution( $piece, $frame ) {
+		wfProfileIn( __METHOD__ );
 
-		if ( array_key_exists( $arg, $inputArgs ) ) {
-			$text = $inputArgs[$arg];
-		} else if (($this->mOutputType == OT_HTML || $this->mOutputType == OT_PREPROCESS ) &&
-		null != $matches['parts'] && count($matches['parts']) > 0) {
-			$text = $matches['parts'][0];
+		$text = false;
+		$error = false;
+		$parts = $piece['parts'];
+		$argWithSpaces = $frame->expand( $piece['title'] );
+		$arg = trim( $argWithSpaces );
+
+		if ( isset( $frame->args[$arg] ) ) {
+			$text = $frame->parent->expand( $frame->args[$arg] );
+		} else if ( ( $this->ot['html'] || $this->ot['pre'] ) && $parts->length > 0 ) {
+			$text = $frame->expand( $parts->item( 0 ) );
 		}
 		if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
-			$text = $matches['text'] .
-				'';
+			$error = '';
 		}
 
+		if ( $text === false ) {
+			$text = '{{{' . $frame->implode( '|', $argWithSpaces, $parts ) . '}}}';
+		}
+		if ( $error !== false ) {
+			$text .= $error;
+		}
+
+		wfProfileOut( __METHOD__ );
 		return $text;
 	}
 
+	/**
+	 * Return the text to be used for a given extension tag.
+	 * This is the ghost of strip().
+	 *
+	 * @param array $params Associative array of parameters:
+	 *     name       DOMNode for the tag name
+	 *     attrText   DOMNode for unparsed text where tag attributes are thought to be
+	 *     inner      Contents of extension element
+	 *     noClose    Original text did not have a close tag
+	 * @param PPFrame $frame
+	 */
+	function extensionSubstitution( $params, $frame ) {
+		global $wgRawHtml, $wgContLang;
+		static $n = 1;
+
+		$name = $frame->expand( $params['name'] );
+		$attrText = is_null( $params['attr'] ) ? null : $frame->expand( $params['attr'] );
+		$content = is_null( $params['inner'] ) ? null : $frame->expand( $params['inner'] );
+
+		$marker = "{$this->mUniqPrefix}-$name-" . sprintf('%08X', $n++) . $this->mMarkerSuffix;
+		
+		if ( $this->ot['html'] ) {
+			$name = strtolower( $name );
+
+			$params = Sanitizer::decodeTagAttributes( $attrText );
+			switch ( $name ) {
+				case 'html':
+					if( $wgRawHtml ) {
+						$output = $content;
+						break;
+					} else {
+						throw new MWException( ' extension tag encountered unexpectedly' );
+					}
+				case 'nowiki':
+					$output = Xml::escapeTagsOnly( $content );
+					break;
+				case 'math':
+					$output = $wgContLang->armourMath(
+						MathRenderer::renderMath( $content, $params ) );
+					break;
+				case 'gallery':
+					$output = $this->renderImageGallery( $content, $params );
+					break;
+				default:
+					if( isset( $this->mTagHooks[$name] ) ) {
+						$output = call_user_func_array( $this->mTagHooks[$name],
+							array( $content, $params, $this ) );
+					} else {
+						throw new MWException( "Invalid call hook $name" );
+					}
+			}
+		} else {
+			if ( $content === null ) {
+				$output = "<$name$attrText/>";
+			} else {
+				$close = is_null( $params['close'] ) ? '' : $frame->expand( $params['close'] );
+				$output = "<$name$attrText>$content$close";
+			}
+		}
+
+		if ( $name == 'html' || $name == 'nowiki' ) {
+			$this->mStripState->nowiki->setPair( $marker, $output );
+		} else {
+			$this->mStripState->general->setPair( $marker, $output );
+		}
+		return $marker;
+	}
+
 	/**
 	 * Increment an include size counter
 	 *
@@ -3367,7 +3731,7 @@ class Parser
 	 * @return boolean False if this inclusion would take it over the maximum, true otherwise
 	 */
 	function incrementIncludeSize( $type, $size ) {
-		if ( $this->mIncludeSizes[$type] + $size > $this->mOptions->getMaxIncludeSize() ) {
+		if ( $this->mIncludeSizes[$type] + $size > $this->mOptions->getMaxIncludeSize( $type ) ) {
 			return false;
 		} else {
 			$this->mIncludeSizes[$type] += $size;
@@ -3386,7 +3750,13 @@ class Parser
 	}
 
 	/**
-	 * Detect __TOC__ magic word and set a placeholder
+	 * Find the first __TOC__ magic word and set a 
+	 * placeholder that will then be replaced by the real TOC in
+	 * ->formatHeadings, this works because at this points real
+	 * comments will have already been discarded by the sanitizer.
+	 *
+	 * Any additional __TOC__ magic words left over will be discarded
+	 * as there can only be one TOC on the page.
 	 */
 	function stripToc( $text ) {
 		# if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
@@ -3413,7 +3783,7 @@ class Parser
 	/**
 	 * This function accomplishes several tasks:
 	 * 1) Auto-number headings if that option is enabled
-	 * 2) Add an [edit] link to sections for logged in users who have enabled the option
+	 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
 	 * 3) Add a Table of contents on the top for users who have enabled the option
 	 * 4) Auto-anchor headings
 	 *
@@ -3469,7 +3839,6 @@ class Parser
 
 		# headline counter
 		$headlineCount = 0;
-		$sectionCount = 0; # headlineCount excluding template sections
 		$numVisible = 0;
 
 		# Ugh .. the TOC should have neat indentation levels which can be
@@ -3484,18 +3853,21 @@ class Parser
 		$prevlevel = 0;
 		$toclevel = 0;
 		$prevtoclevel = 0;
+		$markerRegex = "{$this->mUniqPrefix}-h-(\d+)-{$this->mMarkerSuffix}";
+		$baseTitleText = $this->mTitle->getPrefixedDBkey();
+		$tocraw = array();
 
 		foreach( $matches[3] as $headline ) {
-			$istemplate = 0;
-			$templatetitle = '';
-			$templatesection = 0;
+			$isTemplate = false;
+			$titleText = false;
+			$sectionIndex = false;
 			$numbering = '';
-			$mat = array();
-			if (preg_match("//", $headline, $mat)) {
-				$istemplate = 1;
-				$templatetitle = base64_decode($mat[1]);
-				$templatesection = 1 + (int)base64_decode($mat[2]);
-				$headline = preg_replace("//", "", $headline);
+			$markerMatches = array();
+			if (preg_match("/^$markerRegex/", $headline, $markerMatches)) {
+				$serial = $markerMatches[1];
+				list( $titleText, $sectionIndex ) = $this->mHeadings[$serial];
+				$isTemplate = ($titleText != $baseTitleText);
+				$headline = preg_replace("/^$markerRegex/", "", $headline);
 			}
 
 			if( $toclevel ) {
@@ -3568,32 +3940,41 @@ class Parser
 				}
 			}
 
-			# The canonized header is a version of the header text safe to use for links
+			# The safe header is a version of the header text safe to use for links
 			# Avoid insertion of weird stuff like  by expanding the relevant sections
-			$canonized_headline = $this->mStripState->unstripBoth( $headline );
+			$safeHeadline = $this->mStripState->unstripBoth( $headline );
 
 			# Remove link placeholders by the link text.
 			#     
 			# turns into
 			#     link text with suffix
-			$canonized_headline = preg_replace( '//e',
+			$safeHeadline = preg_replace( '//e',
 							    "\$this->mLinkHolders['texts'][\$1]",
-							    $canonized_headline );
-			$canonized_headline = preg_replace( '//e',
+							    $safeHeadline );
+			$safeHeadline = preg_replace( '//e',
 							    "\$this->mInterwikiLinkHolders['texts'][\$1]",
-							    $canonized_headline );
+							    $safeHeadline );
+
+			# Strip out HTML (other than plain  and : bug 8393)
+			$tocline = preg_replace(
+				array( '#<(?!/?(sup|sub)).*?'.'>#', '#<(/?(sup|sub)).*?'.'>#' ),
+				array( '',                          '<$1>'),
+				$safeHeadline
+			);
+			$tocline = trim( $tocline );
+
+			# For the anchor, strip out HTML-y stuff period
+			$safeHeadline = preg_replace( '/<.*?'.'>/', '', $safeHeadline );
+			$safeHeadline = trim( $safeHeadline );
 
-			# strip out HTML
-			$canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
-			$tocline = trim( $canonized_headline );
 			# Save headline for section edit hint before it's escaped
-			$headline_hint = trim( $canonized_headline );
-			$canonized_headline = Sanitizer::escapeId( $tocline );
-			$refers[$headlineCount] = $canonized_headline;
+			$headlineHint = $safeHeadline;
+			$safeHeadline = Sanitizer::escapeId( $safeHeadline );
+			$refers[$headlineCount] = $safeHeadline;
 
 			# count how many in assoc. array so we can track dupes in anchors
-			isset( $refers[$canonized_headline] ) ? $refers[$canonized_headline]++ : $refers[$canonized_headline] = 1;
-			$refcount[$headlineCount]=$refers[$canonized_headline];
+			isset( $refers[$safeHeadline] ) ? $refers[$safeHeadline]++ : $refers[$safeHeadline] = 1;
+			$refcount[$headlineCount] = $refers[$safeHeadline];
 
 			# Don't number the heading if it is the only one (looks silly)
 			if( $doNumberHeadings && count( $matches[3] ) > 1) {
@@ -3602,29 +3983,33 @@ class Parser
 			}
 
 			# Create the anchor for linking from the TOC to the section
-			$anchor = $canonized_headline;
+			$anchor = $safeHeadline;
 			if($refcount[$headlineCount] > 1 ) {
 				$anchor .= '_' . $refcount[$headlineCount];
 			}
 			if( $enoughToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
 				$toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
+				$tocraw[] = array( 'toclevel' => $toclevel, 'level' => $level, 'line' => $tocline, 'number' => $numbering );
 			}
 			# give headline the correct  tag
-			if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
-				if( $istemplate )
-					$editlink = $sk->editSectionLinkForOther($templatetitle, $templatesection);
-				else
-					$editlink = $sk->editSectionLink($this->mTitle, $sectionCount+1, $headline_hint);
+			if( $showEditLink && $sectionIndex !== false ) {
+				if( $isTemplate ) {
+					# Put a T flag in the section identifier, to indicate to extractSections() 
+					# that sections inside  should be counted.
+					$editlink = $sk->editSectionLinkForOther($titleText, "T-$sectionIndex");
+				} else {
+					$editlink = $sk->editSectionLink($this->mTitle, $sectionIndex, $headlineHint);
+				}
 			} else {
 				$editlink = '';
 			}
 			$head[$headlineCount] = $sk->makeHeadline( $level, $matches['attrib'][$headlineCount], $anchor, $headline, $editlink );
 
 			$headlineCount++;
-			if( !$istemplate )
-				$sectionCount++;
 		}
 
+		$this->mOutput->setSections( $tocraw );
+
 		# Never ever show TOC if no headers
 		if( $numVisible < 1 ) {
 			$enoughToc = false;
@@ -3690,14 +4075,12 @@ class Parser
 			$this->clearState();
 		}
 
-		$stripState = new StripState;
 		$pairs = array(
 			"\r\n" => "\n",
 		);
 		$text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
-		$text = $this->strip( $text, $stripState, true, array( 'gallery' ) );
-		$text = $this->pstPass2( $text, $stripState, $user );
-		$text = $stripState->unstripBoth( $text );
+		$text = $this->pstPass2( $text, $user );
+		$text = $this->mStripState->unstripBoth( $text );
 		return $text;
 	}
 
@@ -3705,7 +4088,7 @@ class Parser
 	 * Pre-save transform helper function
 	 * @private
 	 */
-	function pstPass2( $text, &$stripState, $user ) {
+	function pstPass2( $text, $user ) {
 		global $wgContLang, $wgLocaltimezone;
 
 		/* Note: This is the timestamp saved as hardcoded wikitext to
@@ -3728,7 +4111,7 @@ class Parser
 		$text = $this->replaceVariables( $text );
 
 		# Strip out  etc. added via replaceVariables
-		$text = $this->strip( $text, $stripState, false, array( 'gallery' ) );
+		#$text = $this->strip( $text, $this->mStripState, false, array( 'gallery' ) );
 
 		# Signatures
 		$sigText = $this->getUserSig( $user );
@@ -3778,11 +4161,16 @@ class Parser
 	 * @private
 	 */
 	function getUserSig( &$user ) {
+		global $wgMaxSigChars;
+		
 		$username = $user->getName();
 		$nickname = $user->getOption( 'nickname' );
 		$nickname = $nickname === '' ? $username : $nickname;
-
-		if( $user->getBoolOption( 'fancysig' ) !== false ) {
+		
+		if( mb_strlen( $nickname ) > $wgMaxSigChars ) {
+			$nickname = $username;
+			wfDebug( __METHOD__ . ": $username has overlong signature.\n" );
+		} elseif( $user->getBoolOption( 'fancysig' ) !== false ) {
 			# Sig. might contain markup; validate this
 			if( $this->validateSig( $nickname ) !== false ) {
 				# Validated; clean up (if needed) and return it
@@ -3798,8 +4186,13 @@ class Parser
 		$nickname = $this->cleanSigInSig( $nickname );
 
 		# If we're still here, make it a link to the user page
-		$userpage = $user->getUserPage();
-		return( '[[' . $userpage->getPrefixedText() . '|' . wfEscapeWikiText( $nickname ) . ']]' );
+		$userText = wfEscapeWikiText( $username );
+		$nickText = wfEscapeWikiText( $nickname );
+		if ( $user->isAnon() )  {
+			return wfMsgExt( 'signature-anon', array( 'content', 'parsemag' ), $userText, $nickText );
+		} else {
+			return wfMsgExt( 'signature', array( 'content', 'parsemag' ), $userText, $nickText );
+		}
 	}
 
 	/**
@@ -3823,18 +4216,27 @@ class Parser
 	 * @return string Signature text
 	 */
 	function cleanSig( $text, $parsing = false ) {
-		global $wgTitle;
-		$this->startExternalParse( $wgTitle, new ParserOptions(), $parsing ? OT_WIKI : OT_MSG );
+		if ( !$parsing ) {
+			global $wgTitle;
+			$this->startExternalParse( $wgTitle, new ParserOptions(), OT_MSG );
+		}
 
+		# FIXME: regex doesn't respect extension tags or nowiki
+		#  => Move this logic to braceSubstitution()
 		$substWord = MagicWord::get( 'subst' );
 		$substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
 		$substText = '{{' . $substWord->getSynonym( 0 );
 
 		$text = preg_replace( $substRegex, $substText, $text );
 		$text = $this->cleanSigInSig( $text );
-		$text = $this->replaceVariables( $text );
+		$dom = $this->preprocessToDom( $text );
+		$frame = new PPFrame( $this );
+		$text = $frame->expand( $dom->documentElement );
+
+		if ( !$parsing ) {
+			$text = $this->mStripState->unstripBoth( $text );
+		}
 
-		$this->clearState();
 		return $text;
 	}
 
@@ -3865,6 +4267,11 @@ class Parser
 	/**
 	 * Transform a MediaWiki message by replacing magic variables.
 	 *
+	 * For some unknown reason, it also expands templates, but only to the 
+	 * first recursion level. This is wrong and broken, probably introduced 
+	 * accidentally during refactoring, but probably relied upon by thousands 
+	 * of users. 
+	 *
 	 * @param string $text the text to transform
 	 * @param ParserOptions $options  options
 	 * @return string the text with variables substituted
@@ -3893,6 +4300,7 @@ class Parser
 		$this->setOutputType( OT_MSG );
 		$this->clearState();
 		$text = $this->replaceVariables( $text );
+		$text = $this->mStripState->unstripBoth( $text );
 
 		$executing = false;
 		wfProfileOut($fname);
@@ -3918,6 +4326,15 @@ class Parser
 		$tag = strtolower( $tag );
 		$oldVal = isset( $this->mTagHooks[$tag] ) ? $this->mTagHooks[$tag] : null;
 		$this->mTagHooks[$tag] = $callback;
+		$this->mStripList[] = $tag;
+
+		return $oldVal;
+	}
+
+	function setTransparentTagHook( $tag, $callback ) {
+		$tag = strtolower( $tag );
+		$oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null;
+		$this->mTransparentTagHooks[$tag] = $callback;
 
 		return $oldVal;
 	}
@@ -3947,8 +4364,8 @@ class Parser
 	 * @return The old callback function for this name, if any
 	 */
 	function setFunctionHook( $id, $callback, $flags = 0 ) {
-		$oldVal = isset( $this->mFunctionHooks[$id] ) ? $this->mFunctionHooks[$id] : null;
-		$this->mFunctionHooks[$id] = $callback;
+		$oldVal = isset( $this->mFunctionHooks[$id] ) ? $this->mFunctionHooks[$id][0] : null;
+		$this->mFunctionHooks[$id] = array( $callback, $flags );
 
 		# Add to function cache
 		$mw = MagicWord::get( $id );
@@ -3988,10 +4405,7 @@ class Parser
 	/**
 	 * Replace  link placeholders with actual links, in the buffer
 	 * Placeholders created in Skin::makeLinkObj()
-	 * Returns an array of links found, indexed by PDBK:
-	 *  0 - broken
-	 *  1 - normal link
-	 *  2 - stub
+	 * Returns an array of link CSS classes, indexed by PDBK.
 	 * $options is a bit field, RLH_FOR_UPDATE to select for update
 	 */
 	function replaceLinkHolders( &$text, $options = 0 ) {
@@ -4003,6 +4417,7 @@ class Parser
 
 		$pdbks = array();
 		$colours = array();
+		$linkcolour_ids = array();
 		$sk = $this->mOptions->getSkin();
 		$linkCache =& LinkCache::singleton();
 
@@ -4031,12 +4446,14 @@ class Parser
 
 				# Check if it's a static known link, e.g. interwiki
 				if ( $title->isAlwaysKnown() ) {
-					$colours[$pdbk] = 1;
+					$colours[$pdbk] = '';
 				} elseif ( ( $id = $linkCache->getGoodLinkID( $pdbk ) ) != 0 ) {
-					$colours[$pdbk] = 1;
+					$colours[$pdbk] = '';
 					$this->mOutput->addLink( $title, $id );
 				} elseif ( $linkCache->isBadLink( $pdbk ) ) {
-					$colours[$pdbk] = 0;
+					$colours[$pdbk] = 'new';
+				} elseif ( $title->getNamespace() == NS_SPECIAL && !SpecialPage::exists( $pdbk ) ) {
+					$colours[$pdbk] = 'new';
 				} else {
 					# Not in the link cache, add it to the query
 					if ( !isset( $current ) ) {
@@ -4066,25 +4483,17 @@ class Parser
 
 				# Fetch data and form into an associative array
 				# non-existent = broken
-				# 1 = known
-				# 2 = stub
 				while ( $s = $dbr->fetchObject($res) ) {
 					$title = Title::makeTitle( $s->page_namespace, $s->page_title );
 					$pdbk = $title->getPrefixedDBkey();
 					$linkCache->addGoodLinkObj( $s->page_id, $title );
 					$this->mOutput->addLink( $title, $s->page_id );
-
-					if ( $threshold >  0 ) {
-						$size = $s->page_len;
-						if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
-							$colours[$pdbk] = 1;
-						} else {
-							$colours[$pdbk] = 2;
-						}
-					} else {
-						$colours[$pdbk] = 1;
-					}
+					$colours[$pdbk] = $sk->getLinkColour( $s, $threshold );
+					//add id to the extension todolist
+					$linkcolour_ids[$s->page_id] = $pdbk;
 				}
+				//pass an array of page_ids to an extension
+				wfRunHooks( 'GetLinkColours', array( $linkcolour_ids, &$colours ) );
 			}
 			wfProfileOut( $fname.'-check' );
 
@@ -4123,7 +4532,7 @@ class Parser
 				}
 
 				// process categories, check if a category exists in some variant
-				foreach( $categories as $category){
+				foreach( $categories as $category ){
 					$variants = $wgContLang->convertLinkToAllVariants($category);
 					foreach($variants as $variant){
 						if($variant != $category){
@@ -4180,18 +4589,10 @@ class Parser
 
 								// set pdbk and colour
 								$pdbks[$key] = $varPdbk;
-								if ( $threshold >  0 ) {
-									$size = $s->page_len;
-									if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
-										$colours[$varPdbk] = 1;
-									} else {
-										$colours[$varPdbk] = 2;
-									}
-								}
-								else {
-									$colours[$varPdbk] = 1;
-								}
+								$colours[$varPdbk] = $sk->getLinkColour( $s, $threshold );
+								$linkcolour_ids[$s->page_id] = $pdbk;
 							}
+							wfRunHooks( 'GetLinkColours', array( $linkcolour_ids, &$colours ) );
 						}
 
 						// check if the object is a variant of a category
@@ -4224,19 +4625,15 @@ class Parser
 				$pdbk = $pdbks[$key];
 				$searchkey = "";
 				$title = $this->mLinkHolders['titles'][$key];
-				if ( empty( $colours[$pdbk] ) ) {
+				if ( !isset( $colours[$pdbk] ) || $colours[$pdbk] == 'new' ) {
 					$linkCache->addBadLinkObj( $title );
-					$colours[$pdbk] = 0;
+					$colours[$pdbk] = 'new';
 					$this->mOutput->addLink( $title, 0 );
 					$replacePairs[$searchkey] = $sk->makeBrokenLinkObj( $title,
 									$this->mLinkHolders['texts'][$key],
 									$this->mLinkHolders['queries'][$key] );
-				} elseif ( $colours[$pdbk] == 1 ) {
-					$replacePairs[$searchkey] = $sk->makeKnownLinkObj( $title,
-									$this->mLinkHolders['texts'][$key],
-									$this->mLinkHolders['queries'][$key] );
-				} elseif ( $colours[$pdbk] == 2 ) {
-					$replacePairs[$searchkey] = $sk->makeStubLinkObj( $title,
+				} else {
+					$replacePairs[$searchkey] = $sk->makeColouredLinkObj( $title, $colours[$pdbk],
 									$this->mLinkHolders['texts'][$key],
 									$this->mLinkHolders['queries'][$key] );
 				}
@@ -4343,8 +4740,11 @@ class Parser
 		$ig->setContextTitle( $this->mTitle );
 		$ig->setShowBytes( false );
 		$ig->setShowFilename( false );
-		$ig->setParsing();
+		$ig->setParser( $this );
+		$ig->setHideBadImages();
+		$ig->setAttributes( Sanitizer::validateTagAttributes( $params, 'table' ) );
 		$ig->useSkin( $this->mOptions->getSkin() );
+		$ig->mRevisionId = $this->mRevisionId;
 
 		if( isset( $params['caption'] ) ) {
 			$caption = $params['caption'];
@@ -4361,6 +4761,8 @@ class Parser
 		if( isset( $params['heights'] ) ) {
 			$ig->setHeights( $params['heights'] );
 		}
+		
+		wfRunHooks( 'BeforeParserrenderImageGallery', array( &$this, &$ig ) );
 
 		$lines = explode( "\n", $text );
 		foreach ( $lines as $line ) {
@@ -4392,7 +4794,7 @@ class Parser
 			);
 			$html = $pout->getText();
 
-			$ig->add( new Image( $nt ), $html );
+			$ig->add( $nt, $html );
 
 			# Only add real images (bug #5586)
 			if ( $nt->getNamespace() == NS_IMAGE ) {
@@ -4402,10 +4804,50 @@ class Parser
 		return $ig->toHTML();
 	}
 
+	function getImageParams( $handler ) {
+		if ( $handler ) {
+			$handlerClass = get_class( $handler );
+		} else {
+			$handlerClass = '';
+		}
+		if ( !isset( $this->mImageParams[$handlerClass]  ) ) {
+			// Initialise static lists
+			static $internalParamNames = array(
+				'horizAlign' => array( 'left', 'right', 'center', 'none' ),
+				'vertAlign' => array( 'baseline', 'sub', 'super', 'top', 'text-top', 'middle', 
+					'bottom', 'text-bottom' ),
+				'frame' => array( 'thumbnail', 'manualthumb', 'framed', 'frameless', 
+					'upright', 'border' ),
+			);
+			static $internalParamMap;
+			if ( !$internalParamMap ) {
+				$internalParamMap = array();
+				foreach ( $internalParamNames as $type => $names ) {
+					foreach ( $names as $name ) {
+						$magicName = str_replace( '-', '_', "img_$name" );
+						$internalParamMap[$magicName] = array( $type, $name );
+					}
+				}
+			}
+
+			// Add handler params
+			$paramMap = $internalParamMap;
+			if ( $handler ) {
+				$handlerParamMap = $handler->getParamMap();
+				foreach ( $handlerParamMap as $magic => $paramName ) {
+					$paramMap[$magic] = array( 'handler', $paramName );
+				}
+			}
+			$this->mImageParams[$handlerClass] = $paramMap;
+			$this->mImageParamsMagicArray[$handlerClass] = new MagicWordArray( array_keys( $paramMap ) );
+		}
+		return array( $this->mImageParams[$handlerClass], $this->mImageParamsMagicArray[$handlerClass] );
+	}
+
 	/**
 	 * Parse image options text and use it to make an image
 	 */
-	function makeImage( $nt, $options ) {
+	function makeImage( $title, $options ) {
 		# @TODO: let the MediaHandler specify its transform parameters
 		#
 		# Check if the options text is of the form "options|alt text"
@@ -4417,6 +4859,9 @@ class Parser
 		#  * ___px		scale to ___ pixels width, no aligning. e.g. use in taxobox
 		#  * center		center the image
 		#  * framed		Keep original image size, no magnify-button.
+		#  * frameless		like 'thumb' but without a frame. Keeps user preferences for width
+		#  * upright		reduce width for upright images, rounded to full __0 px
+		#  * border		draw a 1px border around the image
 		# vertical-align values (no % or length right now):
 		#  * baseline
 		#  * sub
@@ -4426,67 +4871,66 @@ class Parser
 		#  * middle
 		#  * bottom
 		#  * text-bottom
+		
+		$parts = array_map( 'trim', explode( '|', $options) );
+		$sk = $this->mOptions->getSkin();
 
+		# Give extensions a chance to select the file revision for us
+		$skip = $time = false;
+		wfRunHooks( 'BeforeParserMakeImageLinkObj', array( &$this, &$title, &$skip, &$time ) );
 
-		$part = array_map( 'trim', explode( '|', $options) );
-
-		$mwAlign = array();
-		$alignments = array( 'left', 'right', 'center', 'none', 'baseline', 'sub', 'super', 'top', 'text-top', 'middle', 'bottom', 'text-bottom' );
-		foreach ( $alignments as $alignment ) {
-			$mwAlign[$alignment] =& MagicWord::get( 'img_'.$alignment );
+		if ( $skip ) {
+			return $sk->makeLinkObj( $title );
 		}
-		$mwThumb  =& MagicWord::get( 'img_thumbnail' );
-		$mwManualThumb =& MagicWord::get( 'img_manualthumb' );
-		$mwWidth  =& MagicWord::get( 'img_width' );
-		$mwFramed =& MagicWord::get( 'img_framed' );
-		$mwPage   =& MagicWord::get( 'img_page' );
-		$caption = '';
 
-		$params = array();
-		$framed = $thumb = false;
-		$manual_thumb = '' ;
-		$align = $valign = '';
-		$sk = $this->mOptions->getSkin();
+		# Get parameter map
+		$file = wfFindFile( $title, $time );
+		$handler = $file ? $file->getHandler() : false;
 
-		foreach( $part as $val ) {
-			if ( !is_null( $mwThumb->matchVariableStartToEnd($val) ) ) {
-				$thumb=true;
-			} elseif ( ! is_null( $match = $mwManualThumb->matchVariableStartToEnd($val) ) ) {
-				# use manually specified thumbnail
-				$thumb=true;
-				$manual_thumb = $match;
-			} else {
-				foreach( $alignments as $alignment ) {
-					if ( ! is_null( $mwAlign[$alignment]->matchVariableStartToEnd($val) ) ) {
-						switch ( $alignment ) {
-							case 'left': case 'right': case 'center': case 'none':
-								$align = $alignment; break;
-							default:
-								$valign = $alignment;
-						}
-						continue 2;
-					}
-				}
-				if ( ! is_null( $match = $mwPage->matchVariableStartToEnd($val) ) ) {
-					# Select a page in a multipage document
-					$params['page'] = $match;
-				} elseif ( !isset( $params['width'] ) && ! is_null( $match = $mwWidth->matchVariableStartToEnd($val) ) ) {
-					wfDebug( "img_width match: $match\n" );
-					# $match is the image width in pixels
+		list( $paramMap, $mwArray ) = $this->getImageParams( $handler );
+
+		# Process the input parameters
+		$caption = '';
+		$params = array( 'frame' => array(), 'handler' => array(), 
+			'horizAlign' => array(), 'vertAlign' => array() );
+		foreach( $parts as $part ) {
+			list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part );
+			if ( isset( $paramMap[$magicName] ) ) {
+				list( $type, $paramName ) = $paramMap[$magicName];
+				$params[$type][$paramName] = $value;
+				
+				// Special case; width and height come in one variable together
+				if( $type == 'handler' && $paramName == 'width' ) {
 					$m = array();
-					if ( preg_match( '/^([0-9]*)x([0-9]*)$/', $match, $m ) ) {
-						 $params['width'] = intval( $m[1] );
-						 $params['height'] = intval( $m[2] );
+					if ( preg_match( '/^([0-9]*)x([0-9]*)$/', $value, $m ) ) {
+						$params[$type]['width'] = intval( $m[1] );
+						$params[$type]['height'] = intval( $m[2] );
 					} else {
-						$params['width'] = intval($match);
+						$params[$type]['width'] = intval( $value );
 					}
-				} elseif ( ! is_null( $mwFramed->matchVariableStartToEnd($val) ) ) {
-					$framed=true;
-				} else {
-					$caption = $val;
+				}
+			} else {
+				$caption = $part;
+			}
+		}
+
+		# Process alignment parameters
+		if ( $params['horizAlign'] ) {
+			$params['frame']['align'] = key( $params['horizAlign'] );
+		}
+		if ( $params['vertAlign'] ) {
+			$params['frame']['valign'] = key( $params['vertAlign'] );
+		}
+
+		# Validate the handler parameters
+		if ( $handler ) {
+			foreach ( $params['handler'] as $name => $value ) {
+				if ( !$handler->validateParam( $name, $value ) ) {
+					unset( $params['handler'][$name] );
 				}
 			}
 		}
+
 		# Strip bad stuff out of the alt text
 		$alt = $this->replaceLinkHoldersText( $caption );
 
@@ -4496,8 +4940,18 @@ class Parser
 		$alt = $this->mStripState->unstripBoth( $alt );
 		$alt = Sanitizer::stripAllTags( $alt );
 
+		$params['frame']['alt'] = $alt;
+		$params['frame']['caption'] = $caption;
+
 		# Linker does the rest
-		return $sk->makeImageLinkObj( $nt, $caption, $alt, $align, $params, $framed, $thumb, $manual_thumb, $valign );
+		$ret = $sk->makeImageLink2( $title, $file, $params['frame'], $params['handler'] );
+
+		# Give the handler a chance to modify the parser object
+		if ( $handler ) {
+			$handler->parserTransformHook( $this, $file );
+		}
+
+		return $ret;
 	}
 
 	/**
@@ -4513,12 +4967,12 @@ class Parser
 	 * Callback from the Sanitizer for expanding items found in HTML attribute
 	 * values, so they can be safely tested and escaped.
 	 * @param string $text
-	 * @param array $args
+	 * @param PPFrame $frame
 	 * @return string
 	 * @private
 	 */
-	function attributeStripCallback( &$text, $args ) {
-		$text = $this->replaceVariables( $text, $args );
+	function attributeStripCallback( &$text, $frame = false ) {
+		$text = $this->replaceVariables( $text, $frame );
 		$text = $this->mStripState->unstripBoth( $text );
 		return $text;
 	}
@@ -4536,7 +4990,7 @@ class Parser
 	/**#@+
 	 * Accessor
 	 */
-	function getTags() { return array_keys( $this->mTagHooks ); }
+	function getTags() { return array_merge( array_keys($this->mTransparentTagHooks), array_keys( $this->mTagHooks ) ); }
 	/**#@-*/
 
 
@@ -4546,119 +5000,116 @@ class Parser
 	 *
 	 * External callers should use the getSection and replaceSection methods.
 	 *
-	 * @param $text Page wikitext
-	 * @param $section Numbered section. 0 pulls the text before the first
-	 *                 heading; other numbers will pull the given section
-	 *                 along with its lower-level subsections.
-	 * @param $mode One of "get" or "replace"
-	 * @param $newtext Replacement text for section data.
+	 * @param string $text Page wikitext
+	 * @param string $section A section identifier string of the form:
+	 *    -  - ... - 

+	 *
+	 * Currently the only recognised flag is "T", which means the target section number
+	 * was derived during a template inclusion parse, in other words this is a template 
+	 * section edit link. If no flags are given, it was an ordinary section edit link. 
+	 * This flag is required to avoid a section numbering mismatch when a section is 
+	 * enclosed by  (bug 6563).
+	 *
+	 * The section number 0 pulls the text before the first heading; other numbers will 
+	 * pull the given section along with its lower-level subsections. If the section is 
+	 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
+	 *
+	 * @param string $mode One of "get" or "replace"
+	 * @param string $newText Replacement text for section data.
 	 * @return string for "get", the extracted section text.
 	 *                for "replace", the whole page with the section replaced.
 	 */
-	private function extractSections( $text, $section, $mode, $newtext='' ) {
-		# strip NOWIKI etc. to avoid confusion (true-parameter causes HTML
-		# comments to be stripped as well)
-		$stripState = new StripState;
-
-		$oldOutputType = $this->mOutputType;
-		$oldOptions = $this->mOptions;
-		$this->mOptions = new ParserOptions();
+	private function extractSections( $text, $section, $mode, $newText='' ) {
+		global $wgTitle;
+		$this->clearState();
+		$this->mTitle = $wgTitle; // not generally used but removes an ugly failure mode
+		$this->mOptions = new ParserOptions;
 		$this->setOutputType( OT_WIKI );
-
-		$striptext = $this->strip( $text, $stripState, true );
-
-		$this->setOutputType( $oldOutputType );
-		$this->mOptions = $oldOptions;
-
-		# now that we can be sure that no pseudo-sections are in the source,
-		# split it up by section
-		$uniq = preg_quote( $this->uniqPrefix(), '/' );
-		$comment = "(?:$uniq-!--.*?QINU)";
-		$secs = preg_split(
-			"/
-			(
-				^
-				(?:$comment|<\/?noinclude>)* # Initial comments will be stripped
-				(=+) # Should this be limited to 6?
-				.+?  # Section title...
-				\\2  # Ending = count must match start
-				(?:$comment|<\/?noinclude>|[ \\t]+)* # Trailing whitespace ok
-				$
-			|
-				
-				.*?
-				<\/h\\3\s*>
-			)
-			/mix",
-			$striptext, -1,
-			PREG_SPLIT_DELIM_CAPTURE);
-
-		if( $mode == "get" ) {
-			if( $section == 0 ) {
-				// "Section 0" returns the content before any other section.
-				$rv = $secs[0];
-			} else {
-			  	//track missing section, will replace if found.
-				$rv = $newtext;
-			}
-		} elseif( $mode == "replace" ) {
-			if( $section == 0 ) {
-				$rv = $newtext . "\n\n";
-				$remainder = true;
-			} else {
-				$rv = $secs[0];
-				$remainder = false;
+		$curIndex = 0;
+		$outText = '';
+		$frame = new PPFrame( $this );
+
+		// Process section extraction flags
+		$flags = 0;
+		$sectionParts = explode( '-', $section );
+		$sectionIndex = array_pop( $sectionParts );
+		foreach ( $sectionParts as $part ) {
+			if ( $part == 'T' ) {
+				$flags |= self::PTD_FOR_INCLUSION;
 			}
 		}
-		$count = 0;
-		$sectionLevel = 0;
-		for( $index = 1; $index < count( $secs ); ) {
-			$headerLine = $secs[$index++];
-			if( $secs[$index] ) {
-				// A wiki header
-				$headerLevel = strlen( $secs[$index++] );
-			} else {
-				// An HTML header
-				$index++;
-				$headerLevel = intval( $secs[$index++] );
-			}
-			$content = $secs[$index++];
-
-			$count++;
-			if( $mode == "get" ) {
-				if( $count == $section ) {
-					$rv = $headerLine . $content;
-					$sectionLevel = $headerLevel;
-				} elseif( $count > $section ) {
-					if( $sectionLevel && $headerLevel > $sectionLevel ) {
-						$rv .= $headerLine . $content;
-					} else {
-						// Broke out to a higher-level section
+		// Preprocess the text
+		$dom = $this->preprocessToDom( $text, $flags );
+		$root = $dom->documentElement;
+
+		//  nodes indicate section breaks
+		// They can only occur at the top level, so we can find them by iterating the root's children
+		$node = $root->firstChild;
+
+		// Find the target section
+		if ( $sectionIndex == 0 ) {
+			// Section zero doesn't nest, level=big
+			$targetLevel = 1000;
+		} else {
+			while ( $node ) {
+				if ( $node->nodeName == 'h' ) {
+					if ( $curIndex + 1 == $sectionIndex ) {
 						break;
 					}
+					$curIndex++;
 				}
-			} elseif( $mode == "replace" ) {
-				if( $count < $section ) {
-					$rv .= $headerLine . $content;
-				} elseif( $count == $section ) {
-					$rv .= $newtext . "\n\n";
-					$sectionLevel = $headerLevel;
-				} elseif( $count > $section ) {
-					if( $headerLevel <= $sectionLevel ) {
-						// Passed the section's sub-parts.
-						$remainder = true;
-					}
-					if( $remainder ) {
-						$rv .= $headerLine . $content;
-					}
+				if ( $mode == 'replace' ) {
+					$outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
+				}
+				$node = $node->nextSibling;
+			}
+			if ( $node ) {
+				$targetLevel = $node->getAttribute( 'level' );
+			}
+		}
+
+		if ( !$node ) {
+			// Not found
+			if ( $mode == 'get' ) {
+				return $newText;
+			} else {
+				return $text;
+			}
+		}
+
+		// Find the end of the section, including nested sections
+		do {
+			if ( $node->nodeName == 'h' ) {
+				$curIndex++;
+				$curLevel = $node->getAttribute( 'level' );
+				if ( $curIndex != $sectionIndex && $curLevel <= $targetLevel ) {
+					break;
 				}
 			}
+			if ( $mode == 'get' ) {
+				$outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
+			}
+			$node = $node->nextSibling;
+		} while ( $node );
+		
+		// Write out the remainder (in replace mode only)
+		if ( $mode == 'replace' ) {
+			// Output the replacement text
+			// Add two newlines on -- trailing whitespace in $newText is conventionally 
+			// stripped by the editor, so we need both newlines to restore the paragraph gap
+			$outText .= $newText . "\n\n";
+			while ( $node ) {
+				$outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
+				$node = $node->nextSibling;
+			}
+		}
+
+		if ( is_string( $outText ) ) {
+			// Re-insert stripped tags
+			$outText = trim( $this->mStripState->unstripBoth( $outText ) );
 		}
-		if (is_string($rv))
-			# reinsert stripped tags
-			$rv = trim( $stripState->unstripBoth( $rv ) );
 
-		return $rv;
+		return $outText;
 	}
 
 	/**
@@ -4668,9 +5119,9 @@ class Parser
 	 *
 	 * If a section contains subsections, these are also returned.
 	 *
-	 * @param $text String: text to look in
-	 * @param $section Integer: section number
-	 * @param $deftext: default to return if section is not found
+	 * @param string $text text to look in
+	 * @param string $section section identifier
+	 * @param string $deftext default to return if section is not found
 	 * @return string text of the requested section
 	 */
 	public function getSection( $text, $section, $deftext='' ) {
@@ -4737,6 +5188,70 @@ class Parser
 		}
 	}
 
+	/**
+	 * Try to guess the section anchor name based on a wikitext fragment 
+	 * presumably extracted from a heading, for example "Header" from 
+	 * "== Header ==".
+	 */
+	public function guessSectionNameFromWikiText( $text ) {
+		# Strip out wikitext links(they break the anchor)
+		$text = $this->stripSectionName( $text );
+		$headline = Sanitizer::decodeCharReferences( $text );
+		# strip out HTML
+		$headline = StringUtils::delimiterReplace( '<', '>', '', $headline );
+		$headline = trim( $headline );
+		$sectionanchor = '#' . urlencode( str_replace( ' ', '_', $headline ) );
+		$replacearray = array(
+			'%3A' => ':',
+			'%' => '.'
+		);
+		return str_replace(
+			array_keys( $replacearray ),
+			array_values( $replacearray ),
+			$sectionanchor );
+	}
+
+	/**
+	 * Strips a text string of wikitext for use in a section anchor
+	 * 
+	 * Accepts a text string and then removes all wikitext from the
+	 * string and leaves only the resultant text (i.e. the result of
+	 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
+	 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
+	 * to create valid section anchors by mimicing the output of the
+	 * parser when headings are parsed.
+	 * 
+	 * @param $text string Text string to be stripped of wikitext
+	 * for use in a Section anchor
+	 * @return Filtered text string
+	 */
+	public function stripSectionName( $text ) {
+		# Strip internal link markup
+		$text = preg_replace('/\[\[:?([^[|]+)\|([^[]+)\]\]/','$2',$text);
+		$text = preg_replace('/\[\[:?([^[]+)\|?\]\]/','$1',$text);
+		
+		# Strip external link markup (FIXME: Not Tolerant to blank link text
+		# I.E. [http://www.mediawiki.org] will render as [1] or something depending
+		# on how many empty links there are on the page - need to figure that out.
+		$text = preg_replace('/\[(?:' . wfUrlProtocols() . ')([^ ]+?) ([^[]+)\]/','$2',$text);
+		
+		# Parse wikitext quotes (italics & bold)
+		$text = $this->doQuotes($text);
+		
+		# Strip HTML tags
+		$text = StringUtils::delimiterReplace( '<', '>', '', $text );
+		return $text;
+	}
+
+	/**
+	 * strip/replaceVariables/unstrip for preprocessor regression testing
+	 */
+	function srvus( $text ) {
+		$text = $this->replaceVariables( $text );
+		$text = $this->mStripState->unstripBoth( $text );
+		$text = Sanitizer::removeHTMLtags( $text );
+		return $text;
+	}
 }
 
 /**
@@ -4769,25 +5284,328 @@ class StripState {
 
 	function unstripGeneral( $text ) {
 		wfProfileIn( __METHOD__ );
-		$text = $this->general->replace( $text );
+		do {
+			$oldText = $text;
+			$text = $this->general->replace( $text );
+		} while ( $text != $oldText );
 		wfProfileOut( __METHOD__ );
 		return $text;
 	}
 
 	function unstripNoWiki( $text ) {
 		wfProfileIn( __METHOD__ );
-		$text = $this->nowiki->replace( $text );
+		do {
+			$oldText = $text;
+			$text = $this->nowiki->replace( $text );
+		} while ( $text != $oldText );
 		wfProfileOut( __METHOD__ );
 		return $text;
 	}
 
 	function unstripBoth( $text ) {
 		wfProfileIn( __METHOD__ );
-		$text = $this->general->replace( $text );
-		$text = $this->nowiki->replace( $text );
+		do {
+			$oldText = $text;
+			$text = $this->general->replace( $text );
+			$text = $this->nowiki->replace( $text );
+		} while ( $text != $oldText );
 		wfProfileOut( __METHOD__ );
 		return $text;
 	}
 }
 
-?>
+/**
+ * An expansion frame, used as a context to expand the result of preprocessToDom()
+ */
+class PPFrame {
+	var $parser, $title;
+	var $titleCache;
+
+	const NO_ARGS = 1;
+	const NO_TEMPLATES = 2;
+	const STRIP_COMMENTS = 4;
+	const NO_IGNORE = 8;
+
+	const RECOVER_ORIG = 11;
+
+	/**
+	 * Construct a new preprocessor frame.
+	 * @param Parser $parser The parent parser
+	 * @param Title $title The context title, or false if there isn't one
+	 */
+	function __construct( $parser ) {
+		$this->parser = $parser;
+		$this->title = $parser->mTitle;
+		$this->titleCache = array( $this->title ? $this->title->getPrefixedDBkey() : false );
+	}
+
+	/**
+	 * Create a new child frame
+	 * $args is optionally a DOMNodeList containing the template arguments
+	 */
+	function newChild( $args = false, $title = false ) {
+		$assocArgs = array();
+		if ( $title === false ) {
+			$title = $this->title;
+		}
+		if ( $args !== false ) {
+			$xpath = false;
+			foreach ( $args as $arg ) {
+				if ( !$xpath ) {
+					$xpath = new DOMXPath( $arg->ownerDocument );
+				}
+
+				$nameNodes = $xpath->query( 'name', $arg );
+				if ( $nameNodes->item( 0 )->hasAttributes() ) {
+					// Numbered parameter
+					$name = $nameNodes->item( 0 )->attributes->getNamedItem( 'index' )->textContent;
+				} else {
+					// Named parameter
+					$name = $this->expand( $nameNodes->item( 0 ), PPFrame::STRIP_COMMENTS );
+				}
+
+				$value = $xpath->query( 'value', $arg );
+				$assocArgs[$name] = $value->item( 0 );
+			}
+		}
+		return new PPTemplateFrame( $this->parser, $this, $assocArgs, $title );
+	}
+
+	/**
+	 * Expand a DOMNode describing a preprocessed document into plain wikitext, 
+	 * using the current context
+	 * @param $root the node
+	 */
+	function expand( $root, $flags = 0 ) {
+		if ( is_string( $root ) ) {
+			return $root;
+		}
+
+		if ( $this->parser->ot['html'] 
+			&& ++$this->parser->mPPNodeCount > $this->parser->mOptions->mMaxPPNodeCount ) 
+		{
+			return $this->parser->insertStripItem( '' );
+		}
+
+		if ( is_array( $root ) ) {
+			$s = '';
+			foreach ( $root as $node ) {
+				$s .= $this->expand( $node, $flags );
+			}
+		} elseif ( $root instanceof DOMNodeList ) {
+			$s = '';
+			foreach ( $root as $node ) {
+				$s .= $this->expand( $node, $flags );
+			}
+		} elseif ( $root instanceof DOMNode ) {
+			if ( $root->nodeType == XML_TEXT_NODE ) {
+				$s = $root->nodeValue;
+			} elseif ( $root->nodeName == 'template' ) {
+				# Double-brace expansion
+				$xpath = new DOMXPath( $root->ownerDocument );
+				$titles = $xpath->query( 'title', $root );
+				$title = $titles->item( 0 );
+				$parts = $xpath->query( 'part', $root );
+				if ( $flags & self::NO_TEMPLATES ) {
+					$s = '{{' . $this->implodeWithFlags( '|', $flags, $title, $parts ) . '}}';
+				} else {
+					$lineStart = $root->getAttribute( 'lineStart' );
+					$params = array( 
+						'title' => $title, 
+						'parts' => $parts, 
+						'lineStart' => $lineStart,
+						'text' => 'FIXME' );
+					$s = $this->parser->braceSubstitution( $params, $this );
+				}
+			} elseif ( $root->nodeName == 'tplarg' ) {
+				# Triple-brace expansion
+				$xpath = new DOMXPath( $root->ownerDocument );
+				$titles = $xpath->query( 'title', $root );
+				$title = $titles->item( 0 );
+				$parts = $xpath->query( 'part', $root );
+				if ( $flags & self::NO_ARGS || $this->parser->ot['msg'] ) {
+					$s = '{{{' . $this->implodeWithFlags( '|', $flags, $title, $parts ) . '}}}';
+				} else {
+					$params = array( 'title' => $title, 'parts' => $parts, 'text' => 'FIXME' );
+					$s = $this->parser->argSubstitution( $params, $this );
+				}
+			} elseif ( $root->nodeName == 'comment' ) {
+				# HTML-style comment
+				if ( $this->parser->ot['html'] 
+					|| ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() ) 
+					|| ( $flags & self::STRIP_COMMENTS ) ) 
+				{
+					$s = '';
+				} else {
+					$s = $root->textContent;
+				}
+			} elseif ( $root->nodeName == 'ignore' ) {
+				# Output suppression used by  etc.
+				# OT_WIKI will only respect  in substed templates.
+				# The other output types respect it unless NO_IGNORE is set. 
+				# extractSections() sets NO_IGNORE and so never respects it.
+				if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] ) || ( $flags & self::NO_IGNORE ) ) {
+					$s = $root->textContent;
+				} else {
+					$s = '';
+				}
+			} elseif ( $root->nodeName == 'ext' ) {
+				# Extension tag
+				$xpath = new DOMXPath( $root->ownerDocument );
+				$names = $xpath->query( 'name', $root );
+				$attrs = $xpath->query( 'attr', $root );
+				$inners = $xpath->query( 'inner', $root );
+				$closes = $xpath->query( 'close', $root );
+				$params = array(
+					'name' => $names->item( 0 ),
+					'attr' => $attrs->length > 0 ? $attrs->item( 0 ) : null,
+					'inner' => $inners->length > 0 ? $inners->item( 0 ) : null,
+					'close' => $closes->length > 0 ? $closes->item( 0 ) : null,
+				);
+				$s = $this->parser->extensionSubstitution( $params, $this );
+			} elseif ( $root->nodeName == 'h' ) {
+				# Heading
+				$s = $this->expand( $root->childNodes, $flags );
+
+				if ( $this->parser->ot['html'] ) {
+					# Insert heading index marker
+					$headingIndex = $root->getAttribute( 'i' );
+					$titleText = $this->title->getPrefixedDBkey();
+					$this->parser->mHeadings[] = array( $titleText, $headingIndex );
+					$serial = count( $this->parser->mHeadings ) - 1;
+					$marker = "{$this->parser->mUniqPrefix}-h-$serial-{$this->parser->mMarkerSuffix}";
+					$count = $root->getAttribute( 'level' );
+					$s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
+					$this->parser->mStripState->general->setPair( $marker, '' );
+				}
+			} else {
+				# Generic recursive expansion
+				$s = '';
+				for ( $node = $root->firstChild; $node; $node = $node->nextSibling ) {
+					if ( $node->nodeType == XML_TEXT_NODE ) {
+						$s .= $node->nodeValue;
+					} elseif ( $node->nodeType == XML_ELEMENT_NODE ) {
+						$s .= $this->expand( $node, $flags );
+					}
+				}
+			}
+		} else {
+			throw new MWException( __METHOD__.': Invalid parameter type' );
+		}
+		return $s;
+	}
+
+	function implodeWithFlags( $sep, $flags /*, ... */ ) {
+		$args = array_slice( func_get_args(), 2 );
+
+		$first = true;
+		$s = '';
+		foreach ( $args as $root ) {
+			if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
+				$root = array( $root );
+			}
+			foreach ( $root as $node ) {
+				if ( $first ) {
+					$first = false;
+				} else {
+					$s .= $sep;
+				}
+				$s .= $this->expand( $node, $flags );
+			}
+		}
+		return $s;
+	}
+
+	function implode( $sep /*, ... */ ) {
+		$args = func_get_args();
+		$args = array_merge( array_slice( $args, 0, 1 ), array( 0 ), array_slice( $args, 1 ) );
+		return call_user_func_array( array( $this, 'implodeWithFlags' ), $args );
+	}
+
+	/**
+	 * Split an  or