* (which in turn the browser understands, and can display).
*
* <pre>
- * There are four main entry points into the Parser class:
+ * There are five main entry points into the Parser class:
* parse()
* produces HTML output
* preSaveTransform().
* produces altered wiki markup.
- * transformMsg()
- * performs brace substitution on MediaWiki messages
* preprocess()
* removes HTML comments and expands templates
+ * cleanSig()
+ * Cleans a signature before saving it to preferences
+ * extractSections()
+ * Extracts sections from an article for section editing
*
* Globals used:
* objects: $wgLang, $wgContLang
* changes in an incompatible way, so the parser cache
* can automatically discard old data.
*/
- const VERSION = '1.6.2';
+ const VERSION = '1.6.4';
# Flags for Parser::setFunctionHook
# Also available as global constants from Defines.php
const COLON_STATE_COMMENTDASH = 6;
const COLON_STATE_COMMENTDASHDASH = 7;
+ // Flags for preprocessToDom
+ const PTD_FOR_INCLUSION = 1;
+
/**#@+
* @private
*/
# Persistent:
var $mTagHooks, $mTransparentTagHooks, $mFunctionHooks, $mFunctionSynonyms, $mVariables,
$mImageParams, $mImageParamsMagicArray, $mStripList, $mMarkerSuffix,
- $mExtLinkBracketedRegex;
-
+ $mExtLinkBracketedRegex, $mPreprocessor, $mDefaultStripList;
+
+
# Cleared with clearState():
var $mOutput, $mAutonumber, $mDTopen, $mStripState;
var $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
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 $mTplExpandCache; // empty-frame expansion cache
var $mTplRedirCache, $mTplDomCache, $mHeadings;
# Temporary
$this->mTransparentTagHooks = array();
$this->mFunctionHooks = array();
$this->mFunctionSynonyms = array( 0 => array(), 1 => array() );
- $this->mStripList = array( 'nowiki', 'gallery' );
+ $this->mDefaultStripList = $this->mStripList = array( 'nowiki', 'gallery' );
$this->mMarkerSuffix = "-QINU\x7f";
$this->mExtLinkBracketedRegex = '/\[(\b(' . wfUrlProtocols() . ')'.
'[^][<>"\\x00-\\x20\\x7F]+) *([^\]\\x0a\\x0d]*?)\]/S';
+ if ( isset( $conf['preprocessorClass'] ) ) {
+ $this->mPreprocessorClass = $conf['preprocessorClass'];
+ } else {
+ $this->mPreprocessorClass = 'Preprocessor_DOM';
+ }
$this->mFirstCall = true;
}
if ( !$this->mFirstCall ) {
return;
}
+ $this->mFirstCall = false;
wfProfileIn( __METHOD__ );
global $wgAllowDisplayTitle, $wgAllowSlowParserFunctions;
$this->setFunctionHook( 'special', array( 'CoreParserFunctions', 'special' ) );
$this->setFunctionHook( 'defaultsort', array( 'CoreParserFunctions', 'defaultsort' ), SFH_NO_HASH );
$this->setFunctionHook( 'filepath', array( 'CoreParserFunctions', 'filepath' ), SFH_NO_HASH );
+ $this->setFunctionHook( 'tag', array( 'CoreParserFunctions', 'tagObj' ), SFH_OBJECT_ARGS );
if ( $wgAllowDisplayTitle ) {
$this->setFunctionHook( 'displaytitle', array( 'CoreParserFunctions', 'displaytitle' ), SFH_NO_HASH );
}
$this->initialiseVariables();
- $this->mFirstCall = false;
+
+ wfRunHooks( 'ParserFirstCallInit', array( &$this ) );
wfProfileOut( __METHOD__ );
}
$this->mUniqPrefix = "\x7fUNIQ" . Parser::getRandomString();
# Clear these on every parse, bug 4549
- $this->mTemplatePath = array();
$this->mTplExpandCache = $this->mTplRedirCache = $this->mTplDomCache = array();
$this->mShowToc = true;
$this->ot = array(
'html' => $ot == OT_HTML,
'wiki' => $ot == OT_WIKI,
- 'msg' => $ot == OT_MSG,
'pre' => $ot == OT_PREPROCESS,
);
}
+ /**
+ * Set the context title
+ */
+ function setTitle( $t ) {
+ if ( !$t || $t instanceof FakeTitle ) {
+ $t = Title::newFromText( 'NO TITLE' );
+ }
+ if ( strval( $t->getFragment() ) !== '' ) {
+ # Strip the fragment to avoid various odd effects
+ $this->mTitle = clone $t;
+ $this->mTitle->setFragment( '' );
+ } else {
+ $this->mTitle = $t;
+ }
+ }
+
/**
* Accessor for mUniqPrefix.
*
* @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;
}
}
$this->mOptions = $options;
- $this->mTitle =& $title;
+ $this->setTitle( $title );
$oldRevisionId = $this->mRevisionId;
$oldRevisionTimestamp = $this->mRevisionTimestamp;
if( $revid !== null ) {
if ( $this->mOptions->getEnableLimitReport() ) {
$max = $this->mOptions->getMaxIncludeSize();
$limitReport =
+ "NewPP limit report\n" .
"Preprocessor node count: {$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";
$this->clearState();
$this->setOutputType( OT_PREPROCESS );
$this->mOptions = $options;
- $this->mTitle = $title;
+ $this->setTitle( $title );
if( $revid !== null ) {
$this->mRevisionId = $revid;
}
wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
$text = $this->replaceVariables( $text );
- if ( $this->mOptions->getRemoveComments() ) {
- $text = Sanitizer::removeHTMLcomments( $text );
- }
$text = $this->mStripState->unstripBoth( $text );
wfProfileOut( __METHOD__ );
return $text;
return $this->mOptions->getInterfaceMessage() ? $wgLang : $wgContLang;
}
+ /**
+ * Get a preprocessor object
+ */
+ function getPreprocessor() {
+ if ( !isset( $this->mPreprocessor ) ) {
+ $class = $this->mPreprocessorClass;
+ $this->mPreprocessor = new $class( $this );
+ }
+ return $this->mPreprocessor;
+ }
+
/**
* Replaces all occurrences of HTML-style comments and the given tags
* in the text with a random marker and returns teh next text. The output
return $text ;
}
- # Remove <noinclude> tags and <includeonly> sections
- $text = strtr( $text, array( '<onlyinclude>' => '' , '</onlyinclude>' => '' ) );
- $text = strtr( $text, array( '<noinclude>' => '', '</noinclude>' => '') );
- $text = StringUtils::delimiterReplace( '<includeonly>', '</includeonly>', '', $text );
-
$text = $this->replaceVariables( $text );
$text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ), false, array_keys( $this->mTransparentTagHooks ) );
wfRunHooks( 'InternalParseBeforeLinks', array( &$this, &$text, &$this->mStripState ) );
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() ) );
}
/**
- * Parse any parentheses in format ((title|part|part)} and return the document tree
+ * Preprocess some wikitext and return the document tree.
* This is the ghost of replace_variables().
*
* @param string $text The text to parse
+ * @param integer flags Bitwise combination of:
+ * self::PTD_FOR_INCLUSION Handle <noinclude>/<includeonly> 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 and the flags.
+ * 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 preprocessToDom ( $text ) {
- wfProfileIn( __METHOD__ );
- wfProfileIn( __METHOD__.'-makexml' );
-
- static $msgRules, $normalRules;
- if ( !$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;
- }
-
- $extElements = implode( '|', $this->getStripList() );
- // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset
- $extElementsRegex = "/($extElements)(?:\s|\/>|>)|(!--)/iA";
-
- $stack = array(); # Stack of unclosed parentheses
- $stackIndex = -1; # Stack read pointer
-
- $searchBase = implode( '', array_keys( $rules ) ) . '<';
-
- $i = -1; # Input pointer, starts out pointing to a pseudo-newline before the start
- $topAccum = '<root>'; # 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
-
- while ( $i < strlen( $text ) ) {
- if ( $i == -1 ) {
- $found = 'line-start';
- $curChar = '';
- } else {
- # 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 {
- $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;
- // Determine element name
- if ( !preg_match( $extElementsRegex, $text, $matches, 0, $i + 1 ) ) {
- // Element name missing or not listed
- $accum .= '<';
- ++$i;
- continue;
- }
- // Handle comments
- if ( isset( $matches[2] ) && $matches[2] == '!--' ) {
- // HTML comment, scan to end
- $endpos = strpos( $text, '-->', $i + 4 );
- if ( $endpos === false ) {
- // Unclosed comment in input, runs to end
- $accum .= htmlspecialchars( substr( $text, $i ) );
- if ( $this->ot['html'] ) {
- // Close it so later stripping can remove it
- $accum .= htmlspecialchars( '-->' );
- }
- $i = strlen( $text );
- continue;
- }
- $accum .= htmlspecialchars( substr( $text, $i, $endpos - $i + 3 ) );
- #$inner = substr( $text, $i + 4, $endpos - $i - 4 );
- #$accum .= '<ext><name>!--</name><inner>' . htmlspecialchars( $inner ) . '</inner></ext>';
- $i = $endpos + 3;
- 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;
- }
- 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 = '<close>' . htmlspecialchars( $matches[0][0] ) . '</close>';
- } else {
- // No end tag -- let it run out to the end of the text.
- $inner = substr( $text, $tagEndPos + 1 );
- $i = strlen( $text );
- $close = '';
- }
- }
- $accum .= '<ext>';
- if ( $attrEnd <= $attrStart ) {
- $attr = '';
- } else {
- $attr = substr( $text, $attrStart, $attrEnd - $attrStart );
- }
- $accum .= '<name>' . htmlspecialchars( $name ) . '</name>' .
- // Note that the attr element contains the whitespace between name and attribute,
- // this is necessary for precise reconstruction during pre-save transform.
- '<attr>' . htmlspecialchars( $attr ) . '</attr>';
- if ( $inner !== null ) {
- $accum .= '<inner>' . htmlspecialchars( $inner ) . '</inner>';
- }
- $accum .= $close . '</ext>';
- }
-
- 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 accumulator to see if it has a proper close
- // No efficient way to do this in PHP AFAICT: strrev, PCRE search with $ anchor
- // and rtrim are all O(N) in total size. Optimal would be O(N) in trailing
- // whitespace size only.
- $m = false;
- $count = $piece['count'];
- if ( preg_match( "/(={{$count}})\s*$/", $accum, $m, 0, $count ) ) {
- // Found match, output <h>
- $count = min( strlen( $m[1] ), $count );
- $element = "<h level=\"$count\" i=\"$headingIndex\">$accum</h>";
- $headingIndex++;
- } else {
- // No match, no <h>, 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
- $count = strspn( $text, $curChar, $i );
-
- # we need to add to stack only if opening brace count is enough for one of the rules
- 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 ) );
- }
- $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;
- $rule = $rules[$piece['open']];
- if ( $count > $rule['max'] ) {
- # The specified maximum exists in the callback array, unless the caller
- # has made an error
- $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, $rule['names'] ) ) {
- --$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;
- }
- $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 {
- # 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>";
- $element .= "<title>$title</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 ) );
- $element .= "<part>$ws1<name>$argName</name>$ws2=$ws3<value>$argValue</value>$ws4</part>";
- } else {
- list( $ws1, $value, $ws2 ) = self::splitWhitespace( $part );
- $element .= "<part>$ws1<name index=\"$argIndex\" /><value>$value</value>$ws2</part>";
- $argIndex++;
- }
- }
- $element .= "</$name>";
- }
-
- # Advance input pointer
- $i += $matchingCount;
-
- # Unwind the stack
- unset( $stack[$stackIndex--] );
- if ( $stackIndex == -1 ) {
- $accum =& $topAccum;
- $findEquals = false;
- $findPipe = false;
- } else {
- $partCount = count( $stack[$stackIndex]['parts'] );
- $accum =& $stack[$stackIndex]['parts'][$partCount - 1];
- $findPipe = $stack[$stackIndex]['open'] != "\n";
- $findEquals = $findPipe && $partCount > 1
- && !isset( $stack[$stackIndex]['eqpos'][$partCount - 1] );
- }
-
- # Re-add the old stack element if it still has unmatched opening characters remaining
- if ($matchingCount < $piece['count']) {
- $piece['parts'] = array( '' );
- $piece['count'] -= $matchingCount;
- $piece['eqpos'] = array();
- # do we still qualify for any callback with remaining count?
- $names = $rules[$piece['open']]['names'];
- $skippedBraces = 0;
- $enclosingAccum =& $accum;
- while ( $piece['count'] ) {
- 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 );
- }
-
- # 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 .= '</root>';
-
- wfProfileOut( __METHOD__.'-makexml' );
- wfProfileIn( __METHOD__.'-loadXML' );
- $dom = new DOMDocument;
- if ( !$dom->loadXML( $topAccum ) ) {
- throw new MWException( __METHOD__.' generated invalid XML' );
- }
- wfProfileOut( __METHOD__.'-loadXML' );
- wfProfileOut( __METHOD__ );
+ function preprocessToDom ( $text, $flags = 0 ) {
+ $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
return $dom;
}
*
* Note that the substitution depends on value of $mOutputType:
* OT_WIKI: only {{subst:}} templates
- * OT_MSG: only magic variables
- * OT_HTML: all templates and magic variables
+ * OT_PREPROCESS: templates but not extension tags
+ * OT_HTML: all templates and extension tags
*
* @param string $tex The text to transform
* @param PPFrame $frame Object describing the arguments passed to the template
wfProfileIn( $fname );
if ( $frame === false ) {
- $frame = new PPFrame( $this );
+ $frame = $this->getPreprocessor()->newFrame();
} elseif ( !( $frame instanceof PPFrame ) ) {
throw new MWException( __METHOD__ . ' called using the old argument format' );
}
$dom = $this->preprocessToDom( $text );
$flags = $argsOnly ? PPFrame::NO_TEMPLATES : 0;
- $text = $frame->expand( $dom, 0, $flags );
+ $text = $frame->expand( $dom, $flags );
wfProfileOut( $fname );
return $text;
* replacing any variables or templates within the template.
*
* @param array $piece The parts of the template
- * $piece['text']: matched text
* $piece['title']: the title, i.e. the part before the |
* $piece['parts']: the parameter array
+ * $piece['lineStart']: whether the brace was at the start of a line
* @param PPFrame The current frame, contains template arguments
* @return string the text of the template
* @private
# 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
$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
+ $isChildObj = false; # $text is a DOM node needing expansion in a child frame
+ $isLocalObj = false; # $text is a DOM node needing expansion in the current frame
# Title object, where $text came from
$title = NULL;
# 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 = '{{' . $frame->implode( '|', $titleWithSpaces, $args ) . '}}';
+ $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
+ $isLocalObj = true;
$found = true;
- $noparse = true;
- $noargs = true;
}
}
# Variables
- if ( !$found && $args->length == 0 ) {
+ if ( !$found && $args->getLength() == 0 ) {
$id = $this->mVariables->matchStartToEnd( $part1 );
if ( $id !== false ) {
$text = $this->getVariableValue( $id );
- $this->mOutput->mContainsOldMagic = true;
+ if (MagicWord::getCacheTTL($id)>-1)
+ $this->mOutput->mContainsOldMagic = true;
$found = true;
- $noparse = true;
- $noargs = true;
}
}
}
wfProfileOut( __METHOD__.'-modifiers' );
- # Save path level before recursing into functions & templates.
- $lastPathLevel = $this->mTemplatePath;
-
# Parser functions
if ( !$found ) {
wfProfileIn( __METHOD__ . '-pfunc' );
# Add a frame parameter, and pass the arguments as an array
$allArgs = $initialArgs;
$allArgs[] = $frame;
- foreach ( $args as $arg ) {
- $funcArgs[] = $arg;
+ for ( $i = 0; $i < $args->getLength(); $i++ ) {
+ $funcArgs[] = $args->item( $i );
}
$allArgs[] = $funcArgs;
} else {
# Convert arguments to plain text
- foreach ( $args as $arg ) {
- $funcArgs[] = trim( $frame->expand( $arg ) );
+ for ( $i = 0; $i < $args->getLength(); $i++ ) {
+ $funcArgs[] = trim( $frame->expand( $args->item( $i ) ) );
}
$allArgs = array_merge( $initialArgs, $funcArgs );
}
$result = call_user_func_array( $callback, $allArgs );
$found = true;
- // The text is usually already parsed, doesn't need triple-brace tags expanded, etc.
- $noargs = true;
- $noparse = true;
-
if ( is_array( $result ) ) {
if ( isset( $result[0] ) ) {
$text = $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 = $result;
$wgContLang->findVariantLink($part1, $title);
}
# Do infinite loop check
- if ( isset( $this->mTemplatePath[$titleText] ) ) {
- $noparse = true;
- $noargs = true;
+ if ( !$frame->loopCheck( $title ) ) {
$found = true;
- $text = "[[$part1]]" . $this->insertStripItem( '<!-- WARNING: template loop detected -->' );
+ $text = "<span class=\"error\">Template loop detected: [[$titleText]]</span>";
wfDebug( __METHOD__.": template loop broken at '$titleText'\n" );
}
+ # Do recursion depth check
+ $limit = $this->mOptions->getMaxTemplateDepth();
+ if ( $frame->depth >= $limit ) {
+ $found = true;
+ $text = "<span class=\"error\">Template recursion depth limit exceeded ($limit)</span>";
+ }
}
}
$text = SpecialPage::capturePath( $title );
if ( is_string( $text ) ) {
$found = true;
- $noparse = true;
- $noargs = true;
$isHTML = true;
$this->disableCache();
}
list( $text, $title ) = $this->getTemplateDom( $title );
if ( $text !== false ) {
$found = true;
- $isDOM = true;
+ $isChildObj = true;
}
}
if ( $this->ot['html'] && !$forceRawInterwiki ) {
$text = $this->interwikiTransclude( $title, 'render' );
$isHTML = true;
- $noparse = true;
} else {
$text = $this->interwikiTransclude( $title, 'raw' );
+ // Preprocess it like a template
+ $text = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
+ $isChildObj = true;
}
$found = true;
}
wfProfileOut( __METHOD__ . '-loadtpl' );
}
- # Recursive parsing, escaping and link table handling
- # Only for HTML output
- if ( $nowiki && $found && ( $this->ot['html'] || $this->ot['pre'] ) ) {
- if ( $isDOM ) {
- $text = $frame->expand( $text );
- }
- $text = wfEscapeWikiText( $text );
- } elseif ( !$this->ot['msg'] && $found ) {
- if ( $noargs ) {
- $newFrame = $frame->newChild();
- } else {
- # Clean up argument array
- $newFrame = $frame->newChild( $args, $title );
- # Add a new element to the templace recursion path
- $this->mTemplatePath[$titleText] = 1;
- }
-
- if ( !$noparse ) {
- if ( $isDOM ) {
- 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 {
- $text = $newFrame->expand( $text );
- }
- } else {
- $text = $this->replaceVariables( $text, $newFrame );
- }
+ # If we haven't found text to substitute by now, we're done
+ # Recover the source wikitext and return it
+ if ( !$found ) {
+ $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
+ wfProfileOut( $fname );
+ return array( 'object' => $text );
+ }
- # strip woz 'ere 2004-07
+ # Expand DOM-style return values in a child frame
+ if ( $isChildObj ) {
+ # Clean up argument array
+ $newFrame = $frame->newChild( $args, $title );
- # Bug 529: if the template begins with a table or block-level
- # element, it should be treated as beginning a new line.
- # This behaviour is somewhat controversial.
- if (!$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
- if ( $isDOM ) {
- $text = $newFrame->expand( $text, 0, PPFrame::NO_TEMPLATES );
+ if ( $titleText !== false && $newFrame->isEmpty() ) {
+ # Expansion is eligible for the empty-frame cache
+ if ( isset( $this->mTplExpandCache[$titleText] ) ) {
+ $text = $this->mTplExpandCache[$titleText];
} else {
- $text = $this->replaceVariables( $text, $newFrame, true );
+ $text = $newFrame->expand( $text );
+ $this->mTplExpandCache[$titleText] = $text;
}
- } elseif ( $isDOM ) {
- $text = $frame->expand( $text );
+ } else {
+ # Uncached expansion
+ $text = $newFrame->expand( $text );
}
- } elseif ( $isDOM ) {
- $text = $frame->expand( $text, 0, PPFrame::NO_TEMPLATES | PPFrame::NO_ARGS );
}
- # Prune lower levels off the recursion check path
- $this->mTemplatePath = $lastPathLevel;
-
- if ( $found && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
+ # 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 ( is_string( $text ) && !$piece['lineStart'] && preg_match('/^(?:{\\||:|;|#|\*)/', $text)) /*}*/{
+ $text = "\n" . $text;
+ }
+
+ if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
# Error, oversize inclusion
$text = "[[$originalTitle]]" .
$this->insertStripItem( '<!-- WARNING: template omitted, post-expand include size too large -->' );
- $noparse = true;
- $noargs = true;
}
- if ( !$found ) {
- wfProfileOut( $fname );
- return '{{' . $frame->implode( '|', $titleWithSpaces, $args ) . '}}';
+ if ( $isLocalObj ) {
+ $ret = array( 'object' => $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 );
- }
- wfProfileOut( __METHOD__ . '-placeholders' );
+ $ret = array( 'text' => $text );
}
- # Prune lower levels off the recursion check path
- $this->mTemplatePath = $lastPathLevel;
-
- if ( !$found ) {
- wfProfileOut( $fname );
- return '{{' . $frame->implode( '|', $titleWithSpaces, $args ) . '}}';
- } else {
- wfProfileOut( $fname );
- return $text;
- }
+ wfProfileOut( $fname );
+ return $ret;
}
/**
return array( false, $title );
}
- # If there are any <onlyinclude> tags, only include them
- if ( !$this->ot['msg'] ) {
- if ( in_string( '<onlyinclude>', $text ) && in_string( '</onlyinclude>', $text ) ) {
- $replacer = new OnlyIncludeReplacer;
- StringUtils::delimiterReplaceCallback( '<onlyinclude>', '</onlyinclude>',
- array( &$replacer, 'replace' ), $text );
- $text = $replacer->output;
- }
- # Remove <noinclude> sections and <includeonly> tags
- $text = StringUtils::delimiterReplace( '<noinclude>', '</noinclude>', '', $text );
- $text = strtr( $text, array( '<includeonly>' => '' , '</includeonly>' => '' ) );
- }
-
- $dom = $this->preprocessToDom( $text );
-
+ $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
$this->mTplDomCache[ $titleText ] = $dom;
if (! $title->equals($cacheTitle)) {
function argSubstitution( $piece, $frame ) {
wfProfileIn( __METHOD__ );
- $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 ) );
+ $nameWithSpaces = $frame->expand( $piece['title'] );
+ $argName = trim( $nameWithSpaces );
+ $object = false;
+ $text = $frame->getArgument( $argName );
+ if ( $text === false && ( $this->ot['html'] || $this->ot['pre'] ) && $parts->getLength() > 0 ) {
+ # No match in frame, use the supplied default
+ $object = $parts->item( 0 )->getChildren();
}
if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
$error = '<!-- WARNING: argument omitted, expansion size too large -->';
}
- if ( $text === false ) {
- $text = '{{{' . $frame->implode( '|', $argWithSpaces, $parts ) . '}}}';
+ if ( $text === false && $object === false ) {
+ # No match anywhere
+ $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts );
}
if ( $error !== false ) {
$text .= $error;
}
+ if ( $object !== false ) {
+ $ret = array( 'object' => $object );
+ } else {
+ $ret = array( 'text' => $text );
+ }
wfProfileOut( __METHOD__ );
- return $text;
+ return $ret;
}
/**
* 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
+ * name PPNode for the tag name
+ * attr PPNode for unparsed text where tag attributes are thought to be
+ * attributes Optional associative array of parsed attributes
* inner Contents of extension element
* noClose Original text did not have a close tag
* @param PPFrame $frame
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'] );
+ $attrText = !isset( $params['attr'] ) ? null : $frame->expand( $params['attr'] );
+ $content = !isset( $params['inner'] ) ? null : $frame->expand( $params['inner'] );
$marker = "{$this->mUniqPrefix}-$name-" . sprintf('%08X', $n++) . $this->mMarkerSuffix;
if ( $this->ot['html'] ) {
- if ( $name == '!--' ) {
- return '';
- }
$name = strtolower( $name );
- $params = Sanitizer::decodeTagAttributes( $attrText );
+ $attributes = Sanitizer::decodeTagAttributes( $attrText );
+ if ( isset( $params['attributes'] ) ) {
+ $attributes = $attributes + $params['attributes'];
+ }
switch ( $name ) {
case 'html':
if( $wgRawHtml ) {
break;
case 'math':
$output = $wgContLang->armourMath(
- MathRenderer::renderMath( $content, $params ) );
+ MathRenderer::renderMath( $content, $attributes ) );
break;
case 'gallery':
- $output = $this->renderImageGallery( $content, $params );
+ $output = $this->renderImageGallery( $content, $attributes );
break;
default:
if( isset( $this->mTagHooks[$name] ) ) {
$output = call_user_func_array( $this->mTagHooks[$name],
- array( $content, $params, $this ) );
+ array( $content, $attributes, $this ) );
} else {
throw new MWException( "Invalid call hook $name" );
}
}
} else {
- if ( $name == '!--' ) {
- $output = '<!--' . $content . '-->';
- } else {
- if ( $content === null ) {
- $output = "<$name$attrText/>";
- } else {
- $close = is_null( $params['close'] ) ? '' : $frame->expand( $params['close'] );
- $output = "<$name$attrText>$content$close";
+ if ( is_null( $attrText ) ) {
+ $attrText = '';
+ }
+ if ( isset( $params['attributes'] ) ) {
+ foreach ( $params['attributes'] as $attrName => $attrValue ) {
+ $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' .
+ htmlspecialchars( $attrValue ) . '"';
}
}
+ 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' ) {
$prevtoclevel = 0;
$markerRegex = "{$this->mUniqPrefix}-h-(\d+)-{$this->mMarkerSuffix}";
$baseTitleText = $this->mTitle->getPrefixedDBkey();
+ $tocraw = array();
foreach( $matches[3] as $headline ) {
$isTemplate = false;
}
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 <h#> tag
if( $showEditLink && $sectionIndex !== false ) {
- if( $isTemplate )
- $editlink = $sk->editSectionLinkForOther($titleText, $sectionIndex);
- else
+ if( $isTemplate ) {
+ # Put a T flag in the section identifier, to indicate to extractSections()
+ # that sections inside <includeonly> should be counted.
+ $editlink = $sk->editSectionLinkForOther($titleText, "T-$sectionIndex");
+ } else {
$editlink = $sk->editSectionLink($this->mTitle, $sectionIndex, $headlineHint);
+ }
} else {
$editlink = '';
}
$headlineCount++;
}
+ $this->mOutput->setSections( $tocraw );
+
# Never ever show TOC if no headers
if( $numVisible < 1 ) {
$enoughToc = false;
*/
function preSaveTransform( $text, &$title, $user, $options, $clearState = true ) {
$this->mOptions = $options;
- $this->mTitle =& $title;
+ $this->setTitle( $title );
$this->setOutputType( OT_WIKI );
if ( $clearState ) {
function cleanSig( $text, $parsing = false ) {
if ( !$parsing ) {
global $wgTitle;
- $this->startExternalParse( $wgTitle, new ParserOptions(), OT_MSG );
+ $this->clearState();
+ $this->setTitle( $wgTitle );
+ $this->mOptions = new ParserOptions;
+ $this->setOutputType = OT_PREPROCESS;
}
# FIXME: regex doesn't respect extension tags or nowiki
$text = preg_replace( $substRegex, $substText, $text );
$text = $this->cleanSigInSig( $text );
$dom = $this->preprocessToDom( $text );
- $frame = new PPFrame( $this );
- $text = $frame->expand( $dom->documentElement );
+ $frame = $this->getPreprocessor()->newFrame();
+ $text = $frame->expand( $dom );
if ( !$parsing ) {
$text = $this->mStripState->unstripBoth( $text );
* @public
*/
function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
- $this->mTitle =& $title;
+ $this->setTitle( $title );
$this->mOptions = $options;
$this->setOutputType( $outputType );
if ( $clearState ) {
}
/**
- * Transform a MediaWiki message by replacing magic variables.
+ * Wrapper for preprocess()
*
- * 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 string $text the text to preprocess
* @param ParserOptions $options options
- * @return string the text with variables substituted
+ * @return string
* @public
*/
function transformMsg( $text, $options ) {
$executing = true;
wfProfileIn($fname);
-
- if ( $wgTitle && !( $wgTitle instanceof FakeTitle ) ) {
- $this->mTitle = $wgTitle;
- } else {
- $this->mTitle = Title::newFromText('msg');
- }
- $this->mOptions = $options;
- $this->setOutputType( OT_MSG );
- $this->clearState();
- $text = $this->replaceVariables( $text );
- $text = $this->mStripState->unstripBoth( $text );
+ $text = $this->preprocess( $text, $wgTitle, $options );
$executing = false;
wfProfileOut($fname);
return $oldVal;
}
+ /**
+ * Remove all tag hooks
+ */
+ function clearTagHooks() {
+ $this->mTagHooks = array();
+ $this->mStripList = $this->mDefaultStripList;
+ }
+
/**
* Create a function, e.g. {{sum:1|2|3}}
* The callback function should have the form:
* found The text returned is valid, stop processing the template. This
* is on by default.
* nowiki Wiki markup in the return value should be escaped
- * noparse Unsafe HTML tags should not be stripped, etc.
- * noargs Don't replace triple-brace arguments in the return value
* isHTML The returned text is HTML, armour it against wikitext transformation
*
* @public
$label = '';
}
- $pout = $this->parse( $label,
- $this->mTitle,
- $this->mOptions,
- false, // Strip whitespace...?
- false // Don't clear state!
- );
- $html = $pout->getText();
+ $html = $this->recursiveTagParse( trim( $label ) );
$ig->add( $nt, $html );
*
* 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. If the section is
- * not found, $mode=get will return $newtext, and
- * $mode=replace will return $text.
- * @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:
+ * <flag1> - <flag2> - ... - <section number>
+ *
+ * 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 <includeonly> (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='' ) {
+ global $wgTitle;
$this->clearState();
+ $this->setTitle( $wgTitle ); // not generally used but removes an ugly failure mode
$this->mOptions = new ParserOptions;
$this->setOutputType( OT_WIKI );
$curIndex = 0;
$outText = '';
- $frame = new PPFrame( $this );
-
+ $frame = $this->getPreprocessor()->newFrame();
+
+ // 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;
+ }
+ }
// Preprocess the text
- $dom = $this->preprocessToDom( $text );
- $root = $dom->documentElement;
+ $root = $this->preprocessToDom( $text, $flags );
// <h> 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;
+ $node = $root->getFirstChild();
// Find the target section
- if ( $section == 0 ) {
+ if ( $sectionIndex == 0 ) {
// Section zero doesn't nest, level=big
$targetLevel = 1000;
} else {
while ( $node ) {
- if ( $node->nodeName == 'h' ) {
- if ( $curIndex + 1 == $section ) {
+ if ( $node->getName() == 'h' ) {
+ if ( $curIndex + 1 == $sectionIndex ) {
break;
}
$curIndex++;
}
if ( $mode == 'replace' ) {
- $outText .= $frame->expand( $node, 0, PPFrame::RECOVER_ORIG );
+ $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
}
- $node = $node->nextSibling;
+ $node = $node->getNextSibling();
}
if ( $node ) {
- $targetLevel = $node->getAttribute( 'level' );
+ $bits = $node->splitHeading();
+ $targetLevel = $bits['level'];
}
}
// Find the end of the section, including nested sections
do {
- if ( $node->nodeName == 'h' ) {
+ if ( $node->getName() == 'h' ) {
$curIndex++;
- $curLevel = $node->getAttribute( 'level' );
- if ( $curIndex != $section && $curLevel <= $targetLevel ) {
+ $bits = $node->splitHeading();
+ $curLevel = $bits['level'];
+ if ( $curIndex != $sectionIndex && $curLevel <= $targetLevel ) {
break;
}
}
if ( $mode == 'get' ) {
- $outText .= $frame->expand( $node, 0, PPFrame::RECOVER_ORIG );
+ $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
}
- $node = $node->nextSibling;
+ $node = $node->getNextSibling();
} while ( $node );
// Write out the remainder (in replace mode only)
// stripped by the editor, so we need both newlines to restore the paragraph gap
$outText .= $newText . "\n\n";
while ( $node ) {
- $outText .= $frame->expand( $node, 0, PPFrame::RECOVER_ORIG );
- $node = $node->nextSibling;
+ $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
+ $node = $node->getNextSibling();
}
}
*
* 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='' ) {
return $text;
}
+ function srvus( $text ) {
+ return $this->testSrvus( $text, $this->mOutputType );
+ }
+
/**
* strip/replaceVariables/unstrip for preprocessor regression testing
*/
- function srvus( $text ) {
+ function testSrvus( $text, $title, $options, $outputType = OT_HTML ) {
+ $this->clearState();
+ $this->mTitle = $title;
+ $this->mOptions = $options;
+ $this->setOutputType( $outputType );
$text = $this->replaceVariables( $text );
$text = $this->mStripState->unstripBoth( $text );
+ $text = Sanitizer::removeHTMLtags( $text );
return $text;
}
-}
-/**
- * @todo document, briefly.
- * @addtogroup Parser
- */
-class OnlyIncludeReplacer {
- var $output = '';
+ function testPst( $text, $title, $options ) {
+ global $wgUser;
+ return $this->preSaveTransform( $text, $title, $wgUser, $options );
+ }
- function replace( $matches ) {
- if ( substr( $matches[1], -1 ) == "\n" ) {
- $this->output .= substr( $matches[1], 0, -1 );
- } else {
- $this->output .= $matches[1];
- }
+ function testPreprocess( $text, $title, $options ) {
+ return $this->testSrvus( $text, $title, $options, OT_PREPROCESS );
}
}
}
/**
- * An expansion frame, used as a context to expand the result of preprocessToDom()
+ * @todo document, briefly.
+ * @addtogroup Parser
*/
-class PPFrame {
- var $parser, $title;
-
- const NO_ARGS = 1;
- const NO_TEMPLATES = 2;
- const RECOVER_ORIG = 3;
-
- /**
- * 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;
- }
-
- /**
- * 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 ) );
- }
-
- $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, $shallowFlags = 0, $deepFlags = 0 ) {
- if ( is_string( $root ) ) {
- return $root;
- }
-
- if ( $this->parser->ot['html']
- && ++$this->parser->mPPNodeCount > $this->parser->mOptions->mMaxPPNodeCount )
- {
- return $this->parser->insertStripItem( '<!-- node-count limit exceeded -->' );
- }
- $flags = $shallowFlags | $deepFlags;
+class OnlyIncludeReplacer {
+ var $output = '';
- if ( is_array( $root ) ) {
- $s = '';
- foreach ( $root as $node ) {
- $s .= $this->expand( $node, 0, $deepFlags );
- }
- } elseif ( $root instanceof DOMNodeList ) {
- $s = '';
- foreach ( $root as $node ) {
- $s .= $this->expand( $node, 0, $deepFlags );
- }
- } 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( '|', 0, $deepFlags, $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->implode( '|', 0, $deepFlags, $title, $parts ) . '}}}';
- } else {
- $params = array( 'title' => $title, 'parts' => $parts, 'text' => 'FIXME' );
- $s = $this->parser->argSubstitution( $params, $this );
- }
- } 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, 0, $deepFlags );
-
- 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' );
-
- // FIXME: bug-for-bug with old parser
- // Lose whitespace for no apparent reason
- // Remove this after differential testing is done
- if ( true ) {
- // Good version
- $s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
- } else {
- // Bad version
- if ( preg_match( '/^(={1,6})(.*?)(={1,6})\s*?$/', $s, $m ) ) {
- if ( $m[2] != '' ) {
- $s = $m[1] . $marker . $m[2] . $m[3];
- }
- }
- }
- $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, 0, $deepFlags );
- }
- }
- }
+ function replace( $matches ) {
+ if ( substr( $matches[1], -1 ) == "\n" ) {
+ $this->output .= substr( $matches[1], 0, -1 );
} else {
- throw new MWException( __METHOD__.': Invalid parameter type' );
- }
- return $s;
- }
-
- function implodeWithFlags( $sep, $shallowFlags, $deepFlags /*, ... */ ) {
- $args = array_slice( func_get_args(), 3 );
-
- $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, $shallowFlags, $deepFlags );
- }
- }
- return $s;
- }
-
- function implode( $sep /*, ... */ ) {
- $args = func_get_args();
- $args = array_merge( array_slice( $args, 0, 1 ), array( 0, 0 ), array_slice( $args, 1 ) );
- return call_user_func_array( array( $this, 'implodeWithFlags' ), $args );
- }
-
- function __toString() {
- return 'frame{}';
- }
-}
-
-/**
- * Expansion frame with template arguments
- */
-class PPTemplateFrame extends PPFrame {
- public $parser, $args, $parent, $serial;
-
- function __construct( $parser, $parent = false, $args = array(), $title = false ) {
- $this->parser = $parser;
- $this->parent = $parent;
- $this->args = $args;
- $this->title = $title;
- }
-
- function __toString() {
- $s = 'tplframe{';
- $first = true;
- foreach ( $this->args as $name => $value ) {
- if ( $first ) {
- $first = false;
- } else {
- $s .= ', ';
- }
- $s .= "\"$name\":\"" .
- str_replace( '"', '\\"', $value->ownerDocument->saveXML( $value ) ) . '"';
+ $this->output .= $matches[1];
}
- $s .= '}';
- return $s;
}
}