'('.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
*
function getOptions() { return $this->mOptions; }
/**
- * Replaces all occurrences of <$tag>content</$tag> in the text
- * with a random marker and returns the new text. the output parameter
- * $content will be an associative array filled with data on the form
- * $unique_marker => content.
+ * 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
+ * parameter $matches will be an associative array filled with data in
+ * the form:
+ * 'UNIQ-xxxxx' => array(
+ * 'element',
+ * 'tag content',
+ * array( 'param' => 'x' ),
+ * '<element param="x">tag content</element>' ) )
*
- * If $content is already set, the additional entries will be appended
- * If $tag is set to STRIP_COMMENTS, the function will extract
- * <!-- HTML comments -->
+ * @param $elements list of element names. Comments are always extracted.
+ * @param $text Source text string.
+ * @param $uniq_prefix
*
- * $output: array( 'UNIQ-xxxxx' => array(
- * 'element',
- * 'tag content',
- * array( 'param' => 'x' ),
- * '<element param="x">' ) )
* @private
* @static
*/
$stripped = '';
$matches = array();
- if( $elements == STRIP_COMMENTS ) {
- $start = '/<!--()()/';
- } else {
- $taglist = implode( '|', $elements );
- $start = "/<($taglist)(\\s+[^>]*|\\s*\/?)>/i";
- }
+ $taglist = implode( '|', $elements );
+ $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?>)|<(!--)/i";
while ( '' != $text ) {
$p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
$stripped .= $p[0];
- if( count( $p ) < 4 ) {
+ if( count( $p ) < 5 ) {
break;
}
- $element = $p[1];
- $attributes = $p[2];
- $inside = $p[3];
-
- // If $attributes ends with '/', we have an empty element tag, <tag />
- if( $element != '' && substr( $attributes, -1 ) == '/' ) {
- $attributes = substr( $attributes, 0, -1);
- $empty = '/';
+ if( count( $p ) > 5 ) {
+ // comment
+ $element = $p[4];
+ $attributes = '';
+ $close = '';
+ $inside = $p[5];
} else {
- $empty = '';
+ // tag
+ $element = $p[1];
+ $attributes = $p[2];
+ $close = $p[3];
+ $inside = $p[4];
}
$marker = "$uniq_prefix-$element-$rand" . sprintf('%08X', $n++);
$stripped .= $marker;
- if ( $empty === '/' ) {
+ if ( $close === '/>' ) {
// Empty element tag, <tag />
$content = null;
$text = $inside;
+ $tail = null;
} else {
- if( $element ) {
- $end = "/<\\/$element\\s*>/i";
+ if( $element == '!--' ) {
+ $end = '/(-->)/';
} else {
- $end = '/-->/';
+ $end = "/(<\\/$element\\s*>)/i";
}
- $q = preg_split( $end, $inside, 2 );
+ $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
$content = $q[0];
- if( count( $q ) < 2 ) {
+ if( count( $q ) < 3 ) {
# No end tag -- let it run out to the end of the text.
- break;
+ $tail = '';
+ $text = '';
} else {
- $text = $q[1];
+ $tail = $q[1];
+ $text = $q[2];
}
}
$matches[$marker] = array( $element,
$content,
Sanitizer::decodeTagAttributes( $attributes ),
- "<$element$attributes$empty>" );
+ "<$element$attributes$close$content$tail" );
}
return $stripped;
}
# Replace any instances of the placeholders
$uniq_prefix = $this->mUniqPrefix;
#$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
+ $commentState = array();
$elements = array_merge(
array( 'nowiki', 'pre', 'gallery' ),
}
- // Strip comments in a first pass.
- // This saves us from needlessly rendering extensions in comment text
- $text = Parser::extractTagsAndParams(STRIP_COMMENTS, $text, $comment_matches, $uniq_prefix);
- $commentState = array();
- foreach( $comment_matches as $marker => $data ){
- list( $element, $content, $params, $tag ) = $data;
- $commentState[$marker] = '<!--' . $content . '-->';
- }
-
$matches = array();
$text = Parser::extractTagsAndParams( $elements, $text, $matches, $uniq_prefix );
foreach( $matches as $marker => $data ) {
list( $element, $content, $params, $tag ) = $data;
- // Restore any comments; the extension can deal with them.
- if( $content !== null) {
- $content = strtr( $content, $commentState );
- }
if( $render ) {
- switch( $element ) {
+ $tagName = strtolower( $element );
+ 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;
$output = $this->renderImageGallery( $content );
break;
default:
- $tagName = strtolower( $element );
if( isset( $this->mTagHooks[$tagName] ) ) {
$output = call_user_func_array( $this->mTagHooks[$tagName],
array( $content, $params, $this ) );
}
} else {
// Just stripping tags; keep the source
- if( $content === null ) {
- $output = $tag;
- } else {
- $output = "$tag$content</$element>";
- }
+ $output = $tag;
+ }
+ if( !$stripcomments && $element == '!--' ) {
+ $commentState[$marker] = $output;
+ } else {
+ $state[$element][$marker] = $output;
}
- $state[$element][$marker] = $output;
}
# 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 ) {
- // Add remaining comments to the state array
- foreach( $commentState as $marker => $content ) {
- $state['comment'][$marker] = $content;
- }
- } else {
+ if ( !$stripcomments ) {
// Put them all back and forget them
$text = strtr( $text, $commentState );
}
wfProfileIn( "$fname-paragraph" );
# No prefix (not in list)--go to paragraph mode
// XXX: use a stack for nestable elements like span, table and div
- $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
+ $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<ol|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
$closematch = preg_match(
'/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
- '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul)/iS', $t );
+ '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul|<\\/ol)/iS', $t );
if ( $openmatch or $closematch ) {
$paragraphStack = false;
# TODO bug 5718: paragraph closed
}
/**
- * Split up a string on ':', ignoring any occurences inside
- * <a>..</a> or <span>...</span>
+ * Split up a string on ':', ignoring any occurences inside tags
+ * to prevent illegal overlapping.
* @param string $str the string to split
* @param string &$before set to everything before the ':'
* @param string &$after set to everything after the ':'
* return string the position of the ':', or false if none found
*/
function findColonNoLinks($str, &$before, &$after) {
- # I wonder if we should make this count all tags, not just <a>
- # and <span>. That would prevent us from matching a ':' that
- # comes in the middle of italics other such formatting....
- # -- Wil
$fname = 'Parser::findColonNoLinks';
wfProfileIn( $fname );
- $pos = 0;
- do {
- $colon = strpos($str, ':', $pos);
-
- if ($colon !== false) {
- $before = substr($str, 0, $colon);
- $after = substr($str, $colon + 1);
-
- # Skip any ':' within <a> or <span> pairs
- $a = substr_count($before, '<a');
- $s = substr_count($before, '<span');
- $ca = substr_count($before, '</a>');
- $cs = substr_count($before, '</span>');
-
- if ($a <= $ca and $s <= $cs) {
- # Tags are balanced before ':'; ok
+
+ $pos = strpos( $str, ':' );
+ if( $pos === false ) {
+ // Nothing to find!
+ wfProfileOut( $fname );
+ return false;
+ }
+
+ $lt = strpos( $str, '<' );
+ if( $lt === false || $lt > $pos ) {
+ // Easy; no tag nesting to worry about
+ $before = substr( $str, 0, $pos );
+ $after = substr( $str, $pos+1 );
+ wfProfileOut( $fname );
+ return $pos;
+ }
+
+ // Ugly state machine to walk through avoiding tags.
+ $state = MW_COLON_STATE_TEXT;
+ $stack = 0;
+ $len = strlen( $str );
+ for( $i = 0; $i < $len; $i++ ) {
+ $c = $str{$i};
+
+ switch( $state ) {
+ // (Using the number is a performance hack for common cases)
+ case 0: // MW_COLON_STATE_TEXT:
+ switch( $c ) {
+ case "<":
+ // Could be either a <start> tag or an </end> tag
+ $state = MW_COLON_STATE_TAGSTART;
+ break;
+ case ":":
+ if( $stack == 0 ) {
+ // We found it!
+ $before = substr( $str, 0, $i );
+ $after = substr( $str, $i + 1 );
+ wfProfileOut( $fname );
+ return $i;
+ }
+ // Embedded in a tag; don't break it.
break;
+ default:
+ // Skip ahead looking for something interesting
+ $colon = strpos( $str, ':', $i );
+ if( $colon === false ) {
+ // Nothing else interesting
+ wfProfileOut( $fname );
+ return false;
+ }
+ $lt = strpos( $str, '<', $i );
+ if( $stack === 0 ) {
+ if( $lt === false || $colon < $lt ) {
+ // We found it!
+ $before = substr( $str, 0, $colon );
+ $after = substr( $str, $colon + 1 );
+ wfProfileOut( $fname );
+ return $i;
+ }
+ }
+ if( $lt === false ) {
+ // Nothing else interesting to find; abort!
+ // We're nested, but there's no close tags left. Abort!
+ break 2;
+ }
+ // Skip ahead to next tag start
+ $i = $lt;
+ $state = MW_COLON_STATE_TAGSTART;
}
- $pos = $colon + 1;
+ break;
+ case 1: // MW_COLON_STATE_TAG:
+ // In a <tag>
+ switch( $c ) {
+ case ">":
+ $stack++;
+ $state = MW_COLON_STATE_TEXT;
+ break;
+ case "/":
+ // Slash may be followed by >?
+ $state = MW_COLON_STATE_TAGSLASH;
+ break;
+ default:
+ // ignore
+ }
+ break;
+ case 2: // MW_COLON_STATE_TAGSTART:
+ switch( $c ) {
+ case "/":
+ $state = MW_COLON_STATE_CLOSETAG;
+ break;
+ case "!":
+ $state = MW_COLON_STATE_COMMENT;
+ break;
+ case ">":
+ // Illegal early close? This shouldn't happen D:
+ $state = MW_COLON_STATE_TEXT;
+ break;
+ default:
+ $state = MW_COLON_STATE_TAG;
+ }
+ break;
+ case 3: // MW_COLON_STATE_CLOSETAG:
+ // In a </tag>
+ if( $c == ">" ) {
+ $stack--;
+ if( $stack < 0 ) {
+ wfDebug( "Invalid input in $fname; too many close tags\n" );
+ wfProfileOut( $fname );
+ return false;
+ }
+ $state = MW_COLON_STATE_TEXT;
+ }
+ break;
+ case MW_COLON_STATE_TAGSLASH:
+ if( $c == ">" ) {
+ // Yes, a self-closed tag <blah/>
+ $state = MW_COLON_STATE_TEXT;
+ } else {
+ // Probably we're jumping the gun, and this is an attribute
+ $state = MW_COLON_STATE_TAG;
+ }
+ break;
+ case 5: // MW_COLON_STATE_COMMENT:
+ if( $c == "-" ) {
+ $state = MW_COLON_STATE_COMMENTDASH;
+ }
+ break;
+ case MW_COLON_STATE_COMMENTDASH:
+ if( $c == "-" ) {
+ $state = MW_COLON_STATE_COMMENTDASHDASH;
+ } else {
+ $state = MW_COLON_STATE_COMMENT;
+ }
+ break;
+ case MW_COLON_STATE_COMMENTDASHDASH:
+ if( $c == ">" ) {
+ $state = MW_COLON_STATE_TEXT;
+ } else {
+ $state = MW_COLON_STATE_COMMENT;
+ }
+ break;
+ default:
+ wfDebugDieBacktrace( "State machine error in $fname" );
}
- } while ($colon !== false);
+ }
+ if( $stack > 0 ) {
+ wfDebug( "Invalid input in $fname; not enough close tags (stack $stack, state $state)\n" );
+ return false;
+ }
wfProfileOut( $fname );
- return $colon;
+ return false;
}
/**
# Use the original $piece['title'] not the mangled $part1, so that
# modifiers such as RAW: produce separate cache entries
if( $found ) {
- $this->mTemplates[$piece['title']] = $text;
+ if( $isHTML ) {
+ // A special page; don't store it in the template cache.
+ } else {
+ $this->mTemplates[$piece['title']] = $text;
+ }
$text = $linestart . $text;
}
}
* @return The old value of the mTagHooks array associated with the hook
*/
function setHook( $tag, $callback ) {
+ $tag = strtolower( $tag );
$oldVal = @$this->mTagHooks[$tag];
$this->mTagHooks[$tag] = $callback;
* 'A tree'.
*/
function renderImageGallery( $text ) {
- # Setup the parser
- $parserOptions = new ParserOptions;
- $localParser = new Parser();
-
$ig = new ImageGallery();
$ig->setShowBytes( false );
$ig->setShowFilename( false );
$label = '';
}
- $pout = $localParser->parse( $label , $this->mTitle, $parserOptions );
+ $pout = $this->parse( $label,
+ $this->mTitle,
+ $this->mOptions,
+ false, // Strip whitespace...?
+ false // Don't clear state!
+ );
$html = $pout->getText();
$ig->add( new Image( $nt ), $html );
if ( $nt->getNamespace() == NS_IMAGE ) {
$this->mOutput->addImage( $nt->getDBkey() );
}
-
- # Register links with the parent parser
- foreach( $pout->getLinks() as $ns => $keys ) {
- foreach( $keys as $dbk => $id )
- $this->mOutput->addLink( Title::makeTitle( $ns, $dbk ), $id );
- }
-
}
return $ig->toHTML();
}
dépôts / lhc / web / wiklou.git / blobdiff