Clarify info action's "search engine status"
[lhc/web/wiklou.git] / includes / parser / Preprocessor_DOM.php
index d0c57ab..c9e16b3 100644 (file)
@@ -72,9 +72,8 @@ class Preprocessor_DOM implements Preprocessor {
                $xml = "<list>";
 
                foreach ( $values as $k => $val ) {
-
                        if ( is_int( $k ) ) {
-                               $xml .= "<part><name index=\"$k\"/><value>" . htmlspecialchars( $val ) ."</value></part>";
+                               $xml .= "<part><name index=\"$k\"/><value>" . htmlspecialchars( $val ) . "</value></part>";
                        } else {
                                $xml .= "<part><name>" . htmlspecialchars( $k ) . "</name>=<value>" . htmlspecialchars( $val ) . "</value></part>";
                        }
@@ -149,26 +148,28 @@ class Preprocessor_DOM implements Preprocessor {
                                        wfDebugLog( "Preprocessor", "Loaded preprocessor XML from memcached (key $cacheKey)" );
                                }
                        }
-               }
-               if ( $xml === false ) {
-                       if ( $cacheable ) {
+                       if ( $xml === false ) {
                                wfProfileIn( __METHOD__ . '-cache-miss' );
                                $xml = $this->preprocessToXml( $text, $flags );
                                $cacheValue = sprintf( "%08d", self::CACHE_VERSION ) . $xml;
                                $wgMemc->set( $cacheKey, $cacheValue, 86400 );
                                wfProfileOut( __METHOD__ . '-cache-miss' );
                                wfDebugLog( "Preprocessor", "Saved preprocessor XML to memcached (key $cacheKey)" );
-                       } else {
-                               $xml = $this->preprocessToXml( $text, $flags );
                        }
-
+               } else {
+                       $xml = $this->preprocessToXml( $text, $flags );
                }
 
+
                // Fail if the number of elements exceeds acceptable limits
                // Do not attempt to generate the DOM
                $this->parser->mGeneratedPPNodeCount += substr_count( $xml, '<' );
                $max = $this->parser->mOptions->getMaxGeneratedPPNodeCount();
                if ( $this->parser->mGeneratedPPNodeCount > $max ) {
+                       if ( $cacheable ) {
+                               wfProfileOut( __METHOD__ . '-cacheable' );
+                       }
+                       wfProfileOut( __METHOD__ );
                        throw new MWException( __METHOD__ . ': generated node count limit exceeded' );
                }
 
@@ -183,6 +184,11 @@ class Preprocessor_DOM implements Preprocessor {
                        // 1 << 19 == XML_PARSE_HUGE, needed so newer versions of libxml2 don't barf when the XML is >256 levels deep
                        $result = $dom->loadXML( $xml, 1 << 19 );
                        if ( !$result ) {
+                               wfProfileOut( __METHOD__ . '-loadXML' );
+                               if ( $cacheable ) {
+                                       wfProfileOut( __METHOD__ . '-cacheable' );
+                               }
+                               wfProfileOut( __METHOD__ );
                                throw new MWException( __METHOD__ . ' generated invalid XML' );
                        }
                }
@@ -355,9 +361,11 @@ class Preprocessor_DOM implements Preprocessor {
                                }
                                // 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.
+
+                                       // To avoid leaving blank lines, when a sequence of
+                                       // space-separated comments is both preceded and followed by
+                                       // a newline (ignoring spaces), then
+                                       // trim leading and trailing spaces and the trailing newline.
 
                                        // Find the end
                                        $endPos = strpos( $text, '-->', $i + 4 );
@@ -369,9 +377,24 @@ class Preprocessor_DOM implements Preprocessor {
                                        } else {
                                                // Search backwards for leading whitespace
                                                $wsStart = $i ? ( $i - strspn( $revText, ' ', $lengthText - $i ) ) : 0;
+
                                                // Search forwards for trailing whitespace
                                                // $wsEnd will be the position of the last space (or the '>' if there's none)
                                                $wsEnd = $endPos + 2 + strspn( $text, ' ', $endPos + 3 );
+
+                                               // Keep looking forward as long as we're finding more
+                                               // comments.
+                                               $comments = array( array( $wsStart, $wsEnd ) );
+                                               while ( substr( $text, $wsEnd + 1, 4 ) == '<!--' ) {
+                                                       $c = strpos( $text, '-->', $wsEnd + 4 );
+                                                       if ( $c === false ) {
+                                                               break;
+                                                       }
+                                                       $c = $c + 2 + strspn( $text, ' ', $c + 3 );
+                                                       $comments[] = array( $wsEnd + 1, $c );
+                                                       $wsEnd = $c;
+                                               }
+
                                                // Eat the line if possible
                                                // TODO: This could theoretically be done if $wsStart == 0, i.e. for comments at
                                                // the overall start. That's not how Sanitizer::removeHTMLcomments() did it, but
@@ -379,14 +402,24 @@ class Preprocessor_DOM implements Preprocessor {
                                                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 );
                                                        }
+
+                                                       // Dump all but the last comment to the accumulator
+                                                       foreach ( $comments as $j => $com ) {
+                                                               $startPos = $com[0];
+                                                               $endPos = $com[1] + 1;
+                                                               if ( $j == ( count( $comments ) - 1) ) {
+                                                                       break;
+                                                               }
+                                                               $inner = substr( $text, $startPos, $endPos - $startPos);
+                                                               $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
+                                                       }
+
                                                        // Do a line-start run next time to look for headings after the comment
                                                        $fakeLineStart = true;
                                                } else {
@@ -397,7 +430,7 @@ class Preprocessor_DOM implements Preprocessor {
 
                                                if ( $stack->top ) {
                                                        $part = $stack->top->getCurrentPart();
-                                                       if ( !(isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 )) {
+                                                       if ( !( isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 ) ) {
                                                                $part->visualEnd = $wsStart;
                                                        }
                                                        // Else comments abutting, no change in visual end
@@ -432,7 +465,7 @@ class Preprocessor_DOM implements Preprocessor {
                                }
 
                                $tagStartPos = $i;
-                               if ( $text[$tagEndPos-1] == '/' ) {
+                               if ( $text[$tagEndPos - 1] == '/' ) {
                                        $attrEnd = $tagEndPos - 1;
                                        $inner = null;
                                        $i = $tagEndPos + 1;
@@ -569,7 +602,7 @@ class Preprocessor_DOM implements Preprocessor {
                                                'open' => $curChar,
                                                'close' => $rule['end'],
                                                'count' => $count,
-                                               'lineStart' => ($i > 0 && $text[$i-1] == "\n"),
+                                               'lineStart' => ( $i > 0 && $text[$i - 1] == "\n" ),
                                        );
 
                                        $stack->push( $piece );
@@ -746,7 +779,7 @@ class PPDStack {
                        $class = $this->elementClass;
                        $this->stack[] = new $class( $data );
                }
-               $this->top = $this->stack[ count( $this->stack ) - 1 ];
+               $this->top = $this->stack[count( $this->stack ) - 1];
                $this->accum =& $this->top->getAccum();
        }
 
@@ -757,7 +790,7 @@ class PPDStack {
                $temp = array_pop( $this->stack );
 
                if ( count( $this->stack ) ) {
-                       $this->top = $this->stack[ count( $this->stack ) - 1 ];
+                       $this->top = $this->stack[count( $this->stack ) - 1];
                        $this->accum =& $this->top->getAccum();
                } else {
                        $this->top = self::$false;
@@ -1014,11 +1047,13 @@ class PPFrame_DOM implements PPFrame {
 
                while ( count( $iteratorStack ) > 1 ) {
                        $level = count( $outStack ) - 1;
-                       $iteratorNode =& $iteratorStack[ $level ];
+                       $iteratorNode =& $iteratorStack[$level];
                        $out =& $outStack[$level];
                        $index =& $indexStack[$level];
 
-                       if ( $iteratorNode instanceof PPNode_DOM ) $iteratorNode = $iteratorNode->node;
+                       if ( $iteratorNode instanceof PPNode_DOM ) {
+                               $iteratorNode = $iteratorNode->node;
+                       }
 
                        if ( is_array( $iteratorNode ) ) {
                                if ( $index >= count( $iteratorNode ) ) {
@@ -1148,9 +1183,7 @@ class PPFrame_DOM implements PPFrame {
 
                                        # Insert a heading marker only for <h> children of <root>
                                        # This is to stop extractSections from going over multiple tree levels
-                                       if ( $contextNode->parentNode->nodeName == 'root'
-                                         && $this->parser->ot['html'] )
-                                       {
+                                       if ( $contextNode->parentNode->nodeName == 'root' && $this->parser->ot['html'] ) {
                                                # Insert heading index marker
                                                $headingIndex = $contextNode->getAttribute( 'i' );
                                                $titleText = $this->title->getPrefixedDBkey();
@@ -1206,7 +1239,9 @@ class PPFrame_DOM implements PPFrame {
                $first = true;
                $s = '';
                foreach ( $args as $root ) {
-                       if ( $root instanceof PPNode_DOM ) $root = $root->node;
+                       if ( $root instanceof PPNode_DOM ) {
+                               $root = $root->node;
+                       }
                        if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
                                $root = array( $root );
                        }