In QueryPage: fixing a misleading comment. value needn't be numeric.
[lhc/web/wiklou.git] / includes / Parser.php
index c25acec..3c539f0 100644 (file)
@@ -42,10 +42,29 @@ define( "OT_MSG", 3 );
 # to strip HTML comments in addition to regular
 # <XML>-style tags. This should not be anything we
 # may want to use in wikisyntax
-define( "STRIP_COMMENTS", "HTMLCommentStrip" );
+define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
 
 # prefix for escaping, used in two functions at least
-define( "UNIQ_PREFIX", "NaodW29");
+define( 'UNIQ_PREFIX', 'NaodW29');
+
+
+# Constants needed for external link processing
+
+define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
+define( 'HTTP_PROTOCOLS', 'http|https' );
+# Everything except bracket, space, or control characters
+define( 'EXT_LINK_URL_CLASS', '[^]\\x00-\\x20\\x7F]' );
+define( 'INVERSE_EXT_LINK_URL_CLASS', '[\]\\x00-\\x20\\x7F]' );
+# Including space
+define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
+define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
+define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
+define( 'EXT_LINK_BRACKETED',  '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/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
+);
 
 class Parser
 {
@@ -225,7 +244,7 @@ class Parser
                                }
                        }
                }
-               
+
                # nowiki
                $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
                foreach( $nowiki_content as $marker => $content ) {
@@ -326,7 +345,7 @@ class Parser
                for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
                        $text = str_replace( key( $state['nowiki'] ), $content, $text );
                }
-               
+
                global $wgRawHtml;
                if ($wgRawHtml) {
                        for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
@@ -373,18 +392,17 @@ class Parser
 
        # This method generates the list of subcategories and pages for a category
        function oldCategoryMagic () {
-               global $wgLang , $wgUser ;
+               global $wgLang ;
                $fname = 'Parser::oldCategoryMagic';
 
                if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 
-               $cns = Namespace::getCategory() ;
-               if ( $this->mTitle->getNamespace() != $cns ) return "" ; # This ain't a category page
+               if ( $this->mTitle->getNamespace() != NS_CATEGORY ) return "" ; # This ain't a category page
 
                $r = "<br style=\"clear:both;\"/>\n";
 
 
-               $sk =& $wgUser->getSkin() ;
+               $sk =& $this->mOptions->getSkin() ;
 
                $articles = array() ;
                $children = array() ;
@@ -409,7 +427,7 @@ class Parser
                        if ( $t != "" ) $t .= ":" ;
                        $t .= $x->cur_title ;
 
-                       if ( $x->cur_namespace == $cns ) {
+                       if ( $x->cur_namespace == NS_CATEGORY ) {
                                array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
                        } else {
                                array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
@@ -437,16 +455,15 @@ class Parser
 
 
        function newCategoryMagic () {
-               global $wgLang , $wgUser ;
+               global $wgLang;
                if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 
-               $cns = Namespace::getCategory() ;
-               if ( $this->mTitle->getNamespace() != $cns ) return '' ; # This ain't a category page
+               if ( $this->mTitle->getNamespace() != NS_CATEGORY ) return '' ; # This ain't a category page
 
                $r = "<br style=\"clear:both;\"/>\n";
 
 
-               $sk =& $wgUser->getSkin() ;
+               $sk =& $this->mOptions->getSkin() ;
 
                $articles = array() ;
                $articles_start_char = array();
@@ -470,7 +487,7 @@ class Parser
                        if ( $t != '' ) $t .= ':' ;
                        $t .= $x->cur_title ;
 
-                       if ( $x->cur_namespace == $cns ) {
+                       if ( $x->cur_namespace == NS_CATEGORY ) {
                                $ctitle = str_replace( '_',' ',$x->cur_title );
                                array_push ( $children, $sk->makeKnownLink ( $t, $ctitle ) ) ; # Subcategory
 
@@ -642,6 +659,9 @@ class Parser
                        '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
                        "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
                        $t);
+
+               $t = str_replace ( "<></>" , "" , $t ) ; # This should fix bug 980557
+
                # Strip javascript "expression" from stylesheets. Brute force approach:
                # If anythin offensive is found, all attributes of the HTML tag are dropped
 
@@ -821,7 +841,6 @@ class Parser
 
                $text = $this->removeHTMLtags( $text );
                $text = $this->replaceVariables( $text, $args );
-               print $text;
 
                $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 
@@ -831,12 +850,14 @@ class Parser
                        $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
                }
                $text = $this->doAllQuotes( $text );
-               // $text = $this->doExponent( $text );
                $text = $this->replaceExternalLinks( $text );
                $text = $this->doMagicLinks( $text );
                $text = $this->replaceInternalLinks ( $text );
                $text = $this->replaceInternalLinks ( $text );
-               //$text = $this->doTokenizedParser ( $text );
+                               
+               $text = $this->unstrip( $text, $this->mStripState );
+               $text = $this->unstripNoWiki( $text, $this->mStripState );
+
                $text = $this->doTableStuff( $text );
                $text = $this->formatHeadings( $text, $isMain );
                $sk =& $this->mOptions->getSkin();
@@ -846,11 +867,11 @@ class Parser
                        $text .= $this->categoryMagic () ;
                        $this->categoryMagicDone = true ;
                }
-
+               
                wfProfileOut( $fname );
                return $text;
        }
-       
+
        /* private */ function &doMagicLinks( &$text ) {
                $text = $this->magicISBN( $text );
                $text = $this->magicGEO( $text );
@@ -887,57 +908,168 @@ class Parser
                $outtext = '';
                $lines = explode( "\n", $text );
                foreach ( $lines as $line ) {
-                       $outtext .= $this->doQuotes ( '', $line, '' ) . "\n";
+                       $outtext .= $this->doQuotes ( $line ) . "\n";
                }
                $outtext = substr($outtext, 0,-1);
                wfProfileOut( $fname );
                return $outtext;
        }
 
-       /* private */ function doQuotes( $pre, $text, $mode ) {
-               if ( preg_match( "/^(.*)''(.*)$/sU", $text, $m ) ) {
-                       $m1_strong = ($m[1] == "") ? "" : "<strong>{$m[1]}</strong>";
-                       $m1_em = ($m[1] == "") ? "" : "<em>{$m[1]}</em>";
-                       if ( substr ($m[2], 0, 1) == '\'' ) {
-                               $m[2] = substr ($m[2], 1);
-                               if ($mode == 'em') {
-                                       return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'emstrong' );
-                               } else if ($mode == 'strong') {
-                                       return $m1_strong . $this->doQuotes ( '', $m[2], '' );
-                               } else if (($mode == 'emstrong') || ($mode == 'both')) {
-                                       return $this->doQuotes ( '', $pre.$m1_strong.$m[2], 'em' );
-                               } else if ($mode == 'strongem') {
-                                       return "<strong>{$pre}{$m1_em}</strong>" . $this->doQuotes ( '', $m[2], 'em' );
-                               } else {
-                                       return $m[1] . $this->doQuotes ( '', $m[2], 'strong' );
-                               }
-                       } else {
-                               if ($mode == 'strong') {
-                                       return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'strongem' );
-                               } else if ($mode == 'em') {
-                                       return $m1_em . $this->doQuotes ( '', $m[2], '' );
-                               } else if ($mode == 'emstrong') {
-                                       return "<em>{$pre}{$m1_strong}</em>" . $this->doQuotes ( '', $m[2], 'strong' );
-                               } else if (($mode == 'strongem') || ($mode == 'both')) {
-                                       return $this->doQuotes ( '', $pre.$m1_em.$m[2], 'strong' );
-                               } else {
-                                       return $m[1] . $this->doQuotes ( '', $m[2], 'em' );
+       /* private */ function doQuotes( $text ) {
+               $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
+               if (count ($arr) == 1)
+                       return $text;
+               else
+               {
+                       # First, do some preliminary work. This may shift some apostrophes from
+                       # being mark-up to being text. It also counts the number of occurrences
+                       # of bold and italics mark-ups.
+                       $i = 0;
+                       $numbold = 0;
+                       $numitalics = 0;
+                       foreach ($arr as $r)
+                       {
+                               if (($i % 2) == 1)
+                               {
+                                       # If there are ever four apostrophes, assume the first is supposed to
+                                       # be text, and the remaining three constitute mark-up for bold text.
+                                       if (strlen ($arr[$i]) == 4)
+                                       {
+                                               $arr[$i-1] .= "'";
+                                               $arr[$i] = "'''";
+                                       }
+                                       # If there are more than 5 apostrophes in a row, assume they're all
+                                       # text except for the last 5.
+                                       else if (strlen ($arr[$i]) > 5)
+                                       {
+                                               $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
+                                               $arr[$i] = "'''''";
+                                       }
+                                       # Count the number of occurrences of bold and italics mark-ups.
+                                       # We are not counting sequences of five apostrophes.
+                                       if (strlen ($arr[$i]) == 2) $numitalics++;  else
+                                       if (strlen ($arr[$i]) == 3) $numbold++;
                                }
+                               $i++;
                        }
-               } else {
-                       $text_strong = ($text == '') ? '' : "<strong>{$text}</strong>";
-                       $text_em = ($text == '') ? '' : "<em>{$text}</em>";
-                       if ($mode == '') {
-                               return $pre . $text;
-                       } else if ($mode == 'em') {
-                               return $pre . $text_em;
-                       } else if ($mode == 'strong') {
-                               return $pre . $text_strong;
-                       } else if ($mode == 'strongem') {
-                               return (($pre == '') && ($text == '')) ? '' : "<strong>{$pre}{$text_em}</strong>";
-                       } else {
-                               return (($pre == '') && ($text == '')) ? '' : "<em>{$pre}{$text_strong}</em>";
+
+                       # If there is an odd number of both bold and italics, it is likely
+                       # that one of the bold ones was meant to be an apostrophe followed
+                       # by italics. Which one we cannot know for certain, but it is more
+                       # likely to be one that has a single-letter word before it.
+                       if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
+                       {
+                               $i = 0;
+                               $firstsingleletterword = -1;
+                               $firstmultiletterword = -1;
+                               $firstspace = -1;
+                               foreach ($arr as $r)
+                               {
+                                       if (($i % 2 == 1) and (strlen ($r) == 3))
+                                       {
+                                               $x1 = substr ($arr[$i-1], -1);
+                                               $x2 = substr ($arr[$i-1], -2, 1);
+                                               if ($x1 == " ") {
+                                                       if ($firstspace == -1) $firstspace = $i;
+                                               } else if ($x2 == " ") {
+                                                       if ($firstsingleletterword == -1) $firstsingleletterword = $i;
+                                               } else {
+                                                       if ($firstmultiletterword == -1) $firstmultiletterword = $i;
+                                               }
+                                       }
+                                       $i++;
+                               }
+
+                               # If there is a single-letter word, use it!
+                               if ($firstsingleletterword > -1)
+                               {
+                                       $arr [ $firstsingleletterword ] = "''";
+                                       $arr [ $firstsingleletterword-1 ] .= "'";
+                               }
+                               # If not, but there's a multi-letter word, use that one.
+                               else if ($firstmultiletterword > -1)
+                               {
+                                       $arr [ $firstmultiletterword ] = "''";
+                                       $arr [ $firstmultiletterword-1 ] .= "'";
+                               }
+                               # ... otherwise use the first one that has neither.
+                               else
+                               {
+                                       $arr [ $firstspace ] = "''";
+                                       $arr [ $firstspace-1 ] .= "'";
+                               }
                        }
+
+                       # Now let's actually convert our apostrophic mush to HTML!
+                       $output = '';
+                       $buffer = '';
+                       $state = '';
+                       $i = 0;
+                       foreach ($arr as $r)
+                       {
+                               if (($i % 2) == 0)
+                               {
+                                       if ($state == 'both')
+                                               $buffer .= $r;
+                                       else
+                                               $output .= $r;
+                               }
+                               else
+                               {
+                                       if (strlen ($r) == 2)
+                                       {
+                                               if ($state == 'em')
+                                               { $output .= "</em>"; $state = ''; }
+                                               else if ($state == 'strongem')
+                                               { $output .= "</em>"; $state = 'strong'; }
+                                               else if ($state == 'emstrong')
+                                               { $output .= "</strong></em><strong>"; $state = 'strong'; }
+                                               else if ($state == 'both')
+                                               { $output .= "<strong><em>{$buffer}</em>"; $state = 'strong'; }
+                                               else # $state can be 'strong' or ''
+                                               { $output .= "<em>"; $state .= 'em'; }
+                                       }
+                                       else if (strlen ($r) == 3)
+                                       {
+                                               if ($state == 'strong')
+                                               { $output .= "</strong>"; $state = ''; }
+                                               else if ($state == 'strongem')
+                                               { $output .= "</em></strong><em>"; $state = 'em'; }
+                                               else if ($state == 'emstrong')
+                                               { $output .= "</strong>"; $state = 'em'; }
+                                               else if ($state == 'both')
+                                               { $output .= "<em><strong>{$buffer}</strong>"; $state = 'em'; }
+                                               else # $state can be 'em' or ''
+                                               { $output .= "<strong>"; $state .= 'strong'; }
+                                       }
+                                       else if (strlen ($r) == 5)
+                                       {
+                                               if ($state == 'strong')
+                                               { $output .= "</strong><em>"; $state = 'em'; }
+                                               else if ($state == 'em')
+                                               { $output .= "</em><strong>"; $state = 'strong'; }
+                                               else if ($state == 'strongem')
+                                               { $output .= "</em></strong>"; $state = ''; }
+                                               else if ($state == 'emstrong')
+                                               { $output .= "</strong></em>"; $state = ''; }
+                                               else if ($state == 'both')
+                                               { $output .= "<em><strong>{$buffer}</strong></em>"; $state = ''; }
+                                               else # ($state == '')
+                                               { $buffer = ''; $state = 'both'; }
+                                       }
+                               }
+                               $i++;
+                       }
+                       # Now close all remaining tags.  Notice that the order is important.
+                       if ($state == 'strong' || $state == 'emstrong')
+                               $output .= "</strong>";
+                       if ($state == 'em' || $state == 'strongem' || $state == 'emstrong')
+                               $output .= "</em>";
+                       if ($state == 'strongem')
+                               $output .= "</strong>";
+                       if ($state == 'both')
+                               $output .= "<strong><em>{$buffer}</em></strong>";
+                       return $output;
                }
        }
 
@@ -948,101 +1080,131 @@ class Parser
        /* private */ function replaceExternalLinks( $text ) {
                $fname = 'Parser::replaceExternalLinks';
                wfProfileIn( $fname );
-               $text = $this->subReplaceExternalLinks( $text, 'http', true );
-               $text = $this->subReplaceExternalLinks( $text, 'https', true );
-               $text = $this->subReplaceExternalLinks( $text, 'ftp', false );
-               $text = $this->subReplaceExternalLinks( $text, 'irc', false );
-               $text = $this->subReplaceExternalLinks( $text, 'gopher', false );
-               $text = $this->subReplaceExternalLinks( $text, 'news', false );
-               $text = $this->subReplaceExternalLinks( $text, 'mailto', false );
-               wfProfileOut( $fname );
-               return $text;
-       }
 
-       /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber ) {
-               $unique = '4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3';
-               $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
-
-               # this is  the list of separators that should be ignored if they
-               # are the last character of an URL but that should be included
-               # if they occur within the URL, e.g. "go to www.foo.com, where .."
-               # in this case, the last comma should not become part of the URL,
-               # but in "www.foo.com/123,2342,32.htm" it should.
-               $sep = ",;\.:";
-               $fnc = 'A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF';
-               $images = 'gif|png|jpg|jpeg';
-
-               # PLEASE NOTE: The curly braces { } are not part of the regex,
-               # they are interpreted as part of the string (used to tell PHP
-               # that the content of the string should be inserted there).
-               $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
-                 "((?i){$images})([^{$uc}]|$)/";
-
-               $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
                $sk =& $this->mOptions->getSkin();
+               $linktrail = wfMsg('linktrail');
+               $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 
-               if ( $autonumber and $this->mOptions->getAllowExternalImages() ) { # Use img tags only for HTTP urls
-                       $s = preg_replace( $e1, '\\1' . $sk->makeImage( "{$unique}:\\3" .
-                         '/\\4.\\5', '\\4.\\5' ) . '\\6', $s );
+               $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
+
+               $i = 0;
+               while ( $i<count( $bits ) ) {
+                       $url = $bits[$i++];
+                       $protocol = $bits[$i++];
+                       $text = $bits[$i++];
+                       $trail = $bits[$i++];
+                       
+                       # If the link text is an image URL, replace it with an <img> tag
+                       # This happened by accident in the original parser, but some people used it extensively
+                       $img = $this->maybeMakeImageLink( $text );
+                       if ( $img !== false ) {
+                               $text = $img;
+                       }
+
+                       $dtrail = '';
+
+                       # No link text, e.g. [http://domain.tld/some.link]
+                       if ( $text == '' ) {
+                               # Autonumber if allowed
+                               if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) { 
+                                       $text = "[" . ++$this->mAutonumber . "]";
+                               } else { 
+                                       # Otherwise just use the URL
+                                       $text = wfEscapeHTML( $url ); 
+                               }
+                       } else {
+                               # Have link text, e.g. [http://domain.tld/some.link text]s
+                               # Check for trail
+                               if ( preg_match( $linktrail, $trail, $m2 ) ) {
+                                       $dtrail = $m2[1];
+                                       $trail = $m2[2];
+                               }
+                       }
+                       
+                       $encUrl = htmlspecialchars( $url );
+                       # Bit in parentheses showing the URL for the printable version
+                       if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $url ) ) {
+                               $paren = '';
+                       } else {
+                               # Expand the URL for printable version
+                               $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
+                       }
+
+                       # Process the trail (i.e. everything after this link up until start of the next link),
+                       # replacing any non-bracketed links
+                       $trail = $this->replaceFreeExternalLinks( $trail );
+
+                       $la = $sk->getExternalLinkAttributes( $url, $text );
+                       
+                       # Use the encoded URL
+                       # This means that users can paste URLs directly into the text
+                       # Funny characters like &ouml; aren't valid in URLs anyway
+                       # This was changed in August 2004
+                       $s .= "<a href=\"{$url}\" {$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
                }
-               $s = preg_replace( $e2, '\\1' . "<a href=\"{$unique}:\\3\"" .
-                 $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
-                 "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
-                 '</a>\\5', $s );
-               $s = str_replace( $unique, $protocol, $s );
 
-               $a = explode( "[{$protocol}:", " " . $s );
-               $s = array_shift( $a );
-               $s = substr( $s, 1 );
+               wfProfileOut( $fname );
+               return $s;
+       }
 
-               # Regexp for URL in square brackets
-               $e1 = "/^([{$uc}{$sep}]+)\\](.*)\$/sD";
-               # Regexp for URL with link text in square brackets
-               $e2 = "/^([{$uc}{$sep}]+)\\s+([^\\]]+)\\](.*)\$/sD";
+       # Replace anything that looks like a URL with a link
+       function replaceFreeExternalLinks( $text ) {
+               $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
+               $s = array_shift( $bits );
+               $i = 0;
 
-               foreach ( $a as $line ) {
+               $sk =& $this->mOptions->getSkin();
 
-                       # CASE 1: Link in square brackets, e.g.
-                       # some text [http://domain.tld/some.link] more text
-                       if ( preg_match( $e1, $line, $m ) ) {
-                               $link = "{$protocol}:{$m[1]}";
+               while ( $i < count( $bits ) ){
+                       $protocol = $bits[$i++];
+                       $remainder = $bits[$i++];
+
+                       if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
+                               # Found some characters after the protocol that look promising
+                               $url = $protocol . $m[1];
                                $trail = $m[2];
-                               if ( $autonumber ) { $text = "[" . ++$this->mAutonumber . "]"; }
-                               else { $text = wfEscapeHTML( $link ); }
-                       }
+                               
+                               # Move trailing punctuation to $trail
+                               $sep = ',;\.:!?';
+                               # If there is no left bracket, then consider right brackets fair game too
+                               if ( strpos( $url, '(' ) === false ) {
+                                       $sep .= ')';
+                               }
 
-                       # CASE 2: Link with link text and text directly following it, e.g.
-                       # This is a collection of [http://domain.tld/some.link link]s
-                       else if ( preg_match( $e2, $line, $m ) ) {
-                               $link = "{$protocol}:{$m[1]}";
-                               $text = $m[2];
-                               $dtrail = '';
-                               $trail = $m[3];
-                               if ( preg_match( wfMsg ('linktrail'), $trail, $m2 ) ) {
-                                       $dtrail = $m2[1];
-                                       $trail = $m2[2];
+                               $numSepChars = strspn( strrev( $url ), $sep );
+                               if ( $numSepChars ) {
+                                       $trail = substr( $url, -$numSepChars ) . $trail;
+                                       $url = substr( $url, 0, -$numSepChars );
                                }
-                       }
 
-                       # CASE 3: Nothing matches, just output the source text
-                       else {
-                               $s .= "[{$protocol}:" . $line;
-                               continue;
-                       }
+                               # Replace &amp; from obsolete syntax with &
+                               $url = str_replace( '&amp;', '&', $url );
 
-                       if( $link == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
-                               $paren = '';
+                               # Is this an external image?
+                               $text = $this->maybeMakeImageLink( $url );
+                               if ( $text === false ) {
+                                       # Not an image, make a link
+                                       $text = $sk->makeExternalLink( $url, $url );
+                               }
+                               $s .= $text . $trail;
                        } else {
-                               # Expand the URL for printable version
-                               $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
+                               $s .= $protocol . $remainder;
                        }
-                       $la = $sk->getExternalLinkAttributes( $link, $text );
-                       $s .= "<a href='{$link}'{$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
-
                }
                return $s;
        }
-
+                               
+       function maybeMakeImageLink( $url ) {
+               $sk =& $this->mOptions->getSkin();
+               $text = false;
+               if ( $this->mOptions->getAllowExternalImages() ) {
+                       if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
+                               # Image found
+                               $text = $sk->makeImage( htmlspecialchars( $url ) );
+                       }
+               }
+               return $text;
+       }
 
        /* private */ function replaceInternalLinks( $s ) {
                global $wgLang, $wgLinkCache;
@@ -1057,7 +1219,6 @@ class Parser
                $sk =& $this->mOptions->getSkin();
 
                $redirect = MagicWord::get ( MAG_REDIRECT ) ;
-               $isRedirect = $redirect->matchStart ( strtoupper ( substr ( $s , 0 , 10 ) ) ) ;
 
                $a = explode( '[[', ' ' . $s );
                $s = array_shift( $a );
@@ -1072,14 +1233,6 @@ class Parser
 
                $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
                # Special and Media are pseudo-namespaces; no pages actually exist in them
-               static $image = FALSE;
-               static $special = FALSE;
-               static $media = FALSE;
-               static $category = FALSE;
-               if ( !$image ) { $image = Namespace::getImage(); }
-               if ( !$special ) { $special = Namespace::getSpecial(); }
-               if ( !$media ) { $media = Namespace::getMedia(); }
-               if ( !$category ) { $category = Namespace::getCategory(); }
 
                $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
 
@@ -1169,17 +1322,20 @@ class Parser
                                        $s .= (trim($tmp) == '')? '': $tmp;
                                        continue;
                                }
-                               if ( $ns == $image ) {
+                               if ( $ns == NS_IMAGE ) {
                                        $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
                                        $wgLinkCache->addImageLinkObj( $nt );
                                        continue;
                                }
-                               if ( $ns == $category && !$isRedirect ) {
+                               if ( $ns == NS_CATEGORY ) {
                                        $t = $nt->getText() ;
-                                       $nnt = Title::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
+                                       $nnt = Title::newFromText ( Namespace::getCanonicalName(NS_CATEGORY).":".$t ) ;
 
                                        $wgLinkCache->suspend(); # Don't save in links/brokenlinks
+                                       $pPLC=$sk->postParseLinkColour();
+                                       $sk->postParseLinkColour( false );
                                        $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
+                                       $sk->postParseLinkColour( $pPLC );
                                        $wgLinkCache->resume();
 
                                        $sortkey = $wasblank ? $this->mTitle->getPrefixedText() : $text;
@@ -1196,11 +1352,11 @@ class Parser
                                continue;
                        }
 
-                       if( $ns == $media ) {
+                       if( $ns == NS_MEDIA ) {
                                $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
                                $wgLinkCache->addImageLinkObj( $nt );
                                continue;
-                       } elseif( $ns == $special ) {
+                       } elseif( $ns == NS_SPECIAL ) {
                                $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
                                continue;
                        }
@@ -1692,7 +1848,7 @@ class Parser
                        if ( isset( $this->mTemplatePath[$part1] ) ) {
                                $noparse = true;
                                $found = true;
-                       }                  
+                       }
                        # set $text to cached message.
                        $text = $this->mTemplates[$part1];
                        $found = true;
@@ -1758,10 +1914,7 @@ class Parser
                         # Add a new element to the templace recursion path
                        $this->mTemplatePath[$part1] = 1;
 
-                       # Run full parser on the included text
-                       $text = $this->internalParse( $text, $newline, $assocArgs );
-                       # I replaced the line below with the line above, as it former seems to cause several bugs
-                       #$text = $this->stripParse( $text, $newline, $assocArgs );
+                       $text = $this->stripParse( $text, $newline, $assocArgs );
 
                        # Resume the link cache and register the inclusion as a link
                        if ( !is_null( $title ) ) {
@@ -2312,10 +2465,8 @@ class Parser
                if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
 
                $text = preg_replace( '/~~~~~/', $d, $text );
-               $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText(
-                 Namespace::getUser() ) . ":$n|$k]] $d", $text );
-               $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText(
-                 Namespace::getUser() ) . ":$n|$k]]", $text );
+               $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
+               $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
 
                # Context links: [[|name]] and [[name (context)|]]
                #