merge latest master into Wikidata branch
[lhc/web/wiklou.git] / includes / parser / Parser.php
index 61b4239..1902443 100644 (file)
@@ -200,6 +200,13 @@ class Parser {
         */
        var $mUniqPrefix;
 
+       /**
+        * @var Array with the language name of each language link (i.e. the
+        * interwiki prefix) in the key, value arbitrary. Used to avoid sending
+        * duplicate language links to the ParserOutput.
+        */
+       var $mLangLinkLanguages;
+
        /**
         * Constructor
         *
@@ -282,6 +289,7 @@ class Parser {
                        $this->mRevisionId = $this->mRevisionUser = null;
                $this->mVarCache = array();
                $this->mUser = null;
+               $this->mLangLinkLanguages = array();
 
                /**
                 * Prefix for temporary replacement strings for the multipass parser.
@@ -491,6 +499,11 @@ class Parser {
                                $PFreport;
                        wfRunHooks( 'ParserLimitReport', array( $this, &$limitReport ) );
                        $text .= "\n<!-- \n$limitReport-->\n";
+
+                       if ( $this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10 ) {
+                               wfDebugLog( 'generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' .
+                                       $this->mTitle->getPrefixedDBkey() );
+                       }
                }
                $this->mOutput->setText( $text );
 
@@ -1948,8 +1961,14 @@ class Parser {
                                # Interwikis
                                wfProfileIn( __METHOD__."-interwiki" );
                                if ( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && Language::fetchLanguageName( $iw, null, 'mw' ) ) {
-                                       // FIXME: the above check prevents links to sites with identifiers that are not language codes
-                                       $this->mOutput->addLanguageLink( $nt->getFullText() );
+                                       // XXX: the above check prevents links to sites with identifiers that are not language codes
+
+                                       # Bug 24502: filter duplicates
+                                       if ( !isset( $this->mLangLinkLanguages[$iw] ) ) {
+                                               $this->mLangLinkLanguages[$iw] = true;
+                                               $this->mOutput->addLanguageLink( $nt->getFullText() );
+                                       }
+
                                        $s = rtrim( $s . $prefix );
                                        $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail;
                                        wfProfileOut( __METHOD__."-interwiki" );
@@ -3702,8 +3721,13 @@ class Parser {
                        return $obj->tc_contents;
                }
 
-               $text = Http::get( $url );
-               if ( !$text ) {
+               $req = MWHttpRequest::factory( $url );
+               $status = $req->execute(); // Status object
+               if ( $status->isOK() ) {
+                       $text = $req->getContent();
+               } elseif ( $req->getStatus() != 200 ) { // Though we failed to fetch the content, this status is useless.
+                       return wfMessage( 'scarytranscludefailed-httpstatus', $url, $req->getStatus() /* HTTP status */ )->inContentLanguage()->text();
+               } else {
                        return wfMessage( 'scarytranscludefailed', $url )->inContentLanguage()->text();
                }
 
@@ -4147,10 +4171,16 @@ class Parser {
                        $safeHeadline = $this->mStripState->unstripBoth( $safeHeadline );
 
                        # Strip out HTML (first regex removes any tag not allowed)
-                       # Allowed tags are <sup> and <sub> (bug 8393), <i> (bug 26375) and <b> (r105284)
-                       # We strip any parameter from accepted tags (second regex)
+                       # Allowed tags are:
+                       # * <sup> and <sub> (bug 8393)
+                       # * <i> (bug 26375)
+                       # * <b> (r105284)
+                       # * <span dir="rtl"> and <span dir="ltr"> (bug 35167)
+                       #
+                       # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
+                       # to allow setting directionality in toc items.
                        $tocline = preg_replace(
-                               array( '#<(?!/?(sup|sub|i|b)(?: [^>]*)?>).*?'.'>#', '#<(/?(sup|sub|i|b))(?: .*?)?'.'>#' ),
+                               array( '#<(?!/?(span|sup|sub|i|b)(?: [^>]*)?>).*?'.'>#', '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|i|b))(?: .*?)?'.'>#' ),
                                array( '',                          '<$1>' ),
                                $safeHeadline
                        );