X-Git-Url: https://git.heureux-cyclage.org/?p=lhc%2Fweb%2Fwiklou.git;a=blobdiff_plain;f=languages%2FLanguageConverter.php;h=24b1d0c07662d8ac7a94ba962b5c50fe1de6b8ab;hp=00bc02db4582e597a5f37f91ee6c9f11a422587b;hb=942728ab20e01bba200031b01eb606ee59279bd2;hpb=fbe78cfa094645b907d0fd2885c5797321f794eb

diff --git a/languages/LanguageConverter.php b/languages/LanguageConverter.php
index 00bc02db45..24b1d0c076 100644
--- a/languages/LanguageConverter.php
+++ b/languages/LanguageConverter.php
@@ -39,6 +39,7 @@ class LanguageConverter {
 	 */
 	static public $languagesWithVariants = [
 		'en',
+		'crh',
 		'gan',
 		'iu',
 		'kk',
@@ -162,6 +163,8 @@ class LanguageConverter {
 
 		$req = $this->getURLVariant();
 
+		Hooks::run( 'GetLangPreferredVariant', [ &$req ] );
+
 		if ( $wgUser->isSafeToLoad() && $wgUser->isLoggedIn() && !$req ) {
 			$req = $this->getUserVariant();
 		} elseif ( !$req ) {
@@ -353,7 +356,6 @@ class LanguageConverter {
 		if ( $this->guessVariant( $text, $toVariant ) ) {
 			return $text;
 		}
-
 		/* we convert everything except:
 		   1. HTML markups (anything between < and >)
 		   2. HTML entities
@@ -376,9 +378,12 @@ class LanguageConverter {
 		$scriptfix = '<script[^>]*+>[^<]*+(?:(?:(?!<\/script>).)[^<]*+)*+<\/script>|';
 		// disable conversion of <pre> tags
 		$prefix = '<pre[^>]*+>[^<]*+(?:(?:(?!<\/pre>).)[^<]*+)*+<\/pre>|';
+		// The "|.*+)" at the end, is in case we missed some part of html syntax,
+		// we will fail securely (hopefully) by matching the rest of the string.
+		$htmlFullTag = '<(?:[^>=]*+(?>[^>=]*+=\s*+(?:"[^"]*"|\'[^\']*\'|[^\'">\s]*+))*+[^>=]*+>|.*+)|';
 
-		$reg = '/' . $codefix . $scriptfix . $prefix .
-			'<[^>]++>|&[a-zA-Z#][a-z0-9]++;' . $marker . $htmlfix . '|\004$/s';
+		$reg = '/' . $codefix . $scriptfix . $prefix . $htmlFullTag .
+			'&[a-zA-Z#][a-z0-9]++;' . $marker . $htmlfix . '|\004$/s';
 		$startPos = 0;
 		$sourceBlob = '';
 		$literalBlob = '';
@@ -386,6 +391,7 @@ class LanguageConverter {
 		// Guard against delimiter nulls in the input
 		// (should never happen: see T159174)
 		$text = str_replace( "\000", '', $text );
+		$text = str_replace( "\004", '', $text );
 
 		$markupMatches = null;
 		$elementMatches = null;
@@ -400,6 +406,13 @@ class LanguageConverter {
 					// We hit the end.
 					$elementPos = strlen( $text );
 					$element = '';
+				} elseif ( substr( $element, -1 ) === "\004" ) {
+					// This can sometimes happen if we have
+					// unclosed html tags (For example
+					// when converting a title attribute
+					// during a recursive call that contains
+					// a &lt; e.g. <div title="&lt;">.
+					$element = substr( $element, 0, -1 );
 				}
 			} else {
 				// If we hit here, then Language Converter could be tricked
@@ -409,11 +422,11 @@ class LanguageConverter {
 				$log = LoggerFactory::getInstance( 'languageconverter' );
 				$log->error( "Hit pcre.backtrack_limit in " . __METHOD__
 					. ". Disabling language conversion for this page.",
-					array(
+					[
 						"method" => __METHOD__,
 						"variant" => $toVariant,
 						"startOfText" => substr( $text, 0, 500 )
-					)
+					]
 				);
 				return $text;
 			}
@@ -427,7 +440,14 @@ class LanguageConverter {
 			if ( $element !== ''
 				&& preg_match( '/^(<[^>\s]*+)\s([^>]*+)(.*+)$/', $element, $elementMatches )
 			) {
+				// FIXME, this decodes entities, so if you have something
+				// like <div title="foo&lt;bar"> the bar won't get
+				// translated since after entity decoding it looks like
+				// unclosed html and we call this method recursively
+				// on attributes.
 				$attrs = Sanitizer::decodeTagAttributes( $elementMatches[2] );
+				// Ensure self-closing tags stay self-closing.
+				$close = substr( $elementMatches[2], -1 ) === '/' ? ' /' : '';
 				$changed = false;
 				foreach ( [ 'title', 'alt' ] as $attrName ) {
 					if ( !isset( $attrs[$attrName] ) ) {
@@ -446,7 +466,7 @@ class LanguageConverter {
 				}
 				if ( $changed ) {
 					$element = $elementMatches[1] . Html::expandAttributes( $attrs ) .
-						$elementMatches[3];
+						$close . $elementMatches[3];
 				}
 			}
 			$literalBlob .= $element . "\000";
@@ -658,29 +678,42 @@ class LanguageConverter {
 		$out = '';
 		$length = strlen( $text );
 		$shouldConvert = !$this->guessVariant( $text, $variant );
-
-		while ( $startPos < $length ) {
-			$pos = strpos( $text, '-{', $startPos );
-
-			if ( $pos === false ) {
+		$continue = 1;
+
+		$noScript = '<script.*?>.*?<\/script>(*SKIP)(*FAIL)';
+		$noStyle = '<style.*?>.*?<\/style>(*SKIP)(*FAIL)';
+		// phpcs:ignore Generic.Files.LineLength
+		$noHtml = '<(?:[^>=]*+(?>[^>=]*+=\s*+(?:"[^"]*"|\'[^\']*\'|[^\'">\s]*+))*+[^>=]*+>|.*+)(*SKIP)(*FAIL)';
+		while ( $startPos < $length && $continue ) {
+			$continue = preg_match(
+				// Only match -{ outside of html.
+				"/$noScript|$noStyle|$noHtml|-\{/",
+				$text,
+				$m,
+				PREG_OFFSET_CAPTURE,
+				$startPos
+			);
+
+			if ( !$continue ) {
 				// No more markup, append final segment
 				$fragment = substr( $text, $startPos );
 				$out .= $shouldConvert ? $this->autoConvert( $fragment, $variant ) : $fragment;
 				return $out;
 			}
 
-			// Markup found
+			// Offset of the match of the regex pattern.
+			$pos = $m[0][1];
+
 			// Append initial segment
 			$fragment = substr( $text, $startPos, $pos - $startPos );
 			$out .= $shouldConvert ? $this->autoConvert( $fragment, $variant ) : $fragment;
-
-			// Advance position
+			// -{ marker found, not in attribute
+			// Advance position up to -{ marker.
 			$startPos = $pos;
-
 			// Do recursive conversion
+			// Note: This passes $startPos by reference, and advances it.
 			$out .= $this->recursiveConvertRule( $text, $variant, $startPos, $depth + 1 );
 		}
-
 		return $out;
 	}
 
@@ -737,7 +770,7 @@ class LanguageConverter {
 							$warningDone = true;
 						}
 						$startPos += 2;
-						continue;
+						break;
 					}
 					// Recursively parse another rule
 					$inner .= $this->recursiveConvertRule( $text, $variant, $startPos, $depth + 1 );