Improve tolerance of illegal XML characters (#1290)

module/VuFind/src/VuFind/RecordDriver/MarcAdvancedTrait.php

@@ -853,11 +853,15 @@ trait MarcAdvancedTrait
     {
         // Special case for MARC:
         if ($format == 'marc21') {
-            $xml = $this->getMarcRecord()->toXML();
-            $xml = str_replace(
-                [chr(27), chr(28), chr(29), chr(30), chr(31), chr(8)], ' ', $xml
+            $sanitizeXmlRegEx
+                = '[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+';
+            $xml = simplexml_load_string(
+                trim(
+                    preg_replace(
+                        "/$sanitizeXmlRegEx/u", ' ', $this->getMarcRecord()->toXML()
+                    )
+                )
             );
-            $xml = simplexml_load_string($xml);
             if (!$xml || !isset($xml->record)) {
                 return false;
             }

module/VuFind/src/VuFind/XSLT/Processor.php

@@ -58,7 +58,10 @@ class Processor
         $xsl = new XSLTProcessor();
         $xsl->importStyleSheet($style);
         $doc = new DOMDocument();
-        if ($doc->loadXML($xml)) {
+        $sanitizeXmlRegEx
+            = '[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+';
+        $cleanXml = trim(preg_replace("/$sanitizeXmlRegEx/u", ' ', $xml));
+        if ($doc->loadXML($cleanXml)) {
             foreach ($params as $key => $value) {
                 $xsl->setParameter('', $key, $value);
             }