Use stricter regex (JabRef#11720)

Author: subhramit

src/main/java/org/jabref/logic/openoffice/ReferenceMark.java

@@ -15,7 +15,7 @@ public class ReferenceMark {
 
     private static final Logger LOGGER = LoggerFactory.getLogger(ReferenceMark.class);
 
-    private static final Pattern REFERENCE_MARK_FORMAT = Pattern.compile("^((?:JABREF_\\w+ CID_\\w+(?:,\\s*)?)+)(\\s*\\w+)?$");
+    private static final Pattern REFERENCE_MARK_FORMAT = Pattern.compile("^(JABREF_\\w+ CID_\\d+(?:, JABREF_\\w+ CID_\\d+)*) (\\w+)$");
     private static final Pattern ENTRY_PATTERN = Pattern.compile("JABREF_(\\w+) CID_(\\w+)");
 
     private final String name;
@@ -24,7 +24,11 @@ public class ReferenceMark {
     private String uniqueId;
 
     /**
-     * @param name Format: <code>JABREF_{citationKey} CID_{citationNumber} {uniqueId}</code>
+     * @param name Allowed formats:
+     * Single entry: <code>JABREF_{citationKey} CID_{citationNumber} {uniqueId}</code>
+     * Group of entries: <code>JABREF_{citationKey1} CID_{citationNumber1}, JABREF_{citationKey2} CID_{citationNumber2}, ..., JABREF_{citationKeyN} CID_{citationNumberN} {uniqueId}</code>
+     * Disallowed: <code>JABREF_{citationKey} CID_{citationNumber}</code> (no unique ID at the end)
+     * Disallowed: <code>JABREF_{citationKey1} CID_{citationNumber1} JABREF_{citationKey2} CID_{citationNumber2} {uniqueId}</code> (no comma between entries)
      */
     public ReferenceMark(String name) {
         this.name = name;