LogosHTML: Fix handling of Strongs without text

When a Strongs number is not associated to any text (which is a
validation error but happens for some input formats), and there are more
than 10 nested hyperlinks, LogosHTML created an output file that could
not be parsed by the nested hyperlinks postprocessor after convertion to
DOCX.

Fix it by also allow empty elements in GrammarHackVisitor.

biblemulticonverter-logos/src/main/java/biblemulticonverter/logos/format/LogosHTML.java

@@ -902,7 +902,7 @@ public GrammarHackVisitor(Visitor<IOException> nextVisitor, Writer writer, Strin
 		public int visitElementTypes(String elementTypes) throws IOException {
 			if (elementTypes == null)
 				return 1;
-			if (elementTypes.equals("t")) {
+			if (elementTypes.equals("t") || elementTypes.isEmpty()) {
 				// only text, let's use our alternative Strings
 				writer.write(altPrefix);
 				suffix = altSuffix;