PDFBOX-6106: catch incorrect type when in strict mode

git-svn-id: https://svn.apache.org/repos/asf/pdfbox/trunk@1930057 13f79535-47bb-0310-9956-ffa450edef68

Author: THausherr

xmpbox/src/main/java/org/apache/xmpbox/xml/DomXmpParser.java

@@ -280,12 +280,20 @@ private void parseDescriptionRootAttr(XMPMetadata xmp, Element description, Attr
             PropertyType type = checkPropertyDefinition(xmp,
                     new QName(attr.getNamespaceURI(), attr.getLocalName()));
 
-            //Default to text if no type is found
-            if( type == null)
+            // PDFBOX-2318, PDFBOX-6106: Default to text if no type is found
+            if (type == null)
             {
-                type = TypeMapping.createPropertyType(Types.Text, Cardinality.Simple);
+                if (strictParsing)
+                {
+                    throw new XmpParsingException(ErrorType.InvalidType, "No type defined for {" + attr.getNamespaceURI() + "}"
+                            + attr.getLocalName());
+                }
+                else
+                {
+                    type = TypeMapping.createPropertyType(Types.Text, Cardinality.Simple);
+                }
             }
-            
+
             try
             {
                 AbstractSimpleProperty sp = tm.instanciateSimpleProperty(namespace, schema.getPrefix(),

xmpbox/src/test/java/org/apache/xmpbox/xml/DomXmpParserTest.java

@@ -83,4 +83,37 @@ void testPDFBox5976() throws XmpParsingException
         Assertions.assertEquals("B", xmp.getPDFAIdentificationSchema().getConformance());
         Assertions.assertEquals((Integer) 3, xmp.getPDFAIdentificationSchema().getPart());
     }
+
+    /**
+     * PDFBOX-6106: Check that "pdf:CreationDate='2004-01-30T17:21:50Z'" is detected as incorrect.
+     * (Only Keywords, PDFVersion, and Producer are allowed in strict mode)
+     *
+     * @throws XmpParsingException
+     */
+    @Test
+    void testPDFBox6106() throws XmpParsingException
+    {
+        // from file 001358.pdf
+        String s = "<?xpacket begin='' id='W5M0MpCehiHzreSzNTczkc9d' bytes='647'?>\n" +
+                    "<rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'\n" +
+                    "         xmlns:iX='http://ns.adobe.com/iX/1.0/'>\n" +
+                    "	<rdf:Description about=''\n" +
+                    "	                 xmlns='http://ns.adobe.com/pdf/1.3/'\n" +
+                    "	                 xmlns:pdf='http://ns.adobe.com/pdf/1.3/'\n" +
+                    "	                 pdf:CreationDate='2004-01-30T17:21:50Z'\n" +
+                    "	                 pdf:ModDate='2004-01-30T17:21:50Z'\n" +
+                    "	                 pdf:Producer='Acrobat Distiller 5.0.5 (Windows)'/>\n" +
+                    "	<rdf:Description about=''\n" +
+                    "	                 xmlns='http://ns.adobe.com/xap/1.0/'\n" +
+                    "	                 xmlns:xap='http://ns.adobe.com/xap/1.0/'\n" +
+                    "	                 xap:CreateDate='2004-01-30T17:21:50Z'\n" +
+                    "	                 xap:ModifyDate='2004-01-30T17:21:50Z'\n" +
+                    "	                 xap:MetadataDate='2004-01-30T17:21:50Z'/>\n" +
+                    "</rdf:RDF><?xpacket end='r'?>";
+        DomXmpParser xmpParser = new DomXmpParser();
+        XmpParsingException ex = Assertions.assertThrows(
+                XmpParsingException.class,
+                () -> xmpParser.parse(s.getBytes(StandardCharsets.UTF_8)));
+        Assertions.assertEquals("No type defined for {http://ns.adobe.com/pdf/1.3/}CreationDate", ex.getMessage());
+    }
 }