XML should not be validated when no schema provided

[pwinckles]

The documentation states:

Note that the concept of validity applies only to XML files that explicitly reference a DTD or XML Schema. JHOVE can determine if either of these conditions are met and if so, it will invoke automatically the SAX2 parser in a validating mode. Otherwise, the parser is invoked in a manner that only checks for well-formedness.

However, this is not what the code actually does. Instead, it will validate any XML document that contains a namespace regardless of whether or not there is a schema or dtd.

Here is a example that demonstrates the problem:

<example xmlns="http://example.com">
  <text>This is an example</text>
</example>

❯ ./jhove -m XML-hul example.xml 
Jhove (Rel. 1.24.1, 2020-03-16)
 Date: 2021-06-24 07:41:51 CDT
 RepresentationInformation: example.xml
  ReportingModule: XML-hul, Rel. 1.5.1 (2019-04-17)
  LastModified: 2021-06-24 07:41:36 CDT
  Size: 82
  Format: XML
  Version: 1.0
  Status: Well-Formed, but not valid
  SignatureMatches:
   XML-hul
  ErrorMessage: SaxParseException: cvc-elt.1: Cannot find the declaration of element 'example'.: Line = 1, Column = 37
   ID: XML-HUL-1
  MIMEtype: text/xml
  XMLMetadata: 
   Parser: org.apache.xerces.jaxp.SAXParserImpl$JAXPSAXParser
   Encoding: UTF-8
   Schemas: 
    Schema: 
     NamespaceURI: http://example.com
     SchemaLocation: 
   Root: example
   Namespaces: 
    Namespace: 
     Prefix: 
     URI: http://example.com

The expected behavior is that JHOVE only checks for well-formedness and does not validate the xml.

[ross-spencer]

Hi Peter - are you anticipating that the result should just read "well formed"? Or is there something else, e.g. "well-formed, validity not checked", i.e. not to imply validity was checked?

[pwinckles]

As far as I can tell, these are the possible outcomes:

1. Well-Formed
2. Well-Formed and valid
3. Well-Formed, but not valid
4. Not well-formed

I believe that covers all possible outcomes, and I do not think a new status needs to be introduced. Well-Formed implies Well-Formed, validity not checked.