XML signature detection does not work

[pwinckles]

JHOVE will not identify an XML document based on having its XML declaration.

For example, suppose you have the following XML:

<?xml version="1.0" encoding="UTF-8"?>
<outer>
  <inner>
    blah
  </outer>
</inner>

And execute JHOVE on it:

❯ ./jhove -s /var/tmp/invalid.xml 
Jhove (Rel. 1.24.1, 2020-03-16)
 Date: 2021-06-30 15:02:25 CDT
 RepresentationInformation: /var/tmp/invalid.xml
  ReportingModule: ASCII-hul, Rel. 1.4.1 (2019-04-17)
  LastModified: 2021-06-29 13:21:30 CDT
  Size: 86
  Format: ASCII
  Status: Well-Formed
  SignatureMatches:
   ASCII-hul
  MIMEtype: text/plain; charset=US-ASCII

(Yes, the XML is not well-formed. That's the point.)

I dug into why this is happening, and the root cause is because a DataInputStream is being used to read the XML file here. DataInputStream cannot be used like this, and a InputStreamReader must be used instead.

Here's a demonstration of the problem:

String xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n" +
        "<outer>\n" +
        "  <inner>\n" +
        "    blah\n" +
        "  </outer>\n" +
        "</inner>\n" +
        "\n";

DataInputStream dataStream = new DataInputStream(new ByteArrayInputStream(xml.getBytes(StandardCharsets.UTF_8)));
System.out.print("DataInputStream: ");
while (true) {
    try {
        System.out.print(dataStream.readChar());
    } catch (IOException e) {
        break;
    }
}
System.out.print("\n\n");

InputStreamReader reader = new InputStreamReader(new ByteArrayInputStream(xml.getBytes(StandardCharsets.UTF_8)));
char c;
System.out.print("InputStreamReader: ");
while ((c = (char) reader.read()) != -1) {
    System.out.print(c);
}

That code produces the following output when run:

DataInputStream: 㰿硭氠癥牳楯渽∱⸰∠敮捯摩湧㴢啔䘭㠢㼾਼潵瑥爾ਠ‼楮湥爾ਠ†⁢污栊†㰯潵瑥爾਼⽩湮敲㸊

InputStreamReader: <?xml version="1.0" encoding="UTF-8"?>
<outer>
  <inner>
    blah
  </outer>
</inner>