XmlReader does not take doctype notation into regards

Question

First time writing on these forums. Been reading them for long though.

I have a problem, trying to validate an Xml file with XmlReader in .Net.

Xml file:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE root [
  <!NOTATION png PUBLIC "-//W3C//NOTATION Portable Network Graphics//EN">
  <!ENTITY mypic SYSTEM "mypic.png" NDATA png>
]>
<root>
  <img ref="mypic" />
</root>

Xsd file:

<?xml version="1.0" encoding="utf-8"?>
<xs:schema id="myschema"
    elementFormDefault="qualified"
    xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:element name="img">
    <xs:complexType>
      <xs:attribute name="ref" type="xs:ENTITY" />
    </xs:complexType>
  </xs:element>

  <xs:element name="root">
    <xs:complexType>
      <xs:sequence>
        <xs:element ref="img" />
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

C# snippet:

XmlReaderSettings settings = new XmlReaderSettings();
settings.ValidationEventHandler += settings_ValidationEventHandler;
settings.ValidationType = ValidationType.Schema;
settings.DtdProcessing = DtdProcessing.Parse;
settings.ValidationFlags |= XmlSchemaValidationFlags.ProcessIdentityConstraints
            | XmlSchemaValidationFlags.ProcessInlineSchema
            | XmlSchemaValidationFlags.ProcessSchemaLocation
            | XmlSchemaValidationFlags.ReportValidationWarnings;

using (XmlReader reader = XmlReader.Create("myschema.xsd")) 
{
    settings.Schemas.Add(XmlSchema.Read(reader, new ValidationEventHandler(settings_ValidationEventHandler)));
}

using (XmlReader reader = XmlReader.Create("mydata.xml", settings))
{
    while (reader.Read()) ;
}

I get a validation error stating:

Reference to an unparsed entity, 'mypic'.

According to other validators it does validate, but not according to XmlReader.

I've tried everything I can think of, so now I am turning to you guys. Any help is much appreciated.

Answer 1

By using Reflector I've found that the NOTATION and ENTITY declarations are indeed parsed as they should.

The validator will find your mypic entity reference and consider it as external and unparsed since it has an non-empty NDATA declaration.

However, I've also found that the validator will always send a Reference to unparsed entity validation error whenever it encounters an unparsed entity reference.

I don't understand why Microsoft is reporting this as an error. In my opinion, the validator should ignore unparsed entity references or report them as warnings, not errors.

It should be safe to ignore these errors from the validator and still consider the XML to be valid unless, of course, any other error is reported.

So, how to determine whether the reported error is an ignorable Reference to an unparsed entity-error?

I see three options:

1. Check if the error message starts with the string Reference to an unparsed entity. This will break if your code runs on a non-english platform.

2. Use reflection to get the value of the internal GetRes property and see if it is equal to Sch_UnparsedEntityRef. This will break if Microsoft decide to change the internal API.

3. Serialize the exception and determine whether the serialized res member is equal to Sch_UnparsedEntityRef. This will break if Microsoft decide to change the serialization format.

All of these options are "hacks". The first one is most likely to break. The third one, however, should be safe. It is not likely that Microsoft changes the serialization format since that would potentially break compatibility with other code.

Here are examples of how you could determine whether to ignore the validation exception that you receive in your settings_ValidationEventHandler method.

Based on error message (not safe):

    static bool IsUnparsedEntityReferenceError_BasedOnMessage(
        XmlSchemaException error)
    {
        return error != null && error.Message.StartsWith(
            "Reference to an unparsed entity", StringComparison.Ordinal);
    }

Based on reflection (quite safe):

    static readonly PropertyInfo GetResProp = typeof(XmlSchemaException)
        .GetProperty("GetRes", BindingFlags.NonPublic | BindingFlags.Instance);

    static bool IsUnparsedEntityReferenceError_BasedOnReflection(
        XmlSchemaException error)
    {
        return error != null && GetResProp != null && 
            "Sch_UnparsedEntityRef".Equals(GetResProp.GetValue(error, null));
    }

Based on serialization format (safest):

    static bool IsUnparsedEntityReferenceError_BasedOnSerializer(
        XmlSchemaException error)
    {
        if (error == null)
        {
            return false;
        }
        else
        {
            SerializationInfo info = new SerializationInfo(
                typeof(XmlSchemaException), new FormatterConverter());

            error.GetObjectData(info, default(StreamingContext));
            return "Sch_UnparsedEntityRef" == info.GetString("res");
        }
    }