My application needs to be able to parse and validate either a DTD or schema
based document without knowing in advance which form of grammar a document is
using. (New documents presented to my system are schema based while the older
ones are DTD - conversion is not an option as these document represent
legally binding contracts and they must be processed as is).

In the .Net 1.1 version of my code I used a XmlValidatingReader instance
configured with ValidationType.Auto which handled both document types. Now
that n I am porting the code to 2.0 I am trying to use the new
XmlReader.Create method but these readers must be configured as either DTD
validating or schema validating and can not do both at once.

I found a several examples on the web of nested XmlReaders and have tried
this approach (see following code) but the data in the <!DOCTYPE(my
application needs the PUBLIC name intact to determine the grammar version)
node of a DTD based instance seems to be lost as it passed through the
readers.

XmlReaderSettings settings;

settings = new XmlReaderSettings ();
settings.XmlResolver = GetResolver ();
settings.ValidationType = ValidationType.DTD;
settings.ProhibitDtd = false;

if (eventHandler != null)
settings.ValidationEventHandler += eventHandler;

XmlReader inner = XmlReader.Create (stream, settings);

settings = new XmlReaderSettings ();
settings.Schemas = GetSchemaSet ();
settings.ValidationType = ValidationType.Schema;
settings.ProhibitDtd = false;

if (eventHandler != null)
settings.ValidationEventHandler += eventHandler;

XmlReader reader = XmlReader.Create (inner, settings);

document.Load (reader);
return (document);

Can someone please confirm if I am using the right approach and whether the
problems with DOCTYPE are bugs, and if so is there a work around. If not then
it looks like I will have to remain with the old and now 'obsolete'
XmlValidatingReader interface which handles this scenario correctly.

The changes to the XML API in 2.0 seem to forget that there is a substantial
amount of information coded using DTD based grammars that will be with us for
many years. Some of the contracts I process represent 10 or 20 year deals.
Not everyone using XML is writing webservices where the XML is simply a
transient wire format. Efficient support for legacy documents and mixed
DTD/schema environments is commercially very important.

"Confused XML hacker" <Confused XML hacker@discussions.microsoft.comwrote
in message news:F7576F6D-6775-4319-A1F8-CD963532AD7B@microsoft.com...

My application needs to be able to parse and validate either a DTD or
schema
based document without knowing in advance which form of grammar a document
is
using. (New documents presented to my system are schema based while the
older
ones are DTD - conversion is not an option as these document represent
legally binding contracts and they must be processed as is).
>
In the .Net 1.1 version of my code I used a XmlValidatingReader instance
configured with ValidationType.Auto which handled both document types. Now
that n I am porting the code to 2.0 I am trying to use the new
XmlReader.Create method but these readers must be configured as either DTD
validating or schema validating and can not do both at once.
>
I found a several examples on the web of nested XmlReaders and have tried
this approach (see following code) but the data in the <!DOCTYPE(my
application needs the PUBLIC name intact to determine the grammar version)
node of a DTD based instance seems to be lost as it passed through the
readers.
>
XmlReaderSettings settings;
>
settings = new XmlReaderSettings ();
settings.XmlResolver = GetResolver ();
settings.ValidationType = ValidationType.DTD;
settings.ProhibitDtd = false;
>
if (eventHandler != null)
settings.ValidationEventHandler += eventHandler;
>
XmlReader inner = XmlReader.Create (stream, settings);
>
settings = new XmlReaderSettings ();
settings.Schemas = GetSchemaSet ();
settings.ValidationType = ValidationType.Schema;
settings.ProhibitDtd = false;
>
if (eventHandler != null)
settings.ValidationEventHandler += eventHandler;
>
XmlReader reader = XmlReader.Create (inner, settings);
>
document.Load (reader);
return (document);
>
Can someone please confirm if I am using the right approach and whether
the
problems with DOCTYPE are bugs, and if so is there a work around. If not
then
it looks like I will have to remain with the old and now 'obsolete'
XmlValidatingReader interface which handles this scenario correctly.
>
The changes to the XML API in 2.0 seem to forget that there is a
substantial
amount of information coded using DTD based grammars that will be with us
for
many years. Some of the contracts I process represent 10 or 20 year deals.
Not everyone using XML is writing webservices where the XML is simply a
transient wire format. Efficient support for legacy documents and mixed
DTD/schema environments is commercially very important.

Could you first attempt it using a schema, and if that fails, using DTD? Or
should I read your post more carefully?

John