Ndtyjky

As you can see in the linked question, when using XDocument in C#, an empty internal subset will be added in the DTD-Header, if non had existed before. The question and answer deal with how to remove this, however, while the answerer points out that that this is well-formatted XML, I ask:

1) Under which circumstances will this cause problems? Which standard changed so that a legacy application might be incompatible with an empty internal subset?

2) Is it preferable to use empty internal subsets going forward with new applications?

When you parse Xml Document by using XDocument with DTD then empty Internal Subset means Square Brackets is automatically inserted.

If you want to remove this Internal Subset then you can set XDocumentType.InternalSubset = null like

XDocument doc = XDocument.Load(@"Path to xml file");

if (doc.DocumentType != null)

    doc.DocumentType.InternalSubset = null;



//Do code with XDocument

Example:

<?xml version="1.0" encoding="utf-8" ?>

<!DOCTYPE book [ <!ENTITY h "hardcover"> ]>

<book genre="novel" ISBN="1-861001-57-5">  

  <title>Pride And Prejudice</title>

  <author>Mark Henry</author>  

</book>

In above example kindly noticed this part [ <!ENTITY h "hardcover"> ]. This is called as Internal Subset.

Does it matter?

No this does not matter. but its a well formed XML if your XML doesn't contain any internal subset then it represent as blank square brackets . it means that your xml doesn't contain any internal subset.

While parsing xml with XDocument with no internal subset then XDocument append blank square brackets instead of display nothing in DOCTYPE.

What does an empty internal subset do?

The basic purpose of an internal entity is to get rid of typing same content (like the name of the organization) again and again. And instead, we can define an internal entity to contain the text and then only you need to use the entity where you want to insert the text. Because the entity is expanded by the parser, you can be assured that you'll get the same text in every location. The parser will also catch if you misspell an entity name.

You can read more about Internal Subset here

It's a little bit strange, but not wrong.

The grammar for DOCTYPE is

doctypedecl ::= '<!DOCTYPE' S Name (S ExternalID)? S? ('[' intSubset ']' S?)? '>'   

intSubset   ::= (markupdecl | DeclSep)*

So you're allowed a sequence of zero-or-more markup declarations between square brackets, and if there aren't any markup declarations then you can omit the square brackets (but you aren't required to).