XmlSerialization - reading the End Element Tag?

Question

I recently read the IXmlSerializable interface information on MSDN and was a little confused about the whole end tag thing.

As I understand it, if you are implementing IXmlSerializable then you do not need to write the end tag, but you do need to read it? E.g. the following would error if my class A does not read the end tag

<A>
   <i>32</i>
</A>

But what happens if the content is

<A i="32"/>

If I attempt to read the end tag here I get an InvalidOperationException. But if I read on to determine what I'm supposed to be reading is that not going to much up the position of the reader?

I think I'm just getting a little lost, as the code I'm working with gets a bit more complex when I have to deal with IXmlSerializable children (which can be of varying types), and collections of IXmlSerializable elements.

Answer 1

I'm recalling XmlReader.IsEmptyElement. On the start tag if this is set then you have <a />. If it is not set then you have <a></a>. The latter you read the end tag, the former you don't.

Answer 2

I think this bit about not reading the end tag is for a scenario where your object is embedded in a larger structure:

public class Wrapper {
    public A Wrapped;
}

<Wrapper>
    <Wrapped><i>32</i></Wrapped>
</Wrapper>

I expect you'll need to read <Wrapped> and </Wrapped> but not </Wrapper>.

Answer 3

Both XmlReader and XPathNavigator implement a property called IsEmptyElement.

From the documentation:

When overridden in a derived class, gets a value indicating whether the current node is an empty element without an end element tag.

So when you see a start-element event, if IsEmptyElement is true you should not look for an end-element event. The start-element event acts as both start and end for that case.