Fixed incorrect logic in XmlReader's example (#43799)

[scale-tone]

(#43799 is now 3 weeks old without even being triaged, so fixing it myself)

The code sample for XmlReader is incorrect. It relies on the fact that the input XML has some formatting in it. Without formatting the code returns wrong results.

In more details, the issue happens because XElement.ReadFrom() reads the current element and positions the reader at the next element. When the XML has formatting, that next element will be of type Whitespace or smth - and everything works. But without formatting that next element will be the next Child element. Then the reader.Read() happens, which moves the reader to yet another next element. Hence the <Child Key=""02""> tag simply gets skipped.

Here is the code to address the issue.

---

Internal previews

📄 File	🔗 Preview link
docs/standard/linq/stream-xml-fragments-xmlreader.md	How to stream XML fragments from an XmlReader (LINQ to XML)

[scale-tone]

@dotnet-policy-service agree

[adegeo]

Converting to review comment

[adegeo]

Hi @scale-tone sorry for the delay. The post US Thanksgiving and Christmas holiday season has a lot of people in and out of office.

Do you know if the VB code also suffers from this bug? It would have to be fixed too. If you're not interested in doing that, I'll schedule some time to take a look and modify your PR with the fixes. Cheers!

[scale-tone]

Do you know if the VB code also suffers from this bug?

It does, and that's why this PR already contains a fix for VB part as well :)

[adegeo]

OH GEEZ did I not spot that! HAHA

[adegeo]

I totally scrolled down to it too and my brain just saw it as all the same snippet of code. 🤣😂

[adegeo]

What do you think about modernizing the C# code for the reader too? The switch seems like the wrong check type here. We can eliminate a lot of indentation by changing using, swapping to if and using pattern matching on the type check. Thoughts?

static IEnumerable<XElement> StreamRootChildDoc(StringReader stringReader)
{
    using XmlReader reader = XmlReader.Create(stringReader);

    reader.MoveToContent();

    // Parse the file and display each of the nodes.
    while (true)
    {
        // If the current node is an element and named "Child"
        if (reader.NodeType == XmlNodeType.Element && reader.Name == "Child")
        {
            // Get the current node and advance the reader to the next
            if (XNode.ReadFrom(reader) is XElement el)
                yield return el;
            else
                continue;
        }
        else if (!reader.Read())
            break;
    }
}

Instead of

static IEnumerable<XElement> StreamRootChildDoc(StringReader stringReader)
{
    using (XmlReader reader = XmlReader.Create(stringReader))
    {
        reader.MoveToContent();
        // Parse the file and display each of the nodes.
        while (true)
        {
            switch (reader.NodeType)
            {
                case XmlNodeType.Element:
                    if (reader.Name == "Child") {
                        XElement? el = XNode.ReadFrom(reader) as XElement;
                        if (el != null)
                            yield return el;

                        // Should only call reader.Read(), if XNode.ReadFrom() was NOT called,
                        // because XNode.ReadFrom() advances the reader to the next element by itself.
                        continue;
                    }
                    break;
            }

            if (!reader.Read())
                break;
        }
    }
}

[adegeo]

OH man, that VB code is soooooooooo old too. I'm going to update that.

[adegeo]

OK that trimmed about 10 lines from the C# code and almost 70 lines from the VB code!

[scale-tone]

Minus two more lines :) :

static IEnumerable<XElement> StreamRootChildDoc(StringReader stringReader)
{
    using XmlReader reader = XmlReader.Create(stringReader);

    reader.MoveToContent();

    // Parse the file and display each of the nodes.
    while (true)
    {
        // If the current node is an element and named "Child"
        if (reader.NodeType == XmlNodeType.Element && reader.Name == "Child")
        {
            // Get the current node and advance the reader to the next
            if (XNode.ReadFrom(reader) is XElement el)
                yield return el;
        }
        else if (!reader.Read())
            break;
    }
}

[adegeo]

Thanks a bunch for the fix! Considering this area is low priority (usually based on page views), I'm unsure if we would have gotten to this. your contribution really matters and forces us to take a look. Thanks again!

[scale-tone]

My pleasure, and happy holidays! :)

[adegeo]

Proper link:

Fixes #43799