Plain XML content in XNA Content Pipeline

During the first stages of Little Racers development, I found the need of parsing XML data during the game load. The XNA Content Pipeline system includes XML Parsing, but it uses it in a very particular way. I won’t go much further in the way XNA parses XML because, to be sincere, I’ve not tried it enough to test wether it’s good or not.
My first impression was that it was too strict (probably due to optimization and type safe issues), moreover during parse I would need access to some data structures that only were available at runtime, so I preferred to store plain XML and parse it during game load.
This can be done by extending the content pipeline and adding a new type. To do this we’ll need to add two new projects to our solution.

XmlSource project

The first project we’ll create is the one which has the class that stores our xml plain text. I’ve called it XmlSource.It will have two classes in total.

XmlSource

This class is a very simple one. It simply has a string with the xml plain text:

public class XmlSource
{
public XmlSource(string xmlCode)
{
this.xmlCode = xmlCode;
}

private string xmlCode;
public string XmlCode { get { return xmlCode; } }
}

XmlSourceReader

This class extends ContentTypeReader and simply creates an XmlSource instance from its binary representation. In our case, the binary representation wil be a simple string, so the read is straightforward.

public class XmlSourceReader : ContentTypeReader<XmlSource>
{
/// <summary>
/// Loads an imported shader.
/// </summary>
protected override XmlSource Read(ContentReader input, XmlSource existingInstance)
{
string xmlData = input.ReadString();

return new XmlSource(xmlData);
}
}

With this, our first project is finished. This project is platform-independant (The Xbox needs to read the binary data) and will have to be referenced from every project that uses the XmlSource class, including the second project we’re making:

XmlSourceImporter

The second project(I called it XmlDocumentImporter) is a project of type “Content Pipeline Extension Library”. This project is referenced from the Content projects, and converts the source text file to binary data that will be saved in the .xnb files.

This project has 2 files:

XmlSourceImporter.cs

This file converts the XML source file into an XmlSource instance, that will be further converted into binary data

[ContentImporter(".xml", DisplayName = "Xml Source Importer")]
class XmlSourceImporter : ContentImporter<XmlSource>
{
public override XmlSource Import(string filename, ContentImporterContext context)
{
string sourceCode = System.IO.File.ReadAllText(filename);
return new XmlSource(sourceCode);
}
}

XmlSourceWriter.cs

This file converts an XmlSource into binary data, as opposed to XmlSourceReader

[ContentTypeWriter]
class XmlSourceWriter : ContentTypeWriter<XmlSource>
{
protected override void Write(ContentWriter output, XmlSource value)
{
/*StringWriter sw=new StringWriter();
value.Save(sw);
string content = sw.ToString();*/
output.Write(value.XmlCode);
}
public override string GetRuntimeType(TargetPlatform targetPlatform)
{
return typeof(XmlDocument).AssemblyQualifiedName;
}
public override string GetRuntimeReader(TargetPlatform targetPlatform)
{
return typeof(XmlDocumentReader).AssemblyQualifiedName;
}
}

Sample of use

Once both projects set up (remember to add reference to XmlSourceImporter from the Content project), we can add .xml files to the content project. Under “Content importer”, we’ll need to select “Xml Source importer”, instead of the default XNA one.

Once done this, the program sould compile ok. Loading XML from the program now is very easy:

XmlSource xs = Content.Load<XmlSource>(AssetName);
XmlDocument xd = new XmlDocument();
xd.LoadXml(xs.XmlCode);

I know this can be a little confusing, i’ll upload some binaries if someone asks for them.

PS: Here they are: Plain Xml Content Importer

Tags: content, csharp, xna