XML Parser

Basic XML parsing using libexpat bindings for .NET |

• Its optimized for memory usage, if you don't listen for any SAX-like events (Start/end tag, comment, etc) will not produce any string from handlers.
• You can really reset parser internal state.
• Don't need check end tag because expat already does.
• Use its internal string pooling (maybe add NameTable would be useful too?).
• No unsafe operations (fixed pointers, etc). All marshaling using a "safe" context.
• Library name hints for multiple platforms. Eg: VCPKG compiles as libexpat.dll, other compilers will produce expat.dll instead, same for Unix (so) and Mac systems (dylib)
• Useful for streaming XML (eg: XMPP protocol). DO NOT try parse an HTML, because HTML syntax is based on SGML (support unclosed tags, some extra chars in attribute names) and XML is an extension of SGML too so. Parsing HTML need an SGML parser instead (except if website have well-formed XML-like syntax, as is, have all opening/closing tags correctly)

Usage

using System.Buffers;
using System.Diagnostics;
using System.Reflection;
using System.Runtime.InteropServices;
using System.Text;
using System.Xml;

static class Program
{
	static void Main(string[] args)
	{
		XmlDocument doc = new();
		XmlElement root = null, current = null;

		using (var parser = new XmlParser())
		{
			parser.OnStartTag += (name, attrs) =>
			{
				Console.WriteLine("StartTag");

				var element = doc.CreateElement(name);

				foreach (var (key, value) in attrs)
					element.SetAttribute(key, value);

				if (root == null)
				{
					root = element;
					doc.AppendChild(root);
				}

				current?.AppendChild(element);
				current = element;
			};

			parser.OnCdata += value =>
			{
				Console.WriteLine("Cdata");
				current?.AppendChild(doc.CreateCDataSection(value));
			};

			parser.OnText += value =>
			{
				Console.WriteLine("Text");
				current?.AppendChild(doc.CreateTextNode(value));
			};

			parser.OnEndTag += name =>
			{
				Console.WriteLine("EndTag");
				if (current.ParentNode is XmlElement e)
					current = e;
			};

			parser.OnComment += value =>
			{
				Console.WriteLine("Comment");
				current?.AppendChild(doc.CreateComment(value));
			};

			using (var fs = File.OpenRead("sample.xml"))
			{
				var bufferSize = 8;
				var nTotalBlocks = Math.Round(fs.Length / (float)bufferSize);
				var nCurrentBlock = 0;

				var buf = new byte[bufferSize];

				while (true)
				{
					int len = fs.Read(buf);

					parser.Parse(buf, len);

					if (len == 0)
						break;

					ReportProgress();
					nCurrentBlock++;
				}

				ReportProgress();

				void ReportProgress()
				{
					Console.Title = $"Write block {(nCurrentBlock)} / {nTotalBlocks} ({((float)fs.Position / fs.Length) * 100f:F2}%)";
				}
			}

		}
		var sb = new StringBuilder();

		var settings = new XmlWriterSettings
		{
			ConformanceLevel = ConformanceLevel.Fragment,
			IndentChars = "  ",
			Indent = false
		};

		using (var writer = XmlWriter.Create(sb, settings))
			root.WriteTo(writer);

		Console.WriteLine(sb);

		Console.ReadKey(true);
	}
}