Blob_RawFile.cs

using System;
using System.Linq;
using System.IO;

namespace BizHawk.Emulation.DiscSystem
{
	partial class Disc
	{
		internal class Blob_RawFile : IBlob
		{
			public string PhysicalPath
			{
				get
				{
					return physicalPath;
				}
				set
				{
					physicalPath = value;
					length = new FileInfo(physicalPath).Length;
				}
			}
			string physicalPath;
			long length;

			public long Offset = 0;

			BufferedStream fs;
			public void Dispose()
			{
				if (fs != null)
				{
					fs.Dispose();
					fs = null;
				}
			}
			public int Read(long byte_pos, byte[] buffer, int offset, int count)
			{
				//use quite a large buffer, because normally we will be reading these sequentially but in small chunks.
				//this enhances performance considerably
				
				//NOTE: wouldnt very large buffering create stuttering? this would depend on how it's implemented.
				//really, we need a smarter asynchronous read-ahead buffer. that requires substantially more engineering, some kind of 'DiscUniverse' of carefully managed threads and such.
				
				const int buffersize = 2352 * 75 * 2;
				if (fs == null)
					fs = new BufferedStream(new FileStream(physicalPath, FileMode.Open, FileAccess.Read, FileShare.Read), buffersize);
				long target = byte_pos + Offset;
				if (fs.Position != target)
					fs.Position = target;
				return fs.Read(buffer, offset, count);
			}
			public long Length
			{
				get
				{
					return length;
				}
			}
		}
	}
}