Implementation of [CommonMark] 1 specification (passes tests from version 0.20) in C# for converting Markdown documents to HTML.

The current version of the library is also [available on NuGet] nuget.

To convert Markdown data in a stream or file:

using (var reader = new System.IO.StreamReader("source.md"))
using (var writer = new System.IO.StreamWriter("result.html"))
{
  CommonMark.CommonMarkConverter.Convert(reader, writer);
}

To convert data stored in a string:

var result = CommonMark.CommonMarkConverter.Convert("**Hello world!**");

See [wiki] extensibility for an example how the parser can be extended with additional logic.

Important: The converter does not include any HTML sanitizing.

The library uses no references except for System - it has no external dependencies. It is cross compiled to:

• .NET Framework 2.0
• .NET Framework 3.5 Client Profile
• .NET Framework 4.0 Client Profile
• .NET 4.0 Portable Class Library
  • .NET Framework 4.0 Client Profile
  • Silverlight 5
  • Windows 8
  • Windows Phone 8.1
  • Windows Phone Silverlight 8
  • Xamarin.Android
  • Xamarin.iOS
• .NET Framework 4.5
• .NET Framework 5.0 Core (also known as ASP.NET vNext Core CLR)

For working with the source code you will need Visual Studio 2013.3 or newer (Community edition is supported) except for the .NET Core project which requires Visual Studio 2015.

Using a simple tool to compare the performance of various Markdown implementations for .NET yields the following results:

 CommonMark.NET 0.4.1      4 ms   7%      

CommonMarkSharp 0.3.2     30 ms   46%
   MarkdownSharp 1.13     55 ms   84%     (might not conform to CommonMark specification)
     MarkdownDeep 1.5      7 ms   11%     (might not conform to CommonMark specification)
  MoonShine (sundown)      3 ms    6%     (wrapper for a native x86 .dll)
             Baseline     65 ms   100%    (used to compare results on different machines)

This benchmark is very simple and tests the processing of the CommonMark specification document itself (a 115 KB file). The results are provided just for a high-level comparison.

The parser uses algorithms that completely avoid recursion so even specifically crafted input documents will not cause exceptions and will be rendered correctly. This is important because a StackOverflowException cannot be caught and will bring down the entire process.

This library is based on a port of the reference implementation in C, available on [jgm/stdm repo] 2. It follows the same approach - the source is parsed into a syntax tree that can be used to add custom processing if the application needs it and then formatted into HTML.

All tests from spec.txt are merged into the unit testing project and can be executed from within Visual Studio.

The project also includes a slightly modified version of runtests.pl for compatibility with the original implementation. Use CommonMark.Console.exe --perltest as the argument to runtests.pl so that the application can correctly handle input from the Perl script.