Which parsers are available for parsing C# code?

Works on source code:

• CSParser: From C# 1.0 to 2.0, open-source
• Metaspec C# Parser: From C# 1.0 to 3.0, commercial product (about 5000$)
• #recognize!: From C# 1.0 to 3.0, commercial product (about 900€) (answer by SharpRecognize)
• SharpDevelop Parser (answer by Akselsson)
• NRefactory: From C# 1.0 to 4.0 (+async), open-source, parser used in SharpDevelop. Includes semantic analysis.
• C# Parser and CodeDOM: A complete C# 4.0 Parser, already support the C# 5.0 async feature. Commercial product (49$ to 299$) (answer by Ken Beckett)
• Microsoft Roslyn CTP: Compiler as a service.

Works on assembly:

• System.Reflection
• Microsoft Common Compiler Infrastructure: From C# 1.0 to 3.0, Microsoft Public License. Used by Fxcop and Spec#
• Mono.Cecil: From C# 1.0 to 3.0, open-source

The problem with assembly "parsing" is that we have less informations about line and file (the informations is based on .pdb file, and Pdb contains lines informations only for methods)

I personnaly recommend Mono.Cecil and NRefactory.

Mono (open source) includes C# compiler (and of course parser)

If you are going to compile C# v3.5 to .net assemblies:

var cp = new Microsoft.CSharp.CSharpCodeProvider(new Dictionary<string, string>() { { "CompilerVersion", "v3.5" } });

http://msdn.microsoft.com/en-us/library/microsoft.csharp.csharpcodeprovider.aspx

I've implemented just what you are asking (AST Parsing of C# code) at the OWASP O2 Platform project using SharpDevelop AST APIs.

In order to make it easier to consume I wrote a quick API that exposes a number of key source code elements (using statements, types, methods, properties, fields, comments) and is able to rewrite the original C# code into C# and into VBNET.

You can see this API in action on this O2 XRule script file: ascx_View_SourceCode_AST.cs.o2 .

For example this is how you process a C# source code text and populate a number of TreeViews & TextBoxes:

    public void updateView(string sourceCode)
    {   
        var ast = new Ast_CSharp(sourceCode);
        ast_TreeView.show_Ast(ast);
        types_TreeView.show_List(ast.astDetails.Types, "Text");
        usingDeclarations_TreeView.show_List(ast.astDetails.UsingDeclarations,"Text");
        methods_TreeView.show_List(ast.astDetails.Methods,"Text");
        fields_TreeView.show_List(ast.astDetails.Fields,"Text");
        properties_TreeView.show_List(ast.astDetails.Properties,"Text");
        comments_TreeView.show_List(ast.astDetails.Comments,"Text");

        rewritenCSharpCode_SourceCodeEditor.setDocumentContents(ast.astDetails.CSharpCode, ".cs");
        rewritenVBNet_SourceCodeEditor.setDocumentContents(ast.astDetails.VBNetCode, ".vb");                                
    }

The example on ascx_View_SourceCode_AST.cs.o2 also shows how you can then use the information gathered from the AST to select on the source code a type, method, comment, etc..

For reference here is the API code that wrote (note that this is my first pass at using SharpDevelop's C# AST parser, and I am still getting my head around how it works):

• AstDetails.cs
• AstTreeView.cs
• AstValue.cs
• Ast_CSharp.cs

http://www.codeplex.com/csparser

SharpDevelop, an open source IDE, comes with a visitor-based code parser which works really well. It can be used independently of the IDE.

Consider to use reflection on a built binary instead of parsing the C# code directly. The reflection API is really easy to use and perhaps you can get all the information you need?

Have a look at Gold Parser. It has a very intuitive IU that lets you interactively test your grammar and generate C# code. There are plenty of examples available with it and it is completely free.

Maybe you could try with Irony on irony.codeplex.com.

It's very fast and a c# grammar already exists.

The grammar itself is written directly in c# in a BNF like way (acheived with some operators overloads)

The best thing with it is that the "grammar" produces the AST directly.

Not in C#, but a full C# 2/3/4 parser that builds full ASTs is available with our DMS Software Reengineering Toolkit.

DMS provides a vast infrastructure for parsing, tree building, construction of symbol tables and flow analyses, source-to-source transformation, and regeneration of source code from the (modified) ASTs. (It also handles many other languages than just C#.)

EDIT (September) 2013: This answer hasn't been updated recently. DMS has long handled C# 5.0

GPPG might be of use, if you are willing to write your own parser (which is fun).