the function ToTextSpan needs to be modified as follows:

    public static TextSpan ToTextSpan(IToken token)
    {
        if (token == null)
            return new TextSpan();

        string tokentext = token.Text ?? string.Empty;
        if (token.Text.Equals("\n") || token.Text.Equals("\r\n"))
            tokentext = string.Empty;

        return new TextSpan()
        {
            iStartLine = token.Line - 1,
            iStartIndex = token.CharPositionInLine,
            iEndLine = token.Line - 1,
            iEndIndex = token.CharPositionInLine + tokentext.Length
        };
    }

Thaks!!

For me, VS2010/MEF was unable to resolve the OutputWindowService. I had to change the property type of the GlobalServiceProvider field from IServiceProvider to SVsServiceProvider (Microsoft.VisualStudio.Shell namespace). [Import] public SVsServiceProvider GlobalServiceProvider;

(Also I had to comment out the code contracts part, because I did not wanted to enable contracts in my project.)

I may implement this in http://solutionsfactory.macaw.nl/ , and I’m looking for good strategies to implement intellisense autocomplete, syntax highlighting and error squiggles.

I’m focusing on designing simple, small languages that require only a simple (combined) antlr grammar for the purpose of generating output (PowerShell statements). These are not C like languages but very simple, only containing a handful of statements, identifiers, literals, and lists and tables of values.

After reading your posts, the Antlr books, and playing around with Antlr and VS2010 Editor extensibility, I have these questions:

1) How would you implement syntax coloring and autocomplete from a simple, combined grammar (output=template) instead of a tree grammar? Is using an AST grammar only needed for performance or is it also the simplest way to get both intellisense and translation? My languages are very simple and do not require an AST for translation to output. I would rather have one grammar than two grammars for a language, as long as that will not make the combined grammar too hard to read.

2) How do you structure your grammar for incomplete/invalid input? Do you create two grammars, one for syntax coloring and autcomplete that handles incomplete, invalid input, and one for validation & translation that requires complete, valid input? Or would you introduce some sort of ‘mode’ parameter for the parser rules to indicate how strict they should be? I want to keep my grammars as uncomplicated as possible.

PS You have not been posting here for a long time, have you been working on VS2010 – Antlr integration after Jan 2010? I’d really like to see more posts on this, working code for Antlr – VS2010 MEF integration with intellisense ‘done right’ would be great.

Thx for any help you may be able to offer.

wonderful work, very helpful, indeed. I was wondering whether you could agree to open up some of the source code. I have been using your previous posts on integrating language services in VS quite a lot, but I still don’t get the hang of the MS way of things. Is it possible to get your ANTLR integration sources, or parts of them?

Regardsm chris

I’m working through the tutorial now on a Windows 7 machine, and I just wanted to let you know that section 1, line 24 of Cache.cs has to be replaced with: LoadStartMenu( System.Environment.GetEnvironmentVariable( "PROGRAMDATA" ) + @"\Microsoft\Windows\Start Menu" ); if you want it to work on Vista or 7

The reference documentation can be found here: http://support.microsoft.com/kb/886549 at step 12

Thanks for the writeup!