Ayende @ Rahien

Nexus commented on Breaking the C# compiler, again

Thu, 12 Aug 2010 13:07:52 GMT

"I just think you're idiot. That's why it keeps crashing. Reading your posts reveal complete lameness and idiocy. You're doing things the hard way and you're thinking the universe spins around you. That said - you're idiot. " --> humpbacked lout : Maybe you should rethink that statement with the following in mind: If you don't understand maybe you're idiot, and stop reading this blog.

humpbacked lout commented on Breaking the C# compiler, again

Thu, 12 Aug 2010 06:40:22 GMT

I just think you're idiot. That's why it keeps crashing. Reading your posts reveal complete lameness and idiocy. You're doing things the hard way and you're thinking the universe spins around you. That said - you're idiot.

Bob commented on Breaking the C# compiler, again

Wed, 11 Aug 2010 06:48:04 GMT

MAkes sure project target is set to .net 4.0?

configurator commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 19:34:10 GMT

Interesting, the lambda-syntax fails here too. This doesn't compile: var q = docs.SelectMany(doc => doc.AssociatedEntities); Neither does this: var q2 = docs.SelectMany((dynamic doc) => doc.AssociatedEntities); Both with the same error: 'object' does not contain a definition for 'AssociatedEntities' and no extension method 'AssociatedEntities' accepting a first argument of type 'object' could be found (are you missing a using directive or an assembly reference?) However this works: var q = docs.Where(doc => doc.AssociatedEntities != null); So you can absolutely use dynamic as lambda parameters. Sometimes.

13xforever commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 18:24:58 GMT

It seems like MS have a lot to fix in their c# 4.0 compiler that is dealing with dynamics. 'Cause as simple code as this will crash the compiler: using System; using System.Collections.Generic; namespace ExcelToRavenDbExporter { using TConverter = Dictionary >; internal class Program { private static void Main() { } } }

Elroy commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 17:40:29 GMT

My day-job includes developing a compiler and I can so well relate to this scenario. But, of course, we're talking about VS here. You would obviously expect a lot from the compiler.

tobi commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 17:16:01 GMT

Obviously a compiler bug as the transparent identifier class names should never be visible. On stack overflow you can find other C# compiler bugs and even bugs in the Jitter. I personally have only hit one single compiler bug in C# at all. Cant complain about Microsofts quality standards at all (in which product or lib that you are using for five years have you found <= 1 bug?).

Scubastard commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 14:15:31 GMT

I think I saw a Rob Conery post where he details issues with dynamic C# classes/properties not being able to be resolved with extension methods. Sounds like thats related to whats happening here and as a result your linq extension methods can no longer be found... thats just a wild guess though.

Tom Pester commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 13:12:52 GMT

The compiler rewrites the syntactic sugar that are query expression to (extension) method calls. This is an early pass that the compiler does to get to plain vanilla c# statements. I think the parser gets seriously confused by your line 8 probably in comination with previous build up state. What happens if your rewrite the query expression to plain method calls?

D. P. Bullington commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 12:47:46 GMT

@David Cuccia I hope that was in reference to: "Ayende is not your helpdesk" It goes both ways. :)

FallenGameR commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 08:10:17 GMT

Stupid question.... =) dynamic is in 4.0 and 4.0 is in 2010. Anyway, the code you provided doesn't crush on my side.

FallenGameR commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 08:08:46 GMT

And Studio 2010 says the same error as LinqPad. What studio & target framework are you using?

FallenGameR commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 08:06:22 GMT

When I ran the query in the LinqPad, I get following error on line 5 (before your error): 5: from entity in doc.AssociatedEntities An expression of type 'dynamic' is not allowed in a subsequent from clause in a query expression with source type 'System.Collections.Generic.IEnumerable

Ayende Rahien commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 07:58:57 GMT

FallenGameR, The issue is in the _compiler_, not at runtime

FallenGameR commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 07:57:33 GMT

Try to attach Reflector Pro - it can step into Framework assemblies.

David Cuccia commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 03:06:14 GMT

Sounds like a good SO question.

Dmitry commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 00:52:55 GMT

Have you tried writing (retyping) the same code in another file. I had a couple of weird issues with whitespace that was a result of copying-and-pasting.

Steve Py commented on Breaking the C# compiler, again

Tue, 10 Aug 2010 00:51:38 GMT

Simple options that come to mind: What happens if you break that query down into separate queries? Have you tried setting up to debug into the ,Net source? It might be easier to step through with the extension method syntax: var q = docs.Where(doc => doc["@metadata"]["Raven-Entity-Name"] == "Cases" && doc.AssociatedEntities != null) .SelectMany( doc => doc.AssociatedEntities) .Where(entity => entity.Tags != null) .SelectMany(entity => entity.Tags) .Where(tag => tag.ReferencedAggregate != null) .Select (tag => new {tag.ReferencedAggregate.Id, doc.__document_id}); Should compile down to effectively the same operations, I just find the extension syntax more consistent to follow in code.