C# 3.0 version support on NetMF v4.0

[Armand R.]

Is there a description of what language features are supported? I think
generics, lambda expression, linq are not supported. But looks like the C#
3.0 initializers are supported. Can we get some more detail on the .Netmf
compiler specs?

Thanks!

-Armand

[Lorenzo Tessiore]

Hi Armand,

we do use the same compiler that any otehr C# project uses: csc.exe. The
difference is the we do postprocess the DLLs into our custom format. We do
suppport all compilers features that do not need a runtime or class library
support we do not have:
e.g. waht you listed below is not supported, but we do support anonymous
methods, since that is just a pure compiler feature.

I will try and make a comprehsive list of what we support and what not once
we have the community web up&running.

Regards
Lorenzo

[Armand R.]

Thanks Lorenzo! Please do let me know when the community site is up. I'd
like to contribute to it.

-A

[Jan Kučera]

Hi Armand

>> > Is there a description of what language features are supported? I
>> > think
>> > generics, lambda expression, linq are not supported. But looks like
>> > the C#
>> > 3.0 initializers are supported.

Actually the only unsupported feature is generics. All stuff which needs
generics is therefore not available, but this does not include LINQ for
example.

C# 3.0
- Automatic Properties SUPPORTED
- Object Initializers SUPPORTED
- Collection Initializers SUPPORTED
- Implicitly Typed Local Variables and Implicitly Typed Arrays SUPPORTED
- Extension Methods SUPPORTED (you need to declare an attribute)
- Lambda Expressions SUPPORTED (if you avoid Func<> declarations)
- Expression Trees REQUIRES GENERICS
- Query Expressions SUPPORTED (but you need to declare them yourself on
IEnumerable)
- Anonymous Types REQUIRES GENERICS

C# 4.0
- Dynamic lookup DLR N/A
- Named parameters SUPPORTED
- Optional parameters SUPPORTED (you need to declare some attributes)
- Variance REQUIRES GENERICS

Hope this helps. If I am wrong, great anybody corrected me.
Jan

[Armand R.]

Jan, thanks for the clarification. This also makes sense based on Lorenzo's
description.

-a

"Jan Kučera" wrote:

> .
>

[MtheK]

Multidimension Arrays also seem to be UNSUPPORTED!

I get an E_FAIL error from the metadata processor when I use theM
int tmp[,,] = { ... };

This is important to know, because VisualStudio/ the Medadata
Processor won't tell you about the location of the error or what kind
of error it is ... just "E_FAIL"

On Nov 25, 8:02 pm, Armand R. <Arma...@discussions.microsoft.com>
wrote:

[Steve Maillet]

Correct, You can have Jagged arrays, which are essentially arrays of arrays
but all arrays are SZARRAYS, SZARRAYS as defined by the CLI spec are single
dimensioned with the LowerBound==0.

(
"MtheK" <m.kle...@gmail.com> wrote in message
news:2e342831-3d62-48c9...@c3g2000yqd.googlegroups.com...