Can I use a .NET 4 feature while targeting .NET 3.5 SP1?

丶灬走出姿态 提交于 2019-12-05 00:16:49

问题


I would like to use some of the features in .NET 4.0 but still target .NET 3.5 within Visual Studio 2010. Basically I want to have something like:

if (.NET 4 installed) then
    execute .NET 4 feature

This is an optional feature, and I would just like it to run if the system has .NET 4.0 installed. If the system only has .NET 3.5 then the feature would not execute as is it not something that is critical to the application.


回答1:


First of all, you have to target the 3.5 version of the framework but make your program loadable by the 4.0 framework by having an App.config that looks like this (from How to force an application to use .NET 3.5 or above?):

<?xml version="1.0" encoding="utf-8" ?>
<configuration>
  <startup>
    <supportedRuntime version="v4.0"/>
    <supportedRuntime version="v2.0.50727"/>
  </startup>
</configuration>

As for how you activate the 4.0 feature, it depends on the feature you want to use. If it's a method on a built-in class, you can just look for it and use it if it exists. Here's an example in C# (it applies equally to VB):

var textOptions = Type.GetType("System.Windows.Media.TextOptions, " +
        "PresentationFramework, Version=4.0.0.0, " +
        "Culture=neutral, PublicKeyToken=31bf3856ad364e35");
if (textOptions != null)
{
    var setMode = textOptions.GetMethod("SetTextFormattingMode");
    if (setMode != null)
         // don't bother to lookup TextFormattingMode.Display -- we know it's 1
        setMode.Invoke(null, new object[] { this, 1 });
}

If you put this in your MainWindow constructor, it will set TextFormattingMode to Display in apps running under the .NET 4.0 framework and do nothing under 3.5.

If you want to use a type that isn't available in 3.5, you have to create a new assembly for it. For example, create a class library project targetting 4.0 called "Factorial" with code like this (you'll have to add a reference to System.Numerics; same C# disclaimer):

using System.Numerics;

namespace Factorial
{
    public class BigFactorial
    {
        public static object Factorial(int arg)
        {
            BigInteger accum = 1;  // BigInteger is in 4.0 only
            while (arg > 0)
                accum *= arg--;
            return accum;
        }
    }
}

Then create a project targetting 3.5 with code like this (same C# disclaimer):

using System;
using System.Reflection;

namespace runtime
{
    class Program
    {
        static MethodInfo factorial;

        static Program()
        {   // look for Factorial.dll
            try
            {
                factorial = Assembly.LoadFrom("Factorial.dll")
                           .GetType("Factorial.BigFactorial")
                           .GetMethod("Factorial");
            }
            catch
            { // ignore errors; we just won't get this feature
            }
        }

        static object Factorial(int arg)
        {
            // if the feature is needed and available, use it
            if (arg > 20 && factorial != null)
                return factorial.Invoke(null, new object[] { arg });
            // default to regular behavior
            long accum = 1;
            while (arg > 0)
                accum = checked(accum * arg--);
            return accum;
        }

        static void Main(string[] args)
        {
            try
            {
                for (int i = 0; i < 25; i++)
                    Console.WriteLine(i + ": " + Factorial(i));
            }
            catch (OverflowException)
            {
                if (Environment.Version.Major == 4)
                    Console.WriteLine("Factorial function couldn't be found");
                else
                    Console.WriteLine("You're running " + Environment.Version);
            }
        }
    }
}

If you copy the EXE and Factorial.DLL into the same directory and run it, you'll get all the first 25 factorials under 4.0 and only the factorials up to 20 along with an error message on 3.5 (or if it can't find the DLL).




回答2:


No you can't. One limited option is to use conditional compilation, like this:

#if NET40
    some 4.0 code 
#else
    some 3.5 code
#endif

but the limitation of this is that it either compiles the code in or it doesn't - you cannot switch execution path at run time. (The conditional compilation symbols can either be declared at the top of the file or in the project properties build tab, or on the command line when compiling the project (so they can be specified as part of an automated build)).

The absolute best thing to do is ensure the .Net 4.0 framework is installed - it's only 49MB for the full version so it isn't huge.




回答3:


Main problem here is, that you can't run code compiled for .NET 3.5 on .NET 4 CLR or vice-versa. You need to recompile again for .NET4.

So you will have 2 executables, one for .NET 3.5 and second for .NET 4. Both will have same code, but you can use Preprocessor Directives, concretly #IF directive, to make differences between those two.

Then you specify specific directive in configuration of both projects.




回答4:


No, because you can't use .NET 4 features without the .NET 4 CLR. The problem is that assemblies are bound at load time, and the assembly is bound to the specific version fo the CLR you compiled for.



来源:https://stackoverflow.com/questions/6119614/can-i-use-a-net-4-feature-while-targeting-net-3-5-sp1