Question

I've been researching how to successfully convert C++ pointer-to-member to C# but I haven't found anything useful yet. Let's say I have this function.

typedef int STRUCT::*DEFINED;

protected static Method(STRUCT* sampleStruct, DEFINED pMember)
    {
        return (sampleStruct->*pMember);
    }

I have learned through my research that ->* is a pointer-to-member. In this case, we send a pointer of a member variable in a struct called STRUCT. Since Method isn't really sure which member was sent as a parameter, it accesses it through pointer-to-member sampleStruct->*pMember.

I think Reflection could help to convert this code to C#, or maybe Delegates, but I really don't know how to implement it, and I haven't found any similar example online. Any help will be appreciated.

Thanks, YT

UPDATE

This is how I have implemented this in C#.

Instead of a struct, I created an enum, and a class to represent the C++ struct, like this:

C++ struct

public struct ServerStats
{
    int serverStat1;
    int serverStat2;
    int serverStat3;
    int serverStat4;
    int serverStat5;
}

Now, in C#:

public enum ServerStatsEnum
{
    serverStat1,
    serverStat2,
    serverStat3,
    serverStat4,
    serverStat5,
}

public class ServerStats
{
    public int[] serverStatsArray;      

    public ServerStats()
    {
        int numElementsInEnum = Enum.GetNames(typeof(ServerStatsEnum)).Length;
        serverStatsArray = new int[numElementsInEnum];
    }
}

}

Now, I can access the elements of the array by calling the specific enum, like this:

public static void Operation(ServerStats server1, ServerStats server2, ServerStatsEnum index)
    {
        Console.WriteLine("serverStatsArray[{0}] in server1 is {1}", index, server1.serverStatsArray[(int)index]);
        Console.WriteLine("serverStatsArray[{0}] in server2 is {1}", index, server2.serverStatsArray[(int)index]);          
    }

It's more code, but it works natively in C# and it's more efficient than other solutions.

Was it helpful?

Solution

You can use delegates to emulate pointer-to-member access.

class C
{
    public int Method1() { return 1; }
    public int Method2() { return 2; }
}

class Program
{
    static void Main(string[] args)
    {
        C myC = new C();
        Func<C, int> ptrToMember1 = (C c) => { return c.Method1(); };
        int i = Method(myC, ptrToMember1 );
    }

    static int Method(C c, Func<C, int> method)
    {
        return method(c);
    }
}

OTHER TIPS

From an implementation standpoint, reflection can do something very similar:

protected static int Method(MyType obj, PropertyInfo member)
{
    return member.GetValue(obj, null);
}

The big difference is on the calling side. Whereas C++ can create a member pointer at compile time, you need to wait until run time to get a PropertyInfo instance, and you need to pass in the member name as a string to get it.

Something like what jyoung suggests with delegates might work:

class Foo
{
    private int m_val;
    public Func<int> GetValGetter()
    {
        return () => { return m_val; };
    }
}

But this is kind of convoluted and not idiomatic, and there's probably a cleaner way to solve it if you look at it from a higher level.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top