Referencing unmanaged librararies from managed code, adventages and disadvantages ?

Question

HI all,

I would like to call from my C# code, unamanaged library functions like presented below. There are two options and the both works. In this moment "Beep" function is simple and have no input/output parameters, pointers, references... I am wondering in more complex cases what would be adventages and disadvantage of both approches ?

Thanks, Milan.

    [DllImport("kernel32.dll")]
    public static extern bool Beep(uint iFreq, uint iDuration);
    public void TestBeep()
    {
        Beep(300, 3000);
    }

---

    internal delegate bool DelegBeep(uint iFreq, uint iDuration);
    [DllImport("kernel32.dll")]
    internal static extern IntPtr LoadLibrary(String dllname);
    [DllImport("kernel32.dll")]
    internal static extern IntPtr GetProcAddress(IntPtr hModule, String procName);
    public void BeepIt()
    {
        IntPtr kernel32 = LoadLibrary("Kernel32.dll");
        IntPtr procBeep = GetProcAddress(kernel32, "Beep");
        DelegBeep delegBeep = Marshal.GetDelegateForFunctionPointer(procBeep, typeof(DelegBeep)) as DelegBeep;
        delegBeep(50, 1000);//Hz,ms
    }

Answer 1

Your second one is much more complicated than the first but achieves the same thing in this case.

If the name of the DLL and the name of the function are known at compile time, then stick with the first approach. If you don't know the name of the DLL and/or function until run time then the LoadLibary/GetProcAddress approach is your only option.

Answer 2

The P/Invoke marshaller finds an entrypoint in a DLL by using LoadLibrary() and GetProcAddress(). And knows how to convert a C# declaration to the equivalent of a delegate declaration.

Doing this yourself has no advantage beyond maybe a wee bit of efficiency. Which you'd better measure, it is no slamdunk.