C++ function exported in dll and loaded from C# - c#

I have in C++:
void __declspec(dllexport) foo(HWND wnd)
And in C#
[DllImport("MyDll.dll", CharSet = CharSet.Ansi)]
public static extern void foo(IntPtr wnd);
When I'm trying to call it I have this error - Additional information: Unable to find an entry point named 'foo' in DLL. I tried to inspect the dll and I have the function with the fallowing definition:
Undecorated C++ Function:
void cdecl foo(struct HWND *)
I searched on several forums and is seems that this is the right way to do this... Do you know why I have this run time error?

You need to disable C++ name mangling. Declare your native function like this:
extern "C" __declspec(dllexport) void foo(HWND wnd)
You can use the dumpbin.exe utility to see DLL exports as well.

Related

Sending Unicode string from C# to C++ working in one project but not another

I have built a test harness solution with two projects in VS2017, a C# project where I am sending a unicode string and a C++ DLL where I receive it and show it in a MessageBox.
My C# code is:
[DllImport(#"TestDLL.dll", CallingConvention = CallingConvention.StdCall, CharSet = CharSet.Unicode)]
private static extern void SendString([MarshalAs(UnmanagedType.LPWStr)] string str);
private void Button_Click(object sender, RoutedEventArgs e)
{
SendString("Test2 String ☻我是美国人");
}
My C++ code is:
__declspec(dllexport) bool __stdcall SendString(wchar_t *IncomingPath)
{
MessageBoxW(
NULL,
IncomingPath,
L"Header",
MB_ICONINFORMATION
);
return true;
}
This works as expected with output like this:
Great. So I transplanted the code into a larger solution, again with the same small C# test project as described, and a larger existing C++ project where I need to include the same string passing functionality.
This time my C# code is very similar:
[DllImport(#"my.dll", CallingConvention = CallingConvention.StdCall, CharSet = CharSet.Unicode)]
private static extern void ProcessOneFile([MarshalAs(UnmanagedType.LPWStr)] string str);
private void Button_Click(object sender, RoutedEventArgs e)
{
ProcessOneFile("Test2 String ☻我是美国人");
}
and C++ code is basically the same too.
__declspec(dllexport) cv::Mat __stdcall ProcessOneFile(wchar_t *FullFilePath) {
MessageBoxW(
NULL,
FullFilePath ,
L"Headervv",
MB_ICONINFORMATION
);
}
Yet I am getting this:
To test, I changed the C++ project from a DLL to an exe and called the ProcessOneFile function locally, using the following code from main() inside the same C++ file:
wchar_t *temp = L"C:/Z2C6BC1C克 - Copy.jpg";
ProcessOneFile(temp);
and I get :
So it seems the data is not surviving the transfer across the boundary from C# to C++.
Both solutions are running on the same PC, both compile in the same Visual Studio (2017) both c++ projects have Properties -> General -> Character Set set to "Use Unicode Character Set".
There must be something project related that is causing this, any help would be appreciated. I am not a C++ guy at all, I am trying to get the functionality I need from code on Github.
Thanks for any advice.
First stop returning the cv::Mat from your function. Exporting a class has never been advisable. Furthermore the OpenCV cv::Mat is not being exported.
Also, to avoid name mangling prefix the function definition with extern "C":
extern "C" __declspec(dllexport) void __stdcall ProcessOneFile(wchar_t *FullFilePath)

C# code can't "find" exported function in a DLL, Works in C++ test program [duplicate]

This question already has answers here:
Unable to find an entry point when calling C++ dll in C#
(3 answers)
Closed 3 years ago.
I created a C++ (native) DLL that exports one function and I'm trying to call in from C# (managed) code and I receive runtime error that the function name can't be found in my DLL.
Here's the code of the DLL:
#include "pch.h"
#include "framework.h"
#include "Device.h"
// This is an example of an exported function.
DEVICE_API bool DeviceAvailable(void)
{
return false;
}
The header exports the function like:
#ifdef DEVICE_EXPORTS
#define DEVICE_API __declspec(dllexport)
#else
#define DEVICE_API __declspec(dllimport)
#endif
DEVICE_API bool DeviceAvailable(void);
In C# I import the function as such:
.. enclosing code ...
internal static class NativeMethods
{
[DllImport("Device.dll", CharSet = CharSet.Auto)]
[return: MarshalAs(UnmanagedType.Bool)]
public static extern bool DeviceAvailable();
[DllImport("user32.dll", CharSet = CharSet.Auto)]
[return: MarshalAs(UnmanagedType.I4)]
public static extern int MessageBox(IntPtr hWnd, String text, String caption, uint type);
}
When I call NativeMethods.DeviceAvailable(), I get an error on invocation:
System.EntryPointNotFoundException
HResult=0x80131523
Message=Unable to find an entry point named 'DeviceAvailable' in DLL 'Device.dll'.
The MessageBox function works fine when called by the program (C#). A C++ test program is able to find and invoke the DLL function.
I guess your problem related to access settings on function level when creating the mentioned functions ic DLL.
here
you can find a useful links could help you to understand the case better and to find the best solution.

using cpp dll in c# code

I'm trying to use native c++ dll inside a c# application.
I followed this article to get started.
Its working fine while the dll using only c files (like in the tutorial), but when using cpp files I cant make it work.
using the dumpbin tool I can see that my exported function names changed.
for example a function call 'next' changed to '?next##YAHH#Z' and when I'm trying to call it in the c# code it cant find it.
my dll code is:
__declspec(dllexport)
int next(int n)
{
return n + 1;
}
the c# code
[DllImport("lib.dll", CallingConvention = CallingConvention.Cdecl)]
extern static int next(int n);
its the same code while using c file or cpp files
Thanks in advance
Amichai
What you are attempting to do is to make a trampoline function, to externalize the use of your C++ objects.
What you are experiencing is called symbol mangling. It is a c++-specific way of naming the symbols (like the functions) in a dynamic library and an executable, to enable polymorphism.
You have to specify that you want to cancel the mangling effect by using an extern "C" statement in your library code, just like this:
extern "C" __declspec(dllexport)
int next(int n)
{
return n + 1;
}
More on this here: Exporting functions from a DLL with dllexport
Please provide the Entry Point
An example below
[DllImport("lib.dll", CallingConvention = CallingConvention.Cdecl, EntryPoint = "next", CharSet = CharSet.Ansi)]
static extern int next(int n);
Hopefully It helps.

How do I use a DLL function in C# without adding the DLL as reference?

I'm trying to use functions from a DLL that can't be added as Reference through Visual Studio (a message saying "reference cannot be added" appears). However, I've gotten instructions from the creator of the DLL and they suggested I use them like this in VB.Net:
Private Declare Function Prn_Init Lib "VAx_VPOS396_APPAPI.dll" () As Integer
That works, but now I want to write a program in C#. How do I "translate" that declaration to C#?
Additional: In C++ the declaration comes in a *.h file with these lines:
#ifndef _VPOS396DLL_API_H
#define _VPOS396DLL_API_H
VPOS396_DLL_API int Prn_Init(void);
You should create the method(s) you want to use in C#, make them extern and add a [DllImport] attribute. For example:
[DllImport("kernel32.dll")]
static extern bool Beep(uint dwFreq, uint dwDuration);
See https://msdn.microsoft.com/en-us/library/aa984739%28v=vs.71%29.aspx
[System.Runtime.InteropServices.DllImport("VAx_VPOS396_APPAPI.dll")]
public static extern unsafe Int32 Prn_Init(void);
("unsafe" is optional)

Call native method from dll

I am trying to use MimeTex.dll in Mixed Mode of C++/CLI project. I include the dll by:
#pragma comment(lib,"MimeTex.dll")
and I tried to call this method:
CreateGifFromEq("expression","path");
but the compiler inform that it doesn't know CreateGifFromEq() method.
I didn't find resources in the web in how to use MimeTex.dll in C++. I just find how to use it with C# in this link by Pinvok like:
[System.Security.SuppressUnmanagedCodeSecurity()]
internal class NativeMethods
{
private NativeMethods()
{ //all methods in this class would be static
}
[System.Runtime.InteropServices.DllImport("MimeTex.dll")]
internal static extern int CreateGifFromEq(string expr, string fileName);
[System.Runtime.InteropServices.DllImport("kernel32.dll")]
internal extern static IntPtr GetModuleHandle(string lpModuleName);
[System.Runtime.InteropServices.DllImport("kernel32.dll")]
[return: System.Runtime.InteropServices.MarshalAs(System.Runtime.InteropServices.UnmanagedType.Bool)]
internal extern static bool FreeLibrary(IntPtr hLibModule);
}
and then call it like:
NativeMethods.CreateGifFromEq(equation, tempGifFilePath);
How I can call it without Pinvok in mixed mode of C++/CLI?
Surely you meant to write:
#pragma comment(lib,"MimeTex.lib")
In other words, you pass the .lib file to the linker rather than the .dll. When you compiled the DLL, a .lib file will have been generated.
But that's not your immediate problem. The compiler has no declaration for CreateGifFromEq. That's because you have not included the library's header file in your C++ code. Doing that should resolve the issue.
If all you need is that one function then it should be trivial to declare it in your C++ code.
__declspec(dllimport) extern int CreateGifFromEq(char *expr, char *fileName);
You probably will need to wrap that in an extern "C" block too.
Actually, having looked at the library, the header file that is supplied with the source does not declare that function, even though it's present in the library.
In C++/CLI you can use P/Invoke as in C#.
If you don't want to use P/Invoke, the other way is to load the DLL with LoadLibrary and get pointers to the functions with GetProcAddress.
Here's the equivalent of the C# code in C++/CLI:
using namespace System;
using namespace System::Runtime::InteropServices;
private ref class NativeMethods abstract sealed
{
internal:
[DllImport("MimeTex.dll")]
static int CreateGifFromEq(String^ expr, String^ fileName);
[DllImport("kernel32.dll")]
static IntPtr GetModuleHandle(String^ lpModuleName);
[DllImport("kernel32.dll")]
[returnvalue: MarshalAs(UnmanagedType::Bool)]
static bool FreeLibrary(IntPtr hLibModule);
};
As you can see it's almost identical to the C# code except for some minor syntax changes.
Here's how you would use LoadLibrary and GetProcAddress (which has the advantage of not requiring marshaling, which is unnecessary in C++ anyway):
HMODULE hModule = LoadLibrary(L"MimeTex.dll");
int (*fpCreateGifFromEq)(WCHAR*, WCHAR*) = GetProcAddress(hModule, "CreateGifFromEq");
(*fpCreateGifFromEq)(L"expression", L"filename");
You don't need to do that for GetModuleHandle and FreeLibrary, because they're from kernel32.dll so you can just include Windows.h and call them normally.
There's a cleaner way of doing things:
Create a C++/CLI project. It will be your interface between C# and C++ (this is what C++/CLI is aimed at).
Make your C# project have a reference to this new assembly.
Inside your new C++/CLI project, configure project options so as you point to the directory with mimetex .h files for the compiler to compile your project, and the directory with .lib/.dll for the linker.
Inside your C++/CLI part, you just write C++ inside the functions. They are exposed to your C# assembly, but in the inside it's just calls to C++ functions as usual.
(You should check a C++CLI tutorial to know what to do with strings to convert them to native char or std::string. I remember I had to look at that.)*
In your C# part, you just call the C++CLI project assembly exposed functions like normal C# functions, without noticing it's native code called behind.

Categories

Resources