How can I specify a [DllImport] path at runtime?

In fact, I got a C++ (working) DLL that I want to import into my C# project to call it's functions.

It does work when I specify the full path to the DLL, like this :

string str = "C:\\Users\\userName\\AppData\\Local\\myLibFolder\\myDLL.dll";
[DllImport(str, CallingConvention = CallingConvention.Cdecl)]
public static extern int DLLFunction(int Number1, int Number2);

The problem is that it's gonna be an installable project, so the user's folder will not be the same (ex : pierre, paul, jack, mum, dad, ...) depending computer/session where it'd be runned on.

So I'd like my code to be a little more generic, like this :

/* 
goes right to the temp folder of the user 
    "C:\\Users\\userName\\AppData\\Local\\temp"
then go to parent folder
    "C:\\Users\\userName\\AppData\\Local"
and finally go to the DLL's folder
    "C:\\Users\\userName\\AppData\\Local\\temp\\myLibFolder"
*/

string str = Path.GetTempPath() + "..\\myLibFolder\\myDLL.dll"; 
[DllImport(str, CallingConvention = CallingConvention.Cdecl)]
public static extern int DLLFunction(int Number1, int Number2);

The big deal is that "DllImport" desire a "const string" parameter for the DLL's directory.

So my question is :: What could be done in this case ?


Solution 1:

Contrary to the suggestions by some of the other answers, using the DllImport attribute is still the correct approach.

I honestly don't understand why you can't do just like everyone else in the world and specify a relative path to your DLL. Yes, the path in which your application will be installed differs on different people's computers, but that's basically a universal rule when it comes to deployment. The DllImport mechanism is designed with this in mind.

In fact, it isn't even DllImport that handles it. It's the native Win32 DLL loading rules that govern things, regardless of whether you're using the handy managed wrappers (the P/Invoke marshaller just calls LoadLibrary). Those rules are enumerated in great detail here, but the important ones are excerpted here:

Before the system searches for a DLL, it checks the following:

  • If a DLL with the same module name is already loaded in memory, the system uses the loaded DLL, no matter which directory it is in. The system does not search for the DLL.
  • If the DLL is on the list of known DLLs for the version of Windows on which the application is running, the system uses its copy of the known DLL (and the known DLL's dependent DLLs, if any). The system does not search for the DLL.

If SafeDllSearchMode is enabled (the default), the search order is as follows:

  1. The directory from which the application loaded.
  2. The system directory. Use the GetSystemDirectory function to get the path of this directory.
  3. The 16-bit system directory. There is no function that obtains the path of this directory, but it is searched.
  4. The Windows directory. Use the GetWindowsDirectory function to get the path of this directory.
  5. The current directory.
  6. The directories that are listed in the PATH environment variable. Note that this does not include the per-application path specified by the App Paths registry key. The App Paths key is not used when computing the DLL search path.

So, unless you're naming your DLL the same thing as a system DLL (which you should obviously not be doing, ever, under any circumstances), the default search order will start looking in the directory from which your application was loaded. If you place the DLL there during the install, it will be found. All of the complicated problems go away if you just use relative paths.

Just write:

[DllImport("MyAppDll.dll")] // relative path; just give the DLL's name
static extern bool MyGreatFunction(int myFirstParam, int mySecondParam);

But if that doesn't work for whatever reason, and you need to force the application to look in a different directory for the DLL, you can modify the default search path using the SetDllDirectory function.
Note that, as per the documentation:

After calling SetDllDirectory, the standard DLL search path is:

  1. The directory from which the application loaded.
  2. The directory specified by the lpPathName parameter.
  3. The system directory. Use the GetSystemDirectory function to get the path of this directory.
  4. The 16-bit system directory. There is no function that obtains the path of this directory, but it is searched.
  5. The Windows directory. Use the GetWindowsDirectory function to get the path of this directory.
  6. The directories that are listed in the PATH environment variable.

So as long as you call this function before you call the function imported from the DLL for the first time, you can modify the default search path used to locate DLLs. The benefit, of course, is that you can pass a dynamic value to this function that is computed at run-time. That isn't possible with the DllImport attribute, so you will still use a relative path (the name of the DLL only) there, and rely on the new search order to find it for you.

You'll have to P/Invoke this function. The declaration looks like this:

[DllImport("kernel32.dll", CharSet = CharSet.Auto, SetLastError = true)]
static extern bool SetDllDirectory(string lpPathName);

Solution 2:

Even better than Ran's suggestion of using GetProcAddress, simply make the call to LoadLibrary before any calls to the DllImport functions (with only a filename without a path) and they'll use the loaded module automatically.

I've used this method to choose at runtime whether to load a 32-bit or 64-bit native DLL without having to modify a bunch of P/Invoke-d functions. Stick the loading code in a static constructor for the type that has the imported functions and it'll all work fine.