How to adjust %PATH% for dynamically loaded native dlls? - c#

I'm loading dynamically a .NET assembly that depends on several bunches of native .dlls
located in various folders. But Windows finds those DLLs only if their folders are in the PATH environment variable when my application is started.
I would like to modify my PATH variable from my program to allow finding the necessary libraries.
According to
MSDN
"the search order is as follows: ...
The directories that are listed in the PATH environment variable."
Which instance of the PATH environment variable is used?
Every process has an instance.
I tried Environment.SetEnvironmentVariable("PATH", ...) but it did not help.
I also tried
SetDefaultDllDirectories()
together with
AddDllDirectory()
but these made no difference either.
The symptom is that when %PATH% contains the necessary folders when starting my .exe
(from a CMD prompt – it is a console application), ProcessMonitor shows that the native .dlls are probed in all the PATH folders, and are eventually found.
But when %PATH% does not contain the necessary folders at the time of starting, then the native .dlls are probed in the .exe folder and in SYSTEM32 only (although %PATH% contains far more),
regardless of the above-mentioned SetEnvironmentVariable()/SetDefaultDllDirectories()/AddDllDirectory() calls.
What is going here? What am I doing wrong? Why I cannot adjust the PATH for my process effectively?
Note: The AppDomain.AssemblyResolve event cannot help me because it's not fired when native .dlls
load other native .dlls.

That is because each process inherits its environment from the process that spawned it. And it didn't occur to the dit at Microsoft that something like PATH might change during the course of execution, so the CLR never refreshes the environment during process execution (and doesn't provide a means for the process to do so itself). See http://social.msdn.microsoft.com/Forums/vstudio/en-US/acf2d0f3-143e-4ba5-acdc-76a70a5c9830/environment-variables-refresh?forum=csharpgeneral for details.
Since the loader is resolving references to unmanaged DLLs via the normal Win32 way, you should probably look at P/Invoking these Win32 functions to alter the DLL search order used by the Win32 LoadLibrary() and LoadLibraryEx():
SetDllDirectory().
Adds a directory to the search path used to locate DLLs for the application.
AddDllDirectory().
Adds a directory to the process DLL search path.
SetDefaultDllDirectories().
Specifies a default set of directories to search when the calling process loads a DLL.
RemoveDllDirectory().
Removes a directory that was added to the process DLL search path by using AddDllDirectory.
See also DLL Search Order.

Related

What makes a console application project [duplicate]

I want to know the exact difference between the dll and exe file.
I don't know why everybody is answering this question in context of .NET. The question was a general one and didn't mention .NET anywhere.
Well, the major differences are:
EXE
An exe always runs in its own address space i.e., It is a separate process.
The purpose of an EXE is to launch a separate application of its own.
DLL
A dll always needs a host exe to run. i.e., it can never run in its own address space.
The purpose of a DLL is to have a collection of methods/classes which can be re-used from some other application.
DLL is Microsoft's implementation of a shared library.
The file format of DLL and exe is essentially the same. Windows recognizes the difference between DLL and EXE through PE Header in the file. For details of PE Header, You can have a look at this Article on MSDN
EXE:
It's a executable file
When loading an executable, no export is called, but only the module entry point.
When a system launches new executable, a new process is created
The entry thread is called in context of main thread of that process.
DLL:
It's a Dynamic Link Library
There are multiple exported symbols.
The system loads a DLL into the context of an existing process.
For More Details: http://www.c-sharpcorner.com/Interviews/Answer/Answers.aspxQuestionId=1431&MajorCategoryId=1&MinorCategoryId=1
http://wiki.answers.com/Q/What_is_the_difference_between_an_EXE_and_a_DLL
Reference: http://www.dotnetspider.com/forum/34260-What-difference-between-dll-exe.aspx
The difference is that an EXE has an entry point, a "main" method that will run on execution.
The code within a DLL needs to be called from another application.
There are a few more differences regarding the structure you could mention.
Both DLL and EXE share the same file structure - Portable Executable, or PE. To differentiate between the two, one can look in the Characteristics member of IMAGE_FILE_HEADER inside IMAGE_NT_HEADERS. For a DLL, it has the IMAGE_FILE_DLL (0x2000) flag turned on. For a EXE it's the IMAGE_FILE_EXECUTABLE_IMAGE (0x2) flag.
PE files consist of some headers and a number of sections. There's usually a section for code, a section for data, a section listing imported functions and a section for resources. Some sections may contain more than one thing. The header also describes a list of data directories that are located in the sections. Those data directories are what enables Windows to find what it needs in the PE. But one type of data directory that an EXE will never have (unless you're building a frankenstein EXE) is the export directory. This is where DLL files have a list of functions they export and can be used by other EXE or DLL files. On the other side, each DLL and EXE has an import directory where it lists the functions and DLL files it requires to run.
Also in the PE headers (IMAGE_OPTIONAL_HEADER) is the ImageBase member. It specifies the virtual address at which the PE assumes it will be loaded. If it is loaded at another address, some pointers could point to the wrong memory. As EXE files are amongst the first to be loaded into their new address space, the Windows loader can assure a constant load address and that's usually 0x00400000. That luxury doesn't exist for a DLL. Two DLL files loaded into the same process can request the same address. This is why a DLL has another data directory called Base Relocation Directory that usually resides in its own section - .reloc. This directory contains a list of places in the DLL that need to be rebased/patched so they'll point to the right memory. Most EXE files don't have this directory, but some old compilers do generate them.
You can read more on this topic # MSDN.
This answer was a little more detailed than I thought but read it through.
DLL:
In most cases, a DLL file is a library. There are a couple of types of libraries, dynamic and static - read about the difference. DLL stands for dynamic link library which tells us that it's a part of the program but not the whole thing. It's made of reusable software components (library) which you could use for more than a single program. Bear in mind that it's always possible to use the library source code in many applications using copy-paste, but the idea of a DLL/Static Library is that you could update the code of a library and at the same time update all the applications using it - without compiling.
For example:
Imagine you're creating a Windows GUI component like a Button. In most cases you'd want to re-use the code you've written because it's a complex but a common component - You want many applications to use it but you don't want to give them the source code You can't copy-paste the code for the button in every program, so you decide you want to create a DL-Library (DLL).
This "button" library is required by EXEcutables to run, and without it they will not run because they don't know how to create the button, only how to talk to it.
Likewise, a DLL cannot be executed - run, because it's only a part of the program but doesn't have the information required to create a "process".
EXE:
An executable is the program. It knows how to create a process and how to talk to the DLL. It needs the DLL to create a button, and without it the application doesn't run - ERROR.
hope this helps....
Both DLL and EXE are Portable Executable(PE) Formats
A Dynamic-link library (DLL) is a library and therefore can not be executed directly. If you try to run it you will get an error about a missing entry point. It needs an entry point (main function) to get executed, that entry point can be any application or exe. DLL binding occurs at run-time. That is why its called "Dynamic Link" library.
An Executable (EXE) is a program that can be executed. It has its own entry point. A flag inside the PE header indicates which type of file it is (irrelevant of file extension). The PE header has a field where the entry point for the program resides. In DLLs it isn't used (or at least not as an entry point).
There are many software available to check header information. The only difference causing both to work differently is the bit in header as shown in below diagram.
EXE file has only single main entry means it is isolated application, when a system launches exe, a new process is created while DLLs have many entry points so when application use it no new process started, DLL can be reused and versioned. DLL reduces storage space as different programs can use the same dll.
Dll v/s Exe
1)DLL file is a dynamic link library which can be used in exe files and
other dll files.
EXE file is a executable file which runs in a separate
process which is managed by OS.
2)DLLs are not directly executable . They are separate files containing functions that can be called by programs and other DLLs to perform computations and functions.
An EXE is a program that can be executed . Ex :Windows program
3)Reusability
DLL: They can be reused for some other application. As long as the coder knows the names and parameters of the functions and procedures in the DLL file .
EXE: Only for specific purpose .
4)A DLL would share the same process and memory space of the calling application while an
EXE creates its separate process and memory space.
5)Uses
DLL: You want many applications to use it but you don't want to give them the source code You can't copy-paste the code for the button in every program, so you decide you want to create a DL-Library (DLL).
EXE: When we work with project templates like Windows Forms Applications, Console Applications, WPF Applications and Windows Services they generate an exe assembly when compiled.
6)Similarities :
Both DLL and EXE are binary files have a complex nested structure defined by the Portable Executable format, and they are not intended to be editable by users.
Two things: the extension and the header flag stored in the file.
Both files are PE files. Both contain the exact same layout. A DLL is a library and therefore can not be executed. If you try to run it you'll get an error about a missing entry point. An EXE is a program that can be executed. It has an entry point. A flag inside the PE header indicates which file type it is (irrelevant of file extension). The PE header has a field where the entry point for the program resides. In DLLs it isn't used (or at least not as an entry point).
One minor difference is that in most cases DLLs have an export section where symbols are exported. EXEs should never have an export section since they aren't libraries but nothing prevents that from happening. The Win32 loader doesn't care either way.
Other than that they are identical. So, in summary, EXEs are executable programs while DLLs are libraries loaded into a process and contain some sort of useful functionality like security, database access or something.
The .exe is the program. The .dll is a library that a .exe (or another .dll) may call into.
What sakthivignesh says can be true in that one .exe can use another as if it were a library, and this is done (for example) with some COM components. In this case, the "slave" .exe is a separate program (strictly speaking, a separate process - perhaps running on a separate machine), but one that accepts and handles requests from other programs/components/whatever.
However, if you just pick a random .exe and .dll from a folder in your Program Files, odds are that COM isn't relevant - they are just a program and its dynamically-linked libraries.
Using Win32 APIs, a program can load and use a DLL using the LoadLibrary and GetProcAddress API functions, IIRC. There were similar functions in Win16.
COM is in many ways an evolution of the DLL idea, originally concieved as the basis for OLE2, whereas .NET is the descendant of COM. DLLs have been around since Windows 1, IIRC. They were originally a way of sharing binary code (particularly system APIs) between multiple running programs in order to minimise memory use.
An EXE is visible to the system as a regular Win32 executable. Its entry
point refers to a small loader which initializes the .NET runtime and tells
it to load and execute the assembly contained in the EXE.
A DLL is visible to the system as a Win32 DLL but most likely without any
entry points. The .NET runtime stores information about the contained
assembly in its own header.
dll is a collection of reusable
functions where as an .exe is an
executable which may call these
functions
An exe is an executible program whereas A DLL is a file that can be loaded and executed by programs dynamically.
● .exe and dll are the compiled version of c# code which are also called as
assemblies.
● .exe is a stand alone executable file, which means it can executed directly.
● .dll is a reusable component which cannot be executed directly and it requires
other programs to execute it.
For those looking a concise answer,
If an assembly is compiled as a class library and provides types for other assemblies to use, then it has the ifle extension .dll (dynamic link library), and it cannot be executed standalone.
Likewise, if an assembly is compiled as an application, then it has the file extension .exe (executable) and can be executed standalone. Before .NET Core 3.0, console apps were compiled to .dll fles and had to be executed by the dotnet run command or a host executable. - Source
Difference in DLL and EXE:
1) DLL is an In-Process Component which means running in the same memory space as the client process. EXE is an Out-Process Component which means it runs in its own separate memory space.
2) The DLL contains functions and procedures that other programs can use (promotes reuability) while EXE cannot be shared with other programs.
3) DLL cannot be directly executed as they're designed to be loaded and run by other programs. EXE is a program that is executed directly.
The major exact difference between DLL and EXE that DLL hasn't got an entry point and EXE does. If you are familiar with c++ you can see that build EXE has main() entry function and DLL doesn't :)

DLL can't be accessed on some systems

Our C# / .NET program is calling a third party DLL (x.dll) in the target folder. x.dll is then calling another third party DLL y.dll in the same folder. At some systems, this does not work, because x.dll can't find y.dll.
The same program with the same dlls works on most other systems, but we can't find out what the differences are to the systems were it does not work. We checked the properties, the dlls are not blocked by Windows. Also the virus-scanner was disabled.
Please go through Dynamic-Link Library Search Order and see which case it applies. the most likely case is standard DLL search order with SafeDllSearchMode enabled:
If SafeDllSearchMode is enabled, the search order is as follows:
The directory from which the application loaded.
The system directory. Use the GetSystemDirectory function to get the path of this directory.
The 16-bit system directory. There is no function that obtains the path of this directory, but it is searched.
The Windows directory. Use the GetWindowsDirectory function to get the path of this directory.
The current directory.
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.
Note that if the DLL is a managed assembly (you do not mention it is) then the .Net assembly search rules apply, see How the Runtime Locates Assemblies. Obviously, GAC would also play a role for managed DLLs.
I'm just giving you the pointers to do your own investigations, since your post is missing the relevant info.

DLL depends another DLL but windows can't find it

I am working on an application where users can develop their own plugins as dll files. They need to put these dll files to a specific path so my application(C++) can load them and run. The problem is, when user develops a dll file depends another dll file and put both files to my dll path, my application can't find second dll file.
Ex:
DllA(C++) depends DllB(C#)
my app loads DllA but can't find DllB and fails.
If user puts DllB into application binary path and DllA to specific dll path, I am able to execute plugin. I also tried to put DllB into various directories such as windows or system32 but it again fails. I also tried to add my dll path to PATH environment variable but no effect.
I am already executing SetDLLDirectory function to load DllA in the specific path. Since I've no control over user code, I can't make them to call this function in DllA code also. My question is that how can I force windows to search my specific dll path?

Gets FileNotFound Exception if managed c# class library dll is not in .exe folder when called form unmanaged code (c++)

I have to use the C# class library code in my native code which is written in c++. For this I added one interface written in c++ add the my C# class library project as dependency to it. This works fine till I put the C# dll in .exe folder , but I put this in libs folders (where we keep our dll's), I get file not found exception when tried to call the C# dll function.
Is there any way to keep c# dll in libs folder and make it work?
Thanks.
Yes, you need to re-add it from the libs folder and it will be fine. Or if you are naming the DLL in your code itself, then just provide the full path along with the DLL name. Then when you recompile, the .exe will contain the path you gave it in the source code and will look for the DLL there.
From the MSDN(LoadLibrary),
Remarks section::
The first directory searched is the directory containing the image
file used to create the calling process (for more information, see the
CreateProcess function). Doing this allows private dynamic-link
library (DLL) files associated with a process to be found without
adding the process's installed directory to the PATH environment
variable. If a relative path is specified, the entire relative path is
appended to every token in the DLL search path list. To load a module
from a relative path without searching any other path, use
GetFullPathName to get a nonrelative path and call LoadLibrary with
the nonrelative path. For more information on the DLL search order,
see Dynamic-Link Library Search Order.
The search path can be altered
using the SetDllDirectory function. This solution is recommended
instead of using SetCurrentDirectory or hard-coding the full path to
the DLL.
If a path is specified and there is a redirection file for
the application, the function searches for the module in the
application's directory. If the module exists in the application's
directory, LoadLibrary ignores the specified path and loads the module
from the application's directory. If the module does not exist in the
application's directory, LoadLibrary loads the module from the
specified directory. For more information, see Dynamic Link Library
Redirection.
If you call LoadLibrary with the name of an assembly
without a path specification and the assembly is listed in the system
compatible manifest, the call is automatically redirected to the
side-by-side assembly.
This is all you need to know about path search of a dll. The easy solution, enter the full path in LoadLibrary call or set your PATH environment variable to contain the alternate path of the dll.

Unable to load dll: Module cannot be found

I am using a managed DLL that referencing an unmanaged dll.
I added the unmanaged dll under /External/foo/ directory, and i changed the "Copy To Output Directory" to always.
The problem is that the dlls are copied to the outputDir/External/foo dir and when the runtime tries to find it, it fails with this error: "Unable to load dll: Module cannot be found"
When i copy it directly to the output dir, everything works well.
How can i set the project runtime to search in the sub directories of the outut dir?
Thanks!
You don't want to do this, DLL Search Hell is something you want to avoid. Windows is pretty inflexible about where it looks for DLLs. Rightfully so, DLL Hell is nothing to mess with. The only good place for unmanaged DLLs is in the same directory as your EXE. If you don't want to store the DLLs in the project directory, so copying them is very easy, then you are going to need a post-build event that uses xcopy.exe to copy the DLL to the output directory.
The alternatives are not pretty. If this DLL is implicitly loaded then the only option is to use an installer to add the directory to the system's PATH environment variable. This is a Really Bad idea, way too easy for your program to break. If it is loaded explicitly, either with LoadLibrary or a pinvoke declaration then you have more attractive options since you can alter the search path in your Main() method before the DLL is needed. Either append the path to your local copy of the PATH environment variable by using Environment.SetEnvironmentVariable() or pinvoke SetDllDirectory(). Do beware that you have a deployment problem as well, somebody is going to have to create this subdirectory and copy the DLL there on the user's machine. All great reasons to not do this.
This is something we did in our last project. Write a batch file which copies this unmanaged dll to the output directory. Call the batch file every time the program compiles. You can control the calling of batch file also using a Config key. So it runs based on the config key value true/false. i hope this helps

Categories

Resources