Ideal way to include headers from a DLL? - c++

I've been testing things with multiple project setups where a console application accesses functions from several DLLs.
I thought about how to include the DLLs' headers in the console application. My current implementation is as follows but it's being a headache to manage and even sometimes runs into errors:
Each DLL project has a folder called "Include"
The Console application project references the Include folder of each DLL project (as suggested by msdn guide to working with DLLs)
Each DLL project contains a single header that includes all the headers in that one project
The console application then #includes these "master headers"
Every project uses the pre-compiled header "stdafx" and each file #includes it.
It worked out fine until I started overloading operators.
I believe the grief is caused by the pre-compiled header somehow, here's and example of my current usage of stdafx:
#define DLL // Found in every DLL, not in the console project
#ifdef DLL
#define DLLEI __declspec(dllexport)
#else
#define DLLEI __declspec(dllimport)
#endif
#include <iostream>
#include <vector>
#include "Include\Engine.h"
using namespace std;
With this I sometimes get some irrelevant random compiler errors that I can hackfix by excluding the header from the "master header" and including the troublemaker separately in the console app.
Suggestions what could be done better?

__declspec(dllexport) and __declspec(dllimport) definitions should be placed in every public Dll include file, or at least in the main Dll public include file, which includes all other files. These definitions should not be in stdafx.h.
#define DLL // Found in every DLL, not in the console project
This is incorrect, every Dll must have unique preprocessor definition. In your case, of one Dll depends on another, it always compiles another Dll functions as __declspec(dllexport)
Ensure that every header has #pragma once in the beginning.
Consider using common Include directory for all projects.
As already mentioned in the comments, using namespace can be used only in source files.

This article answered quite a few of my questions.
http://www.codeproject.com/Articles/6351/Regular-DLL-Tutor-For-Beginners

Related

C++ preprocessor executing both #ifdef and #ifndef

So I'm currently working on something that uses OpenCL. The OpenCL spec provides the users with a directive which must be included before the inclusion of the header (cl.h)
#define CL_TARGET_OPENCL_VERSION 110
Which basically defines the version they want to use. Suppose I'm making a library and I want my users to define this instead of me defining this inside my files. What I did was.
-----main.cpp---
#define CL_TARGET_OPENCL_VERSION 110
#include "library.h"
-------x---------
----library.h-----
#ifdef CL_TARGET_OPENCL_VERSION
#pragma message("def")
#endif
#ifndef CL_TARGET_OPENCL_VERSION
#pragma message("ndef")
#endif
.... include other headers.
--------x---------
And the compiler prints both def and ndef messages. And the OpenCL library also throws a warning that it's undefined. I thought that the library header would get substituted into main and it'd only print the def message. Is there anything I understood wrong?
I'm particularly confused as to where does the preprocessor start? If it starts from main.cpp and goes from top to down, then it surely has defined the macro. After that it sees the library inclusion, then it should only print the def message but it prints both.
This leds me to believe the preprocessor does scan the header file before including it in main? Dunno the reason why. Also I have assured that the library header isn't included elsewhere.
One interesting thing I noticed was, if i did this
-----helper.h---
#define CL_TARGET_OPENCL_VERSION 110
-------x---------
----library.h-----
#include helper.h
#ifdef CL_TARGET_OPENCL_VERSION
#pragma message("def")
#endif
#ifndef CL_TARGET_OPENCL_VERSION
#pragma message("ndef")
#endif
.... include other headers.
--------x---------
It prints the def message "twice". If anybody can explain all this I'd be grateful.
EDIT:- The files I'm compiling are main.cpp library.h and library.cpp
Library.cpp includes library.h from the start as usual. Maybe this other cpp is causing the problem?
In C/C++ programs, the compiler handles each .c and .cpp file separately.
The compilers build each source file (NOT the header files, only .c and .cpp files) independently from each other (this source files are called compilation unit).
Thus, when your main.cpp is built, the compiler finds the #define CL_TARGET_OPENCL_VERSION 110 you have added on top of the main.cpp file, emiting the defmessage.
But when the compiler builds the library.cpp file, it does not find the version define, so it emits the ndef message.
So, following this explanation, it is completely normal that in your last case, when you add the define to the .h file, the compiler emits the def message twice, once for the main.cpp file and once for the library.cpp file.
Now, the problem is where should you add the define, in order to have the program built consistently, with the same version for all the .cpp files.
Usually, all the IDEs have some configuration page where you can add global defines, for all the project, which are "inserted" into all the compilation units before everything else. So when the IDE calls the compiler, it passes the same defines to all the compilation units. You should add this kind of defines in this page.
In your IDE (I am using Code::Blocks, v 17.12), you can find this page in the menu: Project / Build Options
For each type (Debug or Release), you have to go to the tab Compiler Settings, and there to the sub tab #defines. There you can add global defines, which can be different if you are building in Debug or in Release mode (of course, if you set the same in both modes, they would be the same).
Once you have added your define here, please, remove it from the main.cpp, library.h and any other place where you may have added it, in order to avoid duplicities.
From the comments about portability:
You have several options:
Always use Code::Blocks: this would be the easiest way, since you can pass the Code::Blocks project along with the source files, and everything would be already setup.
Use cmake, which is a script build system, where you can set defines and so in the same way as using an IDE. cmake is much widely used than Code::Blocks, so maybe it is a better option.
Add a new options.h header file, where you set all the defines, and include it to all your .c/.cpp. This setup has the additional benefit that for different systems, changing only the options.h file the build can be completely different. This is a manually setup of what the IDE is doing. It has the advantage that does not rely on external tools, but the disadvantage that you have to remember to add it in all the new .cpp files added to the project.
My recommendation is go with cmake, just as the others have said.
Prefer using #ifndef XXXX_h #define XXXX_h #endif over #pragma once
If your #include search path is sufficiently complicated, the compiler may be unable to tell the difference between two headers with the same basename (e.g. a/foo.h and b/foo.h), so a #pragma once in one of them will suppress both. It may also be unable to tell that two different relative includes (e.g. #include "foo.h" and #include "../a/foo.h" refer to the same file, so #pragma once will fail to suppress a redundant include when it should have.
This also affects the compiler's ability to avoid rereading files with #ifndef guards, but that is just an optimization. With #ifndef guards, the compiler can safely read any file it isn't sure it has seen already; if it's wrong, it just has to do some extra work. As long as no two headers define the same guard macro, the code will compile as expected. And if two headers do define the same guard macro, the programmer can go in and change one of them.
#pragma once has no such safety net -- if the compiler is wrong about the identity of a header file, either way, the program will fail to compile. If you hit this bug, your only options are to stop using #pragma once, or to rename one of the headers. The names of headers are part of your API contract, so renaming is probably not an option.
(The short version of why this is problematic to use #pragma is that neither the Unix nor the Windows filesystem API offer any mechanism that guarantees to tell you whether two absolute pathnames refer to the same file.)

C++ Visual Studio DLL File

My objective is to create a dll and lib file so i'm following this guide
1) I created a new win32 Console Application project in VS, chose DLL as 'Application Type' and Emptied Project
2) I'm trying to create a database in C++. So I have 5 headers with ONLY function declarations.
3) Database.h is my top header, it looks like this:
#ifdef DBDLL_EXPORTS
#define DBDLL_API __declspec(dllexport)
#else
#define DBDLL_API __declspec(dllimport)
#endif
#ifndef __Database_H
#define __Database_H
#include "Table.h"
class DBDLL_API Database { ... };
#endif
4) Now with only headers, I tried compiling the project. It compiled but I don't see any DLLs or Libs anywhere in the project folder. How do I create them?
This is because headers are not compiled -- only CPP files are compiled (which pull in the headers -- a source file and all the headers it pulls in is called a "translation unit", which is the thing actually being compiled (independent of other TUs)).
So, in effect, the build system thinks you're building an empty project, and in this case generates nothing at all. Note that even if the headers are pulled in and compiled, unless an exported symbol is actually referenced somewhere, it may be optimized out. In such cases, you will get an (empty) DLL, but no .lib file (which can cause errors down the line if you have dependent projects looking for this .lib before there's anything in the DLL).
You'll need to create some CPP files that implement the functions declared in the headers. If you have everything in the headers and don't need any CPP files at all, then there's no point in having a DLL! (Since you could include the headers in any other project without needing any other dependency.)
Finally, your include guard (#ifndef ...) should encompass the entire header, including the DLL export macros, otherwise you'll eventually get macro redefinition warnings.
I would suggest using #pragma once instead of include guards since it's simpler and recognized by pretty much every compiler. I would also put the DLL macro goop into its own header (e.g. "dllmacros.h") and include that where it's needed (this will reduce duplication).

Dynamic libraries; how to fix useless dependencies?

I have a bunch of Dll's in my project, using VStudio 9.0 compiler, pre-compiled headers are used for all my Dll's. The dll's loading process is done by implicit caller (.dll, .lib and header must be supplied).
I typically create a different macro definition file per dll, for instance if my current Dll is called MYMacroDll I add a file _MyMacroDll.h which contains:
#ifndef _MYMACRODLL_H_
#define _MYMACRODLL_H_
#ifdef _WIN32
#ifdef MYMACRODLL_EXPORTS
#define MYMACRODLL_API __declspec(dllexport)
#else
#define MYMACRODLL_API __declspec(dllimport)
#endif
#define MYMACRODLL_CALL __cdecl
#else
#define MYMACRODLL_API
#define MYMACRODLL_CALL
#endif
#endif
I know the code can be simplified but I keep the last defines for portability, at least, that's what I understood...
The pre-compiled header file pch.h, will always include the macro definition file _MyMacroDll.h, this makes things easier to me, 'cause later I can decide whether a new class or function will be interfaced or not. This so far works correct.
The confusion comes from using the dll interfaces in another dll; let's suppose a second dll ImageLoaderDll. This one uses instances or references of one (or several) of the interfaced classes/functions in _MyMacroDll. At first glance I guessed there was no need of including the _MyMacroDll.h, but when compiling the ImageLoaderDll it complains with
error C2470: '_AnInterfaceClassFromMyMacro' : looks like a function definition, but there is no parameter list; skipping apparent body
Then I have to include the _MyMacroDll.h in the pre-compiler header file of the other Dll, my project is becoming really messy and I find more and more useless dependencies.
What am doing wrong? Is there another way of setting up the macro definitions so I can avoid adding it to the client Dll's? Am not an expert regarding software design, but in this situation the more decoupled the better.
Hope my explanation was good enough.
If you are using the DLL interface from MyMacroDll in ImageLoaderDll, then you do have a dependency. ImageLoaderDll should include _MyMacroDll.h, otherwise you can't call its functions correctly. It's exactly the same as including <string.h> when you want to call strlen.

MFC winsock1 and 2

I'm deep in trouble with my project.
I have to link two radar with my program, but the first has its own library that uses winsock, while in the second I want to use winsock2.
How can i do that?
At the moment i get many redefinition errors from the includes within winsock.h and winsock2.h.
Take into account that the first radar library is already an DLL, i've got only a header and lib file (no source).
Thank you in advance for any answer.
You could possibly work around the compilation problem by structuring your code (and precompiled headers) so that no file includes both winsock.h and winsock2.h, this may mean either not using precompiled headers at all or using them in a more complex way than is normal in MFC projects...
You could wrap each DLL in a COM object and access them via COM from your main program. This has the advantage of completely separating the use of the two DLLs from your main compilation.
You could wrap each of the DLLs in a new DLL (one each) which provides an interface to your program that does not need the winsock headers in the interface headers.
Of course this may simply be an issue with your Windows.h include order, try putting this at the top of your precompiled header...
#ifndef _WINDOWS_
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#undef WIN32_LEAN_AND_MEAN
#endif
#include <winsock2.h>

About inconsistent dll linkage

How can I remove this link warning? You can see code segment that causes this warning.
static AFX_EXTENSION_MODULE GuiCtrlsDLL = { NULL, NULL };
//bla bla
// Exported DLL initialization is run in context of running application
extern "C" void WINAPI InitGuiCtrlsDLL()
{
// create a new CDynLinkLibrary for this app
new CDynLinkLibrary(GuiCtrlsDLL);
// nothing more to do
}
warning C4273: 'InitGuiCtrlsDLL' : inconsistent dll linkage
I have also export and import definitions, like:
#ifdef _GUICTRLS
#define GUI_CTRLS_EXPORT __declspec(dllexport)
#else
#define GUI_CTRLS_EXPORT __declspec(dllimport)
#endif
The purpose of the preprocessor statements:
#ifdef _GUICTRLS
#define GUI_CTRLS_EXPORT __declspec(dllexport)
#else
#define GUI_CTRLS_EXPORT __declspec(dllimport)
#endif
is to make sure that the header file declares the class or function as __declspec(dllexport) in the .dll where it is defined, and as __declspec(dllimport) for any other .dll that might want to use it.
For this to work, _GUICTRLS must be defined when compiling the exporting .dll, and not defined for any other .dll. Generally you would expect _GUICTRLS to be defined in the project properties, under C/C++ -> Preprocessor -> Preprocessor Definitions.
The compiler error you are seeing usually happens because either _GUICTRLS is not defined for the project that is doing the export, or it is defined for multiple projects, usually resulting from cutting an pasting from one project to another. You will also see this if _GUICTRLS is defined in a header file that is included in multiple projects.
There are multiple possibilities:
1) static AFX_EXTENSION_MODULE GuiCtrlsDLL = { NULL, NULL };
You use AFX_EXTENSION_MODULE. This means that you are implementing an MFC extension DLL. For such extension dlls you have to define the preprocessor _AFXEXT. Set this in the C++ compiler settings of your Visual C++ project
see:
How To Use _declspec(dllexport) in an MFC Extension DLL: http://support.microsoft.com/kb/128199
AFX_EXTENSION_MODULE Structure: http://msdn.microsoft.com/en-us/library/sxfyk0zk.aspx
TN033: DLL Version of MFC: http://msdn.microsoft.com/en-us/library/hw85e4bb.aspx
2) It is likely that you have a duplicated definiton/declaration.
In addition to reading the warning message, pay attention to where it occurs if you have multiple projects as part of a workspace.
I wasted time looking for a problem in my DLL which was compiling and linking correctly. The workspace was also building the main application and my error was that I had inadvertently included a new (DLL) source file into the build file list of the application itself.
The main program requires the DLL header mynewdll.h to import things but does not require the source file mynewdll.cpp. (The code is brought in at run time with a DLL.) I have a habit of including header and code files into projects as a pair, and this is where I had gone wrong.
I would have detected the error much sooner if I had been alert and noticed that the DLL project linked with no errors and it was the main program that complained!
My DLL source code and project was error free and it was only the way I tried to build my executable that was faulty.
That warning is usually caused by a duplicate definition of a function with different use of dllimport. Are you sure you didn't do this?
[ CMake inconsistent dll linkage ]
I encountered the following issue + solution with the __declspec(dllexport) + __declspec(dllimport) :
# # #CMakeLists.txt
add_defintions(-DMYLIB=1)
# The above was the solution...
# (MYLIB is used in the standard ifdef + define MYLIB_EXPORT syntax)
# Below: seems to get overruled by other directory's headers:
set_source_files_properties( file1.h file2.h COMPILE_FLAGS "-DMYLIB=1")
This was annoying because a number of sources say to use the 'set source file properties' command to get better granularity but the doc is not clear on what happens to file1.h's declares when included from a different directory... better stick with add_definitions( -DMYLIB=1 ) for now!
To catch this problem: in your Foo.cpp file:
#include "export.h"
#if defined(MYLIB)
#if defined(OTHERLIB)
static_assert(0,"error, check your definitions!");
// OTHER depends on MY; can't have both of these flags being set!
#endif
#endif
struct OTHER_EXPORT foo
{
};
See that you are not defining the exported symbols in a different project. Also clean all the intermediate files by hand and recompile.
To elaborate the answer of damian with an example. I read it but didn't understand at first glance.
You have a shared library with a source file compiled in that contains the function. In a new project you use the library and in addition you compile also the source file to use the function (I forgot that it is already in the library). Within the library the functions label is exported, within the additional compiled source file the functions label is marked to be imported. That's the conflict.
In my case, error C4273 was caused by trying linking to .lib file from a DLL dynamic load tester app in Qt5 by msvc2017_64 toolchain. Removing the reference to .lib file by changing LIBS setting in .pro file have the problem solved.