I have a visual studio project called MyKinectDll which I am compiling as a DLL. I have another project called MyKinectApp which is an executable. I have them in the same solution. I have added a dependency from MyKinectApp to MyKinectDll. I have added a reference from MyKinectApp to MyKinectDll. I have added the location of MyKinectDll to the Additional Include Directories of MyKinectApp. Then in my code I have added #include "MyKinect.h", which is one of the headers in MyKinectDll, and it is recognised. But when I compile, only MyKinectDll builds successfully and I get errors in MyKinectApp complaining that there are missing header files. These header files are being called in MyKinect.h and are present elsewhere in MyKinectDll and building correctly on their own.
I am unsure whether I need to link to the .lib of MyKinectDll in linker settings, but I have tried it and I get the same error. What am I doing wrong?
Many thanks for your help.
You need to add the locations of all the headers used by MyKinect.h to the Additional Include Directories for the MyKinectApp project..
This is because you're linking statically (which will link with the lib file, not the dll). The linker needs to look in the header files of the library you're linking to to see the class/method declarations, as they're not available in the lib file.
If you were linking dynamically, the linker could reference the class/function declarations as exported in the dll interface, so you wouldn't need to include the header directories in this case..
Related
I am a cmake newbie (on Debian/Sid/Linux/x86-64)
I forked libonion on https://github.com/bstarynk/onion to enable customization of malloc with Boehm's garbage collector; see this mail thread.
I added two files there onion/src/low_util.c and onion_src/low_util.h (which is #include-d successfully in several other patched files.
It is compiled but not linked.
set(SOURCES onion.c codecs.c dict.c low_util.c request.c response.c handler.c
log.c sessions.c sessions_mem.c shortcuts.c block.c mime.c url.c ${POLLER_C}
listen_point.c request_parser.c http.c ${HTTPS_C} websocket.c ${RANDOM_C} ${SQLITE3_C})
later:
SET(INCLUDES_ONION block.h codecs.h dict.h handler.h http.h https.h listen_point.h low_util.h log.h mime.h onion.h poller.h request.h response.h server.h sessions.h shortcuts.h types.h types_internal.h url.h websocket.h ${SQLITE3_H})
MESSAGE(STATUS "Found include files ${INCLUDES_ONION}")
but when I build, my file low_util.c got compiled but not linked.
Linking C executable otemplate
CMakeFiles/opack.dir/__/__/src/onion/dict.c.o: In function `onion_dict_new':
dict.c:(.text+0x1bc): undefined reference to `onionlow_calloc'
CMakeFiles/opack.dir/__/__/src/onion/dict.c.o: In function `onion_dict_node_data_free':
dict.c:(.text+0x2ec): undefined reference to `onionlow_free'
CMakeFiles/opack.dir/__/__/src/onion/dict.c.o: In function `onion_dict_node_add':
Notice that libonion is a library (in C, providing HTTP service) and that I just want to add a low_util.c file (wrapping malloc & pthread_create etc... to make Boehm's GC happy: it is calling GC_malloc and GC_pthread_create ....) with its low_util.h header. Surprisingly, they get compiled, but do not seems to be linked. And I am not familiar with cmake and I am not familiar with how D.Moreno (the main author of libonion) has organized his cmake files.
Any clues?
Apply the following patch to make it link. The two executables which are being linked with the symbols generated from the .c file you added are missing and are added in the patch.
http://pastebin.com/mDMRiUQu
Based on what you posted, its hard to tell what could be wrong. The cake source code above says that a variable ${SOURCES} is equivalent to onion.c codecs.c dict.c low_util.c ... ${SQLITE3_C}, and a variable ${INCLUDE_ONION} is equivalent to block.h codecs.h dict.h ... ${SQLITE3_H}. You did not provide any targets or the files included in those targets.
A brief list of things that may help:
where do you define the top level library or executable? If your making a library, you will need the command add_library(). If you are making an executable, you will need the add_executable() command.
Use the command target_link_libraries() to resolve dependencies. Rather than placing all of the source files in a single library, group similar together in a single target (a target is defined by the add_* commands), and use this command to link the targets after compilation.
Use the find_package() to get any libraries which are defined on your system but not in you project. Then, link to that library using the target_link_libraries() command.
In this case, if the onion_dict_* functions are defined within the same library, your not including those files in library. When you use add_library or add_executable, ensure you add those files to the list. If the functions are within your project but not in the same library, use the target_link_libraries() command to link to the library which contains the correct files. If those commands are defined in an external library, then first find the library using find_package(), and then link to the library using target_link_libraries().
I have two existing executables A and T, in the same solution that both ran just fine before I touched them. In executable A is a header defining a class P, and a prototype for a static instance MyP. The definitions are compiled in project A. In executable T, I wanted to call member functions of MyP in project A, so I added dllimport/export macros to the declarations of the class and MyP in the headers (not at the definitions), and included the headers in project T. The dllimport/export macros are standard, and A_EXPORTS is defined in project A, but not in T.
#ifdef A_EXPORTS
#define A_API __declspec(dllexport)
#else
#define A_API __declspec(dllimport)
#endif
//various definitions and includes, defining ENUM_RECORDING_TYPE and ERROR
A_API HFILE viosopen(const _TCHAR *path, ENUM_RECORDING_TYPE rt, int flags);
A_API struct P {
ERROR B(SHORT phraseNum);
};
A_API extern P MyP;
I added project A as a dependency on project T in the solution. A still compiles fine, but T comes up with unresolved external symbol "__declspec(import) <snip> referenced in function <snip> for the function calls, and unresolved external symbol "__declspec(dllimport) class P MyP" <snip> for the static object. I also see in the output log, right after it starts linking: Creating library Debug/A.lib and object Debug/A.exp which seems ominous since it's supposed to be linking against the existing executable.
My question is: how can I tell MSVC 2010 where those are? I thought simply setting A as a dependency would have it figure that out automatically. I can link against the existing executable, right?
To statically link your program you don't need the __declspec() stuff and you don't need a separate project to create a LIB file. I think you can just link using the .obj file from your A project.
Your A project has a header file and presumably has a .cpp file that contains the implementation of the items described in that header. Let's say your header file is foo.h and the associated implementation is foo.cpp. When compiled, there should be a foo.obj intermediate file in the <solutiondir>\A\Debug or <solutiondir>\A\release intermediate folder. That file can be used by the linker.
In project T's properties, find Linker | Input and change the "Additional Dependencies" property to include the foo.obj file. One approach would be to use a relative file path to locate the file - for example ..\A\Debug\foo.obj in your debug configuration. Another approach is to use the simple file name in "Additional Dependencies" - foo.obj - and then use Linker | General | Additional Library Directories" to help the linker find the file - e.g., ..\A\$(IntDir). Using the $(IntDir) macro has the advantage that the same value works for Debug and Release settings.
Remember to set up a build dependency from your T project to your A project to be sure the A project is compiled first. Otherwise the foo.obj file might not exist when the T linker comes to look for it. In the Solution properties, select Project Dependencies and then set Project T depends on Project A.
To dynamically link you need to use the A.LIB file as #ajay said. The __declspec(DllImport) tells the compiler what functions and data you are importing but doesn't tell it where you are importing those things from.
Using the A.LIB file as input to the linker is much the same as using the foo.obj file in the statically linking case except that the lib file ends up in the solution output directory <solutiondir>\Debug instead of the project intermediate directory <solutiondir>\A\Debug.
This walkthrough on creating and using a DLL might be useful background.
I asssume project A is DLL not an EXE, which is successfully producing a LIB file.
You need to use the A.LIB as Linker Input in project B. Just producing LIB file wont make other projects automatically link to it.
I am banging my head against a wall here.
All i want is to link a static .lib file in a cpp windows forms application! So, i have an include folder in my project folder that holds header files for that .lib (lib is Yaml-cpp if someone wonders). And i have a lib folder that has the .lib files for that library.
I tested it on a blank project
1.I make a new windows forms project in VS2008, in C++.
2.I go to project properties - c/c++ general and additional include directories that have yaml-cpp header files
3.I go to linker and add path to my lib directory
4.I go to linker - input and add my .lib file
5.I check linker command line and it contains my .lib file so it must be all set.
6.Then i write the sample code in an onbutton function body (which appears in form1.h fie).
Sample code is from here http://code.google.com/p/yaml-cpp/wiki/HowToParseADocument and its just as simple as:
#include <fstream>
#include "yaml.h"
and then:
std::ifstream fin("test.yaml");
YAML::Parser parser(fin);
YAML::Node doc;
while(parser.GetNextDocument(doc)) {
//do nothing yet
}
7.And then i compile and have 10 "unresolved externals".
I have looked into another project that uses same library, and it has exactly same directory structure, same name of .lib added in linker - input - additional dependencies, same .lib and .h files. And it works for that project- but it doesnt work for me.
What in the world is wrong?
EDIT: I tried making new windows 32 console application and it worked correctly. I then tried to make another fresh windows forms application and it failed.
It is confirmed now. Same steps for linking a library work in console application and fail in windows forms application. WHY?
What you've done seems correct from your explanation, let's do some blind troubleshooting.
Try giving full path instead of relative path for library path and additional include directories. (but as u said that the file is getting read, that should be the problem.)
If you are using a copy of original libyaml-cppmdd.lib and its include file, check whether the header file that gets included is of the same version as the lib.
Check whether there are any functions in any part of the code that is having a declaration but no definition.
Check whether you're accessing any private functions from the library.
Please read this Microsoft article on Troubleshooting UnResolved External Symbol error.
Hope it helps!
I'm new to vc++. I've just built a software and it generated a .dll and a .lib. I need to use functions from this in my code. Do I need to link to both .lib and .dll to build my code? What project properties do I have to alter to do this linking?
Actually, you need only the .dll file. It contains all the necessary code and data to run it's functions. It also contains a table that links the symbolic names of the functions (e.g. the function PrintMe), their ordinals (the number of that function in the DLL) and their addresses in the DLL.
If you want to use only the DLL, you have to "manually" get the symbols resolved:
Let's say you want to use the function PrintMe of the DLL. What you had to do is to resolve it's name (PrintMe) or it's ordinal (PrintMe is the 1st function of the DLL) to it's address. For this, you could use LoadLibrary, GetModuleHandle and GetProcAdress from the Win32 API (aka Windows SDK). Additionally, this method allows you to load the DLL at runtime (see below).
The easier way is to use the MSVC(++) features __declspec(dllexport) and __declspec(dllimport), e.g.
// your DLL
__declspec(dllexport) void PrintMe()
{
printf("Hello World!");
}
// you project to use the DLL
__declspec(dllimport) void PrintMe();
The first one (dllexport) tells the compiler to export the function. The second one (dllimport) is the interesting one: It creates all the necessary code to be able to use the function from the DLL.
For this, you need the .lib file in your project (which wants to use the DLL). The .lib file contains information for the linker to resolve the symbol name (PrintMe) to its address in the DLL. Since the .lib is statically bound, the linker can make use of it - the DLL on the contrary is bound at runtime / loading time, so the linker cannot use it. (Yes, the information in the .lib file is redundant.). Note: You cannot change the whole DLL when using this method w/o rebuilding your project with the new .lib file. Some structure changes affect the addresses of the functions in the DLL, see this SO answer.
One last difference between using the Win32 API (LoadLibrary...) and the MSVC method via __declspec is the loading of the DLL. When you use LoadLibrary, the DLL is loaded at runtime, of course (so you can catch exceptions when it cannot be found and so on). The other method loads the DLL at loading time, so you program will terminate (will not run) when Windows cannot find the DLL.
When you create a project in VS, you can activate the "export symbols" checkbox at the end of a wizard (Win32 project). That gives you some examples of exported symbols. Additionally, it introduces a macro plus a preprocessor defition plus some directives that are very useful:
// DLL header
#ifdef _YOUR_DLL_EXPORTS
#define YOUR_DLL_API __declspec(dllexport)
#else
#define YOUR_DLL_API __declspec(dllimport)
#endif
YOUR_DLL_API PrintMe();
You now can use this header file to build you DLL as your DLL project has that _YOUR_DLL_EXPORTS definition (see project properties page, C++, preprocessor). The project that uses the DLL can use this header, too, but then must not have such a name defined. When you include the header file in the project in which you want to use the DLL, the macro is resolved to __declspec(dllimport). This instructs the linker to look for this function (which is found in the .lib file) and create all the necessary code to load the DLL at runtime and resolve the symbol name.
I am using Visual C++ 2008 Express for the first time for a project. And I can't seem to be able to split the .h & .cpp files for classes I'm writing. I was under the impression that you add a header file and prototype the class in there, and then you add a .cpp file with the implementation into your source files directory. Then when you include the .h it would automatically include the .cpp implementation files. Is this correct or am I missing something?
Not sure if this is the same as in Express version. But you can also add a new C++ class with header (.h) and source (.cpp) at the same time by right clicking on the project -> Add -> Class...
By including the .h file using #include, doesn't mean the actual implementation (in another .cpp file) is also include in your source file. The content of .h file which are class and method prototypes is only included. These prototypes allow you to make use of the classes declared in header file (without including real C++ code.)
Each source files (.cpp) are first compiled into object files. All these object file are then linked together to create single executable file. The referenced symbols in, each object file, are actually linked to their implementation during this linking process (http://www.cprogramming.com/compilingandlinking.html)
I don't remember the rules, but sometimes the IDE assumes you're putting all your code in the header file. This is legal, but not a common preference.