I realise this is a newb question but I've been racking my brains for hours.
So I want to use the STK (Synthesis Toolkit) to generate sine waves etc. I've downloaded the source files from https://ccrma.stanford.edu/software/stk/download.html
I unzipped the tar.gz using 7zip.
I opened up the demo.cpp project file (under /projects/demo/demo.cpp) and whenever I try to compile it, I receive lots of errors, all starting with "undefined reference to"
I have set the compiler search directory to include the root folder of the stk kit (unzipped as "stk-4.5.0"). My understanding is that I also have to find file for the linker, that is of the type .lib? Is that correct? I haven't been able to find a .lib file to link to.
I feel like this is a simple fix - what am I missing?
You've downloaded a source , not the compiled library which might be used by linker (.lib file) There are three solutions:
1. Compile whole src folder of your download ( see instructions on their website). Frankly, I've tried to do that recently but failed, especially under Windows although I don't have any experience in this so you can try yourself.
2. Use precompiled library available in some Linux repositories. I had success with Slacko Puppy linux. Package simply adds libstk.a (linux static library) to /usr/lib and stk headers to /usr/include/stk (notice you have to add exactly that path to compiler, since /usr/lib is not enough). This gave me best results.
3. If you use only several classes, you can copy their .cpp files directly to your project source folder. Remember to add parent classes, too (see STK site -> Classes). Again I recommend that you give the compiler path to include folder you've unzipped, but copying only selected headers should also work.
I've found this article helpful (see also A1): http://www.learncpp.com/cpp-tutorial/a3-using-libraries-with-codeblocks/
Hope I've helped
Related
I am trying out the new linux packaging tool http://snapcraft.io/create/
It does have some documentation but it's not very mature and doesn't contain everything.
What I am trying to do is to generate a package for my binary that is compiled using Qt framework, thus it depends on a load of .so files. I couldn't find in documentation any way to convince snapcraft to pull all these dependencies and add them to the program hierarchy (the mounted squashfs).
So in order to "temporarily fix that" I hacked a very lame shell script that executes ldd on my binary and does this on its own. That "sort of" works because now indeed all the necessary libraries are in the image, but the program can't find them, because most of them have own paths and are not in root lib folder /usr/lib
The example of error I am getting when I try to start my program is:
/snap/huggle/x1/usr/bin/huggle: error while loading shared libraries: libGL.so.1: cannot open shared object file: No such file or directory
But the library is there:
/snap/huggle/x1/usr/lib/x86_64-linux-gnu/mesa/libGL.so.1
It just has no clue how to find it. On other hand, given that snapcraft is advertised as awesome simple stupid packaging tool, I find it quite weird that something so simple as packaging a C++ binary would be so complicated. How should I do this properly?
Greetings, what I'm trying to do is to port an existing Windows application to Linux using CMake with CMakeLists, which I generated with the vcproj2cmake Script (https://github.com/sixman9/vcproj2cmake).
With the CMakeLists.txt and CMake I was able to successfully port a VS 2005 project to VS 2010. Now I try to port the same VS 2005 project to Linux, so that I can edit it with KDevelope.
The project itself is small and I think it would be easier to just create a new project and copy all relevant files (if that would work), the problem is, that it's not only one project but many, hence I was looking for a way to port a project in an as simple as possible way.
In Linux I was able to create a CMakeLists.txt using the vcproj2cmake script. The next step would be creating a KDevelope project using CMake. And this is where I'm stuck.
Everytime I try to run CMake I get the following error:
CMake Error at CMakeLists.txt: 196 (set_property) :
set_property could not find TARGET Test_Project. Perhaps it has not yet been
created.
Test_Project: installing /root/Desktop/Test_Project/vs8/CMakeLists. txt rebuilder (watching /root/Desktop/Test_Project/vs8/Test_Project. vcproj)
Configuring incomplete, errors occurred!
See also “/root/Desktop/Test_Project/vs8/CMakeFiles/CMakeOutput.log”.
I searched for a solution or an approach for quite a time now, but the only results that I get are project-specific, or at least I think they are.
Content of the CMakeLists.txt from line 196:
set_property(TARGET Test_Project PROPERTY PROJECT_LABEL "Test_Project")
v2c_rebuild_on_update(Test_Project "${CMAKE_CURRENT_SOURCE_DIR}/Test_Project.vcproj" ${CMAKE_CURRENT_LIST_FILE} "vcproj2cmake.rb" ".")
include(${V2C_HOOK_POST} OPTIONAL)
Perhaps there are more efficient ways to port many projects form Windows to Linux, I'm open for any suggestion.
It is a bit difficult to help without having your project - since you are using a third party tool to convert a VC project you should ask the author of that tool. :D
If you have many projects which you are going to maintain, I suggest that you select one where you can create a cross platform (Windows/Linux) CMake config for which you can reuse for other projects.
Most of the config should be the same for all platforms you are building for - the difference should be which generator is used and what libraries to link in the final executable (if you are building one that is). The generator is specified when running cmake.
In my opinion you should try an out of source build with the following structure:
/$COMMON_DIR/CMakeLists.txt
/CMakeLists.txt
/src/
/build/
/build/vcX
/build/generate_vcX.bat
/build/linux
/build/generate_linux.sh
$COMMON_DIR should contain the CMake code you expect to be common for all you projects - like a function to create a static library for a module which searches for files in /src and another function which can be used to add an executable for certain platform.
The CMakeLists.txt in the root of your project (repository, I assume) should call the functions defined in $COMMON_DIR and possibly add some project specific defines or link flags.
I got this error using Visual Studio 2008 and I found a solution on the web here and here. But I can't find out how to configure link.exe.
How can I set the /expectedoutputsize:600000000 option for linker.exe in VS2008? I searched in the project properties in the Linker section, but I can't find the place...I searched in the solution and in Visual Studio options. I found the linker.exe.config but I don't know the schema.
My problem is not the disk space, I have plenty of disk space. Any help?
The option must be added in the Project Settings->Librarian->Command Line->Additional options: text box.
Sorry to resurrect this old thread, but I had a similar problem yesterday, and my solution had nothing to do with anything I found online. This is the first SO post that comes up, so I figured I would contribute in case anyone as the same problem.
Here is how I ran into the problem:
I originally had a project that created an exe:
MyProject.vcxproj -> MyProject.exe
I then turned the original project in to a .lib project by splitting main.cpp out to a separate .exe project. I set the target name for the exe project to be the same as the lib, so that we wouldn't change our executable name. I also added a different .exe project that uses the library but has a slightly different main.cpp
MyProject.vcxproj -> MyProject.lib
MyProjectVariant1.vcxproj -> MyProject.exe
MyProjectVariant2.vcxproj -> MyProjectVariant2.exe
The way our solution is currently laid out, all of the projects dump their targets into the same output directory.
The problem was that both the .lib and first .exe share the same target name, so any secondary files (pdb files, iobj, ipdb, etc.) would get overwritten. MyProject.exe would literally overwrite these ancillary files before it could link in the MyProject.lib.
Conclusion:
I "fixed" the problem by using a unique target name for the first variant. We will also review our build strategy to see if we should be using different output directories for each project instead of slamming them all together in the same location. Seems more logical to give them different target directories.
I am pretty experienced writing my own code, but not in C++. I am used to adding .NET assembly references, and I want to get into using C++.
I downloaded a VC++ 6 project presenting a graph algorithm, opened it in Visual Studio 2010 Ultimate and downloaded the required libraries (LEDA .NET assemblies) however, they came with both DLLs and LIB so I thought I would have no problems with the .NET and C++ connection. Also, the LEDA project comes with C++ style .header files, and static libraries, yet frustration persists.
I put my LEDA header files in a directory visible to the project, since I can "Go to definition" of all the custom LEDA data types and see where they are defined in the LEDA include files.
But, I can not compile/link these data types in. The VS finds their very existence repulsive, and refuses to acknowledge they could possibly be found in leda.lib.
Steps I have already taken:
Added additional /LIBPATH path under Properties ->Linker->General
Added explicit lib references under Properties->Linker->Input
Added all the paths containing LEDA header files and libraries to every path location I can find, like include path, reference path, library path.
Copied the DLLs and LIBs to the release folder as well as the base folder of the project.
Added a #pragma comment( lib, "leda" )
Still no joy.
It's quite bad since none of the custom typedefs are recognised the errors balloon to 646 errors.
If anyone is kind enough to offer assistance, thank you.
I can't guarantee this will work but try the following steps (therefore using it's default predefined paths instead of setting your own ones) :
1) Copy *.dll into Windows/System32 or / and (there is some confusion about x64 and x32 default checking in my head :)) Windows/SysWOW64.
2) Copy *.h files into x:\PathToVC2010\include\
3) Copy *.lib files into x:\PathToVC2010\lib\
All was right in the world until I upgraded to Xcode 4 a few days ago. Since then I've had endless problems getting things to work as they should. And I have a crucial update I need to release. I've tried every permutation of settings I can think of, restarted, reinstalled Xcode, reverted to old versions of my files, everything.
My project links to three static libraries, contained in three other projects. I have used the standard processes to link libraries (drag the project files into mine, add their products as target dependencies, add the lib---.a files to the Link Binary With Libraries phase). And actually, I have no problem compiling with the Debug Build Configuration, either for the simulator or my test device.
Where everything goes sideways is when I compile with the Release Build Configuration, or when I try to Archive. I've gotten many different errors depending on my settings, but most are variations on this:
ld: warning: ignoring file
[...]/Build/Products/Debug-iphonesimulator/libGDataTouchStaticLib.a,
file was built for archive which is
not the architecture being linked
(armv6) Undefined symbols for
architecture armv6:
"_OBJC_CLASS_$_GDataSpreadsheetData",
referenced from:
objc-class-ref in ExportViewController.o
I can't understand why it's even looking at Products in the Debug-iphonesimulator directory (I swear, everything I'm linking to reveals itself in the Finder to be in the proper Release-iphoneos directory).
I have put a ridiculous number of hours into fixing this, really need help! Thank you!
Please check this question and answer. I encountered same problem and fixed it.
Xcode4 Linking Problem. File was built for archive which is not the architecture being linked (arm6)
I solved this problem by copying the .a lib files from the Release-iphoneos directory to the Debug-iphonesimulator directory so that the correct files would be found even though Xcode was looking in the wrong directory.
However, then I encountered the problem raised here of a multi-application bundle -- and the solution given didn't work for me. Finally gave up and reinstalled Xcode 3. Compiled, archived, and uploaded to the App Store in 20 minutes. Will be some time, and several dot-releases, before I give Xcode 4 another shot.
You may have -DGDATA_REQUIRE_SERVICE_INCLUDES=1 enabled in your other C flags for the GDataTouchStaticLib target. If so, add the service you need, in this case, spreadsheets, by adding -DGDATA_INCLUDE_SPREADSHEET_SERVICE=1 to your Other C Flags. Or if you don't want GData to require service includes (which will build everything into the static lib, not just what you need), just remove the DGDATA_REQUIRE_SERVICE_INCLUDES flag.