This should be simple, but unfortunately it has given me hours of headache. I have a third party library with multiple DLL files. I want these to be included in the final bazel bin so that I can run my project.
Copying them manually works, but obviously this isn't ideal.
Bazel has docs which use CC_import rule, which seemed to be what I want, except that I need to use linkstatic=0
Using this it tries to build all my packages dynamically into shared object files. This isn't really what I want either , especially since it seems MinGw is failing to link against these.
I'm out of things I can try so any help would be infinitely appreciated. Thanks :)
Hopefully I can save time for someone with this in future, since I finally worked out what my issue was. I think this is an issue with bazel, and so I have raised an issue on their github page.
I was trying to use the cc_import rule but hadn't specified a interface_library. This was having me run into all kinds of issues where it was trying to statically include the DLL. The interface_library is required.
My next issue was I coudln't specfiy a .a file for the interface library, which I believe GCC uses for linking against DLLs. This type isn't supported by bazel. I've raised an issue for this (as above).
My workaround currently is to manually copy the DLLs into the bazel-bin folder. Everything works perfectly if I use a supported file type for the interface_library.
Hope this saves someone time in the future :)
Related
My current task is to set up a working environment for an application using (static) Qt and VTK, under Windows 10.
I successfully installed the Qt 5.11.1 static and now want to build VTK. Any version of VTK does, as long as it is compatible to Qt 5.11.1, and the Qt Modules (VTK_Group_Qt) have to be build. For my version, I decided on VTK 8.1.1, but as said, any version would do.
I used the CMake GUI to configure VTK, using MinGW Makefiles (I use a MinGW64), which worked. Then I wanted to proceed generating the Makefile, having turned on VTK_Group_Qt. The paths and version of Qt were already correctly recognized:
Qt5Core_DIR C:/build_environment/qt-everywhere-src-5.11.1/qtbase/lib/cmake/Qt5Core
Qt5Gui_DIR C:/build_environment/qt-everywhere-src-5.11.1/qtbase/lib/cmake/Qt5Gui
Qt5Widgets_DIR C:/build_environment/qt-everywhere-src-5.11.1/qtbase/lib/cmake/Qt5Widgets
Qt5_DIR C:/build_environment/qt-everywhere-src-5.11.1/qtbase/lib/cmake/Qt5
Made sure that all of those exist. After that, I hit Generate and got the following error:
The imported target "Qt5::Core" references the file
"C:/build_environment/qt-everywhere-src-5.11.1/qtbase/.../qt-everywhere-src-5.11.1/qtbase//mkspecs/win32-g++"
but this file does not exist. Possible reasons include:
* The file was deleted, renamed, or moved to another location.
* An install or uninstall procedure did not complete successfully.
* The installation package was faulty and contained
"C:/build_environment/qt-everywhere-src-5.11.1/qtbase/lib/cmake/Qt5Core/qt5CoreConfigExtras.cmake"
but not all the files it references
Now, we see that it goes one folder upwards from qtbase and then into the very folder it currently is in, which of course is not a legitimate path. That is, for some reason it tries to go up and then down again, but not with the right amount of folders.
I could try to move the contents of my Qt installation so that the path would be legitimate, but I'd consider that to be pretty much of a smell. Fiddling around in a folder structure that should be correct is not what I understand to be a good solution. I'd rather be willed to change one of the .cmake files of Qt, but that would also be more of a makeshift approach than a real solution as far as I am concerned.
One thing I tried so far was to look at all the CMake variables (checking Advanced in the CMake GUI) in order to see if everything appear to be in order, and especially to see if that faulty path to mkspecs is to be found somewhere there. Found nothing of particular interest.
Does anybody have an idea what went wrong? Tell me if additional information is required. In parallel, I will try to build other versions of VTK, but I'm not too confident that this will solve the issue.
I found a solution which I want to present here, but I consider it a makeshift approach rather than an actual solution, and I hope to see a better answer.
I went to qt-everywhere-src-5.11.1/qtbase/lib/cmake/Qt5Core/ and changed the file Qt5CoreConfigExtrasMkspecDir.cmake (after making a copy).
The original file contains a single line
set(_qt5_corelib_extra_includes "${_qt5Core_install_prefix}/.../qt-everywhere-src-5.11.1/qtbase//mkspecs/win32-g++")
which leads to the faulty path. I changed it to
set(_qt5_corelib_extra_includes "${_qt5Core_install_prefix}/mkspecs/win32-g++")
And it now works fine. Still, I don't feel good about having to change a file of Qt, since I consider the answer the question "Is this famous library at fault?" to be "Most likely I used it wrong."
A possible explanation would be that I used the wrong install prefix when configuring Qt, but shouldn't I be allowed to put any valid writeable path in there? Plus, the prefix is still in the cmake file and thus has some validity.
While this is an answer since it fixed the problem, I'd really like to read comments about what could have originally went wrong.
I am new to Unreal Engine. But here is what I planned to do. I want to an experiment project by combining socket.io with Unreal Engine 4. I know somebody may shot at me, for it already had a plugin. But I don't like graph programming at all (I prefer "real" coding)
But as I followed instruction from Installing socket.io C++ and here. It just won't work. The error told me something about File Not Found on multiple headers file from this github repo. So I tried to add a lot of things (Hopefully it won't affect the outcome). And now a tons of errors had popped up. Now they are mostly about Macro errors. And I have absolutely no idea how to fix.
The problem is I don't understand why after I include .lib files to PublicAdditionalLibraries I still needed the original file (Shouldn't it work like any other visual studio projects?)
Also how to tackle loads of errors I am facing
Thanking in advanced
Did you add your include paths with the lib headers to PublicIncludePaths array ( in the StartupModule() method ? it's gonna be needed to use methods from your lib.
Here is an exemple how to use the PublicIncludePaths
string includePath = Path.Combine(ThirdPartyPath, "opencv", "include");
PublicIncludePaths.Add(includePath);
ThirdPartyPath is a member of my module class i retrieve with my getter
Path.GetFullPath(Path.Combine(ModulePath, "../../ThirdParty/"));
I took a look at the github repo of Socket-IO-cpp lib and the only libs compiled for windows are in 32bits, take care you're not tried to compile in 64bits, or every .lib files of your libraries are in 32bits to be linked all together, in case you're using others libs.
If you got some runtime error, take a look at my answer which explain how to indicate .dll when .libs arn't enough.
Hope it helps
I'm lost here and I have no clue how to proceed. This is not a question about how to make my program work, this is a question about how to stop wasting my time.
My programming environment is Visual Studio 2013 on windows, in C++.
I use 3 libraries extensively, namely: boost (using dynamic linking), OpenCV, and Qt.
During the development, I have configured VS to look at those 3 libraries by default for include and .lib. I have also added the 3 folders containing all the dlls to my PATH environment variable.
It works, but it is sometime painful, let me explain you when.
First hassle: Anytime I have a LNK error telling me I miss a function, it is usually on OpenCV since it has only one include file referencing all the functions. I have to look at OpenCV's source code to see what module this function belongs to and to know what I must link my program to.
Second Hassle: When comes the time to deploy my application, I have to ship it with all the relevant dlls. To know which one I need, I open dependency walker and try to forget nothing, I have then to test it on a different computer because 102% of the time I have missed a couple, and then I have to configure my Installer generator to include all those one by one.
Third Hassle: To ease a little bit the process of configuring a new development machine, I have recently switched to NuGet. It is great, I add boost with a couple of clicks to any project. But now my boost DLLs are everywhere, I have one folder per boost library, and since there are dozens of those I can't even add them all at once to my PATH now, so I have to move them manually to the appropriate folder, and that is really not what I want to do with my not-so-precious-but-who-are-you-to-judge time
I have looked around and couldn't find any good practice regarding this issue, maybe because they are too obvious, or too specific to a particular setup.
How do you do? How would you do if you were me?
We put all our external dependencies in version control along with the code. This ensures that all code can build "out of the box" on any of our development machines and also ensures that for any given version of the code, we know exactly which dependencies is has.
The best way to check for missing dependencies is how have a good automated test suite, if you've got comprehensive converge then if your tests pass you must have deployed the required libraries.
In terms of linking to the appropriate libraries, unfortunately, that just sounds like an issue with the structure of OpenCV (I'm not familiar with OpenCV). I tend to use dumpbin under Windows and nm under Linux to easily grep for symbols when I get link errors with an unfamiliar library.
I have written and tested a library in C++. The code even works in my Android application if I add the source files directly. While I do have experience compiling static and dynamic libraries for common operating systems, I have zero experience compiling for a mobile system like this. I've done some research, and I'm still a bit lost as to exactly how to approach this. For example, I am unsure of whether to build a makefile for use with ndk-build or to just invoke one of the Android's compilers directly.
I did see this question, but it does not quite match my situation. I just want to run build and have it spit out libfoo.a (I'd like to produce libfoo.so as well, but libfoo.a is of greater interest to me right now.) The example in that question's winning answer implied that it would build the library as one step/module for building the final application. I tried doing it that way just to see, but I had no luck.
Can anyone please guide me in this endeavor?
CLARIFICATION -- I do not want to build the library and immediately pipe it into an application. I want a .a or .so file that I can link against in multiple future Android applications.
Create a dummy java file with empty code and make sure there is a android_main function in your C++ code. Build using ndk-build. the resulting apk will make your library an application.
See the samples from the android-ndk-r5/samples directory, see the sample native-bitmap to get some idea.
If I understand it correctly that a shared library is not acceptable and you want to be static (but why is that so important?), probably the easiest way to do so is to simply supply source code that can be added to a project.
Ultimately there is nothing special about the ndk build system other than it knowing the right commands to issue to build the necessary files for the assortment of curent android architectures. If you really want to do something different, you can log the operation of the android build system in creating a shared library, and then write your own Makefile that performs the analogous steps to create a static library. HOWEVER, you will have to update your Makefile any time the underlying assumptions or target collections change in a new android version.
I'm teaching myself OpenGL and I'm implementing ttf text rendering using FreeType 2. I downloaded the library from
http://gnuwin32.sourceforge.net/packages/freetype.htm
and after a couple of minor issues I got it running properly. The thing that's bothering me is that I have to place a copy of freetype6.dll in the directory with my executable in order for the thing to run. I generally try to avoid a bunch of unnecessary dll files floating around. I'm sort of new to windows programming, but from what I understand most libraries can be built to run fully from a lib rather than requiring a dll at runtime. Looking through the documentation from FT is making my brain melt, so I thought I would ask here to see if there were any devs that have worked with FT before and if so, do they know how to build the library such that no dll is required at runtime.
Thank you in advance for any advice or support.
Check out this link. See the section Optional: Installing FreeType (by compiling it yourself)
Follow the instructions and you'll be good to go.
you can generate static lib by getting source code of it... then you won't need dll...and i think freetype2's source is available...
It is something DLL generic. All your DLLs should be in PATH similarly to LD_LIBRARY_PATH. Also under Windows (unlike Unix) the current directory is always in the PATH. So you just need to set your PATH variable to point to location of this dll.
Now, for Unix... you probably just have this library installed by default like hundreds of other useful libraries that are not present under Windows by default.
So... No unless you link statically you should use DLL somehow. And my suggestion - use dll.