FindProtobuf.cmake for Google Protocol Buffers creates a set of targets, for example protobuf::libprotobuf-lite one.
Now, in my project, I'm using this protobuf::libprotobuf-lite in target_link_libraries. And this works fine.
However, after upgrading to protobuf v3 I noticed that generated files have small blocks guarded by
#if LANG_CXX11
Yet, as it seems, this LANG_CXX11 is not defined by FindProtobuf.cmake. It seems I have to define it on my own if I want to use C++11 features with protobuf-generated code.
Obviously, I could just do it globally and be done with it. However, I would much more prefer to somehow "bind" the LANG_CXX11 with the protobuf::libprotobuf-lite target, so that I will get it whenever I used it and only where I use it.
target_compile_definitions will not work here since it doesn't accept imported targets while targets created by FindProtobuf.cmake are imported ones. And anyway, it seems like not the right thing to do.
The only idea I have for now (although not yet tested) is to create another custom target that would only link (PUBLIC) with protobuf::libprotobuf-lite and add the LANG_CXX11 define (PUBLIC as well). Then I would use this custom target instead of protobuf::libprotobuf-lite. But it doesn't feel good either.
How to do it properly with CMake? (I'm on version 3.8 if it matters.)
Related
I have big project with cmake. It mostly works.
But recently some combination of compilation server vs test server broke. Investigation found that final compile/link command calls gcc (...) -licudata -licui18n -licuuc (...), this introduces dependency on shared library which is not present on test server.
How do I find out what in my project (my library, imported library, found library, whatever) adds those 3 flags to compile command?
I don't add them explicitly, so something is done automagically and I want to find it. compile_commands.json doesn't have them because linking flags don't belong in it. CMakeCache.txt has those flags in some obscure variable PC_LIBXML_STATIC_LIBRARIES:INTERNAL but removing them there doesn't affect compile/link command.
Note that this question is not about dealing with libicu specifically but about a method for investigation in general (though comments about eventual known problems with libicu would be appreciated too).
I found out that dependency graphs created by cmake can have more details that was configured for our project. Here are all options: https://cmake.org/cmake/help/latest/module/CMakeGraphVizOptions.html I expect GRAPHVIZ_EXTERNAL_LIBS, GRAPHVIZ_SHARED_LIBS are most important to set to true.
We enabled everything that was possible to enable, filtered out nothing and resulting graph was massive (to big for xdot - luckily .dot files are human readable), but showed that Boost::regex uses those 3 libraries.
As far as I understand using bazel I can only produce libtensorflow_cc.so and libtensorflow_framework.so.
I need to produce static libs that are position independent (-fPIC) because I'll link them to a dynamic lib of my own later.
I found this answer which suggest the use of a Makefile included in the project.
I successfully used it to replace the libtensorflow_cc.so but what can I do to replace libtensorflow_framework.so?
Not an actual answer, but too long for a comment.
I managed to do something like what you mention using Bazel on Windows. In particular, I wanted to make a single wrapper DLL with one or two headers (limited in functionality) that I could move around easily. I'll write a summary of the things that I did; it's rather convoluted an customized for our needs, but maybe you find something useful.
I pass --config=monolithic to the bazel build command (besides any other option that you need). That will avoid modularizing the library and thus remove the dependency to a libtensorflow_framework.so (see
tools/bazel.rc).
The goal that I build is not any of the ones in the TensorFlow repository. Instead, I add a very small program that uses my wrapper as a new Bazel target (a C++ file plus my headers headers and a BUILD file). So all of TensorFlow had to be compiled beforehand in order to compile this final dummy program.
When I get that done, I take advantage of the fact that Bazel does already compile every subgoal as a static library. I check a file under the bazel-bin directory generated for my dummy program goal with a name ending .params - there I find the path of all the static libraries that were used to compile it.
I copy all of these intermediate static libraries to somewhere else. Also, I copy a bunch of headers I will need to compile my final wrapper (TensorFlow own's, but also Eigen, Protobuf and Nsync now too). I put all of this in a build area I have prepared before.
I use NMake Makefile to produce my custom DLL, using the static libraries, the copied headers and my own thin wrapper.
And that's about it, I think. I have an ugly Bash script I run on MSYS2 that does everything for me. Usually with every new release I need to tweak one or two things (some option in the configure script, some additional headers I need to copy, etc.), but I do get it to work in the end. It's quite a lot of fiddling though, so I'm not necessarily saying you should use the same approach (but feel free to ask for details about any step if you want).
Using the -2.params files #jdehesa mentioned and bazel verbose output (-s switch), you can even create a link command to eventually statically link these intermediate static libraries. I automated this process for Windows/Linux/macOS and included it to the vcpkg package manager. To use it just run vcpkg install tensorflow:x64-windows-static. If you're interested in the sources, you'll find them here.
I have this project that I have done for experimentation with Qt and shared libraries. This is basically a couple of Qt Widgets from the tutorials for Qt and what I think is the right CMakeLists configuration so a MylibConfig.cmake is automatically generated from a MylibConfig.cmake.in to share the library. The problem is that I don't want the end user to add the dependencies of my library to its own CMakeLists.txt. This is, in my case, the library depends on Qt4, but I want that the end user to not have to do find_package(Qt 4 REQUIRED). Imagine that I want to provide an enclosed functionality to someone that does not need or want to know about what my library is built on. Is there a way in the automatic generation of the MylibConfig.cmake that it automatically finds all necessary packages or is the only option to add the fin package manually in the MylibConfig.cmake.in?
Thank you very much.
In fact, both mentioned projects do find of dependencies from their *Config.cmake files. And nowadays that is the only option -- CMake can't help you to do it "automatically".
So, some way or another, your config module should do the same.
The easy way is to add find_dependency() calls (cuz you know exactly what other packages, your project is based on).
A little bit harder is to do it "automatically" (writing your own helper function) -- for example by inspecting properties of your target(s), "searching" where all that libraries come from and finally generating find_dependency() calls anyway.
The CMake manual of Qt 5 uses find_package and says:
Imported targets are created for each Qt module. Imported target names should be preferred instead of using a variable like Qt5<Module>_LIBRARIES in CMake commands such as target_link_libraries.
Is it special for Qt or does find_package generate imported targets for all libraries? The documentation of find_package in CMake 3.0 says:
When the package is found package-specific information is provided through variables and Imported Targets documented by the package itself.
And the manual for cmake-packages says:
The result of using find_package is either a set of IMPORTED targets, or a set of variables corresponding to build-relevant information.
But I did not see another FindXXX.cmake-script where the documentation says that a imported target is created.
find_package is a two-headed beast these days:
CMake provides direct support for two forms of packages, Config-file Packages
and Find-module Packages
Source
Now, what does that actually mean?
Find-module packages are the ones you are probably most familiar with. They execute a script of CMake code (such as this one) that does a bunch of calls to functions like find_library and find_path to figure out where to locate a library.
The big advantage of this approach is that it is extremely generic. As long as there is something on the filesystem, we can find it. The big downside is that it often provides little more information than the physical location of that something. That is, the result of a find-module operation is typically just a bunch of filesystem paths. This means that modelling stuff like transitive dependencies or multiple build configurations is rather difficult.
This becomes especially painful if the thing you are trying to find has itself been built with CMake. In that case, you already have a bunch of stuff modeled in your build scripts, which you now need to painstakingly reconstruct for the find script, so that it becomes available to downstream projects.
This is where config-file packages shine. Unlike find-modules, the result of running the script is not just a bunch of paths, but it instead creates fully functional CMake targets. To the dependent project it looks like the dependencies have been built as part of that same project.
This allows to transport much more information in a very convenient way. The obvious downside is that config-file scripts are much more complex than find-scripts. Hence you do not want to write them yourself, but have CMake generate them for you. Or rather have the dependency provide a config-file as part of its deployment which you can then simply load with a find_package call. And that is exactly what Qt5 does.
This also means, if your own project is a library, consider generating a config file as part of the build process. It's not the most straightforward feature of CMake, but the results are pretty powerful.
Here is a quick comparison of how the two approaches typically look like in CMake code:
Find-module style
find_package(foo)
target_link_libraries(bar ${FOO_LIBRARIES})
target_include_directories(bar ${FOO_INCLUDE_DIR})
# [...] potentially lots of other stuff that has to be set manually
Config-file style
find_package(foo)
target_link_libraries(bar foo)
# magic!
tl;dr: Always prefer config-file packages if the dependency provides them. If not, use a find-script instead.
Actually there is no "magic" with results of find_package: this command just searches appropriate FindXXX.cmake script and executes it.
If Find script sets XXX_LIBRARY variable, then caller can use this variable.
If Find script creates imported targets, then caller can use these targets.
If Find script neither sets XXX_LIBRARY variable nor creates imported targets ... well, then usage of the script is somehow different.
Documentation for find_package describes usual usage of Find scripts. But in any case you need to consult documentation about concrete script (this documentation is normally contained in the script itself).
I have created a static library containing all my generic classes. Some of these classes use frameworks.
Now I have two projects, one that uses some classes that use frameworks, and one that doesn't use any of the classes that use frameworks.
Because Static Libraries don't support including frameworks (if I am correct). I have to include the frameworks in the project that uses them. But when I compile the project that doesn't use any of the framework-classes the compiler breaks because it still requires the frameworks. Now I know it tries to compile all the (unused) classes from the library because I use the Linker Flag '-ObjC' to prevent 'unrecognized selector' errors.
Does anyone know how to compile only the required source files per project? And prevent from all frameworks having to be included in all projects that use my static library?
First of all, you are right in that a static library cannot include any framework nor other static libraries, it is just the collection of all object files (*.obj) that make up that specific static library.
Does anyone know how to compile only the required source files per project?
The linker will by default only link in object files from the static library that contain symbols referenced by the application. So, if you have two files a.m and b.m in your static library and you only use symbols from a.m in your main program, then b.o (the object file generated from b.c) will not appear in your final executable. As a sub-case, if b.m uses a function/class c which is only declared (not implemented), then you will not get any linker errors. As soon as you include some symbols from b.m in your program, b.o will also be linked and you will get linker errors due to the missing implementation of c.
If you want this kind of selection to happen at symbol rather than at object level granularity, enable dead code stripping in Xcode. This corresponds to the gcc option -Wl,-dead_strip (= linker option -dead_strip in the Build settings Info pane for your project). This would ensure further optimization.
In your case, though, as you correctly say, it is the use of the "-ObjC" linker flag that defeats this mechanism. So this actually depends on you. If you remove the -Objc flag, you get the behavior you like for free, while losing the stricter check on selectors.
And prevent from all frameworks having to be included in all projects that use my static library?
Xcode/GCC support an linking option which is called "weak linking", which allows to lazily load a framework or static library, i.e., only when one of its symbols is actually used.
"weak linking" can be enabled either through a linker flag (see Apple doc above), or through Xcode UI (Target -> Info -> General -> Linked Libraries).
Anyhow, the framework or library must be available in all cases at compile/link time: the "weak" option only affects the moment when the framework is first loaded at runtime. Thus, I don't think this is useful for you, since you would need anyway to include the framework in all of your projects, which is what you do not want.
As a side note, weak_linking is an option that mostly make sense when using features only available on newer SDK version (say, 4.3.2) while also supporting deployment on older SDK versions (say, 3.1.3). In this case, you rely on the fact that the newer SDK frameworks will be actually available on the newer deployment devices, and you conditionally compile in the features requiring it, so that on older devices they will not be required (and will not produce thus the attempt at loading the newer version of the framework and the crash).
To make things worse, GCC does not support a feature known as "auto-linking" with Microsoft compilers, which allow to specify which library to link by means of a #pragma comment in your source file. This could offer a workaround, but is not there.
So, I am really sorry to have to say that you should use a different approach that could equally satisfy your needs:
remove the -ObjC flag;
split your static library in two or more parts according to their dependencies from external frameworks;
resort to including the source files directly.
Abour second part of your question, you can mark a linked framework as Optional :
About first part, it is not clear to me what you intend to do:
A library being declared in a project
A project declaring which files are compiled (via Target > Build phases > Compile sources)
Unless setting complex build rules to include or not files, which if I remember well can be done using .xcconfig files, I don't see any other solutions than splitting your Library. Which I would recommend, for its ease. You should even do several targets in the same project... You could also just use precompiler MACROS (#ifdef...) but that depends on what you want to do.
It sounds like you have library bloat. To keep things small I think you need to refactor your library into separate libraries with minimal dependencies. You could try turning on "Dead Code Stripping" in the "Linker Flags" section of the build target info (Xcode 3.x) to see if that does what you want (doesn't require frameworks used by classes that are dead-stripped.)
When you link against a framework on iOS I don't think that really adds any bloat since the framework is on the device and not in your application. But your library is still a bit bloated by having entire classes that never get used but are not stripped out of the library.
A static library is built before your app is compiled, and then the whole thing is linked into your app. There's no way to include some parts of the library but not others -- you get the whole enchilada.
Since you have the source code for the library, why not just add the code directly to each application? That way you can control exactly what goes into each app. You can still keep your generic classes together in the same location, and use the same code in both apps, but you avoid the hassle of using a library.