Question: Is it possible to join multiple object libraries into a single larger object library?
Consider the following simple project. libA.cpp is compiled into an object library LibA.Obj, and libB.cpp is compiled into an object library LibB.Obj. I want to have an object library, LibAB.Obj, that consists of both LibA.Obj and LibB.Obj, and I want to link against it to produce a single shared library.
MyProject
├── CMakeLists.txt
└── src
├── libA.cpp
└── libB.cpp
If I want to link two libraries, A and B, into a single unit, I can do that by declaring them as object libraries:
cmake_minimum_required(VERSION 3.21)
project(ObjectLibExample)
add_library(LibA.Obj OBJECT src/libA.cpp)
add_library(LibB.Obj OBJECT src/libB.cpp)
add_library(LibAB SHARED $<TARGET_OBJECTS:LibA.Obj> $<TARGET_OBJECTS:LibB.Obj>)
This works as expected, and everything builds.
For reasons I will soon explain, I would like to merge multiple object libraries together into a single unit:
cmake_minimum_required(VERSION 3.21)
project(ObjectLibExample)
add_library(LibA.Obj OBJECT src/libA.cpp)
add_library(LibB.Obj OBJECT src/libB.cpp)
# Using LibAB.Obj as a source should be equivilant to using both LibA.Obj and LibB.Obj
add_library(LibAB.Obj OBJECT $<TARGET_OBJECTS:LibA.Obj> $<TARGET_OBJECTS:LibB.Obj>)
add_library(LibAB SHARED $<TARGET_OBJECTS:LibAB.Obj>)
Here, LibAB.Obj should be an object library collecting the sources of both LibA.Obj and LibB.Obj.
For some reason, CMake acts as though LibAB.Obj provides no sources, and the build fails.
Why does this happen?
Is this a bug?
What workarounds are there?
Motivation / Background
I am currently working on re-organizing an existing project to have a modular structure. It will have modules A, B, and C, each of which will have submodules.
You can imagine the directory structure looking like this:
MyProject
├── A
│ ├── A1
│ ├── A2
│ └── A3
├── B
│ ├── B1
│ ├── B2
│ └── B3
└── C
├── C1
├── C2
└── C3
Each submodule represents a library that forms a part of a greater whole, and the build system should link everything together into a single library at the end.
The reason for this structure is twofold:
each submodule forms a cohesive unit. It has it's own tests, benchmarks, and example code
submodules should provide a clean API for their functionality, even if the internals are kind of hairy.
For example, MyProject/A/A1 looks like this:
MyProject/A/A1
├── bench
├── CMakeLists.txt
├── examples
├── src
└── test
My initial plan was to have each module contain an object library linking all of it's respective submodules, with all modules being statically linked together at root to produce a single library, MyProject.so.
Each submodule would also produce a library specific to that submodule (eg, MyProject/A/A1 would produce MyProject.A.A1.so), against which tests, benchmark code, and examples would be linked. This would allow faster iteration, without needing to re-build the entire project.
Unfortunately, this initial approach failed, for the reasons discussed above.
According to the documentation of add_library(OBJECT):
An object library compiles source files but does not archive or link their object files into a library.
[...]
Object libraries may contain only sources that compile, header files, and other files that would not affect linking of a normal library (e.g. .txt).
You haven't provided any sources that would compile, so a fatal error indicating that the object library isn't used in the intended way is expected.
A workaround possible starting cmake 3.12 would be to create an INTERFACE library and linking the object libraries with INTERFACE visibility.
add_library(LibAB INTERFACE)
target_link_libraries(LibAB.Obj INTERFACE LibA.Obj LibB.Obj)
add_library(LibAB SHARED)
target_link_libraries(LibAB PRIVATE LibAB.Obj)
This does prevent you from using $<TARGET_OBJECTS:LibAB.Obj> though.
I made a Non-Qt project, Plain C++ Application in QT creator. My current folder structure is like this:
.
├── client
│ ├── client.cpp
│ └── client.h
├── CMakeLists.txt
├── CMakeLists.txt.user
├── main.cpp
└── server
├── server.cpp
└── server.h
My CMakeLists.txt file looks like this:
cmake_minimum_required(VERSION 3.5)
project(eshraagh-project LANGUAGES CXX)
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
add_executable(eshraagh-project main.cpp server/server.h server/server.cpp client/client.h client/client.cpp)
In my main file, I included server.h and client.h and made instances of both of them, and it didn't give any errors.
The problem arises when I try to use an external library (this library). I have no idea how to use it.
I tried to copy contents of its src folder (cpnet-export.h, cpnet-network.h and cpnet-network.c) into my project and add them to CMakeLists, just like I did with server and client files, but it doesn't work. I copied the server part of the repository and pasted it in my main, but it gives me errors such as undefined reference to 'cpnet_bind' and more (A little note: The repository uses the functions with net prefix, but QT recommender told me to write cpnet prefix instead. I don't know why this is happening).
Linking a library of an executable in CMake is achieved by using the target_link_libraries function. In your case,
target_link_libraries(eshraagh-project PRIVATE cpnet)
Here, cpnet could be the name of a target that you CMake project knows (otherwise, it is turned into some kind of platform specific -lcpnet linker flag). In order to make CMake know about the cpnet library, you can
add the external library sources to your project (as a git submodule, or by tracking them yourself), for this example let's say they the sources are located at lib/cpnet
tell CMake to include the configuration instructions from that folder. This works if the external project uses CMake, too, and doesn't mess with global variables and settings (this seems to be the case here).
So in your toplevel CMakeLists.txt:
add_subdirectory(lib/cpnet)
Note that this must be done before the target_link_libraries call.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 3 years ago.
Improve this question
Currently I have a directory structure that looks like the following
.
├── CMakeLists.txt
├── Base
│ ├── CMakeLists.txt
│ ├── include
│ │ └── base.h
│ └── src
│ └── base.cpp
├── Derived
│ ├── CMakeLists.txt
│ ├── include
│ │ └── derived.h
│ └── src
│ └── derived.cpp
└── src
└── main.cpp
and the CMakeLists.txt files looking like
./CMakeLists.txt
cmake_minimum_required(VERSION 3.1)
set(CMAKE_CXX_STANDARD 11)
project(MyProj)
add_subdirectory(Base)
add_subdirectory(Derived)
add_executable(main src/main.cpp)
target_link_libraries(main Base)
target_link_libraries(main Derived)
./Base/CMakeLists.txt
add_library(Base STATIC src/base.cpp)
target_include_directories(Base PUBLIC include)
./Derived/CMakeLists.txt
add_library(Derived STATIC src/derived.cpp)
target_include_directories(Derived PUBLIC include)
target_link_libraries(Derived Base)
I want to know if this is an appropriate way to structure a CMake project when using inheritance in C++. If there is a more idiomatic way to structure this, I am open to suggestions.
If your intention is to build two libraries, and an executable, then that is what your structure achieves. I would however recommend thinking why you are doing that, and perhaps consider whether it would be simpler to instead have both classes in a single library.
If I keep extending this, say make another class that inherits from Derived, then I have to make a new CMakeLists.txt file then link and include against these files. I don't have a problem doing this, but it does not seem very maintainable for larger projects
No, you don't have to make a new CMakeLists.txt unless you add a new target. You don't necessarily need more than a single target for a project.
A target (be it executable or a library) is not limited to having only one translation unit.
My personal preference for small project is three (or two) targets: A library containing all functionality, a main executable with int main() that does nothing except invoke the library (if the project is an executable rather than just the library), and a test executable with int main() from a test framework. The library can optionally be split to smaller pieces if it has parts that are reusable in other projects.
set(CMAKE_CXX_STANDARD 11)
prefer target specific properties instead:
set_target_properties(
Base PROPERTIES
CXX_STANDARD 11
# use unless you also support older standards
CXX_STANDARD_REQUIRED ON
)
The dependees will inherit the property, but you can set it explicitly in each for potentially lesser chance of breakage in case you later decide to not use the base library.
Creating a library for each class is overkill and might slow down compilation on some build systems.
On some of my projects, I have over 300 types, not counting templates and lambdas. I cannot imagine creating a library for each classes.
I want to know if this is an appropriate way to structure a CMake project when using inheritance in C++
The features you are using should not change how the code is organised physically. Instead, base your file layout in logical self contained part of your code with clear dependencies between each parts.
The only nitpick: Use this form to link libraries together:
target_link_libraries(main PUBLIC Base) # or private
The rest of your CMake make good use of the target based API. The only thing I would change would be to be less fine grained in term of what constitute a library.
I am new to programming in general so I decided that I would start by making a simple vector class in C++. However I would like to get in to good habits from the start rather than trying to modify my workflow later on.
I currently have only two files vector3.hpp and vector3.cpp. This project will slowly start to grow (making it much more of a general linear algebra library) as I become more familiar with everything, so I would like to adopt a "standard" project layout to make life easier later on. So after looking around I have found two ways to go about organizing hpp and cpp files, the first being:
project
└── src
├── vector3.hpp
└── vector3.cpp
and the second being:
project
├── inc
│ └── project
│ └── vector3.hpp
└── src
└── vector3.cpp
Which would you recommend and why?
Secondly I would like to use the Google C++ Testing Framework for unit testing my code as it seems fairly easy to use. Do you suggest bundling this with my code, for example in a inc/gtest or contrib/gtest folder? If bundled, do you suggest using the fuse_gtest_files.py script to reduce the number or files, or leaving it as is? If not bundled how is this dependency handled?
When it comes to writing tests, how are these generally organized? I was thinking to have one cpp file for each class (test_vector3.cpp for example) but all compiled in to one binary so that they can all be run together easily?
Since the gtest library is generally build using cmake and make, I was thinking that it would make sense for my project to also be built like this? If I decided to use the following project layout:
├── CMakeLists.txt
├── contrib
│ └── gtest
│ ├── gtest-all.cc
│ └── gtest.h
├── docs
│ └── Doxyfile
├── inc
│ └── project
│ └── vector3.cpp
├── src
│ └── vector3.cpp
└── test
└── test_vector3.cpp
How would the CMakeLists.txt have to look so that it can either build just the library or the library and the tests? Also I have seen quite a few projects that have a build and a bin directory. Does the build happen in the build directory and then the binaries moved out in to the bin directory? Would the binaries for the tests and the library live in the same place? Or would it make more sense to structure it as follows:
test
├── bin
├── build
└── src
└── test_vector3.cpp
I would also like to use doxygen to document my code. Is it possible to get this to automatically run with cmake and make?
Sorry for so many questions, but I have not found a book on C++ that satisfactorily answers these type of questions.
C++ build systems are a bit of a black art and the older the project
the more weird stuff you can find so it is not surprising that a lot
of questions come up. I'll try to walk through the questions one by one and mention some general things regarding building C++ libraries.
Separating headers and cpp files in directories. This is only
essential if you are building a component that is supposed to be used
as a library as opposed to an actual application. Your headers are the
basis for users to interact with what you offer and must be
installed. This means they have to be in a subdirectory (no-one wants
lots of headers ending up in top-level /usr/include/) and your
headers must be able to include themselves with such a setup.
└── prj
├── include
│ └── prj
│ ├── header2.h
│ └── header.h
└── src
└── x.cpp
works well, because include paths work out and you can use easy
globbing for install targets.
Bundling dependencies: I think this largely depends on the ability of
the build system to locate and configure dependencies and how
dependent your code on a single version is. It also depends on how
able your users are and how easy is the dependency to install on their
platform. CMake comes with a find_package script for Google
Test. This makes things a lot easier. I would go with bundling only
when necessary and avoid it otherwise.
How to build: Avoid in-source builds. CMake makes out of source-builds
easy and it makes life a lot easier.
I suppose you also want to use CTest to run tests for your system (it
also comes with build-in support for GTest). An important decision for
directory layout and test organization will be: Do you end up with
subprojects? If so, you need some more work when setting up CMakeLists
and should split your subprojects into subdirectories, each with its
own include and src files. Maybe even their own doxygen runs and
outputs (combining multiple doxygen projects is possible, but not easy
or pretty).
You will end up with something like this:
└── prj
├── CMakeLists.txt <-- (1)
├── include
│ └── prj
│ ├── header2.hpp
│ └── header.hpp
├── src
│ ├── CMakeLists.txt <-- (2)
│ └── x.cpp
└── test
├── CMakeLists.txt <-- (3)
├── data
│ └── testdata.yyy
└── testcase.cpp
where
(1) configures dependencies, platform specifics and output paths
(2) configures the library you are going to build
(3) configures the test executables and test-cases
In case you have sub-components I would suggest adding another hierarchy and use the tree above for each sub-project. Then things get tricky, because you need to decide if sub-components search and configure their dependencies or if you do that in the top-level. This should be decided on a case-by-case basis.
Doxygen: After you managed to go through the configuration dance of
doxygen, it is trivial to use CMake add_custom_command to add a
doc target.
This is how my projects end up and I have seen some very similar projects, but of course this is no cure all.
Addendum At some point you will want to generate a config.hpp
file that contains a version define and maybe a define to some version
control identifier (a Git hash or SVN revision number). CMake has
modules to automate finding that information and to generate
files. You can use CMake's configure_file to replace variables in a
template file with variables defined inside the CMakeLists.txt.
If you are building libraries you will also need an export define to
get the difference between compilers right, e.g. __declspec on MSVC
and visibility attributes on GCC/clang.
As a starter, there are some conventional names for directories that you cannot ignore, these are based on the long tradition with the Unix file system. These are:
trunk
├── bin : for all executables (applications)
├── lib : for all other binaries (static and shared libraries (.so or .dll))
├── include : for all header files
├── src : for source files
└── doc : for documentation
It is probably a good idea to stick to this basic layout, at least at the top-level.
About splitting the header files and source files (cpp), both schemes are fairly common. However, I tend to prefer keeping them together, it is just more practical on day-to-day tasks to have the files together. Also, when all the code is under one top-level folder, i.e., the trunk/src/ folder, you can notice that all the other folders (bin, lib, include, doc, and maybe some test folder) at the top level, in addition to the "build" directory for an out-of-source build, are all folders that contain nothing more than files that are generated in the build process. And thus, only the src folder needs to be backed up, or much better, kept under a version control system / server (like Git or SVN).
And when it comes to installing your header files on the destination system (if you want to eventually distribute your library), well, CMake has a command for installing files (implicitly creates a "install" target, to do "make install") which you can use to put all the headers into the /usr/include/ directory. I just use the following cmake macro for this purpose:
# custom macro to register some headers as target for installation:
# setup_headers("/path/to/header/something.h" "/relative/install/path")
macro(setup_headers HEADER_FILES HEADER_PATH)
foreach(CURRENT_HEADER_FILE ${HEADER_FILES})
install(FILES "${SRCROOT}${CURRENT_HEADER_FILE}" DESTINATION "${INCLUDEROOT}${HEADER_PATH}")
endforeach(CURRENT_HEADER_FILE)
endmacro(setup_headers)
Where SRCROOT is a cmake variable that I set to the src folder, and INCLUDEROOT is cmake variable that I configure to wherever to headers need to go. Of course, there are many other ways to do this, and I'm sure my way is not the best. The point is, there is no reason to split the headers and sources just because only the headers need to be installed on the target system, because it is very easy, especially with CMake (or CPack), to pick out and configure the headers to be installed without having to have them in a separate directory. And this is what I have seen in most libraries.
Quote: Secondly I would like to use the Google C++ Testing Framework for unit testing my code as it seems fairly easy to use. Do you suggest bundling this with my code, for example in a "inc/gtest" or "contrib/gtest" folder? If bundled, do you suggest using the fuse_gtest_files.py script to reduce the number or files, or leaving it as is? If not bundled how is this dependency handled?
Don't bundle dependencies with your library. This is generally a pretty horrible idea, and I always hate it when I'm stuck trying to build a library that did that. It should be your last resort, and beware of the pitfalls. Often, people bundle dependencies with their library either because they target a terrible development environment (e.g., Windows), or because they only support an old (deprecated) version of the library (dependency) in question. The main pitfall is that your bundled dependency might clash with already installed versions of the same library / application (e.g., you bundled gtest, but the person trying to build your library already has a newer (or older) version of gtest already installed, then the two might clash and give that person a very nasty headache). So, as I said, do it at your own risk, and I would say only as a last resort. Asking the people to install a few dependencies before being able to compile your library is a much lesser evil than trying to resolve clashes between your bundled dependencies and existing installations.
Quote: When it comes to writing tests, how are these generally organised? I was thinking to have one cpp file for each class (test_vector3.cpp for example) but all compiled in to one binary so that they can all be run together easily?
One cpp file per class (or small cohesive group of classes and functions) is more usual and practical in my opinion. However, definitely, don't compile them all into one binary just so that "they can all be run together". That's a really bad idea. Generally, when it comes to coding, you want to split things up as much as it is reasonable to do so. In the case of unit-tests, you don't want one binary to run all the tests, because that means that any little change that you make to anything in your library is likely to cause a near total recompilation of that unit-test program, and that's just minutes / hours lost waiting for recompilation. Just stick to a simple scheme: 1 unit = 1 unit-test program. Then, use either a script or a unit-test framework (such as gtest and/or CTest) to run all the test programs and report to failure/success rates.
Quote: Since the gtest library is generally build using cmake and make, I was thinking that it would make sense for my project to also be built like this? If I decided to use the following project layout:
I would rather suggest this layout:
trunk
├── bin
├── lib
│ └── project
│ └── libvector3.so
│ └── libvector3.a products of installation / building
├── docs
│ └── Doxyfile
├── include
│ └── project
│ └── vector3.hpp
│_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
│
├── src
│ └── CMakeLists.txt
│ └── Doxyfile.in
│ └── project part of version-control / source-distribution
│ └── CMakeLists.txt
│ └── vector3.hpp
│ └── vector3.cpp
│ └── test
│ └── test_vector3.cpp
│_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
│
├── build
└── test working directories for building / testing
└── test_vector3
A few things to notice here. First, the sub-directories of your src directory should mirror the sub-directories of your include directory, this is just to keep things intuitive (also, try to keep your sub-directory structure reasonably flat (shallow), because deep nesting of folders is often more of a hassle than anything else). Second, the "include" directory is just an installation directory, its contents are just whatever headers are picked out of the src directory.
Third, the CMake system is intended to be distributed over the source sub-directories, not as one CMakeLists.txt file at the top-level. This keeps things local, and that's good (in the spirit of splitting things up into independent pieces). If you add a new source, a new header, or a new test program, all you need is to edit one small and simple CMakeLists.txt file in the sub-directory in question, without affecting anything else. This also allows you to restructure the directories with ease (CMakeLists are local and contained in the sub-directories being moved). The top-level CMakeLists should contain most of the top-level configurations, such as setting up destination directories, custom commands (or macros), and finding packages installed on the system. The lower-level CMakeLists should contain only simple lists of headers, sources, and unit-test sources, and the cmake commands that register them to compilation targets.
Quote: How would the CMakeLists.txt have to look so that it can either build just the library or the library and the tests?
Basic answer is that CMake allows you to specifically exclude certain targets from "all" (which is what is built when you type "make"), and you can also create specific bundles of targets. I can't do a CMake tutorial here, but it is fairly straight forward to find out by yourself. In this specific case, however, the recommended solution is, of course, to use CTest, which is just an additional set of commands that you can use in the CMakeLists files to register a number of targets (programs) that are marked as unit-tests. So, CMake will put all the tests in a special category of builds, and that is exactly what you asked for, so, problem solved.
Quote: Also I have seen quite a few projects that have a build ad a bin directory. Does the build happen in the build directory and then the binaries moved out in to the bin directory? Would the binaries for the tests and the library live in the same place? Or would it make more sense to structure it as follows:
Having a build directory outside the source ("out-of-source" build) is really the only sane thing to do, it is the de facto standard these days. So, definitely, have a separate "build" directory, outside the source directory, just as the CMake people recommend, and as every programmer I have ever met does. As for the bin directory, well, that is a convention, and it is probably a good idea to stick to it, as I said in the beginning of this post.
Quote: I would also like to use doxygen to document my code. Is it possible to get this to automatically run with cmake and make?
Yes. It is more than possible, it is awesome. Depending on how fancy you want to get, there are several possibilities. CMake does have a module for Doxygen (i.e., find_package(Doxygen)) which allows you to register targets that will run Doxygen on some files. If you want to do more fancy things, like updating the version number in the Doxyfile, or automatically entering a date / author stamps for source files and so on, it is all possible with a bit of CMake kung-fu. Generally, doing this will involve that you keep a source Doxyfile (e.g., the "Doxyfile.in" that I put in the folder layout above) which has tokens to be found and replaced by CMake's parsing commands. In my top-level CMakeLists file, you will find one such piece of CMake kung-fu that does a few fancy things with cmake-doxygen together.
Structuring the project
I would generally favour the following:
├── CMakeLists.txt
|
├── docs/
│ └── Doxyfile
|
├── include/
│ └── project/
│ └── vector3.hpp
|
├── src/
└── project/
└── vector3.cpp
└── test/
└── test_vector3.cpp
This means that you have a very clearly defined set of API files for your library, and the structure means that clients of your library would do
#include "project/vector3.hpp"
rather than the less explicit
#include "vector3.hpp"
I like the structure of the /src tree to match that of the /include tree, but that's personal preference really. However, if your project expands to contain subdirectories within /include/project, it would generally help to match those inside the /src tree.
For the tests, I favour keeping them "close" to the files they test, and if you do end up with subdirectories within /src, it's a pretty easy paradigm for others to follow if they want to find a given file's test code.
Testing
Secondly I would like to use the Google C++ Testing Framework for unit testing my code as it seems fairly easy to use.
Gtest is indeed simple to use and is fairly comprehensive in terms of its capabilities. It can be used alongside gmock very easily to extend its capabilities, but my own experiences with gmock have been less favourable. I'm quite prepared to accept that this may well be down to my own shortcomings, but gmock tests tends to be more difficult to create, and much more fragile / difficult to maintain. A big nail in the gmock coffin is that it really doesn't play nice with smart pointers.
This is a very trivial and subjective answer to a huge question (which probably doesn't really belong on S.O.)
Do you suggest bundling this with my code, for example in a "inc/gtest" or "contrib/gtest" folder? If bundled, do you suggest using the fuse_gtest_files.py script to reduce the number or files, or leaving it as is? If not bundled how is this dependency handled?
I prefer using CMake's ExternalProject_Add module. This avoids you having to keep gtest source code in your repository, or installing it anywhere. It is downloaded and built in your build tree automatically.
See my answer dealing with the specifics here.
When it comes to writing tests, how are these generally organised? I was thinking to have one cpp file for each class (test_vector3.cpp for example) but all compiled in to one binary so that they can all be run together easily?
Good plan.
Building
I'm a fan of CMake, but as with your test-related questions, S.O. is probably not the best place to ask for opinions on such a subjective issue.
How would the CMakeLists.txt have to look so that it can either build just the library or the library and the tests?
add_library(ProjectLibrary <All library sources and headers>)
add_executable(ProjectTest <All test files>)
target_link_libraries(ProjectTest ProjectLibrary)
The library will appear as a target "ProjectLibrary", and the test suite as a target "ProjectTest". By specifying the library as a dependency of the test exe, building the test exe will automatically cause the library to be rebuilt if it is out of date.
Also I have seen quite a few projects that have a build ad a bin directory. Does the build happen in the build directory and then the binaries moved out in to the bin directory? Would the binaries for the tests and the library live in the same place?
CMake recommends "out-of-source" builds, i.e. you create your own build directory outside the project and run CMake from there. This avoids "polluting" your source tree with build files, and is highly desirable if you're using a vcs.
You can specify that the binaries are moved or copied to a different directory once built, or that they are created by default in another directory, but there's generally no need. CMake provides comprehensive ways to install your project if desired, or make it easy for other CMake projects to "find" the relevant files of your project.
With regards to CMake's own support for finding and executing gtest tests, this would largely be inappropriate if you build gtest as part of your project. The FindGtest module is really designed to be used in the case where gtest has been built separately outside of your project.
CMake provides its own test framework (CTest), and ideally, every gtest case would be added as a CTest case.
However, the GTEST_ADD_TESTS macro provided by FindGtest to allow easy addition of gtest cases as individual ctest cases is somewhat lacking in that it doesn't work for gtest's macros other than TEST and TEST_F. Value- or Type-parameterised tests using TEST_P, TYPED_TEST_P, etc. aren't handled at all.
The problem doesn't have an easy solution that I know of. The most robust way to get a list of gtest cases is to execute the test exe with the flag --gtest_list_tests. However, this can only be done once the exe is built, so CMake can't make use of this. Which leaves you with two choices; CMake must try to parse C++ code to deduce the names of the tests (non-trivial in the extreme if you want to take into account all gtest macros, commented-out tests, disabled tests), or test cases are added by hand to the CMakeLists.txt file.
I would also like to use doxygen to document my code. Is it possible to get this to automatically run with cmake and make?
Yes - although I have no experience on this front. CMake provides FindDoxygen for this purpose.
In addition to the other (excellent) answers, I am going to describe a structure I've been using for relatively large-scale projects.
I am not going to address the subquestion about Doxygen, since I would just repeat what is said in the other answers.
Rationale
For modularity and maintainability, the project is organized as a set of small units.
For clarity, let's name them UnitX, with X = A, B, C, ... (but they can have any general name).
The directory structure is then organized to reflect this choice, with the possibility to group units if necessary.
Solution
The basic directory layout is the following (content of units is detailed later on):
project
├── CMakeLists.txt
├── UnitA
├── UnitB
├── GroupA
│ └── CMakeLists.txt
│ └── GroupB
│ └── CMakeLists.txt
│ └── UnitC
│ └── UnitD
│ └── UnitE
project/CMakeLists.txt could contain the following:
cmake_minimum_required(VERSION 3.0.2)
project(project)
enable_testing() # This will be necessary for testing (details below)
add_subdirectory(UnitA)
add_subdirectory(UnitB)
add_subdirectory(GroupA)
and project/GroupA/CMakeLists.txt:
add_subdirectory(GroupB)
add_subdirectory(UnitE)
and project/GroupB/CMakeLists.txt:
add_subdirectory(UnitC)
add_subdirectory(UnitD)
Now to the structure of the different units (let's take, as an example, UnitD)
project/GroupA/GroupB/UnitD
├── README.md
├── CMakeLists.txt
├── lib
│ └── CMakeLists.txt
│ └── UnitD
│ └── ClassA.h
│ └── ClassA.cpp
│ └── ClassB.h
│ └── ClassB.cpp
├── test
│ └── CMakeLists.txt
│ └── ClassATest.cpp
│ └── ClassBTest.cpp
│ └── [main.cpp]
To the different components:
I like having source (.cpp) and headers (.h) in the same folder. This avoids a duplicate directory hierarchy, makes maintenance easier. For installation, it is no problem (especially with CMake) to just filter the header files.
The role of the directory UnitD is to later on allow including files with #include <UnitD/ClassA.h>. Also, when installing this unit, you can just copy the directory structure as is. Note that you can also organize your source files in subdirectories.
I like a README file to summarize what the unit is about and specify useful information about it.
CMakeLists.txt could simply contain:
add_subdirectory(lib)
add_subdirectory(test)
lib/CMakeLists.txt:
project(UnitD)
set(headers
UnitD/ClassA.h
UnitD/ClassB.h
)
set(sources
UnitD/ClassA.cpp
UnitD/ClassB.cpp
)
add_library(${TARGET_NAME} STATIC ${headers} ${sources})
# INSTALL_INTERFACE: folder to which you will install a directory UnitD containing the headers
target_include_directories(UnitD
PUBLIC $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}>
PUBLIC $<INSTALL_INTERFACE:include/SomeDir>
)
target_link_libraries(UnitD
PUBLIC UnitA
PRIVATE UnitC
)
Here, note that it is not necessary to tell CMake that we want the include directories for UnitA and UnitC, as this was already specified when configuring those units. Also, PUBLIC will tell all targets that depend on UnitD that they should automatically include the UnitA dependency, while UnitC won't be required then (PRIVATE).
test/CMakeLists.txt (see further below if you want to use GTest for it):
project(UnitDTests)
add_executable(UnitDTests
ClassATest.cpp
ClassBTest.cpp
[main.cpp]
)
target_link_libraries(UnitDTests
PUBLIC UnitD
)
add_test(
NAME UnitDTests
COMMAND UnitDTests
)
Using GoogleTest
For Google Test, the easiest is if its source is present in somewhere your source directory, but you don't have to actually add it there yourself.
I've been using this project to download it automatically, and I wrap its usage in a function to make sure that it is downloaded only once, even though we have several test targets.
This CMake function is the following:
function(import_gtest)
include (DownloadProject)
if (NOT TARGET gmock_main)
include(DownloadProject)
download_project(PROJ googletest
GIT_REPOSITORY https://github.com/google/googletest.git
GIT_TAG release-1.8.0
UPDATE_DISCONNECTED 1
)
set(gtest_force_shared_crt ON CACHE BOOL "" FORCE) # Prevent GoogleTest from overriding our compiler/linker options when building with Visual Studio
add_subdirectory(${googletest_SOURCE_DIR} ${googletest_BINARY_DIR} EXCLUDE_FROM_ALL)
endif()
endfunction()
and then, when I want to use it inside one of my test targets, I will add the following lines to the CMakeLists.txt (this is for the example above, test/CMakeLists.txt):
import_gtest()
target_link_libraries(UnitDTests gtest_main gmock_main)
What's a good directory structure for larger C++ projects using Makefile ?
This is how my directory structure looks at the moment:
lib/ (class implementations *.cpp)
include/ (class definitions *.h)
tests/ (main.cpp for quick tests)
Now, I'm not sure how my Makefile should look like... it doesn't seem to work when .cpp files and .h files aren't in the same directory. Could anyone point me to a common directory structure with an accompanying Makefile so that I don't reinvent the wheel ?
Separating the .cpp of the .h file is not always a good solution. Generally I separate both of them when it is used as a library (public header in include and private header with the source code).
If it is a library, this structure is ok.
lib/ (class implementations *.cpp .h)
include/ (class definitions *.h) <- Only those to be installed in your system
tests/ (main.cpp for quick tests)
doc/ (doxygen or any kind of documentation)
If it is a application
src/ (source for the application)
lib/ (source for the application library *.cpp *.hpp)
include/ (interface for the library *.h)
tests/ (main.cpp for quick tests) <- use cppunit for this part
doc/ (doxygen or any kind of documentation)
Use the flag -I$(PROJECT_BASE)/include to specify the include path for the compilation
If it is a big project, it can be good to use tool like autoconf/automake or cmake to build everything. It will ease the development.
For those who find this question after 2020, an alternative modern and reasoned vision of "Canonical Project Structure" for C++ has been presented by Boris Kolpackov: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2018/p1204r0.html
Briefly - no include/ and src/ split. All headers, sources, modules and unit tests go into one directory. Implementation details may be separated from public API by moving to <name>/<name>/details/ subdirectory.
<name>/
├── <name>/
│ ├── headers...
│ ├── sources...
│ ├── modules...
│ └── unit tests...
└── tests/
├── functional_test1/
├── functional_test2/
├── integration_test1/
├── integration_test2/
└── ...
For example:
bestlib/
├── bestlib/
│ ├── foo.h
│ ├── foo.cpp
│ ├── foo.test.cpp
│ ├── bar.h
│ ├── bar.cpp
│ └── bar.test.cpp
└── tests/
├── functional_test1/
└── integration_test1/
If you have many source files, it may also be a good idea to further subdivide your source directory. For instance, one subdirectory for the core functionality of your application, one for the GUI, etc.
src/core
src/database
src/effects
src/gui
...
Doing so also forces you to avoid unneeded relationships between your "modules", which is a prerequisite to nice and reusable code.
There is no one specific or required directory structure.
You can set it up anyway you like. Your problem is simple to solve. Just instruct Makefile to look into subdirectories or put compiled objects into subdirectories instead of using just current directory.
You would just use in Makefile paths:
%.o : %.cpp
replace with
bin/%.o : %.cpp
So it will check if binary file in directory bin exists and so on, you can apply the same to locations where files are compiled.
There are ways to add/remove/modify paths of source and object files.
Have a look at gnu make manual, specifically section 8.3 Functions for File Names,and the one before that 8.2 Functions for String Substitution and Analysis.
You can do stuff like:
get a list of objects from list of source files in current directory:
OBJ = $(patsubst %.cpp, %.o, $(wildcard *.cpp))
Output:
Application.o Market.o ordermatch.o
If binary objects are in subdirectory bin but source code is in current directory you can apply prefix bin to generated object files:
OBJ = $(addprefix bin/,$(patsubst %.cpp, %.o, $(wildcard *.cpp)))
Output:
bin/Application.o bin/Market.o bin/ordermatch.o
And so on.
This is an old question. But you can consider the Pitchfork Project as a general guide.
https://github.com/vector-of-bool/pitchfork for the project.
Some Documentation here
There is no "good directory structure". Pick a structure you're comfortable with and stick to it. Some like placing source files (headers and implementation files) in a src/ directory, so the root directory of the project has nothing but a makefile, a readme and little else. Some like placing helper libraries under a lib/ directory, unittests under test/ or src/test/, documentation under doc/ etc.
I have yet to hear of anyone splitting header files and implementation files into two distinct directories though. Personally I don't like splitting files into directories much. I usually place all my source in a single directory and all the documentation in another directory. If I rely on good search tools anyway, there's no need for a complex directory structure.
make can deal with the sort of structure where the makefile resides in a different directory than the source. The only thing is that it will invoke the rules from the directory of the makefile -- compilers usually have no problem compiling source that is in some subdirectory. You don't have to specify relative paths in your #includes; just specify the include path with compiler flags (gcc's -I flag etc).
If you haven't seen it before read Recursive Make Considered Harmful.
Short, short version: Though very common the recursive make idiom is less than optimal and gets ever worse as projects grow larger and more complicated. An alternative is presented.
Related link: What is your experience with non-recursive make?