Expressing runtime dependencies in CMake

Code
1

Hi all,

I am working on a project making excessive use of run-time library loading, i.e. dlopen()/LoadLibrary().

This means that if target B uses library A, it needs:

  • at compile time: the include directories of A
  • at link time: nothing
  • at load time: nothing
  • at run time: the shared library A to be present

Currently I do the following:
add_library(A.SharedLib SHARED …)

add_library(A INTERFACE)
target_include_directories(A INTERFACE path/to/A/includes)
add_dependencies(A A.SharedLib) # make sure that if a target depending on A is built, then A.SharedLib is also built so that it is available at run time

target_link_libraries(B PRIVATE A)

This works, but it has the drawback of non-optimal parallelization: A and B could be built in parallel, but thanks to add_dependencies(), A must be completely built and linked before B starts to build.

So: is there a way to tell CMake that target B needs target A only at runtime, or in other words:

  • if B is built, A must be brought up to date as well
  • but B does not need A at compile or link time

We use Visual Studio, Makefile and Ninja as generators. So adding another “dummy” target that depends on both B and A is probably out because developers would still build and run the executable target A from their VS environment and not the dummy target.

2

I think this is the best you can do right now. CMake doesn’t know why that dependency exists, so it assumes that it is needed before anything happens in B (generated headers and the like).

I have written somewhere about runtime usage requirements, but I haven’t done any work towards it (or found the post…probably on a CMake issue).