Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

[question] using `boost/system` recipe with system-wide config

See original GitHub issue

I’ve read the CONTRIBUTING guide.

Hi! My question is about conan recipes for system packages (like gtk/system or xorg/system from CCI). In our ecosystem we want to use boost/system (custom one) together with boost/1.7? from CCI (we need a transparent switching between them in the root of dependency tree). We use CMakeDeps generator. So before build it generates (among other things) Boost-config.cmake, which is very hard to reproduce in boost/system .

Before 1.36 we use the following smart crutch which exploits a bug (?) in generator. Our recipe declare package boost/system with filename (in latest terms cmake_file_name) SystemBoost and with name (in latest terms cmake_target_name) Boost. When somebody (e.g. A) wants to depend on boost, CMakeDeps generates unused SystemBoost-config.cmake and call find_package(Boost ... ) in A-config.cmake. This find_package uses system-wide FindBoost.cmake or config from boost-cmake project and works perfectly.
In 1.36 this bug was fixed. (the method was introduced). So the crutch is not working now.
Could you elaborate this situation?

There are my thoughts about it:

I like the idea to use system-wide config in the specific case of system boost.
Generating conan’s Boost-config.cmake to emulate a system one seems very hard. In CCI the only way to do this is using PkgConfig, but Boost has no official .pc file AFAIK.
Maybe CMakeDeps can take a system-wide boost config somehow (copying or symlinking it instead of generate its own) in this specific case?
Maybe conan can legitimate the crutch described above?

Thanks in advance!