But I’ve rarely seen pkg-config solutions work with a lot of software I work on, we tend to use CMake stuff.
I think it is absolutely the case that pkg-config would need some additional features for it to work better and correctly on the presence of modules.
I do agree that pkg-config without any sort of curation work (like what we do internally on our own "distribution") will very frequently not work out of the box.
pkg-config grew organically out of very specific use cases, so it's not unexpected that it's incomplete.
With that in mind, I wanted to talk about some requirements we’d have
I think if we're taking the dive into package management we definitely need to collect a bunch of requirements papers, like the one in this thread.
It would be very important to have as many build system maintainers as possible to weigh in.
And there lies my concern, which is that we need to solve package management to use modules effectively, when this has been a thorny issue for so long.
I'm not against taking that plunge, but I think it may be worth exploring a solution that doesn't require us solving the hardest problem in the C++ ecosystem today.