CMake improvements

[friendlyanon]

This PR will modernize the CMake usage of this library.

Some things to note:

• For conan the cmake generator should NOT be used, you should be using at least the cmake_find_package to generate targets that can be discovered using CMAKE_MODULE_PATH. The more ideal generator would be cmake_find_package_multi, but conan v2 is not yet out.
• GTest was being compiled with every single CI run, instead now a script pulls in a known good commit of GTest, compiles it, installs it to a prefix location that can be discovered using CMAKE_PREFIX_PATH and caches GTest. While GTest does not support clients using test discovery, CMake does support clients using test discovery via the GoogleTest module, so ctest does not need -VV verbosity level to show a proper test progression summary.

To see how the above affects building the library, please see the CI workflows that have been adapted for this improved process.

This library will also provide a real relocatable CMake package for easy consumption by clients.

For the rationale behind the changes, please see my contributions to FunctionalPlus for related discussions.

[codecov-io]

Codecov Report

Merging #141 (1a0eaca) into master (6b107a7) will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #141   +/-   ##
=======================================
  Coverage   96.94%   96.94%           
=======================================
  Files          67       67           
  Lines         950      950           
=======================================
  Hits          921      921           
  Misses         29       29           

Impacted Files	Coverage Δ
include/hsm/details/chain_actions.h	100.00% <ø> (ø)
include/hsm/details/collect_events.h	100.00% <ø> (ø)
include/hsm/details/collect_initial_states.h	100.00% <ø> (ø)
include/hsm/details/collect_states.h	100.00% <ø> (ø)
include/hsm/details/create_state.h	100.00% <ø> (ø)
include/hsm/details/dispatch_table.h	100.00% <ø> (ø)
include/hsm/details/fill_dispatch_table.h	100.00% <ø> (ø)
...de/hsm/details/flatten_internal_transition_table.h	100.00% <ø> (ø)
include/hsm/details/flatten_transition_table.h	100.00% <ø> (ø)
include/hsm/details/for_each.h	100.00% <ø> (ø)
... and 16 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 6b107a7...1a0eaca. Read the comment docs.

[erikzenker]

Hi @friendlyanon thank you for your contribution. I will start to review your PR and might come up with some question. I think it will take me a while to go through all the changes.

[friendlyanon]

Resolved the conflicts by rebasing the changes on top of the target branch.

[friendlyanon]

Moving the headers to include from src/include is the least surprising. When I see a src or source directory, it's reasonable to assume that a library builds artifacts, i.e. isn't header only. It's also more common to do it this way than not.

[erikzenker]

@friendlyanon thx again for your contribution. Out of curiosity: do you use the hsm somewhere?

[friendlyanon]

I don't use this library yet, but as you might have seen from my recent activity on GitHub, I am the change I want to see in the world. I just want projects to utilize CMake properly, so they are trivial to consume and package.