Fix find (or build) optional dependencies when optional dependency is already installed

[zachcran]

Is this pull request associated with an issue(s)?

Fixes #166.

Description

Main Fix

This PR updates cmaize_find_optional_dependency() (cfod()) and cmaize_find_or_build_optional_dependency() (cfobod()) calls to reference the correct target name when adding the cmaize_option() enabling the optional dependency to the dependency target's target_compile_definitions().

Enhanced Output

While debugging this issue, I had to investigate code spanning most of the life cycle of a target (both dependency targets and otherwise), so I also added more verbose output where possible, which can be enabled by setting the message log level at or finer than VERBOSE (log levels here).

TODOs

• Tested on Sigma and TensorWrapper with both cfod() and cfobod() calls for the Eigen dependency, both when Eigen is installed and not.
• Fix failing unit tests (I am testing with CMake 3.26-rc4).
• Test CMake>=4. Even if the CMake 3.x tests pass, docs/bare_bones_cmake was failing in CMake 4.0.2 and needs to be checked manually.
• Check integration tests on my system.

[zachcran]

Fixed failing tests on CMake 3.26.0-rc4 and the unit tests here are passing. Still need to verify on CMake 4.x.

[zachcran]

On CMake 4.0.2-dirty, tests/docs/bare_bones_cmake fails due to no cmake_minimum_required() command being present. This is not related to the changes made here (CMake 4 requires a minimum version now; see CMP0000), so I vote to move forward with merging these changes regardless of the failing test. I'll move this discovery to a separate issue.

Just to be clear, while this removes a blocker, I still do not consider this PR r2g yet until I have a chance to test the integration tests as well.

[zachcran]

@ryanmrichard This is r2g. Integration tests passed on my system, as well as all of the other reported tests that I conducted.

[github-actions]

🚀 [bumpr] Bumped!
New version:v1.1.7
Changes:v1.1.6...v1.1.7