Fix Instance() method of Singleton classes

[traversaro]

In particular, make sure that all the Instance() methods of the same class always use the same instantiation.

🦟 Bug fix

Fixes conda-forge/gazebo-feedstock#148 .

Summary

This should fix all the cases in which multiple instances of a Singleton classes were observed, as for example in conda-forge/macOS build of Classic Gazebo (see conda-forge/gazebo-feedstock#148 (comment)). The fix was inspired from gazebosim/gz-sim#1725 and from the related PRs:

• Fixed MeshManager Singleton gz-common#451
• Prevent multiple singletons from being instantiated gz-rendering#734

Thanks a lot to @azeey for suggesting the correct fix in conda-forge/gazebo-feedstock#152 (comment) .

Checklist

• Signed all commits for DCO
• Added tests
• Updated documentation (as needed)
• Updated migration guide (as needed)
• Consider updating Python bindings (if the library has them)
• codecheck passed (See contributing)
• All tests passed (See test coverage)
• While waiting for a review on your PR, please help review another open pull request to support the maintainers

Note to maintainers: Remember to use Squash-Merge and edit the commit message to match the pull request summary while retaining Signed-off-by messages.

[azeey]

Sorry for the delay in reviewing this. I have one comment regarding deprecation, otherwise, LGTM!

[traversaro]

What is your opinion regarding merging this @scpeters @azeey ? We just hit a regression involving this patch in conda-forge (see conda-forge/gazebo-feedstock#161), to avoid this in the future it would be great if we could get this eventually in an official release.

[azeey]

+1 for merging.

[scpeters]

thank you, this looks like a good set of changes

I'm rerunning Ubuntu CI and then it will be ready for merge, though I may need to do some testing with ROS before releasing this

[traversaro]

I'm rerunning Ubuntu CI and then it will be ready for merge, though I may need to do some testing with ROS before releasing this

Sure, we can also just wait for the ROS tests before even merging this PR, I am not particulary in a hurry on this one. I am just afraid that at some point the problem that affected conda-forge macOS builds also started to affect other builds (for example homebrew builds as soon as a new version of Apple Clang is released).