[CI] Add clang job, setup concurrency, use rt_tools humble branch

[christophfroehlich]

• This adds a semi-binary job using clang. I fixed all what clang was complaining about.
• use humble branch of realtime_tools for humble CI

Closes #611

Additionally, I added the concurrency settings which now automatically cancel running workflows if anything is pushed to PRs.
https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/control-the-concurrency-of-workflows-and-jobs

[codecov]

Codecov Report

Attention: Patch coverage is 93.75000% with 1 line in your changes missing coverage. Please review.

Project coverage is 87.75%. Comparing base (0ef99fe) to head (61dd1e3).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
controller_manager/src/ros2_control_node.cpp	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1910   +/-   ##
=======================================
  Coverage   87.75%   87.75%           
=======================================
  Files         122      122           
  Lines       12940    12940           
  Branches     1155     1155           
=======================================
  Hits        11355    11355           
  Misses       1159     1159           
  Partials      426      426           

Flag	Coverage Δ
unittests	87.75% <93.75%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...nclude/semantic_components/force_torque_sensor.hpp	100.00% <100.00%> (ø)
...roller_interface/src/controller_interface_base.cpp	82.47% <100.00%> (ø)
hardware_interface/src/component_parser.cpp	94.82% <100.00%> (ø)
...e_interface/src/mock_components/generic_system.cpp	83.96% <100.00%> (ø)
...dware_interface/test/test_component_interfaces.cpp	97.63% <100.00%> (ø)
hardware_interface/test/test_macros.cpp	100.00% <100.00%> (ø)
controller_manager/src/ros2_control_node.cpp	66.66% <0.00%> (ø)

[saikishor]

In general LGTM

I noticed that we don't cancel jobs of pre-commit and docs. Anyway, they are pretty fast.

I'm also stumbled that clang doesn't like auto

[christophfroehlich]

In general LGTM

I noticed that we don't cancel jobs of pre-commit and docs. Anyway, they are pretty fast.

they are cancelled! the other setups are the following: they are only cancelled in PRs, but not if it is pushed to the respective branches

I'm also stumbled that clang doesn't like auto

auto is not the problem, but std::distance etc return an integer.

[saikishor]

LGTM