[Fix] Cleanup Launch Files

[philipchurchley]

Addresses #950 and follows up on #959 by cleaning the remaining launch files. The launch files have been improved to use a declarative programming style, as described by @emersonknapp at ROSCon, which can be viewed in the livestream at 6:10.

Summary

• Converted ROS 2 Python launch files to a consistent declarative style (single
  LaunchDescription([...]), replaced PathJoinSubstitution with PathSubstitution + /,
  used literal "xacro" in Command, and removed RegisterEventHandler/OnProcessExit).

Validation performed

• Ran non-building smoke checks (import + generate_launch_description() for each launch) — all passed.
• Ran a semantic comparator vs master; it flagged 18 files but most differences are cosmetic or expected.

Potential confusions for reviewer attention

• example_15: added publisher_config launch argument in test launches. Please confirm the default file and behavior.
• Removal of RegisterEventHandler entries may affect startup/shutdown ordering for some bringup launches. Please check that this removal doesn't break anything.

Recommendation

• I have no experience with Ros2 so I’m not certain about runtime impacts of the items above. Please review flagged files and run integration/CI launch tests where ordering or parameters matter. If reviewers find real regressions, please recommend minimal fixes (re-introduce ordering, adjust defaults, or correct node fields).

[philipchurchley]

I will be away for the next week, so I apologize if I do not immediately respond to comments.

[christophfroehlich]

Jobs are failing, can you have a look please?

[mergify]

This pull request is in conflict. Could you fix it @philipchurchley?

[christophfroehlich]

Thanks a lot for the big changes.
My first review round below. Please also fix the merge conflict with example_5 now.

[philipchurchley]

I'm not sure what's happening with the wrench transformer errors because that seems to have been added with the merge commit. I resolved the merge conflicts by just accepting my branches version and then I added the new launch argument to the launch file. However, looking at the diff for commit 74a8f6c, the merge commit also seems to have added other stuff that may be the cause for these failing tests.

[Juliaj]

However, looking at the diff for commit 74a8f6c, the merge commit also seems to have added other stuff that may be the cause for these failing tests.

@philipchurchley, the test failures were due to the new wrentch_transformer wasn't launched. Added my suggestion.

[christophfroehlich]

Except for failing pre-commit jobs this is ready for review?

[philipchurchley]

Except for failing pre-commit jobs this is ready for review?

Yes, although I'm not sure what the failing jobs are for. Could it have to do with the missing EventHandlers?

[Juliaj]

Yes, although I'm not sure what the failing jobs are for. Could it have to do with the missing EventHandlers?

@philipchurchley, for the failed pre-commit jobs, run following commnd in your dev env. If you don't have pre-commit installed, you can find instructions from https://control.ros.org/rolling/doc/contributing/contributing.html.

pre-commit run --all-files

fix any issues reported and commit the changes.

[christophfroehlich]

Yes, although I'm not sure what the failing jobs are for. Could it have to do with the missing EventHandlers?

binary builds are failing because this wrench transformer node was released but not synced yet.

[mhubii]

just curious, is there a way for the controller manager to load controllers in the correct order other than event handlers? (for chainable controllers)

[christophfroehlich]

what do you mean by correct order? If they are part of a chain, simply load them from the same spawner. The system will figure out the hierarchy.

[mhubii]

oh cool, didn't know that was possible! :) thx