ament_pep257 and flake8 failures with pydocstyle 5.0.0

[nuclearsandwich]

Many new pep257 errors have come with the release of pydocstyle 5.0.0.

All of these that I've looked at so far have been D401 related and this version of pydocstyle fixed an issue preventing that check from being run in all cases.

D401: First line should be in imperative mood; try rephrasing

Affected packages:

• ros2/launch Use imperative mood in docstrings. launch#362
  • launch
  • launch_testing
  • launch_xml
  • launch_yaml
• ros2/launch_ros Use imperative mood in constructor docstrings. launch_ros#103
  • launch_ros
  • ros2launch
• quality_of_service_demo_py Use imperative mood in docstring. demos#422
• rclpy Update constructor docstrings to use imperative mood. rclpy#480
• ros2topic Use imperative mood in constructor docstring. ros2cli#422
• rosidl_parser Use imperative mood in constructor docstring. rosidl#425
• micro-ROS/ros_tracing https://gitlab.com/micro-ROS/ros_tracing/ros2_tracing/merge_requests/118
  • tracetools_launch
  • tracetools_test

[nuclearsandwich]

PR to address the changes in ros2_tracing, the others are visibule using GitHub cross-references. https://gitlab.com/micro-ROS/ros_tracing/ros2_tracing/merge_requests/118

[nuclearsandwich]

CI for just Linux with the affected packages. The ros2_tracing repo is not included because I can't push branches there.

I forgot to push a few branches so the CI above is incomplete.

[nuclearsandwich]

CI running linter tests for all above PRs except the ros_tracing merge request

[wjwwood]

I don't know if it something we want to fix at all, but just so people are aware, this (and related changes) affect Dashing CI as well.

[nuclearsandwich]

I don't know if it something we want to fix at all, but just so people are aware, this (and related changes) affect Dashing CI as well.

I've started running CI for non-latest rosdistros with linters off. We've generally held back on backporting linter only changes but if they're muddying CI and there's a willing backporter there is no reason not to.

[nuclearsandwich]

The last outstanding issue was resolved with ros2/ros2#838