Add lookahead behavior for GoToPlace and fix various sources of seg faults

[mxgrey]

This PR came from some rapid development cycles while getting a demo to work, so it's a mixed bag of changes.

• Adds a traffic schedule lookahead ability for GoToPlace, which helps to smooth out traffic negotiations for tasks that involve passing through a sequence of waypoints (e.g. patrolling)
• Removes stdout/stderr redirect for the python bindings, since that was found to cause segfaults (this is a known issue which should be fixed in newer versions of pybind11: pythonbuf fix pybind/pybind11#2675)
• Fixes a logic error for the cancel + kill API which prevented those APIs from working unless the user provides a callback
• Provides the most general possible update_position overload to give system integrators maximum flexibility in how they report their robot's position

[codecov]

Codecov Report

Merging #205 (46d1cac) into main (212225e) will decrease coverage by 0.09%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #205      +/-   ##
==========================================
- Coverage   17.91%   17.81%   -0.10%     
==========================================
  Files         420      844     +424     
  Lines       38606    77812   +39206     
  Branches    18500    37372   +18872     
==========================================
+ Hits         6915    13860    +6945     
- Misses      24283    49054   +24771     
- Partials     7408    14898    +7490     

Flag	Coverage Δ
tests	17.81% <ø> (-0.10%)	⬇️

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

Impacted Files	Coverage Δ
...os2/rmf_fleet_adapter/src/lift_supervisor/main.cpp
...f_task_ros2/src/rmf_task_ros2/bidding/Response.cpp
...RxCpp-4.1.0/Rx/v2/src/rxcpp/sources/rx-iterate.hpp
...RxCpp-4.1.0/Rx/v2/src/rxcpp/sources/rx-iterate.hpp
...adapter/src/rmf_fleet_adapter/phases/DockRobot.hpp
...ter/src/rmf_fleet_adapter/events/PerformAction.cpp
...s2/src/rmf_traffic_ros2/schedule/convert_Patch.cpp
.../rmf_fleet_adapter/test/phases/test_IngestItem.cpp
...xCpp-4.1.0/Rx/v2/src/rxcpp/sources/rx-interval.hpp
...ter/src/rmf_fleet_adapter/phases/WaitForCharge.cpp
... and 1254 more

[aaronchongth]

These features look great! The refactoring in Task manager also makes it easier to go through, thanks!

I just have 2 pretty trivial comments, but otherwise LGTM! Should we be updating the change logs for these packages as well? You can also mention the newly added API for cancelling or killing tasks from RobotUpdateHandle, those were extremely useful ✨

[aaronchongth]

Got confused by this for a while and thought it was a typo, we might need to converge to single or double L at some point, I found multiple locations where we use double Ls, https://github.com/open-rmf/rmf_api_msgs/blob/main/rmf_api_msgs/schemas/task_state.json#L51.

This should not block merging this PR though

[mxgrey]

Yeah, the inconsistency between the number of ls in canceled versus cancellation probably creates a high risk of typos, but I chose those spellings as they're the more common spellings in American English, which is usually preferred over British spellings in the global software industry. I'm open to revising the API if people anticipate that it will cause problems.

[aaronchongth]

Are these comments still useful?

[mxgrey]

Thanks for catching that, comments are removed in 9433066

[aaronchongth]

On second thought, since these are using features in 2.1.0 of rmf_task and rmf_task_sequence, I'd say we should probably consider bumping a minor release to match as well. I'm not super well versed in our versioning workflow, let me know what you think.

[youliangtan]

I tested a few scenarios, with queue task cancellation and the new generic update_position(startset) API for some demo use cases. All works perfectly 👍

[mxgrey]

@aaronchongth great idea, CHANGELOG is updated in 46d1cac. Let me know if there's anything else to block this PR.

[aaronchongth]

Thanks for the changes fixes! 🙇LGTM!