Add PoseStampedArray

[tonynajjar]

The need originated from here. @SteveMacenski

[mjcarroll]

No arguments from me.

Should we potentially add a comment here about what the difference between the root std_msgs/Header and the per-message std_msgs/Header mean here, or is that entirely application specific.

[SteveMacenski]

I think the global message stamp could be removed?

[tonynajjar]

I think the global message stamp could be removed?

Would you then also remove the header from this one? -> https://github.com/ros2/common_interfaces/blob/rolling/nav_msgs/msg/Path.msg
Same concept no?

[SteveMacenski]

The path's header is for the path itself, the pose headers are for each of the poses in the path if it were a time denoted trajectory. Admittedly its rarely (if ever) used that way, but it is conceptually the right way for the path to be structured for kinodynamic paths, should uses need to represent one.

Does the application of PoseStampedArray have a similar need? I'm ambivalent, just couldn't think of a reason why we needed it for PoseStampedArray. I think to Michael's point, there's no objections, but if included we should explain how it should be used so putting that reason inline would resolve his concern and provide clarity. If we can't think of a reason, then maybe just remove it

[tonynajjar]

Okay. I also don't see a reason to keep it but I wanted to stay consistent. Your explanation as to why nav_msgs/Path would have a header makes sense. I removed the header

[tonynajjar]

While working with this message we actually found a potential use for the global header. I added a comment but see here for more info. Feel free to propose a reformulation

[SteveMacenski]

@mjcarroll I think this is in the final form with the header description. Do we need another maintainer to review or can we ship?

[tonynajjar]

Thanks for the review @clalancette , changes were made, anything else?

[SteveMacenski]

We're also planning on using this for the NavigateToPose interface - which makes alot more sense than our vector<PoseStamped> for a list of goal representation so we can have a stamp to gauge the age of the original request. We're finding more places to use this by the day :-)

[tonynajjar]

We're also planning on using this for the NavigateToPose interface

NavigateThroughPoses*

[tonynajjar]

Happy new year! (ping in disguise @clalancette 😄 )

[mjcarroll]

Pulls: #262
Gist: https://gist.githubusercontent.com/mjcarroll/e51db32e9e99205be4fa32319946c28b/raw/5e9a971bdd5ed5eb905513bd6477c1a698030d59/ros2.repos
BUILD args:
TEST args:
ROS Distro: rolling
Job: ci_launcher
ci_launcher ran: https://ci.ros2.org/job/ci_launcher/15039

• Linux
• Linux-aarch64
• Linux-rhel
• Windows

[SteveMacenski]

@mjcarroll they've all passed :-)

[gavanderhoorn]

It's surprising to see this merged, especially after the many discussions about extending geometry_msgs with more messages with low semantic value (of which the linked ros/common_msgs#144 is an example).

What changed? Different maintainers, changed opinions, evolution of best-practices? All of the above?

[Swepz]

Friendly reminder to bump tag :)

[tfoote]

I'd like to review the naming and documentation of this before we make a release to be sure we've locked down the semantics of this message. I've added it to the PMC meeting agenda for next week.