Send state message when components are removed

[nkoenig]

Signed-off-by: Nate Koenig [email protected]

🦟 Bug fix

Summary

The WorldCmdPose was being transmitted to the GUI and back to the Server, resulting in bad behavior. See here.

This PR changes the SceneBroadcaster to send a new state message when components are removed.

Before I write tests, I'd like to get some feedback on this approach.

Checklist

• Signed all commits for DCO
• Added tests
• Updated documentation (as needed)
• codecheck passed (See contributing)
• All tests passed (See test coverage)
• While waiting for a review on your PR, please help review another open pull request to support the maintainers

Note to maintainers: Remember to use Squash-Merge

[codecov]

Codecov Report

Merging #1235 (bf374c9) into ign-gazebo6 (4152bdf) will increase coverage by 0.03%.
The diff coverage is 100.00%.

@@               Coverage Diff               @@
##           ign-gazebo6    #1235      +/-   ##
===============================================
+ Coverage        62.23%   62.26%   +0.03%     
===============================================
  Files              275      275              
  Lines            22710    22713       +3     
===============================================
+ Hits             14133    14142       +9     
+ Misses            8577     8571       -6     

Impacted Files	Coverage Δ
include/ignition/gazebo/EntityComponentManager.hh	100.00% <ø> (ø)
src/EntityComponentManager.cc	87.07% <100.00%> (+0.04%)	⬆️
src/systems/scene_broadcaster/SceneBroadcaster.cc	93.75% <100.00%> (ø)
src/SimulationRunner.cc	93.62% <0.00%> (+0.98%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4152bdf...bf374c9. Read the comment docs.

[chapulina]

Nice! We just need a unit test.

[nkoenig]

Nice! We just need a unit test.

Just checking that you don't think there are unforeseen side effects of this?

[chapulina]

Just checking that you don't think there are unforeseen side effects of this?

I can't think of any unwanted effects, the worst I can think of is that we're publishing more state messages, but that should be negligible if there aren't components being removed every iteration.

[azeey]

We have been preferring to zero out components instead of removing them for performance reasons, so I don't think we have any components that get removed every iteration. So I think this approach might work as a work around. My proposal for a more long term solution is to have a scheme where components are requested for removal, just like entities, and their removal happens after all the *Update phases have completed. I think this would remove the need to keep WorldPoseCmd around for an extra iteration.

[nkoenig]

Tests have been added in b2ad027