Adds a basic unit test for PointCloud functionality

[arjo129]

This PR adds a very trivial (happy path) unit test for PointCloud functionality. Hopefully, I will get more time to add more functionality to this test over time.

TODOS:

• Check behaviour with NaNs
• Check mismatched float and pointcloud sizes
• Check malformed point cloud.

Signed-off-by: Arjo Chakravarty [email protected]

[codecov]

Codecov Report

Merging #496 (69b3972) into gz-gui7 (c93dd14) will increase coverage by 2.12%.
The diff coverage is n/a.

❗ Current head 69b3972 differs from pull request most recent head e79ec8d. Consider uploading reports for the commit e79ec8d to get more accurate results

@@             Coverage Diff             @@
##           gz-gui7     #496      +/-   ##
===========================================
+ Coverage    70.91%   73.04%   +2.12%     
===========================================
  Files           45       45              
  Lines         4941     4941              
===========================================
+ Hits          3504     3609     +105     
+ Misses        1437     1332     -105     

see 1 file with indirect coverage changes

[jennuine]

@arjo129 what is the status of this PR? Will you be addressing the TODOs here?

[arjo129]

We can merge this as is and Ill increase coverage over time.

[jennuine]

Could this test be added to PointCloud_TEST instead?

[jennuine]

Left some more minor comments. Could you summarize the reason for the point cloud publisher/msg? In the marker message callback, only the flatMsg data is being checked but not pcMsg. It would be nice to check the point cloud data as well.

[arjo129]

Could you summarize the reason for the point cloud publisher/msg? In the marker message callback, only the flatMsg data is being checked but not pcMsg. It would be nice to check the point cloud data as well.

The rationale for this is that the pcMsg provides data about position of the markers whereas the marker messages provide data about position. The test checks that the marker messages have the correct color values (which is what the point-cloud plugin basically does).

[mjcarroll]

I think that the focal failures here in CI are unrelated to this PR?

addressed