John Stechschulte

I recommend you double-check that the RobotState isn't corrupted. I had a segfault with a nearly-identical stack trace and it was because there were NaN values in the collision body...

The offending call to `clear()` is [here](https://github.com/ros-planning/moveit/blob/736219d8a5d55d952b85eebf52ae6ea4e3220da8/moveit_ros/planning/planning_scene_monitor/src/planning_scene_monitor.cpp#L532). I'm calling this function directly, rather than through [the service](https://github.com/ros-planning/moveit/blob/b35e64084ad0bc2a30aaf663ba54197ad311edb3/moveit_ros/move_group/src/default_capabilities/clear_octomap_service_capability.cpp), but the service doesn't seem to do anything extra that would prevent this.

That sounds reasonable. My MoveIt workaround skips the FCL call if the root is null: https://github.com/ros-planning/moveit/pull/2104/files So far, at least, I've not hit any other segfaults with the octomap with...

A bit more background that I discovered while writing a MoveIt test for this today. Apparently this segfault only happens when doing a collision distance check after clearing the octomap....

Yes, please support ROS 2!

I was just trying to test this, but it's failing to build with a bunch of warnings about `moveit_resources`, here's the first: ``` WARNING: package "moveit_planners_ompl" should not depend on...

Thanks for the rebase. I'd like to see this merged--I think it'll be helpful in a project I'm working on. I built and ran it, and it seems to work...

Yeah, these were the options we considered: 1. Include a bag file somewhere with a point cloud in it, 1. Include the PCL code to generate the point cloud (and...

Thanks, @Dale-Koenig -- I didn't know about that, but it's probably the best way to do it. I'll convert it, and see if I can figure out this CI failure.

I'm using this for a project and I was surprised to see it's still not merged. @v4hn and @rhaschke could you please take a look at this PR?