Paul Bovbel

To that end, you may find the following functions useful: http://docs.ros.org/jade/api/costmap_2d/html/classcostmap__2d_1_1Costmap2DROS.html#a8e23eef773096a1fca61ae038220070d http://docs.ros.org/indigo/api/frontier_exploration/html/namespacefrontier__exploration.html#a9b90012ac183da448a95eec412936066

On and off, but any contributions are certainly welcome.

Hi folks, I apologise, I haven't had the time to bring the melodic branch up to snuff after a sizeable refactor - and hence never cut a melodic releaser. PRs...

I'm not really sure the context of the change, but contributions (in terms of a deeper investigation into the issue, and a potential fix) are very welcome!

I don't have any immediate plans - PRs to integrate the melodic-devel frontier_exploration with turtlebot3, or the existing demo in https://github.com/husky/husky/blob/melodic-devel/husky_navigation/launch/exploration_demo.launch are welcome!

I hadn't had the chance to clean up melodic-devel and the https://github.com/husky/husky/blob/melodic-devel/husky_navigation/launch/exploration_demo.launch demo for a Melodic release. Help with either is greatly appreciated!

Thanks for the report. It looks like there may have been a change upstream in the costmap_2d package that's issuing this warning. I'll try to have a look at this...

There's a two 'features' at play here, both existing to enable frontier search even when the 'initial' search cell is not strictly FREE_SPACE. https://github.com/paulbovbel/frontier_exploration/blob/lunar-devel/src/frontier_search.cpp#L47 tries to find a clear cell...

It's not a memory leak, but an error on shutdown in terms of destruction order.

Thanks @wxmerkt, looks like that's a valid option.