Closed marc-hanheide closed 4 years ago
Hmm... this wasn't me. Perhaps @nilsbore remembers?
This was me. The functionality it provides is always looking at the default_tolerance
parameter before planning. This would allow us to plan to influence areas even if there was an obstacle in the middle. We need it to be dynamically checked before planning because the tolerance for intermediary nodes should be greater than 0 (e.g. 1m) but for target nodes it should be 0.0
Commit below shows the main difference:
https://github.com/strands-project/strands_movebase/commit/cf3b8a3aac4e13221e9227992d56063f97438133
Thanks @bfalacerda I wonder if this is indeed still needed. It's a bit unfortunate this hasn't been a fork of the original, difficult to merge from upstream in this case, I reckon.
yeah, it wasn't the best way to do it... I'd say ignore this package and if we end up needing it we do the fork and re-do the change, it's easy enough.
I think the only thing that need to be changed to stop using this package is https://github.com/strands-project/strands_movebase/blob/indigo-devel/strands_movebase/strands_movebase_params/move_base_params.yaml#L1
see #74
FYI: our melodic release is slowly taking shape... https://github.com/LCAS/rosdistro/wiki/repos-melodic in case @hawesie or @bfalacerda want to update Bob and/or Betty.
as https://lcas.lincoln.ac.uk/buildfarm/job/Mdev__strands_movebase__ubuntu_bionic_amd64/ shows,
strands_navfn
refuses to compile for ROS melodic.The problem is a lot of API changes in
move_base
: https://github.com/paulbovbel/frontier_exploration/issues/38#issuecomment-449415389So, question: What does
strands_navfn
offer that the standardnavfn
doesn't? Can it be removed?