I am attempting to run simple_demo.launch as described in the README, but it fails with:
[ INFO] [1717669508.068559416]: Elevation mapping node initializing ...
[ INFO] [1717669509.069437459]: Done.
[ WARN] [1717669509.072226345]: UPDATE GLOBAL MAP
ERROR: Wrong path to settings
[elevation_mapping_0-1] process has died [pid 270, exit code -11, cmd /catkin_ws/devel/lib/elevation_mapping/elevation_mapping /voxel_grid/output:=/robot0/pointcloud /stereo_grey/left/image_raw:=/robot0/image_rect __name:=elevation_mapping_0 __log:=/root/.ros/log/0a2013fa-23ef-11ef-a15e-0242ac110002/elevation_mapping_0-1.log].
I don't think I configured any path to any settings, nor should I need to, since the example should work out of the box.
The error message should give information about the path (if any) which it received, and why it is "wrong": Does the path not exist? Does the path reference a folder, rather than a file, or vice-versa? Was some other specific path expected? Was the wrong extension given? Was there a problem with the file itself, such as encoding or format? The error message "Wrong path to settings" simply doesn't give any information which a user could use to attempt to troubleshoot or resolve the issue. So this error message needs to be improved.
If a path to some settings needs to be configured, then this should be documented in the README, and isn't.
Since the example is meant to serve as documentation of how to get GEM working with my robot, not being able to get it running blocks me from using GEM at all.
I am using the following Dockerfile which I wrote (the base image requires the NVIDIA Container Toolkit and its tag should be adjusted based on your host CUDA version:
I am attempting to run
simple_demo.launch
as described in the README, but it fails with:Since the example is meant to serve as documentation of how to get GEM working with my robot, not being able to get it running blocks me from using GEM at all.
I am using the following
Dockerfile
which I wrote (the base image requires the NVIDIA Container Toolkit and its tag should be adjusted based on your host CUDA version:... and I made the following shell script to attempt to run the
simple_demo.launch
, which assumes that:test.bag
exists in the current working directory.The full output of the final
docker exec
command is:I don't think the libGL errors are related.
rviz
still works despite them and does show the trajectory, but not the elevation map, of course.