Simplify and combine docker images

Adjuvo / senseglove_ros_ws

Implementation of the standard SenseGlove API into an example ros_workspace. The workspace features a few publishing topics where users can find glove_pose positions, fingertip_positions, glove_angles, etc. Integration with ROS Control for Force Feedback and Vibration.

7 stars 7 forks source link

Simplify and combine docker images #18

Closed robinkrens closed 2 years ago

robinkrens commented 2 years ago

Dockerfile based on osfr/ros image, instead of nvidia cuda (no need to apt-get all ros packages, saves quite some time)
Bash script to create noetic and melodic docker images
Merge of various docker files (I don't think special windows images are needed)

Hardware passthrough X11 acceleration untested due to lack of nvidia gpu on current setup @Deakensbubbles can you check this?

RPKrijnen commented 2 years ago

Your docker files do not pull the current workspaces from github. Should this not also be added in the dockerfile to ensure that the users have the most current version of the repo?

robinkrens commented 2 years ago

For developers it is normally better to have the repository outside of the container, binding the directory with -v:

-v Create a bind mount. If you specify, -v /HOST-DIR:/CONTAINER-DIR, Docker
          bind mounts /HOST-DIR in the host to /CONTAINER-DIR in the Docker
          container.

That way, you can use host dir tooling.

robinkrens commented 2 years ago

Agreed on the run bash script that it is useful, but since you probably need to edit it, I just put it in the readme.

RPKrijnen commented 2 years ago

I noticed that when trying to run standard ros commands the setup.bash file is inaccessible. you're able to make use of ros commands when you use a clean environment and therefore you need to first 'catkin clean' your workspace and 'catkin_make' or 'catkin build' the workspace from within the container. You could add this disclaimer in the README.