Workspace Layout¶
Layout of the public deploy workspace, from the open-deploy-ws README. Top-level scripts and config files are those listed in that README (./init_repo.sh, submodules_visibility.conf, deb_versions.conf, and scripts/).
open-deploy-ws¶
Source of truth
Tree and init flow: README.EN.md. The repo root has init_repo.sh, submodules_visibility.conf, deb_versions.conf, and scripts/ — there is no workspace setup.bash / env installer besides ./init_repo.sh.
open-deploy-ws/
├── src/
│ ├── arms_ros2_control/ # controllers / commands / hardware interfaces / shared libs
│ ├── robot-descriptions/ # common / manipulator / humanoid (hyphen)
│ └── ocs2_ros2/ # only if that module is installed as source
├── init_repo.sh
├── submodules_visibility.conf
├── deb_versions.conf
└── scripts/ # install_core_debs.sh, uninstall_core_debs.sh
Primary path: ./init_repo.sh (visibility + per-module d/s). Do not git submodule update --init --recursive. After init, colcon build (the README does not wrap that). Then source install/setup.bash in that workspace — standard ROS 2 overlay, not a FiveAges env script.
Lean robot branches (dobot-cr5, arx-acone, plus arx-lift2s / panthera-ht used in the how-tos) clone only that scheme’s packages. See the same README.
arms_ros2_control (inside src/)¶
From the arms_ros2_control README:
arms_ros2_control/
├── controller/
│ ├── ocs2_arm_controller/
│ ├── ocs2_wbc_controller/ # private nested (visibility conf)
│ ├── adaptive_gripper_controller/
│ └── basic_joint_controller/ # own package README; Home / Hold / MoveJ
├── libraries/ # ocs2_humanoid (private), …
├── hardwares/ # topic_based_ros2_control, unitree_ros2_control, …
└── command/ # msgs, RViz plugin, arms_target_manager, arms_teleop
Nested public/private lines: Submodules Visibility.
fa-deploy-ws¶
Internal workspace (private). Tree and extra flags are in that repository’s README after you have access. Public users stay on open-deploy-ws.