-
Notifications
You must be signed in to change notification settings - Fork 173
Description
I went through all commits pushed onto the ros2 branch in recent months from PickNik's side and cleaned them up to the standards of this repo.
Many of them I moved to the master branch and eventually merged them back into the ros2 branch.
Using merging instead of cherry-picking, one can more easily see which commits are on which branch.
Attention: I force-pushed the ros2 branch to the new, cleaned version.
Please update your local workspaces, resetting your local ros2 branch: git remote update; git reset --hard origin/ros2.
Only a few commits, whose value I still consider questionable, are left on the ros2 branch. These include:
- Enable parallel planning with PipelinePlanner #450 (I think a dedicated solver is more suitable in the context of MTC)
- Add planner name to trajectory info #490 (this info is not displayed anywhere, integrating it into the comment makes more sense - which was realized in Add planner info to MTC solution comment. #523)
Some others are ros2-specific, for example:
- Run goalcallback asynchronously #496 (only relevant for the ros2 action server)
- Update to the more recent JumpThreshold API #506
For now, I kept the old ros2 branch with name ros2-old. I will delete this branch in a few weeks.
The new branch should™ include all functionality of the old one as well as all improvements from the master branch.
See here for all differences. To highlight my changes even more, I created a cleanup branch, where individual changes can be seen more easily.
In the new ros2 branch, those cleanups are squash-merged into the corresponding main commits, though.
Comparing PickNik's ros2 branch with the old ros2 branch, apart from many missing commits, I noticed larger deviations in ExecuteTaskSolutionCapability. So far, I don't understand, why you, @henningkayser, refactored this code so much. In any case, I think you should consider switching to the new ros2 branch.