Closed DaniGarciaLopez closed 8 months ago
Thanks for the review @rhaschke, changes addressed!
I noticed that the remaining functions also use the argument names from
and to
. For this reason, I reverted the requested change. Sorry for the confusion.
As this change should become part of master branch too, could you please (squash)rebase onto and retarget to master? Thanks.
I also personally believe that using start
and end
is more intuitive. However, to maintain consistency, IMHO sticking with from
and to
seems to be the clearer option.
Commits squashed!
Not sure how to retarget to master with that many commits ahead. Can you retarget the PR or should I open a new one?
I retargeted the PR to master, but you need to rebase onto master. I don't have permissions to force-push your branch.
Ok, thanks. Done!
Attention: Patch coverage is 78.57143%
with 3 lines
in your changes are missing coverage. Please review.
Project coverage is 58.82%. Comparing base (
55c4b52
) to head (faa8486
).
Files | Patch % | Lines |
---|---|---|
core/src/stage.cpp | 78.58% | 3 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Releated to https://github.com/ros-planning/moveit_task_constructor/issues/543. This PR enables the possibility to set independent start and end states when calling the
spawn
method of a Generator stage, while maintaining the previous functionality that uses the same state.I didn't implement this overload as it implies that the trajectory will be empty and it makes no sense to set different start and end states when the robot is not actually moving, right? Or am I missing any specific use case where this could be handy?
I noticed that
connect
uses const lvalue reference (const InterfaceState&
) instead of rvalue (InterfaceState&&
). Are there any design preferences?Is there an easy way to backport this to humble branch or I should cherry-pick another PR?