Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[pr2_base_trajectory_action] remove dependency to pr2_controller_msgs #623

Merged
merged 4 commits into from
Oct 13, 2016

Conversation

furushchev
Copy link
Member

@furushchev furushchev commented Jun 22, 2016

What I did

  • refactor codes
  • remove pr2 related packages from dependencies
  • use control_msgs/FollowJointTrajectoryAction instead of pr2_controller_msgs
  • support publishing feedback message
  • Write test codes
  • Further Test on real robot / gazebo

@k-okada
Copy link
Member

k-okada commented Jun 29, 2016

did you finished

Write test codes
Further Test on real robot / gazebo
??

Cc: @chiwunau

@furushchev
Copy link
Member Author

@k-okada Sorry for late response. All has been done.

@furushchev
Copy link
Member Author

furushchev commented Jul 1, 2016

@k-okada If this is merged, we can remove pr2 related packages from robot-interface.l (not pr2-interface.l). Should we move robot-interface.l to jsk_roseus?
( https://github.com/jsk-ros-pkg/jsk_pr2eus/pull/237/files#diff-15f6920b924513f6027d36b361f85bbdL5 )

@furushchev furushchev changed the title [WIP] [pr2_base_trajectory_action] remove dependency to pr2_controller_msgs [pr2_base_trajectory_action] remove dependency to pr2_controller_msgs Jul 1, 2016
@furushchev
Copy link
Member Author

furushchev commented Jul 1, 2016

travis failed at test/pr2_base_trajectory_action.test.
I doubt that gazebo's base_controller does not work correctly. (at least different behavior from real robot)
@k-okada Can I disable gazebo test in this point?

@furushchev
Copy link
Member Author

@k-okada How do you think of it?

@k-okada
Copy link
Member

k-okada commented Jul 10, 2016

Why you have to disable gazebo test?

On Sunday, 10 July 2016, Furushchev [email protected] wrote:

@k-okada https://github.com/k-okada How do you think of it?


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#623 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAeG3KrWqYFjQ5QUKeCToMjKXnFKMA1Vks5qUN1fgaJpZM4I7YTc
.

◉ Kei Okada

@furushchev
Copy link
Member Author

@k-okada maybe gazebo's realtime controller does not follow the actual /cmd_vel (slower than command), and it always causes failure of following trajectory in this action server.
But I could not yet inspect codes of pr2_gazebo.

@furushchev furushchev closed this Jul 10, 2016
@furushchev furushchev reopened this Jul 10, 2016
@furushchev
Copy link
Member Author

oh, closed wrongly...

@k-okada
Copy link
Member

k-okada commented Jul 10, 2016

In that case you can send command with large error threshold

On Sunday, 10 July 2016, Furushchev [email protected] wrote:

oh, closed wrongly...


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#623 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAeG3HUyZLC7h4LuRAhJKjgGCblsfjvMks5qUObogaJpZM4I7YTc
.

◉ Kei Okada

@furushchev
Copy link
Member Author

@k-okada Yes, we can increase error threshold, but I think it actually does not test anything...just for check if publish /cmd_vel, we dont need gazebo..

@k-okada
Copy link
Member

k-okada commented Jul 11, 2016

even so, it is important to check the code actually publish something and
simulated robot receive them

◉ Kei Okada

On Sun, Jul 10, 2016 at 10:14 PM, Furushchev [email protected]
wrote:

@k-okada https://github.com/k-okada Yes, we can increase error
threshold, but I think it actually does not test anything...just for check
if publish /cmd_vel, we dont need gazebo..


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#623 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAeG3Nkgne8vnnVS4sNFKzAZK0OEa9YOks5qUPAugaJpZM4I7YTc
.

@furushchev
Copy link
Member Author

@k-okada OK. I added note about this problem to test code and enabled test with broader thresholds.

@furushchev furushchev force-pushed the pr2-base-traj branch 3 times, most recently from 7440014 to 1390b12 Compare July 14, 2016 14:31
@furushchev
Copy link
Member Author

please merge after #644

* refactor codes
* remove pr2 related packages from dependencies
* use control_msgs/FollowJointTrajectoryAction
* support feedback message
* added test codes
* some bugfixes of old codes
* checked on real robot
* update launch files
@furushchev
Copy link
Member Author

furushchev commented Aug 11, 2016

@k-okada I used pr2eus to test with gazebo, and it does not work until jsk-ros-pkg/jsk_pr2eus#237 is merged.
There are two choices.

  1. merge [pr2eus/robot-interface.l] use follow joint trajectory action jsk_pr2eus#237
  2. release pr2eus to pass test with USE_DEB=true on this repo.
  3. restart travis on this PR to test with new pr2eus

or.

  1. merge this PR with temporally disable gazebo test
  2. merge [pr2eus/robot-interface.l] use follow joint trajectory action jsk_pr2eus#237
  3. release pr2eus to pass test with USE_DEB=true on this repo.
  4. create new PR to enable gazebo test

how do you think of it? I recommend the latter.

@k-okada
Copy link
Member

k-okada commented Aug 12, 2016

ok, I'll merge jsk-ros-pkg/jsk_pr2eus#237, before that please explain more
detail, because this will be a big change

  1. which method has been changed, specially in robot-interface
  2. since robot-interface is used in other robots, specially rtmros robots,
    so why this change did not affect that robot, for example, I assume they
    have override the information (
    https://github.com/start-jsk/rtmros_common/blob/master/hrpsys_ros_bridge/euslisp/rtm-ros-robot-interface.l#L314),
    so please list up such case in all method described in 1), and then they
    will remove that code from rtm-ros-robot-interface in future.

Cc:@mmurooka

◉ Kei Okada

On Thu, Aug 11, 2016 at 3:42 PM, Furushchev [email protected]
wrote:

@k-okada https://github.com/k-okada I used pr2eus to test with gazebo,
and it does not work until jsk-ros-pkg/jsk_pr2eus#237
jsk-ros-pkg/jsk_pr2eus#237 is merged.
There are two choices.

  1. merge [pr2eus/robot-interface.l] use follow joint trajectory action jsk_pr2eus#237
    [pr2eus/robot-interface.l] use follow joint trajectory action jsk_pr2eus#237
  2. restart travis on this PR to test with new pr2eus

or.

  1. merge this PR with temporally disable gazebo test
  2. merge [pr2eus/robot-interface.l] use follow joint trajectory action jsk_pr2eus#237
    [pr2eus/robot-interface.l] use follow joint trajectory action jsk_pr2eus#237
  3. create new PR to enable gazebo test


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#623 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAeG3ISXH5IbxnsMcp40Sd-9utGV7fd9ks5qesRqgaJpZM4I7YTc
.

@furushchev
Copy link
Member Author

@k-okada Sorry for late reply.

  1. which method has been changed, specially in robot-interface

in robot-interface.l, I only changed to use control_msgs/FollowJointTrajectoryAction instead of pr2_controller_msgs/JointTrajectoryAction. And moved pr2_* messages to pr2-interface.l from robot-interface.l.
changed methods are :action-feedback-cb in controller-action-client class, :add-controller and:default-controller in robot-interface class and :init, :move-trajectory in robot-move-base-interface class.

  1. since robot-interface is used in other robots, specially rtmros robots,
    so why this change did not affect that robot, for example, I assume they
    have override the information (
    https://github.com/start-jsk/rtmros_common/blob/master/hrpsys_ros_bridge/euslisp/rtm-ros-robot-interface.l#L314),
    so please list up such case in all method described in 1), and then they
    will remove that code from rtm-ros-robot-interface in future.

Is there any robots that uses pr2_controller_msgs other than PR2?
In my thought, by this change, we will support wider variations of robot by using control_msgs, generalized messages, and removing pr2_* messages only can affect PR2 users.
And this PR removes dependency to pr2_ messages for even PR2.
(Sorry basically I could not understand your second question)

@k-okada
Copy link
Member

k-okada commented Aug 27, 2016

old robots uses pr2_controller_msgs, for example, hrp2
https://github.com/start-jsk/rtmros_common/blob/master/hrpsys_ros_bridge/src/HrpsysJointTrajectoryBridge.cpp#L223-L228
now they accept both pr2_controller_msgs and controller_msgs.
This is why @snozawa 's originall PR at pr2eus depneds on pr2 -> https://github.com/jsk-ros-pkg/jsk_pr2eus/pull/197/files#diff-15f6920b924513f6027d36b361f85bbdR1066,
but newer PR uses https://github.com/start-jsk/rtmros_common/pull/887/files#diff-d9351e39dede27bec86e5a1c49dcfaa0R312
So, I hope they now switched to controller_msgs, but also afraid that someone still using old message.

@furushchev
Copy link
Member Author

Why not merge this pull request? If there is problem, please complain.
Merging only jsk-ros-pkg/jsk_pr2eus#237 breaks pr2 unsafe moving.

@furushchev
Copy link
Member Author

@k-okada seems answered at jsk-ros-pkg/jsk_pr2eus#237

@k-okada
Copy link
Member

k-okada commented Oct 13, 2016

I think this PR is ok to merge, because it is only base_trajectry_action

@k-okada k-okada merged commit c48c304 into jsk-ros-pkg:master Oct 13, 2016
@furushchev furushchev deleted the pr2-base-traj branch October 13, 2016 15:06
@furushchev
Copy link
Member Author

@k-okada Thank you for review and merging!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants