[WIP] Tell getting new goal from cancel

[pazeshun]

Not tested in real robot yet.
Not tested in real robot yet.

This PR solves the issue written in start-jsk#765 (comment) (this issue is specific to HIRONX as start-jsk#765 (comment)).
With this PR, clearOfGroup or clearJointAnglesOfGroup is not called when new goal comes, while that is called when cancel topic comes.
Though this PR is not necessary for robots with new hrpsys (e.g., HRP2JSKNTS), this PR makes start-jsk#765 more similar behavior to current master branch (difference appears only when cancel topic comes).
I think this is more acceptable.

[pazeshun]

Tested in HIRONX with rebasing on origin/master (https://github.com/pazeshun/rtmros_common/tree/add-cancel-smooth-master)

[pazeshun]

Tested in HRP2JSKNTS

[Affonso-Gui]

@k-okada We have been using this without problems in Hiro, and it was also tested in HRP2. Can we merge it?

[k-okada]

could you add test code to catch this? for example send cancel to the actionlib while monitoring /joint_states and check the velocity. https://github.com/start-jsk/rtmros_common/blob/master/hrpsys_ros_bridge/test/test-samplerobot.py#L151 -- ◉ Kei Okada 2019年9月3日(火) 13:24 Affonso, Guilherme <[email protected]>:

…

@k-okada We have been using this without problems in Hiro, and it was also tested in HRP2. Can we merge it? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[pazeshun]

could you add test code to catch this?

@k-okada I created start-jsk#1070 to test cancelling and overwriting.
In terms of overwriting test, however, it always passes due to start-jsk#765 (comment).
I'll explain detail in start-jsk#1070.

for example send cancel to the actionlib while monitoring /joint_states and check the velocity.

In fact, we cannot get velocity without start-jsk#1069.
At start-jsk#1070, I use position only and that is enough.

[pazeshun]

@k-okada I added one commit (8881caa) to this PR because I thought clear() and clearJointAngles() is swapped in one place.
Is that fix correct for you?

The reason why we couldn't notice it until now is that the wrong part is in fullbody commander of HrpsysJointTrajectoryBridge and that commander is not used.
Instead, fullbody command from roseus is passed to commander of HrpsysSeqStateROSBridge.

[pazeshun]

@Affonso-Gui I created start-jsk#1071. Please help me to test it next week.

[pazeshun]

The contents of this PR were merged to https://github.com/start-jsk/rtmros_common/tree/master via start-jsk#1071