`start_traj_mode` fails if home position is not configured about motoros2

[gavanderhoorn]

Note: this is also an issue with MotoROS1.

[ted-miller]

We need to document this error in the readme.
#201 (comment)

They should be seeing a message on the pendant

When attempting to start a job (using the pendant), they should get error 0380 Check Position

[gavanderhoorn]

Should we instead not have MR2 check whether the home pos is set and raise an appropriate alarm if it isn't?

---

Edit: or at the very least update the result message returned by start_traj_mode et al.?

[ted-miller]

It looks like we already are

motoros2/src/MotionControl.c

Line 1514 in 9026a1d

Ros_ErrorHandling_ErrNo_ToString(rData.err_no), rData.err_no);

So that's apparently broken.

[gavanderhoorn]

You mean we should be seeing:

motoros2/src/ErrorHandling.c

Line 27 in 9026a1d

case 0x3040: return "The home position is not registered";

?

[ted-miller]

Yeah. That's what I would have thought.

Though perhaps 0x3040 doesn't apply if the home position has been registered and just needs to be validated. If that's the case, then I don't know of any mechanism to detect that this scenario.

[gavanderhoorn]

Just ran into this when testing something for #19:

[...] [...]: Ros_MotionControl_StartMotionMode: enter
[...] [...]: Can't start 'INIT_ROS' because: 'The home position is not registered' (0x3040)
[...] [...]: Ros_MotionControl_StartMotionMode: exit
[...] [...]: Ros_ServiceStartPointQueueMode_Trigger: The INIT_ROS job has not started. Please call start_traj_mode or start_point_queue_mode service (108)

seems this isn't completely broken.

[yai-rosejo]

I started looking into this issue. I found that the debug statement is occurring at:

motoros2/src/MotionControl.c

Lines 1503 to 1516 in 10d8f00

	// Start Job
	bzero(&rData, sizeof(rData));
	bzero(&sStartData, sizeof(sStartData));
	sStartData.sTaskNo = 0;
	strncpy(sStartData.cJobName, g_nodeConfigSettings.inform_job_name, MAX_JOB_NAME_LEN);
	ret = mpStartJob(&sStartData, &rData);
	if( (ret != 0) || (rData.err_no !=0) )
	{
	//TODO(gavanderhoorn): special check for "job is not loaded"
	Ros_Debug_BroadcastMsg(
	"Can't start '%s' because: '%s' (0x%04X)", g_nodeConfigSettings.inform_job_name,
	Ros_ErrorHandling_ErrNo_ToString(rData.err_no), rData.err_no);
	goto updateStatus;
	}

The r.Data is not getting out of the BOOL Ros_MotionControl_StartMotionMode function.

This then gets put out into a failure to start in Ros_ServiceStartTrajMode_Trigger:

motoros2/src/ServiceStartTrajMode.c

Lines 56 to 78 in 10d8f00

	if (!Ros_MotionControl_StartMotionMode(MOTION_MODE_TRAJECTORY))
	{
	// update response
	response->result_code.value = Ros_Controller_GetNotReadySubcode();
	if (response->result_code.value == MOTION_READY || Ros_MotionControl_IsMotionMode_PointQueue() || Ros_MotionControl_IsMotionMode_RawStreaming())
	{
	//Motion is ready, but the StartTrajMode service failed
	//because it's already in a different mode.
	response->result_code.value = MOTION_NOT_READY_OTHER_TRAJ_MODE_ACTIVE;
	rosidl_runtime_c__String__assign(&response->message,
	"Another motion mode is already active. Please call 'stop_traj_mode' service and try again.");
	}
	else
	{
	// map to human readable string
	rosidl_runtime_c__String__assign(&response->message,
	Ros_ErrorHandling_MotionNotReadyCode_ToString((MotionNotReadyCode)response->result_code.value));
	}
	Ros_Debug_BroadcastMsg("%s: %s (%d)", __func__,
	response->message.data, response->result_code.value);
	}

Which then calls Ros_Controller_GetNotReadySubcode(), this function doesn't see the warning only that the job has not started via:

motoros2/src/ControllerStatusIO.c

Lines 447 to 449 in 10d8f00

	// Check operating
	if(!Ros_Controller_IsOperating())
	return MOTION_NOT_READY_NOT_STARTED;

[yai-rosejo]

This same problem also occurs in Issue #192 I believe