Collision WorldPose does not get updated.

[osrf-migration]

Original report (archived issue) by Andrei Haidu (Bitbucket: ahaidu).

---

During simulation Collision::GetWolrdPose does not get updated.

Here is a testing piece of code:

#!c++
		physics::Link_V _links = this->world->GetModel("box")->GetLinks();

		for (unsigned int i = 0; i < _links.size(); i++)
		{
			std::cout << "link name: " << _links.at(i)->GetName() << " pose: " << _links.at(i)->GetWorldPose().pos << std::endl;

			physics::Collision_V _colls = _links.at(i)->GetCollisions();

			for (unsigned int i = 0; i < _colls.size(); i++)
			{
				std::cout << "coll name: " << _colls.at(i)->GetName() << " pose: " << _colls.at(i)->GetWorldPose().pos << std::endl;
			}
		}

During the simulation while the object is moving the link poses change, and the collisions maintain their original position.

Testing works with a simple box model.

[osrf-migration]

Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

Does the model have multiple links? I'm asking because I see different Entity::SetWorldPose* functions used for models, canonical links, and everything else
(see end of Entity::Load).

So I'm wondering if SetWorldPoseDefault is broken, in which case non-canonical links should have a static world pose as well.

[osrf-migration]

Original comment by Andrei Haidu (Bitbucket: ahaidu).

---

I first noticed it with a model with multiple links, and then tested it as well on a model with one link one collision. Both had the same issue.

[osrf-migration]

Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

I can confirm that Collision::GetWorldPose() is not returning an updated value.

Maybe
Entity::UpdatePhysicsPose
is where it should happen?
There would be a performance impact, though, to cache the pose of all collisions and visuals in a model.

@nkoenig what do you recommend?

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

I think the Entity::UpdatePhysicsPose is broken. I'll try writing a test, and then a solution.

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

See pull request #1049

[osrf-migration]

Original comment by Jonathan Bohren (Bitbucket: jbohren).

---

I just ran into this, too.

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

• changed state from "new" to "resolved"

Resolved in pull request #1049

[osrf-migration]

Original comment by Andrei Haidu (Bitbucket: ahaidu).

---

The issue still appears for models with multiple links. I have installed gazebo from default (currently version 4.0.0).

I am using for testing a model with 2 links. First link has 2 collision, second link 1 collision:

• model: model_box
• • link1: link_green_blue
• • • collision1: collision_green
• • • collision2: collision_blue
• • link2: link_orange
• • • collision1: collision_orange

The second collision doesn't get updated. Only when manually moving it with the mouse, and then only for one iteration.

Testing code:

#!c++
	// DEBUG collision issue
	physics::ModelPtr box = this->world->GetModel("box");

	std::cout << "model name: '" << box2->GetName()
			<< "' --> world pos: " << box2->GetWorldPose().pos << std::endl;

	// get all links from the model
	physics::Link_V _links = box2->GetLinks();

	// iterate through all the links
	for (physics::Link_V::iterator l_iter = _links.begin();
			l_iter != _links.end(); ++l_iter)
	{
		// get all collisions of the link
		physics::Collision_V _collisions = l_iter->get()->GetCollisions();

		std::cout << "\t link name: '" << l_iter->get()->GetName()
				<< "' --> world pos: " << l_iter->get()->GetWorldPose().pos << std::endl;

		// iterate through all the collisions
		for (physics::Collision_V::iterator c_iter = _collisions.begin();
				c_iter != _collisions.end(); ++c_iter)
		{
			std::cout << "\t\t collision name: '" << c_iter->get()->GetName()
					<< "' --> world pos: " << c_iter->get()->GetWorldPose().pos << std::endl;
		}
	}

Screenshot with the terminal output:

[osrf-migration]

Original comment by Andrei Haidu (Bitbucket: ahaidu).

---

• changed state from "resolved" to "open"

[osrf-migration]

Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

Can you suggest a patch for our regression test? That will make it easier to reproduce. Thanks for the report.

[osrf-migration]

Original comment by John Hsu (Bitbucket: hsu, GitHub: hsu).

---

and reuse issue_1124 branch if possible

Your test world will go into this world folder

[osrf-migration]

Original comment by Roxana Leontie (Bitbucket: rleontie).

---

Issue #1415 was marked as a duplicate of this issue.

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

• set version to "all"

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

• changed state from "open" to "new"