You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Some member functions of physics::Link set update functions for the link’s sdf (ex. https://github.com/osrf/gazebo/blob/63cce1db38286ba373cfb5e54c2a58dfb4d42501/gazebo/physics/Link.cc#L159 ). And these update functions remains after the link has been removed by physics::Model::RemoveChild() or others. If these zombie update functions are called by Update() of a parent model’s sdf or others, a runtime error or segmentation fault occur.
In my opinion, physics::Link::Fini() should make the final call of Update() to write link’s final states to the sdf, and then detach update functions from the sdf.
Finally, this is an issue not only for physics::Link, but also for all classes which set update functions (as of now, Battery, Inertial, Link and Model).
The text was updated successfully, but these errors were encountered:
Original report (archived issue) by Yoshito Okada (Bitbucket: yoshito-okada).
Some member functions of physics::Link set update functions for the link’s sdf (ex. https://github.com/osrf/gazebo/blob/63cce1db38286ba373cfb5e54c2a58dfb4d42501/gazebo/physics/Link.cc#L159 ). And these update functions remains after the link has been removed by physics::Model::RemoveChild() or others. If these zombie update functions are called by Update() of a parent model’s sdf or others, a runtime error or segmentation fault occur.
In my opinion, physics::Link::Fini() should make the final call of Update() to write link’s final states to the sdf, and then detach update functions from the sdf.
Finally, this is an issue not only for physics::Link, but also for all classes which set update functions (as of now, Battery, Inertial, Link and Model).
The text was updated successfully, but these errors were encountered: