Gazebo Splashscreen up forever if plugin blocks during WorldUpdateBegin

gazebosim / gazebo-classic

Gazebo classic. For the latest version, see https://github.com/gazebosim/gz-sim

http://classic.gazebosim.org/

Other

1.16k stars 476 forks source link

Gazebo Splashscreen up forever if plugin blocks during WorldUpdateBegin #2224

Open osrf-migration opened 7 years ago

osrf-migration commented 7 years ago

Original report (archived issue) by Shane Loretz (Bitbucket: Shane Loretz, GitHub: sloretz).

The original report had attachments: CMakeLists.txt, synchronized_model.world, TestModelPlugin.cpp

The gazebo GUI splash screen never goes away if a plugin blocks WorldUpdateBegin. This is used to synchronize gazebo with another tool. The plugin blocks until it receives a signal from the other tool that it is ready for the next timestep.

I've attached a model plugin and world that reproduces the behavior.

Use cmake
run make
from the build directory GAZEBO_PLUGIN_PATH=$(pwd) gazebo --verbose path/to/src/synchronized_model.world

There are workarounds to this issue

If the plugin allows the world to step once, the splash screen will go away. With the attached plugin type "asdf[enter]" to step the world once, and the splash screen will disappear
If the world is started paused --pause the splash screen will go away

osrf-migration commented 7 years ago

Original comment by Louise Poubel (Bitbucket: chapulina, GitHub: chapulina).

I'm not so sure this is a bug in Gazebo, or just something that should be properly handled by the plugin developer.

osrf-migration commented 7 years ago

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

It is difficult for Gazebo to prevent a plugin from doing something bad, like blocking indefinitely. The plugins we develop and release should behave correctly, and we can help other people developing their own plugins. But we can't easily prevent a plugin from blocking, segfaulting, etc.

osrf-migration commented 7 years ago

Original comment by Shane Loretz (Bitbucket: Shane Loretz, GitHub: sloretz).

This could be closed as "wontfix".

The troublesome part is we've recommended blocking during WorldUpdateBegin as a way to synchronize gazebo with another program. I believe it's being done here.

The ~/world_control topic is a way to run the world one step at a time, but someone reported the world would unexpectedly unpause. That could be a separate issue if I can reproduce it.

osrf-migration commented 7 years ago

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

It does look like uctf is blocking during WorldUpdateBegin, via the ProcessModelMsgs function:

https://github.com/osrf/uctf/blob/master/plugins/SASCPlugin.cc#L134-L219