ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

SuperElectric's profile - activity

2014-09-29 16:41:03 -0500 received badge  Supporter (source)
2013-12-08 08:08:12 -0500 received badge  Famous Question (source)
2013-11-17 23:47:34 -0500 received badge  Notable Question (source)
2013-11-17 23:47:34 -0500 received badge  Popular Question (source)
2013-03-27 22:00:02 -0500 received badge  Famous Question (source)
2013-03-12 12:24:40 -0500 received badge  Notable Question (source)
2013-02-07 10:03:24 -0500 received badge  Notable Question (source)
2012-12-30 14:58:56 -0500 received badge  Popular Question (source)
2012-12-20 12:09:47 -0500 asked a question rviz doesn't start when run remotely, throws GLXUnsupportedPrivateRequest error

This is similar to http://answers.ros.org/question/10588/rviz-does-not-start-prints-glxunsupportedprivaterequest/ (this question), in that when I run rviz, I get the following error:

Xlib: extension "NV-GLX" missing on display "localhost:10.0". The program 'rviz' received an X Window System error. This probably reflects a bug in the program. The error was 'GLXUnsupportedPrivateRequest'.
(Details: serial 20 error_code 176 request_code 154 minor_code 16)

In my case, this happens only when I'm running rviz on my desktop while connected to it remotely, with X forwarding, using ssh -Y. When I'm at the desktop in person, rviz works just fine.

A bit more info:

I'm actually connecting to my desktop from an Ubuntu virtual machine running on a Windows laptop (I know, it's horribly baroque, but it's the only way I could get linux running on my company laptop).

This virtual machine doesn't have access to the laptop's graphics card, so it can't run rviz locally. However, I thought that it would still be able to run rviz on a remote machine, since the remote machine would do the rendering, and the laptop would just display the rendered screen images that it receives over X. Was I wrong?

2012-12-09 13:27:43 -0500 asked a question How can I render many objects in rviz without slowing down to a halt? (or, how do I switch scene managers?)

I would like to use rviz to display a large SLAM graph with tens of thousands of nodes and edges. When I do so by using LINE_ARRAY markers to render edges, rviz grinds to a stop for any graph with more than a handful (< 1000?) of edges.

I believe this may be because rviz uses OGRE's default scene manager, which is bad at doing culling on flat scene hierarchies:

From Ogre3D's wiki:(http://www.ogre3d.org/tikiwiki/SceneManagersFAQ#Generic_Scene_Manager_default_)

The default generic manager (the one that's built in) isn't that good. It can do hierarchal frustum culling, but relies on you to build the hierarchy. If you add 1000 nodes with entities to the root scene node, each has to be tested for culling every frame.

I would therefore like to use some other scene manager, such as the octree scene manager (see above link). How do I tell rviz to switch scene managers?

2012-11-28 04:32:41 -0500 received badge  Famous Question (source)
2012-11-26 07:15:22 -0500 commented answer Is there a way to get ROS_INFO, ROS_ERROR, etc to output line numbers?

Hmm; this isn't working for me. I'm sure I could get it working, but before I invest energy in that, I'd like to revisit the original question: is there any way to get ROS_INFO to print the line number (without need of external tools like rxconsole)?

2012-11-06 05:59:09 -0500 received badge  Notable Question (source)
2012-10-26 09:58:42 -0500 received badge  Popular Question (source)
2012-10-25 11:28:58 -0500 asked a question Is there a way to get ROS_INFO, ROS_ERROR, etc to output line numbers?

The title says it all. Is there a flag somewhere I can set to get the ROS_INFO/ROS_ERROR/... macros to output line numbers, or is that something I'm going to have to feed to ROS_INFO myself by using __LINE__?

2012-10-25 03:07:26 -0500 received badge  Popular Question (source)
2012-10-24 14:22:04 -0500 received badge  Scholar (source)
2012-10-24 10:40:47 -0500 received badge  Student (source)
2012-10-24 06:59:25 -0500 received badge  Editor (source)
2012-10-24 06:50:01 -0500 asked a question How safe is it to directly modify message member vectors in c++?

I have a custom message class A that contains vectors (of other message classes, but I think that's immaterial).

Is it safe to mess around with the vectors directly, by resizing them, inserting things into them, etc?

My concern is that A's header might contain some metadata about A's vector, which becomes out-of-date when I modify A's vector.

-- Matt