tlinder's profile - activity

Hi all,

recently i noticed that rxgraph has some problems and produces the following error message:

 File "/opt/ros/fuerte/lib/python2.7/dist-packages/rxgraph/impl.py", line 125, in run
    info_text = get_info_text(self.selection_url)
  File "/opt/ros/fuerte/lib/python2.7/dist-packages/rxgraph/impl.py", line 59, in get_info_text
    return rosnode.get_node_info_description(node_name) + rosnode.get_node_connection_info_description(node_api)
TypeError: get_node_connection_info_description() takes exactly 2 arguments (1 given)

This Error is new to me and appeared first time after a ROS updated (ca. 10'th of July). After this error rxgraph is not updatimng any longer the screen, i.e. you can still see the old graph displayed but now response at all to updates on the ROS-Network-Graph.

Here my question: Does any one else have this issue and if yes which OS and ROS version are involved?

As written in the title for me this is Ubuntu 12.04 64Bit fuerte.

Thank you Lorenz, you pointed on a really important point. OS`s which do not support VMs will not be used much for professional development. From my point of view Mac OSX will never be really in focus of ROS as there is even no embedded Mac hardware which you can use for building a commercial robot.

Hi, can you ping the localhost ? If not check your Hosts-file in /etc/hosts and may reconfigure it to default. Also it is not realy needed to make use of the ROS_HOSTNAME settings, so may do not use this command for the next time.

So i assume that you started Eclipse from an Icon/GUI and not from the terminal. The Problem here is that the environment Variables of the .bashrc are not in use. Therefore you need to execute also Steep 5 or just start Eclipse from a Terminal instead. If you start it from terminal the environment Variables are in use and the packages can be found (of course assuming that your ROS_PACKAGE_PATH contains them, can be checked by roscd + A_PACKAGE or "echo $ROS_PACKAGE_PATH")

The first time looks like your current local time while the second on looks like the time recorded within the bag-file

Also if it is true that sensor_msg::PointCloud is easier to understand, it is still not any longer standard. PointCloud2 is the current standard in ROS. To use the PCL is of course an option, but it takes more load and computational power if you need to convert a lot in between ROS and PCL.

Hi all,

currently I have installed a fresh Ubuntu 12.04 LTS 6 4Bit and ROS fuerte. Unfortunately I noticed several problems by using it. The most weirdest issue is given by executing the roscore which is prompting python failure messages "AttributeError("'_DummyThread' object has no attribute '_Thread__block'",)".

So if I just start the core by :

roscore

the system prompts following results:

... logging to /home/tlinder/.ros/log/c84ccda0-bc8c-11e1-986c-0024d76d2f84/roslaunch-fhg-thorsten-2827.log
Checking log directory for disk usage. This may take awhile.
Press Ctrl-C to interrupt
Done checking log file disk usage. Usage is <1GB.

started roslaunch server http://localhost:52473/
ros_comm version <unversioned>

SUMMARY
========

PARAMETERS
 * /rosdistro
 * /rosversion

NODES

auto-starting new master
Exception AttributeError: AttributeError("'_DummyThread' object has no attribute '_Thread__block'",) in <module 'threading' from '/usr/lib/python2.7/threading.pyc'> ignored
process[master]: started with pid [2852]
ROS_MASTER_URI=http://localhost:11311/

setting /run_id to c84ccda0-bc8c-11e1-986c-0024d76d2f84
Exception AttributeError: AttributeError("'_DummyThread' object has no attribute '_Thread__block'",) in <module 'threading' from '/usr/lib/python2.7/threading.pyc'> ignored
process[rosout-1]: started with pid [2865]
started core service [/rosout]

So as you can see, some python components are crashing and producing this errors:

Exception AttributeError: AttributeError("'_DummyThread' object has no attribute '_Thread__block'",) in <module 'threading' from '/usr/lib/python2.7/threading.pyc'> ignored

Edit1:

It seems to be a known bug of python 2.7 which is still open.

Take a look in the migration web page here.

Eigen is in fuerte differently used then in prior versions:

Starting with Fuerte, the new pattern to find Eigen in a CMakeLists.txt file is:

find_package(Eigen REQUIRED)

include_directories(${EIGEN_INCLUDE_DIRS})

add_definitions(${EIGEN_DEFINITIONS})

yes it is possible and quit common. Just check this link and you will find a description of the data type PointCloud2.

Hi Benoit, in cases were you migrated (e.g. from diamond/electric to fuerte) you need do ensure that your CMakeFiles/Caches and co are removed. They contain paths to the used ROS-Version, i.e. paths to older versions which are not any longer available on your system.

Try rosmake --pre-clean or even better inside the package in question try rm -r CMakeFiles cmake_install.cmake CMakeCache.txt && rosmake

Well i do not know what is causing this problem but i can give you some hints which i us if rviz fails.

1. If rviz was once working and is not any longer then something is wrong with the config file within your home folder. Delete the config file in /home/USERNAME/.rviz will help.

2. IF rviz fails after software updated, then it can be that the graphi card driver needs to be reinstalled once more

Hope this helps

PTAM can work for both, indoor and outdoor.

Ptam from the ETH stack is may an option for you

Well Hector is able to work without any odometry support. In such a case Hector uses the laser data only to calc. its location in respect to the map. As long as the laser data are meaning full hector will work well in this mode, i.e. the range, noise, updated rate and vehicle speed stay within some bounds. For example, the hokuyo URG-04LX has a range of 4-7m (if i remember correctly) which means your environment should not have spots larger the 4-7m free space. As rule of the thump you can say that at least 50% of the beams should stay within half of the max range. Therefore your outdoor conditions are easily violating this rule of the thump.

Hope this helps.

"Should I be looking for a different virtual machine? " Well may but maybe it is just better to install a real Linux. You will always have Problems like such which you need to solve if you use a VM.

If you have this information (location, frame and frame relation) than you can use the code above.

Well i am not really familiar with Yaml Maps. Can you giv me an example of one or two objects within that format? I assumed that you have the objects, there location and if the location is not already within the map frame the knowledge of how the object-frame is linked to the map-frame.

As far as i understand you you need only to read from each object its location. As you staid you have the location within the yaml map. So read it with a node and then create for each on a TF object which you publish. So, somethink like this for each object:

  static tf::TransformBroadcaster br;
  tf::Transform transform;
  transform.setOrigin( tf::Vector3(object->x, object->y, object->z) );
  transform.setRotation( tf::Quaternion(object->rot->x, object->rot->y, object->rot->z) );
  br.sendTransform(tf::StampedTransform(transform, ros::Time::now(), "world", object->name));

In cases where your object coordinates are not within the world/map frame you may need to change also the "world" label.

I hope this helps a little.

P.S.: Check this tutorials

Double Posting! pls. Follow this Topic kinect_calibration

nice question but your topic is not really self explaining. Next time (or if possible edit also this question) use a titel which highlights the related Problem to attract the right experts. Regarding your question i do not know and will follow this topic to get enlightened ;-)