jespestana's profile - activity

I think so too. Thanks. Actually, I just checked it. And yes, you get the same address on both sides of the message tra

I think so too. Thanks.

Thanks for the in-depth answer! I have a follow-up comment/question about question 3: would it not be possible to chec

The advantage of standalone nodelets, is that you program the nodelet but can still run your module as a node. This mayb

I had the same problem. I was publishing the image and the camera_info correctly, but the image would still not show up.

You can ros-launch a node and make it start with gdbserver. The node is not started until you connect to it using Qt. To do so, try adding the launch-prefix "xterm -e gdbserver localhost:10000", and connect to it with Qt (both things from the local machine). More info here: http://bit.ly/1OEaVce

Thanks a lot! More info about first solution: http://askubuntu.com/a/148882/136863 , so I see that I can recover the original more secure behavior doing echo 1 | sudo tee /proc/sys/kernel/yama/ptrace_scope . About your second solution, more info here: http://askubuntu.com/a/153970/136863 .

Hey, there is another post in ROS Answers addressing how to debug a rosnode in Qtcreator: http://answers.ros.org/question/34966...

I think that the answer can be summarized to:

1. compile the node/package that you want to compile in debug mode (add "set(CMAKE_BUILD_TYPE Debug)" to your CMakeLists.txt)
2. start your roscore and everything else as usual, except for the node you want to debug
3. start qtcreator with "sudo" (remember to source correctly your catkin workspace)
4. in qtcreator start the debugging mode for the rosnode by going to "Debug"->"Start Debugging"->"Attach to unstarted Application...", look for your compiled node, it should be in: "${CATKIN_WORKSPACE_FOLDER}"/devel/lib/package_name/node_name and click "Start Watching".
5. start the node in a terminal either with rosrun or roslaunch
6. enjoy...

Note: you need to run qtcreator with "sudo", otherwise you get a "ptrace operation not allowed" problem. Note2: if you try to run qtcreator as a normal user afterwards, you will have problems accessing some of the Qt configuration files of your home folder, run these commands to get this back to normal:

sudo chown -R ${USER}:${USER} .qt
sudo chown -R ${USER}:${USER} .config/

@Stefan Kohlbrecher thanks it worked! As a side note, I had to run qtcreator with "sudo" for the debugging from Qtcreator to work; otherwise it works perfectly.

For anybody getting this error "ERROR: cannot launch node of type [rosbag/rosbag]: can't locate node [rosbag] in package [rosbag]". The launchfile should be changed to "<node pkg="rosbag" type="record" ...".<="" p="">

I have had problems with C++11 and ROS in the past. If you have a robot with Ubuntu 12.04 preinstalled you are going to have a bad time using C++11, because the boost version in 12.04 is not compatible with C++11. I recommend using C++03 which is the standard: http://www.ros.org/reps/rep-0003.html

Wow, thanks a lot for this answer! Then: catkin + vstools is a good option, I'll try it out soon! As a side note, it would be really worth it that there was a tutorial on how to use catkin outside of the ROS ecosystem (I tried to find it).

Hello everyone,

I would like to have advice on how to manage a family of packages in various robotics projects. I have already gathered information about alternatives, including: catkin + wstool, git-submodules, git-subtrees, and other software management options. I think that these links summarize the options:

• git-submodules: http://blogs.atlassian.com/2013/05/al...
• git-subtrees: http://blogs.atlassian.com/2013/03/gi...
• software build/dependency management tools: http://blogs.atlassian.com/2014/04/gi...
• other possibilities: http://stackoverflow.com/questions/81...

In my research group, we have worked on various ROS and non-ROS Robotics projects in the last few years, and naturally we have modules which are used in many of these projects at the same time. Until now, our code management approach was to have everything in a single git repository (there are 40+ ROS and non-ROS packages, as we try keep granularity/modularity in our libraries and executables). As we are updating our source structure to be catkin compliant, we are also thinking on separating our git repository into chunks (no necesarily at the package level). We hope that by managing our projects diferently we won't need to download our whole software repository in every platform.

I believe that catkin + vcstools/wstool are a very good solution since:

• They are both independent of the ROS ecosystem: check section 2.2 here http://wiki.ros.org/catkin/conceptual... , and this http://wiki.ros.org/wstool
• Catkin can manage build/compilation dependencies
• wstool allows to set the git repository for modules/packages inside the workspace, including specifying a branch or a specific commit
• also: wstool update, wstool diff, [...] are very useful

So, my questions are relating to the usage of catkin and wstool outside of the ROS ecosystem:

1. How can I create a catkin workspace outside of the ROS ecosystem? From what I understand, creating a workspace requires to load the setup.bash of a ROS dristro.
2. Any advice of a different set of tools which is suitable for working in ROS-like projects, where many modules are reused between projects?
3. As a side question regarding the use of catkin + wstool + rosdep : Is it possible to overlay rosdep's package lists with private repositories?

Thanks for your advice!

Hello everyone,

I just wanted to point out that the link to the rosws tutorials are broken in these websites:

• ROS > /wiki/ROS/Tutorials/InstallingandConfiguringROSEnvironment
• ROS > /wiki/rosws

The link that does not work is:

• ROS > /doc/api/rosinstall/html/rosws_tutorial.html

I have tried to find the correct link to the tutorial, but I have had no success. I do not know what to do about it (I am starting to use ROS), so I decided to post this question. Sorry for the cryptic webpage links, I do not have enough karma to post them. Thanks!

Regarding the stereo camera baseline (if anybody reads this question) is stored in the matrixes in the cameraInfo of the left and right cameras (see the answers to this question: http://answers.ros.org/question/29050/stereo-camera-calibration-fundamental-matrix/ )

Thanks, I am publishing the tf of the cameras using static_transform_publisher nodes ( http://www.ros.org/wiki/tf#static_transform_publisher ). I had to be careful to name the camera frames equally in the cameraInfo and when launching the static_transform_publisher nodes. Now everything works!

I have a dataset of images on disk, along with a set of camera parameters (though they are not in ROS camera_info file format). I am trying to make a node that reads and publish de images in round robin (when it finishes reading the images in the folder, it starts again). My main objective is to make it in such a way that I can "fool" other nodes, making them think that the node is a camera.

I have being trying to do it in the following way:

• Publish the images in a topic /camera/image_rect_color
• Publish sensor_msgs/CameraInfo in the topic /camera/camera_info
• I have tried to visualize the images in rviz using a camera display, but rviz complains because it says no /camera_link frame is published

At this point I believe there has to be an easier way to do this... Does anybody has suggestions about how to achieve this?

Thanks in advance!

I have made my own version of image_stereo_sequence_publisher.py, image_stereo_sequence_publish.launch based on your files. The main difference is that I am publishing stereo images simultaneously from the same node; and that I can set the name of the topics stereo/left/image_rect, etc.Thanks!

Thanks a lot, I am going to study and try your code!