Jose Luis Blanco's profile - activity

Pure CMake package: what is needed to append a path to PYTHONPATH? Context: Working on a pure cmake package (mrpt/packag

Should CMAKE_INSTALL_LIBDIR be ignored in REP136 packages? For example, from the build log for nodeletlib here: cd obj-

This (likely seems)-->seemed to be caused by only partially following the instructions of REP 136. Will come back wit

This (likely seems)-->seemed to be caused by only partially following the instructions of REP 136. Will come back wit

This (likely seems)-->seemed to be caused by only partially following the instructions of REP 136. Will come back wit

This (likely seems)-->seemed to be caused by only partially following the instructions of REP 136. Will come back wit

This likely seems to be caused by only partially following the instructions of REP 136. Will come back with the outcome

This likely seems to be caused by only partially following the instructions of REP 136. Will come back with the outcome

This likely seems to be caused by only partially following the instructions of REP 136. Will come back with the outcome

@Vincent-rabaud @WilliamWoodall I recently found an almost identical issue, please let me know if the bug you mention ma

@Vincent I recently found an almost identical issue, please let me know if the bug you mention may still be the cause? h

@vincent-rabaud I recently found an almost identical issue, please let me know if the bug you mention may still be the c

Did you find a solution? Basically I found the same issue, here: https://answers.ros.org/question/290505/jenkins-error-d

Jenkins error: dpkg-shlibdeps can't find libs building non-catkin pkg Hi, I hope you can help me finding what's wrong

Hello,

I am struggling with a persistent error with one of my packages: in the past we have done several releases (bloom-release) without problems, but since a few weeks ago Jenkins is persistently complaining about:

Invoking 'git clone --branch debian/ros-kinetic-mrpt-ekf-slam-3d_0.1.1-0_jessie --depth 1 --no-single-branch https://github.com/mrpt-ros-pkg-release/mrpt_slam-release.git /tmp/sourcedeb/source'
Cloning into '/tmp/sourcedeb/source'...
fatal: Remote branch debian/ros-kinetic-mrpt-ekf-slam-3d_0.1.1-0_jessie not found in upstream origin

Indeed, that branch does not exist: I can only see these tags:

debian/ros-kinetic-mrpt-ekf-slam-3d_0.1.1-0_xenial
debian/ros-kinetic-mrpt-ekf-slam-3d_0.1.1-0_willy

Here is the complete log: http://build.ros.org/job/Ksrc_dJ__mrp...

My actual questions are:

1) Is there something that I might have done wrong while releasing the package with bloom-release?

2) If not, how can I fix it, apart of doing a "fake" new release as a workaround?

Yes, mrpt_localization relies on odometry, but if it is not present it should still be able to track the robot just from laser scans (increase the "uncertainty parameters" a bit to make the particles spread more than usual). If it doesn't work open an issue on GitHub or try to fix and pull-request

@kost9: Great!

@sealguy77: Just clone the branch named "indigo", it should work on hydro too. In the future we'll probably rename the branch to "master" or alike to avoid confusion.

The best solution is to remove the system version of mrpt, which is quite old in Ubuntu 12.04:

sudo apt-get remove libmrpt-dev mrpt-apps

then installing a more recent version. You have 2 options:

1) Build from sources, either cloning master from GitHub or downloading the last stable version. Building instructions can be found here.

2) Installing with apt-get via the PPA repository.

Cheers,
JL

As usual, you're right... Without further changes it now builds. Thanks!

tl;dr: Due to bad planning, I bloom-released a metapkg and now I want to get one of the pkgs out to become an independent pkg. Already updated rosdistro/indigo/distribution.yaml and forced farm rebuilds a couple times. Build farms still tries to clone OLD_REPO-release instead of NEW_REPO-release. What else should I try?

Details:

1) Metapackage is mrpt_navigation. See rosdistro entry

2) Yesterday it contained a package named pose_cov_ops, I had to manually remove it because otherwise bloom complained (correctly) when trying to release the new package because it was already in the metapkg. See commit.

3) After that, I manually finished the PR that bloom failed to do. See commit

4) bloom succeeds in subsequent releases. See commit.

5) This is the current entry of the new pkg In rosdistro:

  pose_cov_ops:
    doc:
      type: git
      url: https://github.com/mrpt-ros-pkg/pose_cov_ops.git
      version: master
    release:
      tags:
        release: release/indigo/{package}/{version}
      url: https://github.com/mrpt-ros-pkg-release/pose_cov_ops-release.git       ## <<<< NEW, CORRECT REPO URI (this comment is not on rosdistro!!)
      version: 0.1.5-0
    source:
      type: git
      url: https://github.com/mrpt-ros-pkg/pose_cov_ops.git
      version: master
    status: maintained

6) But the build farm still believes the release repository is the OLD one (see complete log):

+ cd /home/rosbuild/hudson/workspace/ros-indigo-pose-cov-ops_sourcedeb/workspace && gbp-clone https://github.com/mrpt-ros-pkg-release/mrpt_navigation-release.git
Fetching targets
+ cd /home/rosbuild/hudson/workspace/ros-indigo-pose-cov-ops_sourcedeb/workspace/mrpt_navigation-release && git tag -l debian/*
No matching tag found. Are you sure you pointed to the right repository or the version is right?, expected debian/ros-indigo-pose-cov-ops_0.1.5-0_saucy:

Notice it uses the OLD repo URI: https://github.com/mrpt-ros-pkg-relea... instead of the one provided in rosdistro... Why?

I tried to trace where the incorrect repository URI comes from and found it to come from:

• the input argument "repo_uri" of ros-infrastructure/buildfarm/scripts/generate_sourcedeb

• which comes from RELEASE_URI in source_build.sh

• from "doit(release_uri,..." in release_jobs.py

But: WHO calls release_jobs.py? I couldn't find it.

Any ideas?

I wrote a short tutorial on this (and also on using CodeBlocks, but QtCreator seems much more responsive).

In short:

• Replace the src/CMakeLists.txt symlink with an actual file
• Open qtcreator from the console.
• File -> “Open File or project…“, then pick ~/catkin_ws/src/CMakeLists.txt and set the build dir to ~/catkin_ws/build/

• Select the tab Projects -> Build Settings and add these CMake arguments:

  -DCMAKE_INSTALL_PREFIX=../install -DCATKIN_DEVEL_PREFIX=../devel

• Close & reopen qtcreator (from a console where you sourced ROS setup.bash!) to make sure it generates the project.

• Ready to go!

Hi all,

PoseWithCovariance has a 6x6 covariance matrix which, according to REP 103, corresponds to these variables in this order: (x, y, z, rotation about X axis, rotation about Y axis, rotation about Z axis). With rotation around fixed axes, not like in the "common" yaw/pitch/roll various conventions.

I'm working in a pkg for covariance transformations (pose_cov_ops) so it would be great if I could compare my numerical results with any other existing ROS packages that output PoseWithCovariance at present. Checks would go in the package unit tests.

More explicitly: I'm looking for some "stable" ROS package which computes covariances in 6D, so I can peek its source code and make sure I'm following the right convention.

Any recommendation?

If there're none, I'll assume the 3x3 rotation matrix is as follows:

R = R_x(roll) * R_y(pitch) * R_z(yaw)

R = R_z(yaw) * R_y(pitch) * R_x(roll)

where matrices are in the inverse same order than the non-global axis, yaw-pitch-roll convention.

PS: wouldn't it have made more sense to keep a 7x7 covariance for the unit quaternion form? In general, equations for uncertainty propagation in this form are far simpler than for Euler angles. Is there room for a PoseWith7x7CovarianceStamped yet?