Switch to tf2 for Kinetic #458
Conversation
vrabaud
force-pushed
the
kinetic-devel
branch
2 times, most recently
from
43394c1 to
058469c
Compare
|
I think you had a problem with that PR at some point, a year or two ago. Tests in Kinetic apparently got tighter and helped uncover a but on my end in AMCL. You should probably give it a new try if you have the time. |
|
@dbking77 , enjoy |
vrabaud
force-pushed
the
kinetic-devel
branch
2 times, most recently
from
356b601 to
e7fe270
Compare
|
rebased, tests are passing: it shows as failed because there was a GCC warning (that I cannot narrow down to). |
|
Just tried this out on my Kinetic simulation -- and it seems that Fetch cannot turn (except in the recovery behaviors)... going to try and look into it more later |
|
Thx for not giving up on that. Remember it is the first step to get rid of PCL. |
|
Yeah, I had rebased on top of head of kinetic for testing, I'm not sure why it doesn't work at runtime -- I do plan to look into it now that I have a 16.04 setup, but ran out of time the other day |
|
👍 for this, but... out of curiosity: given tf1 indigo implementation wraps a tf2 listener for backward compatibility, can we expect any performance improvement with this? |
vrabaud
force-pushed
the
kinetic-devel
branch
from
c4ee16c to
054dd63
Compare
|
@corot, no idea about performance, I don't use that code. |
vrabaud
force-pushed
the
kinetic-devel
branch
from
054dd63 to
7c086f7
Compare
|
Sooooo, how do we move with this one? I understand it is scary and things were not working for you. |
|
If this is passing... Could be nice! |
|
I can cherry-pick to Lunar, please advise. |
vrabaud
force-pushed
the
kinetic-devel
branch
3 times, most recently
from
88f0835 to
8a5f900
Compare
vrabaud
force-pushed
the
kinetic-devel
branch
2 times, most recently
from
e58a0fd to
ab3243a
Compare
This is to get test_rosie_multilaser.xml to work
Cherry-pick a4a3da9 from ros-planning#458 & @vrabaud Original PR branch was aimed at Kinect, thus a straight cherry-pick does not compile. This fixes all conflicts detected by Git. Not Compiling: move_base, global_planner. Due to changes merged since the original ros-planning#458. I will fix those issues in a new commit. Conflicts: amcl/package.xml amcl/src/amcl_node.cpp amcl/test/basic_localization.py base_local_planner/include/base_local_planner/local_planner_util.h base_local_planner/package.xml base_local_planner/src/local_planner_util.cpp base_local_planner/src/trajectory_planner_ros.cpp carrot_planner/package.xml clear_costmap_recovery/package.xml costmap_2d/CMakeLists.txt costmap_2d/include/costmap_2d/costmap_2d_ros.h costmap_2d/package.xml costmap_2d/src/costmap_2d_ros.cpp costmap_2d/test/footprint_tests.cpp dwa_local_planner/package.xml fake_localization/fake_localization.cpp fake_localization/package.xml global_planner/package.xml move_base/package.xml move_base/src/move_base.cpp nav_core/package.xml navfn/package.xml robot_pose_ekf/CMakeLists.txt robot_pose_ekf/include/robot_pose_ekf/odom_estimation.h robot_pose_ekf/include/robot_pose_ekf/odom_estimation_node.h robot_pose_ekf/package.xml robot_pose_ekf/src/odom_estimation.cpp robot_pose_ekf/src/odom_estimation_node.cpp rotate_recovery/package.xml rotate_recovery/src/rotate_recovery.cpp
This may have already been forward ported with PR ros-planning#728 But it was originally done as part of PR ros-planning#458. Which is why there was a conflict. I am cherry-picking for completeness. get code to work with frames starting with "/" This is to get test_rosie_multilaser.xml to work Conflicts: amcl/src/amcl_node.cpp
This finishes the compile errors from cherry-picking ros-planning#458. This commit ports the changes from ros-planning#601 to tf2.
This a cherry-pick from https://github.com/ros-planning/navigation/pull/322/files