When comparing the get_path/goal/goal/target_pose and get_path/result/result/path/poses[-1] (both in the same coordinate frame map), I noticed that the poses were slightly different sometimes (in orientation). I traced it down to being caused by transformPose "transforming" the pose from map to map; skipping the transformation for matching frames and just returning the input pose fixed the above issue.
When comparing the
get_path/goal/goal/target_poseandget_path/result/result/path/poses[-1](both in the same coordinate framemap), I noticed that the poses were slightly different sometimes (in orientation). I traced it down to being caused bytransformPose"transforming" the pose frommaptomap; skipping the transformation for matching frames and just returning the input pose fixed the above issue.