MoveRelative: Fix bug in calculation of target pose

[JafarAbdi]

3b83598 introduced a bug where MoveRelative tries to move the link's origin to ik_frame despite link_T_ik_frame isn't the identity matrix, the problem is in the calculation we assume target_eigen to be world_T_ik_frame_new_pose which isn't the case (see move_relative.cpp#L270-L273 and move_relative.cpp#L243-L249)

I think the bug existed even before the refactor, it's just that we didn't have a case where link_T_ik_frame isn't the identity matrix

[codecov]

Codecov Report

Merging #318 (be1c59e) into master (b4a9e20) will decrease coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #318      +/-   ##
==========================================
- Coverage   53.00%   52.99%   -0.00%     
==========================================
  Files         102      102              
  Lines        7599     7598       -1     
==========================================
- Hits         4027     4026       -1     
  Misses       3572     3572              

Impacted Files	Coverage Δ
core/src/stages/move_relative.cpp	69.40% <ø> (-0.16%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b4a9e20...be1c59e. Read the comment docs.

[v4hn]

This patch is still broken because the direction property will be computed based on link_pose instead of ik_pose_world. I will provide a proper fix + test in a bit.