* rtabmap_slam tests and doc
* another round of review of the doc
* Disable by default use_intra_process_comms on latched/transient publishers
* Added test to catch not unlocked mutex from early error exit
* updated coverage settings
* Using UScopeMutex on all tryLock()
* fix(rtabmap_slam): release scan mutex on conversion failure
* test(rtabmap_slam): cover scan transform recovery in native tests
Replace the standalone reproducer with registered C++ GoogleTests for 2D and 3D scan recovery. Use real ROS inputs, separate DDS domains and bounded timeouts, and include rtabmap_slam in supported CI and coverage jobs.
Validated on Jazzy with RTAB-Map core 9ed83a7: both negative controls fail map recovery without the unlocks; both pass with the fix, followed by five repeated runs per case (10/10 passed). Broader distro CI has not been run locally.
* chore(rtabmap_slam): keep upstream submission focused on mutex fix
Remove the regression test and CI integration from the proposed diff. Retain the reproducer in commit 254ca3b so it can be offered separately with the validation evidence.
---------
Co-authored-by: Matthew Kotara <[email protected]>
Co-authored-by: matlabbe <[email protected]>
* Lyrical support
* removed dep not available on lyrical
* updated docker lyrcical
* making cmake less verbose, disabled tests on rtabmap_python
* cleanup not used tests in rtabmap_python
* addressing some lyrical deprecated warnings
* cancel pr jobs on recommit
* small refactor
* disabled lyrical bin docker
* restored python test and fixed error
* trigger ci
* fixing docker lyrical
* disabled rtabmap_ros lyrical
* pytest
* fixing packages-skip
* bump 0.23.7
* Migrating tf to tf2
* Added ci action to test PR on ros1
* updated dev container with nvidia working
* backward compatibility with topics having frame_id with leading slash not allowed with tf2
* backward compatibility of leading slash for other tf2 buffers
* updated comment
* Added stale_update_detection parameter to detect if upstream is stale for too long, trigger new map
* On too small guess motion, still publish odom topic along tf
* Added parameters when guess odom topic is sent
* fixed stale disabled logic
* Added info when odometry update is skipped.
* Making vo publishing the expected output data even if odometry update was skip by not enough motion
* fixed skipping odom
* refactored tooOldPreviousData to work with guess
* renamed stale_update_detection to staleness_factor
* cleanup
* reverting a change
---------
Co-authored-by: mathieu86 <[email protected]>
* Odom reset on time jump
* Adding more logs to debug
* refactored
* dont skip frame on clock jump
* Making clock check independent of the topic stamp check
* fixed post check
* Making diagnostic more robust to time jump
* Added node name to warning
* make sync warning msg working in case of time jump
* not need to reset timer
* reset timer
* timer auto reset already
* typo
* Added more time checks to make sure we don't republish a tf frame with stamp from a topic in the future
* dont send tf if time jump happened while processing
* fixed errors
* addressing comments
* fixing time comparison
* Adding ROS2 Kilted to CI
* updated ros2 ci workflow
* updated version of action-ros-ci
* Updated dev containers, fixed kilted build
* Added git autocompletion
matlabbe