Skip to content

Drain covered LiDAR frames before reading more input - #189

Merged
mehermvr merged 2 commits into
PRBonn:masterfrom
dddd-jh:fix/drain-ready-lidar-buffer
Oct 5, 2026
Merged

mehermvr merged 2 commits into
PRBonn:masterfrom
dddd-jh:fix/drain-ready-lidar-buffer

Conversation

@dddd-jh

@dddd-jh dddd-jh commented Oct 5, 2026 •

Copy link
Copy Markdown
Contributor

The sequencer can drop buffered LiDAR frames at end of input even when the final IMU message already covers them. After emitting one covered frame it resumes reading the input iterator; if that iterator is exhausted, it exits before emitting the remaining ready frames.

This adds continue after a frame is emitted so the next iteration checks the buffer again before reading more input. The existing strict IMU coverage condition remains unchanged.

Per the maintainer's feedback, the PR now contains only the one-line change in the package. The regression tests used for the original validation are retained locally rather than included in this patch.

Validation: the original 11 regression cases had three failures before the fix and all passed afterward; the complete 53-case Python suite and Ruff checks passed with the added local tests. That validation used the current Python source with the published RKO-LIO 0.4.0 Windows native extension; I have not rebuilt the full C++ extension or run a ROS node locally.

@mehermvr

mehermvr commented Oct 5, 2026

Copy link
Copy Markdown
Member

Thanks for the fix. I'll just take the one line change in the package proper. no need for a test.

@dddd-jh

dddd-jh commented Oct 5, 2026

Copy link
Copy Markdown
Contributor Author

Thanks! I've reduced the PR to the one-line change in the sequencer and kept the regression checks locally.

@mehermvr
mehermvr merged commit 51bc8bb into PRBonn:master Oct 5, 2026
23 checks passed
@dddd-jh

dddd-jh commented Oct 5, 2026 via email

Copy link
Copy Markdown
Contributor Author

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants