Fix zombie processes on Mac tests

[Dreamsorcerer]

Kill processes that may have been left over from last run.
Log errors from failed processes.

[github-actions]

go2 replay realtime (arm64)

Details

Benchmark suite	Current: fd49a5d	Previous: afc6a47	Ratio
peak memory	1607.148 MB	1570.027 MB	1.02
peak threads	358 threads	363 threads	0.99
network (transport)	2722.047 MB	2730.099 MB	1.00
disk write	1.629 MB	1.727 MB	0.94
instructions	115.617 G	115.659 G	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The reviewed changes are safe to merge; no blocking issues remain.

T-Rex evidence

Evidence from the check

• The executed harness retrieves Apple references, extracts the workflow block, and runs isolated before/after tests with the Linux emulation boundary explicitly stated.

Evidence from the check

• The baseline shell preserves the merge-base absence of cleanup, allowing the SIGTERM-ignoring fixture to remain alive.

Evidence from the check

• The extracted workflow block runs with a labeled Linux adapter for Apple’s documented quiet option, making the exact execution reproducible.

Command output from the check

• The baseline test exited successfully while its isolated fixture remained alive, establishing the before condition.

Command output from the check

• The traced cleanup completed 30 polls and killed the SIGTERM-ignoring fixture with SIGKILL, showing no early exit under the documented quiet-option semantics.

Command output from the check

• The recorded curl response contains Apple’s older getopt string and quiet-option handler, contradicting the claim that macOS lacks the flag.

Command output from the check

• The recorded Apple manual documents quiet output for the q option, confirming support in the older release.

Command output from the check

• The recorded curl response contains the Apple-specific option parser and quiet handler, directly contradicting the candidate’s flag-support premise.

Command output from the check

• The recorded Apple manual explicitly documents pgrep quiet mode, establishing that the flag is supported.

Command output from the check

• Executed commands capture Linux platform details, runtime tool availability, and the exact workflow change, establishing the native macOS validation limitation.

Command output from the check

• The driver capture records the executed command, working directory, successful exit, downloaded references, and both process runs, preserving the full execution record.

Command output from the check

• The final git diff command exited zero with no output, confirming that validation left tracked files unchanged.

Summary

The PR adds macOS cleanup before tests and includes stderr when a Livox test process exits early. The Livox tests now track their spawned processes while collecting messages. No blocking issues were found in the reviewed changes.

• macOS CI clears leftover workspace processes before tests.
• Livox tests show why native processes exit early.

_{Reviews (1) · Last reviewed commit: "Conditional sleep" · Reviewed by Greptile}

[codecov]

Codecov Report

❌ Patch coverage is 85.10638% with 7 lines in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
dimos/hardware/sensors/lidar/livox/test_e2e.py	85.10%	4 Missing and 3 partials ⚠️

@@            Coverage Diff             @@
##             main    #4567      +/-   ##
==========================================
- Coverage   81.21%   80.54%   -0.68%     
==========================================
  Files        1651     1670      +19     
  Lines      154436   156743    +2307     
  Branches    13017    13301     +284     
==========================================
+ Hits       125431   126247     +816     
- Misses      25424    26909    +1485     
- Partials     3581     3587       +6     

Components	Coverage Δ
Tests	94.60% <85.10%> (-1.45%)	⬇️

Flag	Coverage Δ
OS-macos-latest	75.69% <23.40%> (-0.02%)	⬇️
OS-ubuntu-24.04-arm	75.45% <23.40%> (-0.01%)	⬇️
OS-ubuntu-latest	75.81% <23.40%> (-0.01%)	⬇️
Py-3.10	75.81% <23.40%> (-0.01%)	⬇️
Py-3.11	75.80% <23.40%> (-0.01%)	⬇️
Py-3.12	75.80% <23.40%> (-0.01%)	⬇️
SelfHosted-Linux	40.04% <82.97%> (+<0.01%)	⬆️
SelfHosted-macOS	40.34% <85.10%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
dimos/hardware/sensors/lidar/livox/test_e2e.py	95.90% <85.10%> (-2.84%)	⬇️

... and 44 files with indirect coverage changes

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.