Skip to content

Add regression tests for lock poison-recovery paths and Whitaker CI/Makefile wiring #159

Description

@coderabbitai

Summary

Edge-harvest, recipe, MST, and invariant changes have strong tests, but the new poison-recovery paths (PoisonError::into_inner usages) and the CI/Makefile Whitaker lint wiring lack direct regression coverage.

Required changes

  • Add focused unit tests that deliberately poison the touched mutexes/locks (e.g. DistanceCache, chutoro-test-support/src/lib.rs::RecordingLayer, chutoro-core/src/session/clock.rs::FixedMonotonicClock, chutoro-cli/src/logging.rs::init_logging) and assert recovery behaviour.
  • Add contract tests (or CI smoke checks) for the new lint-clippy/lint-whitaker Makefile targets, the whitaker-installer install/cache steps in .github/workflows/ci.yml, and the dataset_recipe_phase_order nextest timeout overrides in .config/nextest.toml.

Affected areas

  • chutoro-core/src/hnsw/distance_cache.rs
  • chutoro-core/src/session/clock.rs
  • chutoro-cli/src/logging.rs
  • chutoro-test-support/src/lib.rs
  • Makefile, .github/workflows/ci.yml, .config/nextest.toml

Acceptance criteria

  • New tests exercise poisoned-lock recovery for each touched mutex and pass.
  • CI/Makefile lint wiring has an automated check (e.g. a CI job step or script) verifying the targets/steps run as expected.

References

  • PR: $pr_url
  • Requested by: @$assignee

Metadata

Metadata

Assignees

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions