Findings: BR-005, BR-019, BR-023, BR-028.
Establish branch and critical-module coverage thresholds, a bounded warning allowlist, exact
accounting invariant bounds, and a dedicated non-coverage performance-regression job.
Primary surfaces: pytest and coverage configuration, invariant helpers, warning filters,
performance workflow and retained samples.
Acceptance:
- CI fails below agreed statement, branch, and critical-module thresholds.
- Clean Python 3.12-3.14 runs emit no repeated configuration warning and fail on new project
warnings.
- Performance tests retain interleaved samples and do not run under coverage instrumentation.
- Repeated runs stay within the 10 percent regression rule without the retained flake.
Dependencies: Issues 2 and 15.
Plan reference: /home/stefan/ml4t/libraries/ml4t-backtest-dev/.workspace/work/backtest-stable-release/milestone-3/plan.md - Issue 16 in milestone 0.1.0-m3 - Compatibility and release enforcement.
Findings: BR-005, BR-019, BR-023, BR-028.
Establish branch and critical-module coverage thresholds, a bounded warning allowlist, exact
accounting invariant bounds, and a dedicated non-coverage performance-regression job.
Primary surfaces: pytest and coverage configuration, invariant helpers, warning filters,
performance workflow and retained samples.
Acceptance:
warnings.
Dependencies: Issues 2 and 15.
Plan reference:
/home/stefan/ml4t/libraries/ml4t-backtest-dev/.workspace/work/backtest-stable-release/milestone-3/plan.md- Issue 16 in milestone0.1.0-m3 - Compatibility and release enforcement.