Skip to content

feat(gates): wire gate6 classifier into receipts#26

Open
gowdaharshith1998-lang wants to merge 1 commit into
m2-phase5-gate6-probesfrom
m2-phase6-gate6-classifier
Open

feat(gates): wire gate6 classifier into receipts#26
gowdaharshith1998-lang wants to merge 1 commit into
m2-phase5-gate6-probesfrom
m2-phase6-gate6-classifier

Conversation

@gowdaharshith1998-lang

Copy link
Copy Markdown
Owner

Summary

  • Adds Gate 6 probe classification and overall evaluation for all seven required buckets.
  • Wires Gate 6 evaluation into rebuild receipts so all six gates now emit real statuses.
  • Records passed/inconclusive/failed Gate 6 details, including probe_count, divergence_count, accepted_with_note_count, and diverging_input for failures.

EARS Coverage

  • R-6.1: classify_results assigns exactly one bucket per probe.
  • R-6.2: value, exception, and deterministic stdout/stderr divergences fail Gate 6.
  • R-6.3: replay-based stochastic observable differences are accepted with note.
  • R-6.4: double relative tolerance uses 2^-52.
  • R-6.5: passed Gate 6 receipts include required counts.

Verification

  • python -m ruff check src/omnix/gates/gate6_equivalence src/omnix/rebuild/runner.py tests/gates/test_gate6_classifier.py tests/rebuild/test_runner.py
  • python -m mypy src/omnix/gates/gate6_equivalence src/omnix/rebuild/runner.py
  • pytest tests/gates/test_gate6_classifier.py tests/rebuild/test_runner.py -q
  • pytest tests/gates/ tests/rebuild/test_runner.py tests/semantic/java/test_vendor_integrity.py -q
  • real Gate 6 evaluate on identical StringUtils.reverse sample: passed, 26 probes, 0 divergences
  • pytest tests/ -q --tb=line
  • npx gitnexus detect-changes --scope staged: MEDIUM risk, 4 rebuild flows affected

Notes

  • Stacked on Phase 5 PR because earlier M2 PRs are still awaiting merge.
  • GitHub Actions may not start due the known account billing/spending-limit block documented in the dispatch.

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.

1 participant