[Chaffelson]

Summary

Patch release candidate for 1.8.1. Fixes #419 and fixes #418.

• AffectedComponentEntity.reference_type rejects STATELESS_GROUP, breaking parameter provider fetch/apply on flows with stateless groups #419: AffectedComponentEntity.reference_type rejects STATELESS_GROUP. NiFi returns STATELESS_GROUP for components in stateless process groups. The NiFi 2.12.0 OpenAPI spec omits it from AffectedComponentEntity.referenceType, though AffectedComponentDTO has it. So fetching or applying a parameter provider on a flow with stateless groups raised ValueError while deserialising the response, before the apply was submitted.
  • Upstream: NIFI-16464, fixed by apache/nifi#11811 (merged to main, so in a future NiFi release).
  • Fix: restore the spec augmentation nifi_affected_component_reference_type.py, which mirrors the DTO enum into the entity, then make augment-openapi gen-clients against the committed 2.12.0 base spec.
  • The regeneration changes only nipyapi/nifi/models/affected_component_entity.py, by one line. The resulting values and order match the merged upstream annotation.
  • The augmentation is a no-op once the upstream spec has the value. Its docstring says to remove it at the regeneration that picks up NIFI-16464.
  • We chose this over generating from the upstream merge commit: main is a 2.13.0-SNAPSHOT that is 126 commits past 2.12.0, with new model fields and endpoints a 2.12 server doesn't have.
• ci.verify_config returns port failures as strings; controller and processor failures are dicts #418: ci.verify_config port failures were plain strings. Port validation has returned port_results[*].failures as plain strings since it was introduced in 1.7.0, while controller and processor results use {"step", "explanation"} dicts. We are changing ports to the same shape ("step": "Port validation") for consistency, so callers can handle every result row the same way.
  • Impact is narrow. It only affects callers that read the individual port failure items, and only when a port is invalid. Checks on verified, failed_count or a row's success are unaffected, and valid ports still return failures: [].
  • No other nipyapi code reads port failures. Callers that parsed the strings should read explanation; this is noted in the history.

docs/history.rst has the dated 1.8.1 (2026-10-24) entry.

Test plan

• make lint: 10.00/10. All pre-commit hooks pass (black, isort, flake8, pylint).
• make augment-openapi gen-clients: regenerated diff is the single AffectedComponentEntity enum line.
• New tests/test_generated_models.py::test_affected_component_accepts_stateless_group.
• tests/test_ci.py::test_verify_config_with_invalid_port now asserts the failure shape.
• make test NIPYAPI_PROFILE=single-user (NiFi 2.12.0): 777 passed, 33 skipped, 1 failed.
  • The failure, test_versioning.py::test_update_git_flow_ver_to_latest, passes when run alone. It is order-dependent in the GitHub-registry tests, and the versioning code is untouched.
• CI matrix (single-user py3.9–3.12, LDAP, mTLS), CodeQL, codecov and the RTD preview are all green.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.30%. Comparing base (3743318) to head (63d8cf3).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #420      +/-   ##
==========================================
+ Coverage   78.27%   78.30%   +0.02%     
==========================================
  Files          41       41              
  Lines        5115     5116       +1     
==========================================
+ Hits         4004     4006       +2     
+ Misses       1111     1110       -1     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.