diff --git a/tests/test_brief_11b_step2_verb_classification.py b/tests/test_brief_11b_step2_verb_classification.py index 33f5fd8f..9b5131d0 100644 --- a/tests/test_brief_11b_step2_verb_classification.py +++ b/tests/test_brief_11b_step2_verb_classification.py @@ -44,6 +44,11 @@ EXPECTED_FILLER_CASES = frozenset( { "gsm8k-train-sample-v1-0002", "gsm8k-train-sample-v1-0016", + # 0021 joined the set as the auditor's verb classification evolved; it + # refuses with rows[0]=pre_frame_filler_sentence and still refuses after + # the filler sentence is elided (zero-lift holds — verified by the + # parametrized test_post_skip_simulation_still_refuses). + "gsm8k-train-sample-v1-0021", "gsm8k-train-sample-v1-0025", "gsm8k-train-sample-v1-0028", "gsm8k-train-sample-v1-0030", @@ -64,6 +69,10 @@ EXPECTED_POST_SKIP = { "unresolved_pronoun", "pronoun_resolution", ), + "gsm8k-train-sample-v1-0021": ( + "unknown_word", + "lexicon_entry", + ), "gsm8k-train-sample-v1-0025": ( "unknown_word", "lexicon_entry", diff --git a/tests/test_cli_test_suites.py b/tests/test_cli_test_suites.py index 58828b9f..ae6a8dbc 100644 --- a/tests/test_cli_test_suites.py +++ b/tests/test_cli_test_suites.py @@ -73,12 +73,15 @@ def test_core_test_suite_accepts_pytest_flags_without_separator(monkeypatch) -> rc = cli.main(["test", "--suite", "packs", "-q"]) assert rc == 0 + # Assert against the live suite definition rather than re-hardcoding the + # (growing) packs file list: the contract under test is that a curated + # suite expands to its files followed by the forwarded "-q", with no "--" + # separator needed. assert calls[0] == ( cli.sys.executable, "-m", "pytest", - "tests/test_core_semantic_seed_pack.py", - "tests/test_adr_0127_pack_ratification.py", + *cli._TEST_SUITES["packs"], "-q", ) diff --git a/tests/test_recognizer_registry.py b/tests/test_recognizer_registry.py index 90248781..6645c28f 100644 --- a/tests/test_recognizer_registry.py +++ b/tests/test_recognizer_registry.py @@ -287,7 +287,11 @@ def test_live_proposal_log_has_phase_c_proposals() -> None: f"expected {len(PHASE_C_PROPOSAL_IDS)} accepted Phase C proposals, " f"got {accepted_count}: {[(pid[:12], state[pid]['state']) for pid in PHASE_C_PROPOSAL_IDS]}" ) - # Registry exposes the ratified set. - assert len(load_ratified_registry(log)) == len(PHASE_C_PROPOSAL_IDS) + # Registry exposes the ratified set, which has grown past round 1 as later + # ratification rounds (flywheel-demo, rat1-cli-seed, ME-waves) accepted more + # recognizers. The contract here is that the Phase C three remain present + # and ratified — not that they are the only entries. + registry_proposal_ids = {r.proposal_id for r in load_ratified_registry(log)} + assert set(PHASE_C_PROPOSAL_IDS) <= registry_proposal_ids