Knowledge Wiki Orchestrata Ultimate

Orchestrata Ultimate › Tests › Reference › scripts › test_check_feature_loss

Tests reference — tests/scripts/test_check_feature_loss.py

Module of tests/scripts, split out because the directory page would be too large.

Prose pending.

tests/scripts/test_check_feature_loss.py

source · 67 tests · 2414 lines

Tests for scripts/check_feature_loss.py — the feature/requirement-loss

Covers: core.base.traces

Kind Name Line Description
function test_not_a_git_repo_is_invalid_never_pass 145 Fail-closed, the deliberate opposite of check_wiring.py's SKIP=0
function test_empty_ratified_corpus_is_invalid 159 A repo with no specs/ at all examines 0 requirement ids — INVALID,
function test_clean_repo_passes 179 Everything ratified is currently implemented — PASS, exit 0, zero
function test_deleted_traced_symbol_orphans_its_requirement 212 The historical-catch proof: a symbol carrying @traces is deleted while
function test_symbol_deletion_not_orphaned_when_covered_elsewhere 263 A requirement is NOT a loss when another surviving symbol still
function test_unratified_deleted_symbol_dropped_not_failed 292 A deleted symbol citing a requirement id that NO current spec defines
function test_never_implemented_reported_as_info_not_failure 337 A requirement with zero current evidence AND zero deletion history —
function test_never_implemented_never_double_counts_orphaned 372 A requirement that IS orphaned (check 1) must appear there, never
function test_many_never_implemented_ids_summarised_not_dumped 398 Above the inline cap, the never-implemented section prints a COUNT and
function test_deleted_spec_doc_still_cited_fails 433 The spec/ADR document defining a requirement is deleted while a
function test_deleted_spec_doc_uncited_is_not_a_loss 484 Ordinary spec housekeeping: the doc is gone and NOTHING cites its old
function test_delete_then_restore_is_not_a_false_positive 518 Reproduces the documented noisy-history shape: a commit mass-deletes
function test_gate_ok_comment_does_not_suppress_a_loss 565 The # gate-ok: waiver convention (core/gate_waivers.py, honoured by
function test_many_losses_are_never_truncated 600 Unlike the never-implemented INFO section, every LOSS is printed in
function test_traces_from_text_extracts_multiple_reqs 638
function test_traces_from_text_ignores_placeholder_and_syntax_error 650
function test_dedupe_latest_keeps_first_seen_per_path 661
function test_json_output_is_well_formed 672
function test_main_in_process_pass_and_json 702
function test_main_in_process_fail_and_invalid 730
function test_render_report_invalid_status_directly 761
function test_render_scan_scope_empty_list_is_a_defensive_fallback 769 _render_scan_scope is called with a real, non-empty scan_scope on
function test_render_scan_scope_groups_multiple_roots_by_language 779
function test_non_python_file_deletion_under_src_is_ignored 811 Check 3 only ever looks at .py deletions — a deleted data file
function test_non_markdown_file_deletion_under_specs_is_ignored 836 Check 4 only ever looks at .md deletions.
function test_deleted_spec_doc_dual_defined_elsewhere_is_not_a_loss 860 A requirement id defined in TWO spec files (a legitimate, documented
function test_non_traces_call_decorator_is_ignored 891 A decorator that IS a call but is not named traces (or
function test_find_deletions_against_non_repo_returns_empty 904 git log genuinely erroring (a real, non-mocked returncode != 0,
function test_symbol_qualname_builds_the_evidence_files_dotted_form 923
function test_record_symbol_qualname_bare_and_dotted_forms_agree 936 SCHEMA CORRECTION (2026-08-19): the evidence-writing agents' real
function test_record_symbol_qualname_defensive_fallback_for_non_src_file 957 A file field that is not a src/orchestrata_ultimate/ path at
function test_absent_evidence_file_behaves_exactly_as_today 965 No file at plans/feature-loss-invalid-citations.json — "no
function test_evidenced_pair_reclassified_unevidenced_pair_still_fails 999 The combined proof the brief asks for: an EXACT evidenced
function test_stale_evidence_record_reported_not_silently_ignored 1091 A record whose pair matches nothing current (the finding was
function test_render_report_always_states_reclassified_count_and_path 1141 Property 4 of the brief: the output must always SAY how many were
function test_malformed_evidence_record_makes_the_whole_gate_invalid 1206 Property 3 + "refuse it rather than silently honour a half-filled
function test_methods_field_accepts_either_string_or_list_shape 1238 The positive counterpart of the parametrized rejection tests: both
function test_evidence_file_unreadable_is_invalid 1300 A path that EXISTS but cannot be read as text — here, it is a
function test_evidence_file_invalid_json_is_invalid 1329
function test_evidence_file_wrong_schema_version_is_invalid 1352
function test_evidence_records_not_a_list_is_invalid 1372
function test_evidence_top_level_not_an_object_is_invalid 1396
function test_json_output_includes_reclassification_fields 1415
function test_main_in_process_reclassification_json 1459 In-process (coverage-visible) counterpart of the subprocess JSON
function test_module_does_not_import_gate_waivers 1493 Structural guard mirroring test_check_generated_artifacts.py's own
function test_ts_only_traced_requirement_is_not_never_implemented 1537 A requirement with ZERO Python implementation but a real TS @traces
function test_deleted_python_symbol_covered_by_surviving_ts_symbol_is_not_a_loss 1567 The exact shape of the REQ-021-195..198 incident: a Python symbol
function test_genuine_loss_still_fails_with_unrelated_ts_tree_present 1600 Scanning TS/JS must never create a false NEGATIVE: an unrelated web
function test_scan_scope_reports_absent_web_roots_as_not_scanned 1631 A repo with no opencode/ or vizint-ui/ tree at all (the common case
function test_scan_scope_reports_present_web_root_as_scanned 1669 An opencode/ tree that DOES exist in this repo must be reported as
function test_json_output_includes_scan_scope_field 1698
function test_real_opencode_session_status_ids_produce_no_finding_at_all 1721 Canary against the real live tree (not just synthetic fixtures,
function test_surface_loss_flagship_reproduces_2883703a 1833 The acceptance fixture named in the brief, reconstructed: run the
function test_surface_loss_tier1_clears_via_same_class_property 1900 Isolates the RENAME-HANDLING claim: a field renamed OFF a dataclass
function test_surface_loss_evidence_reclassifies_a_genuine_relocation 1957 A field that Tier 1 CANNOT clear (relocated to a differently-named
function test_surface_loss_stale_evidence_reported_not_silently_ignored 2046 A surface_records entry matching no current candidate (fixed,
function test_absent_surface_records_key_is_not_an_error 2106 Today's real plans/feature-loss-invalid-citations.json has a
function test_malformed_surface_evidence_record_makes_the_whole_gate_invalid 2172
function test_surface_records_not_a_list_is_invalid 2203
function test_json_output_includes_surface_loss_fields 2229
function test_main_in_process_surface_loss_json 2260 In-process (coverage-visible) counterpart of the subprocess JSON test
function test_extract_public_dataclass_fields_decorator_forms_and_privacy 2285 Unit coverage of _extract_public_dataclass_fields /
function test_extract_public_dataclass_fields_ignores_syntax_error 2319
function test_current_class_members_finds_fields_properties_and_methods 2323 Unit coverage of _current_class_members: a plain field, a
function test_module_still_does_not_import_gate_waivers_with_surface_loss 2349 The SURFACE-LOSS mechanism, like RECLASSIFICATION before it, must
function test_load_surface_reclassification_evidence_invalid_json_is_invalid 2368 Direct unit coverage of load_surface_reclassification_evidence's
function test_render_report_surface_loss_over_cap_is_summarised_not_dumped 2387 Above _SURFACE_LOSS_INLINE_CAP, the SURFACE-LOSS section prints a