Orchestrata Ultimate › Tests › Reference › scripts › test_gates
Tests reference — tests/scripts/test_gates.py
Module of tests/scripts, split out because the directory page would be too large.
Prose pending.
tests/scripts/test_gates.py
source · 25 tests · 315 lines
| Kind | Name | Line | Description |
|---|---|---|---|
| function | test_every_gate_names_a_command |
62 | |
| function | test_every_gate_script_exists_on_disk |
71 | A gate whose script is absent is a declared mechanism that cannot run. The |
| function | test_gate_input_roots_exist |
85 | |
| function | test_worktree_envs_gate_wires_the_poison_class_only |
101 | The RUL-021 detector existed but was wired into NOTHING — how ship |
| function | test_always_run_sentinel_is_never_a_fingerprint_input |
115 | _fingerprint must never be handed the sentinel: it would hash to the |
| function | test_interesting_accepts_analysed_file_types |
143 | |
| function | test_interesting_rejects_noise |
156 | |
| function | test_fingerprint_is_stable_for_unchanged_content |
163 | |
| function | test_fingerprint_changes_when_content_changes |
168 | |
| function | test_fingerprint_ignores_mtime_churn |
176 | Content-addressed, deliberately: a file rewritten with identical bytes must |
| function | test_fingerprint_covers_untracked_files |
186 | It walks the tree rather than asking git what changed, so a brand-new |
| function | test_fingerprint_ignores_uninteresting_files |
194 | |
| function | test_fingerprint_skips_absent_roots |
200 | A declared root that does not exist is skipped, not an exception — the |
| function | test_fingerprint_distinguishes_identical_content_at_different_paths |
208 | The path is hashed alongside the bytes, so moving a file changes the |
| function | test_cache_roundtrips |
220 | |
| function | test_absent_cache_reads_as_empty |
225 | |
| function | test_corrupt_cache_reads_as_empty |
229 | A corrupt cache must never block a run — it degrades to "nothing cached", |
| function | test_save_leaves_no_temp_file |
237 | The write goes through a .tmp + replace so a killed run never leaves a |
| function | test_save_overwrites_previous_contents |
244 | |
| function | test_run_gate_reports_success |
253 | |
| function | test_run_gate_reports_failure_with_output |
261 | |
| function | test_run_gate_captures_stderr_too |
270 | A gate that reports only on stderr must not appear to have said nothing. |
| function | test_run_gate_summary_is_bounded |
284 | Summaries are truncated so one chatty gate cannot bury the other fifteen. |
| function | test_runtime_call_count_is_zero_without_a_snapshot |
296 | |
| function | test_runtime_call_count_walks_nested_structures |
301 | The snapshot nests functions inside modules inside a list, so the counter |