Orchestrata Ultimate › Tests › Reference › agents › test_kernel_handler_ack_only
Tests reference — tests/agents/test_kernel_handler_ack_only.py
Module of tests/agents, split out because the directory page would be too large.
Prose pending.
tests/agents/test_kernel_handler_ack_only.py
source · 21 tests · 1002 lines
Covers: agents.kernel_handler, agents.volk_registry, core.base.traces, core.paths, kernel.pipeline, orchestrator.intent_broker_factory, plugin.registries, postman.postman, core.packet/ (package)
| Kind | Name | Line | Description |
|---|---|---|---|
| function | test_ack_returns_before_slow_volk_completes_then_reply_arrives |
82 | (a) A slow real plugin volk: the ack returns immediately (accepted=True, |
| function | test_failing_volk_emits_exactly_one_failure_packet |
138 | (b) A raising real plugin volk: exactly one terminal failure packet |
| function | test_governance_denial_stays_synchronous_no_background_task |
191 | (c) A sink IS wired, but governance denies the intent before the |
| function | test_status_subject_stays_synchronous_with_sink_set |
230 | (d) status never backgrounds, even with a sink wired. |
| function | test_sink_unset_stays_fully_synchronous_back_compat |
260 | (e) No sink wired -> byte-compatible synchronous receipt with the |
| function | test_process_in_background_is_noop_without_sink |
293 | Defensive guard: _process_in_background is only ever scheduled |
| function | test_background_blocked_intent_emits_no_reply_terminal |
307 | (D-H5, amends the parent design's D3) A governance-blocked |
| function | test_no_match_round_emits_no_reply_terminal |
356 | (D-H5) A response sent for a thread_id the runner has never seen: |
| function | test_second_round_never_reemits_first_rounds_reply |
402 | (P-H3, D-H3) Two same-thread rounds: round A gets a genuine reply, |
| function | test_no_reply_terminal_is_machine_detectable |
477 | (D-H5, SPEC-032 G243) _build_no_reply_packet's terminal is sent |
| function | test_failure_packet_carries_top_level_assistant_status |
507 | (D-H5, backlog [403]) _build_failure_packet's terminal mirrors |
| function | test_reply_to_propagated_to_reply_and_failure |
525 | P-B2 (D6): an inbound packet stamped metadata.reply_to gets that |
| function | test_reply_to_propagated_synchronous_path |
632 | The same P-B2 propagation applies to the fully-synchronous path (no |
| function | test_run_intent_response_subject_reuses_existing_thread |
676 | Direct unit coverage of _run_intent's response branch (used by |
| function | test_shutdown_drains_inflight_round |
709 | (P-H5, D-H6) A background round still running when shutdown is |
| function | test_shutdown_cancels_and_sinks_failure |
758 | (P-H5, D-H6) A background round still running past the drain grace |
| function | test_shutdown_is_noop_with_no_inflight_rounds |
810 | shutdown with an empty _bg_tasks set (nothing in flight, or |
| function | test_shutdown_falls_back_to_default_grace_on_invalid_env |
829 | (F-1 hardening) An unparseable KERNEL_BG_DRAIN_GRACE_S (garbage, |
| function | test_accept_during_drain_gets_draining_receipt |
856 | (P-H10, D-H11) A packet arrives after shutdown has begun draining |
| function | test_draining_receipt_preserves_thread_and_reply_to |
924 | (P-H10, D-H11) The synchronous KERNEL_DRAINING receipt echoes the |
| function | test_non_spawning_subject_unaffected_while_draining |
956 | (P-H10, D-H11) A non-spawning subject (status) behaves identically |