Radio State Pipeline Regression Matrix

Status: MOR-348 risk-to-coverage index Date: 2026-06-03

This matrix maps the original state-pipeline risks from docs/superpowers/specs/2026-06-02-radio-state-pipeline-design.md to current coverage. It is an index, not the release checklist. Run final gate commands from docs/internals/radio-state-pipeline-validation.md.

CI coverage must remain hardware-free. Hardware validation remains separate and requires explicit opt-in.

Coverage Primitives

• rigplane.core.state_pipeline_contracts: FieldPath, Observation, ChangeSet, freshness transitions, command intent/lifecycle contracts, and capability metadata.
• rigplane.core.state_store.StateStore: canonical state/freshness revisions, observation sequence, immutable snapshots/deltas, and stale-field repair hints.
• rigplane.core.acquisition_scheduler: backend-neutral ensure-fresh requests, dedupe/priority/cadence policy, external CAT pause handling, and meter coalescing.
• Web delivery: HTTP and WebSocket state projections from StateStore, legacy revision aliasing stateRevision, and additive transportSeq.
• Runtime/rigctld adapters: CI-V observation projection, poller command readback, rigctld GET projection, and scoped pending overlays.

Risk-To-Coverage Matrix

Area	Original symptom or risk	Current coverage	Remaining release gate
StateStore invariants	Consumers could not distinguish semantic changes, freshness changes, and observation traffic.	tests/test_state_store.py::test_noop_observations_do_not_advance_state_revision, test_freshness_expiration_advances_freshness_without_state_change, test_full_snapshot_and_delta_projection_agree_after_observation_sequence, test_snapshot_output_cannot_mutate_store_owned_state, test_returned_changes_cannot_mutate_delta_history, test_direct_writer_api_is_not_exposed; tests/test_state_pipeline_contracts.py::test_observation_and_changeset_serialization_round_trip.	Focused state pipeline batch plus full pytest gate.
Meters	S-meter/meter samples mutated state but waited for unrelated revision bumps before Web saw them.	tests/test_state_store.py::test_meter_delta_is_visible_without_unrelated_follow_up_revision; tests/test_state_pipeline_diagnostics.py::test_civ_meter_write_records_diagnostic_without_notify_or_revision, test_web_meter_write_delivers_state_store_revision_without_unrelated_trigger; tests/test_web_server_coverage.py::test_state_store_s_meter_change_broadcasts_without_legacy_revision_event, test_meter_only_state_store_change_emits_web_delta_without_legacy_revision; tests/test_civ_rx_coverage.py::test_meter_coalescing_applies_latest_due_sample_and_records_diagnostics, test_same_value_coalesced_meter_flush_completes_scheduler_request; tests/test_radio_poller_coverage.py::test_high_tier_emits_s_meter_on_rx_for_consecutive_cycles, test_high_tier_rotates_pwr_swr_alc_on_tx, test_low_tier_emits_at_expected_stride_for_lan.	Live v2 Playwright plus IC-7610 LAN hardware for real cadence.
Frequency latency	Slow serial or request/response radios could show delayed panel tuning updates.	tests/test_radio_poller_coverage.py::test_set_freq_readback_is_applied_as_state_store_observation, test_set_freq_readback_uses_original_command_correlation, test_set_freq_readback_reconciles_only_matching_websocket_session_for_reused_command_id, test_scheduler_polling_does_not_starve_user_command_queue; tests/test_rigctld_handler.py::test_get_freq_projects_state_store_snapshot, test_get_freq_ensure_fresh_requests_bounded_projection, test_get_freq_stale_store_value_falls_through_to_radio_readback, test_set_freq_enters_command_service_and_applies_observation.	Fake live Web v2 and X6200 serial hardware latency notes.
Stale fields	A missed or malformed unsolicited frame could leave old state looking authoritative.	tests/test_state_store.py::test_dropped_event_marks_stale_and_requests_reconciliation, test_freshness_service_marks_stale_and_queues_reconciliation_without_web, test_observation_refreshes_stale_field_without_semantic_state_change; tests/test_acquisition_scheduler.py::test_stale_unsolicited_field_queues_reconciliation_and_accepts_readback, test_model_service_queues_when_field_observation_max_age_expired_without_mark_stale; tests/test_civ_rx_coverage.py::test_same_value_stale_refresh_notifies_state_store_changed; tests/test_rigctld_handler.py::test_get_freq_headless_freshness_service_stale_value_queues_reconciliation.	Hardware missed-event recovery when available.
Snapshots	HTTP snapshots and initial WebSocket state could disagree when built from different revision sources.	tests/test_web_server_coverage.py::test_http_snapshot_matches_initial_ws_full_state_for_same_store_revision, test_initial_full_state_envelope_revisions_match_ingested_legacy_state, test_http_and_ws_full_state_share_post_sync_legacy_snapshot, test_state_response_includes_revision_and_updated_at; tests/test_delta_encoder.py::TestDeltaEncoder::test_transport_sequence_is_split_from_canonical_state_revision.	Live v2 Playwright.
WebSocket reconnect	Reconnect/reset logic could confuse canonical state revision with transport sequence.	tests/test_delta_encoder.py::TestDeltaEncoder::test_transport_sequence_is_split_from_canonical_state_revision; tests/test_web_server_coverage.py::test_initial_full_state_envelope_does_not_consume_broadcast_delta, test_on_radio_reconnect_enables_scope_without_waiting_for_broadcast; tests/test_websocket_coverage.py::test_keepalive_loop_sends_ping, test_close_idempotent, frame parsing/close-path tests.	Live v2 Playwright reconnect observation; hardware optional.
Web delivery cleanup	MOR-347 removed Web poller-owned public revision; regression could reintroduce poller revisions or legacy sync in delivery.	tests/test_state_pipeline_contracts.py::test_web_poller_no_longer_exposes_public_revision_delivery_api, test_web_server_state_change_callback_does_not_bump_poller_revision, test_web_delivery_methods_snapshot_only, test_legacy_state_sync_is_not_called_from_web_server_delivery; docs/internals/legacy-state-writer-inventory.md.	Full pytest and review of changed Web files.
rigctld GET	rigctld GET could read stale local cache instead of shared confirmed state.	tests/test_rigctld_handler.py::test_get_freq_projects_state_store_snapshot, test_get_mode_projects_state_store_snapshot, test_get_ptt_projects_state_store_snapshot, test_get_vfo_projects_state_store_active_slot, test_get_level_strength_projects_state_store_snapshot, test_get_level_af_projects_state_store_snapshot, test_get_func_projects_state_store_snapshot, dual-RX state-store projection tests near test_dual_rx_get_freq_vfob_projects_state_store_receiver_1 and test_dual_rx_get_mode_vfob_projects_state_store_receiver_1.	External rigctld/Hamlib hardware or fake integration when claiming provider readiness.
rigctld SET / read-after-write	SET behavior could depend on unscoped pending/cache state.	tests/test_rigctld_handler.py::test_get_freq_keeps_optimistic_value_until_radio_state_catches_up, test_get_mode_read_after_write_uses_pending_overlays_until_reconciled, test_get_mode_pending_overlays_are_scoped_to_rigctld_session, test_get_ptt_keeps_optimistic_state_until_radio_state_catches_up, test_set_freq_enters_command_service_and_applies_observation; tests/test_radio_poller_coverage.py::test_queued_command_failure_emits_failed_lifecycle_and_expires_overlay, test_queued_core_timeout_emits_timed_out_lifecycle_and_expires_overlay.	rigctld semi-integration plus Hamlib/provider hardware when in scope.
Pending overlays	Optimistic values could leak between Web, HTTP, rigctld sessions, or reused command IDs.	tests/test_web_server_coverage.py::test_websocket_reused_command_ids_are_scoped_per_connection, test_http_command_batch_timeout_marks_command_timed_out_and_expires_overlay; tests/test_radio_poller_coverage.py::test_set_freq_readback_reconciles_only_matching_websocket_session_for_reused_command_id, test_mark_queued_command_failed_scopes_reused_command_ids_by_source, test_mark_queued_command_failed_scopes_reused_command_ids_by_session; tests/test_rigctld_handler.py::test_get_mode_pending_overlays_are_scoped_to_rigctld_session.	Live v2 command audit.
Dropped unsolicited events	Push-capable backends can lose an event, so revision alone is not proof of physical state.	tests/test_state_store.py::test_dropped_event_marks_stale_and_requests_reconciliation; tests/test_acquisition_scheduler.py::test_stale_unsolicited_field_queues_reconciliation_and_accepts_readback; tests/test_civ_rx_coverage.py::test_scheduler_active_freq_mode_request_completes_from_civ_rx_loop, test_update_state_cache_records_scheduler_result_for_matching_pending_request.	Hardware missed-event recovery when available.
Adaptive acquisition	Background polling could delay commands or flood slow links.	tests/test_acquisition_scheduler.py::test_duplicate_requests_coalesce_with_highest_priority_and_urgent_deadline, test_same_family_requests_share_one_acquisition_request, test_user_facing_requests_preempt_background_telemetry, test_due_polling_emits_only_pollable_cadence_fields_and_groups_by_existing_key, test_due_polling_does_not_reemit_pending_cadence_request, test_unchanged_acquisition_result_decays_cadence_exponentially_and_caps, test_semantic_change_resets_adaptive_cadence_to_base_policy, test_diagnostics_include_cadence_next_due_pending_pressure_and_counts; tests/test_radio_poller_coverage.py::test_scheduler_due_request_sends_supported_civ_query_once, test_scheduler_polling_does_not_starve_user_command_queue.	X6200/Yaesu hardware or simulator notes for slow-link cadence.
Command responses	Command acknowledgements could be mistaken for confirmed state, or confirmed responses could bypass state revisions.	tests/test_state_pipeline_contracts.py::test_command_intent_and_lifecycle_event_serialization_round_trip; tests/test_radio_poller_coverage.py::test_command_queue_entries_preserve_command_correlation_metadata, test_set_freq_readback_is_applied_as_state_store_observation, test_observable_queue_commands_apply_state_store_observations, test_compatibility_mirror_commands_reconcile_from_state_store_observation, test_select_vfo_readback_uses_original_command_correlation; tests/test_civ_rx_coverage.py::test_update_state_cache_applies_command_response_observation_once.	Live v2 command audit and full pytest.
Mode/buttons	Mode, VFO, split, PTT, and button-like controls could diverge between legacy state and shared state.	tests/test_rigctld_handler.py::test_get_mode_projects_state_store_snapshot, test_get_split_vfo_reads_state_store_split_and_protocol_local_tx_vfo, test_set_mode_uses_core_contract_string_values; tests/test_radio_poller_coverage.py::test_execute_quick_dw_trigger_equalizes_then_enables_dw, test_execute_quick_split_trigger_equalizes_then_enables_split, test_select_vfo_readback_uses_original_command_correlation; tests/test_bsr_band_switching.py::TestSetBandBSRRecall::test_bsr_recall_updates_state, test_bsr_recall_emits_events.	Live v2 Playwright.
Encoder-style controls	Knob/encoder controls could overwrite wrong receiver/global paths or bypass command correlation.	tests/test_civ_rx_coverage.py::test_update_radio_state_cmd14_receiver_dsp_levels, test_update_radio_state_cmd14_global_dsp_levels, test_update_radio_state_operator_toggle_family, test_update_state_cache_uses_slot_specific_observation_path_when_override_active; tests/test_radio_poller_coverage.py::test_execute_receiver_routed_set_commands_use_backend_receiver_and_target_state, test_execute_set_filter_width_dispatches_to_radio_protocol, test_execute_set_scope_edge_updates_state, test_execute_set_scope_vbw_updates_state, test_execute_set_scope_rbw_updates_state.	Live v2 Playwright and hardware operator notes.
External rigctld client / Hamlib	Hamlib-backed responses must behave like observations, not a separate cache; direct libhamlib remains out of Core.	tests/test_rigctld_handler.py state-store projection/read-after-write coverage; tests/test_rigctld_server.py::TestSemiIntegrationSerialMockRadio::test_get_and_set_frequency_flows_through_core, test_get_and_set_mode_flows_through_core; tests/integration/test_rigctld_wsjtx.py when integration tests are explicitly run. Boundary decision documented in docs/internals/hamlib-provider-rollout.md.	External rigctld hardware/provider checklist before Hamlib readiness claim.
Cleanup guards	Waiters, watchdogs, stale backlog, and lifecycle servers could leak or deliver stale data after migration.	tests/test_civ_rx_coverage.py::test_cleanup_stale_civ_waiters_logs_cleaned_count, test_civ_rx_loop_sheds_stale_scope_backlog_but_keeps_control_packets, watchdog stop/reconnect tests; tests/test_lifecycle_diagnostics.py WebServer/SerialMockRadio lifecycle diagnostics; tests/test_state_pipeline_contracts.py static cleanup guards.	Full pytest; manual check that fake live server is stopped after Playwright.
Frontend v2 and i18n	Frontend could consume stale revisions, private live URLs, regress localization/UI smoke, or audit controls that only mirrored RadioState rather than observation-backed live state.	tests/test_frontend_e2e_config.py::test_frontend_e2e_does_not_default_to_private_live_backend; frontend/src/lib/i18n/__tests__/*; frontend/tests/e2e/i18n/i18n-visual.spec.ts; live frontend/tests/e2e/v2-ui-interactive.impl.ts via RIGPLANE_V2_URL. MOR-437 made the in-scope Icom v2 field families observation-backed via CI-V emitters in runtime/_civ_rx.py, and the fake SerialMockRadio is now ObservationPollable, so the live v2 audit exercises live controls through the observation pipeline rather than asserting against a RadioState mirror.	npm run build, npm run check, npx vitest run, npm run test:e2e:i18n, live v2 Playwright. Residual non-blocking items: the mode-gated DSP/CW Pitch audit case is a known-fail (CW Pitch only writable in CW modes); Yaesu/rigctld-client provider observation gaps remain open under MOR-424.

Hardware Validation Matrix

Hardware cases are outside CI. Mark each run as blocker-pass, blocker-fail, waived, or informational.

Hardware case	Purpose	Release status
IC-7610 LAN CI-V unsolicited meter/frequency burst	Compare real CI-V ingress rate with state revision and Web delivery rate.	Blocker for IC-7610 LAN release readiness unless explicitly waived.
IC-7610 LAN missed-event recovery	Drop or ignore one captured unsolicited frame and verify reconciliation read corrects shared state.	Blocker when claiming missed-event repair on IC-7610 hardware.
X6200 serial panel tuning latency	Measure bounded frequency update latency after external VFO tuning with and without unsolicited frames.	Blocker for serial-readiness claim; informational otherwise.
Yaesu CAT polling profile	Verify request/response polling updates shared state without model-specific state delivery branches.	Blocker for backend-neutral polling claim.
External Hamlib rigctld client	Verify real rigctld GET/SET responses publish observations and preserve read-after-write/read-only behavior.	Blocker for Hamlib-provider readiness claim.
WebSocket reconnect against live radio	Verify reconnect full-state envelope preserves canonical state/freshness revisions while transport sequence resets.	Informational unless reconnect regression appears in fake live or hardware run.
Operator smoothness notes	Capture perceived meter jitter, knob feel, and UI smoothness.	Informational only; convert to blocker only with reproducible failing steps.

Residual Non-Blocking Compatibility Shims

• Legacy Web revision remains and aliases stateRevision.
• WebSocket transportSeq remains additive ordering metadata and is not a state/freshness revision.
• sync_state_store_from_radio_state(...) remains a one-way compatibility adapter at explicit ingress/startup points only.
• Mutable RadioState, StateCacheCapable, and rigctld cache fallback remain for older public/backend compatibility.
• MOR-437 migrated the in-scope Icom v2 field families to observation-backed CI-V emitters and made the fake SerialMockRadio ObservationPollable. Some non-v2 Icom families remain compatibility-only deferred_follow_up mirrors (see docs/internals/legacy-state-writer-inventory.md), and the Yaesu/rigctld-client provider observation gaps stay open under MOR-424.
• Hardware latency is measured evidence unless a release explicitly promises a numeric hardware target.