[gdluck]

Requested by A

Summary

Before: OWNd has no WHO 1004 (thermoregulation autodiagnostics) classes, no WHO 16 FM tuner step or station-store builders, and no WHO 22 manual-step builders; a WHO 22 dimension 11 report with preset 0 is indistinguishable from one without a preset field.

After: OWNHeatingDiagnosticEvent / OWNHeatingDiagnosticCommand decode and request WHO 1004 diagnostics with the libwhere.so / libdiag.so grammar; OWNSoundCommand.step_frequency_up / step_frequency_down (WHAT 5001..5015 / 5101..5115) and store_station (1..5) build the WHO 16 frames the firmware ACKs; OWNSoundDiffusionCommand.seek_up / seek_down (bare WHAT 5 / 6) and step_frequency_up / step_frequency_down (5#STEP / 6#STEP, 1..15) build the WHO 22 frames; OWNSoundDiffusionEvent exposes frequency_step, the seek / step flags and is_preset_cleared.

Context. Companion of OpenWebNet-HA/MyHOME#839 (andrea-mag's H/L4684 6.0.11 captures on OpenWebNet-HA/MyHOME#466), for 2.0.0b13; the CHANGELOG bullets sit under a new [Unreleased] / Added above the shipped 2.0.0b12 section. Revision 3b76ab1 (on master 360e4f4, after #152 merged) implements @GreenGrassBlueOcean's corrected, firmware-verified specification (review 5738866331, findings F1..F7), replacing the first spec from #152; #152 ships 2.0.0b12 without it.

Changes

1. OWNd/message/heating.py (F1, F2, F3): OWNHeatingDiagnosticEvent scopes per OPN_ThermoWhereTransl: #0 / #0#C -> central; #Z / #Z#C -> zone (zone_number); Z 1..99 -> probe; NZZ -> slave_probe (zone_number, probe_number); [1-9]00 -> external_probe; Z#N -> actuator (zone_number, actuator_number); anything else (0, 901, 66#0, #0#0, #66#1#2) -> None. unique_id is 1004-<full WHERE> on both classes (1004-#0#2, 1004-#66#2, 1004-66#1), so central-unit selectors and actuator numbers no longer collide; 1004 stays out of OWNMessage.interface. diagnostic_mask: 24-bit 0 / 1 string for dimensions 7, 8, 11, 12 on central and the addressed devices, 16-bit for dimensions 21 and 22 on zone, exactly one value, else None; cleared_bits 1-based from the left, has_fault; the docstring notes the WHO_4 v2.0.0 pp. 69-74 active-low bit tables while keeping the mask raw. Dimension 23 on central with two ASCII-digit values -> non_responding_units, faulty_units. slave_count, scalar faulty_zone and the dimension 13 "local offset" claim are gone. Builders, only the keys libdiag.so registers: request_central_diagnostic(where="#0", *, dimension=7) (7 / 8 / 11 / 12, _validate_central_where), request_device_diagnostic(where, *, dimension=7) (Z, NZZ, [1-9]00, Z#N), request_zones_with_failures (*20##), request_zone_diagnostic (*21##; #0, #0#C, 1..99 -> #Z, #Z, #Z#C), request_fault_counts (*23##). No *22## request (R22P0 does not exist). Every integer validator rejects bool and non-int / str before int(). The is_who4_dimension_22 / is_who4_dimension_22_crash_form negative guards for *#1004*#66*22## stay.
2. OWNd/message/sound.py, WHO 16 (F4, F5): step_frequency_up -> *16*{5000+s}*W##, step_frequency_down -> *16*{5100+s}*W## (1..15); OWNSoundEvent.frequency_step_forward / frequency_step_backward from 5001..5015 / 5101..5115 (4001 / 4101 no longer decoded); store_station restricted to 1..5 with the note that 6..15 need WHO 22 store_preset; the dimension 10 event decode is gone (bt_difson never emits it).
3. OWNd/message/sound.py, WHO 22 (F6, F7a): step_frequency_up(where="6", step=1) -> *22*5#{s}*W##, step_frequency_down -> *22*6#{s}*W## (1..15); WHAT 7 / 8 removed; frequency_step (1..15, else None) with is_step_frequency_up / is_step_frequency_down on 5#STEP / 6#STEP (a step of 0, 16 or 01 sets no flag); is_seek_up / is_seek_down for bare WHAT 5 / 6; is_preset_cleared kept as defensive parsing of *#22*…*11*1*<freq>*0##, without the F500-trace attribution.
4. Citations (F7b): the three kb_who_* slugs are replaced everywhere by the ownkb:* ids of the review. scripts/sync_documentation.py / README.md: the 4 / 1004 catalogue row describes the corrected scopes and dimensions. CHANGELOG.md: three bullets under [Unreleased] / Added.

State Diagram

stateDiagram-v2
    direction TB
    state "WHO 1004 WHERE (OPN_ThermoWhereTransl) -> scope" as Scope {
        [*] --> Central : #0, #0#C (dims 7 8 11 12 -> 24 bits, 20 request, 23 -> two counts)
        [*] --> Zone : #Z, #Z#C (dims 21, 22 -> 16 bits)
        [*] --> Probe : Z 1..99 (24 bits)
        [*] --> SlaveProbe : NZZ, N 1..8 (24 bits, probe_number)
        [*] --> ExternalProbe : 100..900 (24 bits)
        [*] --> Actuator : Z#N, N 1..9 (24 bits, actuator_number)
        [*] --> Unknown : 0, 901, 66#0, #0#0, #66#1#2 -> scope None, mask None
    }
    state "WHO 16 / WHO 22 tuner (golden oracle rows)" as Tuner {
        [*] --> W16Step : *16*5001..5015*W## / *16*5101..5115*W## ack on 6 gateways
        [*] --> W16Store : *#16*W*#10*1..5## ack on 6 gateways
        [*] --> W22Seek : *22*5*W## / *22*6*W## ack on 8 gateways
        [*] --> W22Step : *22*5#1..15*W## / *22*6#1..15*W## ack on 8 gateways
    }

Loading

Where are the traces

Subsystem	Capture	Plant	Frames verified by this PR
WHO 1004	OpenWebNet-HA/MyHOME#839 tests/fixtures/traces/issue_466/myhome_trace_HL4684_all_2026-11-01T11-30-19.json	H/L4684 6.0.11 + 3550 central unit + F430/4, zone 66	*#1004*66*11*111111111111111111111111## (probe), *#1004*#66*22*1111111111111111## (zone), *#1004*#0*11*111111111111111111111111## (central), verbatim in the tests
WHO 16 / WHO 22 F500 tuner	same PR, ...T11-48-15.json	H/L4684 6.0.11 + F500	*#16*101*6*0*105300##, *#16*101*7*0*5##, *#22*5#2#1*5*1*10530##, *#22*5#2#1*11*1*10530*5##. The off-preset *0## shape is not in the trace and is tested as a spec-shaped frame only.

What does the Oracle say

tests/golden/firmware_oracle.json, pinned by the tests:

Frame	Result
*16*5001*101##, *16*5101*101##	ack / out $0494018FA1 / $0494018FB1 on F453, F453AV, F454, H4684, MH200, MH200N; nack / silent on the other six. The old 4001 / 4101 spellings are asserted absent.
*#16*101*#10*5##	ack / out $0493018FD4 on the same six
*22*5*2#1##, *22*6*2#1##, *22*5#1*2#1##, *22*6#1*2#1##	ack / out $0494018FA0 / FB0 / FA1 / FB1 on F453, F453AV, F454, F459, F460, F461, MH200N, MyHomeServer1; MH202 nack / out; F450, H4684, MH200 nack / silent
any WHO 1004 frame	not replayed yet; the decoders rest on the physical frames above and on libdiag.so as the review documents

Testing

• tests/test_who1004_heating_diagnostic.py: every scope on the physical frames and on edge WHEREs (0, 100, 901, 66#0, #0#0, #66#1#2, 10#1, 166, 800), the per-scope mask dimensions (a 16-bit mask on a probe or a 24-bit mask on a zone gives None), cleared_bits / has_fault, dimension 23 counts, unique_id with every WHERE form, all five builders with rejected inputs (bool, float, out of range, wrong scope), the dimension 22 guard regressions, OWNEvent.parse == OWNMessage.parse.
• tests/test_sound.py: the 5001 / 5101 steps and the 1..5 store with rejected inputs, the event decode, and the golden-row pins above. tests/test_sound_diffusion.py: the 5#STEP / 6#STEP builders and decode, frequency_step edges (0, 16, 01), bare seeks, is_preset_cleared, the four oracle rows in the WHO 22 conformance parametrisation plus a positive pin on all eight ack gateways.
• Full suite: 2786 passed; scripts/verify_coverage.py 6,986 / 6,986 statements; branches 100 % on heating.py, sound.py adds no partial branch (its 11 pre-exist on master); ruff check, mypy OWNd, scripts/verify_library_standards.py and scripts/sync_documentation.py --check pass.

Judgement calls

• is_step_frequency_up / is_step_frequency_down are True only with a valid step 1..15 (5#0 aliases the seek on bt_multi, 5#16 overflows; neither sets a flag).
• request_central_diagnostic defaults to dimension 7 (the review's ownkb:chunk:r000205); the first revision defaulted to 11.
• The 4 / 1004 catalogue row text was rewritten because its old wording contradicted the corrected scopes; row structure and dispatch keys are unchanged.
• First-listed names used wherever the review offered two (zone, slave_probe, request_device_diagnostic).

🤖 Generated with Claude Code

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[GreenGrassBlueOcean]

Hi @gdluck — thank you for putting this PR together so cleanly and thoroughly, and huge credit for flagging the three oracle/trace discrepancies in your PR body:

1. *16*4001*101##, *16*4101*101##, *22*7*2#1##, *22*8*2#1##, and the cited firmware versions being absent from firmware_oracle.json, while noting that the functional docs list *22*5#STEP*2#S## / *22*6#STEP*2#S##;
2. *#16*101*#10*5## emitting $0493018FD4 ($0494018FA0 being *22*5*2#1##); and
3. *#22*5#2#1*11*1*10530*0## not appearing in myhome_trace_HL4684_all_2026-11-01T11-48-15.json.

Every single one of those catches was right on the money. I audited the implementation and my earlier specification comment (#152 (comment)) directly against libdiag.so, libwhere.so, bt_difson, and bt_multi disassemblies, tests/golden/firmware_oracle.json, and the OpenWebNet-Encyclopedia Machine KB (ownkb-build-0.9.0). All of the protocol defects below originated in my #152 spec comment — my apologies for handing you an unverified spec.

Let's reset #152 back to 76b9550c so 2.0.0b12 ships cleanly right away, and rework #153 on top of master for 2.0.0b13 using the verified specification below. (OWNd/message/__init__.py, scripts/sync_documentation.py, WHO 22 seek_up/seek_down + oracle rows, and the is_who4_dimension_22 negative guards are already clean and reusable as-is.)

---

1. Summary of Protocol Findings (F1–F7)

ID	Severity	Location	Finding & Primary-Source Proof
F1	Critical	heating.py (request_zone_actuator_diagnostic)	Emits *#1004*#Z*22##. Across libdiag.so (F453, F454, F450, MH200N, MH202, MyHomeServer1, F460, F461), OPN_OpenScsDiagTermoTr registers only R20P0, R21P0, and R23P0 on top of base R0P0..R12P0. R22P0 does not exist ("Termo Object command not Found for key: R22P0" -> *#*0## NACK); DIMENSION 22 is spontaneous s2o-only (ownkb:chunk:r000205, "No request defined in this flow"). Single-zone queries use *#1004*#Z*21## (R21P0).
F2	Critical	heating.py (OWNHeatingDiagnosticEvent)	Wrong wire semantics for DIM 7, 13, 20, 21, 23: • DIM 7 (s2o_diagnosticsA, ownkb:claim:c004423): 24-bit '0'/'1' diagnostic mask on #0/#0#C and addressed devices (mirroring OWNLightingDiagnosticEvent _MASK_DIMENSIONS = (7, 11)), never 0..8 slave_count (DIM 8 & DIM 12 are also 24-bit masks). • DIM 21 (s2o_autoDiagnostics in libdiag.so): zone == 0 is rejected ('Invalid zone value', so #0*21*0 is never emitted); 1 <= zone <= 99 jumps to the exact same 16-bit '0'/'1' mask formatter as DIM 22, emitting *#1004*#ZONE*21*BIT16## on #Z / #Z#C (currently fails closed to None). • DIM 23 (s2o_autoDiagnostics zone == 100, ownkb:claim:c004442): emitted on #0 / #0#C as two decimal integer counts *#1004*#0*23*NO_RESPONSE*FAULTS##, not per-actuator 1..9 on #Z. • DIM 20 (R20P0) is a request-only trigger returning zero or more *#1004*#ZONE*21*BIT16## frames; DIM 13 in OPN_OpenScsDiagIdTr is a generic 32-bit hex Device ID, not WHO 4 local offset.
F3	Medium	heating.py (WHERE & unique_id)	OPN_ThermoWhereTransl (libwhere.so) accepts #0/#0#C, #Z/#Z#C (zone Z under 4-zone central unit C), Z (1..99 master probes), NZZ (101..899 slave probes, N=1..8, ZZ=01..99), [1-9]00 (100..900 external probes), and Z#N (Z=1..99, N=1..9 zone actuators such as 66#1). Also, because 1004 is (rightly) excluded from OWNMessage.interface, inherited OWNMessage.unique_id drops _where_param, colliding #0/#0#1/#0#2, #66/#66#2, and 66/66#1.
F4	Critical	sound.py (WHO 16 step_frequency_*)	Emits/decodes WHAT 4001..4015 / 4101..4115, which bt_difson NACKs on every WHO 16 gateway. Per ownkb:chunk:r000371 (ownkb:claim:c002096, c002098) and bt_difson o2s, manual frequency steps are 5001..5015 / 5101..5115 (5000/5100 = auto-seek) — and *16*5001*101## / *16*5101*101## are already in tests/golden/firmware_oracle.json (ack / out $0494018FA1 / $0494018FB1 across F453, F453AV, F454, H4684, MH200, MH200N).
F5	High	sound.py (WHO 16 store_station & DIM 10)	OWNSoundCommand.store_station allows 6..15, which bt_difson (cmp r0, #5; bgt NACK) NACKs on every WHO 16 gateway (ownkb:claim:c002128: "For DIMENSION 10, Established detail is station range 1 through 5"; 6..15 requires WHO 22 OWNSoundDiffusionCommand.store_preset *22*33#P*2#S##). And bt_difson s2o only formats dimensions 1, 6, 7, 8, 11 — never DIMENSION 10 (ownkb:claim:c002135), and *#16*101*10## is NACKed.
F6	High	sound.py (WHO 22 step_frequency_*)	Emits/decodes WHAT 7 / WHAT 8, which bt_multi logs as "Increment/Decrement subSource - not implemented yet" and NACKs (*#*0##), while leaving real manual steps *22*5#STEP*2#S## / *22*6#STEP*2#S## (1..15, ownkb:claim:c002729) with is_step_frequency_up == False. *22*5#1*2#1## and *22*6#1*2#1## are already in tests/golden/firmware_oracle.json (ack / out $0494018FA1 / $0494018FB1 across all 8 WHO 22 gateways).
F7	Medium-High	sound.py, CHANGELOG.md, tests	(a) sound.py docstrings/comments and CHANGELOG.md attribute *#22*5#2#1*11*1*10530*0## to (H/L4684 6.0.11 F500 trace, MyHOME#839) (bt_multi emits DIM 5 and skips DIM 11 when SCS station nibble is 'F', and *0## is absent from that trace). (b) kb_who_04_thermoregulation_diagnostics, kb_who_16_what_frequency_step, and kb_who_16_dim_station (13 occurrences across 5 files) do not exist in the Machine KB (ownkb:* IDs listed below).

flowchart LR
    subgraph WHO16["WHO 16 (bt_difson)"]
        W16Step["*16*5001..5015*101## / *16*5101..5115*101##"] -->|"ACK + bus"| BusStep["$0494018FA1..FAF / FB1..FBF"]
        W16Store["*#16*101*#10*1..5## (write-only)"] -->|"ACK + bus"| BusStore["$0493018FD0..FD4"]
    end
    subgraph WHO22["WHO 22 (bt_multi)"]
        W22Seek["*22*5*2#1## / *22*6*2#1##"] -->|"ACK + bus"| BusSeek["$0494018FA0 / FB0"]
        W22Step["*22*5#1..15*2#1## / *22*6#1..15*2#1##"] -->|"ACK + bus"| BusStep
        W22Store["*22*33#1..15*2#1##"] -->|"ACK + bus"| BusStore22["$0493018FD0..FDE"]
    end
    subgraph WHO1004["WHO 1004 (libdiag.so)"]
        D7["*#1004*WHERE*7## (R7P0)"] -->|"24-bit mask"| E7["*#1004*WHERE*7*BIT24##"]
        D20["*#1004*#0*20## (R20P0)"] -->|"0..N faulty zones"| E21["*#1004*#Z*21*BIT16##"]
        D21["*#1004*#Z*21## or #0*21## (R21P0)"] -->|"16-bit mask"| E21
        D22["Spontaneous s2o only (no R22P0)"] -->|"16-bit mask"| E22["*#1004*#Z*22*BIT16##"]
        D23["*#1004*#0*23## (R23P0)"] -->|"2 decimal counts"| E23["*#1004*#0*23*NO_RESPONSE*FAULTS##"]
    end

Loading

---

2. Corrected Specification for #153

(Note across all new integer validators in heating.py and sound.py: guard if isinstance(val, bool) or not isinstance(val, (int, str)): raise ValueError(...) so True / False / 1.9 never silently coerce via int().)

A. WHO 1004 Thermoregulation Autodiagnostics (OWNd/message/heating.py + OWNd/message/__init__.py)

• Verified Machine KB citations: ownkb:chunk:r000204–r000209, ownkb:claim:c004423, ownkb:claim:c004429–c004448 (WHO_4 v2.0.0 pp. 69–74), plus libdiag.so (OPN_OpenScsDiagGenTr, OPN_OpenScsDiagTermoTr) and libwhere.so (OPN_ThermoWhereTransl).
• WHERE grammar & scopes (OPN_ThermoWhereTransl):
  1. Central unit (#0 or #0#<1..99>, _WHO4_CENTRAL_WHERE_REGEX) -> scope == "central", zone_number is None.
  2. Central-unit zone (#<1..99> or #<1..99>#<1..99> — zone Z on central unit / 4-zone central unit C, matching heating.py:116-120) -> scope == "zone" (or "zone_actuators"), zone_number == Z.
  3. Addressed bus devices:
     • Master probe 1..99 (Z) -> scope == "probe", zone_number == Z.
     • Slave probe NZZ (N = 1..8, ZZ = 01..99, i.e. 101..199, 201..299, …, 801..899) -> scope == "slave_probe" (or "probe"), zone_number == ZZ, probe_number == N.
     • External probe [1-9]00 (100, 200, …, 900) -> scope == "external_probe", zone_number is None.
     • Zone actuator Z#N (Z = 1..99, N = 1..9, e.g. 66#1) -> scope == "actuator", zone_number == Z, actuator_number == N.
     • Anything else (0, 901, 66#0, #0#0, #66#1#2, etc.) -> scope is None, zone_number is None.
• unique_id collision fix:
  • Keep 1004 excluded from OWNMessage.interface.
  • Override unique_id on both OWNHeatingDiagnosticEvent and OWNHeatingDiagnosticCommand using the full reconstructed WHERE (#0#2 -> "1004-#0#2", #66#2 -> "1004-#66#2", 66#1 -> "1004-66#1") so _where_param is never dropped.
• OWNHeatingDiagnosticEvent properties (fail-closed):
  • diagnostic_mask: str | None:
    • 24-bit '0'/'1' mask when dimension in (7, 8, 11, 12) and scope in ("central", "probe", "slave_probe", "external_probe", "actuator"), len(dimension_value) == 1, len(v) == 24, set(v) <= {"0", "1"}.
    • 16-bit '0'/'1' mask when dimension in (21, 22) and scope == "zone" (or "zone_actuators"), len(dimension_value) == 1, len(v) == 16, set(v) <= {"0", "1"}.
    • Otherwise None.
  • cleared_bits: tuple[int, ...] | None (1-based positions from left where bit == '0') and has_fault: bool | None (bool(cleared_bits) when diagnostic_mask is not None, else None). Note in docstring (ownkb:chunk:r000206–r000207) that WHO_4 v2.0.0 pp. 69–74 documents active-low bit conditions (13..16, 21..24 for 24-bit central masks; 11..16 for 16-bit zone masks) while keeping diagnostic_mask raw until hardware fault injection confirms bit-order indexing on the bus.
  • DIMENSION 23 fault-count summary on scope == "central" (*#1004*#0*23*NO_RESPONSE*FAULTS##, ownkb:claim:c004442): when dimension == 23, scope == "central", len(dimension_value) == 2, and both values are ASCII digits, expose non_responding_units: int | None and faulty_units: int | None (else None).
  • Remove slave_count, scalar faulty_zone, and DIM 13 local-offset claims.
• OWNHeatingDiagnosticCommand builders (only keys registered in libdiag.so: R7P0, R8P0, R11P0, R12P0, R20P0, R21P0, R23P0 — remove request_zone_actuator_diagnostic / *22##):
  • request_central_diagnostic(cls, where: str = "#0", *, dimension: int = 7) -> OWNHeatingDiagnosticCommand -> *#1004*{target}*{dimension}## (dimension in (7, 8, 11, 12), defaulting to 7 per ownkb:chunk:r000205, validated with _validate_central_where).
  • request_device_diagnostic(cls, where: int | str, *, dimension: int = 7) -> OWNHeatingDiagnosticCommand (or request_probe_diagnostic, dimension in (7, 8, 11, 12) on valid addressed bus devices Z, NZZ, [1-9]00, Z#N).
  • request_zones_with_failures(cls, where: str = "#0") -> OWNHeatingDiagnosticCommand -> *#1004*{target}*20## (R20P0, _validate_central_where).
  • request_zone_diagnostic(cls, where: int | str = "#0") -> OWNHeatingDiagnosticCommand -> *#1004*{target}*21## (R21P0: accepts #0, #0#C, 1..99 normalized to #Z, #Z, or #Z#C).
  • request_fault_counts(cls, where: str = "#0") -> OWNHeatingDiagnosticCommand -> *#1004*{target}*23## (R23P0, _validate_central_where).

B. WHO 16 FM Tuner Steps & Preset Store (OWNd/message/sound.py)

• Verified citations: ownkb:chunk:r000371, ownkb:chunk:r000376, ownkb:claim:c002096, ownkb:claim:c002098, ownkb:claim:c002128, ownkb:claim:c002135, plus tests/golden/firmware_oracle.json (sound-who16-source).
• Frequency steps (5001..5015 / 5101..5115):
  • OWNSoundCommand.step_frequency_up(cls, where: str | int, step: int = 1) -> *16*{5000 + s}*{target}## (1 <= s <= 15).
  • OWNSoundCommand.step_frequency_down(cls, where: str | int, step: int = 1) -> *16*{5100 + s}*{target}## (1 <= s <= 15).
  • OWNSoundEvent: decode 5001 <= self._state <= 5015 -> frequency_step_forward = self._state - 5000 and 5101 <= self._state <= 5115 -> frequency_step_backward = self._state - 5100.
  • Pin *16*5001*101## and *16*5101*101## positively against tests/golden/firmware_oracle.json (ack / out $0494018FA1 and $0494018FB1 on F453, F453AV, F454, H4684, MH200, MH200N).
• Station store (1..5 write-only):
  • Restrict OWNSoundCommand.store_station(cls, where: str | int, station: int) to 1 <= station <= 5 (ownkb:claim:c002128, bt_difson ceiling), noting in the docstring that presets 6..15 require WHO 22 OWNSoundDiffusionCommand.store_preset (*22*33#<P>*2#<S>##).
  • Revert OWNSoundEvent back to elif self._dimension == 7: (DIMENSION 10 is write-only and never emitted by bt_difson). (We also cleaned up the matching downstream MyHOME check in OpenWebNet-HA/MyHOME#844.)

C. WHO 22 Seek, Frequency Steps & is_preset_cleared (OWNd/message/sound.py)

• Verified citations: ownkb:chunk:r000478–r000480, ownkb:claim:c002729, plus tests/golden/firmware_oracle.json (sound-who22).
• Keep OWNSoundDiffusionCommand.seek_up (*22*5*{target}##) and seek_down (*22*6*{target}##) + is_seek_up / is_seek_down (bare WHAT 5 / 6 with no parameter) as-is.
• Manual frequency steps (*22*5#<1..15>*W## / *22*6#<1..15>*W##):
  • Change OWNSoundDiffusionCommand.step_frequency_up(cls, where: str | int = "6", step: int = 1) -> *22*5#{s}*{target}## and step_frequency_down(cls, where: str | int = "6", step: int = 1) -> *22*6#{s}*{target}##, validating 1 <= s <= 15 (bt_multi 5#0 aliases seek FA0, >=16 overflows %.1X).
  • Add *22*5#1*2#1## and *22*6#1*2#1## to the WHO 22 oracle conformance parametrisation (ack / out $0494018FA1 / $0494018FB1 across all 8 WHO 22 gateways in tests/golden/firmware_oracle.json).
  • Remove WHAT 7 / WHAT 8 from OWNSoundDiffusionEvent and wire is_step_frequency_up / is_step_frequency_down + frequency_step: int | None (1..15, failing closed outside 1..15) to WHAT 5#<step> / 6#<step>.
• is_preset_cleared & KB slugs:
  • Keep is_preset_cleared as defensive parsing for explicit *#22*...*11*1*<freq>*0## frames, removing the (H/L4684 6.0.11 F500 trace, MyHOME#839) attribution from sound.py and CHANGELOG.md.
  • Replace all kb_who_* slugs across CHANGELOG.md, heating.py, sound.py, and tests with the verified ownkb:* IDs above.

[gdluck]

Reworked to the corrected specification (review 5738866331), revision 3b76ab1: four commits on master 360e4f4 (after #152 merged), CHANGELOG bullets under a new [Unreleased] / Added, for 2.0.0b13.

• F1 / F2 / F3: OPN_ThermoWhereTransl scopes (central, zone, probe, slave_probe + probe_number, external_probe, actuator + actuator_number, else None); unique_id = 1004-<full WHERE> on both classes; 24-bit masks on dimensions 7 / 8 / 11 / 12 for the central unit and addressed devices, 16-bit on 21 / 22 for a zone; dimension 23 -> non_responding_units / faulty_units; slave_count, scalar faulty_zone and the dimension 13 claim removed; builders request_central_diagnostic(dimension=7), request_device_diagnostic, request_zones_with_failures, request_zone_diagnostic, request_fault_counts; no *22## request; bool / non-int guards on every validator.
• F4 / F5: WHO 16 steps 5001..5015 / 5101..5115 (builders and decode; 4001 / 4101 asserted absent from the golden file), store_station 1..5 with the WHO 22 store_preset note, dimension 10 decode removed; *16*5001*101##, *16*5101*101##, *#16*101*#10*5## pinned positively against tests/golden/firmware_oracle.json.
• F6: WHO 22 step_frequency_up/down -> *22*5#S*W## / *22*6#S*W## (1..15) with frequency_step and the step flags; WHAT 7 / 8 gone; *22*5#1*2#1## / *22*6#1*2#1## join the conformance parametrisation with the seeks, plus a positive pin on all eight ack gateways.
• F7: the F500-trace attribution of the *0## frame removed; all kb_who_* slugs replaced by the ownkb:* ids you listed.

2786 passed, verify_coverage.py 6,986 / 6,986, ruff, mypy, library standards and sync_documentation.py --check clean. Judgement calls are in the body (step flags only for 1..15; request_central_diagnostic defaults to dimension 7; the 4 / 1004 catalogue row text rewritten to the corrected scopes).

[GreenGrassBlueOcean]

Thank you @gdluck — exemplary rework on 3b76ab1f. Every finding (F1–F7) and all four judgement calls match libdiag.so / libwhere.so / bt_difson / bt_multi, tests/golden/firmware_oracle.json, and the Machine KB (ownkb:*), with comprehensive fail-closed and non-integer/bool guards, synced WHO catalog docs, and all 8 CI gates (including the MyHOME suite) green. Approved and merging!

[GreenGrassBlueOcean]

@gdluck Since #153 is now merged (fb0acd7d) before 2.0.0b12 was published to PyPI, could you open a quick chore/release-2.0.0b13 PR (chore(release): bump version to 2.0.0b13) on top of master (fb0acd7d) so I can approve and cut v2.0.0-b13 directly to PyPI?

1. OWNd/__init__.py: __version__ = "2.0.0b13"
2. CHANGELOG.md: promote the current [Unreleased] section to ## [2.0.0b13] - 2026-11-02 (keeping a fresh empty ## [Unreleased] above it and leaving ## [2.0.0b12] - 2026-11-01 intact below it).