[gdluck]

Requested by A

What is the problem

OpenWebNet-HA/MyHOME pins OWNd==2.0.0b11 and waits for a release that carries the thirteen PRs merged since v2.0.0-b11 (2026-10-20), so that its pin bump can drop the two pending_decoder_rx dimension 27 frames and regenerate the discovery table for the F461 / MyHomeServer1 profiles without WHO 16. Requested by @GreenGrassBlueOcean on #139.

---

How is it solved

1. OWNd/__init__.py: __version__ = "2.0.0b12" (pyproject.toml reads it; tests/test_metadata.py checks PEP 440 and the installed distribution).
2. CHANGELOG.md: ## [Unreleased] promoted to ## [2.0.0b12] - 2026-11-01, with a fresh empty ## [Unreleased] header above it. The section covers feat(lighting): auto-detect pre-09W29 F414 coarse-only dimmers via speed 240 (0xF0) and emit WHAT 2..10 presets #134, fix(lighting): WHAT 39 is the end of movement of a WHO 1 motion detector #135, fix(connection): refuse WHO 4 dimension 22 in every send path #136, fix(heating): set_central_mode requires a central-unit address #137, fix(message): parse '#'-prefixed dimension 14 setpoint values #138, feat(gateway): WHO 13 dimension 27 model-name decoder #139, feat(heating): OWNHeatingCommand.actuators_status(where) builder for *#4*WHERE*20## #140, fix(profiles): drop WHO 16 from the F461 and MyHomeServer1 profiles #142, feat(lighting): WHO 1001 dimension 11 diagnostic event #145, fix(profiles): alias H/L4684 & HC4684 to H4684Profile and enable WHO 22 on FW >= 6.0.0 #147, docs(changelog): describe send_to_gateway's dimension 22 refusal as it is #148, ci(pytest): set asyncio_default_fixture_loop_scope so pytest-asyncio 1.4.1 does not abort the suite #150 and fix(heating): refuse WHO 4 DIMENSION 22 only without an actuator token in WHERE (oracle thermo-dim22) #151. (The MyHOME#839 companion was folded in on 2026-11-01 and taken out again on the maintainer's review 5738875005: it ships reworked in 2.0.0b13 through feat(message): WHO 1004 thermoregulation diagnostics, WHO 16 tuner steps and station store, WHO 22 seek and step builders (companion of MyHOME#839) #153.)

No other file carries the version string; the only other 2.0.0b11 mention is a historical note in a heating.py comment ("the detector of OWNd 2.0.0b11"), left as is.

Changes callers can notice (already described in the section)

• WHO 4 dimension 22 is sent only in its published split form 3#Z#N (fix(connection): refuse WHO 4 dimension 22 in every send path #136, docs(changelog): describe send_to_gateway's dimension 22 refusal as it is #148, fix(heating): refuse WHO 4 DIMENSION 22 only without an actuator token in WHERE (oracle thermo-dim22) #151).
• The central-unit builders require #0 / #0#C (fix(heating): set_central_mode requires a central-unit address #137).
• F461Profile and MyHomeServer1Profile no longer list WHO 16 (fix(profiles): drop WHO 16 from the F461 and MyHomeServer1 profiles #142).
• WHO 13 dimension 27 model names and #-prefixed dimension 14 setpoints parse (feat(gateway): WHO 13 dimension 27 model-name decoder #139, fix(message): parse '#'-prefixed dimension 14 setpoint values #138).

Testing

• pytest tests on this head with the package reinstalled: 2516 passed (tests/test_metadata.py included); ruff check clean.

🤖 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 — quick request alongside this release bump!

We just opened OpenWebNet-HA/MyHOME#839 following @andrea-mag's two new H/L4684 (FW 6.0.11) bus captures on OpenWebNet-HA/MyHOME#466 (comment), which gave us:

1. Our first physical WHO 1004 thermoregulation diagnostic frames (*#1004*66*11*...##, *#1004*#66*22*...##, *#1004*#0*11*...##) alongside a physical H/L4684 WHO 1001 lighting diagnostic frame (*#1001*91*11*111111111111111111111111##); and
2. Our first physical dual-stack WHO 16 + WHO 22 F500 FM radio tuner session (105.30 MHz, Station 5, RDS "MY HOME " -> "DIM*SOFT").

Since @fedem95 is currently unavailable, I cannot approve a PR that I open myself on OWNd — whereas I can review and approve a companion PR if you author it. Would you be willing to write this companion OWNd PR?

(Note: MyHOME#839 already works cleanly on OWNd==2.0.0b11 / 2.0.0b12 using raw frame fallbacks where builders don't exist yet, so this does not need to block merging #152 — whether you prefer to ship 2.0.0b12 as-is and open a follow-up PR for 2.0.0b13, or fold it in first, is completely up to you. I'm ready to review and approve #152 and the companion PR right away.)

Below is the exact, self-contained specification mapped directly to OWNd/message/ so nothing is left ambiguous:

---

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

Mirroring your OWNLightingDiagnosticEvent (OWNd/message/lighting.py:967-1042) and OWNGatewayDiagnosticEvent / OWNGatewayDiagnosticCommand (OWNd/message/gateway.py:643-709), add OWNHeatingDiagnosticEvent(OWNEvent) and OWNHeatingDiagnosticCommand(OWNCommand) in OWNd/message/heating.py, register both via:

register_event_parser(1004, OWNHeatingDiagnosticEvent)
register_command_parser(1004, OWNHeatingDiagnosticCommand)

and re-export both in OWNd/message/__init__.py.

Evidence & Wire Shapes

• Encyclopedia KB: kb_who_04_thermoregulation_diagnostics (WHO 1004, OPEN.db / libdiag.so s2o_selfDiagnosticsA).
• Physical H/L4684 (FW 6.0.11) + 3550 central unit + F430/4 zone 66 trace (tests/fixtures/traces/issue_466/myhome_trace_HL4684_all_2026-11-01T11-30-19.json in MyHOME#839):
  • *#1004*66*11*111111111111111111111111## — Probe / zone 66, DIMENSION 11, 24-bit mask ('1' = healthy, '0' = active fault).
  • *#1004*#66*22*1111111111111111## — Central-unit zone 66 actuators (WHERE = "#66"), DIMENSION 22, 16-bit mask.
  • *#1004*#0*11*111111111111111111111111## — 3550 central unit (WHERE = "#0"), DIMENSION 11, 24-bit mask.
• Safety / Parser Notes:
  • Do not add 1004 to OWNMessage.interface in OWNd/message/base.py:333 ([0, 1, 2, 14, 15, 1001]) unless guarded against #0 / #Z, because WHO 1004 uses #0 (#0#C) for central units and #1..#99 for zone actuators. (base.py:105-110 already parses self._where as "#0", "#66", "66" cleanly.)
  • is_who4_dimension_22 (OWNd/message/heating.py:93-110) anchors on \*#\s*\+?0*4\s*\*, so *#1004*#66*22## is already not blocked by the WHO 4 Dimension 22 crash guard — worth adding an explicit regression assertion assert not is_who4_dimension_22("*#1004*#66*22##") and assert not is_who4_dimension_22("*#1004*#66*22*1111111111111111##").

OWNHeatingDiagnosticEvent(OWNEvent) Scope & Properties

Three WHERE addressing scopes per kb_who_04_thermoregulation_diagnostics:

1. Probe / standalone zone (WHERE = "1".."99") -> scope == "probe", zone_number == int(where):
   • DIMENSION 7: slave_count: int | None (0..8 from single integer dimension_value[0], else None).
   • DIMENSION 8: 24-bit '0'/'1' slave-probe diagnostic mask.
   • DIMENSION 11: 24-bit '0'/'1' master-probe diagnostic mask (*#1004*66*11*111111111111111111111111##).
   • DIMENSION 13: local offset failure (raw preserved in dimension_value).
2. Central unit (WHERE = "#0" or #0#C matching _WHO4_CENTRAL_WHERE_REGEX) -> scope == "central", zone_number is None:
   • DIMENSION 11: 24-bit '0'/'1' central unit diagnostic mask (*#1004*#0*11*111111111111111111111111##).
   • DIMENSION 20: zone failure summary (raw in dimension_value).
   • DIMENSION 21: active fault-zone discovery (faulty_zone: int | None — 0 when no faulty zones, 1..99 when reporting a faulty zone).
3. Zone actuators (WHERE = "#1".."#99", optionally #Z#C) -> scope == "zone_actuators" (or "zone"), zone_number == int(where[1:]):
   • DIMENSION 22: 16-bit '0'/'1' actuator diagnostic mask (*#1004*#66*22*1111111111111111##).
   • DIMENSION 23: per-actuator 1..9 diagnostic status (raw in dimension_value).

Properties (matching OWNLightingDiagnosticEvent fail-closed semantics):

• scope: str | None ("probe" | "central" | "zone_actuators" | None)
• zone_number: int | None (1..99 for probe / zone actuators, None for #0 or unrecognized WHERE)
• diagnostic_mask: str | None — exact '0'/'1' string only when len(dimension_value) == 1, set(value) <= {"0", "1"}, and len(value) == 24 for dimension in (8, 11) or len(value) == 16 for dimension == 22; otherwise None (fail closed on wrong length, trailing *val, or non-binary chars).
• cleared_bits: tuple[int, ...] | None — 1-based positions from the left where bit == "0" (() when all '1', None when diagnostic_mask is None).
• has_fault: bool | None — bool(self._cleared_bits) when self._cleared_bits is not None, else None.
• slave_count: int | None — int 0..8 on scope == "probe" and dimension == 7, else None.
• faulty_zone: int | None — int 0..99 on scope == "central" and dimension == 21, else None.

OWNHeatingDiagnosticCommand(OWNCommand) Builders

• request_probe_diagnostic(cls, zone: int | str) -> OWNHeatingDiagnosticCommand -> *#1004*{z}*11## (validate 1 <= z <= 99 integer, rejecting #0/#Z).
• request_probe_slave_count(cls, zone: int | str) -> OWNHeatingDiagnosticCommand -> *#1004*{z}*7## (1 <= z <= 99).
• request_probe_slave_diagnostic(cls, zone: int | str) -> OWNHeatingDiagnosticCommand -> *#1004*{z}*8## (1 <= z <= 99).
• request_central_diagnostic(cls, where: str = "#0") -> OWNHeatingDiagnosticCommand -> *#1004*{target}*11## (validate with existing _validate_central_where in heating.py:126).
• request_faulty_zones(cls, where: str = "#0") -> OWNHeatingDiagnosticCommand -> *#1004*{target}*21## (validate with _validate_central_where).
• request_zone_actuator_diagnostic(cls, zone: int | str) -> OWNHeatingDiagnosticCommand -> *#1004*#{z}*22## (accept 1..99 or "#1".."#99", normalize to #{z} with 1 <= z <= 99).

---

2. WHO 16 FM Tuner Frequency Step & Station Preset Store (OWNd/message/sound.py)

Firmware Oracle & Encyclopedia Evidence

• Encyclopedia KB: kb_who_16_what_frequency_step (WHAT 4001..4015 / 4101..4115), kb_who_16_dim_station (DIMENSION 7 & DIMENSION 10 station read/store on F500 / F500N).
• Firmware Oracle (bt_difson):
  • *16*4001*101## (step_frequency_up, 1..15 steps -> 4001..4015) -> ACK (*#*1##) on MH200N (040209), F454 (020055), H4684 (020054), emitting SCS 0xD2 subcommand 0x04 <step>.
  • *16*4101*101## (step_frequency_down, 1..15 steps -> 4101..4115) -> ACK (*#*1##) on MH200N, F454, H4684, emitting SCS 0xD2 subcommand 0x14 <step>.
  • *#16*101*#10*5## (store_station, preset 1..15) -> ACK (*#*1##) on MH200N (040209), F454 (020055), H4684 (020054), emitting SCS EC 01 09 05 ($0494018FA0\r).

Additions in OWNd/message/sound.py

1. OWNSoundCommand (alongside seek_up / seek_down / select_track around sound.py:384-426, using step: int = 1 to match next_track / previous_track at sound.py:356-381):
   • step_frequency_up(cls, where: str | int, step: int = 1) -> OWNSoundCommand:
     • Validate _require_non_empty_where(where) and 1 <= int(step) <= 15 (else ValueError), emit *16*{4000 + s}*{target}##.
   • step_frequency_down(cls, where: str | int, step: int = 1) -> OWNSoundCommand:
     • Validate _require_non_empty_where(where) and 1 <= int(step) <= 15 (else ValueError), emit *16*{4100 + s}*{target}##.
   • store_station(cls, where: str | int, station: int) -> OWNSoundCommand:
     • Validate _require_non_empty_where(where) and 1 <= int(station) <= 15 (else ValueError, matching the 1..15 extended F500N preset range documented in select_track), emit *#16*{target}*#10*{st}##.
2. OWNSoundEvent (sound.py:79-126):
   • Decode 4001 <= self._state <= 4015 -> frequency_step_forward: int | None (self._state - 4000) and 4101 <= self._state <= 4115 -> frequency_step_backward: int | None (self._state - 4100), matching track_step_forward / track_step_backward (6001..6015 / 6101..6115 at sound.py:85-94).
   • At sound.py:116, extend elif self._dimension == 7 and self._dimension_value: to elif self._dimension in (7, 10) and self._dimension_value: so WHO 16 DIMENSION 10 station reports populate track / station identically to DIMENSION 7.

---

3. WHO 22 Tuner Seek/Step Builders & Dimension 11 Cleared-Preset Flag (OWNd/message/sound.py)

Firmware Oracle & Physical Trace Evidence

• Firmware Oracle (bt_multi): On both MH200N (040209) and MyHomeServer1 (021602), *22*5*2#1## (seek_up), *22*6*2#1## (seek_down), *22*7*2#1## (step_frequency_up), and *22*8*2#1## (step_frequency_down) all return *#*1## ACK and put EC 01 05/06/01/02 01 on the SCS bus. (store_preset *22*33#<P>*<where>## is already implemented at sound.py:1077!)
• Physical H/L4684 (FW 6.0.11) F500 trace (tests/fixtures/traces/issue_466/myhome_trace_HL4684_all_2026-11-01T11-48-15.json in MyHOME#839):
  • When manually tuned off-preset or before selecting a station, the F500 emits *#22*5#2#1*11*1*10530*0## (preset == 0), and when Station 5 is selected at the same frequency, it emits *#22*5#2#1*11*1*10530*5## (preset == 5).
  • Right now in OWNSoundDiffusionEvent (sound.py:594-595), values[2] == "0" leaves self._preset = None, which makes an explicit "no preset (0)" Dimension 11 report indistinguishable from frames that don't carry a preset field at all (such as Dimension 5 *#22*2#1*5*1*10530##).

Additions in OWNd/message/sound.py

1. OWNSoundDiffusionCommand (alongside next_station / previous_station at sound.py:1195-1207, using where: str | int = "6" for consistency):
   • seek_up(cls, where: str | int = "6") -> OWNSoundDiffusionCommand -> *22*5*{target}##
   • seek_down(cls, where: str | int = "6") -> OWNSoundDiffusionCommand -> *22*6*{target}##
   • step_frequency_up(cls, where: str | int = "6") -> OWNSoundDiffusionCommand -> *22*7*{target}##
   • step_frequency_down(cls, where: str | int = "6") -> OWNSoundDiffusionCommand -> *22*8*{target}##
2. OWNSoundDiffusionEvent (sound.py:506, 551-596):
   • Add is_preset_cleared: bool (default False), set to True when self._dimension == 11, self._frequency_khz is not None, and values[2] == "0" (self._preset remains None for backwards compatibility, while consumers like MyHOME can check event.is_preset_cleared directly instead of inspecting dimension_value[2]).
   • Decode self._what in (5, 6, 7, 8) into boolean properties is_seek_up (what == 5), is_seek_down (what == 6), is_step_frequency_up (what == 7), is_step_frequency_down (what == 8) + human-readable log lines.

Let me know what you think — happy to review & approve as soon as you're ready!

[gdluck]

Happy to. I will author the companion as a separate PR on top of master (WHO 1004 OWNHeatingDiagnosticEvent / OWNHeatingDiagnosticCommand, WHO 16 step_frequency_up / step_frequency_down / store_station plus the 4001..4015 / 4101..4115 and dimension 10 decode, WHO 22 seek_up / seek_down / step_frequency_up / step_frequency_down plus is_preset_cleared and the WHAT 5..8 flags), with the MyHOME#839 traces and the oracle rows as evidence, so #152 can ship 2.0.0b12 as is and the companion goes into the next version; if you would rather fold it into b12, say so and I rebase it onto this branch instead.

[GreenGrassBlueOcean]

Awesome, thank you @gdluck!

Let's fold it into 2.0.0b12 — otherwise we'd immediately need to cut another 2.0.0b13 release right after anyway.

Whichever workflow is cleanest for you works for me:

• Either open the companion PR against master first (I'll review and approve it right away so it merges cleanly into master), and then rebase #152 on top of it so 2.0.0b12 includes everything (and I'll approve #152 right after);
• Or fold/rebase it directly into this #152 branch (chore/release-2.0.0b12) with the updated CHANGELOG.md so a single review & approval ships 2.0.0b12 in one go.

Standing by to review and approve as soon as it's up!

[gdluck]

The companion is up as a draft: #153 (WHO 1004 OWNHeatingDiagnosticEvent / OWNHeatingDiagnosticCommand, WHO 16 step_frequency_up / step_frequency_down / store_station + WHAT 40xx / 41xx and dimension 10 decode, WHO 22 seek_up / seek_down / step_frequency_up / step_frequency_down + is_preset_cleared and the WHAT 5..8 flags), on master so #152 can ship as is. Three things worth your eye, all in the body: (1) *16*4001*101##, *16*4101*101##, *22*7*2#1##, *22*8*2#1## and the WHO 1004 frames are not in the oracle index I can see, and the firmware versions you cite (MH200N 040209, F454 020055, MyHomeServer1 021602) are not among its twelve gateways, so those builders rest on your statement; (2) the *#16*101*#10*5## bus telegram in the index is $0493018FD4 ($0494018FA0 is *22*5*2#1##); (3) the off-preset *#22*5#2#1*11*1*10530*0## is not in the T11-48-15 trace, only the *5## form, so it is a spec-derived test input. Say if you want it folded into b12 instead.

[gdluck]

Folded in, option two: the three feature commits of #153 are cherry-picked unchanged onto this branch (d4ff9e4, f99b890, 3da84b6) and 1a86d28 moves their CHANGELOG bullets under [2.0.0b12], so one review ships everything. On this head: 2676 passed with the package reinstalled, verify_coverage.py 6,919 / 6,919, sync_documentation.py --check, ruff and mypy clean. #153 stays open only as the reference for the evidence and the three oracle caveats (the WHO 16 / WHO 22 step frames and WHO 1004 are not in the oracle index I can see; the *#16*101*#10*5## telegram in the index is $0493018FD4; the off-preset dimension 11 report is a spec-derived test input); close it with this one or tell me to.

[GreenGrassBlueOcean]

Hi @gdluck — thank you for putting #153 together so quickly and folding it in here, and especially for flagging the three oracle/trace discrepancies in your comment (*16*4001*101## / *22*7*2#1## absent from the oracle index, $0493018FD4 vs $0494018FA0, and *#22*5#2#1*11*1*10530*0## absent from the T11-48-15 trace).

Your instincts on all three were spot-on: I ran a full binary disassembly + openwebnet-mcp / own-firmware-oracle audit against libdiag.so, bt_difson, and bt_multi, and my earlier specification comment (#152 (comment)) contained several protocol errors that carried straight into d4ff9e47..1a86d28a (4001..4115 instead of 5001..5115, WHO 16 DIM #10 6..15 NACKed by bt_difson + unreachable DIM 10 event decode, WHO 22 WHAT 7/8 unimplemented in bt_multi instead of 5#STEP/6#STEP, WHO 1004 R22P0 absent in libdiag.so + misdecoded DIM 7/21/23 + unique_id collisions, and non-existent kb_who_* slugs). That is 100% on me for giving you an unverified spec.

Action for #152 (2.0.0b12): Reset back to 76b9550c to ship right away

Could you please reset chore/release-2.0.0b12 back to your clean version-bump commit 76b9550c4c2f19bf766efe2a7a7a8f74a935fed7 and force-push?

git checkout chore/release-2.0.0b12
git reset --hard 76b9550c4c2f19bf766efe2a7a7a8f74a935fed7
git push --force-with-lease

• (Note: I didn't push the reset to your branch myself because master's ruleset enforces require_last_push_approval, which would prevent me from approving #152 afterward.)
• 76b9550c (OWNd/__init__.py + CHANGELOG.md only) is completely clean, and OpenWebNet-HA/MyHOME currently pins OWNd==2.0.0b11 in manifest.json without calling any #153 symbols (with the matching WHO 16 store_station / DIM 10 cleanup opened in OpenWebNet-HA/MyHOME#844), so as soon as #152 is back on 76b9550c I will approve and merge 2.0.0b12 immediately.
• Let's then rework the companion changes cleanly on #153 for 2.0.0b13 — I have posted the full primary-source audit and corrected, firmware-verified specification directly on OpenWebNet-HA/OWNd#153 (review).

[gdluck]

Reset: the branch is back at 76b9550 (OWNd/__init__.py + CHANGELOG.md only), force-pushed with lease, body restored. The companion is being reworked on #153 to your corrected specification for 2.0.0b13.

[GreenGrassBlueOcean]

Thanks @gdluck — verified head is back on \76b9550c4c2f19bf766efe2a7a7a8f74a935fed7\ (\CHANGELOG.md\ + \OWNd/init.py\ only) and all 8 CI / MyHOME checks are green. Approving and merging \2.0.0b12.