[gdluck]

Requested by A

What is the problem

manifest.json pins OWNd==2.0.0b11. OWNd 2.0.0b13 (OpenWebNet-HA/OWNd#154, after #152) carries the changes MyHOME has been waiting for behind transition windows: the F461 / MyHomeServer1 profiles without WHO 16 (OWNd#142), the WHO 13 dimension 27 model-name decoder (OWNd#139), the #-prefixed dimension 14 setpoint (OWNd#138), the H4684 aliases and the firmware-derived WHO 22 (OWNd#147), the narrowed dimension 22 guard (OWNd#151) and the MyHOME#839 companion (OWNd#153). Offered on OWNd#139 / #153; the maintainer asked to finish the pin bump together with the release.

CI cannot be green until OWNd 2.0.0b13 is on PyPI (OpenWebNet-HA/OWNd#154 is open; 2.0.0b12 was never published): every job that installs the manifest pin (the conftest fallback install, the HA container smoke, hassfest, the OWNd smoke --target pinned gate) fails on the missing package. The OWNd-from-git coverage job is the one to watch meanwhile; locally the whole suite is green with OWNd master on PYTHONPATH (that is what 2.0.0b13 will be). Draft until the package exists.

---

What is the solution

1. custom_components/myhome/manifest.json: OWNd==2.0.0b13.
2. Discovery rows follow the pinned profiles (scripts/update_gateway_profiles.py): F461 and MyHomeServer1 pinned_whos [1, 2, 4, 18], sweep_omits [*#4*0##]; the transition helpers of chore(profiles): derive the discovery rows from the OWNd profile (prep for OWNd #142) #808 / chore(scripts): align discovery doc check across OWNd==2.0.0b11 WHO 16 transition #831 (PRE_WHO16_OWND_PIN, manifest_pins_pre_who16) go; docs/getting-started/hardware-compatibility.md and docs/architecture/anti-drift-safeguards.md regenerated with scripts/sync_documentation.py --update. verify_profiles_against_oracle keeps zero WHO 16 tolerance and the Generic-row rejection (tests renamed to the post-bump contract).
3. tests/test_trace_oracle_validation.py: pending_decoder_rx (the two WHO 13 dimension 27 frames *#13**27*70#52#54#49## and the MyHomeServer1 one) and PRE_MODEL_NAME_OWND_PIN removed; replaced by a test that the pinned OWNd decodes them (model_name F461 / MyHomeServer1).
4. OWNd#147 shims removed (gateway.py _MODEL_SPELLINGS / _profile_for_model fallback and the _firmware_at_least WHO 22 shim in _profile_supports_who, topology.py WHO 22 shim; the TODO(OWNd>2.0.0b11) markers): the pinned OWNd aliases HL4684 / H/L4684 / HC4684 to H4684Profile and derives WHO 22 from firmware >= 6.0.0 inside the profile, so those branches were unreachable; the brand-prefixed "BTicino H/L4684" spelling (no OWNd profile) keeps its one substring key in supports_zone_status with a corrected comment; tests/test_issue_466_hl4684_trace.py now asserts the pinned OWNd's own aliasing. No runtime behaviour changes with the pinned OWNd.
5. .github/release-notes/2.0.0b17.md: "on OWNd 2.0.0b13".

Not changed on purpose: pyproject.toml [test] floor OWNd>=2.0.0b11 (CI force-installs OWNd from git after it), historical "on OWNd 2.0.0b11" lines in the roadmap and the trace fixtures' ownd_version fields.

---

State Diagram

N/A — dependency pin, generated docs, transition-window test cleanup, dead-shim removal.

---

Hardware Verification

N/A (no protocol change). The underlying evidence is on the OWNd PRs listed above.

---

How is it tested

• Full suite with OWNd master (fb0acd7, the 2.0.0b13 tree) on PYTHONPATH: 5188 passed (5 snapshots passed).
• tests/test_gateway_profiles.py, tests/test_sync_documentation.py, tests/test_trace_oracle_validation.py: 119 passed; shim-affected modules after the removal: 453 passed.
• ruff check ., mypy custom_components/myhome (76 files), scripts/verify_ha_standards.py, scripts/sync_documentation.py --check: all pass.
• Against the installed 2.0.0b11 (no PYTHONPATH) the same branch fails by design: the docs oracle / discovery checks (RULE_DOCS_ANTI_DRIFT), test_gateway_profiles / test_sync_documentation (7 tests) and the OWNd smoke Gate 1 metadata check.

---

Where are the traces

N/A. The dimension 27 frames in the replaced test are the live F461 2.0.11 / MyHomeServer1 3.87.13 captures of #799.

---

What does the Firmware Oracle / Encyclopedia say

1. OpenWebNet Encyclopedia: N/A.
2. Firmware Oracle: the F461 020011 and MyHomeServer1 028206 WHO 16 refusals behind OWNd#142 (findings/ of own-firmware-oracle) are what the regenerated discovery rows encode.

---

Checklist

• PR Description: Follows the structured sections above.
• Trace-Backed (if protocol/device change): N/A.
• V2 Architecture: N/A, no entity change.
• Tests & Linters: pytest, ruff, mypy and scripts/verify_ha_standards.py pass with OWNd master on PYTHONPATH; CI waits for the PyPI package.

🤖 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!

[gdluck]

Added cc493cb (normal push): topology.gateway_supported_whos no longer swallows an ImportError from the profile lookup. A parallel merge of this PR with the others had left a stale _firmware_at_least import behind, and the except Exception turned it into an empty WHO set for every gateway, which surfaced as eighteen unrelated topology / multi-gateway / config-flow failures. A profile lookup error still degrades to "no WHOs" (now logged at debug, the existing test_gateway_supported_whos_profile_exception keeps passing); an ImportError propagates, with a test. Full suite with OWNd master on PYTHONPATH: 5189 passed, 1 deselected (the pin-bump signal). The mypy-ratchet job stays red until OWNd==2.0.0b13 is on PyPI.