[gdluck]

Requested by A

Before: seven sentences in the trace READMEs and test docstrings of #799, #810 and #815 stated numbers or frames the fixtures do not hold (a request count double-counted, a dimension 14 "catching up 0.3 s later" that is never pushed, 34 zones for 33, 44 actuator frames for 43, "17 zones in 0.5 s" for 8), and three replay tests passed for the wrong reason.

After: every sentence states the fixture fact (each re-derived from the rows with jq before committing), and the three tests assert what their names say.

What is the problem

Docs / test-only follow-up on my trace PRs, found by an adversarial re-read after @GreenGrassBlueOcean's #140 review on OWNd. How it was found: each README sentence and docstring was read against the fixture rows and the question asked was "would this test still pass if the sentence were false?":

• f461_live_2026-10-23/README.md:57: run 1 "165 requests mirrored from the oracle suites … plus 31 plant reads" (= 196); run 1 has 165 result rows in total (134 oracle + 31 plant reads).
• f461_live_2026-10-23/README.md:72 and test_f461_live_2026_10_23_trace.py:413: "*#4*2*#14*0235*3## ACKed in 78 ms, dimension 12 follows at once, dimension 14 catches up 0.3 s later": no dimension 14 follows the write; the only 14 of run 3 is in the reply to the probe's own *#4*2## 2.8 s later, and the 0.3 s is the 12 → 14 gap inside that reply; the assertion evt.index(dim12) < evt.index(dim14) is satisfied by the poll order alone.
• myhomeserver1_live_2026-10-24/README.md:18: "34 zones report a temperature": 33 do (zone 43 has an actuator, no temperature); the PR's own test asserts 33; 34 is the actuator count.
• test_myhomeserver1_live_2026_10_24_trace.py:217: "ACKed with 44 actuator frames"; the body asserts 43 after that request (44 only with the earlier zone-35 reply).
• test_f461_live_2026_10_23_trace.py:631 (run 4b): "23.5 in force and commanded, heating season, valve off": run 4b has no dimension 19 / 20 and the test asserted neither _target_temperature nor hvac_action.
• test_trace_oracle_validation.py:166: "*#4*0*60## … ACK, 17 zones in 0.5 s": ACK at 0.499 s with 8 replies; the 17th zone at 0.97 s.
• test_f461_live_2026_10_24_trace.py:577 (test(trace): ingest live F461 2.0.11 2026-10-24 climate setpoint capture (knob, antifreeze split, presets, bezel, HA write) #810 knob turn): "dimension 14 changes nothing", but every dimension 14 of run 1 carries 18.0 = the dimension 12 value, so the assertions cannot tell which dimension drove the display (a scratch replay with the fix(climate): show the setpoint in force (dimension 12), keep dimension 14 as the commanded value #791 hold disabled passed them).
• test_f461_actuators_live_2026_10_25_trace.py:529 (test(trace): ingest the F461 and MyHomeServer1 captures of 2026-10-25 (actuator status, fan speed and humidity per zone) #815): test_replay_action_sequence_matches_the_recorded_entity_states compared replay and record as two separate literals.

@GreenGrassBlueOcean a quick confirmation is enough here; nothing in the fixtures changes.

---

What is the solution

1. READMEs and docstrings (5c7474d): "165 read-only requests: 134 mirrored … plus 31 plant reads"; "no dimension 14 is pushed; the next *#4*2## reads 14 = 23.5 (0.3 s after the dimension 12 of its own reply)"; "33 zones report a temperature (34 answer the actuator request)"; "43 actuator frames (44 in the run with the zone 35 reply)"; "ACK after 0.50 s with 8 replies; 17 zones by 0.97 s"; "valve off" dropped.
2. Assertions: run 3 asserts not any(f.startswith("*#4*2*14*") for f in _evt_between(rows, RUN3, write)); run 4b asserts _target_temperature == 23.5; the knob-turn test asserts _target_temperature == 18.0 and _local_target_reported is True (a flag only a dimension 12 report sets) and names test(trace): ingest the F461 2.0.11 heating comfort / eco preset capture (dimension 12 split from 14) #812's presets replay as the discriminating one; the test(trace): ingest the F461 and MyHomeServer1 captures of 2026-10-25 (actuator status, fan speed and humidity per zone) #815 test compares the recorded actions with the replay's first changes and says the trailing idle was read back, not recorded.

---

State Diagram

N/A — docs / test-text change only.

---

Hardware Verification

• N/A (no code under custom_components/ changes). The 2026-10-26 live reads on the F461 and the MyHomeServer1 reproduce the same per-zone dimension 12 / 13 / 14 shapes the fixtures hold.

---

How is it tested

• tests/test_f461_live_2026_10_23_trace.py (run 3 write, run 4b), tests/test_myhomeserver1_live_2026_10_24_trace.py, tests/test_trace_oracle_validation.py, tests/test_f461_live_2026_10_24_trace.py::test_replay_knob_turn_shows_dimension_12, tests/test_f461_actuators_live_2026_10_25_trace.py::test_replay_action_sequence_matches_the_recorded_entity_states.
• pytest tests -n 4 --cov=custom_components/myhome: 4991 passed, 1 skipped, TOTAL 16810 0 100.0%; ruff check . clean; scripts/typing_ratchet.py mypy --strict: 0 error(s); scripts/verify_ha_standards.py SUCCESS; scripts/sync_documentation.py --check SUCCESS; scripts/manage_translations.py check no drift.
• Kimi (independent review): the corrected counts and timings agree with the fixture facts; the run-3 assertion is bounded to the window after the write and cannot be vacuous; the recorded-states comparison is non-vacuous; it asked for the _local_target_reported assertion on the knob-turn test (taken).

---

Where are the traces

• tests/fixtures/traces/f461_live_2026-10-23/, tests/fixtures/traces/f461_live_2026-10-24/, tests/fixtures/traces/myhomeserver1_live_2026-10-24/, tests/fixtures/traces/f461_actuators_live_2026-10-25/ (unchanged).

---

What does the Firmware Oracle / Encyclopedia say

1. OpenWebNet Encyclopedia: N/A.
2. Firmware Oracle: N/A (text about our own captures).

🤖 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]

@gdluck Confirmed — verified all 6 fixture readings, docstrings, and assertions directly against the trace rows (trace_payload["rows"] / frames):

1. f461_live_2026-10-23 Run 1 & Run 3: 165 read-only requests in total (134 oracle + 31 plant), and *#4*2*#14*0235*3## pushes Dim 12 immediately with no pushed Dim 14 before the next *#4*2## poll (not any(f.startswith("*#4*2*14*") ...)).
2. f461_live_2026-10-23 Run 4b: Carries no Dim 19/20 actuator frame and leaves zone2._target_temperature == 23.5.
3. myhomeserver1_live_2026-10-24: 33 zones report Dim 0 temperature while 34 answer *#4*#0*20## (zone 43 has 43#1 only), and 43 actuator frames follow *#4*#0*20## (44 across the run including the earlier *#4*35*20## reply).
4. test_trace_oracle_validation.py (*#4*0*60##): ACK on MyHomeServer1 at 0.50 s with 8 replies on the command session and 17 zones by 0.97 s.
5. #810 & #815 assertions: zone._local_target_reported is True cleanly discriminates Dim 12 from Dim 14 when both equal 18.0, and [a.value for a in changes[:len(recorded)]] == [s["hvac_action"] for s in recorded] ties the replay directly to the recorded states.

Merging now (followed by #820 and #817).