Repository navigation
docs(debug-journal): fold the 2026-10-29..11-04 queue slice (13 of 15 candidates) - #5993
Merged
Merged
Conversation
… candidates) Add balancer, Huawei-template and discovery-catalogue notes; the rest are folded into existing rows (Solis, Solis Modbus, Fox, Kraken, Sigenergy, web MCP, charger control, manual-API, history cache) and two trap bullets. Two candidates dropped as overtaken by merged fixes (#5836/#5870, #5842). Co-Authored-By: Claude Code <noreply@anthropic.com>
springfall2008
marked this pull request as ready for review
November 6, 2026 16:29
springfall2008
enabled auto-merge (squash)
November 6, 2026 16:29
springfall2008
disabled auto-merge
November 6, 2026 16:31
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Automated journal update from the triage bot — folds the next queue slice into
tools/debug-journal.mdafter re-verifying each candidate against current main. Docs-only change; the quality gate (coverage/run_pre_commit) is green (all hooks passed, quick suite: all tests passed). No existing entries needed correcting this time — the journal was compacted yesterday (#5988) and every claim it makes about main held at ff6eccf.Slice: the first 15 candidates in filename sort order (
5556-…then5834-…–5849-…). 91 candidates remain queued.Per candidate:
B0072/B0673fall throughSOLIS_API_CODES_NO_RETRYand get no datalogger cooldown, and B0673 carries B0600's message text so classification must key on the code. Verified against main by reading the three code tables insolis.py(no B0072/B0673 hits repo-wide). Two of the candidate's other points were dropped as per-issue: the feat(solis): write a parallel master's slot current for the whole group (#5556) #5909 diff detail (the Solis row already documents the parallel-pair gap and PR feat(solis): write a parallel master's slot current for the whole group (#5556) #5909) and the "name which commits a field test covered" note (covered by the version-drift trap bullet).energy_control_switch) — folded into the Solis Modbus row: with the switch unset,write_solis_energy_control()writes nothing andwarn_missingfires only on a real charge cycle, so freeze/hold cycles degrade silently; an S6-style separate work-mode select is not a valid target. Verified on main:adjust_charge_immediate()still writes withwarn_missing=(target_soc > 0 and not freeze).utils.pybalance_inverters())": check the snapshot builder first, the three similarly-named reserve attributes,adjust_reserve()not refreshing the_currentattributes (re-verified on main today), and the fixture trap (the execute-path fixture still stores a percent into the kWh field, seen again on main).reserve_percent_current) and PR fix(inverter): seed the dummy reserve of a register-less inverter in percent, not kWh (#5850) #5870 (register-less dummy reserve seeded in percent, not kWh) both merged, so the kWh-echo mechanism no longer describes main. Its still-true kernel lives in the new Balancer area via the Inverter balancing compares SoC percent against the reserve in kWh #5835 fold.inverter_limit_charge_dcis inCONFIG_API_OVERRIDEbut still missing fromdocs/manual-api.md's list — re-verified on main today (docs list hasinverter_limit/_charge/_discharge/_exportonly).hass.py), the web log tab is substring (classify_log_line()), plus the de-noise technique. Verified both classifiers and PR fix(inverter): warn once, not every cycle, while battery_scaling keeps reading invalid #5841's merged shape._arg()passes norequired_unit" at the 5 Nov compaction.now_utcnaming) — the naming trap itself was already covered by the journal's clocks bullet; folded only what was missing: a Kraken-row bullet on the rate-window bounds (REST = true local-midnight instants since PR fix(kraken): bound the REST rates request to the window Predbat reads back #5843, GraphQL fallback still true UTC midnight) and theminute_data()±(forecast_days+1)-days-either-side read-back bound (verified on main), plus a clause in the clocks bullet on setting the clock after construction (ComponentBase'smidnight_utcis a property). The candidate's correction of an old triage comment was dropped as per-issue.onboard_statusvocabulary) — folded into the Sigenergy row: open string vocabulary, published verbatim, cached in Storage, consumed by predbat-saas so the SaaS UI lags, and failed-run statuses stay in memory until the poll-cadence save. Verified the publish path and cache keys on main.templates/huawei.yamlonly)":inverter_limit_charge_dcis model-only, execution rewrites the wiredcharge_rateto the AC cap every idle cycle, the huawei_solar "maximum charging power" register is the battery's overall ceiling not a grid-only cap, and the chat-agent trap. Verified: nohuawei.py, the template wiring, the docs sentence, and that no execution path consumesbattery_rate_max_charge_dc(prediction only).mcp_secretbehind four mechanisms, JWT key derived from it, scope never enforced; defaults (enable off,predbat_mcp_secret, binds0.0.0.0) re-verified on main. The candidate's line numbers were stripped per the journal's house rules.is_charger: false) — folded into the shared charger-control row: the components re-pointoctopus_intelligent_slotwith a plainset_argregardless of apps.yaml,is_charger: falseleaves the charger alone and skipsplan_car_charging()(both by design); umbrella Octopus Intelligent: correlate car chargers with Octopus devices instead of assuming car 0 #5726. Verified the set_arg site, the is_charger veto and the plan-skip comment on main. Dropped: the reporter-dump-only claim about write cadence (not independently verifiable here).infostrings-only) — folded as a new area, "Discovery catalogue (coordinator.py)": the typed container map, the dropped-value warn signature, and the per-branch round-trip-test need. The stub part of the candidate is fixed on main — the marker now lives inflags: ["stub"](landed with the reporter features) — so the entry keeps only the still-true container mechanics.No existing entry was corrected or removed: HEAD is yesterday's journal compaction (#5988), no code merges have landed since, and each of the candidates' conflicts with the journal resolved in favour of the merged fixes above rather than a stale journal line.