fix: recover zero-slot Z-Wave locks, keep failed locks visible, stop clobbering saved options#1344
Merged
Merged
Conversation
…clobbering saved options Three fixes for the re-included-lock failure mode reported on the community forum (a factory-reset BE469ZP that LCM "wouldn't accept"): 1. zwave_js: when the capability probe reports a PIN type with zero slots (User Code CC interview data missing from the driver value DB, e.g. the interview completed while the lock was asleep after a re-inclusion), invoke the User Code CC API getUsersCount once and re-read capabilities before raising. The solicited UsersNumberReport persists supportedUsers in the driver's value DB; refreshValues cannot repair this state because it reads the same cached value. 2. providers: a structural validation failure during initial setup now degrades the lock (same outcome as the reconnect path) instead of propagating and getting the lock popped from runtime data with no retry path and no UI surface. The capability cache is invalidated on validation failure so revalidation on the provider integration's LOADED transition re-probes instead of serving the degenerate read. 3. init: the data->options migration now merges options-preferred instead of overwriting options with stale data. An options-flow save made while the entry was failed (no update listener registered) was silently discarded on the next successful setup. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Entire-Checkpoint: 06bd50e78ebb
github-actions
Bot
added
python
5 tasks
raman325
added a commit
that referenced
this pull request
Jul 16, 2026
…-repair * origin/main: fix: recover zero-slot Z-Wave locks, keep failed locks visible, stop clobbering saved options (#1344) build(deps-dev): bump markdownlint-cli from 0.49.0 to 0.49.1 (#1343) build(deps-dev): bump matter-python-client from 1.2.4 to 1.2.5 (#1342) [pre-commit.ci] pre-commit autoupdate (#1340) build(deps-dev): bump matter-python-client from 1.2.1 to 1.2.4 (#1338) build(deps): bump mypy from 2.2.0 to 2.3.0 (#1339) build(deps): bump the homeassistant group with 2 updates (#1337) # Conflicts: # custom_components/lock_code_manager/providers/_base.py # tests/providers/test_base.py
raman325
added a commit
that referenced
this pull request
Jul 16, 2026
…stale options Follow-ups from the post-#1344/#1345 review: - sync: a lock whose provider setup has not validated (deferred or structurally failed) now suspends its sync managers instead of attempting writes. Writes against such a lock could only fail on the capability probe, landing in the generic error handler with a misleading "report this bug" suspension repair (or, on matter, disabling the slot) stacked on the accurate lock_setup_failed repair -- and, on Z-Wave, re-firing the slot-count recovery device query on every attempt. Reconciliation resumes via request_sync_check once the LOADED-transition revalidation succeeds. New public BaseLock.provider_setup_succeeded property backs the gate. - providers: reset _setup_succeeded when revalidation fails on the reconnect path, so a previously-healthy lock that degrades (e.g. a re-interview zeroed its slot count) re-gates sync instead of keeping stale success state. - config_flow: the reauth write now consumes pending options (options-preferred, its own input winning) and clears them. An options-flow save made while the entry was failed would otherwise override the newer reauth fix on the next load, because the data->options migration merges options-preferred. - zwave_js: guard the recovery's root-endpoint lookup with .get(0) so a KeyError cannot escape past the typed handlers and drop the lock. - tests: collapse the duplicated zero-slot/healthy capability dict builders into one _user_code_caps(num_slots) factory. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
5 tasks
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #1344 +/- ##
=======================================
Coverage 96.97% 96.98%
=======================================
Files 53 53
Lines 6454 6470 +16
Branches 473 473
=======================================
+ Hits 6259 6275 +16
Misses 195 195
Flags with carried forward coverage won't be shown. Click here to find out more.
🚀 New features to boost your workflow:
|
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.
Proposed change
Three fixes for the re-included-lock failure mode reported on the community forum (a factory-reset BE469ZP that LCM "wouldn't accept," which also "dropped" a sibling lock):
zwave_js: one-shot recovery before concluding zero PIN slots. The driver derives a User Code CC lock's slot count from the cached
supportedUsersvalue, defaulting to 0 when it is missing from the value DB — the state a battery lock lands in when its interview completes while asleep (common right after a factory-reset re-inclusion, [ISSUE] _uc_fallback_capabilities gives up silently when User Code CC values are missing from the DB (incomplete interview) → misleading "does not advertise PIN credential support" #1298).UserCodeCC.refreshValuesreads that same cached value rather than re-querying it, so the only primitive that repopulates it is the CC APIgetUsersCountdevice query: its solicited UsersNumberReport persistssupportedUsers. The provider now invokes it once and re-reads capabilities before raising the actionable structural error.providers: structural validation failures degrade instead of dropping. A
LockCodeManagerProviderErrorduring initial setup used to propagate, making_async_setup_new_lockspop the lock from runtime data — invisible in the UI, no repair issue, and no retry path until a full reload (the config-entry state listener was never registered). Initial setup now matches the reconnect path: the failure is logged, the coordinator/entities/recovery listener are still created, and the LOADED transition revalidates. The capability cache is invalidated on validation failure so revalidation re-probes instead of serving the degenerate read forever.init: the data→options migration merges options-preferred instead of clobbering. An options-flow save made while the entry was failed (e.g. right after a Z-Wave exclusion removed a configured lock entity — no update listener registered in that state) sat unprocessed in
options, and the next successful setup overwrote it wholesale with staledata, silently discarding the user's fix. The migration now merges with options winning, matchingEntryConfig.from_entryprecedence.A follow-up PR adds a repair issue so structural setup failures are visible in the UI.
Type of change
Additional information
🤖 Generated with Claude Code