You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
zwave-lock: Schlage sub-driver drops the UserCode Report that confirms addCredential (no commandResult, slot reused) #3365
On SLGA-migrated Schlage locks (seen on BE469ZP 003B-0001-0469, and on the 003B-6341-5044, 003B-6349-5044 and 003B-634B-504C models), addCredential emits no commandResult at all. SmartThings accepts the command and lock/keypad events keep arriving, but the driver never confirms or rejects the add.
Cause
After a UserCode:Set, these locks confirm with an occupied UserCode Report whose code is masked (**********) or blank. In schlage-lock/init.lua, is_user_code_report_mfr_specific classifies that report as manufacturer-specific, and user_code_report_handler consumes it, only acting on index 0. The generic handler never sees it, so:
set_credential_report_helper isn't called. There's no success and no credential entry, and the busy state lapses after 10 s without a result.
tables.next_index keeps returning the same slot, so the next addCredential writes to that same slot. If the PIN landed, it is overwritten.
#3232 / #3318 fixed this only for CODES_FROM_LOCK. #3364 handles the refused (0xFE) case. Nothing yet covers the success case for an add or update.
The pre-migration handlers treated the same report as "code set": the legacy Edge Schlage handler (code_set_event), and the Groovy DTH (zwave-lock.groovy: "Schlage locks sends a blank/empty code during code creation… Some Schlage locks send '**********' during code creation also").
Evidence
Across about 200 migrated, online Schlage locks we integrate with, 22 of the 23 locks that went silent on addCredential in one day were Schlage. On affected locks, keypad unlocks keep arriving with userIndex set to slot numbers that aren't in the driver's credentials table, so PINs are on the lock that the driver never recorded. The hubs report driver build 2026-09-02. main has the same add-path logic.
-- in user_code_report_handler, after the CODES_FROM_LOCK branchlocalin_progress=device:get_field(consts.DRIVER_STATE.COMMAND_IN_PROGRESS)
if (in_progress==consts.LOCK_CREDENTIALS.ADDorin_progress==consts.LOCK_CREDENTIALS.UPDATE)
andcmd.args.user_id_status==user_id_status.ENABLED_GRANT_ACCESSthenlocalargs=device:get_field(consts.DRIVER_STATE.CREDENTIAL_ARGS_IN_USE) or {}
ifcmd.args.user_identifier==args.credentialIndexthenzwave_handlers.user_code_report(self, device, cmd)
returnendend
A test mirroring test_schlage_lock.lua would queue a masked occupied report for the pending slot and expect addCredential success.
Activity
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …
Z-Wave Lock: Treat a masked Schlage code report as confirmation of a …