Releases: phoinixgrr/sigma_connect_ha
Release v1.2.4
Fixed
- Custom partition names no longer break alarm status parsing — When a custom partition name is configured via the AXION installer menu (e.g.
Τμήμα 1 (ΣΠΙΤΙ):), the status regex failed to match, resulting inunknownalarm state. The parser now handles optional parenthesized text between the partition number and colon. Closes #4.
Added
- Unit tests for custom partition name parsing — Covers disarmed, armed, various name formats, and regression tests for the default partition format.
Confirmed Compatible
- S-PRO 64 with Axion v1.4.0
Full Changelog: v1.2.3...v1.2.4
Release v1.2.4-dev01
Full Changelog: v1.2.3...v1.2.4-dev01
Release v1.2.3
Fixed
-
Action failures no longer persist until HA restart — When the panel session went stale, all 5 retry attempts reused the same dead session, causing every arm/disarm/stay action to fail. Now the client re-authenticates (
logout+login) before each retry attempt, allowing recovery on attempt 2+. -
Session expiration detected from response body — If the panel silently rejects an action by returning a login page, it's now detected (via
gen_inputtoken orlogin.htmlredirect) and triggers an immediate re-authentication + retry. -
Partition page selected before each action — Previously the last visited page before triggering
arm.html/disarm.html/stay.htmlwas the zones page, not the partition page. Nowselect_partition()is called right before every action trigger. -
Retry loop no longer crashes on transient errors — If
login(),select_partition(), or the HTTP request threw an exception inside the retry loop, the entire action would crash — wasting all remaining attempts. These are now caught and the next attempt proceeds normally. -
Initial status check no longer fatal — The "already in desired state?" check before the retry loop could crash on a stale session, killing the action before it even started. Now treated as non-fatal — if it fails, the action proceeds to the retry loop.
-
Lock acquisition has a 60-second timeout — Previously, if another action was stuck,
perform_action()would block the executor thread indefinitely. Now times out after 60s and returns cleanly.
Changed
- Simplified settings UI to essential options only
- Improved config option labels for clarity
- Removed unused constants
Diagnostics
- Action response body (first 500 chars) is now logged at DEBUG level on every attempt for future troubleshooting
Full Changelog: v1.2.2...v1.2.3
Release v1.2.2
Fixed
- Previous version v1.2.1 , can cause a deadlock, rendering arm/disarm commands unusable.
Ha Core logs throw:
2026-01-08 20:38:52.723 DEBUG (SyncWorker_34) [custom_components.sigma_connect_ha.sigma_client] [LOCK] Waiting to perform 'arm'
This is a fix release for this
Full Changelog: v1.2.1...v1.2.2
Release v1.2.1
Added
- PIN support: separated the Web Interface Password from the actual Alarm PIN used for keypad auth (v1.2.1-dev-02) (tackles: #1 )
- Updated setup UI to ask for Alarm PIN and added translations (EN/EL)
- Coordinator/client updated to use PIN safely with fallback (so existing installs keep working)
Changed
- Improved GitHub Release workflow for dev + stable tags
- Switched to HACS zip_release flow
- Updated release packaging so the manifest version can match the tag
Full Changelog: v1.1.3...v1.2.1
Release v1.1.3
Full Changelog: v1.1.2...v1.1.3
Release v1.1.2
- fixed a race condition when arming/disarming
Full Changelog: v1.1.1...v1.1.2
Release v1.1.1
Completely revisited parse logic. We now have super fast updates!
- Fully switched to parsing all alarm data (status, battery, AC power, zones) from zones.html
- Removed all legacy parsing from partition page (get_part_status)
- Introduced
try_zones_directly()to attempt reuse of the authenticated session - Reuse existing HTTP session if authenticated and zones.html is complete
- Fallback to full login flow only when session reuse fails or data is incomplete
- Improved logging: clearly indicate when session is reused or re-authenticated
- Updated
perform_action()logic to use zones.html for state checks - Ensured session reuse works across both polling and action flows
Full Changelog: v1.0.11...v1.1.1
Release v1.0.11
- revisiting default DEFAULT_MAX_CONSECUTIVE_FAILURES to 3. To be fault tolerant
Full Changelog: v1.0.9...v1.0.11
Release v1.0.10
- marking alarm unavailable after N failed attemps funcionality
Full Changelog: v1.0.9...v1.0.10