Skip to content

CHANGELOG.md

Latest commit

 

History

History
183 lines (132 loc) · 6.16 KB

CHANGELOG.md

File metadata and controls

183 lines (132 loc) · 6.16 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.8.4] - 2026-03-24

Fixed

  • Poll interval not respected: Inverter data was published as fast as the Modbus read completed (~5-8s) instead of respecting the configured poll interval
    • Root cause: poll_interval was read from config but never applied in the main loop — the previous ~20-30s cycle time was an accidental side effect of broken TCP teardown in v1.8.3 (now fixed), which unintentionally throttled the loop
    • Added cycle_start timestamp and asyncio.sleep(remaining) after each successful cycle to wait out the configured interval
    • Duplicate error_tracker instantiation removed (dead code)
    • ImportError fallback for pymodbus marked with # pragma: no cover

Tests

  • Fixed auto_detect_slave_id → modbus_auto_detect_slave_id attribute name mismatch in 7 test locations
  • Added test_main_loop_waits_poll_interval to verify poll interval sleep behavior

[1.8.3] - 2026-03-23

Fixed

  • Auto Slave ID Detection: Request cancelled outside library error for subsequent Slave IDs (2, 100) after a failed first attempt (#15)
    • Added INTER_ATTEMPT_DELAY (2s) between attempts to allow TCP teardown
    • Added timeout to client.stop() to prevent blocking on broken clients
    • CancelledError now correctly propagates instead of being swallowed

[1.8.2] - 2026-03-02

Changed

  • CI Pipeline: Full migration to uv (replaces pip)
    • Benefits: 40% faster builds, reproducible environments (uv.lock)
    • Jobs affected: Lint, Test, Type-Check, Config-Validation
    • Example: 
      - name: uv sync & test
  run: |
    curl -LsSf https://astral.sh/uv/install.sh | sh
    uv sync --dev --frozen
    uv run pytest tests/

Performance

Job pip uv Δ
Test 45s 26s -42%
Type Check 12s 4s -67%
Total 120s 72s -40%

[1.8.1] - 2026-02-15

Fixed

  • Slave ID 0 Auto-Detection Issue: Removed broadcast address from auto-detection sequence
    • Root cause: Slave ID 0 is reserved as broadcast address in Modbus specification (write-only, no responses expected)
    • Impact: Home Assistant 2025.1+ enforces Modbus specification more strictly, causing timeouts during auto-detection
    • Solution: Updated auto-detection sequence from [0, 1, 2, 100] to [1, 2, 100]
    • Benefits:
      • Eliminates timeout errors on HA 2025.1+
      • Faster auto-detection (one less attempt)
      • Compliant with Modbus specification
    • Fixes compatibility with Home Assistant 2025.1 and later versions

Technical Details

Before:

SLAVE_IDS_TO_TRY =   # Slave ID 0 caused timeouts [github](https://github.com/arboeh/huABus)

After:

SLAVE_IDS_TO_TRY =   # Slave ID 0 removed [community.home-assistant](https://community.home-assistant.io/t/app-huabus-huawei-solar-modbus-to-mqtt-sun2-3-5-000-mqtt-home-assistant-auto-discovery/958230)

Affected Versions:

  • Home Assistant 2025.1 and later with stricter Modbus handling

Workaround (if using older addon version):

modbus:
  auto_detect_slave_id: false
  slave_id: 2 # Use your actual Slave ID

Credits

Thanks to HANT for the detailed bug report! 🙏

[1.8.0] - 2026-02-10

Added

  • Automatic Slave ID Detection: No more guessing! The addon now automatically detects the correct Slave ID

    • Tries common values (0, 1, 2, 100) and uses the first working one
    • New config option: modbus_auto_detect_slave_id (enabled by default)
    • UI toggle in add-on configuration for easy enable/disable
    • Fallback to manual Slave ID if auto-detection disabled
    • Detailed logging shows which Slave IDs were tried
    • Eliminates "Timeout while waiting for connection" errors for new users

  • Dynamic Register Count Display: Startup logs now show exact number of registers being read

    • Calculated dynamically from ESSENTIAL_REGISTERS constant
    • Example: INFO - Registers: 63 essential

Changed

  • Improved Error Messages: More helpful guidance for common connection issues

    • Connection errors now suggest trying different Slave IDs
    • Better context in log messages (shows attempted Slave ID)
    • Clearer distinction between timeout and connection refused errors

  • Configuration UI: Reorganized for better user experience

    • Auto-detect option prominently displayed
    • Manual Slave ID clearly marked as "only used when auto-detection disabled"
    • Better descriptions with practical examples

Fixed

  • MQTT Auto-Configuration: Restored automatic credential detection from Home Assistant MQTT service
    • Feature was accidentally removed in previous development iterations
    • Now properly uses Home Assistant MQTT service credentials when available
    • Falls back to custom credentials from config if specified
    • Clear logging indicates whether using HA service or custom config

Technical Details

Logging Examples:

# With auto-detection:
INFO - Inverter: 192.168.1.100:502 (Slave ID: auto-detect)
INFO - Trying Slave ID 0... ⏸️
INFO - Trying Slave ID 1... ✅
INFO - Connected (Slave ID: 1)

# Manual configuration:
INFO - Inverter: 192.168.1.100:502 (Slave ID: 1)
INFO - Connected (Slave ID: 1)

Configuration:

# New option (default: true)
modbus_auto_detect_slave_id: true

# Only used when auto_detect = false
slave_id: 1

Backward Compatibility:

  • Existing configurations without modbus_auto_detect_slave_id default to true
  • Existing slave_id values preserved and used when auto-detection disabled
  • No configuration migration needed

Performance:

  • Auto-detection adds 0-3 seconds to startup (depends on how many IDs tried)
  • Once connected, no performance difference

When to Disable Auto-Detection:

  • You know your exact Slave ID and want faster startup
  • You use a non-standard Slave ID not in the auto-detect list
  • For debugging purposes