v0.11.0

Added

• Built-in connection-fault notifications, configurable per inverter and for
  the inverter group: off, a persistent notification, or a push notification
  through any registered notify.* service (offered as a dropdown). An
  inverter defaults to inheriting the group's setting; explicitly setting an
  inverter's own mode overrides the group for that inverter only, since the
  group has no connection of its own and exists purely as a shared default.
  A timing choice sends the notification immediately on fault classification
  or after the same 5-minute threshold as the repair issue, and an optional
  "notify when back online" setting sends a second notification once a fault
  that was actually announced clears.

Full Changelog: v0.10.2...v0.11.0

v0.10.2

Fixed

• ConnectionEngine._request_with_retry() now also retries once after a
  LinkClosed (e.g. a refused reconnect, ECONNREFUSED), the same way it
  already retried once after a LinkTimeout. Debug logs from a real
  installation showed the endpoint actively refusing brief reconnect
  attempts thousands of times over a few days -- almost always absorbed
  transparently by the link's own single reconnect-and-resend, but
  occasionally two refusals landed back to back within one poll cycle and
  surfaced as a visible connection fault. A second attempt should catch
  most of those before they escalate.

Full Changelog: v0.10.1...v0.10.2

v0.10.1

Changed

• Renamed the integration's display name from "Solarmax Inverter" to
  "SolarMax" in manifest.json and hacs.json, matching how comparable
  official integrations name themselves after just the manufacturer (e.g.
  "SolarEdge"). Corrected the "Solarmax" → "SolarMax" capitalization
  inconsistency throughout config flow titles/descriptions (all
  languages), the manufacturer field on every device, docstrings, and
  documentation. The ha_solarmax domain, entity unique IDs, and Python
  identifiers are unaffected — no reconfiguration needed for existing
  installations.

Full Changelog: v0.10.0...v0.10.1

v0.10.0

Added

• Debug-level logging throughout the connection retry path (connect
  timeouts/failures, response timeouts, peer-closed reconnects, and the
  engine's own retry-once attempts), each tagged with host, port, and
  (where relevant) inverter address. Previously a sustained daytime fault
  left almost no trace in the log beyond the single "unreachable (fault)"
  warning at the moment it started, making it impossible to tell after
  the fact whether repeated automatic reconnect attempts were timing out,
  getting refused, or something else -- particularly unhelpful for a
  fault that a Home Assistant restart clears but the integration's own
  retries do not, since that points at state stuck in the running
  process rather than the network.

Fixed

• Corrected stale troubleshooting docs claiming HOLD/ZERO-policy sensors
  stay unavailable after a restart at night; 0.5.4/0.6.3/0.6.4 fixed
  that.

Full Changelog: v0.9.3...v0.10.0

v0.9.3

Fixed

• The "Host" → "IP address" field rename in 0.9.0 missed the Repairs
  flow's own copy of that label ("Repair inverter connection" dialog),
  which kept showing "Host" in all three languages.

Full Changelog: v0.9.2...v0.9.3

v0.9.2

Fixed

• Fixed TypeError: find_endpoint_conflict() missing 1 required positional argument: 'address' when submitting the "Repair inverter connection"
  dialog from Repairs. Both call sites in repairs.py were missing the
  inverter's address, a pre-existing bug that only surfaced once the
  repair flow was actually exercised.

Full Changelog: v0.9.1...v0.9.2

v0.9.1

Changed

• Enabled by default for newly created entities: Grid frequency (TNF)
  and Inverter temperature (TKK) — previously opt-in. Both are single,
  broadly useful values (grid stability, overheating) unlike their
  per-phase/per-string/secondary siblings (TNH/TNL, TK2/TK3,
  etc.), which stay opt-in diagnostics. Only affects entities Home
  Assistant has not created yet; existing disabled entities keep their
  current state.

Full Changelog: v0.9.0...v0.9.1

v0.9.0

Changed

• Config flow wording cleanup:
  • The setup menu's "Add an inverter" / "Add the inverter group" options
    dropped the redundant "Add" (the dialog is already an add flow) and
    the group option now says it sums the selected inverters, not
    every configured one, matching the options step added in 0.7.0.
  • The Host field is now labeled IP address (IP-Adresse /
    Adresse IP), and no longer suggests 192.168.1.100 by default --
    the field starts empty.
  • The inverter group's setup step gained its own Device name field
    (it previously had none, always using the fixed English "Inverter
    Group"), and both it and the single-inverter setup step now suggest
    a name in the active Home Assistant language ("Wechselrichter" /
    "Wechselrichtergruppe" for German, "Onduleur" / "Groupe d'onduleurs"
    for French, "Inverter" / "Inverter Group" otherwise) instead of a
    fixed English default -- editable either way. Dropped the redundant
    "Solarmax" prefix from both defaults, matching the earlier
    integration-name simplification (the manufacturer already shows
    separately in the device info).

Full Changelog: v0.8.0...v0.9.0

v0.8.0

Changed

• The inverter group no longer sums voltages, frequencies, temperatures, or
  relative power (UDC/UL1-UL3/UD01-UD03, ULH/ULL, TNF,
  TNH/TNL, TKK/TK2/TK3, PRL), and also drops operating hours
  (KHR). These are intensive quantities -- the same regardless of how
  many inverters are in the group -- so summing them across inverters
  produced a meaningless multiple rather than a real total (e.g. three
  inverters on the same grid summing to a nonsensical "690V"). The group
  now only sums registers where a total genuinely means something: power,
  current, energy, installed power, and start count. Existing group sum
  entities for the dropped registers are removed from the entity registry
  on the next setup instead of lingering as permanently-unavailable ghosts.

Full Changelog: v0.7.1...v0.8.0

v0.7.1

Fixed

• Fixed Handler OptionsFlow doesn't support step group_members when
  opening the inverter group's options (0.7.0). Home Assistant's flow
  manager resolves a step by looking up a method literally named
  async_step_<step_id> on the handler; the group's member-selection
  step used that step_id but was implemented as a private
  _async_step_group_members() helper instead, so HA couldn't find it.
  Renamed to async_step_group_members() to match.

Full Changelog: v0.7.0...v0.7.1