[WIP] Fix UtilResponse to distinguish WebSocket completion from errors

[Copilot]

Thanks for asking me to work on this. I will get started on it and keep this PR's description up to date as I form a plan and make progress.

---

This section details on the original issue you should resolve

<issue_title>UtilResponse cannot distinguish a clean WebSocket completion from an error/abnormal close</issue_title>
<issue_description>In mistapi/device_utils/__tools/__ws_wrapper.py (v0.62.0), a completed UtilResponse exposes no signal that the device-utility WebSocket errored, closed abnormally, or failed to start — so a consumer cannot tell a confirmed result from an unconfirmed one.

Where the information is lost

• WebSocketWrapper.start() wires ws.on_error(lambda error: LOGGER.error("Error: %s", error)) — the error is logged and discarded.
• WebSocketWrapper._on_close(code, msg) logs the close code but stores nothing on the UtilResponse.
• start_with_trigger._run(): when the WS factory returns None, the trigger raises, or the trigger is non-200, it calls self.util_response._closed.set() (→ done=True) without setting ws_required or recording why it ended.

Impact

After await / wait(), a UtilResponse only carries trigger_api_response, ws_required, ws_data, done. As a result these three outcomes are indistinguishable to a caller:

Scenario	trigger_api_response	ws_required	ws_data	done
Healthy trigger-only command	200	False	[]	True
WS-backed command, WebSocket failed to start	200	False	[]	True
WS-backed command, WebSocket errored mid-stream	200	True	partial	True

A caller that persists a durable "executed/confirmed" record (e.g. an approval or audit flow) can therefore record an unconfirmed device action as confirmed.

Proposed change (additive, backward-compatible)

1. UtilResponse.__init__: add self.ws_error: str | None = None and self.ws_close_code: int | None = None.
2. WebSocketWrapper.start(): replace the discard lambda with a handler that records self.util_response.ws_error = str(error) (still logs).
3. WebSocketWrapper._on_close(self, code, msg): set self.util_response.ws_close_code = code; if code indicates an abnormal close, also set ws_error.
4. start_with_trigger._run() failure fall-throughs: set self.util_response.ws_error = "<reason>" (e.g. "WebSocket factory returned None", f"trigger error: {e}", f"trigger failed: {status_code}") before _closed.set().

Then ws_error is not None is a reliable "did not cleanly confirm" signal, and happy paths are unchanged (ws_error stays None). A PR implementing this follows.</issue_description>

• Closes UtilResponse cannot distinguish a clean WebSocket completion from an error/abnormal close #29

Comments on the Issue (you are @copilot in this section)