laundry: fix four findings from the final review of guided setup

The one that could run the wrong program: guided setup accepted a name
another program already had. The duplicate check only compared names within
the form it was handed, and guided setup submits one program at a time, so
it never saw the others. Two programs sharing a label resolve to whichever
option comes first, so picking the second would have run the first one's
payload -- the exact failure the check exists to prevent, working correctly
in the bulk form and blind in the guided one. The taken set now includes
every other named program, excluding the slot being edited so confirming an
unchanged name doesn't reject itself.

The rest:

- The timeout screen's copy interpolates the same counters as the other two
  but was shown without placeholders, so it rendered literal braces.
- The probe reports whatever payload is loaded, while observe() declines one
  whose slot the device doesn't advertise. Guided setup could reach the name
  form for such a slot, take a name, and silently discard it -- there is no
  record to hang it on and no payload to replay. It now waits instead.
- The prefilled Download course came from the raw candidate list while the
  dropdown filters to courses the appliance still offers, so a stale
  candidate prefilled a value the selector rejects and the form failed
  validation on something the user never chose.

custom_components/localthings/config_flow.py

@@ -1005,7 +1005,16 @@ class LocalThingsOptionsFlow(config_entries.OptionsFlow):
         deadline = time.monotonic() + _CLOUD_WAIT_TIMEOUT_S
         while time.monotonic() < deadline:
             slot = await coord.async_probe_cloud_courses()
-            if slot is not None and slot != self._cloud_baseline:
+            # Only a slot the store actually recorded. The probe reports
+            # whatever payload is loaded, while observe() declines one whose
+            # slot the device doesn't advertise -- offering to name that would
+            # take a name and silently discard it, since there is no record to
+            # hang it on and no payload to replay.
+            if (
+                slot is not None
+                and slot != self._cloud_baseline
+                and coord.cloud_courses.snapshot()["slots"].get(slot)
+            ):
                 return slot
             await asyncio.sleep(_CLOUD_PROBE_INTERVAL_S)
         return None
@@ -1104,11 +1113,17 @@ class LocalThingsOptionsFlow(config_entries.OptionsFlow):
 
     @property
     def _cloud_course(self) -> str | None:
+        """The best observed candidate, narrowed to what the selector offers.
+
+        Unfiltered, a candidate the appliance's own course list no longer
+        contains would prefill a dropdown that rejects it, and the form would
+        fail validation on a value the user never chose.
+        """
         coord = self._coordinator()
         if coord is None:
             return None
-        candidates = coord.cloud_courses.download_candidates()
-        return candidates[0] if candidates else None
+        available = cycle_options(coord.canonical_resources(MAIN))
+        return next((c for c in coord.cloud_courses.download_candidates() if c in available), None)
 
     def _cloud_course_selector(self, coord):
         """The appliance's own course codes, observed candidates first.
@@ -1134,9 +1149,13 @@ class LocalThingsOptionsFlow(config_entries.OptionsFlow):
         """Nothing was selected in time. Offer another round rather than
         dropping the user out of the flow -- and an explicit finish, for
         anyone who doesn't think to close the dialog."""
+        coord = self._coordinator()
+        if coord is None:
+            return self.async_abort(reason="not_loaded")
         return self.async_show_menu(
             step_id="cloud_timeout",
             menu_options=["cloud_guided", "cloud_finish"],
+            description_placeholders=self._cloud_progress_placeholders(coord),
         )
 
     async def async_step_cloud_finish(
@@ -1204,7 +1223,18 @@ class LocalThingsOptionsFlow(config_entries.OptionsFlow):
         courses come first, so a shared label resolves to the real cycle).
         """
         names = {slot: str(user_input.get(f"name_{slot}", "")).strip() for slot in known}
+        stored_slots = coord.cloud_courses.snapshot()["slots"]
         taken = {name.casefold() for name in self._device_course_names(coord)}
+        # Programs this form isn't editing. The bulk form edits every slot at
+        # once so this adds nothing there, but guided setup submits one at a
+        # time -- without it, naming two programs the same was accepted, and
+        # the select resolves a shared label to whichever option comes first,
+        # so picking the second would run the first one's payload.
+        taken |= {
+            record["name"].casefold()
+            for slot, record in stored_slots.items()
+            if record["name"] and slot not in known
+        }
         for name in names.values():
             if not name:
                 continue