[Android][Test gap] Issue36108 regression test doesn't exercise ShellItemWrapperFragment.OnDestroyView _displayedPage reset

## Context

Follow-up from PR #36133 (fixes #36108). That PR made two changes:

1. **Fix 1** — restored the `StackNavigationManager.Connect(IView)` 1-arg binary-compat overload (the binary break). ✅ Correct.
2. **Fix 2** — added `_handler._displayedPage = null;` to `ShellItemWrapperFragment.OnDestroyView` (`src/Controls/src/Core/Handlers/Shell/ShellItemWrapperFragment.Android.cs`) so that on back-stack view recreation, `SwitchToSection → AddDisplayedPageObserver → UpdateDisplayedPage` re-runs fully instead of early-returning on the stale `_displayedPage == page` guard (leaving the toolbar/back-button unconfigured). ✅ The code fix is correct.

To guard Fix 2, PR #36133 also added a UI test `Issue36108` (`TestCases.HostApp/Issues/Issue36108.cs` + `TestCases.Shared.Tests/Tests/Issues/Issue36108.cs`) that navigates **FlyoutItem A → B → A**.

## The gap

The added test **does not exercise the `OnDestroyView` code path that Fix 2 changes**, so it would pass with or without the fix (it is not a valid regression guard for the specific bug).

Tracing the flyout-item navigation:

- `ShellHandler.SwitchToItem` (`ShellHandler.Android.cs`, ~L176–181): after the first item, subsequent `ShellItem` switches return early via `SwitchToShellItem()` — *"No fragment transaction — the wrapper fragment stays in navigationlayout_content."*
- The wrapper fragment is `.Replace()`d **once** into `navigationlayout_content` with **no `AddToBackStack`** (~L224–227), so `ShellItemWrapperFragment.OnDestroyView` never fires during A → B → A.
- `SwitchToShellItem` **independently** nulls `_displayedPage` at `ShellItemHandler.Android.cs:368` (and `DisconnectHandler` nulls it at ~L828), regardless of the `OnDestroyView` change.

Net: the flyout round-trip is handled entirely by `SwitchToShellItem`, which already resets `_displayedPage`. Reverting the new `OnDestroyView` line would **not** make `Issue36108` fail.

## What's needed

A regression test that actually triggers `ShellItemWrapperFragment.OnDestroyView` (fragment **view** destroyed while the fragment instance stays alive) and asserts the toolbar / back-button is correctly reconfigured on view recreation. Candidate triggers to investigate:

- Push/detail navigation that returns via the Android back stack, or
- A configuration change (rotation) / process-death-restore of the hosting fragment.

## Acceptance criteria

- [ ] Identify the concrete user scenario that invokes `ShellItemWrapperFragment.OnDestroyView` followed by `OnViewCreated` with the same displayed page.
- [ ] Add a test reproducing that scenario and asserting the toolbar/back-button state refreshes.
- [ ] **Verify the test FAILS when the `_handler._displayedPage = null;` line in `OnDestroyView` is reverted, and PASSES with it** (true regression guard).
- [ ] (Optional) Keep or adapt the existing `Issue36108` flyout round-trip test as coverage for the `SwitchToShellItem` reconfiguration path, clearly labeled as such.

## Notes

- The `_displayedPage` code fix in #36133 is correct and can merge as-is; this issue only tracks strengthening the test coverage so a future regression in `OnDestroyView` is caught.
- Reference: `UpdateDisplayedPage` early-return guard at `ShellItemHandler.Android.cs:597` (`if (_displayedPage == page) return;`); `AddDisplayedPageObserver` invokes the callback synchronously (`ShellSection.cs`).

Start by tracing ShellHandler.Android.cs, ShellItemHandler.Android.cs, and ShellItemWrapperFragment.Android.cs to identify a lifecycle path that invokes OnDestroyView followed by OnViewCreated with the same displayed page. Review the existing Issue36108 tests and investigate back-stack navigation or configuration changes. Done means a UI test verifies toolbar/back-button refresh and fails when the OnDestroyView reset is reverted.

Written by the indexing model from the issue text.