## Summary

After the Omarchy 4.0.3 update, the Quickshell lock screen intermittently freezes: after unlocking/waking the display the session stays locked with no visible lock surface, mouse and keyboard input are dead, and the only recovery is a hard reboot.

## Environment

- Omarchy 4.0.3-1, omarchy-settings 4.0.3-1, quickshell 0.3.1-1
- Hyprland 0.56.2-2, hyprutils 0.14.2-1, hyprland-guiutils 0.2.2-3, qt6-base 6.11.2-3
- Intel i5-8600K, NVIDIA GTX 1070 Ti (proprietary driver), kernel 7.2.3-arch1-3
- Host monitor config: DP-2 (LG ULTRAGEAR 2560x1440) active, HDMI-A-1 disabled
- idle.lock=300, idle.screensaver=150

## Steps to reproduce

1. Sit idle until the lock engages (or lock manually).
2. The lock engages normally (`omarchy lock lock-requested` -> `secure=true`, lock surface renders).
3. Wait for the display to blank (~5s, display off), then move the mouse to wake.
4. The screen stays frozen/black: no lock UI, keyboard and mouse input dead.
5. The shell restarts and gets stuck in a "no-real-screen" state; only a reboot restores the session.

## What happens (from \`journalctl --user\`)

Boot -2:

```
11:22:50 omarchy-shell: omarchy lock ... lock-requested
11:22:50 omarchy-shell: omarchy lock ... lock-pending: screen-stabilizing
11:22:50 omarchy-shell: omarchy lock ... secure=true
11:25:17 omarchy-shell: qt.qpa.wayland: There are no outputs - creating placeholder screen
11:25:17 omarchy-shell: Not creating lock surface for screen QScreen(..., name="") as it is not backed by a valid wayland output.
... repeated ...
11:25:19 omarchy-shell: quickshell.hyprland.ipc: Got removal for monitor "FALLBACK" which was not previously tracked.
11:26:55 omarchy-shell: Omarchy shell exited with status 143; relaunching.
11:26:57 omarchy-shell: omarchy lock ... lock-stranded: recovering
11:26:57 omarchy-shell: omarchy lock ... lock-pending: no-real-screen
```

Boot -1 (same pattern):

```
11:56:55 omarchy lock ... lock-requested / secure=true
12:07:17 qt.qpa.wayland: There are no outputs - creating placeholder screen
12:07:17 Not creating lock surface for screen QScreen(...) as it is not backed by a valid wayland output.
12:07:19 quickshell.hyprland.ipc: Got removal for monitor "FALLBACK" which was not previously tracked.
```

## Root-cause analysis

When the display is woken while the session lock is held, all Wayland outputs are briefly removed (Hyprland drops to its `FALLBACK` monitor). Quickshell destroys the lock surfaces for the real screens and, once they come back, refuses to create new lock surfaces because Qt/QQuick still reports a placeholder screen (`name=""`) that "is not backed by a valid wayland output". The session lock stays held (`secure=true`) but no lock surface exists: frozen screen, dead input, forced reboot.

Suspects in the new (4.0.2/4.0.3) code path:
- The idle-lock display-blank/wake cycle (`omarchy-brightness-display off` after lock + `omarchy-system-wake` on input) races with `omarchy-hyprland-monitor-watch`'s `recover_modeless` loop, which issues `hyprctl reload` while the session is locked — this is what drops all outputs to `FALLBACK`.
- The lock service's `lock-pending: no-real-screen` path (`screens` count of 0) means a reloaded shell never re-arms `sessionLock.locked=true` once real screens return.

This did not happen before the Sep 12 update (Omarchy 4.0.1 / hyprland 0.56.2-1 / qt6-base 6.11.2-2).

## Notes

- Intermittent: boot -3 of the same day unlocked normally; boots -2 and -1 froze.
- Workaround used meanwhile: set \`idle.lock\` to a large value (e.g. 43200) in \`~/.config/omarchy/shell.json\` to disable auto-lock until fixed.
- The user has a monitor that is turned off via \`hl.dsp.dpms({ action = \"disable\" })\`, external-only setup (no laptop panel).