### System details

Omarchy 4.0.1-1, Quickshell 0.3.1-1, Qt 6.11.2, Hyprland 0.56.2, NVIDIA 580.178.04, Linux 7.1.9, dual DisplayPort monitors

### What's wrong?

  ## Summary

  During the normal idle screensaver → lock transition, the Omarchy Quickshell
  process failed to initialize an OpenGL graphics backend and deliberately aborted
  via `QMessageLogger::fatal()`.

  The supervisor restarted Quickshell six times, producing seven SIGABRT
  coredumps, before giving up to avoid an endless crash loop.

  ## Expected behavior

  The screensaver should transition to the Omarchy lock screen without terminating
  or repeatedly restarting the shell.

  ## What happened

  The initial process was:

  - PID: 226666
  - Command: `quickshell -n -p /usr/share/omarchy/shell`
  - Signal: SIGABRT
  - First failure: 2026-08-26 20:17:47 CEST

  Relevant timeline:

  ```text
  20:15:16.618 idle-cycle-start: screensaver=150 lock=300
  20:17:46.618 lock-system: lock-timeout
  20:17:46.767 lock-requested
  20:17:46.767 lock-pending: screen-stabilizing
  20:17:47.281 QWaylandGLContext::makeCurrent: eglError: 0x3003
  20:17:47.281 QRhiGles2: Failed to make context current. Expect bad things to happen.
  20:17:47.281 Failed to create RHI (backend 2)
  20:17:47.281 FATAL: Failed to initialize graphics backend for OpenGL.

  The crash occurred immediately after the two Omarchy screensaver windows were
  closed and the lock surface was requested.

  coredumpctl list quickshell recorded seven SIGABRT dumps between approximately
  20:17:47 and 20:18:06. The supervisor then reported:

  Giving up on the Omarchy shell after 6 relaunches in under a minute.

  ## Backtrace shape

  The crashing thread shows an intentional Qt fatal abort while Qt Quick was
  processing a Wayland window expose event:

  abort
  QMessageLogger::fatal
  libQt6Quick
  QWindow::event
  QGuiApplicationPrivate::processExposeEvent
  QtWaylandClient::QWaylandWindow::sendExposeEvent
  QtWaylandClient::QWaylandWindow::updateExposure

  The Quickshell crash report ends with:

  FATAL: Failed to initialize graphics backend for OpenGL.

  ## Resource checks

  This was not an OOM kill:

  - No OOM-killer event appeared in the kernel journal.
  - No NVIDIA Xid, GPU-reset, or NVKMS allocation error appeared around the crash.
  - Afterward, approximately 28 GiB system RAM was available and swap was unused.

  EGL_BAD_ALLOC proves that the EGL implementation could not allocate a required
  graphics resource, but it does not by itself establish whether the underlying
  cause is Qt Wayland, Quickshell surface lifecycle, or the NVIDIA driver.

  ## System information

  - Omarchy: 4.0.1-1
  - Kernel: 7.1.9.arch1-2
  - Hyprland: 0.56.2-1
  - Quickshell: 0.3.1-1
  - qt6-base: 6.11.2-2
  - qt6-declarative: 6.11.2-1
  - egl-wayland: 1.1.21-1
  - GPU: NVIDIA GeForce GTX 1080
  - Driver: nvidia-580xx 580.178.04
  - Display 1: 3440×1440 @ 100 Hz, DisplayPort
  - Display 2: 2560×1440 @ 144 Hz, DisplayPort
  - VRR was disabled on both displays.

  Quickshell and Qt were installed/upgraded the previous day, but there is not
  enough evidence to claim that update caused the failure.

  ## Impact and recovery

  No application data was lost. The Omarchy shell and lock-screen UI stopped
  working after the restart loop. A later session restart restored the bar and
  background surfaces.

  ## Related issues

  These have related symptoms but different recorded mechanisms:

  - https://github.com/basecamp/omarchy/issues/6628
    Lock shell dies during idle-to-lock, but without a coredump or identified fatal
    EGL error.

  - https://github.com/basecamp/omarchy/issues/7380
    Wake-time crash caused by a fatal Wayland protocol error involving a FALLBACK
    monitor, rather than EGL_BAD_ALLOC.