What caused the crash
I was working on a new plugin when Quickshell crashed, I got a notification which told me to file this issue.
This is what Claude analised:
Crash report: quickshell SIGSEGV, 2026-09-13 20:33:17
What crashed and what it was doing
/usr/bin/quickshell -n -p /usr/share/omarchy/shell (your Omarchy shell) segfaulted once. The crashing thread was in the middle of live-reloading QML — the stack shows QQmlIncubationController::incubateFor → QQmlIncubatorPrivate::incubate → QQmlObjectCreator::finalize, i.e. Quickshell was in the process of instantiating/finalizing a freshly-reloaded QML component when it died.
Timeline correlation (the key evidence)
- 20:32:58.95 — ~/.config/omarchy/plugins/dutchbase.webapps/Panel.qml was touched (your own custom local plugin, not part of Omarchy — confirmed by git log: last real edit was your dbb9140 fix: always render a web glyph... commit).
- 20:33:13–15 — the shell log shows it reloading the whole plugin tree (many "IpcHandler ... registered but will not be used" warnings for omarchy.agents, omarchy.indicators, dutchbase.webapps, etc.) — this is Quickshell's config hot-reload in progress.
- 20:33:17 — crash, ~18s after the Panel.qml file change, right in the middle of that reload's object incubation.
- No Omarchy shell files (/usr/share/omarchy/shell) changed around this time — only your own plugin did.
- 20:34:22 onward — a fresh quickshell instance started, reloaded dutchbase.webapps again (this time cleanly), and has been running fine since.
Mechanism — proven vs. inferred
- Proven: the crash occurred during QML incubation/finalize, seconds after your custom Panel.qml was modified on disk, while Quickshell's hot-reload was actively re-registering handlers for that same plugin.
- Proven: right before the fault, quickshell's own code called __dynamic_cast (RTTI cast) and then explicitly called raise() (the signal was delivered via SI_TKILL, i.e. self-inflicted, not a genuine wild-pointer fault) — this is consistent with an internal assertion/fatal-path in Quickshell's C++ QML-type glue, not a random memory corruption.
- Inferred (not provable without Quickshell debug symbols, which aren't published via debuginfod): a dynamic_cast on a QML object that was mid-destruction/re-creation during the hot-reload of your plugin failed or hit a stale pointer, and Quickshell's own crash path turned that into a deliberate SIGSEGV.
- Quickshell binary frames are unsymbolized (no debuginfo available for it); Qt/glibc frames symbolized cleanly via debuginfod.
Data loss
None. This is a shell (bar/panel) process, not a data-holding app — it auto-restarted seconds later and your session was otherwise unaffected. No core files or temp artifacts were left behind (I dumped the core to a private tmpfile and deleted it after inspection).
Recurrence / is this an Omarchy bug
This is not an Omarchy bug — Omarchy's own shell config was untouched at the time. The trigger was editing/saving your own custom dutchbase.webapps plugin while quickshell had it loaded; the crash itself happened one directory-watch reload cycle later. It's most likely a Quickshell hot-reload edge case (destroying/recreating QML objects for a changed local plugin can race with an in-flight incubation) rather than a defect in your Panel.qml. It's a single occurrence so far (coredumpctl list shows nothing else), so it may simply be an occasional hot-reload race rather than something that reproduces every edit.
This is the plugin I was working on: https://github.com/dutchbase/omarchy-webapps
Report file
report.txt
Configuration
No response
Backtrace
No response
What caused the crash
I was working on a new plugin when Quickshell crashed, I got a notification which told me to file this issue.
This is what Claude analised:
Crash report: quickshell SIGSEGV, 2026-09-13 20:33:17
What crashed and what it was doing
/usr/bin/quickshell -n -p /usr/share/omarchy/shell (your Omarchy shell) segfaulted once. The crashing thread was in the middle of live-reloading QML — the stack shows QQmlIncubationController::incubateFor → QQmlIncubatorPrivate::incubate → QQmlObjectCreator::finalize, i.e. Quickshell was in the process of instantiating/finalizing a freshly-reloaded QML component when it died.
Timeline correlation (the key evidence)
Mechanism — proven vs. inferred
Data loss
None. This is a shell (bar/panel) process, not a data-holding app — it auto-restarted seconds later and your session was otherwise unaffected. No core files or temp artifacts were left behind (I dumped the core to a private tmpfile and deleted it after inspection).
Recurrence / is this an Omarchy bug
This is not an Omarchy bug — Omarchy's own shell config was untouched at the time. The trigger was editing/saving your own custom dutchbase.webapps plugin while quickshell had it loaded; the crash itself happened one directory-watch reload cycle later. It's most likely a Quickshell hot-reload edge case (destroying/recreating QML objects for a changed local plugin can race with an in-flight incubation) rather than a defect in your Panel.qml. It's a single occurrence so far (coredumpctl list shows nothing else), so it may simply be an occasional hot-reload race rather than something that reproduces every edit.
This is the plugin I was working on: https://github.com/dutchbase/omarchy-webapps
Report file
report.txt
Configuration
No response
Backtrace
No response