Extend the clamshell gate to /etc/pam.d/sudo, not just polkit-1. When the
lid is shut the reader is unreachable, so a terminal sudo would block on
"Place your finger" until pam_fprintd timed out before letting you type the
password. The same pam_exec gate (success=1 skips fingerprint when the lid
is closed) now runs ahead of pam_fprintd in the sudo stack as well.
setup and removal share one gate definition across sudo and polkit; the
migration now gates both stacks on existing installs.
Resolves the clamshell case in #856 and supersedes #6003.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a first-run notification, alongside the keybindings/Wi-Fi toasts, that
invites anyone with a fingerprint sensor to enable it. Clicking launches
omarchy-setup-security-fingerprint in a floating terminal.
Detection lives in a new omarchy-hw-fingerprint helper that reads sysfs
(device product descriptor plus a fingerprint-vendor allowlist), so it
works before fprintd/usbutils are installed and without nagging machines
that have no reader. The setup script reuses the same helper as an early
gate, bailing before installing any packages when no reader is found
(replacing the old post-install fprintd-list probe).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a first-run notification, alongside the keybindings/Wi-Fi toasts, that
invites anyone with a fingerprint sensor to enable it. Clicking launches
omarchy-setup-security-fingerprint in a floating terminal.
Detection lives in a new omarchy-hw-fingerprint helper that reads sysfs
(device product descriptor plus a fingerprint-vendor allowlist), so it
works before fprintd/usbutils are installed and without nagging machines
that have no reader. The setup script reuses the same helper as an early
gate, bailing before installing any packages when no reader is found
(replacing the old post-install fprintd-list probe).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Bring the fingerprint affordance to the Quickshell lock screen and polkit
dialog, matching what hyprlock did on master.
Lock screen: render the md-fingerprint glyph inside the password field's
right edge when a sensor is enrolled, reserving space so long passwords
never run under it.
Polkit dialog: show one method at a time. When a sensor is enrolled and
the reader is reachable, the dialog is just the centered fingerprint icon
(square card); the moment PAM asks for a password it switches to the
password field. Detects pam_fprintd anywhere in the auth stack now that a
gate can precede it.
Lid awareness: a closed lid means the reader is unreachable, so both
surfaces fall back to the password. polkit gets a pam_exec clamshell gate
(auth [success=1 default=ignore] before pam_fprintd) so a shut lid drops
straight to the password prompt instead of blocking on the reader for the
pam_fprintd timeout; the lock screen hides the icon and skips scanning.
The gate points at the fixed /usr/bin path the package always provides so
it survives switching between package installs and dev-link. A migration
adds the gate for existing fingerprint setups.
New helper omarchy-hw-laptop-closed (pure lid state); omarchy-hw-clamshell
now composes it with the external-monitor check.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The menu card re-centered on every resize, so searching or moving into
a submenu made it jump around. It still opens centered like before, but
the first search keystroke or submenu move freezes the top line where
it sits — from then on the card only grows and shrinks downward.
Closing unfreezes. Also drop the 220px minimum height so the card
always hugs its content instead of leaving a void under short lists.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Model rows stored the live Notification object in a ref role. When the
server destroyed the notification (sender close, DND untrack, dismiss)
the role became a dangling C++ pointer, and the next read segfaulted in
QQmlListModel::data — typically when replaying history over IPC. Track
live notifications in a JS map keyed by originalId instead, cleaned up
on close and untrack, where a stale reference degrades to a catchable
error instead of a crash.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Bring back per-theme Plymouth unlock theming, dropped when walker +
elephant were removed. Selection now uses the same image carousel as
the theme and background switchers instead of a name list.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Quickshell's IconImage decodes at logical size, so on HiDPI displays
PNG icons were rendered from a texture at half the needed resolution
and looked blurry next to SVG icons. Use plain Image with sourceSize
scaled by Screen.devicePixelRatio, as the notification widgets
already do.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Prefer the site's own apple-touch-icon (typically 180-512px) over
Google's favicon service, which often serves an upscaled 16-32px
favicon. Validate downloads are actual images so an HTML error page
never gets saved as an icon.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Spawning quickshell directly from omarchy-restart-shell leaked transient
environment variables from the calling terminal, SSH connection, or
development tool into the fresh shell. Dispatch the launch through
Hyprland instead so it inherits the canonical session environment, same
as autostart.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Runs inside a live Omarchy session (typically a VM installed by
omarchy-iso-test) and verifies the desktop actually works: session
health, application launches by window class, and shell surfaces via
IPC — including typing into the launcher and launching the top hit.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
omarchy bar transparent now accepts toggle, flipping the current value
in one jq update (absent counts as opaque, matching the shell).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Static submenus and links now inherit visibility from their descendants,
so hardware sections like Touchpad Haptics disappear on machines without
the hardware instead of opening empty. Laptop Display and Mirror Display
are guarded by omarchy-hw-laptop.
The guard evaluator brace-wraps every when/checked condition before
silencing it, so conditions in the jsonc no longer need their own
>/dev/null plumbing, and parents no longer need to repeat their
children's guards.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A freshly-started hyprsunset applies its default temperature at the end
of boot, silently overriding a set sent while it was still starting.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
With quickshell-git, qs kill blocks until the instance has fully
exited, so restarting no longer needs to resolve pids and wait for them
to die. Kill in a loop until none remain - each call takes the oldest,
and duplicates from stale sessions are the reason this script exists.
The timeout bounds a wedged shell that can't process the quit message:
its kill would otherwise report success after Qt's 30s wait gives up,
re-killing the same undead instance forever. The test mock now kills
real processes synchronously to match.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
quickshell 0.3.0's qs kill returns before the instance exits, forcing
restart tooling to babysit pids. Upstream fixed kill to block until the
process is gone, but only on master, so the Omarchy repo now carries
quickshell-git pinned past that fix. New installs get it from the base
packages; existing installs migrate with a single --ask 4 transaction so
the conflicting quickshell package is replaced in place, then restart
the shell after the update to pick up the new binary.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
qs parses -n before the ipc subcommand as its top-level --no-duplicate
launch flag and silently ignores it, so IPC still went to the oldest
matching instance - exactly the stale one left behind around a restart.
The --newest flag belongs to the ipc subcommand: qs ipc -n. Verified
against two live instances; the test now pins the flag position.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The bar-null-shell-null-typeof-function ternary guarded against our own
shell missing a method it always defines. bar?.shell?.firstPartyServiceFor()
handles the only real case, delayed bar injection, in one line. The
typeof checks that probe genuinely third-party plugin objects stay.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Same treatment as Stay Awake: the indicator polled the toggle CLI over
a Process with a timer to paper over the race after clicking, and the
CLI ended by asking the shell to refresh every indicator over IPC. A new
omarchy.nightlight service owns hyprsunset instead - it probes the
temperature on startup, applies changes itself for in-shell toggles, and
answers on the nightlight IPC target. The indicator becomes a plain
binding. The CLI still drives hyprctl directly so keybindings, the menu,
and ssh work without the shell, but now just nudges the service to
re-probe since hyprsunset has no state file to watch.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Deriving the tray's implicit size from childrenRect fed layout results
back into the bindings that produced them, tripping implicitWidth
binding loop warnings. Compute it from the active and inactive blocks'
own implicit sizes instead, and center the loaded blocks rather than
anchor-filling containers that are sized by their content. The contract
test now instantiates the tray to check the collapse/expand cycle and
fails on any implicitWidth binding loop in the log.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The indicator polled omarchy-toggle-idle over a Process and re-ran it on
a timer to toggle, while the CLI called back into the shell over IPC to
apply and refresh the state it had just changed. Now the indicator binds
straight to the idle service's stayAwake property and flips it in
process. The CLI only touches the state file, which the service already
watches, so toggling from keybindings and scripts still reaches the
shell without any reentrant IPC.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Hyprland only takes scales where the mode divides into whole logical
pixels (in 1/120 steps), so picking 3x on a 1280x800 QEMU display threw
a red error overlay and silently kept the old scale. Round the request
up to the nearest clean divisor instead, so 3x becomes 3.2x.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Hyprland 0.56 fixed premultiplication in the color-management shader
(hyprwm/Hyprland#14403), so the same opacity values now render visibly
more transparent than on 0.55. Raise each value to its perceptual
equivalent (a^(1/2.2)) to restore the previous look.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Hyprland 0.56 fixed premultiplication in the color-management shader
(hyprwm/Hyprland#14403), so the same opacity values now render visibly
more transparent than on 0.55. Raise each value to its perceptual
equivalent (a^(1/2.2)) to restore the previous look.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Hyprland 0.56.0 emits invalid JSON from `hyprctl -j binds`: the new
allow_input_capture field was added to the format template at slot 16
but its argument at position 8, shifting every value from
has_description onward and landing unquoted strings in numeric slots.
The jq parse failure emptied the keybindings menu down to the two
static web-app entries — and the empty result got cached under a
stable key, permanently.
Parse the plain-format output instead, which is unaffected (and also
immune to the older quotes-in-args JSON breakage), refuse to cache a
record set with zero dynamic binds, and bump the cache version so
poisoned caches invalidate.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adds named reboot and shutdown OSD icons, and gives the touchscreen
toggle a proper gesture-tap glyph instead of the restart arrow it
was accidentally using.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Transient systemd timers default to AccuracySec=1min, so the 2s
reboot/shutdown could fire up to a minute late.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Both services fire one status poll when the shell starts, but after a fresh
boot that lands before tailscaled has connected or dropboxd has finished its
respawn dance, leaving the bar icons stale for a full 30s/60s refresh
interval. Re-poll every 2s until the service shows up (or ~30s pass), then
hand off to the normal periodic refresh.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Use omarchy-display-text-size 11 instead of a lone GTK text-scaling-factor,
so the XPS 13 nudge moves shell, GTK, and terminal text together.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A macOS-style slider snapping to curated stops (9-20px), applied via
omarchy-display-text-size. Suppresses hover-driven focus changes while a
size change reflows the panel, so h/l on the slider can't let a shifted
row steal keyboard focus.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
One knob for apparent text size, anchored to the 12px shell default: it sets
the shell font base-size (~/.config/omarchy/shell.toml), GNOME/GTK's
text-scaling-factor (12px -> 1.0), and the terminal font point size
(12px -> 9pt) in lockstep. Accepts 9-20px; reset returns all three to
default. Terminals live-reload where they can (alacritty/kitty/ghostty);
foot gets a single restart nudge that replaces itself instead of stacking.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Optional tickCount cuts evenly-spaced marks into the track for stepped
sliders. Defaults to 0, so existing sliders are unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Read ~/.config/omarchy/shell.toml as a user override merged over the active
theme's shell values, so a personal font/spacing/bar tweak survives theme
switches. The file is watched, so edits re-flow the shell live.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Each panel's hero on/off toggle (Tailscale, Dropbox Wi-Fi, Bluetooth,
network, audio mute) now has a reachable "header" cursor section, a focus
ring around the hero icon, Enter/Space activation, a letter shortcut, and
hover parity. The hero is inset by heroRingPad so the ring stays inside
the Flickable/ScrollView clip box instead of being cut off.
Bluetooth and audio get a virtual "header" section above their device
sections, so the adapter/mute can be toggled by keyboard even when no
device rows exist. Network repurposes its previously dead header section.
For the PanelHero-based panels (Tailscale, Dropbox), route panel state
through the wrapper's `header` id: inside a PanelHero iconComponent `root`
resolves to PanelHero, not the Panel, so `root.headerHasCursor` and
`root.setHeaderCursor()` silently referenced the wrong object.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The hero icon is now the optimistic on/off toggle with a tooltip,
replacing the connect block and the progress lines beneath the hero.
The disconnected cross is drawn grey rather than red, and action
errors expire after a couple of seconds instead of sticking forever.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Click the Dropbox hero icon to pause or resume the daemon (dropbox-cli
stop/start), with a tooltip on hover. When paused, both the hero and the
bar icon grey out and the phrase reads "Syncing paused". The greying is
optimistic so it reacts the instant you click, then reconciles with the
real daemon state once it settles (reverting if the command failed).
Also fix status.py's running detection, which was stuck reporting true:
this dropbox-cli inverts the `running` exit code and reports "isn't
running" (not "not running") when stopped, so the paused state never
showed. Detection now keys off the status text alone.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Suppression was enabled only after cleanup_legacy_user_paths had already
moved ~/.local/share/omarchy aside, so the live session could auto-reload
during the swap window and pop "source= globbing error: found no match"
for every legacy hyprland.conf source line. Enable suppression as the
first step of the upgrade sequence instead; the EXIT trap still restores
it if the upgrade dies early.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Extract run_as_user_omarchy and run_as_user_wayland_session helpers to
replace ten hand-rolled run_as_user env blocks
- Declare the retired user-unit list once instead of three times
- Use git ls-files --others to find user-added theme backgrounds instead
of diffing against the tracked list by hand
- Stop rewriting the SDDM 10-theme/10-wayland confs the omarchy package
already installs, and drop the duplicate [Theme] section from
99-omarchy-login.conf
- Drop the legacy-path rewrite in the theme hyprland.conf shim; the user
transition already rewrote those references
- Remove dead code: the unreachable channel re-validation, the declare -F
guard in cleanup_on_exit, and the unset block at the end of the
user-transition script
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The user-transition script is fed to bash on stdin, so the browser-close
gum confirm read the script itself: it auto-declined and swallowed the
remaining script text, leaving bash to fail on a mid-construct syntax
error. The outer confirm and reboot prompts had the same exposure when
the script runs via curl | bash. Read all three from /dev/tty.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The Copy URL shortcut repair checked the outer script's yes variable,
which never reached the user-transition heredoc's bash process, so the
upgrade died with "yes: unbound variable" under set -u whenever a
browser was running with the old extension id in its Preferences.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
graphical.target was gated on network-online.target: cups-browsed orders
itself After=network-online.target and sits in multi-user.target, so the
session waited for NetworkManager-wait-online (DHCP/Wi-Fi association)
before coming up. On the bench VM that was ~140ms, but it's the classic
multi-second boot stall on real Wi-Fi.
Mask NetworkManager-wait-online.service so network-online.target is
reached immediately and nothing in the boot waits for connectivity;
NetworkManager still comes up and cups-browsed browses once the link is
live. Mirrors the existing systemd-networkd-wait-online mask. Enabled at
install (enable-services.sh) and masked on existing installs via
migration.
Verified on a fresh boot: the graphical.target critical chain no longer
passes through network-online.target / NetworkManager-wait-online.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
gum_env.lua is fed into the compositor environment via hl.env at login,
but that environment is captured once and is not refreshed on a theme
switch. Terminals launched through uwsm-app inherit the caller's
environment (scopes don't read systemctl --user set-environment), so a
floating terminal would show the login-time theme's gum colors.
Extract the export logic into a sourced omarchy-restart-gum helper and
source it from the presentation terminal launcher so gum widgets match
the active theme.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
system-config-printer ships /etc/xdg/autostart/print-applet.desktop, so
installing it for the printer GUI also starts a ~24 MiB Python tray applet
in every session whose only job is print-job notifications. Override it
with a Hidden=true entry in config/autostart (same mechanism already used
for fcitx5 and limine-snapper-notify), so the applet no longer autostarts.
Printing and the system-config-printer GUI are unaffected.
Verified: with the override in place, systemd's xdg-autostart generator no
longer produces app-print\x2dapplet@autostart.service.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Guards the dock-gating in omarchy-hyprland-monitor-watch: the poll runs
only when omarchy-hw-laptop and an external monitor are both present, via
sync_poll_state, with the event loop fed by process substitution so it
can start/stop the poll. Behaviourally simulated (stubbed hw checks +
synthetic monitor events): the poll runs only while docked and stops on
undock; a desktop or undocked laptop never polls.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The media service ran a 500ms repeat timer calling syncPlayingOrder() for
the whole session — two CPU wakeups a second even with nothing playing.
syncPlayingOrder only depends on the set of MPRIS players and each
player's isPlaying state, both of which the Mpris service already signals.
Replace the timer with onPlayersChanged (players appearing/disappearing)
plus an Instantiator that connects isPlayingChanged for each live player,
and a Component.onCompleted for the initial sync.
Verified with a test MPRIS player: play/pause and player add/remove all
update the active player with no periodic timer.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The reconciliation poll (added in "Fix clamshell display recovery") is a
backstop for lid transitions the Hyprland switch binds can miss across
suspend/resume. But the internal panel is only ever disabled while the
laptop is docked — lid shut with an external monitor active — so the poll
has nothing to reconcile when undocked. Running it 24/7 on every laptop
wakes the CPU every 2s for no reason, which is exactly the case where
battery matters most.
Start the poll when an external monitor is present and stop it when the
last one goes away (driven by the socat monitor-add/remove watch), gated
on omarchy-hw-laptop. Undocked laptops and desktops now never poll; a
docked laptop keeps the 2s recovery poll. Switch the event reader from a
pipe to process substitution so it runs in the main shell and can manage
the poll's lifetime.
Desktop verified: no poll runs even with external monitors attached. The
docked-laptop start/stop path needs a check on real laptop hardware.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
omarchy-hyprland-monitor-watch ran poll_clamshell_state() every 2s for
the whole session, and each pass forks omarchy-hyprland-monitor-clamshell
plus its hyprctl/jq/monitor-* children. Clamshell (lid) handling can only
ever apply to a machine with a lid, so on desktops and VMs this was pure
churn — the single largest remaining source of idle process wakeups
(~3 forks/sec on an otherwise idle desktop).
Gate the poll behind a new omarchy-hw-laptop helper (lid button, or a
laptop DMI chassis type as a fallback). The event-driven socat watch on
Hyprland monitor add/remove is unchanged, so monitor hotplug still
reconciles everywhere; only the periodic lid poll is now laptop-only.
Verified on a desktop: steady-state omarchy-hyprland-* forks drop from
~3/sec to zero, socat watch still present.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The pill was populated by polling omarchy-network-status (previously
every 3s, then 10s). Derive kind/signalStrength directly from the
Quickshell.Networking service instead: wired-connected → ethernet
(preferred, matching the default-route device), else a connected Wi-Fi
network → wifi with its signal, else disconnected. The service's
properties are change-notified, so the pill updates on connect/
disconnect and signal changes with no polling and no subprocess.
Removes the last periodic poll from the bar; the label/frequency pill
fields were write-only and are dropped.
Verified live: pill tracks ethernet → disconnected → ethernet in real
time as the wired device goes down and up.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Each `bash -lc` starts a login shell that re-sources the profile
(mise activation, /etc/profile.d) on every invocation — ~16 forks per
call versus ~2 for `bash -c` — which taxes every menu/panel/launcher
action the shell shells out for. The session already exports PATH and
env to the shell, so omarchy commands resolve fine under `bash -c`.
Switch the internal/omarchy-owned spawns (theme+background switches,
brightness, monitor scaling, DNS, lock/fingerprint, keyboard-layout
probe, voxtype status, and the `:`/printf state-file writes) to
`bash -c`. Leave `bash -lc` on the sites that run user-configurable
commands (custom bar-widget exec, menu provider/guard scripts,
launcher scan commands, configurable idle/screensaver command), where
a user's command may rely on their login environment.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The monitor panel polled `omarchy-monitor-state` every 5s (26 forks per
call) and the network panel polled `omarchy-network-status` every 3s,
both with `running: true` rather than gating like the sibling panels.
Monitor: gate the poll on `opened` (brightness/scale are only shown in
the open panel, which already refreshes on open) and drive the bar glyph
from Quickshell.screens instead of the polled display list, so monitor
count stays correct without polling.
Network: the bar pill needs live-ish state, so keep polling but at 10s
instead of 3s. (Deriving the pill from the native Networking service
would remove the poll entirely; left as a follow-up since it needs Wi-Fi
state testing.)
Idle fork rate drops ~20/s to ~14/s.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The indicators widget broadcast a refresh every 2s, fanning out to four
status subprocesses (nightlight, idle, reminder, screen-recording), and
NightLight/StayAwake each ran an additional 5s poll. On an idle desktop
this was the dominant source of process churn (~33 of ~53 forks/sec in a
VM). The state-changing commands already push `omarchy.indicators
refresh` over IPC, so the polling was redundant.
Drop the 2s broadcast and the two 5s timers; indicators now refresh at
startup and on the IPC push. Also fix three callers that pushed to the
wrong target `Indicators` instead of `omarchy.indicators` (screen
recording, notification silencing, and the omarchy-shell help example) —
those pushes silently failed and only appeared to work because the poll
masked them.
Idle fork rate drops ~53/s to ~20/s.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The background plugin ran a 100ms Timer that spawned a readlink
subprocess ~10x/sec for the life of the shell, purely to notice
wallpaper changes. Wallpaper updates already arrive over IPC
(omarchy-theme-bg-set calls `background set`, omarchy-theme-set calls
`background themeTransition`), and the initial wallpaper is read once
in Component.onCompleted, so the poll was redundant.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
omarchy-update-user-notify.path used PathExistsGlob= on the packaged
migrations directory. That directive is level-triggered: systemd re-checks
it every time the triggered unit deactivates and fires again while the glob
still matches. Since applied migrations stay on disk forever (state lives in
~/.local/state/omarchy/migrations), the glob always matches, so the oneshot
notifier re-triggered itself in a tight loop (~26-66 starts/sec) — burning
about a core and flooding the journal for the whole session.
The loop existed since the unit was introduced, but the default start-rate
limit killed it after 5 iterations, taking the .path unit down with
'unit-start-limit-hit'. That symptom was reported as #6174 and fixed
yesterday by setting StartLimitIntervalSec=0 — which removed the only brake
and turned the capped hiccup into an unbounded busy-loop.
Fix the actual cause instead:
* Drop PathExistsGlob= from the .path unit, keeping the edge-triggered
PathModified= watch for updates that land mid-session.
* Revert the StartLimitIntervalSec=0 override; with the level trigger gone
there is no self-re-fire to trip the limit, and the default limit is a
useful backstop again.
* Preserve the once-per-login pending check the glob used to provide by
giving the service its own WantedBy=graphical-session.target, enabled at
first-run alongside the other user units.
* Add a migration that daemon-reloads, revives a rate-limit-killed .path,
restarts the watcher, and enables the login-time notifier on existing
installs.
Verified with transient path/service units: the old config runs the service
200 times in 3 seconds; the new config runs it zero times while idle and
exactly once when a new migration file lands.
Thanks to @HANCORE-Linux for finding and diagnosing the problem.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The tree is wholly package-owned, but files can land there unowned
(in-place extension work, script-written files), and pacman then aborts
the entire upgrade on the file conflict — as happened when copy-url's
service worker was renamed to background-2.js. The conflict check runs
before any hooks or scriptlets, so the package itself can't recover;
the update command has to allow the overwrite.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
