Omarchy 3.5.1 shipped hardware-specific mic mute handling for Dell XPS
(bin/omarchy-cmd-mic-mute-xps) which made the original PR #4938 approach
obsolete — that PR added a global omarchy-cmd-mic-mute that handled the
toggle inline, but upstream now owns that dispatcher.
This follows the same pattern to add ThinkPad support by extending
omarchy-hw-match to also check product_family (ThinkPad systems report
"ThinkPad T14s Gen 2a" there, not in product_name) and adding
omarchy-cmd-mic-mute-thinkpad which uses brightnessctl to sync the
platform::micmute LED, simpler than the ALSA approach needed for XPS.
Note: as more vendors are added this per-vendor dispatcher will get
unwieldy. A future refactor could move hardware LED sync into
install/config/hardware/ scripts that drop a hook or config file,
keeping omarchy-cmd-mic-mute thin and vendor-agnostic.
pactl set-default-sink only updates PipeWire's default.audio.sink
metadata, but WirePlumber immediately overrides it back because
default.configured.audio.sink remains unchanged. wpctl set-default
updates both, making the switch actually persist.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* Normalize screen recording audio
Screen recordings capture audio at whatever the system level is set to,
which often results in very quiet audio. Add a post-processing step that
normalizes audio to -14 LUFS using ffmpeg loudnorm after the first-frame
trim. Video is copied without re-encoding. Silently skipped for
recordings with no audio track.
* Do both trimming and normalizing in a single step so we don't have to process the video twice
* Simplify conditional
---------
Co-authored-by: David Heinemeier Hansson <david@hey.com>
chromium/brave --refresh-platform-policy blocks forever when the
browser is not already running, hanging omarchy-update entirely.
Guard both calls with pgrep so the policy file is still written
(picked up on next launch) but the blocking refresh is skipped.
Closes#4772
omarchy-refresh-hyprland overwrites input.conf with the default
template, which has kb_layout and kb_variant commented out. Non-US
users lose their keyboard layout with no obvious recovery path.
Re-run detect-keyboard-layout.sh after the refresh so the system
layout from /etc/vconsole.conf is re-applied, matching what the
initial install does.
Closes#2507
* feat: add touchpad toggle keybinding and script
Binds XF86TouchpadToggle/On/Off to a new leenium-toggle-touchpad script
that enables/disables the touchpad via hyprctl and shows OSD feedback.
* fix: make XF86TouchpadOn/Off enforce state instead of toggling
Add on/off/toggle argument support to omarchy-toggle-touchpad.
XF86TouchpadOn now always enables and XF86TouchpadOff always disables,
so firmware that emits separate On/Off keysyms behaves correctly regardless
of current state. XF86TouchpadToggle retains the invert behavior.
* feat: add touchpad toggle to menu and persist state as hyprland config
- Show touchpad option in toggle menu only when a touchpad/trackpad device is detected
- Replace /tmp state file with a persistent hyprland device config at
~/.local/state/omarchy/toggles/hypr/touchpad-disabled.conf
* fix: avoid unsafe && || toggle logic for Touchpad state
Replace `[[ -f $STATE_CONF ]] && enable || disable` with an explicit if/else.
The previous logic was not a true ternary: if `enable` failed (e.g. OSD command error),
`disable` would run unintentionally, causing unreliable toggle behavior.
* Clean up orphaned voxtype status processes on exit
Waybar's custom/voxtype module runs omarchy-voxtype-status, which spawns
a long-lived `voxtype status --follow` child. When Waybar reloads (config
change, display hotplug, Hyprland reload), it kills the wrapper but the
child survives as an orphan. Over time dozens accumulate.
`trap 'kill 0' EXIT` signals the entire process group when the wrapper
exits, ensuring the voxtype child is cleaned up.
* Explain purpose of trap
---------
Co-authored-by: David Heinemeier Hansson <david@hey.com>
* Add keybindings to cycle through monitors
* Unified language around focus
* Group focus keybindings
* More natural sounding
---------
Co-authored-by: David Heinemeier Hansson <david@hey.com>
Redirect stderr to /dev/null alongside stdout when refreshing
browser platform policy, preventing shared_memory_switch errors
from cluttering the terminal.
