1 +#
Quickshell Bar Wheel Validation Attempt — 2026-06-17 5 +Attempted to close the Phase 2B gap for volume and brightness wheel interactions on the current Quickshell bar.
9 +Temporary `XDG_CONFIG_HOME` named config was used; no live home symlink mutation.
11 +Current bar screenshot:
13 +- `/tmp/nugu-qs-bar-locate.png`
15 +Vision/crop analysis found the visible Quickshell bar widget centers:
22 +Old Ironbar was visible above the Quickshell test bar; Quickshell bar was the lower bar and coordinates targeted the Quickshell widgets.
24 +##
Commands attempted 26 +State was saved before testing and restored afterward:
28 +- PipeWire sink volume
29 +- PipeWire sink mute state
30 +- Backlight raw value
32 +Synthetic wheel input used `wdotool --backend wlr-protocols`:
35 +wdotool --backend wlr-protocols mousemove 1178 52
36 +wdotool --backend wlr-protocols scroll 0 1
37 +wdotool --backend wlr-protocols scroll 0 -1
39 +wdotool --backend wlr-protocols mousemove 1254 52
40 +wdotool --backend wlr-protocols scroll 0 1
41 +wdotool --backend wlr-protocols scroll 0 -1
46 +Before attempting a code change:
49 +vol_before=Volume: 0.30 [MUTED]
50 +vol_after_down=Volume: 0.30 [MUTED]
51 +vol_after_up=Volume: 0.30 [MUTED]
53 +bright_after_down=96%
57 +A hypothesis was tested that `WheelHandler` was not receiving synthetic layer-shell wheel events while the `MouseArea` owned the hit region. The minimal code change routed wheel handling through `MouseArea.onWheel`. Retest still produced no state change:
60 +vol_before=Volume: 0.30 [MUTED]
61 +vol_after_scroll_pos=Volume: 0.30 [MUTED]
62 +vol_after_scroll_neg=Volume: 0.30 [MUTED]
64 +bright_after_scroll_pos=96%
65 +bright_after_scroll_neg=96%
68 +The unproven code change was reverted.
73 +INFO: Configuration Loaded
76 +No QML errors, binding loops, reference errors, or lingering `qs` process were observed.
78 +PipeWire emitted rtkit/session-bus warnings from the agent shell, consistent with the existing agent DBus limitations, but `wpctl` state reads and explicit restores completed.
82 +Wheel parity remains **unverified**. Current evidence points to synthetic wheel injection not proving the behavior in this harness. Do not count this row as accepted until either:
84 +- a live manual wheel test is observed, or
85 +- a better automation path can deliver Wayland wheel events to the layer-shell widget and show state changes.
89 +- Ironbar stayed active.
90 +- Mako stayed active.
91 +- Quickshell service stayed inactive/not live.
92 +- Volume, mute state, and brightness were restored after testing.