Skip to content

Fix empty-desktop bar reposition by Top-only expand (#11915) - #12646

Open
paulogeyer wants to merge 2 commits into
omacom:quattrofrom
paulogeyer:fix/11915-empty-desktop-bar-top-expand
Open

paulogeyer wants to merge 2 commits into
omacom:quattrofrom
paulogeyer:fix/11915-empty-desktop-bar-top-expand

Conversation

@paulogeyer

@paulogeyer paulogeyer commented Sep 20, 2026 •

Copy link
Copy Markdown
Contributor

Summary

  • Bug: Reposition-dragging the bar fails on an empty desktop (Dragging the bar only works when there is at least 1 window (broken on empty desktop) #11915). With a window open it works; without one, the gesture starts then dies when the pointer leaves the strip.
  • Fix: While moving, expand the owning BarPanel to fill its output on WlrLayer.Top, so the same CenterGestureArea keeps receiving moves and the release. One panel-level gesture area per orientation; the strip (sized by the bar's notch-aware thickness) stays pinned to the original edge, and the move-edge ghosts stay Overlay and input-transparent.
  • Notifications: Toast cards sit on Overlay above the expanded bar, so their input region now empties while the bar reports a move. The cards stay visible, and a drag across them on an empty desktop no longer ends where they start.
  • Indicator peek: The move only force-collapses the peek when the bar relocates or the pointer ended off the strip, which covers a pointer the panel shrinks out from under. A press, hold and release in place keeps the peek, decided by the real hover state.

Relation to #12184

#12184 is still open and takes a different handoff/overlay approach. Live soak on my laptop: that PR still fails on empty desktop. This PR is an intentional competing fix that passed empty + windowed press-hold-drag-release soak on my laptop (ASUS Vivobook, Intel Arc).

Updated for current quattro

Re-done on quattro 246f33a after #15517 rather than rebased. The strip wraps #15517's widget pools and CenterSections/CenterBox unchanged, the panel-level CenterGestureArea is the only one (the one inside CenterModules is gone), and clearBarMove no longer touches the removed centerSectionHovered. Instead, when the bar relocates or the move ends off the strip, the moving panel's HoverHandler gives back the hover it counted in barHoverCount, and it tracks what it counted so a late leave never subtracts twice. A press-and-hold released in place keeps the peek.

Trade-off

While expanded, exclusion is Ignore, so tiled clients reflow into the strip area for the drag (including on occupied workspaces and cancelled drags), and again on release. I prefer this to the Overlay re-layer blink.

Known limitation: moves that leave the strip before the expanded bar takes effect

The fix relies on the expanded bar surface being in place before the pointer leaves the bar strip. Any move that leaves the strip before that happens can drop, and how often depends on how quickly the machine draws the expanded surface. That covers quick flicks, and a slower first draw such as right after a display wake. In review, a 1.5 s hold moved the bar 8/8, while a 0.4 s hold dropped 6 of 12 moves. On my laptop (Hyprland 0.56.2, one output), with a 0.3–0.4 s hold, the first drag after a wake worked 19/19, and a flick moving within about 20 ms of the press always failed. quattro has the same race (0/3 on a quick flick with a window open, 3/3 with a normal hold), so this change doesn't introduce it. Expanding on press would narrow the race, giving the compositor the hold time to draw the larger surface, but not close it, and it reflows tiled windows on every click on the bar, so I've left that out of this PR.

What this deliberately avoids

  • HyprlandFocusGrab / seat grab
  • HoverHandler pressedButtons overlay handoff
  • Flipping the bar Top → Overlay during the move (visible blink)
  • Hover auto-dock timers

Test plan

  • bar-test.sh 269 ok, bar-notch-test.sh 99, notifications-test.sh 150, clock-test.sh 167, bar-drag-test.sh, bar-click-clipping-test.sh, bar-text-color-test.sh pass. Added checks: each anchor inside the BarPanel anchors block, the bottom/right strip offsets, a pointInBarStrip table and the toast input mask, once each. On quattro with these tests, bar, notch and notifications fail.
  • Virtual mouse, empty workspace: on quattro, 0/8 across top→left→bottom→right→top twice. With this PR, 8/8, and 3/3 across toast cards.
  • Live, virtual mouse plus screenshots: a press-hold-release in place keeps the indicator peek, and it collapses once the pointer leaves. A drag released out on the desktop with no edge change collapses it, and hovering again reopens it. Across toast cards: 3/3 with this PR, 0/3 with only Service.qml reverted. Glassy background: 4/4 drags (the blur itself not checked).
  • Reviewer: multi-monitor if available

Fixes #11915

Fixes #7923

@llstrk

llstrk commented Sep 23, 2026

Copy link
Copy Markdown

No production defect was identified in the source review. The empty-desktop drag fix still needs independent live validation.

Verified by source inspection: only the moving panel expands, remains on the Top layer, and keeps its strip content at the original edge. The gesture area grows with the panel; release/cancel clears the move. The Overlay previews remain input-transparent.

Idle strip -> moving panel fills its output -> release/cancel restores strip
              same MouseArea, same Top layer

Optional test improvement: the four-anchor assertion can match moveExpanded occurrences after the anchors block. An isolated assertion check demonstrated:

Input Current assertion Block-scoped check
PR source Pass Pass
Diagnostic copy with expansion removed from all four anchors Pass Fail

Constrain the match to the anchors block and check top, bottom, left and right individually. This catches removal of the mechanism the test is intended to protect. These are assertion-only results, not a complete-suite run or a production regression.


Review information

Test scope: Head 8aeea9e. Source inspection and isolated source-assertion checks. The full test suite, live empty/windowed desktop drags, visual continuity and multi-monitor behavior were not independently tested. Temporary tiled-client reflow is an acknowledged trade-off of the expansion.

Community review: Independent automated community review, unaffiliated with the Omarchy team, intended to help prepare PRs for their review.

Automated AI review: Astra Medium initial inspection, independent Opus 5.5 High and GPT 6 Sol Xhigh assessments, followed by coordinator verification of the cited source and assertion results.

@omarchybot omarchybot added the bug Something isn't working label Sep 27, 2026
@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 1, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed at 8aeea9e by Claude Opus 5.5, with Codex Medium as a second opinion, and run on a worker VM built from the Omarchy ISO (Hyprland 0.56.2, one output, a virtual mouse that presses, holds, drags and releases).

The fix works for the reported case. On quattro (8b4eae6), dragging the top bar to the left edge fails on an empty workspace and works with a window open, which matches #11915. On this head the same drag works on an empty workspace, and so do top→bottom, bottom→top, right→bottom, bottom→right and left→bottom, with and without a window. A short drag released near the current edge leaves the bar where it was. With only Bar.qml reverted on this head, the empty-desktop drag fails again. ./test/cli passes, and so do bar-test.sh (113/113) and the other seven shell tests that touch the bar. I've linked #11915 so it closes with this.

Three things are still open:

  • The drag still dies when it crosses a notification on an empty desktop. With the first-run notification cards stacked at the top right, dragging the top bar toward the right edge ends partway: CenterGestureArea.onReleased fires at the left edge of the cards while the button is still held, so the move finishes as a no-op. It works with a window open, and it works on an empty workspace once the notifications are dismissed. This is the same focus loss as Dragging the bar only works when there is at least 1 window (broken on empty desktop) #11915, triggered by a different surface: notifications sit on the Overlay layer, and their input mask covers the whole toast column (shell/plugins/notifications/Service.qml:978), so they win the hit test over the expanded Top-layer bar. Codex traced that mechanism and the worker results agree. The smallest fix it suggests is to make that mask empty while shell.bar.barMoveActive is true. Other interactive overlays would still be a limit.
  • A press-and-hold that ends where it started now collapses the indicator peek under the pointer. Hover empty bar space so the indicators reveal, then press, hold and release without moving. On quattro the indicators stay revealed. On this head they collapse while the pointer is still on the bar, because clearBarMove sets centerSectionHovered = false (Bar.qml:521-523). The HoverHandler still thinks it's hovered, so nothing reveals the peek again until the pointer leaves and comes back. I confirmed this with screenshots on the worker. If the stuck peek this works around needs handling, it should come from the real hover state, not a forced un-hover.
  • Part of the diff doesn't affect the bug. beginBarMove on quattro never checked for mapped clients, and the bar's own window always has a screen. So the barMoveEligible({ mappedClients: [] }) call, resolveBarMoveScreen and the BarModel.js helpers behind them change nothing, and the comment saying client-count gating caused the failure doesn't match the base. Most of the new bar-test.sh checks match source text or reject identifiers from Fix bar reposition-drag on an empty desktop (#11915) #12184. They would still pass if the bottom/right strip offsets were removed, so they don't guard the gesture. The proportionate change is the panel expansion, the strip wrapper and the moved CenterGestureArea.

The exclusion change is a visible trade-off, as the description says. While the bar is expanded it reserves no space, so tiled windows resize at the start of every move, including on occupied workspaces and cancelled drags, and again when it ends.

Compared with #12184, both reviews prefer this approach. It keeps the original press on one surface rather than handing the gesture to an overlay that can't see the held button. The choice between them is still the maintainer's. This now waits on @paulogeyer for the notification and peek issues, then on the maintainer.

@omarchybot omarchybot removed the verified Omarchy Triage has verified that this issue is ready for final review label Oct 4, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

This no longer merges into quattro. The head is still 8aeea9e, so the two findings from the last review are still open. I re-read both sides of the conflict at quattro 50d687a. Codex Medium checked the same reading and agreed (it can read this session, so the agreement isn't guaranteed to be independent). Nothing was run this time.

What conflicts. In shell/plugins/bar/Bar.qml, #14267 (e86c8f1, "Reveal indicators only before the clock") put each center HoverHandler inside an Item that ends at the clock (centerAnchorModule.left horizontally, .top vertically), or fills the section when there is no clock. It also kept CenterGestureArea { anchors.fill: parent } on the center section (quattro Bar.qml:1566 and 1618). This PR removes those two gesture areas and leaves the HoverHandler filling the section. In test/shell.d/bar-test.sh, #11015 (454b67d) added startup-stall checks at the same spot as your drag checks.

What a resolution keeps. Keep quattro's bounded hover Item, including its no-clock fallback. Keep your single panel-level CenterGestureArea per orientation, and don't restore the center-section ones. Keep both sets of test checks. Nothing else on quattro since this branched touches beginBarMove/clearBarMove, the panel's layer or exclusion, or the notification input mask, so the approach still applies and #11915 is still broken on quattro.

Still open after the merge. Crossing the notification cards on an empty desktop still ends the drag, because their Overlay input mask is unchanged (shell/plugins/notifications/Service.qml:999-1022 on quattro). A stationary press-hold-release still forces centerSectionHovered = false (Bar.qml:523), so it collapses a peek under the pointer. After #14267 that happens only in the space before the clock, and the bar-test.sh check that requires the forced collapse would change with any fix.

This waits on @paulogeyer to bring the branch up to date and address the two findings. After that it goes to the maintainer, who still has to choose this approach.

@jandrusk

Copy link
Copy Markdown

Fixes #11915 (bar reposition/drag dies on an empty desktop) via Top-only expand while moving. Approach makes sense.

Currently CONFLICTING on Bar.qml/BarModel.js — needs a rebase before review. Happy to re-check the gesture edge cases (multi-monitor, vertical bar) once it’s green.

@paulogeyer
paulogeyer force-pushed the fix/11915-empty-desktop-bar-top-expand branch from 8aeea9e to b7aba24 Compare October 14, 2026 13:27
@paulogeyer

Copy link
Copy Markdown
Contributor Author

Thanks for the detailed reviews. I rebased onto current quattro (84cfb93) and addressed both findings:

  • Conflict: kept Reveal indicators only before the clock #14267's bounded hover Item (with the no-clock fallback), one panel-level CenterGestureArea per orientation, and no center-section gesture areas. Both sets of bar-test.sh checks are kept. The expanded strip now uses the notch-aware thickness and Glassy colour, and keeps the centerBesideRight module list.
  • Notifications: the toast popup's input mask is empty while shell.bar.barMoveActive is true (Region { item: service.barMoveActive ? null : popupColumn }). The cards stay visible but no longer take the pointer mid-drag.
  • Peek: clearBarMove no longer un-hovers unconditionally. It drops the peek only when the bar relocates or the last move update left the pointer off the strip (BarModel.pointInBarStrip). A press, hold and release in place leaves it to the real hover state.
  • Smaller diff: dropped barMoveEligible / resolveBarMoveScreen and their checks, as you suggested.
  • Tests: applied @llstrk's suggestion, checking top/bottom/left/right individually inside the anchors block, and added checks for the bottom/right strip offsets. The new assertions fail on the rebase without the fixes and pass with them.

I've live-soaked it on my laptop, dragging the bar between all edges with and without windows, and had no issues. The description is updated.

Thanks @jandrusk for the review, and for offering to re-check multi-monitor and vertical bar layouts. It's green and ready for that whenever you have time.

@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 14, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed again at b7aba24 against quattro 84cfb93 by Claude Opus 5.5, with Codex Medium as a second opinion. I ran it on a worker VM built from the Omarchy ISO (Hyprland, one 1280×800 output, a virtual mouse that presses, drags and releases).

The fix works, and both earlier findings are fixed. On quattro an empty-desktop move fails every time (0/8, in a round of top→left, left→bottom, bottom→right and right→top) and works with a window open (8/8). On this head the same round passes 12/12 on an empty desktop and 8/8 with a window. With only Bar.qml reverted, the empty-desktop move fails again. Dragging the top bar across the toast cards to the right edge now works on an empty desktop (3/3), and with only Service.qml reverted it fails (0/3), so the empty input mask is what fixes it. A press-hold-release in place keeps the indicator peek, the same as on quattro. ./test/cli passes, and so do bar-test.sh (120/120), notifications-test.sh (150/150), bar-notch-test.sh (96/96) and the ten other shell tests that touch the bar or notifications.

Two things are still open:

  • The first drag after the display wakes can still drop on an empty desktop. After hyprctl dispatch 'hl.dsp.dpms({ on = true })', the first top→left drag failed 5 of 5 times when it started about a second later. After a real off→on wake it failed 1 of 3. A drag right after that always worked, and so did every drag with a window open. Sampling the layers during a failed drag shows the bar expanding to 1280×800 and the move ghost appearing. About 0.1s later both are gone, as the pointer leaves the strip with the button still held. That is Dragging the bar only works when there is at least 1 window (broken on empty desktop) #11915's focus loss again, in a narrower window. Codex reads it the same way: the fix holds only if the expanded surface takes effect before the pointer leaves the strip, and something about the first expansion after an output event can be late. ScreenMoveRemap watches only screen x/y, so it doesn't explain a redundant dpms on. Neither of us has confirmed the cause. If it isn't reasonable to fix here, the description should say it as a known limitation.
  • The test additions are bigger than the change. The Bar.qml and Service.qml changes are the right size for the bug. The 137 lines added to test/shell.d/bar-test.sh mostly match source text, and several checks are made twice, once in the bash half and again in the node half: ownership, exclusion, layer, gesture placement and release (lines 56–146 and 415–443). The checks at lines 94, 117 and 122, and their node copies at 428 and 432, ban identifiers that never existed on quattro. They lock out an abandoned approach rather than guard this one. An equivalent refactor of Bar.qml would fail them, and the wake failure above passes them all. Keeping the per-edge anchor check, the strip offsets, the pointInBarStrip table and the notification mask check, once each, would cover what these tests can actually catch.

Codex Medium also checked the peek bookkeeping, the bottom/right coordinate offset during the expand, Region { item: null } (an empty input region in Quickshell, restored when the move ends) and the multi-monitor behaviour, and found no defect. It agreed with the rest. It can read this session, so that agreement isn't guaranteed to be independent.

This waits on @paulogeyer to trim the tests, and either to look into the wake case or to name it in the description. After that it goes to the maintainer.

@omarchybot omarchybot removed the verified Omarchy Triage has verified that this issue is ready for final review label Oct 15, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

This no longer merges into quattro. The head is still b7aba24, so the two points from the last review (the first drag after a display wake, and the size of the bar-test.sh additions) are still open. I read both sides of the conflict at quattro 246f33a, and Codex Medium read them separately. It agreed and added the point about relocation below. It can read this session, so that agreement isn't guaranteed to be independent. Nothing was run this time.

What conflicts. All four conflicts are in shell/plugins/bar/Bar.qml, and they come from #15517 (b7e900c, "Rework bar dragging, sections and widget orientation"). In horizontalBar and verticalBar, quattro now builds the modules from a WidgetPool with window/pool wiring, and puts CenterSections beside the right section under a notch. This PR wraps the old module list in the strip Item under one panel-level CenterGestureArea. Further down, quattro replaced the anchor-based center trees, and the bounded hover Item from #14267, with CenterSections and CenterBox.

What breaks even where it merges cleanly.

  • clearBarMove sets centerSectionHovered = false, and Rework bar dragging, sections and widget orientation #15517 removed that property. When the bar relocates, finishBarMove calls clearBarMove(true) before setBarPosition(edge). So after a merge that only resolves the markers, the write throws and the bar never moves to the new edge. The bar-test.sh regex at line 102 requires that assignment, so it has to change too.
  • The peek now runs on a single HoverHandler on the panel's Loader, which counts bar hover through setBarHovered/barHoverCount. While the panel is expanded, the Loader fills the screen, so the whole desktop counts as hovering the bar. When the panel shrinks back without a leave event, that count can get stuck and hold the peek open. That's the case your clearBarMove peek handling covered, and it now has to settle barHoverCount instead.
  • quattro's CenterModules has its own CenterGestureArea { anchors.fill: parent } again. With your panel-level one as well, a press on the strip reaches the strip-sized one first, so drop the inner one.

What a resolution keeps. From this PR: the owning panel's expand to all four anchors, Ignore exclusion while expanded, the transparent panel with the edge-positioned strip, the one panel-level gesture area, and the empty toast input mask during a move. From quattro: the widget pools, the window/pool/drawn/active wiring and CenterSections/CenterBox, inside your strip in place of the old center trees. #15501 (41d2f9d, clipped click forwarding) merges cleanly and should stay.

Still broken on quattro. beginBarMove, updateBarMove, clearBarMove and finishBarMove are unchanged, and so are the panel's anchors, exclusion and layer. The bar-move gesture is still a strip-sized area under the center modules. The drag work in #15517 is about moving widgets, so #11915 is still broken and this approach still applies.

This waits on @paulogeyer to bring it onto current quattro, and on the two earlier points. After that it gets run on a worker again: empty and windowed desktops, all four edges, press-and-hold in place, and the peek after a move.

@paulogeyer
paulogeyer force-pushed the fix/11915-empty-desktop-bar-top-expand branch from b7aba24 to f3ab44a Compare October 15, 2026 13:41
@paulogeyer

Copy link
Copy Markdown
Contributor Author

Rebuilt on quattro after #15517, not rebased. There's now one panel-level gesture area (the one inside CenterModules is gone). centerSectionHovered is gone. When the bar relocates or a move ends off the strip, the moving panel gives back the hover it counted in barHoverCount, so the peek can't stick.

The test additions are trimmed to the four checks you named, once each: the per-edge anchor check, the strip offsets, the pointInBarStrip table and the toast mask.

Virtual mouse on an empty workspace on my laptop: quattro 0/8, this head 8/8. Across toast cards: 3/3, and 0/3 with only Service.qml reverted.

I couldn't reproduce the wake case on my laptop: the first drag after a dpms wake worked 19/19 with a normal hold. Flicks that move within about 20 ms of the press fail on quattro too, even with a window open. The description notes this as a known limitation.

@greptile-apps

greptile-apps Bot commented Oct 15, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerView in GreptileConfidence Score: 5/5

[Medium impact] The PR appears safe to merge; the previous cloned-toast issue is fixed.

Summary

The PR expands the moving bar across its screen on the Top layer and disables toast input during the move.

  • An active bar drag keeps its gesture area across the output.
  • Toast cards stop taking pointer input during bar moves.

Diagram

sequenceDiagram
  participant Bar
  participant Shared as PluginBarStateApi
  participant Toasts as Cloned notifications
  Bar->>Bar: Start move and expand
  Bar->>Shared: barMoveActive becomes true
  Shared->>Toasts: Disable toast input
  Bar->>Bar: Finish move and shrink
  Bar->>Shared: barMoveActive becomes false
  Shared->>Toasts: Restore toast input
Loading

Reviews (3) · Last reviewed commit: "Pass the bar move state to cloned notifi..." · Reviewed by Greptile

@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 15, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed again at f3ab44a against quattro 8a787b6 by Claude Opus 5.5, with Codex Medium as a second opinion. I ran it on a worker VM built from the Omarchy ISO (Hyprland 0.56.2, one 1280×800 output, no GPU acceleration, a virtual mouse that presses, holds, drags and releases).

Both points from the last review are addressed. The test additions are down to the four checks, and the rebuild onto #15517 keeps one panel-level gesture area and settles barHoverCount through counted. Codex Medium went through release on and off the strip, relocation, cancel, monitor unplug and two monitors, and found no path that subtracts twice or leaves the count stuck. It also found no regression in the notch, vertical, Glassy or widget-drag paths.

What works. On quattro an empty-desktop move fails every time (0/8) and works with a window open (4/4). On this head, with the press held 1.5 s before moving, every empty-desktop move works (8/8 across top→left, left→right, right→top and top→bottom), and with only Bar.qml reverted they all fail again (0/4). With a window open it works at any pace. Dragging the top bar across three toast cards to the right edge works (3/3), and fails with only Service.qml reverted (0/2). A press-hold-release in place keeps the indicator peek. ./test/cli passes, and so do bar-test.sh (269), bar-notch-test.sh (99), notifications-test.sh (150), clock-test.sh (167), dock-test.sh (39) and nine other bar and notification tests.

What still fails. With the press held 0.4 s, half the empty-desktop moves still drop (6/12). I added logging to see why. Press-and-hold starts the move at 200 ms, and the panel then takes 113–178 ms to reach the full output (once it still hadn't after 188 ms). Every failed move ended in onReleased with the button still held down, at coordinates inside the strip, just as the pointer first left the 26 px strip. So a move survives only if the expanded surface takes effect before the pointer leaves the strip. That's the same race as the wake case. The rebuild didn't introduce it: b7aba24 does the same on this worker today, though it passed 12/12 at the same pace two days ago. How often it fails depends on how fast the machine draws the expanded surface, which fits it showing only on flicks on your laptop and on ordinary drags on this VM. Codex Medium reads the code the same way: the expansion is requested only once the move is recognised, and nothing keeps the pointer while the compositor configures the larger surface and the first frame at the new size is drawn.

Codex Medium added two points that follow from this:

  • The Known limitation section describes quick flicks and the first drag after a wake. More accurately, any move that leaves the strip before the expanded surface takes effect can end, including after a press-and-hold, and a slower machine makes that window wider.
  • Starting the expansion on a press over empty bar space, only when the workspace has no windows, would give it about 200 ms more without the reflow you wanted to avoid, since Dragging the bar only works when there is at least 1 window (broken on empty desktop) #11915 only happens where nothing is tiled. It narrows the race but doesn't close it.

It goes to the maintainer now, to decide whether to take it with the race as it stands or ask for the earlier start. Either way, the description should state the limitation as above. Codex Medium agreed with the rest of this. It can read this session, so that agreement isn't guaranteed to be independent.

@paulogeyer

Copy link
Copy Markdown
Contributor Author

Reworded the Known limitation section: any move that leaves the strip before the expanded bar takes effect can drop, it depends on draw speed, and quattro has the same race. Left expand-on-press out of this PR.

@omarchybot

Copy link
Copy Markdown
Collaborator

Read again at f3ab44a by Claude Opus 5.5, after the description update. The head hasn't moved, so the worker results from the last review still stand. quattro has moved on by one commit (51bfd1d, the screenshot binding), which touches none of these files, and this still merges cleanly. CI is green.

The Known limitation section now states the race the way the last review asked. It says any move that leaves the strip before the expanded surface takes effect can drop, that how often depends on draw speed, and it gives the 0.4 s figures, which cover the press-and-hold case. One wording point, not a blocker: expanding on press would narrow the race, not close it. The compositor still has to configure and draw the larger surface before the pointer leaves the strip, and expanding on press only gives it the hold time to do that.

No second opinion ran this time. Codex Medium answers, but today's review budget is spent, so this read is mine alone.

This waits on the maintainer to decide whether to take it with the race as it stands or ask for the earlier start. Before it can be marked ready, the second opinion also needs one round at this head.

@omarchybot omarchybot removed the verified Omarchy Triage has verified that this issue is ready for final review label Oct 20, 2026
While a bar is being moved, its panel fills its output on the Top layer, so
the one panel-level gesture area that took the press keeps getting moves and
the release after the pointer leaves the strip (omacom#11915). The strip stays on
its own edge, exclusion is Ignore for the move, and the inner gesture area
in CenterModules is gone.

Toast cards take no input during a move, so a drag can cross them. A move
that relocates the bar or ends off the strip settles the moving panel's hover
in barHoverCount; a press-and-hold released in place keeps the peek.
@paulogeyer
paulogeyer force-pushed the fix/11915-empty-desktop-bar-top-expand branch from f3ab44a to 2709fa7 Compare October 21, 2026 02:29
Comment thread shell/plugins/notifications/Service.qml Resolved
@omarchybot omarchybot added verified Omarchy Triage has verified that this issue is ready for final review ready Good to merge labels Oct 21, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed again at b325725 against quattro 087f73d by Claude Opus 5.5, with Codex Medium as a second opinion. I ran it on a worker VM built from the Omarchy ISO (Hyprland, one 1280×800 output, a virtual mouse that presses, holds, drags and releases).

The fix works on the reworked notification deck, and nothing is open. With the press held 1.5 s before moving, an empty-desktop move fails every time on quattro (0/8, two rounds of top→left, left→right, right→top and top→bottom) and works every time on this head (8/8). With only Bar.qml reverted, all eight fail again. With a window open, this head moves 4/4. Dragging the top bar across three toast cards to the right edge works on this head (3/3) and fails on quattro (0/3) and with only Service.qml reverted (0/3), so the deck mask is what carries the move past the notification surface. ./test/cli passes, and so do bar-test.sh (269), bar-notch-test.sh (99), notifications-test.sh, clock-test.sh (167), dock-test.sh (39) and nine other bar and notification tests. The new checks in bar-test.sh, bar-notch-test.sh and notifications-test.sh fail on quattro.

Greptile's point about cloned notifications plugins holds, and b325725 fixes it: those plugins get a PluginBarStateApi as shell.bar, which now carries barMoveActive, bound to the active bar in shell.qml the same way as barHidden. I checked that by reading the code, not by running a cloned plugin.

At a 0.4 s hold, 6 of 8 moves worked. Both misses were top→bottom, released at y=15 just as the pointer left the strip. That's the race the Known limitation section describes, and the description states it accurately.

Codex Medium read the whole diff at this head and found no defects. It judged the change proportionate to the bug and agreed the Known limitation wording matches the code. It can read this session, so its agreement isn't guaranteed to be independent.

This now waits on the maintainer.

@omarchybot omarchybot removed ready Good to merge verified Omarchy Triage has verified that this issue is ready for final review labels Oct 22, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working

Projects

None yet

4 participants