fix(lab): SearchField's tap claim takes a pen on touch screens and lets go after the click guard by vjeux · Pull Request #8689 · meta/astryx

vjeux · 2026-10-27T12:42:54Z

Depends on #8224, #8665 and #8666. Review only this PR's change: vjeux/astryx@4fd7d5b...2370b6c

User impact

On iPhone and iPad, SearchField's bottom and tab placements claim a tap on the field's text so Safari doesn't pan the page (spec:AST-099 FR4). Now the press after a claimed tap loses its click, and an Apple Pencil's tap isn't claimed.

Problem and solution fit

What led to it: the UIKit-parity program moved gestures to every pointer (#8665, spec:AST-094 FR43). The field swallows a claimed tap's late mouse events until their click, but #8665's click guard now takes that click first, so the swallow never ends. And the claim took only a finger.

Options:

Rely on the guard alone: without it, late events would focus natively and Safari would pan.
End the swallow on a timer: a slow tap leaks.
Claim every pointer: a mouse click would lose its click and caret (DEC-21), and a desktop pen tablet's click would be swallowed (spec:AST-017 FR59).

Chosen: the swallow ends once the guard has the click (else at the click); the claim takes a finger and a pen, only under a coarse primary pointer.

Expected behavior and authority

spec:AST-099 FR4; spec:AST-094 FR43 and DEC-21; spec:AST-017 FR59. Owner question: FR43 names no mouse exception for a tap; may a recognizer serving only the touch keyboard's reveal stay off for a mouse?

Smallest restoration

A swallow condition, the pen, a coarse-pointer gate, and an isClickGuardArmed re-export in internal/touch.

Size (gzip, base → head): SearchField JS 60.35 → 60.38 KB, CSS 1.54 → 1.54 KB.

Evidence

Reproduction before the change: Testing 1 to 3 failed.
Result after the change: they pass.
Representative unchanged path: the stacked field; mouse and trackpad.
Regression test or other mutation-sensitive proof: undoing each part fails its own browser test.

On iPad #8665 cancels a claimed pen's touchend (real-device check owed). The ci-duration red here was a GitHub 502, fixed by #8744.

Scope

One defect is restored.
Separable contradictory or unsettled tagalongs were removed or split.
No new public API, visual representation, default, or interaction model is hidden under the bug-fix label.
Public text and artifacts contain no internal Meta context.

Testing

What I ran and what it showed, in Lab/SearchField's bottom story:

Chromium and WebKit, touch: two taps on the text; the second tap's click was eaten before and lands now.
Chromium, test(interaction-harness): wheel and pen input drivers for every-pointer proofs #8666's pen driver: a pen tap is claimed as a finger's (WebKit has no pen).
Chromium as a desktop: no claim; mouse and pen keep the browser's focus and click.
SearchField's browser suite passes in both engines.

…tory's host on public APIs) into uikit/c09-search-bottom

…uikit/c09-search-tab # Conflicts: # apps/storybook/stories/SearchField.stories.tsx

…of the document) into meta#8270

…sture system (spec:AST-094 FR24, FR43) utils/pressFeedback.ts reaches gestures/coarsePointer.ts, so a page whose only touch work is press feedback carries a gesture module, and a bundle check that counts gestures/ code finds it in the root and hooks entries. The coarse-pointer query belongs with the environment queries. This test walks the controller's import graph and fails until it reaches no module under gestures/.

…ries in utils/ (spec:AST-094 FR43) gestures/coarsePointer.ts becomes utils/coarsePointer.ts, beside utils/touchQueries.ts, with its test; the gesture host's selection lock, the press controller and the pointer test helper import it from there. The press controller's import graph now reaches no module under gestures/, so a page whose only touch work is press feedback carries none of the gesture system. No behavior changes.

…/) into meta#8309

…/) into meta#8315

CI's lint rejected the three Array<T> types the previous commit added (@typescript-eslint/array-type); the test now names its record once (WatchTransition) and reuses it.

…e arena's reset The arena resets a recognizer to possible after it ends, as UIKit does, and the transition listener records that too, so the last transition is possible: the test now asserts the ended-then-possible tail.

…/) into meta#8336

…/) into meta#8320

…finger, meta#8589's watcher test, its lint fix; meta#8076 3bfd7c3; the coarse-pointer query in utils/) into meta#8293

…nter query in utils/) into meta#8358

…nter query in utils/) into meta#8356 # Conflicts: # packages/core/src/gestures/GestureHost.ts # packages/core/src/gestures/fingerScroll.ts # packages/core/src/gestures/pageHold.ts # packages/core/src/gestures/selectionLock.ts # packages/core/src/utils/pressFeedback.ts

…nter query in utils/) into meta#8360

… FR19, FR24) Under a fine primary pointer: each element that uses a gesture keeps one passive pointerdown listener and the page none while idle; a mouse or a pen pans, its page listeners lasting only the press; a press consumer takes a mouse press with no feedback-only mode (DEC-21); a pan writes its touch-action for a touchscreen laptop's finger (FR18); perEventPointer is taken as any attach, with no warning; a secondary button starts nothing. All but the last fail until the host attaches for every pointer.

… the scroll brake for a finger only (spec:AST-094 FR11, DEC-21) A mouse or a pen press paints at once, inside a scroller too; a mouse press on a scroller a flick set moving still presses, as the brake is a finger's. The brake case fails until the tracker brakes only a touch.

… (spec:AST-094 FR43) A drag from the edge band begins a screen-edge pan for a finger only, unless the caller allows another pointer. The mouse and pen case fails until the recognizer defaults to touches.

… FR19, FR24, DEC-21) Recognizers and press consumers attach for a mouse, a pen and a finger alike. Where the primary pointer is fine, each element with an enabled recognizer or consumer keeps one passive pointerdown listener and the page has none: a press on it (a mouse's primary button, a pen, or a finger) installs the host's page listeners until that press's contacts end. The action group's own fine-pointer path folds into this one, so a group is served like every attachment. - touch-action is written for every pointer, a kept press's manipulation too (FR18); the element's blocking touchmove stays a coarse pointer's, so under a fine one the element keeps its one listener (FR24). - A press consumer takes a mouse press as a press (DEC-21): no feedback-only mode; PressTracker brakes only a finger's press (FR11). - The screen-edge pan allows touches only unless its caller allows another pointer (FR43). - perEventPointer is accepted and does nothing; the fine-pointer warning goes. Tests that encoded the coarse-only design now state FR43's; a finger's pointerdown comes before its touchstart in them, as in a browser.

…th real mouse drags (spec:AST-094 FR43, FR24) Under a desktop's mouse the host still writes touch-action, leaves the page no listener while idle and warns for nothing; a real mouse drag pans both the bare page's pan and the fixture's horizontal pan, and the page's touch listeners are gone again after the release.

…(spec:AST-094 FR18, DEC-13) A screen-edge pan attached to a stack container, as wide as the screen, with a touch 4 px from its edge: inside the 40.2 px band the pan decides at UIKit's 15 px whatever the element's width, and then tracks the travel past it. From several edges too. Both fail while a wider element decides at FR19's 10 px.

…'s width (spec:AST-094 FR18, DEC-13) FR18: inside the 40.2 px band a screen-edge pan decides at UIKit's 15 px screen-edge hysteresis; outside it the author's scrolling is untouched and FR19's 10 px governs every other claim. The pan decided at 10 px on any element wider than the band, so NavigationStack's edge pan, attached to the stack container, began 5 px early. The edge zone still decides only the touch-action; a touch outside every band still fails at once. The two cases that pinned the 10 px decision go.

…px whatever the element's width) into the experimental touch exports

… move keeps that finger until it lifts) into the experimental touch exports

…than the band (spec:AST-094 FR18, DEC-22) FR18 as amended in meta#8631 (DEC-22): the band's pan-x claim and 15 px start hold only where the screen-edge pan is attached to an element within the band; on an element that reaches beyond it, whose touch-action cannot claim per touch, a touch in the band decides at FR19's 10 px like every other claim. This restores the 10 px decision 3d85c30 removed and keeps 15 px on an element within the band. The cases pin DEC-22 both ways (a wide element with a touch at x=4 begins at 10, an in-band strip at 15, and from several edges at 10 on a wide element); the 10 px cases fail on 3d85c30.

…he edge band decides at 10 px again, DEC-22) into the experimental touch exports

…ouch decides at 15 px on an element within the band and at 10 px beyond it, DEC-22; meta#8641 bec2283) into meta#8270

…eta#8309

…eta#8216

…eta#8224

…age's selection The drag selection hold cleared the page's selection when any mouse or pen contact was claimed, but an action group's press claims at its pointerdown, before it has moved or selected anything, so a plain click on a group's button left its Copy or Cut with nothing to act on. A claim before any movement now keeps the page's selection: the hold still refuses a new one, and puts the page's back over a selection that appears while it runs. A claim after movement clears what its drag selected, as before (spec:AST-094 FR41). The case before the new one now waits for its hold to end a task after the lift, so no hold outlives that test.

…t every pointer does - GestureHost's header, the gestures README and coarsePointer.ts no longer describe the coarse-only host: a finger or a pen takes the touch-down lock under any primary pointer, a kept press's finger is kept during a fine pointer's press too, and recognizers attach for every pointer while the page listeners last a press (spec:AST-094 FR19, FR24, FR41, FR43). - The fine-pointer selection case asserted the replaced rule and passed only because it read iOS WebKit's inline lock without iOS WebKit: it now asserts the lock a fine pointer's finger and pen take, and that a mouse takes none. - FR41's sentence and its verification row say that a press claimed at its own pointerdown, before it moves, keeps the page's selection, while a drag that moved before its claim loses what it selected.

…ion, in the present tense The sentence said a drag that moved before its claim loses the selection it started, but the hold clears the page's selection then, whatever made it; and an unclaimed mouse drag kept "today's" behavior, a phrase that ages. It now says a claim after movement clears the page's selection, a press claimed at its own pointerdown keeps it, and an unclaimed mouse drag keeps its released behavior.

…6c02fc, the harness README, spec and index.ts as in 801dfa6750 (meta#8666 on meta#8360)

…task ended at once, however late the first read WebKit under load from other tasks: the call's task ends at once, another task holds the main thread, and the browser commits the animation after it, dated at the call. Nothing in the call's task delayed it, but the first read still moved its start to the read, about 95 ms late on a loaded box, so whatever chains on its end (the second half of a Reduce Motion cross-fade, a cancel fading on the previous fade's end) ran that much late and the cross-fade missed design:reduced-motion DR2's 200 ms. Fails now: the start moves to 1100.

e7b0676 moved a call's start from rest to its first read whenever WebKit had dated it more than two frames before that read. Under load from other tasks the first read comes late with nothing in the call's task delaying it, so the start moved about 95 ms on a loaded box, and whatever chained on the call's end ran that late: the second half of a Reduce Motion cross-fade, so SearchField's cancel missed DR2's 200 ms. The call's task now records when it ended, at its first microtask (or a read in those microtasks is that time), and only a call whose own task ran past the call by more than two frames, the block e7b0676 fixed, moves its start; a read other tasks delay keeps the browser's start.

…s its start past the block Core Animation dates an animation at the commit, when the calling task ends, however short the block before it. A 20 ms block in the call's task, with WebKit's start dated at the call, left the start at the call: the two-frame bound skipped it. Fails now.

… task, with no threshold Core Animation dates an animation at the commit, when the calling task ends, whatever the block's length. The call's task already records when it ended, at its first microtask, so a start the browser dated before that (WebKit's last frame time before the task) now moves to it: a task that ended at once leaves the start at the call, a blocked one moves it past the block, however short, and a read other tasks delay changes nothing. The two-frame bound, which told a read's own frame from a block before the task's end was known, goes, and the start no longer moves to the read.

…'s click guard has taken the click The claimed tap swallows its own late mouse events until their click. With meta#8360, the gesture host arms its click guard at the claimed lift, and the guard takes that click first, on the window's capture phase, so the swallow never saw it and stayed armed: the next press's mousedown, mouseup and click never reached the page. The swallow now notes the guard while it is armed and lets go once the guard has (its click, the next press, or its window); with no guard it ends at the click, as before. `internal/touch` re-exports `isClickGuardArmed` by name. Red-first: "a claimed tap swallows its own late mouse events, until the next touch" fails on the previous head (the second press's events are swallowed) and passes in Chromium and WebKit with this change.

…ick guard) into meta#8224

…eta#8666 08a2a28) into the web-input drivers helper The harness files are 8d35612's verified union, byte for byte, plus meta#8535's change to the backward self-test, which this line carries; the two helpers hold the same content apart from that change. This line's own harness additions leave it, as no open pull request builds on them: the earlier busy-window control and pooled proof (now meta#8906), openStory waiting for Storybook to finish a render, the touch driver's synthesizesClicks and SCRIPTED_CLICK_SKIP, and the wheel-at-DPR-3 self-test.

… own task's end) into meta#8181

…ta#8666 08a2a28) into meta#8224 1425523, as meta#8689's base

…c0e0a5) into meta#8689

…few ms, and one task's calls start together Browser start times are frame times, while a task's end is a live clock, so on every call they differ by up to a frame plus the task's own length. Moving every start from rest to its task's end shifted each one by that length, about 5 ms in a component's effect after a tap, and the shift added up along a chained sequence: SearchField's Reduce Motion cancel measured 204-207 ms against design:reduced-motion DR2's 200. The first case fails now (the start moves to 1010). The second keeps one motion's parts together when they are separate calls in one task: a block measured per call would split them by a frame.

…n on more than two frames, and all the task's calls move together fc42d64 moved every start from rest to the end of the call's task. Browser start times are frame times while that end is a live clock, so on every call they differ by up to a frame plus the task's own length: each start slid by that length, and a chained sequence added it up, about 5 ms per link in SearchField. The rule from e6876ca is back, with its bound now named for that quantization: only a task that ran on more than LATE_START_MS (two frames) past its first call blocked anything, and its calls' starts move to their first read. The block is now measured per task, not per call, and every call of a blocked task takes the same read, so one motion's parts that are separate calls in one task (a component's effect) start together instead of splitting by a frame. The 20 ms case keeps the browser's start, as no clock here tells it from quantization.

…art moves only past a real block) into meta#8181 [skip ci]

…harness branch 4c0e0a5)

…c0e0a5) into meta#8689 [skip ci]

astracat-bot

Approval waits on two spec decisions: spec:AST-094 FR43 names no exception for this tap claim staying off for a mouse and attaching only under a coarse primary pointer, and the FR41 claimed-drag selection rule carried from #8665 is new text that main does not state, so an engineering owner needs to settle each, on main or by commenting /approve-spec <full head SHA> here (the author may do so if they are an engineering owner). Whichever way those go, two changes are still needed:

Land #8224, #8665, #8666 and the pull requests beneath them first and rebase, so this diff is the 207-line, two-file fix instead of about 37,000 lines across 216 files.
Add Safari-on-iOS evidence (Simulator or device) that a claimed Pencil tap focuses the bottom field without the page panning, as spec:AST-099 asks of its iOS claims; the Chromium pen proof cannot show it, because Chromium sends the pen's compatibility mousedown before the tap is recognized.

The title and description still describe the click-guard swallow fix, which now arrives with #8216, so they should be updated to the pen and the coarse-pointer gate.

[Automated review]

astracat-bot · 2026-10-30T03:13:38Z

+  (`spec:AST-017` FR59; DEC-9). A mouse or pen drag that a recognizer claims
+  takes no new text selection for that drag only. A claim made after the
+  drag has moved clears the page's selection; a press claimed at its own
+  pointerdown, before it moves, keeps it. An unclaimed mouse drag keeps its
+  released behavior.


This adds a requirement main does not state: FR41 there keeps a mouse at its released behavior, and its verification row fails any change to mouse selection. It arrives with #8665 and needs an engineering owner's decision, by /approve-spec on this head or by landing it on main with #8665.

vjeux and others added 30 commits October 26, 2026 22:09

Merge meta#8181 (f829200: meta#8562 abb3e0f, meta#8076 3bfd7c3, the s…

5bc6c48

…tory's host on public APIs) into uikit/c09-search-bottom

Merge meta#8216 (5bc6c48: meta#8562 abb3e0f, meta#8076 3bfd7c3) into …

aa3f4a9

…uikit/c09-search-tab # Conflicts: # apps/storybook/stories/SearchField.stories.tsx

Merge facebook#8076 (3bfd7c3: the animator reads no style for a node …

2119cfa

…of the document) into meta#8270

Merge meta#8270 (meta#8076 3bfd7c3; the coarse-pointer query in utils…

9db6af5

…/) into meta#8309

Merge meta#8309 (meta#8076 3bfd7c3; the coarse-pointer query in utils…

999afb7

…/) into meta#8315

test(gestures): the watcher's browser test names its record type (lint)

b57ac07

CI's lint rejected the three Array<T> types the previous commit added (@typescript-eslint/array-type); the test now names its record once (WatchTransition) and reuses it.

Merge meta#8589 (the watcher browser test's lint fix) into the host line

6cca0de

Merge meta#8336 (meta#8589's lint fix) into the arena fixes

ea8531c

Merge meta#8315 (meta#8076 3bfd7c3; the coarse-pointer query in utils…

790c65e

…/) into meta#8336

Merge meta#8336 (meta#8076 3bfd7c3; the coarse-pointer query in utils…

2fd9697

…/) into meta#8320

Merge meta#8320 (cixzhang's ea8531c: the scroll-edge hold for a lone …

45b25e4

…finger, meta#8589's watcher test, its lint fix; meta#8076 3bfd7c3; the coarse-pointer query in utils/) into meta#8293

Merge meta#8356 (meta#8320 ea8531c; meta#8076 3bfd7c3; the coarse-poi…

e95a4e2

…nter query in utils/) into meta#8358

Merge meta#8358 (meta#8320 ea8531c; meta#8076 3bfd7c3; the coarse-poi…

741787a

…nter query in utils/) into meta#8360

test(gestures): the screen-edge pan stays off under a mouse and a pen…

f9a93c3

… (spec:AST-094 FR43) A drag from the edge band begins a screen-edge pan for a finger only, unless the caller allows another pointer. The mouse and pen case fails until the recognizer defaults to touches.

Merge meta#8061 (3d85c30: an in-band screen-edge touch decides at 15 …

9d0f6f2

…px whatever the element's width) into the experimental touch exports

Merge meta#8641 (bec2283: a gesture that cancels itself on a finger's…

11dbebf

… move keeps that finger until it lifts) into the experimental touch exports

Merge meta#8061 (d9a1e1a: an in-band touch on an element wider than t…

e6ab0b9

…he edge band decides at 10 px again, DEC-22) into the experimental touch exports

Merge meta#8076 (e6ab0b9: meta#8061 d9a1e1a, an in-band screen-edge t…

e7b5b3c

…ouch decides at 15 px on an element within the band and at 10 px beyond it, DEC-22; meta#8641 bec2283) into meta#8270

Merge meta#8270 (meta#8076 e6ab0b9: FR18 per DEC-22, meta#8641) into m…

e718566

…eta#8309

vjeux added 13 commits October 29, 2026 12:29

Merge meta#8181 (f4d4325: meta#8360 045f666, meta#8340 339a352) into m…

c43f3d6

…eta#8216

Merge meta#8216 (c43f3d6: meta#8360 045f666, meta#8340 339a352) into m…

32b0b78

…eta#8224

Merge meta#8666 08a2a28 (wheel and pen input drivers) into meta#8665 c…

8d35612

…6c02fc, the harness README, spec and index.ts as in 801dfa6750 (meta#8666 on meta#8360)

Merge meta#8216 (d0bca4a: the late-event swallow lets go after the cl…

5693872

…ick guard) into meta#8224

vjeux marked this pull request as draft October 29, 2026 22:39

vjeux added 12 commits October 29, 2026 16:10

Merge meta#8340 (fc42d64: a start from rest is never earlier than its…

1c4a7ab

… own task's end) into meta#8181

Merge meta#8181 (1c4a7ab: meta#8340 fc42d64) into meta#8216

5b18722

Merge meta#8216 (5b18722: meta#8340 fc42d64) into meta#8224

1425523

Merge the input-driver harness branch (4c0e0a5: meta#8665 c6c02fc, me…

8a5649e

…ta#8666 08a2a28) into meta#8224 1425523, as meta#8689's base

Merge 8a5649e (meta#8224 1425523 with the input-driver harness branch 4…

9c77cf5

…c0e0a5) into meta#8689

Merge meta#8340 (26ecfe7: calls made in one task start together; a st…

ee835dc

…art moves only past a real block) into meta#8181 [skip ci]

Merge meta#8181 (ee835dc: meta#8340 26ecfe7) into meta#8216 [skip ci]

fc37107

Merge meta#8216 (fc37107: meta#8340 26ecfe7) into meta#8224 [skip ci]

15991be

Merge meta#8224 15991be into meta#8689's base (with the input-driver …

f8c46ee

…harness branch 4c0e0a5)

Merge f8c46ee (meta#8224 15991be with the input-driver harness branch 4…

e72a432

…c0e0a5) into meta#8689 [skip ci]

github-actions Bot added the needs:spec-owner-review Current knowledge records await owner approval label Oct 30, 2026

vercel Bot had a problem deploying to Preview October 30, 2026 02:50 Failure

astracat-bot Bot reviewed Oct 30, 2026

View reviewed changes

vjeux mentioned this pull request Nov 1, 2026

feat(lab): SearchField, iOS 26's search field: stacked, bottom and tab placements #9399

Open

4 tasks done

Navigation Menu

Repository navigation

fix(lab): SearchField's tap claim takes a pen on touch screens and lets go after the click guard - #8689

fix(lab): SearchField's tap claim takes a pen on touch screens and lets go after the click guard#8689
vjeux wants to merge 1999 commits into
meta:mainfrom
vjeux:uikit/c09-search-fr43

vjeux commented Oct 27, 2026 •

edited

Loading

Uh oh!

astracat-bot Bot left a comment

Uh oh!

astracat-bot Bot Oct 30, 2026

Uh oh!

This branch had an error being deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

vjeux commented Oct 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

User impact

Problem and solution fit

Expected behavior and authority

Smallest restoration

Evidence

Scope

Testing

Uh oh!

astracat-bot Bot left a comment

Choose a reason for hiding this comment

Uh oh!

astracat-bot Bot Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

This branch had an error being deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

vjeux commented Oct 27, 2026 •

edited

Loading