On a slow machine Escape could close the full-screen search and have it
open again for good. A close hands focus back to the input — Escape does,
and so does the view's focus trap as it lets go — and `focused()` told
that return from someone coming to search by a 250ms wall-clock window.
Both returns run on frames, and a runner painting a few frames a second
took longer than that, so the returning focus opened the view again. A
`returning` flag now covers the close until the hand-back has actually
run, on the same frame, and the constant is gone.
`hold()` times the full-screen layout from the exit's own duration token,
the one Alpine's x-transition holds `display` for, instead of waiting a
frame or two for the exit's transitions to appear: `getAnimations()` is
empty both before an engine creates them and after they end, and a loaded
engine can leave a second between frames.
The browser tests were racing the same slow machine, reproduced in a
Linux container like the runner, with its Playwright Firefox and
WebKitGTK and two cores kept busy:
- The browser plugin retries every script and action with a one-second
attempt until the budget runs out. An in-page sleep longer than that
only ever passed on the last attempt, which is where the 47-50s tests
came from, and a script that clicks was run again against a page that
had moved on. `onceInPage()` runs such a script once however often it
is retried; the long sleeps are plain retried conditions now.
- Under load WebKitGTK paints no frame while a tight setTimeout loop
runs, so a sample loop saw the start value and then nothing.
`caughtMidExit()` samples on animation frames, for 2.5s.
- "No animations running" is also true before an opening transition
exists, so a close could be sampled from a scrim at 4% opacity.
`settled()` waits two frames before it asks.
The workflow no longer uploads failure screenshots: Gitea's artifact
service timed out on every attempt, two minutes per red run, and the job
logs are readable without it.
Feature 1159 passed. Browser 299 passed on Chrome, Firefox and WebKit on
macOS, and on Firefox and WebKitGTK in the Linux container under load,
twice each.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The browser suite passes on every engine here and fails on the runner,
which takes three times as long: each failing test triggers a close and
then samples for an in-between state — a rail part-way out, a scrim
part-way faded, a menu's exit copy part-way sunk — and a starved runner
takes its one sample after the 150-650ms exit has already finished.
So the tests that assert *that* something animates now stretch every
motion duration token to three seconds first (`slowMotion()`, beside
`ready()` in tests/Pest.php; ActionsTest's own copy of it goes). The two
polling helpers grew their budget to match: the rail panel and the sheet
slide on emphasized-accelerate, which is under 1% of its travel at a
quarter of the way through, so a 300ms window no longer reached the
threshold once the exit itself was three seconds long.
The two full-screen date picker tests waited for a resize through
click()'s own retry, which ate the whole 15s budget on Firefox; they now
wait for the new width and a settled document first. The bottom sheet's
preset test waits for its entry to finish before pressing the grip.
`hold()` in search.js reads the view's animations a frame after the
closed state, as the rail's settle() does, but one frame is not always
enough: an engine that starts them on its next tick shows none, and the
full-screen layout would end at once, mid-exit. An empty list is now
asked again on the following frame.
Feature 1159 passed. Browser 299 passed on Chrome, Firefox and WebKit,
and again on WebKit under ten spinning cores.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
An over-engineering audit of the whole tree, applied in five reviewed
batches. Behaviour stays the same except where UPGRADE.md says otherwise.
PHP: the showcase and error-page stylesheets are prebuilt into
resources/dist by bin/stylesheets.mjs, through Vite's own postcss-import
(first occurrence kept, the order an application's build gives), instead
of Stylesheets::bundle() inlining imports on every request; only the
import walk DesignGuard needs stays. SchemeStylesheet::withProfiles()
replaces three copies of the scheme-plus-profiles loop, material:scheme
leaves spec and contrast checks to the node script that already made
them, and the error page's scheme cache, the hashed view namespace, the
translations path with no lang/ folder and DesignGuard's 1.x-name hints
are gone.
JS: the androidx shape port progress.js and both bin scripts each carried
lives once in resources/js/shapes.js (the generated SVGs are unchanged);
util.js holds ringIndex(), ms(), reopenGuard() and remember(), which
were written out several times; listeners are released through
AbortController; tooltip.js's hoverPopover() serves the rich tooltip too.
CSS: every rule for an element inside the navigation rail queries
`--md-navigation-rail-value` instead of repeating the seven collapsed
conditions under five media branches; badge, alert, progress, slider and
button read one non-inheriting colour-role table (components/color.css);
the dialog chrome, the submenu's popover chrome, the chip's state layer
and touch target, and the visually-hidden inputs use the shared rules
they copied; foundation/tokens.css is folded into foundation.css.
Views: Support\Field and Support\Link replace the error-key, bound-value
and link-attribute blocks copied into the fields and link components;
the timepicker period group, the menu filter and the showcase head are
partials; the datepicker's steppers and entry fields are loops; component
docblocks no longer restate SKILL.md.
Tests and tooling: one dataset-driven ComponentStylesheetsTest replaces
four per-group files, DesignGuardTest and the layout-component tests use
datasets, browser tests share one ready() helper, CSS parsing lives in
ComponentStylesheet alone. docs/audits and the finding IDs citing it are
removed, as are pestphp/pest-plugin-laravel, the unused composer scripts
and check:font; the lint job runs in the feature job, which now installs
node packages so the prebuilt-stylesheet staleness test runs in CI.
Feature suite 1177 passed, Chrome browser suite 299 passed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`fullScreen` followed `open`, so the moment a full-screen search closed,
`data-md-full-screen` went with it: the fixed header bar went back to its
resting pill — or to `display: none`, behind the search icon — and the
view, still fading out under Alpine's hold, dropped to the docked layout
for its exit. In every engine the bar vanished on the first frame while
the view went on fading somewhere else.
search.js: a close from full screen now `hold()`s the layout — `leaving`
keeps `fullScreen`, and with it `data-md-full-screen`, the back arrow and
the fixed bar, for the view's own closing duration, read from its
computed style a frame on (zero under reduced motion); reopening lets a
pending end go by. search.css fades the header bar out with the view on
the view's spring and keeps the root above the page while it leaves,
and the icon trigger's bar stays displayed until the layout settles. The
focus trap now binds to `open && fullScreen`, so it still lets go at the
close and its return of focus lands inside RETURN_GUARD_MS rather than
reopening the view at the end of the exit.
PickingTest samples both exits in the page — the icon trigger, and the
bar trigger on a compact window — for a moment where the root is still
full screen with a fixed bar and both bar and view part-way through their
fade, then checks the settled layout and focus back on the icon or the
field; both fail on the previous code in Chrome, Firefox and Safari. The
tests wait for the entry's own transitions first: Firefox reads a
transition's end value until its next refresh tick.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A sheet-at-compact menu focused its first item, or its filter field, in
$nextTick, and the icon entry point of <x-search> handed focus back to
its trigger the same way. Both targets are x-show'n, and Alpine shows an
element on the animation frame after its condition turns true, not in the
tick: the focus reached an element still display: none, which Firefox and
WebKit refuse. The sheet's focus trap then put the focus on the drag
handle, and a closed search left it on the page body; in Chrome the
frame or the trap's own return happened to cover for it. The comment in
openSheet() assumed Alpine holds $nextTick until the transition begins,
which it only does with x-transition, and the sheet uses CSS transitions.
Both now wait for that frame after the tick, running after Alpine's own
show in it, as search.js's expand() already did.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Plan step 36 (containment group): <x-card>'s class lists move into
resources/css/components/card.css, keyed on data-md-card (per variant),
data-md-card-figure/-body/-header/-heading/-title/-subtitle/-menu/
-content/-actions. A row (data-md-list-row) renders the shared
md-state-layer and md-focus-ring classes for its tint and ring
(foundation/interaction.css) instead of copying their rules; card.css
adds only the per-state elevation the shared class has no opinion on
(ElevatedCardTokens.kt/FilledCardTokens.kt/OutlinedCardTokens.kt), the
press/hover exclusion for a card's own nested buttons, and the ring for
a non-actionable row whose focus lands on its opener rather than the
card. data-md-dragged is read automatically by the shared class's own
16% tint; card.css only adds its elevation levels 3/4.
Hooks renamed data-list-row -> data-md-list-row, data-list-open ->
data-md-list-open, data-list-actionable -> data-md-list-actionable,
data-dragged -> data-md-dragged, updated in the same commit:
resources/js/search.js and list-rows.js's consumers, table.css's
comment, the showcase (index, data and containment sections),
tests/Browser/ContainmentTest.php, and the development skill's card
and table examples.
Imported from the Containment block of components.css.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qwx5USif3wFFmxtHg5U1g9
<x-search> renders data-md-search with data-md-trigger and binds
data-md-open and data-md-full-screen; its parts become data-md-search-*
and the empty row data-md-search-empty. search.css moves into
material.components on px, spacing and state tokens (plan step 36).
search.js and the showcase layout follow the renamed hooks.
Browser tests cover the polite result count as it changes, the icon
entry point expanding full screen and returning focus, and suggestions
swapping to results on the first key.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qwx5USif3wFFmxtHg5U1g9
M3 § Search names three entry points and only the bar existed; suggestions
before the first keystroke were missing too (plan step 24, audit
docs/audits/m3-alignment/inputs.md § Missing). `trigger="icon"` is the search
icon button — search as a secondary action, one 48px button that expands into
the full-screen view at any width, since an icon button has nothing to dock
under, and takes its focus back on close. The `suggestions` slot stands where
the results do until something is typed; the live region counts whichever of
the two is on screen and names suggestions as suggestions.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qwx5USif3wFFmxtHg5U1g9
M3's search accessibility page asks that results be announced when they
appear and read as a list. The bar now wraps its input in a
role="combobox" that carries aria-expanded and aria-controls — ARIA gives
a bare textbox neither — the results container is a role="list", and a
polite live region counts the results whenever the view's DOM settles.
Plan step 13, finding IN-01.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qwx5USif3wFFmxtHg5U1g9
Plan step 16. `navigation.js` takes its standard-rail threshold from
`from('expanded')` (840px) instead of the 64rem it had, and `navigation.css`'s
three 64rem queries follow, so the whole expanded class gets the in-layout
collapsible rail M3 asks for rather than one that opens over a scrim (N-06);
its two 40rem queries become 37.5rem, the compact/medium boundary the bar's own
container query already used (N-07). `search.js` and `datepicker.js` swap their
39.99rem media strings for `upTo('medium')`, so the full-screen search view and
the modal date picker end at 600px, not 640px.
The time picker's landscape layout stops keying on width at all: M3 swaps it on
orientation and viewport height, so it is now landscape plus a window too short
for the upright dial (35rem, the dialog's own height), and the two dial-shrink
rules key on height alone.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qwx5USif3wFFmxtHg5U1g9
M3's assist, filter, input and suggestion chips with chip sets; choices
as filter chips or a searchable combobox whose list is an anchored
popover; and the search bar that opens into a docked or full-screen
search view.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V9NnLxnPp8vaaurb3Z1MFy
