[vjeux]

Maintainer impact

Three busy-proof reads decode one screencast more than once: once per control with trackControl, and once more for the marker with trackMarker.
On a loaded runner each decode pass over a cast takes seconds.
Since #10808, one trackMarkers call can read each control in its own strip.

Intended invariant

Each of these three reads decodes its frames once, however many colours it finds.
What each proof reads and asserts is unchanged.
P14's whole-cast marker read, the colour proof's brightness read and the toggle's reads still decode on their own.

Change

• Switch's thumb clock, which two proofs share, reads both controls and the marker with one trackMarkers call.
• Switch's track colour proof reads both controls with one call.
• Tooltip's closed-trigger anchor proof reads the marker and the control with one call.

Evidence

• Failure before the change: no timeout; the cost is time. On the same cast in Chromium, the separate reads took a median of 1.7 s, 0.6 s and 0.9 s.
• Success after the change: one read took 1.0 s, 0.3 s and 0.7 s, and found the same tracks, frame for frame.
• Product/runtime behavior verified unchanged: test only.
• CI impact (for a change to workflows, actions or CI scripts): none.

Scope

Components: Switch, Tooltip (interaction specs only)

• No intended public API, product behavior, visual direction, or policy change.
• Product changes discovered during the work were removed or split.
• Public text and artifacts contain no internal Meta context.

Testing

All figures are at the CI profile with one worker, on this head, 10f7314.

Same cast, median, the separate reads against one decode (d221877, not for landing):

Read	Engine	Runs	Separate	One decode
Switch thumb clock	Chromium	8	1712 ms	1031 ms
Switch track colour	Chromium	8	561 ms	305 ms
Switch track colour	WebKit	8	1528 ms	782 ms
Tooltip anchor	Chromium	4	858 ms	682 ms
Tooltip anchor	WebKit	4	1718 ms	1201 ms

Each run times both forms on their own copies of the frames, in alternating order. One decode was faster in all 32 runs.

Identity commit, not for landing: 5603d7f.
Each converted read also runs the separate calls on copies of the frames and asserts equal tracks.
It passes 12 of 12 runs: 8 in Chromium and 4 in WebKit.

The four proofs at this head pass 12 of 12 runs in Chromium and 6 of 6 in WebKit.
The two Switch thumb-clock proofs run in Chromium only.

pnpm --filter storybook build
node_modules/.bin/playwright test --config playwright.interaction.config.ts --workers=1 \
  packages/core/src/Switch/__tests__/Switch.interaction.browser.spec.ts \
  packages/core/src/Tooltip/__tests__/TooltipClosedTriggerAnchor.interaction.browser.spec.ts

Prettier, ESLint (no new warnings), the core and harness type checks and review:pre are clean.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
astryx	Ready	Preview	Nov 10, 2026 7:47pm UTC

[astracat-bot]

AI review status for this pull request.

Review status	Updated
🟢 Reviewed (for maintainers only)	Nov 10, 2026, 7:51 PM UTC

[github-actions]

PR Analysis Report

🔎 Preview · 10f7314

Storybook · Sandbox · built from this PR's head

Modified components

Component	Size	LoC	Visual · A11y · RTL
Switch · Storybook	104.7 kB (0) · 9.9 kB CSS (0)	1,734 (0)	✓ · ✓ · ✓
Tooltip · Storybook	30.4 kB (0) · 4.4 kB CSS (0)	870 (0)	✓ · ✓ · ✓

Package size: no published package changed; no new dependencies.

API changes: none in this pull request

Committed API report behind this build: cli 36 rows, core 36 rows; run pnpm api:report.

Touched components checked; add ci:run-visual, ci:run-a11y or ci:run-rtl to run the full suites.

---

_{Generated by PR Enrichment workflow | View full report}

[astracat-bot]

Changes requested: the combined reads look right, and one wording fix is needed before approval.

• The title and the stated invariant say Switch and Tooltip busy proofs decode each cast once, but three Switch busy proofs still decode frames twice (P14's whole-cast trackMarker read at Switch.interaction.browser.spec.ts:2578 before the thumb clock at :2605, the track colour proof's trackBrightness read at :3316, and the P10 toggle's sub-pixel and control reads at :2990 and :2998), so please retitle to the three reads this changes and scope the invariant to them.

Each combined trackMarkers call returns the same tracks the separate reads did, and the type checks, ESLint and Prettier are clean on the changed lines. CI has not run on this draft head yet.

[Automated review]

[astracat-bot]

Approved pending CI: CI has not run on this head yet. It runs once the pull request is marked ready for review and a maintainer adds the ci:run label.

The new title and intended invariant match the change: each of the three converted reads now decodes its frames once, and the description names the reads that still decode on their own. Each combined trackMarkers call returns the same tracks as the separate trackControl and trackMarker calls it replaces, and the type checks, ESLint and Prettier are clean on this head.

[Automated review]