refactor(ios): gesture-response factory + performGesture wrapper

[thymikee]

What

Collapse the repeated gesture-command boilerplate in CommandExecution.swift's executeOnMain switch behind two small seams:

• performGesture(app, idleTimeout:) → (timing, outcome) — folds measureGesture + the scroll idle-timeout/quiescence-skip wrap. Parameterized: touch gestures wrap (default); synthesis gestures (pinch/rotate/transform) pass idleTimeout: false because RunnerSynthesizedGesture governs its own timing (this distinction was previously implicit/easy to get wrong).
• gestureResponse(message:timing:frame:) over a GestureFrame (none/touch/drag) — one factory for the success DataPayload (message + gesture timing + optional visualization frame).

13 gesture cases now use the seams; ~120 lines of the repeated quartet (measureGesture + idle-wrap + unsupportedResponse + 8-field DataPayload) collapse. CommandExecution.swift 949 → 867.

Why

Each new gesture command re-derived the same 8-field payload + timing/idle quartet by hand — drift-prone and ~9× of the file's bulk. One factory + one wrapper make the switch far cheaper to read and make the timing/idle distinction explicit.

Behavior-preserving

Identical outcomes, timing capture, DataPayload fields/messages, and the same touch-wrapped vs synthesis-unwrapped distinction. No wire change — the factory emits the already wire-locked DataPayload. mouseClick (throws) and swipe (optional frame) keep their bespoke measure/wrap but route the success payload through gestureResponse.

Verification

• xcodebuild build-for-testing → TEST BUILD SUCCEEDED (no warnings).
• Pure refactor (field-for-field preserved). Gesture execution is exercised by examples/test-app/replays/gesture-lab.ad (pinch/rotate/pan/fling) — happy to run it on a sim for extra confidence if wanted; the gesture-payload shape is unchanged by construction.

Scope

This was the #1 candidate from the post-merge refactor re-scan (verified deep: complexity concentrates behind two seams serving every gesture). Steers clear of the gesture synthesis (already consolidated in #645) and the Command/DataPayload wire-lock.

[github-actions]

Size Report

Metric	Base	Current	Diff
JS raw	1.1 MB	1.1 MB	0 B
JS gzip	360.7 kB	360.7 kB	0 B
npm tarball	462.0 kB	462.4 kB	+369 B
npm unpacked	1.5 MB	1.5 MB	-2.0 kB

Startup median (7 runs, lower is better):

Scenario	Base	Current	Diff
CLI --version	28.1 ms	28.5 ms	+0.4 ms
CLI --help	43.7 ms	44.6 ms	+1.0 ms

Top changed chunks: no changes in the largest emitted chunks.

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-06-02 01:32 UTC

[thymikee]

Thanks — applied the doc nits (7… follow-up commit), all comment-only:

• idleTimeout: false guardrail — added a NOTE to the performGesture doc: a new synthesis gesture must pass idleTimeout: false; the default true would wrap it in the scroll idle-timeout and change runtime behavior.
• swipe / mouseClick bypass — added a one-line comment on each explaining why they keep raw measureGesture (swipe returns an optional frame; mouseClick throws — neither produces a RunnerInteractionOutcome), and that both still route the success payload through gestureResponse.

On the gesture-lab recommendation: agreed it's the residual risk. This is a field-for-field preserving refactor (timing window, idle-wrap on/off, unsupportedResponse paths, messages, and DataPayload fields all match), so I verified by compile + diff. I'm happy to run examples/test-app/replays/gesture-lab.ad on an iOS sim before merge for empirical confirmation (pinch/rotate/pan/fling) — it needs the test-app Metro bring-up on the default port. Want me to run it, or are you good merging on field-for-field + green CI?