Reduce collector overhead

[mattcosta7]

Summary

• scope style mutation and layer-promotion work to the active story root
• coalesce pointer sampling into one cancellable RAF pipeline and scan layer candidates in bounded idle chunks
• remove global forced-reflow prototype patches in favor of native LoAF forced-style/layout evidence
• add opt-in internal callback, aggregation, serialization, scan, and pending-work telemetry plus a read-only benchmark comparator
• preserve deprecated metric aliases while clarifying their native replacement semantics
• escape generated attribution selectors so inspect actions remain valid for special-character DOM identifiers

Benchmark

Compared with the tracked baseline from #161:

• visible lifecycle mean: -62.1%
• visible React commit mean: -19.2%
• visible raw DOM mean: +7.8% in the final one-shot run (+0.7% in the preceding run), so this remains noisy/inconclusive
• telemetry probe processed all 60 layer candidates and reported zero current RAF, idle-scan, and DOM-count work after cleanup

The tracked baseline was not modified.

Validation

• 921 browser tests across Chromium, Firefox, and WebKit
• npm run tsc
• npm run lint
• npm run build
• attw
• publint
• full stacked adversarial review using the requested three-persona workflow; its concrete selector-integrity finding was fixed and regression-tested

Stacked on #164. Keep this PR in draft while the v1.3 stack is reviewed.

[Copilot]

Pull request overview

Reduces collector overhead by scoping DOM work, coalescing asynchronous sampling, and replacing prototype instrumentation with native LoAF evidence.

Changes:

• Scopes mutation and layer scans to the story root.
• Adds cancellable RAF/idle pipelines and opt-in telemetry.
• Adds native forced-layout metrics and benchmark comparison tooling.

Show a summary per file

File	Description
.changeset/reduce-collector-overhead.md	Records the minor release.
README.md	Documents metrics, telemetry, and benchmarking.
performance-panel.tsx	Displays native forced-layout LoAF counts.
package.json	Adds the benchmark comparator command.
core/preview-core.ts	Integrates optional telemetry.
core/performance-types.ts	Adds the native LoAF metric and alias semantics.
core/overhead-telemetry.ts	Implements overhead telemetry.
core/__tests__/overhead-telemetry.browser.test.ts	Tests telemetry aggregation.
collectors/README.md	Updates collector documentation.
collectors/attribution.ts	Escapes generated selectors.
collectors/collector-manager.ts	Scopes collectors and aggregates native metrics.
collectors/element-timing-collector.ts	Instruments callback overhead.
collectors/forced-reflow-collector.ts	Removes prototype instrumentation.
collectors/frame-timing-collector.ts	Tracks RAF callback and pending work.
collectors/input-collector.ts	Coalesces pointer RAF sampling.
collectors/layout-shift-collector.ts	Instruments observer callbacks.
collectors/long-animation-frame-collector.ts	Counts forced-layout LoAFs.
collectors/main-thread-collector.ts	Instruments long-task processing.
collectors/paint-collector.ts	Adds scoped, chunked layer scans.
collectors/react-profiler-collector.ts	Instruments profiler callbacks.
collectors/style-mutation-collector.ts	Scopes mutation observation.
collectors/__tests__/attribution.browser.test.ts	Tests selector escaping.
collectors/__tests__/collector-manager.browser.test.ts	Tests scoping, aliases, and telemetry.
collectors/__tests__/forced-reflow-collector.browser.test.ts	Removes obsolete instrumentation tests.
collectors/__tests__/input-collector.browser.test.ts	Tests RAF coalescing and cancellation.
collectors/__tests__/long-animation-frame-collector.browser.test.ts	Tests forced-layout counts.
collectors/__tests__/paint-collector.browser.test.ts	Tests scoped chunked scans.
collectors/__tests__/style-mutation-collector.browser.test.ts	Tests scoped mutation collection.
__tests__/performance-decorator-universal.browser.test.ts	Tests serialization telemetry.
__benchmarks__/overhead.browser.bench.tsx	Adds the telemetry probe.
__benchmarks__/compare-results.ts	Adds read-only benchmark comparison.

Review details

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 31/31 changed files
• Comments generated: 1
• Review effort level: Balanced