// ❌ Blocks JS thread during scroll
const parseData = (raw) => JSON.parse(raw);

// ✅ Offloaded to Web Worker / Native Module
import { runOnJS, useWorklet } from 'react-native-reanimated';
const parseWorklet = useWorklet((raw) => {
  'worklet';
  const result = JSON.parse(raw);
  runOnJS(updateState)(result);
});

import { FlashList } from '@shopify/flash-list';

const MemoizedFeedItem = React.memo(({ item, onPress }) => {
  return (
    <TouchableOpacity onPress={onPress}>
      <FastImage source={{ uri: item.thumbnail }} style={styles.image} />
      <Text numberOfLines={2}>{item.title}</Text>
    </Toucha

bleOpacity> ); });

// In parent const handlePress = useCallback((id) => { /* ... */ }, []); return <FlashList data={items} renderItem={({ item }) => <MemoizedFeedItem item={item} onPress={() => handlePress(item.id)} />} estimatedItemSize={120} />;

### Phase 3: State Management & Re-Render Control

Context API triggers full subtree re-renders on any value change. For high-frequency updates (scroll position, WebSocket streams, form inputs), external stores with selector-based subscriptions are mandatory.

**Architecture Decision:** Use Zustand or Redux Toolkit with `useSelector` and `shallow` equality. Avoid `useContext` for state that changes >60Hz.

```javascript
import { create } from 'zustand';

const useFeedStore = create((set) => ({
  items: [],
  loading: false,
  appendItems: (newItems) => set((state) => ({
    items: [...state.items, ...newItems],
  })),
}));

// Component subscribes only to specific slice
const items = useFeedStore((state) => state.items);

Phase 4: Native Bridge & Fabric Migration

The legacy bridge serializes all JS↔Native calls. Fabric replaces it with a synchronous C++ renderer and TurboModules enable direct method invocation.

Implementation:

• Enable bridgeless mode in AppDelegate.mm and MainApplication.java.
• Migrate critical native modules to TurboModules using the New Architecture codegen.
• Use react-native-mmkv or react-native-sqlite-storage for disk I/O instead of AsyncStorage.

// Android: MainApplication.java
@Override
protected boolean isBridgelessEnabled() {
  return true;
}

// iOS: AppDelegate.mm
- (BOOL)bridgelessEnabled {
  return YES;
}

Profiling Workflow

Performance optimization without measurement is guesswork. Establish a profiling pipeline:

1. JS Thread: React DevTools Profiler → record interactions → identify wasted renders.
2. Native Thread: Android Studio Profiler / Xcode Instruments → trace UI thread block time.
3. Bridge/Architecture: systrace or Perfetto → measure frame composition time and GC pauses.
4. Memory: Flipper Memory Plugin → snapshot heap → detect retained references and image leaks.

Run baseline profiles before and after every optimization. Track FPS stability, JS block time, and memory delta.

Pitfall Guide

1. Treating useEffect as a render trigger
   useEffect runs after paint. Using it for data fetching or state derivation causes double renders and layout thrashing. Prefer useMemo for synchronous derivation and React Query/SWR for async data.

2. Rendering full-resolution images in lists
   Loading 4K images into memory for 100px thumbnails spikes heap usage and triggers GC pauses. Always use react-native-fast-image with explicit resizeMode, cacheControl, and CDN-resized URLs.

3. Ignoring Hermes bytecode compilation
   Hermes compiles JS to bytecode at build time, eliminating V8 startup overhead and reducing memory fragmentation. Disabling it for debugging in production builds increases cold start by 30–40%.

4. Synchronous bridge calls blocking UI
   Passing large arrays, base64 strings, or unfiltered objects across the bridge freezes the JS thread. Serialize to JSON, chunk payloads, or move logic to native modules.

5. Missing keyExtractor or using array index as key
   Index keys break reconciliation when items reorder or filter. React re-renders entire lists unnecessarily. Always use stable, unique identifiers.

6. Over-engineering with Context for high-frequency updates
   Context triggers O(n) re-renders. For scroll offsets, keyboard state, or real-time streams, use external stores or useRef + custom event emitters.

7. Skipping memory profiling for async tasks
   Unresolved promises, lingering event listeners, and uncleared intervals retain references. Use useRef for timers, cleanup in useEffect, and snapshot memory before/after navigation.

Production Bundle

Action Checklist

• Enable Hermes bytecode compilation and verify with hermes --version
• Replace FlatList with FlashList and set estimatedItemSize
• Wrap all list items and expensive components in React.memo
• Migrate high-frequency state to Zustand/Redux with selector subscriptions
• Configure Metro for aggressive minification, inline requires, and tree shaking
• Implement react-native-fast-image with CDN-resized thumbnails
• Set up systrace/Perfetto baseline profiling pipeline and track FPS/block time

Decision Matrix

Strategy	Use When	Avoid When	Performance Impact
Hermes vs JSC	Production builds, memory-constrained devices	Debugging native C++ crashes	↓ Cold start 30%, ↓ GC pauses 40%
FlashList vs FlatList	Lists >50 items, dynamic heights, media	Static <20 item lists	↓ Memory 60%, ↑ FPS stability
Context vs External Store	Low-frequency app config/theme	Scroll, form, WebSocket data	↓ Re-renders 70% with selectors
Bridge vs Bridgeless	Legacy native modules, quick POC	Production, gesture-heavy apps	↓ Latency 80%, ↑ UI thread responsiveness
AsyncStorage vs MMKV	Simple key-value, offline config	High-frequency state, large payloads	↓ Read/Write latency 90%

Configuration Template

metro.config.js

const { getDefaultConfig, mergeConfig } = require('@react-native/metro-config');

const defaultConfig = getDefaultConfig(__dirname);

const config = {
  transformer: {
    getTransformOptions: async () => ({
      transform: {
        experimentalImportSupport: false,
        inlineRequires: true,
      },
    }),
  },
  resolver: {
    sourceExts: [...defaultConfig.resolver.sourceExts, 'mjs'],
  },
};

module.exports = mergeConfig(defaultConfig, config);

android/app/build.gradle (Hermes + ProGuard)

project.ext.react = [
    enableHermes: true,
    hermesFlagsRelease: ["-O", "-output-source-map"],
]

def enableProguardInReleaseBuilds = true

android {
    buildTypes {
        release {
            minifyEnabled enableProguardInReleaseBuilds
            proguardFiles getDefaultProguardFile("proguard-android.txt"), "proguard-rules.pro"
        }
    }
}

babel.config.js (Production Optimizations)

module.exports = {
  presets: ['module:@react-native/babel-preset'],
  plugins: [
    ['react-native-reanimated/plugin', { relativeSourceLocation: true }],
    'react-native-paper/babel',
  ],
  env: {
    production: {
      plugins: [
        'react-native-paper/babel',
        ['@babel/plugin-transform-runtime', { regenerator: true }],
      ],
    },
  },
};

Quick Start Guide

1. Baseline Profile
   Run systrace or React DevTools Profiler on your current app. Record FPS, JS block time, and memory usage for your heaviest screen.

2. Enable Hermes & Metro Optimizations
   Set enableHermes: true, enable inlineRequires, and run a release build. Verify bytecode compilation with hermes --version.

3. Swap List Implementation
   Replace FlatList with FlashList. Provide estimatedItemSize and wrap items in React.memo. Remove index-based keys.

4. Stabilize State & Handlers
   Migrate high-frequency state to an external store. Wrap event handlers in useCallback. Profile again and compare deltas.

5. Iterate with Native Modules
   Identify JS thread bottlenecks via systrace. Migrate image decoding, parsing, or crypto to react-native-reanimated worklets or TurboModules. Re-measure.

Performance in React Native is not a framework limitation; it is an architecture discipline. By enforcing threading boundaries, constraining re-renders, adopting bridgeless execution, and profiling continuously, teams consistently achieve stable 60 FPS, sub-700ms cold starts, and predictable memory footprints across mid-tier and flagship devices.