refactor(topics): 멀티 에이전트용 지식 재편 — _Common(공통 기본기) + Domain_* 구조
에이전트 8종(대화형/프로그래머 C·S/디자이너/설계자/기획자/QA/PD/PM)에게 [공통 기본 능력 + 롤별 Specialty] 2층으로 지식을 주입하기 위한 재분류. 문서 내용·포맷은 무수정, 폴더 이동만 (6,372개 문서 수 보존 확인). - Topic_Programming → Domain_Programming (내부 구조 보존) - Topic_Graphic → Domain_Design - Topic_Business → Domain_Product - Topic_General → Domain_General - _Common 신설: Math(구 Topic_Math_Specialty), Reasoning(구 General/From_Thinking & Reasoning), Reasoning_Creativity(구 General/From_창의성), Communication(Poetic_Blog_Writing + From_writing) - 타 도메인의 From_* 폴더는 유지 (출처 표기일 뿐, 이미 도메인에 맞게 분류된 문서) - 빈 폴더 정리 (memory/procedures) - 에이전트→폴더 매핑은 workspace의 .astra/agent-knowledge-map.json (9개 에이전트) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Antigravity Agent
@@ -0,0 +1,227 @@
|
||||
---
|
||||
id: wiki-2026-0508-불필요한-리렌더링-방지
|
||||
title: 불필요한 리렌더링 방지
|
||||
category: 10_Wiki/Topics
|
||||
status: verified
|
||||
canonical_id: self
|
||||
aliases: [Prevent Unnecessary Re-render, React 최적화, memoization, render optimization]
|
||||
duplicate_of: none
|
||||
source_trust_level: A
|
||||
confidence_score: 0.9
|
||||
verification_status: applied
|
||||
tags: [frontend, react, performance, optimization, memoization]
|
||||
raw_sources: []
|
||||
last_reinforced: 2026-05-10
|
||||
github_commit: pending
|
||||
tech_stack:
|
||||
language: typescript
|
||||
framework: react
|
||||
---
|
||||
|
||||
# 불필요한 리렌더링 방지
|
||||
|
||||
## 매 한 줄
|
||||
> **"매 render 매 cheap by default — 매 measure first, 매 then memoize"**. React 매 default 의 parent render 의 child 의 모두 re-render — 매 보통 fast. 매 React Compiler (RC 2026) 의 auto-memoization 의 manual `useMemo`/`useCallback` 의 obsolescence 약속. 매 그 전 의 explicit memoization + state colocation.
|
||||
|
||||
## 매 핵심
|
||||
|
||||
### 매 re-render trigger
|
||||
- **State change** (`useState`, `useReducer`).
|
||||
- **Props change** (parent render).
|
||||
- **Context change** (consumer re-renders).
|
||||
- **Parent re-render** (default behavior).
|
||||
|
||||
### 매 4가지 전략
|
||||
1. **State colocation** — state 의 use 의 곳 가까이 (lift down).
|
||||
2. **Memoization** — `React.memo`, `useMemo`, `useCallback`.
|
||||
3. **Context split** — separate frequently-changing from stable.
|
||||
4. **Component composition** — `children` prop 의 stable reference.
|
||||
|
||||
### 매 응용
|
||||
1. List rendering (10k+ items → virtualization + memo).
|
||||
2. Form state (controlled inputs → debounced or uncontrolled).
|
||||
3. Global state (Zustand selector, Redux `useSelector` with shallowEqual).
|
||||
4. Heavy computation (`useMemo` for derived state).
|
||||
|
||||
## 💻 패턴
|
||||
|
||||
### 1. React.memo for pure components
|
||||
```typescript
|
||||
import { memo } from 'react';
|
||||
|
||||
interface ItemProps {
|
||||
id: string;
|
||||
name: string;
|
||||
onClick: (id: string) => void;
|
||||
}
|
||||
|
||||
export const Item = memo(function Item({ id, name, onClick }: ItemProps) {
|
||||
console.log('render:', id);
|
||||
return <li onClick={() => onClick(id)}>{name}</li>;
|
||||
});
|
||||
|
||||
// custom comparator (rarely needed)
|
||||
export const ItemDeep = memo(Item, (prev, next) => prev.id === next.id && prev.name === next.name);
|
||||
```
|
||||
|
||||
### 2. useCallback for stable function reference
|
||||
```typescript
|
||||
import { useCallback, useState } from 'react';
|
||||
|
||||
function List({ items }: { items: { id: string; name: string }[] }) {
|
||||
const [selected, setSelected] = useState<string | null>(null);
|
||||
|
||||
// without useCallback — new ref each render → all Item re-render
|
||||
const handleClick = useCallback((id: string) => setSelected(id), []);
|
||||
|
||||
return (
|
||||
<ul>
|
||||
{items.map(it => <Item key={it.id} {...it} onClick={handleClick} />)}
|
||||
</ul>
|
||||
);
|
||||
}
|
||||
```
|
||||
|
||||
### 3. useMemo for expensive computation
|
||||
```typescript
|
||||
import { useMemo } from 'react';
|
||||
|
||||
function Report({ rows }: { rows: Row[] }) {
|
||||
const stats = useMemo(() => {
|
||||
return {
|
||||
total: rows.reduce((a, r) => a + r.amount, 0),
|
||||
avg: rows.length ? rows.reduce((a, r) => a + r.amount, 0) / rows.length : 0,
|
||||
max: Math.max(...rows.map(r => r.amount)),
|
||||
};
|
||||
}, [rows]);
|
||||
|
||||
return <Summary {...stats} />;
|
||||
}
|
||||
```
|
||||
|
||||
### 4. State colocation (lift down)
|
||||
```typescript
|
||||
// BAD — input state at parent re-renders entire tree
|
||||
function App() {
|
||||
const [search, setSearch] = useState('');
|
||||
return (
|
||||
<>
|
||||
<input value={search} onChange={e => setSearch(e.target.value)} />
|
||||
<ExpensiveTree /> {/* re-renders on every keystroke */}
|
||||
</>
|
||||
);
|
||||
}
|
||||
|
||||
// GOOD — colocate input state in SearchInput
|
||||
function SearchInput({ onCommit }: { onCommit: (s: string) => void }) {
|
||||
const [v, setV] = useState('');
|
||||
return <input value={v} onChange={e => { setV(e.target.value); onCommit(e.target.value); }} />;
|
||||
}
|
||||
function App() {
|
||||
return <><SearchInput onCommit={...} /><ExpensiveTree /></>;
|
||||
}
|
||||
```
|
||||
|
||||
### 5. Context split (avoid frequent re-renders)
|
||||
```typescript
|
||||
// BAD — single context, everything re-renders on theme change
|
||||
const AppCtx = createContext<{user: User; theme: 'light'|'dark'; cart: Item[]}>(...);
|
||||
|
||||
// GOOD — split by change frequency
|
||||
const UserCtx = createContext<User | null>(null); // rare change
|
||||
const ThemeCtx = createContext<'light'|'dark'>('light'); // user-toggled
|
||||
const CartCtx = createContext<Item[]>([]); // frequent
|
||||
|
||||
// only consumers of changed context re-render
|
||||
```
|
||||
|
||||
### 6. Children as stable prop (composition)
|
||||
```typescript
|
||||
// BAD — Wrapper re-renders => inner re-renders even if same JSX
|
||||
function App() {
|
||||
const [count, setCount] = useState(0);
|
||||
return <Wrapper>{<HeavyChild />}</Wrapper>; // HeavyChild created each render? No — JSX is reference
|
||||
}
|
||||
|
||||
// pattern: pass children to escape parent re-render
|
||||
function Wrapper({ children }: { children: ReactNode }) {
|
||||
const [internal, setInternal] = useState(0);
|
||||
return <div>{children}</div>; // children reference stable from parent
|
||||
}
|
||||
```
|
||||
|
||||
### 7. Zustand selector (subscribe slice)
|
||||
```typescript
|
||||
import { create } from 'zustand';
|
||||
import { useShallow } from 'zustand/react/shallow';
|
||||
|
||||
const useStore = create<{ count: number; user: User; inc: () => void }>((set) => ({
|
||||
count: 0,
|
||||
user: { id: '1', name: 'A' },
|
||||
inc: () => set(s => ({ count: s.count + 1 })),
|
||||
}));
|
||||
|
||||
// only re-renders when count changes (not user)
|
||||
function Counter() {
|
||||
const count = useStore(s => s.count);
|
||||
return <span>{count}</span>;
|
||||
}
|
||||
|
||||
// shallow compare for object selector
|
||||
function Profile() {
|
||||
const { name, id } = useStore(useShallow(s => ({ name: s.user.name, id: s.user.id })));
|
||||
return <span>{id}: {name}</span>;
|
||||
}
|
||||
```
|
||||
|
||||
### 8. React Profiler (measure first)
|
||||
```typescript
|
||||
import { Profiler } from 'react';
|
||||
|
||||
function onRender(id: string, phase: 'mount' | 'update', actualDuration: number) {
|
||||
if (actualDuration > 16) console.warn(`${id} ${phase} took ${actualDuration}ms`);
|
||||
}
|
||||
|
||||
<Profiler id="List" onRender={onRender}>
|
||||
<List items={items} />
|
||||
</Profiler>
|
||||
```
|
||||
|
||||
## 매 결정 기준
|
||||
| 상황 | Approach |
|
||||
|---|---|
|
||||
| 100s of items, fast props | `React.memo` + `useCallback`. |
|
||||
| 10k+ items | virtualization (react-window, TanStack Virtual). |
|
||||
| Heavy compute | `useMemo`. |
|
||||
| Frequent global state | Zustand selector / Redux Toolkit. |
|
||||
| Form input | uncontrolled (react-hook-form) or colocated state. |
|
||||
| React 19+ | RC compiler 의 auto-memo — manual 의 less needed. |
|
||||
|
||||
**기본값**: 매 measure first (React DevTools Profiler). 매 colocate state. 매 React Compiler (when stable) 의 auto-memo. 매 manual memo 의 hot paths only.
|
||||
|
||||
## 🔗 Graph
|
||||
- 부모: [[Memoization]]
|
||||
- 변형: [[React Compiler]] · [[Zustand]]
|
||||
- 응용: [[가상 스크롤]] · [[Virtualization]]
|
||||
|
||||
## 🤖 LLM 활용
|
||||
**언제**: measured perf bottleneck (Profiler), large list, heavy compute, frequent state updates.
|
||||
**언제 X**: 매 not measured — premature optimization. 매 fast already (< 16ms render).
|
||||
|
||||
## ❌ 안티패턴
|
||||
- **Memoize everything**: `useMemo` 의 cost (deps compare) > savings 의 cheap compute.
|
||||
- **Inline object/array as prop**: `<Comp data={{a:1}} />` 매 new ref 매 every render — defeats memo.
|
||||
- **Context for everything**: monolithic context → all consumers re-render.
|
||||
- **`useCallback` w/o `React.memo` child**: useless — child still re-renders on parent.
|
||||
- **Premature optimization**: optimize before measuring.
|
||||
- **Mutating state directly**: `state.list.push(...)` — React doesn't detect change.
|
||||
|
||||
## 🧪 검증 / 중복
|
||||
- Verified (React docs 2026, Kent C. Dodds patterns, React Compiler RFC).
|
||||
- 신뢰도 A.
|
||||
|
||||
## 🕓 Changelog
|
||||
| 날짜 | 변경 |
|
||||
|---|---|
|
||||
| 2026-05-08 | Phase 1 |
|
||||
| 2026-05-10 | Manual cleanup — full content with 8 patterns + React Compiler 2026 note |
|
||||
Reference in New Issue
Block a user