Using React Refs for DOM Access and Edge Cases

Your component is re-rendered... Then what? Most developers stop thinking here. And that’s where the real magic starts. This is Post 3 of the series: 𝗥𝗲𝗮𝗰𝘁 𝗨𝗻𝗱𝗲𝗿 𝘁𝗵𝗲 𝗛𝗼𝗼𝗱 Where I explain what React is actually doing behind the scenes. React does NOT update the DOM immediately. Because touching the DOM is expensive. Instead, it follows a 3-step process: Render → Diff → Commit First, React creates something called a 𝗩𝗶𝗿𝘁𝘂𝗮𝗹 𝗗𝗢𝗠. Not the real DOM. Just a JavaScript object in memory. A blueprint of your UI. Every time your component runs, React builds a new tree of this Virtual DOM. And it already has the previous tree stored. Now comes the interesting part. React compares: Old tree vs New tree This step is called 𝗗𝗶𝗳𝗳𝗶𝗻𝗴. It finds exactly what changed. Not everything. Just the difference. Then comes the final step: 𝗖𝗼𝗺𝗺𝗶𝘁 phase React takes only those changes… …and updates the real browser DOM So no, React is NOT re-rendering your whole page. It’s doing surgical updates. The key insight: React separates thinking from doing. Thinking = Render + Diff Doing = Commit That’s why React is fast. Because touching the DOM is expensive. So React minimizes it. Flow looks like this: ✅️ Component runs ✅️ Virtual DOM created ✅️ Changes calculated ✅️ DOM updated ✅️ Browser paints UI Once you see this, you stop fearing re-renders. And start understanding them. In Post 4 of React Under the Hood: How does React actually compare trees? (The diffing algorithm 👀) Follow Farhaan Shaikh if you want to understand react more deeply. 👉 Read in detail here: https://lnkd.in/dTYAbM6n #React #Frontend #WebDevelopment #JavaScript #SoftwareEngineering

💡 Understanding a subtle React concept: Hydration & “bailout” behavior One interesting nuance in React (especially with SSR frameworks like Next.js) is how hydration interacts with state updates. 👉 Hydration is the process where React makes server-rendered HTML interactive by attaching event listeners and syncing state on the client. When a page is server-rendered, the initial HTML is already in place. During hydration, React attaches event listeners and syncs the client state with that UI. Here’s the catch 👇 👉 If the client-side state matches what React expects, it may skip updating the DOM entirely. This is due to React’s internal optimization often referred to as a “bailout”. 🔍 Why this matters In cases like theme handling (dark/light mode): If the server renders a default UI (say light mode ☀️) And the client immediately initializes state to dark mode 🌙 React may still skip the DOM update if it doesn’t detect a meaningful state transition 👉 Result: UI can temporarily reflect the server version instead of the actual state. 🧠 Conceptual takeaway A more reliable pattern is: ✔️ Start with an SSR-safe default (consistent with server output) ✔️ Then update state after hydration (e.g., in a layout effect) This ensures React sees a real state change and updates the UI accordingly. 🙌 Why this is fascinating It highlights how deeply optimized React is — sometimes so optimized that understanding its internal behavior becomes essential for building predictable UI. Grateful to the developer community for continuously sharing such insights that go beyond surface-level coding. 🚀 Key idea In SSR apps, correctness isn’t just about what state you set — it’s also about when React observes the change. #ReactJS #NextJS #FrontendDevelopment #WebDevelopment #JavaScript #Learning

React Fiber: what it actually means for your code Most React devs know what Fiber is by name. Fewer know why it matters day to day. The old reconciler worked synchronously. State changes, React walks the whole tree, diffs it, writes to the DOM. One pass, no stopping. On a large component tree that blocked the main thread long enough for users to notice — animations stuttered, inputs lagged. Fiber fixed this by replacing the recursive call stack with a linked list. Each fiber is a plain object: component type, props, state, and pointers to parent, first child, next sibling. React now owns the execution, so it can pause whenever something more urgent comes in. Work splits into two phases, and this is the part worth understanding. The render phase builds a work-in-progress tree in memory without touching the DOM. React can stop mid-tree, hand control back to the browser for a keypress or a frame, then continue. If a higher-priority update arrives, it discards the in-progress work and starts over. The commit phase doesn't pause. Once the tree is ready, React applies all DOM mutations in one go. This part has to be synchronous — a half-applied DOM update breaks the UI. Here's where it connects to your actual code. useTransition works because React can genuinely deprioritize a state update and interrupt it if needed. Without Fiber, marking something as "non-urgent" would be meaningless — the old reconciler couldn't stop anyway. Suspense pauses a subtree while data loads for the same reason. React keeps the current tree on screen, builds the new one in the background, swaps when ready. Slow renders that only appear under load, inputs that lag when a big list re-renders, Suspense boundaries not behaving as expected — all of these trace back to how Fiber schedules and interrupts work. When something's off and the profiler gives you a confusing flamegraph, knowing that React works in two phases (one interruptible, one not) usually tells you where to look. #react #frontend #javascript #webdev #reactjs #frontenddevelopment #softwaredevelopment

Why does your 'ref' return 'null' the moment you wrap your input in a custom component? It is one of those React quirks that trips up even experienced developers. By default, refs do not behave like props. They are treated as special instances that stop at the component boundary to protect encapsulation. This is exactly why we need 'forwardRef'. It acts as a bridge that allows a parent component to reach through a child and grab a direct handle on an internal DOM node. The most practical application is building a reusable UI library. If you create a custom 'Input' or 'Button' component, your parent component might need to call '.focus()' or '.scrollIntoView()'. Without 'forwardRef', your parent is just holding a reference to a React internal object, not the actual HTML element. Another real-world scenario involves Higher-Order Components (HOCs). If you wrap a component with a 'withAnalytics' or 'withTheme' wrapper, that wrapper will 'swallow' the ref by default. By using 'forwardRef' in your HOC implementation, you ensure the reference passes through the wrapper and lands on the component that actually needs it. It is about making your abstractions transparent and ensuring that the parent component maintains the control it expects over the physical DOM. Using this pattern sparingly is key. It is an escape hatch for those moments where declarative state isn't enough to manage physical browser behavior. It keeps your component interfaces clean while providing the necessary access for specialized UI interactions. #ReactJS #SoftwareEngineering #Frontend #WebDevelopment #Javascript #CodingTips #CodingBestPractices

My list UI was behaving weirdly… items were changing unexpectedly 😅 Yes, seriously. At first, I thought it was a state issue. But the problem was something very simple. 💡 I was doing this: {items.map((item, index) => ( Seemed fine… right? ❌ ⚠️ This caused: • Wrong UI updates • Items swapping incorrectly • Hard-to-debug bugs 💡 Then I fixed it: 👉 Used a unique and stable key 🧠 Example: {items.map((item) => ( ✅ Result: • Correct UI updates • No unexpected behavior • Predictable rendering 🔥 What I learned: Keys are not just for React warnings 👉 they control how React updates the UI 💬 Curious: Do you still use index as key… or avoid it? #ReactJS #FrontendDeveloper #JavaScript #CodingTips #WebDevelopment

“Reconciliation” sounds like one of those scary React terms most developers avoid. But it’s easier to understand with a simple analogy 👇 Think of your UI like a **shopping list on a whiteboard**. You have an old list: - Milk   - Bread   - Eggs  Now you update it: - Milk   - Eggs   - Butter  Instead of erasing the whole board and rewriting everything, you: - Remove “Bread”   - Add “Butter”   - Keep “Milk” and “Eggs” as they are  That comparison step is **diffing**.   Actually updating the board is **reconciliation**. --- React works the same way. Whenever state or props change: - It creates a new Virtual DOM (updated list)   - Compares it with the previous Virtual DOM (old list)   - Figures out the minimal set of changes   - Updates only those parts in the real DOM  --- But React doesn’t do a perfect comparison — it uses smart shortcuts: - **Element type matters**    If a `<div>` becomes a `<span>`, React rebuilds that part completely  - **Keys in lists are critical**    Keys are like unique IDs in your shopping list    Using index as key can confuse React when items move  - **No deep comparison**    React avoids checking everything deeply to keep things fast and predictable  --- ⚠️ Real example where things break: You render a list with index as key → reorder items →   inputs lose focus or state jumps around  Reason: React couldn’t track which item is actually which  --- What changed for me after understanding this: - I stopped relying on React as “magic”   - Started using stable and meaningful keys   - Kept component structure consistent   - Avoided unnecessary prop changes  --- Because at the end of the day, React is fast —   but only if you play by its rules. --- #ReactJS #Frontend #VirtualDOM #Performance #JavaScript #WebDevelopment

🚀 Controlled vs Uncontrolled Components in React — Real-World Perspective Most developers learn: 👉 Controlled = React state 👉 Uncontrolled = DOM refs But in real applications… 👉 The choice impacts performance, scalability, and maintainability. 💡 Quick Recap 🔹 Controlled Components: Managed by React state Re-render on every input change 🔹 Uncontrolled Components: Managed by the DOM Accessed via refs ⚙️ The Real Problem In large forms: ❌ Controlled inputs → Too many re-renders ❌ Uncontrolled inputs → Hard to validate & manage 👉 So which one should you use? 🧠 Real-world Decision Rule 👉 Use Controlled when: ✔ You need validation ✔ UI depends on input ✔ Dynamic form logic exists 👉 Use Uncontrolled when: ✔ Performance is critical ✔ Minimal validation needed ✔ Simple forms 🔥 Performance Insight Controlled input: <input value={name} onChange={(e) => setName(e.target.value)} /> 👉 Re-renders on every keystroke Uncontrolled input: <input ref={inputRef} /> 👉 No re-render → better performance ⚠️ Advanced Problem (Most devs miss this) 👉 Large forms with 20+ fields Controlled approach: ❌ Can slow down typing 👉 Solution: ✔ Hybrid approach ✔ Use libraries (React Hook Form) 🧩 Industry Pattern Modern apps often use: 👉 Controlled logic + Uncontrolled inputs internally Example: ✔ React Hook Form ✔ Formik (optimized patterns) 🔥 Best Practices ✅ Use controlled for logic-heavy forms ✅ Use uncontrolled for performance-critical inputs ✅ Consider form libraries for scalability ❌ Don’t blindly use controlled everywhere 💬 Pro Insight (Senior Thinking) 👉 This is not about “which is better” 👉 It’s about choosing the right tool for the problem 📌 Save this post & follow for more deep frontend insights! 📅 Day 17/100 #ReactJS #FrontendDevelopment #JavaScript #ReactHooks #PerformanceOptimization #SoftwareEngineering #100DaysOfCode 🚀

Built a to-do app with a glassmorphic UI as a mini project. Kept it simple — vanilla HTML, CSS, and JS. No frameworks. A few things I learned along the way: — Local Storage API to persist tasks on refresh — DOM manipulation with vanilla JS — How CSS backdrop-filter actually works for the glass effect — JSON.stringify and JSON.parse for saving arrays Small project but learned a lot from it. More coming. repo link: https://lnkd.in/dEtj35Us #WebDev #JavaScript #Frontend #BuildInPublic

🌐 Exciting Launch! My First CLI Tool — **amdad-react-cli** Setting up a production-ready React project often takes time—installing dependencies, configuring routing, adding UI libraries, and repeating the same setup across projects. To solve this, I built something to streamline the process. Introducing **amdad-react-cli** A powerful CLI tool that scaffolds modern React applications with a clean, scalable, and production-ready setup in seconds. ✨ What’s included? • Vite-powered React setup for ultra-fast performance • Tailwind CSS v4 + DaisyUI for modern UI out of the box • React Router DOM pre-configured for routing • Lucide + React Icons ready to use • React Hot Toast integrated for notifications • Clean, scalable folder architecture • Premium starter UI with layout and hero section included ⚙️ Get started in seconds: ```bash npx amdad-react-cli@latest ``` I created this tool to save developers time and provide a strong foundation for building real-world React applications. I would really appreciate your feedback  Try it out and let me know your thoughts! 🔗 npm package: https://lnkd.in/gHFG3XVn #ReactJS #Vite #OpenSource #WebDevelopment #JavaScript #DeveloperTools #Frontend