Md Owais Farhan Akhter’s Post

𝗧𝗵𝗲 𝗕𝗼𝗼𝗹𝗲𝗮𝗻 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 You spent hours debugging a loading spinner. The UI showed a spinner and an error message at the same time. This happened because of three booleans: - isLoading - isError - data These booleans created eight possible combinations. But only four of these combinations are valid. The problem is not specific to TypeScript or React. It's a programming issue. You can avoid this problem by using union types instead of booleans. Here's an example: - Instead of using booleans, you can use an enum to define the state of a widget. - You can use a union type to define the state of a loading process. Using union types makes it impossible to create invalid states. This approach is used in many programming languages, including Rust, Python, Swift, and Kotlin. It's not about being trendy, it's about writing correct code. So, what can you do? - Use union types to define states - Avoid using booleans as function parameters - Use enums to define valid states By following these principles, you can make your code more robust and less prone to errors. Source: https://lnkd.in/gBavVtF7

Learn in Public #4 ✅ Just added Real-time Video Calling to my Chat Application! Now users can not only text chat but also make video calls between devices. I tested it by calling from my laptop to my mobile phone — and it works smoothly! Watch the demo: - User 1 clicks Video Call button - User 2 gets incoming call popup and accepts - Camera & Microphone permissions - Live two-way video + audio - Mute & Camera toggle (on/off with single click) - End call from either side Built with: • Backend: FastAPI (Python) • Frontend: Simple HTML, CSS & Vanilla JavaScript • Video Calling: WebRTC with custom signaling I'm more comfortable with Python, so using FastAPI felt natural for the backend. Live Demo: https://lnkd.in/g8AuFeks GitHub Repository:   https://lnkd.in/g3DRrNMW What should I build next?   Screen sharing? Group calls? Or something else? Drop your suggestions in the comments 👇 #LearnInPublic #WebRTC #FastAPI #Python #JavaScript #ChatApp #FullStack

Claude Code just leaked its own source code. Yes, you read that right. Anthropic accidentally shipped the full source map inside their @anthropic-ai/claude-code npm package 512,000 lines. 1,900 files. Fully reconstructable TypeScript. The cherry on top? The codebase contains a subsystem literally called "Undercover Mode" designed to stop Claude from leaking Anthropic internals in git commits. The internet, naturally, did not wait. Developers reverse-engineered the architecture and got to work. Here's what's already out: Claw Code — A clean-room Python reimplementation that captures the full agent harness architecture of Claude Code, built entirely on stdlib with zero proprietary dependencies. 👉 https://lnkd.in/gFvnhG3p free-code — An open-source version of Claude Code that lets you run the same agentic coding loop without the Anthropic paywall. 👉 https://lnkd.in/g3cyuimu Whether you're a security researcher, a curious engineer, or just someone who likes watching billion-dollar companies accidentally open-source themselves this is a moment worth studying.

Most devs missed what Claude Code actually revealed. It ships as a Bun executable. Not Node. Not Python. That one detail points to something bigger: a shift in the runtime layer. So I asked myself — do I actually understand Bun beyond the hype? I built BunEx to find out. It's a clean, hands-on way to explore all four things Bun replaces at once: → Runtime → Package manager → Bundler → Test runner Here's what changes the game: most JS stacks are stitched together from separate tools. Bun collapses that into one — and it's significantly faster. But let's be honest about the tradeoffs: — Not 100% Node-compatible yet — Real adoption needs to be gradual The lowest-friction entry point? Swap npm install for bun install in your next project. That's it. This isn't a rewrite-everything argument. It's pattern recognition: tooling consolidation is the direction the ecosystem is heading, and Bun is the clearest signal we have right now. Full guide in the first comment 👇

"Code is cheap, but performance is expensive." I built this CDN Audit Tool to solve a specific problem: Analyzing asset delivery at scale without the "GUI lag" or slow execution typical of many Python automation scripts. Check out the video below to see it in action! 👇 What's happening under the hood? Concurrency: Managed via ThreadPoolExecutor, allowing the tool to check dozens of assets simultaneously while the main thread stays 100% responsive. Thread-Safe UI: Implemented a queue.Queue "drain" system to pump background scan data into the Tkinter interface every 180ms without crashing the main loop. Architecture: Used a custom "Glow" widget system I built from scratch using the Tkinter Canvas API for a modern, dark-mode aesthetic. Efficiency: A global deduplication cache ensures we never ping the same asset twice, even if it appears on every page of a site. For me, being a developer isn't just about making things work—it's about making them fast, scalable, and user-friendly. #Python #SoftwareDevelopment #WebPerformance #GUIManagement #SoftwareEngineer #CodingProject

DSA series... 🚀 Chapter 3: "Sliding Window Dynamic" LeetCode #3 (Level: Medium) Longest Substring Without Repeating Characters... 🧐 At first glance, this looks like a string problem … Especially substring problem, without any doubts move to Sliding Window (Dynamic) ... 🔥 Sliding Window (Dynamic) ... ⚡ - Just move your right pointer to expand until you hit your goal, then pull your left pointer to shrink and find the optimal size. Explanation ... ✍️ 1⃣ Get the length and check the edge condition. 2⃣ Use two pointers `left` and `right`. Keep a set/map to track characters existence. 3⃣ Move `right` pointer to add characters on the string when the character is not duplicate one. 4⃣ If duplicate found, move `left pointer` to remove characters until duplicate is gone. 5⃣ Track max length during the process. --- Difference between Sliding Window Fixed vs Dynamic Programming ... 👇 ** On Fixed, we decrement 'left' pointer only once when contition met. Use if statement. ** On Dynamic Programming, we decrement 'left' pointer untill reach the condition fully met on the current window. Use while statement. Note: But sometimes it may be differ... You can see that difference from the previous chapter and the current one ... ❗ --- Easy Trigger to Remember for this pattern usage ... 😎 👉 “Longest substring without repeating characters”, 👉 “At most / no duplicate”. --- Key Insights ... 🎯 - Don’t restart the loop when duplicate appears, - Just shrink the window smartly that saves time (O(n)). Just put on your thoughts or suggestions ... 💬 Stay tuned always ... 😎 #LeetCode #SlidingWindow #DSA #CodingInterview #Java #ProblemSolving #Algorithms #TechLearning #SoftwareEngineering

59.8 MB JavaScript source leak of Claude Code. Surprisingly, it was a human error that happened during the release packaging. No customer data, API credentials, or model weights were exposed. But, 512,000 lines of TypeScript across ~1,900 files were exposed. It includes the query engine, tool system, multi-agent orchestration logic, and context compaction. But the amazing part? Someone noticed this as an opportunity. The guy's name is Sigrid Jin, a Korean Developer. He developed Claw Code, a clean-room Python rewrite of Claude Code's agent harness. He published it on GitHub where the repo reached 50k stars in just 2 hours. If you want the GitHub URL, comment "Claw Code" and I will share it with you.

My CLAUDE.md was 281 lines. Claude loaded all of it every session - whether I was writing a post or debugging a Python script. Post rules. Format guides. Note conventions. Voice guidelines. All in context, all the time. Even when none of it was relevant. Longer context = lower adherence. The docs say it plainly. I ignored it until I ran this prompt: --- *Review my CLAUDE.md against https://lnkd.in/g6UNtsC7 and suggest specific improvements - what to add, what to move to .claude/rules/ files, and what to cut.* --- Claude read the official docs, compared them against my file, and came back with a diagnosis. The biggest one was `.claude/rules/` - a directory most people don't know exists. Here's what it told me to do: Rules in `.claude/rules/` load just like CLAUDE.md - but you can scope them to specific file paths. ``` paths: - "posts/**" ``` That one frontmatter block means post rules only enter context when Claude is working inside that folder. Format guides, note conventions - same thing. Zero tokens wasted when they don't apply. After the restructure: - CLAUDE.md: 281 lines → 69 lines - Always-loaded context: ~180 lines total - 3 rule files load on demand, only when relevant The path-scoped rules are in the official docs at https://lnkd.in/g6UNtsC7. Almost nobody uses them. Run `/memory` in Claude Code to see exactly what's loaded in your current session. If it's over 200 lines, that's where the drift is coming from. What rule do you keep repeating to Claude because it keeps forgetting? #ClaudeAI #AITools #BuildInPublic #ProductManagement #DeveloperTools

Your Claude Code token bill drops 36% the day you stop dumping your repo into context. Your agent's token usage drops 27x on the same task. Same model. Different retrieval. Here's what's broken right now. Your README is from six months ago. Your architecture doc predates the rewrite. Your agent quotes both like they're scripture and burns tokens doing it. Repowise(https://lnkd.in/gxjrWD8T) scores every doc by confidence. Every git diff updates the score. When a function changes, the docs that referenced it get flagged before your agent ever reads them. Four layers feed retrieval. Git history. Code graph. Versioned docs. Decision records. Your agent pulls from the layer that matches the code you shipped yesterday, with a confidence number attached. Stale wiki page from 2024 scores low. Decision record from last week scores high. Your agent picks accordingly. Seven MCP tools. Self-hosted. Code never leaves your machine. Works with Claude Code, Cursor, any MCP client. pip install repowise Python 3.11+ v0.3.0 AGPL-3.0 Benchmarks run against naive full-repo context on Claude Code. 36% cheaper. 27x more token efficient. Numbers in the repo.

Someone built a self-hosted video downloader in 150 lines of Python. ReClip. Paste links from YouTube, TikTok, Instagram, Twitter/X, Reddit, and 1000+ other sites. Download as MP4 or MP3. What's inside: → 1000+ supported sites via yt-dlp → MP4 video or MP3 audio extraction → Quality/resolution picker → Bulk downloads - paste multiple URLs at once → Automatic URL deduplication → Clean, responsive UI - no frameworks, no build step → Single Python file backend (~150 lines) The stack: → Backend: Python + Flask (~150 lines) → Frontend: Vanilla HTML/CSS/JS (single file, no build step) → Download engine: yt-dlp + ffmpeg → Dependencies: 2 (Flask, yt-dlp) Self-host with one command: git clone → ./reclip[.]sh → done. Or Docker: docker build → docker run → done. MIT licensed. Open-source. GitHub: https://lnkd.in/ghXzwgwq