Milan Medvec’s Post

Most of the time, we just run a Playwright test and look at the result. But behind that single command, a full flow is happening... browser launch, fresh context creation, page actions, auto-waits, validations, and cleanup. Understanding this end-to-end process makes debugging much easier. Sharing a simple visual that helped me connect all the pieces together.

Docker Medium: You created a Dockerfile to speed up the build process, but the resulting Docker image size has grown to 2.5 GB. What strategies would you apply to reduce the image size while maintaining functionality? [answer in comment]

Just discovered a killer combo for exploring MCP servers 👇 ``` npx @modelcontextprotocol/inspector npx chrome-devtools-mcp@latest --autoConnect ``` This spins up an MCP server + auto-connects an inspector UI where you can: 🔎 browse available tools 🧩 inspect schemas ⚡ run them live 📡 see real inputs/outputs It’s basically 𝐏𝐨𝐬𝐭𝐦𝐚𝐧 𝐟𝐨𝐫 𝐌𝐂𝐏 — but for AI-native tooling. Why this is cool: 🧠 𝘇𝗲𝗿𝗼 𝗱𝗼𝗰𝘀 𝗻𝗲𝗲𝗱𝗲𝗱 → introspect everything 🚀 𝐬𝐮𝐩𝐞𝐫 𝐟𝐚𝐬𝐭 prototyping 🛠️ perfect for building + debugging MCP integrations Feels like the 𝗺𝗶𝘀𝘀𝗶𝗻𝗴 𝗱𝗲𝘃𝘁𝗼𝗼𝗹 for the MCP ecosystem ✨

One thing that always saves me hours of debugging.. and now it's just habit to me, is adding one simple line at the end of every execution prompt: ".. before building this, go on the internet and check the latest versions of dependencies and their docs that will be used in the project"

Just Rewritten entire Lockr, and created a new release today. +Using Svelte instead of React +Better IPC calls between Binary and frontend +Proper Error handling +Better logging on both , OS and Svelte Side +Modular Interface that can easily be replaced by a generic droping +Even OS Errors are being logged to Console For Release v1.0: [Github] https://lnkd.in/eydJtQyj

Built a bounded multi-producer multi-consumer (MPMC) queue in Rust 🦀 Implemented with: • Mutex<VecDeque<T>> + Condvar • Blocking / non-blocking push & pop APIs • Full contention stress tests • Throughput benchmarking across thread counts / capacities / asymmetric workloads A great deep dive into Rust concurrency, synchronization primitives, and benchmark-driven systems design. Excited to keep pushing further into low-level backend / systems engineering. #RustLang #Concurrency #SystemsProgramming #BackendEngineering #PerformanceEngineering github: https://lnkd.in/gJ_Cj35T

In a world of Move fast and break things, the Dockerfile is the only thing keeping your infrastructure from falling apart! Imagine you spent 2 hours setting up a container, installing 15 different tools, and then you accidentally delete it. How do you get it back? If you did it manually, you have to remember every single command you typed. Here comes the #Dockerfile which is a text document that has all the command line instructions needed to assemble an image Docker reads the file and builds the #image automatically. A Dockerfile follows a predefined format and set of instructions. A #Docker image is made up of read-only layers, each representing a Dockerfile instruction. The layers are stacked, and each one is a delta of the previous layer’s changes. Dockerfile consists of specific commands that guide you on how to build a specific Docker image. FROM — Creates a layer from the ubuntu:18.04 PULL — Adds files from your Docker repository. RUN — Executed when image is being build. CMD — Specifies what command to run within the #container.

What if you could control your feature flags directly in your code editor? With the PostHog Extension you can! It gives you live intelligence into your PostHog SDK, showing feature flags rollouts, smart event completions and session replay data. Available in the VSCode and Cursor extension marketplace 🦔 https://lnkd.in/d6Xhtrvy

Debugging workflows are often slowed down by constant context switching. With Appwrite Sites, build logs are now available directly in the CLI. You can view logs from your terminal without leaving your development environment. This keeps your workflow uninterrupted and makes debugging faster and more efficient. Watch till the end for a small bonus 👀

❤️ Give rtk a little love! It's a proxy that sits in between your coding agent and your command line and reduces token usage of CLI call responses by stripping out unnecessary details. Windsurf, for example, changed from credit-based billing to token-based billing around two weeks ago - which increased prices something between 2x to 4x. At work, we are trying to mitigate this to avoid cost explosion. rtk can be a factor here. It's still being expanded to cover more CLI commands but when this is stable, you should strongly consider making it a go-to dev tool. https://lnkd.in/d5Cp-dNe