Maxim Anatsko’s Post

"Preserves your macros" sounds like a marketing promise. Let me tell you what it actually means technically — so you can evaluate it yourself. A VBA macro in an .xls file is stored in a binary stream called `vbaProject.bin`. When you convert to .xlsm, this stream needs to be extracted, re-serialized, and embedded in the new format's Open XML container — without corrupting the bytecode or breaking module/procedure references. On top of that, many .xls macros have 32-bit Windows API declarations

Three things shipped to Claude Code recently that nobody put on a marquee. PostToolUse hooks now expose duration_ms — actual tool execution time, excluding permission prompts and the PreToolUse hook chain. PostToolUse hooks can also replace tool output across all tools now, not only MCP. Plugin MCP dedup fixed servers stuck "connecting" when an unauthenticated claude.ai connector got duplicated. PowerShell tool commands can be auto-approved in permission mode, matching Bash. None of those are headline features. All of them are the difference between a tool you keep using and one you abandon. I run Claude Code against the Laaimylorrie codebase from a Windows box. PowerShell auto-approve on its own saves me roughly forty manual confirmations per long session. The duration_ms field finally lets me build a real "slow tool" telemetry view — I can see which MCP server is the actual bottleneck without instrumenting it externally. The ergonomics-vs-capability split is real. Claude Opus 4.7 is the headline. The reason teams stay is that the CLI keeps removing friction one undramatic patch at a time. What's the smallest Claude Code change that has changed the way you work in the last month? #ClaudeCode #DeveloperTooling #AnthropicAI #BuildingInPublic

🚀 Hi All and Anthropic. Erick D. Machado Alvarez and I — we just open-sourced a small tool we built for our own workflow: claude_widget_win — a lightweight Windows overlay that monitors your Claude Code usage in real time. It sits in the corner of your screen and shows you: 📊 Token consumption (input, output, cache) and estimated cost 📈 Session and weekly usage limits with reset times 🧠 Active model and context window breakdown by category Built with PyQt6 for a clean, frameless UI with per-pixel transparency. Under the hood it reads Claude's local JSONL session files directly and spawns PTY sessions to pull /usage and /context from the CLI — so there's no API key needed and no extra backend. Free, MIT licensed, and open to contributions. 👉 https://lnkd.in/eT3vGQ6x #ClaudeCode #Python #OpenSource #DeveloperTools #Productivity

Three details from Claude Code 2.1.117–2.1.121 that most coverage will gloss over: 1. The Opus 4.7 context bug wasn't just cosmetic. Auto-compact fires when /context percentage hits a threshold. Computing percentage against 200K instead of 1M means Claude was summarizing your session 5x too aggressively, losing detail you didn't need to lose. If you've felt sessions "forget" things faster than they used to, this is why. 2. Native builds on macOS and Linux replaced the Glob and Grep tools with embedded bfs and ugrep, available through the Bash tool. The win is the eliminated tool round-trip – Claude doesn't context-switch to a separate search tool for what's effectively a `find` or `grep`. 3. The find file-descriptor crash on large directory trees was a host-wide stability bug, not a Claude Code-specific bug. macOS and Linux native builds were exhausting the user's open-file limit during recursive search, taking down other processes. 2.1.121 reduced peak FD usage during find significantly. Full breakdown:

One of the projects I've gotten to work on while I've been back has been focused on the Windows UI stack. After Mr. Snover's most excellent post around UI Frameworks, this felt like a good way to join the conversation 🙂... First, the team has been hard at work at shoring up the fundamentals of WinUI (we are dropping the "3"), improving performance, compatibility, interoperability, and working on filling out the set of controls. More to come on all of those. Second, in parallel a small group has been working on a modernized programming model in C# and Rust for building apps in Windows using the same WinUI foundation. We just flipped the repos to "public", as we intend to do the development in the open - these are definitely experimental, expect a lot of churn, we will be adding/removing/changing things on a daily basis. class CounterExample : Component {   public override Element Render()   {     var (count, setCount) = UseState(0);     return VStack(12,       TextBlock($"Count: {count}").FontSize(20).SemiBold(),        HStack(8,          Button("- 1", () => setCount(count - 1)),         Button("Reset", () => setCount(0)),         Button("+ 1", () => setCount(count + 1))       )     ).Padding(24);   } } Repo (C#): https://lnkd.in/gnuQD4_k Repo (Rust): https://lnkd.in/gwzH9-bk

📊 SQL Window Functions Practice 🚀 Recently, I worked on SQL Window Functions and explored concepts like: ✔️ OVER() ✔️ PARTITION BY ✔️ ROW_NUMBER() ✔️ RANK() ✔️ DENSE_RANK() ✔️ CASE Statements ✔️ JOIN Operations These functions are very useful for advanced analytics, ranking, reporting, and business insights. Through hands-on practice, I improved my understanding of: 🔹 Data Analysis using SQL 🔹 Employee Salary Ranking 🔹 Partition-based Calculations 🔹 Query Optimization & Reporting Logic 💻 GitHub Project: https://lnkd.in/dfDAMXzD I’m continuously improving my skills in: SQL | Python | Power BI | Excel | Data Analytics #SQL #WindowFunctions #DataAnalytics #Python #PowerBI #GitHub #LearningJourney #BusinessIntelligence

If you are keen to try your hands at debugging but facing technical challenges, then try our MEX extensions for debugger. It can help you simplify common debugger tasks, and provides powerful text filtering capabilities to the debugger. Tip: For a full listing of commands type: !mex.help https://lnkd.in/g2KpkbVm #Windows #WinDbg #Troubleshooting #Developers #TechCommunity #WindowsDebugging #DebuggingTools #MEXExtensions #LearnDebugging

Tip of the day - Edge Collections - gone but hopefully not forgotten if like me you used this useful feature to organise research and made the mistake of relying on it this might help but the answer is technical I'm afraid .. 1. Scan your user folder under windows for collectionsSQLite. 2. if its there get SQLLite browser (https://sqlitebrowser.org/) 3. run this sql view SELECT   c.title AS collection_name,   i.title AS item_title,   i.text_content AS item_description,   json_extract(i.source, '$.websiteName') AS website_name,   json_extract(i.source, '$.url') AS url FROM collections c JOIN collections_items_relationship cir   ON cir.parent_id = c.id JOIN items i   ON i.id = cir.item_id ORDER BY c.title, i.date_created; 4. copy paste or whatever - I put mine into OneNote table which hopefully will be around longer than me. thanks for reading

💡 Daily Tech Tip - Batch Rename Files for Manageability If you need to rename a large number of files, instead of doing it individually, you can rename them in bulk. On your Windows PC, go to the folder with files you want to rename. Select all the files, right-click on the first file and choose 'Rename'. Enter the name you wish the files to have and press enter. All the selected files will be renamed in sequence with that new name followed by an incremented number. This is a great time-saver strategy, especially for large data sets and photo folders. 🟢 #Intermediate | #Files #TechTip #DailyTip #Technology

What's actually inside a Windows shell shortcut? Pavel Yosifovich just published a deep dive into .lnk files — what they are, why they're binary COM objects, and how to work with them programmatically in C++. He covers the IShellLink and IPersistFile interfaces, walks through loading and modifying an existing shortcut with code, and explains how to use OleView to discover what interfaces a COM class actually implements. There's also a practical security section on .lnk persistence — why it works, and what to watch for defensively. Free on the TrainSec knowledge library: https://lnkd.in/gkZgeGkp