Prune

Prune Devlog #2: Building a Crash-Proof Clipboard Engine

While Prune v1.1.0 worked great in isolated tests, everyday computer use quickly exposed a massive flaw: copying a file or taking a screenshot caused the background thread to crash due to unhandled non-text exceptions. To fix this, Prune v1.2.0 introduces defensive error handling that bypasses binary data like images and files, ensuring stability. Alongside this crash-proofing, I completely optimized the engine by moving the polling interval to a battery-friendly 0.3 seconds, from the faster 0.2 seconds, rewriting the URL pattern to prevent CPU spikes, and ensuring the script only triggers a clipboard write if tracking data was actually removed. Finally, I expanded filter definitions to aggressively strip out modern, invasive referral tracking from platforms like Amazon, Twitter/X, and TikTok. I also added a setup script on Github instead of the portable version that I had in v1.0, so that it is easier to use. That time didn’t track on Wakatime though, so I’ll say it took around 30 more minutes.

Prune began as a small utility to make sharing links easier by automatically removing tracking parameters from URLs copied to the clipboard. After building the URL cleaning logic, I added background clipboard monitoring, packaged the project into a standalone Windows executable with PyInstaller, and integrated a system tray icon so it could run quietly in the background. Version 1.0 delivers the core functionality, with plans to add more customization and features in future changes.