3 ms·
Pretty sleek! The fact that you have to manually add folders to watch destroys the magic a bit IMO. Would it be possible to just watch the whole user folder and
by merl1n 7mo ago
Pretty sleek! The fact that you have to manually add folders to watch destroys the magic a bit IMO. Would it be possible to just watch the whole user folder and what would be the performance implications?
Also can it differentiate between a user renaming a file and a potential background job programmatically modifying files? I could imagine this leading to some weird results when multiple applications are "fighting" over the same file ...
- imfavourite 7mo agoThanks! Glad you like it. Regarding watching the whole user directory: surprisingly, performance isn't an issue at all. macOS handles file system events very efficiently, so watching ~/ wouldn't cause any noticeable overhead. The decision to make it manual is purely about security and user control. I wanted a strict "opt-in" approach so users know exactly where the app is operating, preventing any unexpected accidental conversions in sensitive folders. It's all about the principle of least surprise. As for background jobs vs. manual renaming: Morpholder is designed to be highly specific. It only triggers when a file is renamed to a supported target extension (e.g., specifically to .jpg or .mp3) or when one of the exact keywords (like _nobg or _min) is appended. Because the trigger conditions are so narrow, the chances of conflicting with a random background script are very low. That being said, you bring up a really good point about edge cases with file locking and multiple apps interacting simultaneously. I'll definitely run some additional stress tests on those scenarios. Really appreciate the feedback and food for thought!