A while ago I watched this talk from Lucas Meijer on how he develops with Pi. Although I haven't really used Pi for anything, something I did take away from this video was the terminal emulator he was using. It's called [Supacode](https://supacode.sh). I really was intrigued by it's operating paradigm. Supacode (gosh, I really don't like the name though. So cheesy!) Is built around worktrees. Worktrees have become all the rage with coding agents, but I'd already been using them for managing multiple feature branches before all this AI stuff came along. I loved having such clean separation between what I was working on. If a P1 bug was reported in the middle of developing a feature, pivoting was trivial, and it didn't require futzing around with git stash, and accidentally losing my working changes because I couldn't remember it's position in the stash. Worktrees let me just keep all the working changes in the branch untouched, because I will just fix the bug in another branch. Another…
No comments yet. Log in to reply on the Fediverse. Comments will appear here.