The best debugging habit is leaving a clean trail

talk
1

I’ve been thinking the habit that scales isn’t “be fast,” it’s “leave yourself a clean trail back to the bug.” When I come back to a weird issue the next day, the thing that saves me is usually a tiny repro, a note about what I already ruled out, or even just a screenshot of the broken state.

That feels way more durable than heroic memory. What do people here actually keep around so the next round of debugging doesn’t start from zero?

2

I get saved by one dumb thing: I paste the exact command and inputs that triggered it (or the curl request) into the ticket/notes, even if it feels obvious in the moment. Second best is logging the “known good” baseline right next to the broken one (versions, flags, env vars). Half the time the bug is just “you weren’t actually running what you thought you were, ” and future-you needs proof.