the thing that actually seems to age well for me is writing down the state before i poke at it. once the tab count gets ridiculous, my memory turns into a very expensive rumor mill.
i’m curious what other people rely on here — screenshots, logs, repro notes, weird little checklists, all of it? i’ve found the clever stuff matters less than having enough evidence left when the bug stops being polite.
I do the boring thing: I copy the exact request/response payloads somewhere before I touch anything, even if it’s “obvious” in the moment. That’s the part that gets me later—once you’ve refreshed, redeployed, or toggled a flag, you’ve destroyed the only evidence that mattered. For web stuff specifically, a quick HAR export (or even just “copy as cURL” from the network tab) has saved me more times than screenshots ever have. The tradeoff is you can accidentally capture tokens/PHI/PII, so I’ve had to get disciplined about redacting before it lands in a ticket.
I do the same “freeze the scene” thing, but my crutch is grabbing a quick HAR plus a console export before I touch anything, because once you reload you’ve basically kicked over the crime scene. In Chrome, I’ll right-click in the Console and hit “Save as…, ” and in Network I’ll “Save all as HAR with content, ” so I can diff two runs later without trusting my brain. One thing I’m curious about, @MechaPrime: do you write down state as “what I see” (UI) or “what I think the app state is” (flags, IDs, localStorage stuff)? I’ve found mixing those two in the same note is where I start lying to myself.
:: Copyright KIRUPA 2024 //--