pull down to refresh

Yeah. Its very tempting to hit "merge and fuck y'all"

What helps me in moments of "I don't want to read all that", is to just break it up. Ask the llm to break it up if its huge and you don't wanna.

112 sats \ 5 replies \ @k00b 5h

Good rec. In this case the change is small by line count but delicate. Best use of an LLM in that case (I've found) is asking for a review matrix so I have a little more working memory.

reply
101 sats \ 4 replies \ @optimism 5h

I do that a lot actually. Also cov (in verbose no-hit mode, not coverage %) is helping me figure out what is not looked at in unit test suites. Too bad its all open source because I'd go long test framework stonks.

So then I can choose to review tests or untested code based on mood. When doing refactors, test regimen has helped a lot. Especially when you get something that smells a bit sloppy. Instead of asking to build more code, I ask for some tests and then I get "oh btw I also found 3 bugs"

reply
112 sats \ 3 replies \ @k00b 5h

I find myself wanting automated tests now that I'm not writing as much of the code. Tests can create a false sense of security (not unlike vibe coding) is my main gripe with them, but I should be able to get over that.

edit: what i mean - tests are certainly better to have than not all being equal. but that all being equal part is where things can break down ime.

reply
101 sats \ 2 replies \ @optimism 4h

The nice thing imo about unit tests is mainly to protect against regressions. The worst thing that ever happened to me though, and this was with humans, was that someone deleted test cases in a big update and I missed it in my review. This came back to bite me a few years later. So I'm super allergic to there not being tests, and to changes to tests without explanation. Especially with the bots.

reply
112 sats \ 1 reply \ @k00b 4h

I create my fair share of regressions so I want/need tests. But I also want/need to keep max fear so I'm forced to compress and make sense of the code constantly.

reply
101 sats \ 0 replies \ @optimism 4h

Yeah. I get that. For meatcoding I've often felt the tests were for other people, not myself. Until the codebase becomes big. My memory is leaky, I guess, so then its largely that I don't want to forget about interaction between component x and y.

With a team, or nonmeat coders, I will mess up at least once a day and I find my mistakes the costliest, but that's probably because I'm not letting go of control much

reply