That code doesn’t really do anything that has changed in multiple recent versions of zig.
Curious how close it would get if you asked it to make it more idiomatic zig. Mostly thinking of the write callbacks being changed to use writers.
Relevant quote from the recent Bun is joining Anthropic post:
Over the last several months, the GitHub username with the most merged PRs in Bun’s repo is now a Claude Code bot. We have it set up in our internal Discord and we mostly use it to help fix bugs. It opens PRs with tests that fail in the earlier system-installed version of Bun before the fix and pass in the fixed debug build of Bun. It responds to review comments. It does the whole thing.
Im guessing robobun? This is probably more of an Ad given the announcement but still interesting to see what others are doing in this space…
Damn that’s clever. Google just got backlash for “contributing” to innocent open source projects. Here they instead pay the developers to help train their ai bot/agent into a proven useful tool. On top bun as a platform will let them make a vibe coding foundation that might actually be useful for non technical people.
Unit tests are a pretty good use case for AI. Nobody cares if the test is written like slop as long as it fails in the old code and passes in the new code. And, if it’s AI generated, you won’t feel any remorse for taking the test out to the woodshed and shooting it after it exists for a couple of releases.
Of course, this is also the kind of task that you used to assign to junior devs in order to bring them up to speed on the system architecture, programming language, build/CI/CD system, etc. So, that’s a downside.
I don’t completely agree. More than once I have examined test cases to dive into a code base. I would rather read clear, well-structured testing code, than slop.
4 Likes
Same same.
I jump to reading test cases first, because my brain needs to see examples of use before reading the docs most times.
1 Like