We’ve been searching for a memory-safe programming language to replace C++ in Ladybird for a while now. We previously explored Swift, but the C++ interop never quite got there, and platform support outside the Apple ecosystem was limited. Rust is a different story. The ecosystem is far more mature for systems programming, and many of our contributors already know the language. Going forward, we are rewriting parts of Ladybird in Rust.
Porting LibJS
I used Claude Code and Codex for the translation. This was human-directed, not autonomous code generation. I decided what to port, in what order, and what the Rust code should look like. It was hundreds of small prompts, steering the agents where things needed to go. After the initial translation, I ran multiple passes of adversarial review, asking different models to analyze the code for mistakes and bad patterns.
Actually not a terrible way to use AI. Hopefully the port goes well
There will be people defending ladybird no matter what huh
I read the comment to only refer to the way AI was used.
I came to say this sounds like a good example of how to use ai. Its going to be used but the big problem is folks who will take the output and say good enough.
Yeah, using AI to help you rewrite code (which is tedious) in a scenario where there are clear test cases (it either works or it doesn’t) is basically the ideal usecase.
…no… this is a bad idea. Especially in a browser, given the myriad safety and security concerns in play. This is an objectively bad way to make changes to the project. I can just about guarantee they’re going to introduce some horrid subtle flaws.
I may be misinformed but the article says
which I took to mean that the new code behaves 100% identically to the old code, like a byte-perfect decompilation. If that’s not the case and it’s just the same in their suite of tests then I’d agree with you.
It depends on how extensive (and correct) the tests are, which can be pretty difficult. There is also maintainability of generated code.