Hacker News Reader

Comment by 1238127 | Hacker News Reader

[flagged]

If the original thing is improved by being written in Rust as everyone proclaims, then this would be a good thing. However, I have doubts that the years upon years of updates fixing odd behavior, and then forgotten about the whys&hows of those updates that the same issues will not be introduced in the Rust version.

> and then forgotten about the whys&hows of those updates that the same issues will not be introduced in the Rust version.

Ideally that's what test suites are for, although I'm sure some deviations/gaps will be caught by users. For uutils they are preserving all the edge cases, replicating the original behaviour.

There are benefits to rewrites (far less often in my experience), like doas replacing sudo in BSD distros and culling previous unused and insecure behaviour.

And in the face of gaps of the test suite there's no assurance that consistent output will be preserved across releases of the same application. This is somewhat mitigated by the slow release schedule that these kind of projects usually have, couples with a feeling of "being done" meaning that releases shouldn't have much green field feature development work. But still fixing one bug could cause a regression elsewhere.

You could find a lot of them with your favorite search engine. But you chose to post an inflamatory comment instead.

Not a rust zealot, but https://github.com/firecracker-microvm/firecracker/tree/main... comes to mind.

Reply on news.ycombinator.com