Comment by KRAKRISMOTT

I'm not the person you're replying to, but I don't really mind the kernel (or anything else) being written in C as such, but interacting with C from $other_language can be rather painful for various reasons. In Go the reason they do what they do is because integrating it with how the Go scheduler works is tricky, and getting rid of libc makes cross-compilation a lot easier.

So in short, it's a pragmatic thing more than anything else.

(Aside: that thread is a little bit outdated by the way, as I do believe there's a reasonably complete Rust coreutils implementation now. Not that I think Rust is a good fit for OpenBSD though – I wish people would stop conflating "safety" with "Rust".)

Wow, that discussion does not warm me at all to OpenBSD. A lot of those responses seemed way more abrasive than they needed to be.

It's interesting, though, that at least initially, the only examples of memory-safe languages were ones like Java or Python, which are both very high-level and come with a lot of runtime tradeoffs that would practically make them unsuitable to tools like ls or cat that need to be run quickly and a lot. (Also interesting: this never seems to be given as a reason not to use memory-safe languages.) Later on Rust and Ada are discussed, but it feels much more cursory, except in the final comment.

The big reason to use C seems to be that it provides a signal that the developer is competent - you need to be at least good enough to learn C in order to contribute to OpenBSD. And more specifically, if you would rather write code in a non-C, memory-safe language, you are probably a bad developer who should not contribute to OpenBSD. To me, this intuitively feels like quite a weak argument (are there so many people clamouring to contribute to OpenBSD that they need this kind of gatekeeping mechanism?) but it's pretty much the main one that gets repeated. The more convincing argument at the end is that, practically speaking, most of the benefits of memory-safe, low-level languages like Rust or Ada can be provided by static C linters/checkers, and so they don't add much value overall. This seems very much the opposite of what some other companies are saying, e.g. Google and Microsoft, which is that even with static analysis of C/C++, Rust is still dramatically reducing the number of memory safety issues they're running into.

I find it interesting that they don't bring up what seems to be the more obvious reasons not to bring in memory-safe languages, which are mainly the classic "rewrites are hard, here be dragons" arguments - the existing tools have been hardened through years of use, and rewriting them is more likely to add new bugs than remove old ones; new languages/tools means new knowledge which means having people on hand familiar with the new languages to understand what's going on; reimplantation tends to be slow initially, which eats resources and prevents ongoing maintenance and development; etc.

My main takeaways from that discussion are (a) I really hope I never have to work with the OpenBSD maintainers, and (b) their main arguments seem more about maintaining the exclusivity of their club than about the technical details of any transition.

It does matter. You don't have to link with the kernel, or support the C ABI. You're still tied to C with libc. It's especially bad on Linux because glibc is so awful.

Forcing libc has absolutely nothing to do with type safety. How do you think they invoke syscalls in libc? There's no more type checking of syscalls there than there is in Go's standard library. From a user's point of view it's identical, except you don't have to worry about endless glibc version errors with Go.