I have no problem with Rust But redoing all of this is going to end up shattering Linux into even more bits, and I'm all for it except for the fact that I fear support will then weaken further. Because once one thing's redone why not redo another? (And another, and another, and so on) - ThreadSky

novaaaaaaa.bsky.social • 9 days ago

I have no problem with Rust
But redoing all of this is going to end up shattering Linux into even more bits, and I'm all for it except for the fact that I fear support will then weaken further.
Because once one thing's redone why not redo another? (And another, and another, and so on)

Reposted from Zorin =^o.o^=

Damnit Ubuntu, stop trying to recreate software that works well and has for decades because your developers have an itch in their underpants to reinvent the world.

There's a reason coreutils has endured for decades. They work. They've been vetted. Stop this.

Comments

transgirlonfire.bsky.social•9 days ago

I mean…. I think systemd would be another prime target for a rewrite…

treeshateorcs.bsky.social•9 days ago

some things in systemd/ are already in rust - like this https://github.com/systemd/zram-generator

transgirlonfire.bsky.social•9 days ago

Damn, I actually did not know that. Thanks, I‘ll have to check it out!

sussus.guhcat.com•8 days ago

i suppose if nothing else it supports the most recent no boilerplate video about rust becoming the new C...

sussus.guhcat.com•8 days ago

it seems like the support problem is more of a general OSS downside tho (as few as i'd say there are), no? or is that not really the case? i fear i've not been in the OSS community for too long so i've not really had the chance to see how things play out much :p

nekotachi.bsky.social•9 days ago

You know GNU coreutils aren't the original coreutils, right? they're not even the one ones commonly used on linux, busybox has been a thing for for a while. not to mention bsd coreutils, or any of the others. i think there's room for one more.

novaaaaaaa.bsky.social•9 days ago

Yes but main concern

nekotachi.bsky.social•8 days ago

I would argue software diversity is a good thing, actually.

Hell, one common criticism of rust is the lack of alternative compilers (despite the fact other compilers do actually exist)

drankinatty.bsky.social•9 days ago

The rust fad will fade like the 20 before it. C/C++ were never the problem, it was C/C++ programmers that can't count or can't protect allocation bounds. A rewrite is never as good as the original. A 5MB rust "Hello World" does little to impress. C/C++ will be better for the focus on memory safety.

hjvt.dev•9 days ago

But have you considered that to err is human?

drankinatty.bsky.social•8 days ago

Deeply. One of the axioms I lived by is "to err is human, but to really foul things up requires a computer..." (choose your favorite 'f' adjective) And, admittedly, that is something almost always used in jest.

nekotachi.bsky.social•9 days ago

Yeah, I'm sure Rust will fade into obscurity just like C++. Or C before it. Who even needs macros in assembly? Who even needs assembly? What's even the point of memory protection, seems like a waste of hardware. What's even the point of a computer, log books work just fine.

thalesfragoso.bsky.social•9 days ago

"It was C/C++ programmers that can't count or can't protect allocation bounds."

I guess the programmers behind chrome and android are pretty shitty then /s

drankinatty.bsky.social•8 days ago

I see the argument for both sides. As codebases have grown from tens of thousands to tens of millions of lines, you no longer have one hand or a small group of programmers involved. A failure to validate some bound in a struct member can surface years later. Memory safety has promise there.

Comments

Posting Rules

Reply