59
you are viewing a single comment's thread
view the rest of the comments
[-] hellinkilla@hexbear.net 4 points 6 days ago* (last edited 6 days ago)

Maybe I should make a new thread. But what is the communist line on rust?

In Linux I have seen there is controversy with some rust being introduced in a foundational way. Replacing C. some people say it's trash. Some people say its "safe" and then others dispute that.

In terms of little tools I do find the rust ones are nice to have on hand. Ag silver searcher is much faster than grep. But grep being more mature and feature complete I would never get rid of or replace it. That is probably a question of modernity vs hours spent on it. If both grep and ag started from scratch in 2025 idk how it'd work out.

Is rust like some sort of neoliberal intervention? That's kind if the vibe I get but I can't justify it or contemplate any adverse outcomes.

[-] Kopfrkingl@hexbear.net 6 points 6 days ago

The average Rust rewrite (of GPL projects) very often sets the license to MIT/BSD which is slightly exasperating. Other than that I'm indifferent to it.

[-] hellinkilla@hexbear.net 2 points 5 days ago

I guess that's nice for corpos who would like to integrate them.

[-] doodoo_wizard@lemmy.ml 5 points 5 days ago* (last edited 5 days ago)

Rust/js/python are the mill for programmers. They make the programmers work output increase but are being used by capital along with a bunch of other technologies to change the structure of technology and society in a way that’s good for capital and “good” for the people who don’t write software but harms those who do.

Divorced from the marketplace, rust people are generally just annoying and unwilling to accept being beholden to any larger project they want to “contribute” to. Theres a bunch of reasons for this, it’s not like people with the wrong skull shape choose rust or js at some pivotal point in their brains development and go on to behave in the wrong way because they’re incapable of anything else.

On the one hand they’re taught both that these languages are acceptable and valid to work in (which is true) and that a phenomenal way to pad one’s resume is to have worked on a bunch of open source stuff (which was once true) so now when they show up with a bunch of commits that the rest of the project doesn’t understand or have time to learn a whole new language in order to understand and get told to fuck off until someone has the time to deal with them it’s both an attack on their skills and abilities and on their future prospects.

Anyway the structure of how big maintainers of open source projects are expected to be compensated is changing and the old maintainers material interests are aligned against the new contributors material interests and they don’t have a shared labor to bind them let alone some trade organization and the new programmers are literally under socialized and have spent sixteen years without any arts or literature or history education but I’m tired of typing you read capital volume 3 I’m sure it’s clear.

[-] Imnecomrade@hexbear.net 4 points 5 days ago

Divorced from the marketplace, rust people are generally just annoying and unwilling to accept being beholden to any larger project they want to “contribute” to.

This applies to nearly every software project under capitalism.

[-] doodoo_wizard@lemmy.ml 4 points 5 days ago

You are correct however I think there are unique conditions that rust/js/python devs are living under that makes it more prevalent.

[-] invalidusernamelol@hexbear.net 2 points 5 days ago

Why is Python being lumped into this? It's used in almost every project for managing builds and environments. Also really good for CLI/API stuff

[-] doodoo_wizard@lemmy.ml 4 points 5 days ago

Massive institutional presence in education, mit/bsd license culture, unexamined reliance on package/library manager.

I’m not mounting a criticism of the languages themselves, their usefulness or goodness (whatever that means), just using them to triangulate on the cohort of developers who are in a particular situation.

[-] invalidusernamelol@hexbear.net 3 points 5 days ago

I can see it, there's definitely a type of Python dev that just uses jupyter and numpy/Pandas only.

[-] hello_hello@hexbear.net 5 points 6 days ago

Taking communist to mean materialist:

Rust is CPP but RAII memory model is enforced by the compiler and there is no C subset (just foreign function interface). It became popular because it basically absorbed a lot of saavy high level language constructs into something that could compile to machine code rather than needing a runtime on the system. In the same vein it became very popular because the tooling was all first party (linting, package management, static analysis).

The compiler requires LLVM (which is complicated on some platforms), and there are no shared libraries so every rust program is statically compiled with hundreds of libraries and so compilation takes much longer. There is gccrs which is an effort to port the language to GCC.

Now my personal opinion TM

I dislike rusts large dependency tree for even trivial programs and its npmification. The lack of ABI and reliance on static compilation is also a deal breaker for me.

But also I feel like rust programs generally tend to be of higher quality and they actually are great at porting and claiming what they preach. Rust basically disproved Javas insistence on a virtual machine runtime and showed that compiling to machine code was still always possible.

I just wish everyone used lisp instead.

[-] ziggurter@hexbear.net 4 points 6 days ago* (last edited 6 days ago)

Some people say its "safe" and then others dispute that.

It's not safe. It just makes it easier to write safe code. Important difference. Just like when people claim you "can't have a memory leak in a garbage-collected language", they are also full of shit; all you have to do is keep a reference to something you never plan to use again, and you have as bad of a memory leak as in any environment where you can forget to call free/delete.

I'm also using "easier" here with mild sarcasm. Sometimes it's difficult to get the constructs right in Rust so that shit will compile. Once you do, it's less likely you'll do stuff like leak memory than if you get a program in another language to compile. So "easier" is really more like shorthand for "somewhat harder, most of the time, to let your brain worms escape to the runtime stage" (i.e. production).

[-] footfaults@lemmygrad.ml 3 points 6 days ago

Is rust like some sort of neoliberal intervention?

No it's just the same annoying people that bandwagoned Golang back in the 2010's now are bandwagoning Rust. There's just a group of people that chase trends and are insufferable.

But what is the communist line on rust

Honestly I hope there isn't one. It doesn't really make sense to me to have a political stance weigh in on a programming language.

[-] ANarcoSnowPlow@hexbear.net 1 points 6 days ago

I'm not sure why it's pushed so hard other than I suspect the compiler is compromised or something by some alphabet agency.

If that's not it, I'm about to get real curmudgeonly. People tout it being "safe by design" and "better than c" because of memory safety being built in, etc.

I'm no rust expert, though I'm arguably a (embedded at least) c expert, which biases me to some extent at least.

My take is that for situations where memory safety was already critical, my understanding is that rust mechanisms would have to be bypassed anyway and the safety of C is ensured by processes proven over decades...

So basically it feels like the CISA people trying to push "modern languages with modern safety" either because they don't understand how we actually do things or because they want us to use it for another reason... Both of which are equally believable to me.

[-] hellinkilla@hexbear.net 2 points 6 days ago

What does it mean for a compiler to be compromised?

[-] ANarcoSnowPlow@hexbear.net 2 points 6 days ago* (last edited 6 days ago)

Theoretically speaking... It could be possible for the compiler to recognize certain patterns and inject arbitrary instructions into the compiled code of interest. If it were really smart it would probably be limited to some specific platforms of interest, be some otherwise harmless looking instructions, that might do something to allow consistent exploitation under some specific circumstances. I'm just spit balling here, I've not put much thought into this past "I'm sure there could be some nasty shit you could do if you wanted to."

Another option might be hiding some information about the author and the system doing the compilation in binaries.

You're trusting the compiler to convert human readable code into machine readable code. I suspect you could sneak some "unreachable" code in there or something, and if it doesn't look scary it'd be easy to write it off as a quirk of optimization or something.

Edit: I have no evidence this is being done or has ever actually been done. I'm just saying that it's theoretically possible.

[-] footfaults@lemmygrad.ml 1 points 6 days ago

Read "Reflections on trusting trust"

this post was submitted on 25 Dec 2025
59 points (98.4% liked)

technology

24149 readers
254 users here now

On the road to fully automated luxury gay space communism.

Spreading Linux propaganda since 2020

Rules:

founded 5 years ago
MODERATORS