Open Source Software Community - it's about ethics in Code of Conducts

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
Can anyone think of any "convert some ancient shit to Rust for no reason at all" projects that actually made anything better?
 
AnOminous said:
Can anyone think of any "convert some ancient shit to Rust for no reason at all" projects that actually made anything better?
Click to expand...
Uhm, ackshuakly all of them, chud? Because Rust is newer, shiny and better. Stop questioning things and bitching about future progress, bigot.
 
AnOminous said:
Can anyone think of any "convert some ancient shit to Rust for no reason at all" projects that actually made anything better?
Click to expand...
I dont think so, I stand with vaxry the hyprland dev on this:
vaxry.jpg
 
AnOminous said:
Can anyone think of any "convert some ancient shit to Rust for no reason at all" projects that actually made anything better?
Click to expand...
the only one i can think of that comes close is ripgrep and that's not really a conversion, though it is meant to replace things like ack (heh). and ripgrep itself is really only useful in niche cases at best.
systemd should be rewritten in rust entirely actually, that'd keep these losers busy for a while.
 
skunt said:
the only one i can think of that comes close is ripgrep and that's not really a conversion, though it is meant to replace things like ack (heh). and ripgrep itself is really only useful in niche cases at best.
systemd should be rewritten in rust entirely actually, that'd keep these losers busy for a while.
Click to expand...

The vibe of this thread is that C can be easily written in a memory-safe way if you just git gud but even the best of us make mistakes in projects too big to fit in the human brain easily, and they can be especially hard to spot if you weren't the one who designed the project. Rewriting the coreutils in Rust is pointless at best, but something as big as systemd would actually be a project that would benefit from being written in a memory-safe language. Unfortunately, rewriting a project like that would take actual work (whereas releasing half-assed broken rewrites of relatively simple software like the coreutils is something any transsexual with a keyboard can do), so I'm not holding my breath for that to happen.

I have a conspiracy theory I half-believe in that systemd, a project of enormous scope, is intentionally written in a language like C to enable a security flaw or two to slip in "accidentally." If you were the NSA and wanted to introduce a vulnerability into a common enterprise OS, you probably wouldn't target the Linux kernel very seriously, since the kernel changes so much that even if you do get a vulnerability in it could be removed within an update with no one even noticing. But PID 1 is the next best thing, there's very few people actually reading and studying the entire source code like there are the kernel, C is an easy language to slip up in and any discovered flaw could seem accidental, systemd is pushed with the weight of Red Hat (whose #1 customer is coincidentally the U.S. government), and it's such complex software with so many dependencies and interlocking parts that you could easily design an exploitable bug that would be pretty hard to test for.
 
dec05eba said:
I dont think so, I stand with vaxry the hyprland dev on this:
View attachment 8157190
Click to expand...
How fucking DARE you to even MENTION Hyprland. That is a racist, bigoted, misogynist, xenophobic, Nazi, islamophobic, transphobic, anglophobic, arachnophobic, hippopotomonstrosesquippedaliophobic piece of shit. I WILL ban you from ALL my repo's, BLOCK you on Bluesky and want to never see your disgusting, hateful face ever again. I hope you think before you post next time or suffer the consequences of your actions.
 
1762932529928.png

Why in the fuck are these advertisements for this electronslop cancer at the top of everyone's fucking readmes? Before they even tell me what the fuck the thing I'm looking at does? A readme should not be a goddamn advertisement!
It's fucking everywhere! Even the vim repository has this shit in it
1762932756333.png
1762933073722.png
 
Cunted said:
View attachment 8158289
Why in the fuck are these advertisements for this electronslop cancer at the top of everyone's fucking readmes? Before they even tell me what the fuck the thing I'm looking at does? A readme should not be a goddamn advertisement!
It's fucking everywhere! Even the vim repository has this shit in it
View attachment 8158302
View attachment 8158310
Click to expand...

Checks out
IMG_3689.jpeg

Would you think the AI bubble is going to pop somewhere soon? People are getting really tired of seeing it everywhere where it isn’t supposed to be. Maybe I just have algorithmic bias where I only see fellow people hating it instead of a general populous on it. All the faggots with Clippy profiles sure does seem to align with that theory.
 
Cooked Rabbit said:
best of us make mistakes
Click to expand...
and the vibe you get from the rusty trannies is that with rust, you cannot make mistakes no matter if you are "gud" or not, which is just patiently untrue. And that kind of programmer usually has the attitude of "move fast, break things" because he doesn't understand why systems are the way they are and has a lot of dunning kruger going on in thinking he can do it better, which encourages things like an extensive 3rd party library system (e.g. cargo) because the vibe coder cannot be bothered to figure out mathematically if a number is even or odd and then Chen somehow gains hold of one of these tiny repositories that carry 80% of all projects and suddenly you have a backdoor in all systems. Does that sound safe to you? How can you be confident of the safety of a given program if half of the code written is not even be written or vetted by the programmer? And that's just one aspect of how rust can be just as unsafe in practice.

It shouldn't even be argued that being good at programming and fully understanding the theory of operation of computer systems and given operating systems isn't some kind of prerequisite to writing good and safe programs and it's frankly ridiculous that this at the heart of it all is being done in so many places. Also, the implication that it's impossible for a good programmer to write a safe program in a given environment says more about the environment then about the programmer, me thinks.

Hakase Shinonome said:
Would you think the AI bubble is going to pop somewhere soon?
Click to expand...
What makes you think the bubble popping is going to make the technology go away? The personal computer and internet bubble popped too, didn't make these technologies disappear either. I hear this often in combination with AI, that the bubble bursting and a bunch of companies/investors going broke on dead-ends of the technology will somehow magically remove AI stuff from the world. Never happened before and not gonna happen here either. The companies go all-in here because AGI will be very unlike every technology breakthrough humanity ever had and will make the people in control of it very powerful. It is a bubble but I don't suspect it going to end anytime soon and even if it does, it will not be the end of AI. The underlying technology, even in it's current state, has already proven too useful for that and is already too integrated in too many things, even if people might not be aware of it.
 
AmpleApricots said:
What makes you think the bubble popping is going to make the technology go away?
Click to expand...
Correct. The bubble is not going to pop.
AI is here to stay.
Very junior dev can very cheaply produce a LOT of code that looks good at first glace
so they will do that. So companies will go hard all-in for vibe coding.
Because having a cheap jeet produce thousands of lines of code a day is super attractive from a $/line metric. The SVP that decrees All code shall be vibe coded might even get a bonus due to how much cheaper he made the code to produce.

The cost of this is externalized, as always.

Unskilled and low-cost jeets already today write code they do not understand, or even attempt to try to understand. Today they just cut-n-paste from stack-overflow but it then requires a minimum amount of skills to take that code and integrate it into your project. You at least have to get it to compile in the surrounding code.
With AI the jeet from fiverr don 't even need to do that. Just click the AI button then click commit because the code most of the time will compile.

Two things will come from this:
1, Junior developers now become just AI click button monkeys so they have no opportinity anymore to hone their skills and eventually become senior developers.
A whole new class of data-entry junior developers that will never become skilled or senior.
2, Senior developers will transition from being architects or making the overreaching designs to instead do careful code review all day, every day.
Identifying and fixing the subtle bugs in the AI code.
It is going to suck being a senior dev going forward. Your primary job will basically be to go in and cleanup and rewrite AI slop. And there will be a near infinite amount of AI slop so that is what you will do, all day, every day.
 
skunt said:
the only one i can think of that comes close is ripgrep and that's not really a conversion, though it is meant to replace things like ack (heh). and ripgrep itself is really only useful in niche cases at best.
systemd should be rewritten in rust entirely actually, that'd keep these losers busy for a while.
Click to expand...
Also fd. I use that all the time. Probably more often than rg. I use those when I will notice a speed difference over the old implementations.
 
Angie Wormwood said:
I dug into it and found out that it basically allowed for random memory extracts. Considering that you have billions to trillions of memory registers in a system, the chances of getting something useful when executing it is basically zero.
Click to expand...
"Repeatable random contiguous 64KB extracts from somewhere near the library that happens to handle communication security for all your clients" is a bit more dire than a random register.
 
Hakase Shinonome said:
All the faggots with Clippy profiles sure does seem to align with that theory.
Click to expand...
I've seen AI-generated Clippy PFPs, even Louis self-admittedly uses chatgpu. My personal opinion is that Redmond and all the metastases of this plague should be glassed over, but I don't think the Clippies are gonna be of much help - feels like a modern day analog of the facebook PFP frames.
 
AmpleApricots said:
and the vibe you get from the rusty trannies is that with rust, you cannot make mistakes no matter if you are "gud" or not, which is just patiently untrue. And that kind of programmer usually has the attitude of "move fast, break things" because he doesn't understand why systems are the way they are and has a lot of dunning kruger going on in thinking he can do it better, which encourages things like an extensive 3rd party library system (e.g. cargo) because the vibe coder cannot be bothered to figure out mathematically if a number is even or odd and then Chen somehow gains hold of one of these tiny repositories that carry 80% of all projects and suddenly you have a backdoor in all systems. Does that sound safe to you? How can you be confident of the safety of a given program if half of the code written is not even be written or vetted by the programmer? And that's just one aspect of how rust can be just as unsafe in practice.

It shouldn't even be argued that being good at programming and fully understanding the theory of operation of computer systems and given operating systems isn't some kind of prerequisite to writing good and safe programs and it's frankly ridiculous that this at the heart of it all is being done in so many places. Also, the implication that it's impossible for a good programmer to write a safe program in a given environment says more about the environment then about the programmer, me thinks.
Click to expand...
Sure, Rust has shortcomings, but memory safety isn't intrinsically a bad goal to work toward. I don't think it's unrealistic to say that large codebases written in languages like C will occasionally have memory-related bugs accidentally introduced. Once projects grow past the ability of a single mind to keep all its relevant pieces in mind at once, I think do think this is somewhat inevitable unless you're extremely strict and diligent about how you write software -- much more so than a typical business would consider worthwhile, and more than even very good hobbyists would typically consider worth their time.

But that enormous and complex software is difficult to maintain and modify is part of why the Unix philosophy became popular in the first place, and a big part of why people criticize systemd. It's easier for humans to understand and achieve complex behavior built from small, individually comprehensible components than monoliths existing only unto themselves. Unfortunately corporate interests have largely taken over computing and people sometimes need to use stuff like systemd, and virtually everyone needs to use web browsers, which because of their fundamental goals and shit design of the Web, rather than the skill of their programmers, are introducing and fixing new bugs virtually every single day. Given that we live in such a world already and can't just fix the direction of modern computing, having compilers and languages that catch subtle bugs seems like a good thing to have. I agree that the culture and ecosystem of Rust is awful, but it's nothing new; stuff like Python and JS have the same problems. So though I write most of my personal projects in C, and I hate the way stuff like the coreutils rewrites or Debian introducing a dependency on Rust is being pushed down people's throats, I do think Rust or efforts like it have a place for big corporate software -- not because big corporate software is good, but that it causes the kind of problems Rust is trying to solve.
 
You must log in or register to reply here.
Back
Top Bottom