Thread 107161461

What makes Rust worse for the end-user?

>>107161461 (OP)
oh no, program I received has better memory safety noooooo

>>107161461 (OP)
>What makes Rust worse for the end-user?
Nothing. It's a programming language, it just happens that vast majority of projects using it aren't good.

>>107161461 (OP)
the mental illness that comes with using it

>>107161461 (OP)
The ensuing meltdown you receive from the dev as "she" responds to your bug report about her perfect, blazingly fast, safe, and fearless program.

>>107161461 (OP)
you're already using software written fully or partially in rust.
if you didn't notice, then it doesn't matter for when end_user==(you), positively or negatively.

>>107161461 (OP)
I hate when python packages that should be simple wheels have rust dependencies. Just make the whole thing rust at that point.

>>107161849

the file at /etc/portage/package.mask/disable_rust says otherwise xD

>>107161876
>supposedly a gentoo user
>thinks everything can be put behind a use flag, even when it's a hard dependency
cool story bro

>>107161940
>package.mask
>useflag

>>107162033
what browser are you posting these comments from?

>>107162293
lil nigga thought he got someone but I'm nta >>107161876 and unfortunately I have r*st in my gentoo system

>>107161617
End users don't make bug reports

>>107161461 (OP)
What makes monkey wrench worse for the end user?

>>107161535
Here is a (You) good xir!

Users don't have to see the fugly code so nothing really

>>107161461 (OP)
It sucks at abstractions

https://stackoverflow.com/questions/39541312/function-returning-a-closure-not-working-inside-my-filter

>>107161461 (OP)
>What makes Rust worse for the end-user?
Lack of backwards compatibility.
On Gentoo, I have four versions of rust-bin installed, because different packages need to be built with different compiler versions.

>>107161461 (OP)
Software that was working perfectly fine being replaced by incomplete Rust ports

https://askubuntu.com/questions/1557259/sort-command-ignores-locale-rules-for-capitalization-accentuation-and-special/1557262#1557262

Only one remotely good argument ITT so far. Damn this board's dead.

>>107164874
maybe because it actually has no drawback for the end user anon.

>>107161461 (OP)
no locale support

rust is a fine language
the worst part about rust are rust programmers (their evangelism, their desire to rewrite everything in C, their hatred of the GPL, etc)

>>107164916
locales are easy to support, and are supported by many rust projects.
it's the broken-by-design-and-implementation impossible-to-get-reliably-from-the-outside-of-internal-libc-functions locale info retrieval that is not supported.

>>107165034
all I need is correct sorting and date-time format
>many rust projects
I think you're lying

>>107165062
Oh, and case conversion

>>107165062
feel free to provide a way to retrieve locale settings reliably outside of libc.
related:
https://desuarchive.org/g/thread/104831348/#q104850560
>I think you're lying
see picrel. solve the above C tardation, and let's see if locale support continues to be a problem in rust projects. the floor is yours.
or accept a new way of setting locales that is not inherently broken.
further reading:
https://github.com/mpv-player/mpv/commit/1e70e82baa9193f6f027338b0fab0f5078971fbe

>>107165350
>he's been doing this since march
damn
anyway, rust already depends on libc for malloc, threading, signals, etc. it can wrap locale-aware functions as well
>see picrel
uutils sort still can't sort i between ı and j no matter what
>above C tardation
that code doesn't make any sense
>further reading
you can tell it's a skill issue from the first sentence

>>107165521
>anyway, rust already depends on libc for malloc, threading, signals, etc.
then show how to get locale settings from libc's public API.

>>107161461 (OP)
trannies

>>107161461 (OP)
cargo is god damn awful

>>107165565
nl_langinfo(D_T_FMT) for example
what stops rust from doing this?

>>107164892
No, I know plenty of good arguments against Rust, but /g/ is just too fucking retarded (because of /pol/fags) to come up with them.

>>107165598
how to get system collate settings using nl_langinfo()?

>>107165685
answer the question first

>>107165717
nothing.
now answer, how to get currently set collate settings using public C api, to use that info in non-C sort implementations.

>>107165521
>anyway, rust already depends on libc for malloc, threading, signals, etc.
There is threading and signals in libc? Where?

>>107166121
I don't know. How about the headers literally called <signal.h> and <threads.h>?

>>107164916
>there are people on /g/ right now who defend C locales
holy shit.

>>107165656
>I totally do know, trust me bro, I just don't wanna tell

>>107166373
I unironically like watching fa/g/gots trip over themselves because they have no idea why they hate what they hate, and I'd rather not interfere too directly.

>>107166121
signals are, which is comical because M$ probably had to do a lot of contortions to add them to ucrt. literally one of the dumbest libc additions.
both C++ and C11 pretty much standardized around pthreads.

>>107166277
to be fair, both are absolute garbage. even thread is technically optional and on Linux (or other POOSIX), you'd just use pthread directly and on Windows, well you have a rich threading library available as well.
the reality is, C++ threads are just better anyway and C just tacked them on, along with the atomics model.

>>107165350
uh... C bros? can someone explain that archive link?

>>107161461 (OP)
the trannys shilling for it

>>107161461 (OP)
It makes it better on average actually, it's not even about the memory safety.
>>107161535
If you think the only reason to use rust is memory safety you are a retard that doesn't know what he's talking about.

>>107165521
> depends of libc
Entirely optional dependency, likewise for llvm, you can build without it.

>>107161461 (OP)
more logic errors because the developers are all troons that have no interest in maintaining their troonware. even if they did care, they'd all end up acking anyway. that's why you see rust devs transitioning existing stable C/C++ based programs that end up being slower or flat out broken on edge cases.

>>107161461 (OP)
>What makes Rust worse for the end-user?
Don't care I'm still learning HolyC

>>107164868
The rust initiative in Ubuntu is fronted by a glownigger named Jon Seagar btw.
https://jnsgr.uk/cv/
>british army surveillance specialist who worked for defence contractor before canonical

>>107168899
The creepiest part is that his name never shows up in the archives. Search it yourself, no one talks about this literal fed who created the oxidising Ubuntu initiative
https://www.rustnationuk.com/speaker/jon-seager

>>107161461 (OP)
Memory safety won't save the end user from shit design decisions by the programmers.

>>107161461 (OP)
>git clone <rust project>
>cargo build
>du -sh .
10GB

>>107166277
Oh, I thought these are part of Unix/Linux API, not standard C libraries.

>>107166314
there's nothing wrong with them from the end user's perspective

>>107169031
>I thought these are part of Unix/Linux API
UNIX is an IDE for C

>>107169031
A simple google search would easily tell you this is not true:
https://en.wikipedia.org/wiki/C_standard_library

signals.h has been in C since forever and threads.h was introduced with C11 though it is just a new standard to use instead of the older pthread.h.

There is no Unix/Linux without C and the language would not have become popular without Unix. You cannot have one without the other, they are theoretically a single entity.

>>107165911
>nothing
wrong, incompetence
>how to get currently set collate settings using public C api
wcscoll()
or locate and parse the locale definition

>>107161461 (OP)
>the end-user
so you mean the executable? nothing, it's made by compiler anyway. LLVM doesn't give a fuck if input is in rust or other lang.

>>107169314
>>how to get currently set collate settings using public C api
>wcscoll()
>or locate and parse the locale definition
the goal is NOT to compare using a C function.
the problem is NOT how to use collate data.
the problem is WHAT collate data to use.
this is shown in the archive link from >>107165350.
you got close with the "locate" above.
now show WITH SPECIFIC WORKING CODE how to get currently set collate settings using public C api. or if you want it in simpler terms, show how to "locate" the currently active/set collate info. which should be no problem to you, since you're such a competent senpai.
and try not to sound like a dumb retard this time...who am i kidding. lol.

>>107168899
are defense contractors bad?
APOLOGIZE TO NIX TRANNIES RIGHT NOW.

>>107168985
>made up target sizes are bad
>for the END USER
the /g/eets are not sending their best

>>107169031
they are.
"modern C" is irrelevant and no one uses it, including rust's core/alloc/std crates.
/g/eets are retarded and know nothing, and half of them just copy-paste from LLMs, and that's the good half.

>>107170401
>LLVM doesn't give a fuck
The official LLVM yes, but Rust people uses their own fork which is 10k+ commits behind
https://github.com/rust-lang/llvm-project/tree/rustc/21.1-2025-08-01
Think Rust people are bad at writing Rust code? Imagine what they could be doing to a C++ code base

>>107170608
>cargo requires an internet connection
>pulls in mossad spyware despite not being asked to
nothing personnel, goy

>>107171224
>zero evidence
>vendoring is a built-in feature, where every single source file is hashed and verified
>pretending like anyone builds software with sneakernet-acquired sources
>still zero relevance to end users
maybe you're confusing it with golang with its modules-are-urls "design", and don't know what end-user means.
but i will leave it to you to pick what kind of retard are you.

>>107171224
>cargo requires an internet connection
end user = running the executable
the program is already compiled at this point. try harder.

>>107171326
>>zero evidence
your denial is based on what? idiocy, or expertise?
https://blog.trailofbits.com/2025/09/24/supply-chain-attacks-are-exploiting-our-assumptions/
https://internals.rust-lang.org/t/about-supply-chain-attacks/14038
and this from 2 months ago
https://kerkour.com/rust-supply-chain-attacks
>In a recent analysis, Adam Harvey found that among the 999 most popular crates on crates.io, around 17% contained code that didn't match their code repository.
>>107171326
>golang with its modules-are-urls "design"
from the same article
>Go is today the gold standard for supply chain security by using an hybrid decentralized / centralized architecture.

>>107171224
>>pulls in mossad spyware
how dare you defy rust and its antics? to the gulag.

>>107171755
>>In a recent analysis, Adam Harvey found that among the 999 most popular crates on crates.io, around 17% contained code that didn't match their code repository.
thankfully, the xz release source did match the code repository (broken english not mine).
>>Go is today the gold standard for supply chain security by using an hybrid decentralized / centralized architecture.
must be a nice article. shame i will never read it lmao.
i will leave the the "caring about the possibility of supply chain attacks means the mossad are inserting spyware" (with the implicit implication that the mossad has super powers, that's always a nice addition;)) talking point without reply for others to see. but these comments from another thread come to mind:
>>107170768
>>107171278
very "organic" (people like that word)

>>107171755
>In a recent analysis, Adam Harvey found that among the 999 most popular crates on crates.io, around 17% contained code that didn't match their code repository.
So codegen? That's a nothingburger, that's how you make wrappers for C libraries for example.

>>107169228
>there's nothing wrong with them from the end user's perspective
you're literally retarded if you believe this. the fact that steam client on specific builds of glibc started more reliably failing due to setenv() bullshit should tell you everything wrong with how C locales work in practice.

just because they aren't completely annoying doesn't mean they aren't broken.

>>107171755
>>Go is today the gold standard for supply chain security by using an hybrid decentralized / centralized architecture.
Most go cuckolds realize that go get and friends fetch code from the Google proxy right? literally it's just crates.io with extra superfluous bullshit and forces a specific VCS on you on top of it. You still need to set shit like GOPRIVATE, which is, IMO much more clunky than defining a custom package mirror like you would in literally any other language.

this.

and this.

>>107172238
>I'm just going to bury my head in the sand and not acknowledged that Rust is insecure due to supply chain issues because I don't like what you're saying but I have no real retort to it.

>>107161461 (OP)
that it doesn't work at all

>>107172428
>remove bloat
>seething issue with no actual argument
Kek

>>107172421
bcachefs is C. It has 0.1%-1.4% of Rust code. It never "transitioned" to Rust.
https://github.com/koverstreet/bcachefs
https://github.com/koverstreet/bcachefs-tools

It became such a bloated project of constant feature creep, not because of that 1% of Rust code.

>>107172498
>Rust is insecure due to supply chain issues
How is it more insecure than your average Linux distro package manager?

>>107161461 (OP)
Performance tanking

>>107172634
Got benchmarks?

>>107172625
it's not a bloated project dumbass, but it is trice as hard to compile because of that 1%
>>107172629
my distro packages aren't maintained by retards; crates.io is a mess simular to npm or aur.

>>107172668
>my distro packages aren't maintained by retards
What distro that is? Was it safe from xz and other recent attacks?

>but it is trice as hard to compile because of that 1%
Got any stats for that? It compiles fine on my system (Arch Linux)

>>107172629
it isn't. people still believe the delusions that software repositories are vetted. every single system that pulls code you have not reviewed or paid professionals to review is a risk, whether sneakernet'd into your build system, from your distro package manager, from shithub, npm, PyPI or crates.io or whatever other bullshit repo. it's all the same shit. I'd argue in some cases at least micro deps are usually so small that their testable interface should be easier, but then you carry the burden of going through a big list of deps. everything is trade-offs and I'm sick of all the retards who keep claiming langauge package managers are bad when literally every other language, other than C-like shitwares (outside of M$ land. vcpkg, buck2 and some other meme cross-platform ones exist, gl with those), has one (or two if you're Java for some reason) bespoke ones.

>>107170618
<signal.h> has been in ISO C since the beginning, you absolute mouth-breathing retard. And the one and only reason not everyone uses C11 or later is knuckle-dragging boomer geriatric college teachers who smell like sour milk and still teach programming like it's 2004.

>>107173372
no one gives a fuck about the total of six signals ISO tards defined.
<signal.h> is a POSIX implementation in any real OS.
no one is going to use anything "modern" from ISO niggers (some literally so), be it C or C++.
you can go cry in a corner, ISO nigger.

>>107172707
>Was it safe from xz and other recent attacks?
yes, i don't use systemd, though most mainstream systemd distros was also not vulnerable.
>got any stats
read the fucking article, jesus christ

>>107173896
>the pozzed package which i got wasn't effective that time
>this is proof that the system was rock solid and safe from supply chain attacks
>also read the random search results i provided with all the.. ZERO.. successful attacks mentioned

>but but typo-squating
lol

(i'm going to call this type of retardation the jeet hasbara from now on)

I think the biggest issue is the false sense of security that is conveyed in the advertisements. Apart from that, the wasted rewriting effort that could have been spent doing something productive

>>107168985
It's either that or
>error: a library needed for this project conflicts with one used for another project. If you fix this one, that one breaks.

>>107173896
>yes, i don't use systemd, though most mainstream systemd distros was also not vulnerable.
So what's the name of the distro?

>read the fucking article
I did

>>107173633
>POSIX exists therefore the <signal.h> header isn't ISO
JFC you're desperate.

>>107172541
Oh no, rust doesn't work with alien languages! To the dumpster it goes.

>>107175654
>t. never used cargo

>>107175970
even those who didn't know, anyone with a brain would have already checked out "man 0p signal.h", you dumb nocoding retard.

>check thread
>vast majority doesn't even know what "end-user" means
pottery

>>107176763
Meant to say that's the C alternative to cargo taking lots of space

>>107170583
Where did the pretend cnile oracle go?
I was actually hoping he would provide us with a way we didn't know about.

Why can't rust faggots just bridge or wrap working C code? Rewriting it to be shittier and broken is retarded. Security critical tools and interfaces can be written in rust.

>>107161876
It is not possible to compile firefox (and many other packages) without rust without heavy per-package patching.

>>107165350
>solve above c code
There is nothing to solve, as there is no problem to be solved. You can't just cram in random jibberish and call it retarded, you retard.

>>107177162
I though it meant the end user of the written in rust, compiled and distributed program. Is that not what that meant?

>>107161461 (OP)
Rust based OS tooling is going to crash and burn.

Rust is bloated.
It's written in a language hardly anyone knows and hardly anyone can modify.
It's got a CoC so it's fucked and cucked.
Glowniggers love Rust and the Rust foundation owns the compiler, meaning they can hijack your software later.
The language is stupid and slow as dogshit.

>>107179489
Yep

>>107179458

>>107165565

Why even go through the humiliation to pretend that you care about locales in 2025? Everybody knows that it's broken garbage.

>>107182554
rust can't into localization in general

>>107182554
let the pretend C oracles dig their own graves and just watch man.

>>107182637
localization is a different matter, and the rust ecosystem has fluent (something you never heard of) support in the ecosystem without issue.
but you specify what you mean by localization, and we will go from there.

>>107182669
let's start with sorting files according to my language's alphabet

>>107182678
keep up retard. this has all been already discussed.
https://docs.rs/icu_collator

>>107182678
Learn into unicode.

>>107182728
>>107182729
provide a concrete example. uutils sort and ls, eza, fd, etc. I tested all none of them can do it

>>107182739
https://github.com/houseabsolute/omegasort
anything else?

>>107172625
>bcachefs is C. It has 0.1%-1.4% of Rust code. It never "transitioned" to Rust.
It is planning to, though.
Even uses the "transition" verbiage.

>>107183839
Plans do not break builds.

>>107184055
But culture might, judging by Rust 1.80 time crate mess and the recent Rust-Ubuntu woes.

>>107184542
>might
So it's just a theory