The first thing I always look for is mutability (combined with some other tricky feature), e.g. "Now it's a Just(3):Maybe<Int>, now it's a Nothing:Maybe<Int>, now it's a Nothing:Maybe<String>."
I wonder if TFA's example could be reproduced without mutability.
so what I take from this is that the rust compiler does not guarantee the stuff it claims to. Because new features of the type system is just being yolod instead of formally proven
There never was formal verification of the entire rust compiler, or even its complete borrow checker/type system. There has been verifications on _subsets_ of them.
Also, a lot of those soundness issues aren't related to new features _at all_, but usually due to poor assumptions made in optimizations.
The soundness bugs _are_ taken seriously though, and are slowly being patched out. A lot of the links in OP are now closed because Rust either changed how it generates code to avoid the issues, or fixed LLVM.
not of the implementation. Of the type system. And a verification of subsets, but not a verification of their interactions == the whole system is unverified.
No mainstream compiler is formally proven. If you want a verified compiler, you pretty much have to choose between using a subset of standardml or a subset of C and the one proven compiler for each.
Spot on, but this makes me think of a larger problem.
You can be annoyed by Rust fanboys, sure, but the more annoying are the Rust haters, because the hate always comes from a place of ignorance.
The hate only works if you don't go beyond a surface level understanding of Rust.
E.g. a niche case on the borrow checker is slightly broken, so borrow checking doesn't make sense.
Or the "oh but Rust needs unsafe, so it's just like C++".
Yeah, but by that measure all of C++ is unsafe++, even unsafe has more guarantees than standard C++.
Only if your whole understanding of the concept is that the keyword is called "unsafe", does that argument make sense.
A bit of a rant, but if you hate Rust with a passion, ask yourself first where that opinion comes from and if it can stand on its own. In most cases it can't.
> You can be annoyed by Rust fanboys, sure, but the more annoying are the Rust haters, because the hate always comes from a place of ignorance.
For what it's worth, Rust fanboys also seem just a vulnerable to their ignorance of C/C++ compiler extensions and tooling (e.g. sanitizers, compiler extensions, standard library hardening modes and so on.) which are just normal part of professional C/C++ use today.
Yes. When people describe Rust as "a memory-safe programming language" they are describing an aspirational vision, and not a programming language that has actually been designed with sufficient rigor to guarantee memory safety. Any failure of rustc to actually guarantee this is just a bug in rustc itself, a failure to properly implement the Rust spec, because the closest thing Rust has to a spec is literally just the vague ideal that the compiler should guarantee safety somehow.
I don't think the definition of "memory safe" has ever required full end to end stack conformance with a proof. Very few languages have achieved that, though some have. Most people call Java memory safe, C#, Python, Typescript, etc - all of those fall short of a formally verified, sound type system with a verified toolchain.
Not to talk about Go, which has intentional unchecked-at-runtime UB (data races) yet many argue to consider it memory safe (and I have a feeling that the same people arguing Rust isn't memory safe are also there. Just a feeling though).
I'm not an expert on the Java specification, but my understanding is that Java as a language is well-defined enough that we can say things like "Java's type system is unsound", independently from any specific implementation of Java.
You can't do that with Rust because "Rust's type system" is literally just the vague concept of "a type system that is safe". The implementation of this ideal in rustc can have flaws, but Rust itself cannot, because Rust itself has no flaws by definition.
It's surprising this bug went unnoticed for so long, but at least it's easy to fix, since it's not a bug with the type system itself; it's just a bug in the compiler. The compiler should always have been rejecting that code according to lifetime rules of Rust's type system, so the fix is to just make it reject the code, rather than having to change the fundamental rules of lifetimes.
This is remarkable but not super unusual. There are more than 100 issues labeled unsound currently. Some are long standing but from what I understand the new trait solver will unblock fixes for some of them (but probably not the one in OP).
Finding unsoundness bugs isn't the most unusual, yes. But this one is unusual in how simple the code that causes it can be (see the versions of the reproducer code in the comments). For other unsoundness bugs, the code would need to contort itself a lot to avoid the compiler checks.
Perhaps ironically, AI-based searching of the Rust compiler for bugs is likely to find more of these, pushing Rust closer and closer to the point of complete reliability.
The eventual formalization of Rust and a creation of a formally proven compiler is presumably going to be the end point of this process.
There is a formally proven compiler for a large subset of C, so I don't see why formalizing (a large subset of) rust and making a formally proven compiler shouldn't be possible
Also unlike C, it should be possible to formalize all of Rust, not just a subset. For two reasons: Rust evolution (not initially but today) is strongly tied with formalizing it and the language team consults with formalization experts (which are part of the Rust project themselves) all the time (and formalization or semi-formalization projects are ongoing), and also the mathematical understanding of system languages itself has advanced.
There is a severe misunderstanding here. Beside being only a subset, CompCert definitely has UB. It just formalizes codegen for codebases not containing UB, which obviously won't be enough for Rust.
I asked myself, How many comments to do I have to collapse before someone says something about AI. 3.
I used to compare this hype to crypto hype, but its so much worse. Can we just try not to shove llms into every single conversation.
Is there a hackernews type space where the word is banned? Even lobste.rs has this problem. I need some sort or way to just hide all content on certain sites where LLMs/AI are mentioned.
Took me a good couple minutes to wrap my head around. Man, what a nasty bug. I guess this specific use-case is rare. Time to see if any of the big libs are affected...
And nasty and rare bug can and will be exploited once it is found. The only thing necessary is for malicious actors to start adding this case in innocent looking code and there you have.
What the hell. This is not the esoteric, crazy, impractical constellation I expected to find with a title like this. The minimal example is so simple and is definitely something you'd easily come across in normal code.
Afaik loop {} has never been UB, however because llvm started as a c++ compiled where it is, this has been causing miscompilations in rust and C for… nearly two decades (llvm issue 965).
I wonder if the reason it works is because you'd easily come across it in normal code, and someone didn't realize that saving the effort of implementing a check for this would result in unsoundness.
Not really. I've worked with rust for 4 years, systems side, and never had to write a custom drop implementation. That's the gotcha here, not the rigamarole with the static+Send option type (which _would_ be something you'd encounter in prod).
A custom `Drop` appears very quickly with FFI code. Suppose for example you interface with some C library that expects you to call a function to release some resources.
- "Not really". Implying that your experience generalizes
I've written a bunch of rust programs without custom Drop. I believe your first statement. There are large classes of programs that don't need it. There are also many that do. I think your statement says a lot more about what kind of programs you work on than about custom Drop not being common. People are downvoting the implication that custom Drop is not a common pattern.
No we're in a situation where a bug in the type checker was found and will be fixed and a new release will be made.
That it took 11 years for someone to come across this is a fairly intense counterargument to your attempt to paint this as a language with dangerous footguns.
Worked on a variance-related lint once and the scary part is how rarely anyone writes down variance by hand, so bugs there just sit untested for a decade.
Old C software is generally more trustworthy than new C software. I don't think this rule of thumb transfers across languages very well. And the definition of new vs old has a fairly large effect. If a language has been under constant change it doesn't matter when it was started a large part of the codebase is "new".
struct Q<T: 'static>(Option<T>);
impl<T: 'static> Drop for Q<T> {
fn drop(&mut self) {
// Code in here has access to a non-'static &String
// but assumes that it's a &'static String
}
}
fn main() {
let l = String::new();
let mut q = Q(None);
q.0 = Some(&l);
}
I wonder if the certified compiler for safety critical uses (edit: Ferrocene) has the same issue.
Rust is supposed to guarantee statically that a reference cannot outlive its referent. The `T: 'static` bound means that the value that Q holds must have a 'static lifetime (e.g. either be an owned value or a reference to a static value). So this program should not compile since a reference to l is not 'static. This code is unsound because rust will drop the string l before q so when the Q destructor runs the underlying string has already been freed yet we have a valid reference to it.
So Q is a "tuple-struct". In rust you can define a struct like
struct Q<T> {
value: Option<T>
}
with a named field or you can define it like
struct<T> Q(Some<T>)
which is the same struct layout but without an explicit name for the value. In that case, the field is implicitly names `0`
So q.0 = Some(foo) is equivalent to q.value = Some(foo).
The Some(..) is just one variant of the Option enum:
enum Option<T> {
None,
Some(T)
}
Option represents a value that may or may not be initialized. For uninitialized values you use the None variant and then you can initialize it to Some(value).
Some(...) is just constructor for an option type as seen in other languages e.g. java's Optional, haskell's Maybe, c++'s std::optional. I assume that it's needed here because the field can't be a null pointer otherwise.
q.0 is the first field of q, which is anonymous. They could have just as well called it x and written q.x.
chatgpt and grok linked to the the soundness issue, claude I am in usage timeout, and gemini shit the bed as usual: "The compiler rejects the program with error[E0597] on the assignment q.0 = Some(&l)"
Here's the weakness of compiled programs. Does this compiled code have this vulnerability? You have no idea. At least with a VM or interpreter you can scan for vulnerable versions and replace the runtime or the specific library.
I think it has no relevance here at all. Actix was unsound because of prolific use of unsafe. That was a poor design decision but a conscious one. TFA is just a bug.
Rust is good, I generally do not dislike it - but it does have flaws.
I will happily continue to write things in Golang first. I don't enjoy refactors in Rust, and it's rare that I see things that genuinely need its raw performance.
Cloudflare's Oxy is a stellar example of build in Golang till your design is settled and then port over for performance when you need it.
This is not the first soundness bug. Some bugs aren't fixed for years. The Rust community is well aware that such bugs exist, will be found, and address them just like other high impact bugs. In other words, if you paid the slightest attention to the development of Rust language, you would know this is surprising but not unexpected, and has happened many times.
Given that TFA is currently #1 on HN, I suspect I'm not the only person who is aware of Rust's borrow checker but not aware, until now, that it has a history of serious soundness bugs.
It looks like 18 of the 30 issues listed in table 4 had already been fixed when that paper was released, and there is activity in most of the rest. One was recently flagged as being fixed by -Znext-solver, and another has the submitter write that it is fixed, but the issue is not closed. And that's before you consider the impact of the individual issues; how many of the remaining issues impact "real" code?
So I don't think it's fair to claim that "no-one cared".
I don't think "no one cared" is a fair characterisation. The paper doesn't support this conclusion, and from experience I'd say that the two difficulties in fixing unsoundness are doing so without breaking existing code (Rust guarantees stability unless the code is unsound, and introducing previously missing checks can cause sound code to stop compiling, which is a problem if it occurs often) and doing so in a reliable way that can be supported long-term. Rust is currently in the progress of stabilizing next-solver and Polonius, both of which fix multiple long-standing bugs and have been worked on for years now due to intrinsic complexity, performance considerations, compatibility issues, etc. It's not that no one cares, it's just that this stuff is hard and takes time.
Rust is not a memory-safe language. cve-rs proved it years ago but commenters always argue that cve-rs is too contrived to matter, or that it's a compiler bug so it doesn't count.
Literally nothing can be anything in absolute terms, what's your point? The fact is that Rust is miles better than any other language in terms of memory safety & performance ratio. That's all that matters.
But bugfixes are also allowed. So the fair conclusion is that the implementation is broken, but the language is not - because it's possible to fix, and it's possible for all known soundness issues (including the cve-rs one). For this one specifically, I've heard people are already working on a fix.
Lifetimes can be effectively erased via covariant assignment.
It's in the title.
Even a beginner could understand these assignments are not in spec?
Are you just being contrary to be contrary?
If that's the definition you're using, then C is also a safe programming language. In fact, C is even safer, because it actually does have a specification that lists all the invariants that the programmer has to uphold themselves because the compiler doesn't check them.
Specifically the `q.0 = Some(&l);` violating 4.15.2:40 ("An assignment expression requires that the type of its value operand is a subtype of the type of its assignee operand.")
I totally understand that people have favourite languages and that rust, like any language and any ecosystem around that language, has idiosyncrasies that are not for everyone, but how would one ever come to the genuine conclusion that rust is not being taken seriously? By whom?
I gave it a quick scan for Rust examples; there are 11 samples among the following pages:
The first thing I always look for is mutability (combined with some other tricky feature), e.g. "Now it's a Just(3):Maybe<Int>, now it's a Nothing:Maybe<Int>, now it's a Nothing:Maybe<String>."I wonder if TFA's example could be reproduced without mutability.
Also, a lot of those soundness issues aren't related to new features _at all_, but usually due to poor assumptions made in optimizations.
The soundness bugs _are_ taken seriously though, and are slowly being patched out. A lot of the links in OP are now closed because Rust either changed how it generates code to avoid the issues, or fixed LLVM.
hm, there's an effort but scoped only to the std
It's terrifically complex: https://doc.rust-lang.org/beta/nightly-rustc/rustc_middle/ty...
You can be annoyed by Rust fanboys, sure, but the more annoying are the Rust haters, because the hate always comes from a place of ignorance.
The hate only works if you don't go beyond a surface level understanding of Rust.
E.g. a niche case on the borrow checker is slightly broken, so borrow checking doesn't make sense.
Or the "oh but Rust needs unsafe, so it's just like C++".
Yeah, but by that measure all of C++ is unsafe++, even unsafe has more guarantees than standard C++.
Only if your whole understanding of the concept is that the keyword is called "unsafe", does that argument make sense.
A bit of a rant, but if you hate Rust with a passion, ask yourself first where that opinion comes from and if it can stand on its own. In most cases it can't.
For what it's worth, Rust fanboys also seem just a vulnerable to their ignorance of C/C++ compiler extensions and tooling (e.g. sanitizers, compiler extensions, standard library hardening modes and so on.) which are just normal part of professional C/C++ use today.
- If a compiled C++ program has unsound behavior, the program needs to be fixed
- If a compiled safe Rust program has unsound behavior, the compiler needs to be fixed
1) Unsoundnesses in Java's type system are checked at runtime.
2) They (or at least some) are intentional or at least won't-fix bugs.
Both are not true for Rust.
Rust cannot fail, it can only be failed.
“Please build the perfect system. Because I specified ‘perfect’ any flaws are due to your incomplete implementation, not my awful brief.”
If the person writing the rust spec was a client I would have to drop them because they would be setting themselves up for dissatisfaction.
You can't do that with Rust because "Rust's type system" is literally just the vague concept of "a type system that is safe". The implementation of this ideal in rustc can have flaws, but Rust itself cannot, because Rust itself has no flaws by definition.
Sweet. But why should anyone care..
The eventual formalization of Rust and a creation of a formally proven compiler is presumably going to be the end point of this process.
I used to compare this hype to crypto hype, but its so much worse. Can we just try not to shove llms into every single conversation.
Is there a hackernews type space where the word is banned? Even lobste.rs has this problem. I need some sort or way to just hide all content on certain sites where LLMs/AI are mentioned.
https://www.lesswrong.com/posts/zHhGa3iQBQcdSifFL/a-critical...
The reality is that if this were actually code that people wrote, well, it wouldn't have sat around for a decade.
The fact that nobody reported this before means it's not, by definition.
This is the problem when a language advertises with “guarantees” and it turns out it never had any due to many false assumptions by their promoters.
So what other soundness bugs are there in the language that we need to know about?
For some reason this is controversial, but it remains true.
- you've never written a custom Drop
- "Not really". Implying that your experience generalizes
I've written a bunch of rust programs without custom Drop. I believe your first statement. There are large classes of programs that don't need it. There are also many that do. I think your statement says a lot more about what kind of programs you work on than about custom Drop not being common. People are downvoting the implication that custom Drop is not a common pattern.
But not custom drops that are doing variadic type stuff like this, usually.
That it took 11 years for someone to come across this is a fairly intense counterargument to your attempt to paint this as a language with dangerous footguns.
Maybe they wrote it in C :-)
On a serious note, old software is generally more trustworthy than new software, because bugs are discovered over time and with usage.
It uses the same code, so yes.
(Although I don't know what are their guidelines, it might be that they require avoiding code relying on implied bounds, for example).
struct Q<T> { value: Option<T> }
with a named field or you can define it like
struct<T> Q(Some<T>)
which is the same struct layout but without an explicit name for the value. In that case, the field is implicitly names `0`
So q.0 = Some(foo) is equivalent to q.value = Some(foo).
The Some(..) is just one variant of the Option enum:
enum Option<T> { None, Some(T) }
Option represents a value that may or may not be initialized. For uninitialized values you use the None variant and then you can initialize it to Some(value).
The C rough equivalent would be
struct Q { void* value; }
struct Q q = q { NULL }; q->value = ...
q.0 is the first field of q, which is anonymous. They could have just as well called it x and written q.x.
"q.0" is not really needed, it's a struct with unnamed fields. Here is the same with named fields:
Most LLMs would probably say this piece of code won't compile.
With this you have to recompile everything
Having ideals does not mean there are never bugs.
I will happily continue to write things in Golang first. I don't enjoy refactors in Rust, and it's rare that I see things that genuinely need its raw performance.
Cloudflare's Oxy is a stellar example of build in Golang till your design is settled and then port over for performance when you need it.
ETA: Downvoters, I salute you. Without your selfless courage, the forces of light humour would surely prevail.
https://github.com/rust-lang/rust/issues/25860
EDIT: I don’t even think HN cared either up until now. If you search this paper on HN, guess who posted it?
Lets hope that this will be the only occurrence of soundness bugs in Rust and perhaps it needs a tiny disclaimer on claims of “safety guarantees”.
[0] https://arxiv.org/abs/2608.28713
So I don't think it's fair to claim that "no-one cared".
Safe rust is safe, granting `unsafe` and compiler errors. This isn't a secret.
https://github.com/rust-lang/rust/labels?q=unsound
it has been discussed bilion times
Specifically the `q.0 = Some(&l);` violating 4.15.2:40 ("An assignment expression requires that the type of its value operand is a subtype of the type of its assignee operand.")