News for hackers. Stuff that matters.
Re:11 year old unsoundness bugs found in Rust borrow checker by ahartmetz on Tuesday October 13, 2026 @09:36AM (#50071552)
At least Rust has a chance to be largely correct because it has much greater effort per amount of code (in the compiler) invested into it than most any application code.
Re:11 year old unsoundness bugs found in Rust borrow checker by throwawayffffas on Tuesday October 13, 2026 @10:14AM (#50072130)
"Largely correct" my 2010 python 2 code is "largely correct".
Re:11 year old unsoundness bugs found in Rust borrow checker by ahartmetz on Tuesday October 13, 2026 @10:32AM (#50072395)
You are saying that there are only incorrect and mathematically proven with nothing meaningful in between?
Re:11 year old unsoundness bugs found in Rust borrow checker by AKrumbach on Tuesday October 13, 2026 @11:17AM (#50073070)
Not the grandparent you were replying to, but from the note added to the original linked issue (emphasis mine):

> Rust's type system is *intended to prevent memory unsafety. Upholding this promise is something we take very seriously*, .... It is a large priority of the Types Team to *fix all type system unsoundnesses* and to work towards there being a solid foundation to minimize the risk of unknown ones.

So, yes, if you have a group who "takes very seriously" the job to "fix all type system unsoundness", I would expect them to utilize some manner of formal proof verification that their type system is mathematically correct up to some defined limit.

Re:11 year old unsoundness bugs found in Rust borrow checker by thayne on Tuesday October 13, 2026 @11:33AM (#50073343)
Re:11 year old unsoundness bugs found in Rust borrow checker by ahartmetz on Tuesday October 13, 2026 @01:14PM (#50075221)
That work is ongoing even if the RustBelt project ended. It's been a / the major focus of Ralf Jung's work for over 10 years and he's an assistant professor at ETH Zürich now, still working on it. https://www.ralfj.de/blog/categories/research.html
Re:11 year old unsoundness bugs found in Rust borrow checker by hyperman1 on Tuesday October 13, 2026 @11:18AM (#50073085)
Mathematically correct is not enough.

Java's byte code verification algorithm had a proof. This was important for running untrusted java in the browser. If it said the byte code was ok, it was.

Turns out, nobody made a proof for the verifier actually finishing. So it could return true, or false, or run longer than the universe has time. So around Java 6(?), someone wrote some bytecode that basically took forever to validate, DOSing the validator. Sun had to change the file format and add extra information to get the guarantees back.

Re:11 year old unsoundness bugs found in Rust borrow checker by throwawayffffas on Tuesday October 13, 2026 @11:19AM (#50073105)
No. I am saying you don't need rust for largely correct.

PS: The term you are looking for is formally verified not mathematically proven.

Re:11 year old unsoundness bugs found in Rust borrow checker by throwawayffffas on Tuesday October 13, 2026 @11:22AM (#50073163)
Oh and btw their borrow checker is the kind of system that you would want formally verified.
Re:11 year old unsoundness bugs found in Rust borrow checker by mamcx on Tuesday October 13, 2026 @12:21PM (#50074221)
No, Python has not the level of "correct" that rust has.

`sum(a:int, b:int)`

not guaranteed ANYTHING.

The gulf between how much correct Rust is and the vas majority of languages is HUGE.

Re:11 year old unsoundness bugs found in Rust borrow checker by throwawayffffas on Wednesday October 14, 2026 @07:02AM (#50085704)
`sum(a:int, b:int)`

That's why I specified python 2. No type annotations at all.