41–50 of 89 posts
> As an aside, remember that the only difference to C/c++ is that if you write a “basic linked list” in them, all of your code will be unsafe. I stopped reading here.
In the rust world "unsafe" is synonymous with "isn't proven to be safe by the compiler". Under this definition, every C/C++ program that uses pointers is "unsafe" because the languages make no memory safety guarantees.
I think this misses the problem. Should you write your own data-structures? Not unless absolutely necessary. But mostly everyone knows that. So why are people complaining about data-structures? For me, writing the data-structures is the canary in the mine. It's the next step from hello world when trying to pick up a new language. Most importantly, trying to write a few simple data-structures hints at the difficulty t…
Earlier quoted context omitted.
> For people like me, if writing trivial data-structures in Rust is a great challenge, it shows that writing other, less trivial things in Rust will be much more challenging than might be preferred. And this is the real issue, because your assumption doesn't really have any real basis. I don't mean to belittle your point, I can completely understand where you're coming from, but assuming that because writing data str…
>Data structures are a very specific domain, often very far away from what you'll be doing in the average "real" program. I disagree. Choosing the right data structure often simplifies coding implementation and provides memory/performance guarantees. And it's not only me that thinks that way either. Here's proof by Linus Torvalds (and a whole bunch of people on HN that agree with that sentiment) https://news.ycombina…
"Show me your flowcharts and conceal your tables, and I shall continue to be mystified. Show me your tables, and I won’t usually need your flowcharts; they’ll be obvious."
I think this misses the problem. Should you write your own data-structures? Not unless absolutely necessary. But mostly everyone knows that. So why are people complaining about data-structures? For me, writing the data-structures is the canary in the mine. It's the next step from hello world when trying to pick up a new language. Most importantly, trying to write a few simple data-structures hints at the difficulty t…
Data structures are exactly the abstractions built on top of raw memory. Rust is not really geared toward working with raw memory; it's geared toward working with abstractions that hide the raw memory (i.e., data structures). That's why writing data structures in Rust is hard, and it's also why that fact doesn't imply that "other things" (i.e., code that is not part of a data structure implementation) are hard.
In other words, the fundamental flaw with the assumption that there is a relationship between the difficulty of implementing data structures in Rust and the difficulty of writing applications in Rust is that the two involve very different programming models, and Rust has much more ergonomic support for one than the other.
You can do linear lists and trees in Rust. The problem is backlinks. If you refcount everything, you can have backlinks, but otherwise there's a safety problem. Backlinks require an invariant which covers two variables, and you can't express that in Rust. Back pointers are a special sort of pointer from an ownership perspective. They don't carry ownership, but are locked in an invariant relationship with the pointer…
This is basically dependent types. Rust doesn't really want to go that far from a language level.
> Handling these two classes of unsafe code takes care of a sizable fraction of the unsafe code really needed in Rust.
Not really. All the other datastructures in the stdlib would still need unsafe. Not really a good metric.
The Rustbelt project at MIP-SWS is doing formal verification of Rust including unsafe code, and that approach will be far better IMO. It won't let you eliminate unsafe from your code, but might make it easier to prove that your unsafe code is performing safe operations.
Earlier quoted context omitted.
> For people like me, if writing trivial data-structures in Rust is a great challenge, it shows that writing other, less trivial things in Rust will be much more challenging than might be preferred. And this is the real issue, because your assumption doesn't really have any real basis. I don't mean to belittle your point, I can completely understand where you're coming from, but assuming that because writing data str…
Again, the issue isn't data-structures or about them. The issue is: Can I write something in scratch in this language 1) at all and 2) with some amount of daily progress. Writing data-structures is a test case of my ability to thing and program in the context of Rust. It serves to answer the question: Can I write a program that implements a well defined, well understood construct so that I may learn the building bloc…
The point is that it's not the best test case. A better test case would be writing the programs you generally write. It'll still be hard to be honest, but really not quite as hard. I have enough familiarity with Rust to write application code myself, but I'd probably still struggle a lot with writing a data structure if I tried.
I think this misses the problem. Should you write your own data-structures? Not unless absolutely necessary. But mostly everyone knows that. So why are people complaining about data-structures? For me, writing the data-structures is the canary in the mine. It's the next step from hello world when trying to pick up a new language. Most importantly, trying to write a few simple data-structures hints at the difficulty t…
Creating complex object graphs is also not trivial in Haskell, yet that language seems to do fine with real-world problems.
Earlier quoted context omitted.
In the rust world "unsafe" is synonymous with "isn't proven to be safe by the compiler". Under this definition, every C/C++ program that uses pointers is "unsafe" because the languages make no memory safety guarantees.
Raw pointers, maybe. But not std::unique_ptr , if I understand Rust's concept here.
Earlier quoted context omitted.
In the rust world "unsafe" is synonymous with "isn't proven to be safe by the compiler". Under this definition, every C/C++ program that uses pointers is "unsafe" because the languages make no memory safety guarantees.
Raw pointers, maybe. But not std::unique_ptr , if I understand Rust's concept here.
You can do linear lists and trees in Rust. The problem is backlinks. If you refcount everything, you can have backlinks, but otherwise there's a safety problem. Backlinks require an invariant which covers two variables, and you can't express that in Rust. Back pointers are a special sort of pointer from an ownership perspective. They don't carry ownership, but are locked in an invariant relationship with the pointer…
> The checker needs a few simple theorems, such as "valid_array(A,0,n) and valid_element(A[n+1)) implies valid_array(A,0,n+1)" to check this. This is basically dependent types. Rust doesn't really want to go that far from a language level. > Handling these two classes of unsafe code takes care of a sizable fraction of the unsafe code really needed in Rust. Not really. All the other datastructures in the stdlib would…
The sentence before your quote mentions ref-counting. This is referring to Weak.