Hacker News Reader

Comment by mjg59 | Hacker News Reader

That seems to be a reference to https://lwn.net/Articles/893346/ - I'll let people come to their own conclusion about whether Drew's framing matches that, but I will say that I've never written a line of Rust in my life and certainly can't be described as representing the Rust community in any way. My position here is based on spending my days dealing with the consequences of bugs that we know how to get rid of - I just don't see the excuse for building something that isn't at least as good as solutions that already exist, especially when the consequences are potentially so significant.

View on HN

> I just don't see the excuse for building something that isn't at least as good as solutions that already exist,

Why does anyone need an excuse to build anything? He’s doing this in his free time. He doesn’t need the internet’s permission.

Hare is presented with the expectation that it would be included in Linux distros. With that framing, criticism on matters that are relevant to distro-included software is indeed relevant.

Where was it claimed or implied that criticism of this project isn’t relevant? My comment points out the flaws in the presumption that independently written software needs an excuse to exist.

If you've never written a line of Rust in your life, then your position could be more informed. I've written a lot of Rust, and see how much one sacrifices to get the memory safety guarantees without losing speed or memory safety. You can't make basic observers, RAII, dependency injection, back-references, or a lot of other useful patterns, without sacrificing speed or safety.

I'm experienced in security and rust and I agree with mjg. I've probably written 100s of thousands of lines of Rust, and I've been in security for well over a decade.

None of what you said is even true, all of those patterns are trivial, except "back-references" which are very slightly non-trivial.

And none of it is relevant. We don't need another memory unsafe language. It's fine if it's a toy, but this obviously isn't. It causes real harm.

Everything I said is true.

They all rely on having a mutable member reference to affect the outside world, which the borrow checker rejects. You can try to make it happen with a generic lifetime parameter for your structs, but you end up making invisible the thing you're pointing at, making it useless.

One can use unsafe to have shared mutability, or sacrifice speed with Rc/RefCell (increments/decrements) or Cell (copying, especially bad when copying Vecs which causes heap allocations).

Basic observers aren't possible. The closest thing we can get is some sort of modified observer-like substance that returns a command, which requires a lot of wiring and incidental complexity. [0] [1]

Dependency injection (the pattern, not the framework) isn't viable for the same reasons as observers: you can't have a mutable reference field without causing a lot of headaches elsewhere.

RAII isn't possible without sacrificing speed or safety. Most usages we see are backed by unsafe (FFI) or RefCell. We can't have multiple objects whose drop() affects the outside world, because we can't have multiple extant &mut references, and we can't just pass them in via parameters (because drop takes none).

Back references aren't possible because of the circular problem (having a mutable reference to your owner means nobody else can read it).

I'm not advocating for Hare, I'm just addressing the "I just don't see the excuse" remark, which can be ignorant of the costs of the borrow checker. The borrow checker is a great step forward, but not always a good tradeoff. An architect needs to be aware of these sacrifices before going all-in on a paradigm.

[0] https://stackoverflow.com/questions/37572734/how-can-i-imple...

[1] https://www.reddit.com/r/rust/comments/pwqju6/is_there_an_un...

RAII is built into Rust. DI is trivial and I use it all the time, idk what you're trying to say there. Graphs require Rc, backreferences are trivial with Rc::weak.

idk what tot ell you

You are fundamentally misunderstanding the relationship between Rust's unique and shared mutabilities and the usual mutability in, say, C. The Rust equivalent of C mutability is Cell. Conversely, the C equivalent of Rust's unique references is a `restrict` pointer.

Cell is meant to be applied at the "leaves" of a type where individual assignments take place. When used this way, there is no additional copying relative to the straightforward C version. (Bringing up Vec and heap allocations here is also complete nonsense; Cell has nothing to do with Clone.)

Once you wrap your head around that, all your shared mutability examples translate over to Rust trivially. All the "sacrifices" in speed you are imagining are relative to `restrict`/`&mut T`, not the usual baseline of a simple mutable object.

You can't use RAII in Rust? What on earth could this possibly mean? RAII is an extremely pervasive pattern in Rust and is fundamental to many of the safe APIs in the standard library.

To clarify, I was talking about making RAII, not using RAII. And it surprised me too, when I learned that the borrow checker rejects it.

To see it in action: Have a Database object, and try to have multiple Transaction objects that might commit something to it, in their drop().

It's unfortunately not possible, because they can't all have a &mut Database as struct fields.

We can sacrifice speed (by using Cell's copying or Rc's counting) or safety (by using unsafe). Most RAII we see uses unsafe FFI under the hood, which is why it was so surprising to me.

RAII is a general pattern for tying resource management to the lifetime of objects such that resource allocation is tied to value construction and resource deallocation is tied to value destruction. The smart pointers for allocation in the Rust standard library (Box, Rc, and Arc) are examples of the RAII pattern, since memory allocation happens at creation time (Box::new()) and memory deallocation happens when the Box goes out of scope (in drop()). Another example of RAII in the standard library is File: opening a file means creating a value of type File, and dropping that value means closing the file. Yet another example are the smart-pointer guards used for accessing RefCell and Mutex: RefCell::borrow() returns a Ref, and Mutex::lock() returns a MutexGuard; the underlying value can only be accessed while the guard exists, and access is relinquished when the guard is dropped. Given all this, it's absurd to say that Rust doesn't support RAII — RAII is fundamental to the design of many of Rust's safe APIs.

The very specific API design that you've described is not possible in Rust, but it is strange to equate this with the entirety of RAII. In any case, there are many alternative APIs (some with no sacrifice in speed or safety!) that are perfectly possible in Rust.

i think the real mistake was to have "exclusive references" be called "mutable references" in the language. I've taken the habit of saying "mut" as "mutually exclusive" for references. Of course you can't have each Statement keep an exclusive reference to a db object. They're exclusive!

You need shared references for your DB, implying you need interior mutability. This is how Statements are implemented in real-world rust database drivers such as rusqlite (any operation on a db is done through a shared reference). The fact that a very real package is doing it proves that the pattern you're talking about is, in fact, possible.

Rust is actually right, you cannot have multiple mutable references to a Database object without things going down the drain. (This is related to the fact that, like other comments said, &mut is an exclusive reference).

However, achieving something like what you want is still more than possible in Rust. You can do this with the pattern of 'interior mutability', which in its simplest form is just a Mutex. This allows upgrading a shared reference to an exclusive reference, so that you can safely mutate an object while upholding the expectations that a mutable reference is exclusive, and a non-mutable reference does not change from under your feet.

Of course, for a database, you will probably want a more advanced implementation of interior mutability, so that you can commit multiple transactions at the same time. (Or not, it seems to work quite well for SQLite.)

Reply on news.ycombinator.com