Hacker News Reader

Comment by tyingq | Hacker News Reader

Ahh, okay. So maybe a "c++ common approach" vs "rust best practices" FAQ? Like "maybe you don't need a c++ like doubly linked list, and here's why" write-up?

View on HN

Linked lists are in the standard library[1]. The source code is linked to from that page (`src`, on the top-right corner), although the implementation is optimized so it's probably not a great learning resource.

If, for some reason, you actually need a linked list there are tons of implementations available on crates.io[2].

It's not that there aren't linked lists, and it's not that they can't be implemented as efficiently as in C, it's that implementing them is an inherently tricky problem where the obvious implementation can't be proved by Rust's type system so they're an extremely unpleasant way to learn Rust.

1: https://doc.rust-lang.org/std/collections/struct.LinkedList.... 2: https://crates.io/search?q=linked%20list

I'm not sure that really addresses my question. You seem to be saying C++ folks need to bear the burden of figuring it out. That's fine, but it explains the conflict. If the Rust's team position is that C++ diehards need to fully understand Rust first, then the conflict seems expected. On the other hand, if the Rust team is interested in serious evangelism, something seems missing.

>"maybe you don't need a c++ like doubly linked list, and here's why" write-up?

Done and done, you're welcome:

http://cglab.ca/~abeinges/blah/too-many-lists/book/

I think it's more that there's one in the standard library, so you don't actually need to write one in Rust, but if you look at how they did it you'll find that it was written in unsafe Rust because that's how you do things that rely on pointers for their semantics.

A write-up for the latter: https://isocpp.org/blog/2014/06/stroustrup-lists

High level summary: linked lists are usually not the best data structure.

Reply on news.ycombinator.com