Hacker News Reader

Comment by jen20 | Hacker News Reader

The `dirs` crate has been recently forked and revived - the new home is here: https://github.com/xdg-rs/dirs (and the crate is named `dirs-next`).

Given the relatively sparse nature of the standard library, and the cost of making mistakes (std::sync::mpsc, anyone?!) I can see why this is omitted. IMO it’s a fair criticism that pulling in lots of crates is less than ideal, though.

View on HN

> std::sync::mpsc, anyone?!

Can anyone expand on this? I'm not well versed in Rust lore and have just used that module in a recent learning project.

Found this post https://stjepang.github.io/2019/03/02/new-channels.html

with a link to a comment on GitHub: https://github.com/rust-lang/rust/pull/42397#issuecomment-31...

Yes - sorry, this is the comment I was referring to. I'm not aware of _bugs_ in `std::sync::mpsc`, more concerns around things which could have been done better which are now not possible to retrofit because of the guarantees of the standard library with respect to backwards compatibility.

It's cool, and was included probably because Go has channels and that user-base is valuable. But now there's crossbeam-channel, which is significantly faster, has handles that are more shareable, and it removes the `Result` wrapper around all the operations. So a drop-in replacement is technically possible, but doesn't get rid of `unwrap` everywhere, and there's no steam behind replacing the guts of it because the lesson that was learned says 'just deprecate it, let it live in a crate'.

Reply on news.ycombinator.com