Update to edition 2018

[kornelski]

This crate is one of the last ones on the edition 2015.

[sdroege]

Thanks!

[sdroege]

Looks like this needs a MSRV-specific Cargo.lock that has to be maintained manually for the foreseeable future 😢

[sdroege]

There is no lock file. Can you add a Cargo.lock.msrv or so that has versions of everything that would compile with 1.31, switch this here back to 1.31 and for the 1.31 build copy that lock file in place so it's used?

[kornelski]

I'm unable to generate a lockfile in a format old enough — the oldest Rust that exists for my platform is 1.46.

BTW, 73% of all crates are incompatible with <1.36, 97% of recently updated crates require 1.56+, and 99.5% crates.io requests are from Rust 1.65 or later.

https://lib.rs/stats#rustc-usage

[sdroege]

Ok, I'll take a look. Don't worry about that :)

I'm aware of those numbers and it's also ridiculous to me that some people can update random crates but not their toolchain, but here we are

[kornelski]

I don't think there's a single person able to use Rust 1.30 today. This is way below even Debian oldstable.

Even if such user exists, they already must have frozen all their dependencies and won't be able update any of them, so they won't update pkg-config either. There's only 6% of crates that work with 1.30 in their latest version, and majority of them are complete junk (reserved names, hello ferris, winapi crates that don't compile with Rust 1.x).

Rust 1.36 is still barely usable at all, and is not supported by any recent crates. Much later 1.56 is the baseline in practice, and even that is painful and requires careful management of lockfiles to downgrade most deps.

Rust 1.80 broke most versions of the time crate, so now Rust is incompatible with most non-trivial projects having Cargo.lock compatible with Rust <1.67, and I expect that to push MSRV of most crates even higher.

syn is 1.61, bitflags is 1.56, tokio is 1.70, unicode-width 1.62, base64 is 1.48, serde_json 1.56.

[sdroege]

I agree with all you said. If it was for me, I would always just support latest stable. But I don't have the energy anymore to argue with people that have stupid requirements and somehow need a much older version, so if it's easy enough (like here) I rather do that.