What's everyone working on this week (29/2021)?

New week, new Rust! What are you folks up to?

Working on better performance for num: GitHub - Lemmih/num-criterion: Benchmarking for the num crates (rust).

There's room for a 10x improvement for small (less than, say, 192 bit) numbers. Large numbers are close to optimal when not targeting specific hardware.

Updating my oui library and probably doing some stuff with TLS 1.3

Still working on a uom (type-safe zero-cost dimensional analysis) PR to update to quickcheck 1.0. I hoping that that the last thing needed is fixes to TryFrom impls for Time<-->Duration.

I've compiled a bunch of stats about crates.io crates:

popular rust categories

10 Likes

Upgrading a gtk+ app to gtk4-rs. In the periods I split the backend logic into a separate crate.

Trying to convince people that Replace error-returning builders with typestate-based builders · Issue #330 · hyperium/http · GitHub is a good thing to fix. Lots of people use that crate, and I think it would be a positive chage for many, but most people use http through a whole stack of different crates and few looks at this crate itself. So if you use any web client or server library in rust, this may well affect you. Go there and read the issue, and have your say about if it should be done or not (and if the linked PR is a good way of doing it or not).

Also, I wrote a silly little tool: GitHub - kaj/mondae: An inverse daemon runner, for running things in the foreground just because I needed it and didn't know how to search for any existing implementation.

Continuing my research and prototyping a project I want to implement to parse and add features to my Markdown notes.

Related conversation: Text parsing with text replacement on match.

I have been goofing around with wasmtime/crates/lightbeam at main · bytecodealliance/wasmtime · GitHub and wasmer/lib/compiler-singlepass at master · wasmerio/wasmer · GitHub , trying to see how hard it is to add 64-bit memory address to the load/store operations.

Compiler-wise, wasmer-singlepass looks much more complete than wasmtime-lightbeam (currently does not look like it will pass wasm test quite). Complexity-wise, it looks much easier to hack on lightbeam (route from *.wast file -> x86_64 function is fairly straight forward: wast -> wasm -> lightbeam IR -> x86_64) compared to wasmer-singlepass (still digging through layers of indirections).

This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.