New week, new Rust! What are you folks up to?
I'm hoping to finalise a few proc-macros(try_v2, exit_safely) that I started on in true "pass the salt (xkcd)" style. Maybe I'll also break out the half-finished implementation of Diagnostic for proc-macro2 that I embedded into them too...
Then I'll take another look at trying to get an async backend and an event-loop front-end to play nicely together in a way that compiles to wasm (a personal project I put on the back burner 9 months ago). I only managed to get 2 out of 3 last time I looked at it ;(
I got around to updating Rust Learning to account for various Rust developments...
- Builtin supertrait upcasting
- Precise capturing and changes to default RPIT captures
_being available to inferconstparametersTypeId::ofbeing aconst fn- Being able to omit
where Self: Sizedtrait methods in unsized type impls - Being able to coerce away the
dynprincipal trait mismatched_lifetime_syntaxeswarn-by-default lintelided_lifetimes_in_associated_constantfuture compatibility lint
Later this week I'll reread the guide with fresher eyes in an attempt to clean up the various typos and poor wordings I inevitably also introduced.
Have been extending/improving pstd, the crate doc now starts:
Crate with parts of Rust std library ( different implementations, features not yet stabilised etc ), in particular Box, Vec, Rc, String and collections::{ BTreeMap, BTreeSet, HashMap, HashSet }.
Box, Rc and String currently only have minimal methods/traits implemented.
HashMap and HashSet are imported from the hashbrown crate.
RcStr is a reference-counted string.
The localalloc module has fast thread-local bump allocators.
I have been building a JSON module for Python using only Rust and extremely minimal unsafe. It already passes a JSON test suite and benchmarks as about 2.4x faster than Python's json module, which I'm quite happy with considering how little time I have spent optimizing.
I am building a high performance, multi-modal database - NodeDB. Still have a long way to go though.
I'm trying to review and update bincode-next async-fiber module to make it more robust and secure. Though it is a hard job because the module heavily rely on unsafe and even naked_asm.