jhpratt

jhpratt, 6 days ago to rust

Hit an edge case in the #Rust compiler that appears to be quadratic at best and potentially exponential. Trivial to make happen with real-world code. Investigating...

#RustLang #programming

jhpratt, 6 days ago to rust

Now that it's public, I can say that I will be speaking (again) at #RustConf in Montréal! Catch me there or online in September.

#Rust #RustLang #programming

jhpratt, 7 days ago to firefox

I like #Firefox, but it is definitely trailing behind other browsers. PWAs, WebUSB, stacked/grouped tabs, and far more are present in a number of browsers and are all incredibly useful. Firefox has none of these. I thought translation was an extension because FF communication is crap.

I refuse to use Chrome out of principle, and Vivaldi is the only browser that even seems remotely what I want. But whenever I've tried it, I quickly ran into trivial, obvious bugs that I had reported years prior.

jhpratt, 12 days ago to rust

Preview of what I have been working on recently. The core of this crate is a mere two traits. The crate will ship with a number of parsers and combinators, none of which rely on anything not exposed to downstream users.

I've attached a real-world situation, taken verbatim from the test suite. Parsing integers isn't as efficient as it could be yet, as it's using a naïve method.

Parsing in general compiles to be extremely efficient, and using it is ergonomic.

#Rust #RustLang #programming

pub trait Combinator<'input, P> where P: ByteParser<'input>, { type NewParser: ByteParser<'input>; fn apply_to(self, parser: P) -> Self::NewParser; }
#[test] fn parse_date() -> parcom::Result<()> { let parser = ascii::int:: .limit_length(4) .and(verbatim(b"-").discard()) .and(ascii::int::.limit_length(2)) .and(verbatim(b"-").discard()) .and(ascii::int::.limit_length(2)); let (seq!(year, _, month, _, day), remaining) = parser.parse("2021-07-04")?.into_parts(); assert!(remaining.is_empty()); assert_eq!(year, 2021); assert_eq!(month, 7); assert_eq!(day, 4); Ok(()) }

jhpratt, 15 days ago to rust

So now that -Zcheck-cfg is stabilized and enabled by default, there's no way to tell the compiler about custom cfgs other than a build script?

I routinely use a custom cfg for rustdoc so that I can use nightly features when publishing (both with docs.rs and CI). #[cfg(coverage)] is also used in some places.

Apparently spurious warnings are extremely wide-spread, so I'm not clear on why this was stabilized with such a glaring hole in it.

#Rust #RustLang

jhpratt, 21 days ago to rust

In #Rust, how can I leverage the type system to enforce that one object originates from another? Even requiring the lifetime be exact ('a: 'b and 'b: 'a) wouldn't work.

Basically, I have struct Bytes<'a>(&'a [u8]); and want to prevent someone from creating an arbitrary value that can be swapped in for the correct one. Yet at the same time it is essential that arbitrary values can be created.

Basically I'm trying to enforce my own version of provenance…

#RustLang

jhpratt, 21 days ago to random

Yet another round of layoffs at Tesla, and this time I'm part of it.

If anyone is aware of full time work that I may be interested in, let me know! I'm located in the US (upstate NY specifically) and an open to relocation, though I would prefer remote.

#FediHire

jhpratt, 22 days ago to random

As of today, Donald Trump is a convicted criminal.

ekuber, 1 month ago to rust

Somewhat cursed #RustLang idea: "attributes at a distance"

#![attr(target = crate::foo::Bar)]

mod foo {  
 struct Bar;  
}  

What would you use this for?

jhpratt, 1 month ago to random

What happened with #xz is precisely why I ceased support of serde versions that had pre-compiled binaries included. It doesn't matter how trustworthy someone seems or how long they've been maintaining a project.

I caught flak from a number of people, including one person (in person, and respectfully!) saying that I was "policing the ecosystem". Which...yes, that's the point. Unknown, unverified binaries being executed is unacceptable and dangerous until proven safe, not the other way around.

jhpratt, 2 months ago to rust

Looking back at some of the #Rust crates that I've published as spin-offs from time:

• deranged (range-bounded integers)
• num-conv (avoiding as casts)
• num_threads (thread count for current process)
• powerfmt (making it easier to implement Display while following formatting flags)
• standback (backporting the standard library to older compilers)

Only standback is no longer maintained, as I don't need it any more. All the others are maintained, with num_threads having a co-maintainer.

jhpratt, 2 months ago to rust

For those that wanted a comparison between the time and chrono crates, here's an excellent overview of some differences between the crates:

https://users.rust-lang.org/t/the-state-of-time-in-rust-leaps-and-bounds/107620?u=jhpratt

#Rust #RustLang #programming

jhpratt, 3 months ago to rust

Anyone down for writing up a comparison between time and chrono? I'm not aware of a full comparison anywhere, and it's far from a priority for me. I can check the accuracy of anything on time's end of things.

If you're interested, please respond here or https://github.com/time-rs/time/issues/308.

#Rust #programming

jhpratt, 3 months ago to rust

#RustConf videos have been posted! Check out my talk among others.

https://www.youtube.com/watch?v=MTnIexTt9Dk&list=PL85XCvVPmGQgR1aCC-b0xx7sidGfopjCj

#Rust #RustLang #programming #conference #presentation