01 February 2023
Nightly Rust now has support for async functions in traits, so long as you limit yourself to static dispatch. That’s super exciting! And yet, for many users, this support won’t yet meet their needs. One of the problems we need to resolve is how users can conveniently specify when they need an async function to return a
Send future. This post covers some of the background on send futures, why we don’t want to adopt the solution from the
async_trait crate for the language, and the general direction we would like to go. Follow-up posts will dive into specific solutions.
20 January 2023
When I started working on Rust in 2011, my daughter was about three months old. She’s now in sixth grade, and she’s started growing rapidly. Sometimes we wake up to find that her clothes don’t quite fit anymore: the sleeves might be a little too short, or the legs come up to her ankles. Rust is experiencing something similar. We’ve been growing tremendously fast over the last few years, and any time you experience growth like that, there are bound to be a few rough patches. Things that don’t work as well as they used to. This holds both in a technical sense — there are parts of the language that don’t seem to scale up to Rust’s current size — and in a social one — some aspects of how the projects runs need to change if we’re going to keep growing the way I think we should. As we head into 2023, with two years to go until the Rust 2024 edition, this is the theme I see for Rust: maturation and scaling.(read more...)
22 September 2022
I’ve been thinking about what “Rust 2024” will look like lately. I don’t really mean the edition itself — but more like, what will Rust feel like after we’ve finished up the next few years of work? I think the answer is that Rust 2024 is going to be the year of “everywhere”. Let me explain what I mean. Up until now, Rust has had a lot of nice features, but they only work sometimes. By the time 2024 rolls around, they’re going to work everywhere that you want to use them, and I think that’s going to make a big difference in how Rust feels.(read more...)
21 September 2022
After my last post on dyn async traits, some folks pointed out that I was overlooking a seemingly obvious possibility. Why not have the choice of how to manage the future be made at the call site? It’s true, I had largely dismissed that alternative, but it’s worth consideration. This post is going to explore what it would take to get call-site-based dispatch working, and what the ergonomics might look like. I think it’s actually fairly appealing, though it has some limitations.(read more...)
19 September 2022
Re-reading my previous post, I felt I should clarify why I called it the “soul of Rust”. The soul of Rust, to my mind, is definitely not being explicit about allocation. Rather, it’s about the struggle between a few key values — especially productivity and versatility1 in tension with transparency. Rust’s goal has always been to feel like a high-level but with the performance and control of a low-level one. Oftentimes, we are able to find a “third way” that removes the tradeoff, solving both goals pretty well. But finding those “third ways” takes time — and sometimes we just have to accept a certain hit to one value or another for the time being to make progress. It’s exactly at these times, when we have to make a difficult call, that questions about the “soul of Rust” starts to come into play. I’ve been thinking about this a lot, so I thought I would write a post that expands on the role of transparency in Rust, and some of the tensions that arise around it.
I didn’t write about versatility in my original post: instead I focused on the hit to productivity. But as I think about it now, versatility is really what’s at play here — versatility really meant that Rust was useful for high-level things and low-level things, and I think that requiring an explicit dyn adaptor is unquestionably a hit against being high-level. Interestingly, I put versatility after transparency in the list, meaning that it was lower priority, and that seems to back up the decision to have some kind of explicit adaptor. ↩
18 September 2022
In the last few months, Tyler Mandry and I have been circulating a “User’s Guide from the Future” that describes our current proposed design for async functions in traits. In this blog post, I want to deep dive on one aspect of that proposal: how to handle dynamic dispatch. My goal here is to explore the space a bit and also to address one particularly tricky topic: how explicit do we have to be about the possibility of allocation? This is a tricky topic, and one that gets at that core question: what is the soul of Rust?(read more...)