Jake Goulding

aarch64-apple-darwin is now a Tier-2 #rustlang target. You can use it with rustup, cross-compile from x86-64, or compile natively on appropriate hardware! #issuecomment-710101535" target="_blank" rel="nofollow noopener">github.com/rust-lang/rust…

If you know anyone at @GoogleAds, there's an actively exploited ad bug that obfuscates a malicious link as a legit one. The google ad ID is DChcSEwjIk--7mtb7AhVU-FEKHZLhCqMYABAAGgJ3cw Also anyone at @Atlassian, please get the repo taken down.

@JonTDean Make the meetings low key. Our last IRL meetup before the pandemic was weekly at a Panera. Bring a laptop and work on something, or talk to the person next to you, or help them work on something. Big presentations rarely if ever.

It’s kind of neat when you see something you wrote end up as the title of someone’s blog post. reddit.com/r/programming/…

@Pgh311 green light out at corner of Baum and Graham.

@erichdongubler @yaahc_ @ekuber @seritools I actually prefer impl trait in arg position _when it applies_. I think it most matches what people expect (actually would be ok with `foo: SomeTrait` eventually meaning `impl SomeTrait` instead of olde `dyn SomeTrait`). However inline bounds can go to recycle can in the sky.

@yaahc_ @ekuber Big +1 to this. CC @JakeGoulding for originally corrupting me on `where`-only (I can't even remember the PR now), and @seritools who I know would be...engaged with this hot take. 🤓

@yaahc_ @rustlang Only if you restrict yourself to homogenous success values, yeah? play.integer32.com/?version=stabl…

@JakeGoulding @rustlang you can kinda get most of the way there with `Itertools::partition_result` play.rust-lang.org/?version=stabl…

Hey @rustlang community, show me your best or worst error handling code in rust I'm doing an analysis of the strengths, weaknesses, best pratices, and footguns of rust error handling for the error handling project group and could use some more examples.

@yaahc_ @rustlang I really want something akin to async’s `join` for errors: run these N things concurrently and wait for them all to succeed or error. If any errored, combine them and bubble up, else get all the success values.

@yaahc_ @rustlang Amusingly, a place I’m not happy with Error handling is in the implementation of SNAFU itself. Like others have mentioned, there’s times you want to collect as many errors as possible. #L194" target="_blank" rel="nofollow noopener">github.com/shepmaster/sna…

@yaahc_ @colinfinck @eugeeeeny @rustlang @ekuber @pnkfelix Perhaps, but that example of anonymous enums would run counter to SNAFU’s principle that an io::Error from opening a file and an io::Error from writing to a socket are semantically dissent. It only allows for a type to be in the enum once.

@JakeGoulding @colinfinck @eugeeeeny @rustlang there's a vaguely related long conversation from @ekuber and @pnkfelix yesterday on anon enums and subtyping twitter.com/ekuber/status/…

@colinfinck @eugeeeeny @yaahc_ @rustlang I want to figure out how to enable this with SNAFU #issuecomment-730103994" target="_blank" rel="nofollow noopener">github.com/shepmaster/sna…

@eugeeeeny @yaahc_ @rustlang This. Had a debate with a coworker about one custom error enum for the entire lib vs. one for each function. Latter style is currently the only one to exhaustively tell the app what errors can happen and what not. But it requires more code, conversions, and sometimes duplications

Both excited and terrified about this github.com/integer32llc/r… The #rustlang playground is now running on axum

@shereifHawary I’m not at a computer now, but my gut says that the string allocation can be reduced/avoided. The easy fix is to reuse `input` across calls to `next`. The harder one is removing the string from the split.

@JakeGoulding I had my first problem solving submission with #rust codeforces.com/contest/69/sub… one thing, the same problem in java8 got 300ms time while in rust got 60 ms (other solutions got 30ms)

@DasSurma For Rust, this was deliberate. github.com/rust-lang/rfcs… I’m quite surprised I didn’t include JS in the comparison though.

I am *particularly* annoyed by JavaScript today

@minigamedev I've just been working on various things here and there; time is limited after all 😅

@minigamedev "Don't call it a comeback, I been here for years"

I just released SNAFU version 0.7, my #rustlang error definition and creation library. Thanks to hard work from many great contributors, SNAFU now has improved learnability and ergonomics! users.rust-lang.org/t/snafu-0-7-re…

@smortimerk @loige Can you describe what’s surprising about it? A few people have been manually doing namespacing like this, but I don’t think anyone has brought up prefixing before.

@JakeGoulding @loige I suppose the `snafu(module)` approach is recommended in this case, though it reads a little surprising to me.

Having fun with the `thiserror` crate in #Rust. It makes creating comprehensive error types soooo much easier. Check it out 😇 docs.rs/thiserror/ 🦀👀🦀

@smortimerk @loige It isn’t possible or planned right now (which is different from planned to *not* happen!)

@JakeGoulding @loige Wow, that really does handle a lot of new cases; I'm not sure how long suffixing has been available, but that's a welcome addition. Quick question about multiple error enums, is it possible/planned to use the enum name as a prefix/suffix instead?

@loige @smortimerk Make sure to use version 0.7 (currently in beta) as it smooths some rough edges especially felt by new users. docs.rs/snafu/0.7.0-be…

@smortimerk @JakeGoulding Thanks for the suggestion, Seán! I never used snafu but it looks pretty interesting with that context approach (it would save me from a few map_err). I also found someone requesting the same feature for thiserror: github.com/dtolnay/thiser… 😅