I feel I gotta point out it’s a pretty funny example—email comes up so frequently as a thing that you’re recommended to neither parse nor validate, just try to send an email to the address and see if it works. If you need to know that it was received successfully, a link to click is the general method.
But “parse, don’t validate” is still a generally good idea, no matter the example used. :)
Yeah, you generally just want the same auto-stuff done as would be enforced in CI anyway.
… all the other stuff you could fix but wind up just ignoring because your team ignores it will just glare at you until you sneak it in somehow
It’s just a monoid object in a category of endofunctors, no biggie
Afaik they’re hoping to land it on nightly in 2025H1.
Between that, work on the next-generation trait solver and promoting parallel frontend, there’s some stuff to look forward in the compiler this year.
Do also note that by saying that some % of crates use unsafe, it’s not implied that 100% of the code in that crate is marked unsafe. It could be as little as one line; it could be a whole lot; it could be well-documented and tested; it might not be. (This is part of what the talk is about.)
It’s also rather to be expected that there’s more unsafe in embedded. As Steve Klabnik gets into in How to Do Embedded Development with Rust (GOTO 2023), it’s used when you e.g. want to set a certain memory address to a certain value, which in a lot of contexts is nonsense, but in some contexts makes a LED light up.
Yeah, the article comes off as needing so much context that the article itself is sus. Like
Hejlsberg stated the obvious when saying that TS isn’t the fastest language. Although it can laughably run Doom at 0.0000009645 fps.
… which is referencing an implementation of Doom in the TS type system. It’s a funny idea, but an arbitrary reader who doesn’t know about that and doesn’t bother clicking through will get a very wrong impression.
The reimplementation (which they’ve done partially automated; Go apparently lets them do a very simple translation while Rust or C# would require more work to fit) should be a boon for TS devs, but not noticeable for those who just run stuff that happens to be written in TS.
Would be kinda interesting to see the effect if stuff targeted deno rather than node, though.
>What is C++? A miserable huge pile of "should"s
Parsing is a way of “validating early”. You either get a successful parse and the program continues working on known-good data with that knowledge encoded in the type system, or you handle incorrect data as soon as it’s encountered.