It is (or should be) read as slightly hyperbolic, as there isn't to my knowledge a language that actually works that well. I could water down my message by saying "it mostly works", but then I fail to convey the main reason why I posted in the first place.
I could also instead spend a lot of time and words to explain things about Rust's design that I presume anyone who is slightly interested in the language would already know (specifically: lack of null, strict ownership checking, having to explicitly deal with errors), in order to explain that the language succeeds at solving some of the problems it was specifically designed to solve. But then I would just be repeating things that HN readers presumably already know. Instead I can use a shortcut in my communication which is perfectly understandable if you assume minimal intelligence, and take the other person's comment in good faith.
See, by the time we've reached the bottom of this wall of text, anyone who read this far through my intentional rambling has presumably forgotten my initial point: I had a positive experience with the language.
And at the same time, I've heard people say "when it compiles, it works" about much weaker type systems, like Go's. It just seems to mean "this language catches more errors at compile time than the previous language I used."
> It just seems to mean "this language catches more errors at compile time than the previous language I used."
That is a fair interpretation. I've programmed mostly in C# and Java because that's what was required at the time. I also know enough C and C++ to aim at my toes instead of the entire foot. So the comparison is between strongly typed imperative programming languages which are syntactically close to Rust.
Also, I like that Rust isn't OoP, but try to not get baited into that discussion.
That's not even true for languages with dependent types, which Rust lacks.