3 ms·

I'm not sure how easy Rust claims to be... The main idea here is " I love that Haskell-esque feeling where a compiling program is usually a working program." Ru

I'm not sure how easy Rust claims to be... The main idea here is " I love that Haskell-esque feeling where a compiling program is usually a working program." Rust will probably get easier, but it's main goal is safety. The main problem I have with a pure safety focus is that the warnings often feel like overkill for 99% of programs. For example, when gcc yells about comparing signed and unsigned ints, I find that there rarely is a true safety concern involved. I hope Rust is better than that.

pjmlp 9y ago
Well gcc cannot know if your signed with unsigned comparison is going to blow up on a critical path of a security related function.
grayrest 9y ago
> For example, when gcc yells about comparing signed and unsigned ints, I find that there rarely is a true safety concern involved. I hope Rust is better than that. In Rust these are treated as two completely different types so this is a compiler error and not a warning. You need to explicitly cast one of them to make the comparison and integer types are not automatically converted in any way. You have to do the same for any other operation involving the two types (e.g. addition). You could certainly define the `ParitalOrd` trait for the two types to make the comparison possible (I don't know the implications off the top of my head). The language defaults to safe and explicit.
- lomnakkus 9y ago
  > You could certainly define the `ParitalOrd` trait for the two types to make the comparison possible I don't believe that's possible. Unless something changed relatively recently, the trait implementation has to be either a) in the module where the trait itself is defined (so the standard library in this case), or b) in the module where the type you're implementing the trait for is defined (again, the standard library). Since you can just modify the standard library, implementing PartialOrd would not be possible for e.g. comparing u16/s16 (or whatever).
jcelerier 9y ago
> For example, when gcc yells about comparing signed and unsigned ints, I find that there rarely is a true safety concern involved except for the small detail that stuff like "-1 < 1" is false when comparing int and uint because the int gets casted as the latter ? -Wsign-compare should be used as -Werror=sign-compare

pjmlp 9y ago

Well gcc cannot know if your signed with unsigned comparison is going to blow up on a critical path of a security related function.

grayrest 9y ago

> For example, when gcc yells about comparing signed and unsigned ints, I find that there rarely is a true safety concern involved. I hope Rust is better than that. In Rust these are treated as two completely different types so this is a compiler error and not a warning. You need to explicitly cast one of them to make the comparison and integer types are not automatically converted in any way. You have to do the same for any other operation involving the two types (e.g. addition). You could certainly define the `ParitalOrd` trait for the two types to make the comparison possible (I don't know the implications off the top of my head). The language defaults to safe and explicit.

lomnakkus 9y ago
> You could certainly define the `ParitalOrd` trait for the two types to make the comparison possible I don't believe that's possible. Unless something changed relatively recently, the trait implementation has to be either a) in the module where the trait itself is defined (so the standard library in this case), or b) in the module where the type you're implementing the trait for is defined (again, the standard library). Since you can just modify the standard library, implementing PartialOrd would not be possible for e.g. comparing u16/s16 (or whatever).

jcelerier 9y ago

> For example, when gcc yells about comparing signed and unsigned ints, I find that there rarely is a true safety concern involved except for the small detail that stuff like "-1 < 1" is false when comparing int and uint because the int gets casted as the latter ? -Wsign-compare should be used as -Werror=sign-compare