[arithmetic_side_effects] Fix #11266 #12710
Conversation
rustbot
added
the
S-waiting-on-review
c410-f3r
changed the title
tests/ui/arithmetic_side_effects.rs
Outdated
| pub fn issue_11266<const N: usize>() { | ||
| let _ = N + 1; | ||
| } |
There was a problem hiding this comment.
Why don't we want to lint this? It doesn't seem like rustc generates a warning for the example code, or am I missing something?
Something like N + random() could still run into the same issues in that it could overflow without the compiler possibly knowing it, can't it?
There was a problem hiding this comment.
Changed to only ignore when literals are involved.
There was a problem hiding this comment.
I see. But even then, rustc doesn't seem to catch the case with literals in my example and doesn't issue a warning for it.
Maybe I'm misunderstanding why this change is done, but the way I understood it was, "this lint doesn't need to warn const generic operations because rustc has its own arithmetic_overflow lint that is more accurate for consts and will issue a warning/error if it sees overflow after monomorphization when all const generic parameters are instantiated with concrete values", but it doesn't seem to do that, so now there is no warning for N + 1 emitted from either side even for usize::MAX.
There was a problem hiding this comment.
I see. But even then, rustc doesn't seem to catch the case with literals in my example and doesn't issue a warning for it.
There was a problem hiding this comment.
The top one is only a runtime panic though. And the bottom one will be gone with this change presumably? So that would leave us with no compile time warnings/errors
There was a problem hiding this comment.
Yes, you are right. Thanks for pointing this out
Fix #11266
changelog: [
arithmetic_side_effects]: Let rustc deal with possible overflows involving constant paramenters