feat(manual_ilog2): new lint #15865
feat(manual_ilog2): new lint #15865
Conversation
rustbot
added
needs-fcp
S-waiting-on-review
|
@Jarcho you might be interested in this as you reviewed the original PR |
|
r? clippy |
clippy_lints/src/manual_ilog2.rs
Outdated
| // Whether `leading_zeros` is an inherent method, i.e. doesn't come from some unrelated trait | ||
| && cx.ty_based_def(right).opt_parent(cx).is_impl(cx) |
There was a problem hiding this comment.
I believe this wouldn't catch that since a trait impl is also a DefKind::Impl. I think it's fine without the check though since inherent methods come first in method resolution
There was a problem hiding this comment.
Oh, it looks like you're right; I'll remove the call then.
The reason I wrote it that way is that that's what is_inherent_method_call was replaced with in https://github.com/rust-lang/rust-clippy/pull/15682/files#diff-9d3f62135af85bbb426d0e358666790f639f6777c8dc504a7527e2bc7f830fb1R741, and so I assumed it would only match, well, inherent method calls. @Jarcho maybe this should be a separate method, something like is_inherent_impl?
There was a problem hiding this comment.
TypeckResults always has things resolved to either the inherent item or the trait item, but never a trait impl item. Checking that the parent is an impl is enough because of this.
Alexendoo
added
S-final-comment-period
This comment has been minimized.
This comment has been minimized.
clippy_lints/src/manual_ilog2.rs
Outdated
| && ilog.ident.name == sym::ilog | ||
| && let ExprKind::Lit(lit) = two.kind | ||
| && let LitKind::Int(Pu128(2), _) = lit.node | ||
| && cx.typeck_results().expr_ty(recv).is_integral() |
There was a problem hiding this comment.
This should be the adjusted type. You want the type passed in to the method.
This comment has been minimized.
This comment has been minimized.
| && self.msrv.meets(cx, msrvs::ILOG2) | ||
| && !is_from_proc_macro(cx, expr) => |
There was a problem hiding this comment.
I'd move this to the start of the function, since they're really not part of the 31 - x.leading_zeros() case.
Idem below
There was a problem hiding this comment.
That would remove some duplication, yes, but the problem is that both of these checks are rather expensive:
- the MSRV check requires walking up through all the attributes, and I think also going through the Clippy config
- the proc macro check is very expensive I'm told (though I'm not completely sure why), and so it should generally be the last
ifof any lint
There was a problem hiding this comment.
These checks aren't first in general as a perf thing. They're bad filters in general and neither are fast.
There was a problem hiding this comment.
Thanks for the explanation, sorry for the noise.
This comment has been minimized.
This comment has been minimized.
|
This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed. Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers. |
6 participants
Revival of #13331
Part of #12894
changelog: [
manual_ilog2]: new lint