Auto merge of rust-lang#12258 - tgross35:similiar-names-update, r=Alexendoo

[`similar_names`]: don't raise if the first character is different

A lot of cases of the "noise" cases of `similar_names` come from two idents with a different first letter, which is easy enough to differentiate visually but causes this lint to be raised.

Do not raise the lint in these cases, as long as the first character does not have a lookalike.

Helps with rust-lang/rust-clippy#10926 (does not fix)

This is per-commit reviewable, the first commit is just refactoring.

changelog: [`similar_names`]: don't raise if the first character is different

Author: bors

clippy_lints/src/item_name_repetitions.rs

@@ -385,7 +385,6 @@ impl LateLintPass<'_> for ItemNameRepetitions {
         assert!(last.is_some());
     }
 
-    #[expect(clippy::similar_names)]
     fn check_item(&mut self, cx: &LateContext<'_>, item: &Item<'_>) {
         let item_name = item.ident.name.as_str();
         let item_camel = to_camel_case(item_name);

clippy_lints/src/non_expressive_names.rs

@@ -119,7 +119,6 @@ impl<'a, 'tcx> SimilarNamesLocalVisitor<'a, 'tcx> {
 
 // this list contains lists of names that are allowed to be similar
 // the assumption is that no name is ever contained in multiple lists.
-#[rustfmt::skip]
 const ALLOWED_TO_BE_SIMILAR: &[&[&str]] = &[
     &["parsed", "parser"],
     &["lhs", "rhs"],
@@ -132,6 +131,14 @@ const ALLOWED_TO_BE_SIMILAR: &[&[&str]] = &[
     &["iter", "item"],
 ];
 
+/// Characters that look visually similar
+const SIMILAR_CHARS: &[(char, char)] = &[('l', 'i'), ('l', '1'), ('i', '1'), ('u', 'v')];
+
+/// Return true if two characters are visually similar
+fn chars_are_similar(a: char, b: char) -> bool {
+    a == b || SIMILAR_CHARS.contains(&(a, b)) || SIMILAR_CHARS.contains(&(b, a))
+}
+
 struct SimilarNamesNameVisitor<'a, 'tcx, 'b>(&'b mut SimilarNamesLocalVisitor<'a, 'tcx>);
 
 impl<'a, 'tcx, 'b> Visitor<'tcx> for SimilarNamesNameVisitor<'a, 'tcx, 'b> {
@@ -189,7 +196,6 @@ impl<'a, 'tcx, 'b> SimilarNamesNameVisitor<'a, 'tcx, 'b> {
         }
     }
 
-    #[expect(clippy::too_many_lines)]
     fn check_ident(&mut self, ident: Ident) {
         let interned_name = ident.name.as_str();
         if interned_name.chars().any(char::is_uppercase) {
@@ -219,71 +225,28 @@ impl<'a, 'tcx, 'b> SimilarNamesNameVisitor<'a, 'tcx, 'b> {
             if allowed_to_be_similar(interned_name, existing_name.exemptions) {
                 continue;
             }
-            match existing_name.len.cmp(&count) {
-                Ordering::Greater => {
-                    if existing_name.len - count != 1
-                        || levenstein_not_1(interned_name, existing_name.interned.as_str())
-                    {
-                        continue;
-                    }
-                },
-                Ordering::Less => {
-                    if count - existing_name.len != 1
-                        || levenstein_not_1(existing_name.interned.as_str(), interned_name)
-                    {
-                        continue;
-                    }
-                },
-                Ordering::Equal => {
-                    let mut interned_chars = interned_name.chars();
-                    let interned_str = existing_name.interned.as_str();
-                    let mut existing_chars = interned_str.chars();
-                    let first_i = interned_chars.next().expect("we know we have at least one char");
-                    let first_e = existing_chars.next().expect("we know we have at least one char");
-                    let eq_or_numeric = |(a, b): (char, char)| a == b || a.is_numeric() && b.is_numeric();
-
-                    if eq_or_numeric((first_i, first_e)) {
-                        let last_i = interned_chars.next_back().expect("we know we have at least two chars");
-                        let last_e = existing_chars.next_back().expect("we know we have at least two chars");
-                        if eq_or_numeric((last_i, last_e)) {
-                            if interned_chars
-                                .zip(existing_chars)
-                                .filter(|&ie| !eq_or_numeric(ie))
-                                .count()
-                                != 1
-                            {
-                                continue;
-                            }
-                        } else {
-                            let second_last_i = interned_chars
-                                .next_back()
-                                .expect("we know we have at least three chars");
-                            let second_last_e = existing_chars
-                                .next_back()
-                                .expect("we know we have at least three chars");
-                            if !eq_or_numeric((second_last_i, second_last_e))
-                                || second_last_i == '_'
-                                || !interned_chars.zip(existing_chars).all(eq_or_numeric)
-                            {
-                                // allowed similarity foo_x, foo_y
-                                // or too many chars differ (foo_x, boo_y) or (foox, booy)
-                                continue;
-                            }
-                        }
-                    } else {
-                        let second_i = interned_chars.next().expect("we know we have at least two chars");
-                        let second_e = existing_chars.next().expect("we know we have at least two chars");
-                        if !eq_or_numeric((second_i, second_e))
-                            || second_i == '_'
-                            || !interned_chars.zip(existing_chars).all(eq_or_numeric)
-                        {
-                            // allowed similarity x_foo, y_foo
-                            // or too many chars differ (x_foo, y_boo) or (xfoo, yboo)
-                            continue;
-                        }
-                    }
-                },
+
+            let existing_str = existing_name.interned.as_str();
+
+            // The first char being different is usually enough to set identifiers apart, as long
+            // as the characters aren't too similar.
+            if !chars_are_similar(
+                interned_name.chars().next().expect("len >= 1"),
+                existing_str.chars().next().expect("len >= 1"),
+            ) {
+                continue;
             }
+
+            let dissimilar = match existing_name.len.cmp(&count) {
+                Ordering::Greater => existing_name.len - count != 1 || levenstein_not_1(interned_name, existing_str),
+                Ordering::Less => count - existing_name.len != 1 || levenstein_not_1(existing_str, interned_name),
+                Ordering::Equal => Self::equal_length_strs_not_similar(interned_name, existing_str),
+            };
+
+            if dissimilar {
+                continue;
+            }
+
             span_lint_and_then(
                 self.0.cx,
                 SIMILAR_NAMES,
@@ -302,6 +265,57 @@ impl<'a, 'tcx, 'b> SimilarNamesNameVisitor<'a, 'tcx, 'b> {
             len: count,
         });
     }
+
+    fn equal_length_strs_not_similar(interned_name: &str, existing_name: &str) -> bool {
+        let mut interned_chars = interned_name.chars();
+        let mut existing_chars = existing_name.chars();
+        let first_i = interned_chars.next().expect("we know we have at least one char");
+        let first_e = existing_chars.next().expect("we know we have at least one char");
+        let eq_or_numeric = |(a, b): (char, char)| a == b || a.is_numeric() && b.is_numeric();
+
+        if eq_or_numeric((first_i, first_e)) {
+            let last_i = interned_chars.next_back().expect("we know we have at least two chars");
+            let last_e = existing_chars.next_back().expect("we know we have at least two chars");
+            if eq_or_numeric((last_i, last_e)) {
+                if interned_chars
+                    .zip(existing_chars)
+                    .filter(|&ie| !eq_or_numeric(ie))
+                    .count()
+                    != 1
+                {
+                    return true;
+                }
+            } else {
+                let second_last_i = interned_chars
+                    .next_back()
+                    .expect("we know we have at least three chars");
+                let second_last_e = existing_chars
+                    .next_back()
+                    .expect("we know we have at least three chars");
+                if !eq_or_numeric((second_last_i, second_last_e))
+                    || second_last_i == '_'
+                    || !interned_chars.zip(existing_chars).all(eq_or_numeric)
+                {
+                    // allowed similarity foo_x, foo_y
+                    // or too many chars differ (foo_x, boo_y) or (foox, booy)
+                    return true;
+                }
+            }
+        } else {
+            let second_i = interned_chars.next().expect("we know we have at least two chars");
+            let second_e = existing_chars.next().expect("we know we have at least two chars");
+            if !eq_or_numeric((second_i, second_e))
+                || second_i == '_'
+                || !interned_chars.zip(existing_chars).all(eq_or_numeric)
+            {
+                // allowed similarity x_foo, y_foo
+                // or too many chars differ (x_foo, y_boo) or (xfoo, yboo)
+                return true;
+            }
+        }
+
+        false
+    }
 }
 
 impl<'a, 'b> SimilarNamesLocalVisitor<'a, 'b> {

clippy_lints/src/operators/double_comparison.rs

@@ -8,7 +8,6 @@ use rustc_span::Span;
 
 use super::DOUBLE_COMPARISONS;
 
-#[expect(clippy::similar_names)]
 pub(super) fn check<'tcx>(cx: &LateContext<'tcx>, op: BinOpKind, lhs: &'tcx Expr<'_>, rhs: &'tcx Expr<'_>, span: Span) {
     let (lkind, llhs, lrhs, rkind, rlhs, rrhs) = match (&lhs.kind, &rhs.kind) {
         (ExprKind::Binary(lb, llhs, lrhs), ExprKind::Binary(rb, rlhs, rrhs)) => {

clippy_lints/src/operators/op_ref.rs

@@ -11,7 +11,7 @@ use rustc_middle::ty::{self, Ty};
 
 use super::OP_REF;
 
-#[expect(clippy::similar_names, clippy::too_many_lines)]
+#[expect(clippy::too_many_lines)]
 pub(crate) fn check<'tcx>(
     cx: &LateContext<'tcx>,
     e: &'tcx Expr<'_>,

clippy_utils/src/ast_utils.rs

@@ -2,7 +2,7 @@
 //!
 //! - The `eq_foobar` functions test for semantic equality but ignores `NodeId`s and `Span`s.
 
-#![allow(clippy::similar_names, clippy::wildcard_imports, clippy::enum_glob_use)]
+#![allow(clippy::wildcard_imports, clippy::enum_glob_use)]
 
 use crate::{both, over};
 use rustc_ast::ptr::P;
@@ -296,7 +296,7 @@ pub fn eq_item<K>(l: &Item<K>, r: &Item<K>, mut eq_kind: impl FnMut(&K, &K) -> b
     eq_id(l.ident, r.ident) && over(&l.attrs, &r.attrs, eq_attr) && eq_vis(&l.vis, &r.vis) && eq_kind(&l.kind, &r.kind)
 }
 
-#[expect(clippy::too_many_lines)] // Just a big match statement
+#[expect(clippy::similar_names, clippy::too_many_lines)] // Just a big match statement
 pub fn eq_item_kind(l: &ItemKind, r: &ItemKind) -> bool {
     use ItemKind::*;
     match (l, r) {

clippy_utils/src/hir_utils.rs

@@ -132,7 +132,6 @@ impl HirEqInterExpr<'_, '_, '_> {
     }
 
     /// Checks whether two blocks are the same.
-    #[expect(clippy::similar_names)]
     fn eq_block(&mut self, left: &Block<'_>, right: &Block<'_>) -> bool {
         use TokenKind::{Semi, Whitespace};
         if left.stmts.len() != right.stmts.len() {
@@ -247,7 +246,7 @@ impl HirEqInterExpr<'_, '_, '_> {
         res
     }
 
-    #[expect(clippy::similar_names, clippy::too_many_lines)]
+    #[expect(clippy::too_many_lines)]
     pub fn eq_expr(&mut self, left: &Expr<'_>, right: &Expr<'_>) -> bool {
         if !self.check_ctxt(left.span.ctxt(), right.span.ctxt()) {
             return false;
@@ -463,7 +462,6 @@ impl HirEqInterExpr<'_, '_, '_> {
         }
     }
 
-    #[expect(clippy::similar_names)]
     fn eq_qpath(&mut self, left: &QPath<'_>, right: &QPath<'_>) -> bool {
         match (left, right) {
             (&QPath::Resolved(ref lty, lpath), &QPath::Resolved(ref rty, rpath)) => {
@@ -1158,7 +1156,6 @@ pub fn hash_expr(cx: &LateContext<'_>, e: &Expr<'_>) -> u64 {
     h.finish()
 }
 
-#[expect(clippy::similar_names)]
 fn eq_span_tokens(
     cx: &LateContext<'_>,
     left: impl SpanRange,

clippy_utils/src/qualify_min_const_fn.rs

@@ -389,7 +389,6 @@ fn is_const_fn(tcx: TyCtxt<'_>, def_id: DefId, msrv: &Msrv) -> bool {
         })
 }
 
-#[expect(clippy::similar_names)] // bit too pedantic
 fn is_ty_const_destruct<'tcx>(tcx: TyCtxt<'tcx>, ty: Ty<'tcx>, body: &Body<'tcx>) -> bool {
     // FIXME(effects, fee1-dead) revert to const destruct once it works again
     #[expect(unused)]

tests/ui/approx_const.rs

@@ -1,5 +1,4 @@
 #[warn(clippy::approx_constant)]
-#[allow(clippy::similar_names)]
 fn main() {
     let my_e = 2.7182;
     //~^ ERROR: approximate value of `f{32, 64}::consts::E` found