float_cmp: Allow named constants by path.

Author: Jarcho

clippy_config/src/conf.rs

@@ -657,7 +657,7 @@ define_Conf! {
     ///     x == VALUE
     /// }
     /// ```
-    (float_cmp_allowed_constants: Vec<String> = true),
+    (float_cmp_allowed_constants: Vec<String> = Vec::new()),
     /// Lint: FLOAT_CMP
     ///
     /// Whether to ignore comparisons which have a constant result.

clippy_lints/src/lib.rs

@@ -816,7 +816,7 @@ pub fn register_lints(store: &mut rustc_lint::LintStore, conf: &'static Conf) {
     store.register_late_pass(move |_| Box::new(manual_rem_euclid::ManualRemEuclid::new(conf)));
     store.register_late_pass(move |_| Box::new(manual_retain::ManualRetain::new(conf)));
     store.register_late_pass(move |_| Box::new(manual_rotate::ManualRotate));
-    store.register_late_pass(move |_| Box::new(operators::Operators::new(conf)));
+    store.register_late_pass(move |tcx| Box::new(operators::Operators::new(tcx, conf)));
     store.register_late_pass(|_| Box::<std_instead_of_core::StdReexports>::default());
     store.register_late_pass(move |_| Box::new(instant_subtraction::InstantSubtraction::new(conf)));
     store.register_late_pass(|_| Box::new(partialeq_to_none::PartialeqToNone));

clippy_lints/src/operators/float_cmp.rs

@@ -2,7 +2,7 @@ use clippy_utils::consts::{constant, Constant};
 use clippy_utils::diagnostics::span_lint_and_then;
 use clippy_utils::sugg::Sugg;
 use clippy_utils::visitors::{for_each_expr_without_closures, is_const_evaluatable};
-use clippy_utils::{get_item_name, is_expr_named_const, path_res, peel_hir_expr_while, SpanlessEq};
+use clippy_utils::{get_item_name, get_named_const_def_id, path_res, peel_hir_expr_while, SpanlessEq};
 use core::ops::ControlFlow;
 use rustc_errors::Applicability;
 use rustc_hir::def::Res;
@@ -14,7 +14,7 @@ use super::{FloatCmpConfig, FLOAT_CMP};
 
 pub(crate) fn check<'tcx>(
     cx: &LateContext<'tcx>,
-    config: FloatCmpConfig,
+    config: &FloatCmpConfig,
     expr: &'tcx Expr<'_>,
     op: BinOpKind,
     left: &'tcx Expr<'_>,
@@ -56,8 +56,8 @@ pub(crate) fn check<'tcx>(
             return;
         }
 
-        if config.ignore_named_constants
-            && (is_expr_named_const(cx, left_reduced) || is_expr_named_const(cx, right_reduced))
+        if get_named_const_def_id(cx, left_reduced).is_some_and(|id| config.allowed_constants.contains(&id))
+            || get_named_const_def_id(cx, right_reduced).is_some_and(|id| config.allowed_constants.contains(&id))
         {
             return;
         }

clippy_lints/src/operators/mod.rs

@@ -28,6 +28,7 @@ use clippy_utils::def_path_def_ids;
 use rustc_hir::def_id::DefIdSet;
 use rustc_hir::{Body, Expr, ExprKind, UnOp};
 use rustc_lint::{LateContext, LateLintPass};
+use rustc_middle::ty::TyCtxt;
 use rustc_session::impl_lint_pass;
 
 declare_clippy_lint! {
@@ -789,7 +790,7 @@ pub struct Operators {
     float_cmp_config: FloatCmpConfig,
 }
 impl Operators {
-    pub fn new(conf: &'static Conf) -> Self {
+    pub fn new(tcx: TyCtxt<'_>, conf: &'static Conf) -> Self {
         Self {
             arithmetic_context: numeric_arithmetic::Context::default(),
             verbose_bit_mask_threshold: conf.verbose_bit_mask_threshold,
@@ -859,7 +860,7 @@ impl<'tcx> LateLintPass<'tcx> for Operators {
                 float_equality_without_abs::check(cx, e, op.node, lhs, rhs);
                 integer_division::check(cx, e, op.node, lhs, rhs);
                 cmp_owned::check(cx, op.node, lhs, rhs);
-                float_cmp::check(cx, self.float_cmp_config, e, op.node, lhs, rhs);
+                float_cmp::check(cx, &self.float_cmp_config, e, op.node, lhs, rhs);
                 modulo_one::check(cx, e, op.node, rhs);
                 modulo_arithmetic::check(
                     cx,

clippy_utils/src/lib.rs

@@ -245,14 +245,15 @@ pub fn is_inside_always_const_context(tcx: TyCtxt<'_>, hir_id: HirId) -> bool {
     }
 }
 
-/// Checks if the expression is path to either a constant or an associated constant.
-pub fn is_expr_named_const<'tcx>(cx: &LateContext<'tcx>, e: &'tcx Expr<'_>) -> bool {
-    matches!(&e.kind, ExprKind::Path(p)
-        if matches!(
-            cx.qpath_res(p, e.hir_id),
-            Res::Def(DefKind::Const | DefKind::AssocConst, _)
-        )
-    )
+/// If the expression is path to either a constant or an associated constant get the `DefId`.
+pub fn get_named_const_def_id<'tcx>(cx: &LateContext<'tcx>, e: &'tcx Expr<'_>) -> Option<DefId> {
+    if let ExprKind::Path(p) = &e.kind
+        && let Res::Def(DefKind::Const | DefKind::AssocConst, id) = cx.qpath_res(p, e.hir_id)
+    {
+        Some(id)
+    } else {
+        None
+    }
 }
 
 /// Checks if a `Res` refers to a constructor of a `LangItem`

tests/ui-toml/float_cmp/change_detect/clippy.toml

@@ -1 +1,6 @@
 float-cmp-ignore-change-detection = false
+float-cmp-allowed-constants = [
+    "core::f32::EPSILON",
+    "f32::EPSILON",
+    "test::F32_ARRAY",
+]

tests/ui-toml/float_cmp/const_cmp/clippy.toml

@@ -1 +1,6 @@
 float-cmp-ignore-constant-comparisons = false
+float-cmp-allowed-constants = [
+    "core::f32::EPSILON",
+    "f32::EPSILON",
+    "test::F32_ARRAY",
+]

tests/ui-toml/float_cmp/named_const/clippy.toml

@@ -1 +0,0 @@
-float-cmp-ignore-named-constants = false

tests/ui-toml/float_cmp/test.change_detect.stderr

@@ -1,5 +1,5 @@
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:15:21
+  --> tests/ui-toml/float_cmp/test.rs:17:21
    |
 LL |             let _ = x == y;
    |                     ^^^^^^ help: consider comparing them within some margin of error: `(x - y).abs() < error_margin`
@@ -11,130 +11,130 @@ LL | #![deny(clippy::float_cmp)]
    |         ^^^^^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:16:21
+  --> tests/ui-toml/float_cmp/test.rs:18:21
    |
 LL |             let _ = x != y;
    |                     ^^^^^^ help: consider comparing them within some margin of error: `(x - y).abs() > error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:17:21
+  --> tests/ui-toml/float_cmp/test.rs:19:21
    |
 LL |             let _ = x == 5.5;
    |                     ^^^^^^^^ help: consider comparing them within some margin of error: `(x - 5.5).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:18:21
+  --> tests/ui-toml/float_cmp/test.rs:20:21
    |
 LL |             let _ = 5.5 == x;
    |                     ^^^^^^^^ help: consider comparing them within some margin of error: `(5.5 - x).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:41:21
+  --> tests/ui-toml/float_cmp/test.rs:43:21
    |
 LL |             let _ = x == y;
    |                     ^^^^^^ help: consider comparing them within some margin of error: `(x - y).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:42:21
+  --> tests/ui-toml/float_cmp/test.rs:44:21
    |
 LL |             let _ = x != y;
    |                     ^^^^^^ help: consider comparing them within some margin of error: `(x - y).abs() > error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:43:21
+  --> tests/ui-toml/float_cmp/test.rs:45:21
    |
 LL |             let _ = x == 5.5;
    |                     ^^^^^^^^ help: consider comparing them within some margin of error: `(x - 5.5).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:44:21
+  --> tests/ui-toml/float_cmp/test.rs:46:21
    |
 LL |             let _ = 5.5 == x;
    |                     ^^^^^^^^ help: consider comparing them within some margin of error: `(5.5 - x).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:67:21
+  --> tests/ui-toml/float_cmp/test.rs:69:21
    |
 LL |             let _ = x == y;
    |                     ^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:68:21
+  --> tests/ui-toml/float_cmp/test.rs:70:21
    |
 LL |             let _ = x == [5.5; 4];
    |                     ^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:69:21
+  --> tests/ui-toml/float_cmp/test.rs:71:21
    |
 LL |             let _ = [5.5; 4] == x;
    |                     ^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:70:21
+  --> tests/ui-toml/float_cmp/test.rs:72:21
    |
 LL |             let _ = [0.0, 0.0, 0.0, 5.5] == x;
    |                     ^^^^^^^^^^^^^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:71:21
+  --> tests/ui-toml/float_cmp/test.rs:73:21
    |
 LL |             let _ = x == [0.0, 0.0, 0.0, 5.5];
    |                     ^^^^^^^^^^^^^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:87:21
+  --> tests/ui-toml/float_cmp/test.rs:89:21
    |
 LL |             let _ = x == y;
    |                     ^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:88:21
+  --> tests/ui-toml/float_cmp/test.rs:90:21
    |
 LL |             let _ = x == [5.5; 4];
    |                     ^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:89:21
+  --> tests/ui-toml/float_cmp/test.rs:91:21
    |
 LL |             let _ = [5.5; 4] == x;
    |                     ^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:90:21
+  --> tests/ui-toml/float_cmp/test.rs:92:21
    |
 LL |             let _ = [0.0, 0.0, 0.0, 5.5] == x;
    |                     ^^^^^^^^^^^^^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:91:21
+  --> tests/ui-toml/float_cmp/test.rs:93:21
    |
 LL |             let _ = x == [0.0, 0.0, 0.0, 5.5];
    |                     ^^^^^^^^^^^^^^^^^^^^^^^^^
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:109:21
+  --> tests/ui-toml/float_cmp/test.rs:111:21
    |
 LL |             let _ = x == y;
    |                     ^^^^^^ help: consider comparing them within some margin of error: `(x - y).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64` arrays
-  --> tests/ui-toml/float_cmp/test.rs:115:21
+  --> tests/ui-toml/float_cmp/test.rs:117:21
    |
 LL |             let _ = x == y;
    |                     ^^^^^^
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:131:21
+  --> tests/ui-toml/float_cmp/test.rs:132:21
    |
-LL |             let _ = C * x == x * x;
-   |                     ^^^^^^^^^^^^^^ help: consider comparing them within some margin of error: `(C * x - x * x).abs() < error_margin`
+LL |             let _ = f32::EPSILON * x == x * x;
+   |                     ^^^^^^^^^^^^^^^^^^^^^^^^^ help: consider comparing them within some margin of error: `(f32::EPSILON * x - x * x).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64`
-  --> tests/ui-toml/float_cmp/test.rs:132:21
+  --> tests/ui-toml/float_cmp/test.rs:133:21
    |
-LL |             let _ = x * x == C * x;
-   |                     ^^^^^^^^^^^^^^ help: consider comparing them within some margin of error: `(x * x - C * x).abs() < error_margin`
+LL |             let _ = x * x == f32::EPSILON * x;
+   |                     ^^^^^^^^^^^^^^^^^^^^^^^^^ help: consider comparing them within some margin of error: `(x * x - f32::EPSILON * x).abs() < error_margin`
 
 error: strict comparison of `f32` or `f64`
   --> tests/ui-toml/float_cmp/test.rs:158:17